{
  "name": "crisis-monitor-web-kb",
  "version": "1.0.0",
  "generated": "2026-05-08",
  "sourceEntries": 164,
  "stats": {
    "totalEntries": 164,
    "totalCategories": 11,
    "totalSubcategories": 16,
    "totalTags": 869
  },
  "triggerCategories": [],
  "sections": [
    {
      "id": "frameworks",
      "title": "FRAMEWORKS & THEORY",
      "subtitle": "The theoretical foundations of crisis prediction — from Minsky's instability hypothesis to network contagion models.",
      "entryIds": [
        0,
        1,
        2,
        3,
        4,
        115,
        116,
        127,
        128,
        140,
        141,
        142,
        143,
        154,
        155,
        156,
        157,
        158,
        159,
        160,
        161,
        162,
        163
      ],
      "keyStats": [
        {
          "stat": "75-100yr",
          "label": "Dalio's long debt cycle length"
        },
        {
          "stat": "3",
          "label": "Minsky financing regimes: hedge, speculative, Ponzi"
        },
        {
          "stat": "Robust yet fragile",
          "label": "Network contagion paradox"
        }
      ]
    },
    {
      "id": "composite-indices",
      "title": "COMPOSITE STRESS INDICES",
      "subtitle": "Real-time and periodic composite measures of financial system stress from major institutions.",
      "entryIds": [
        5,
        6,
        7,
        8,
        9,
        10,
        52
      ],
      "keyStats": [
        {
          "stat": "33",
          "label": "Variables in OFR Financial Stress Index"
        },
        {
          "stat": "Daily",
          "label": "OFR FSI and Cleveland SRI update frequency"
        },
        {
          "stat": "5",
          "label": "Severity levels in Cleveland Fed CFSI"
        }
      ]
    },
    {
      "id": "leading-indicators",
      "title": "LEADING INDICATORS",
      "subtitle": "Individual metrics with documented track records for predicting financial crises and recessions.",
      "entryIds": [
        11,
        12,
        13,
        14,
        15,
        16,
        17,
        48,
        49,
        50,
        51,
        53,
        54,
        55,
        56,
        57,
        58,
        59,
        60,
        61,
        62,
        66,
        67,
        78,
        79,
        80,
        81,
        82,
        83,
        84,
        85,
        86,
        87,
        88,
        89,
        90,
        91,
        92,
        121,
        124,
        125,
        126,
        131,
        132,
        133,
        134,
        144,
        145,
        146,
        147,
        148,
        149,
        150,
        152
      ],
      "keyStats": [
        {
          "stat": "10mo-3yr",
          "label": "Yield curve inversion lead time"
        },
        {
          "stat": "2-5yr",
          "label": "Credit-to-GDP gap lead time"
        },
        {
          "stat": "2-3yr",
          "label": "Mispricing risk indicator lead time"
        }
      ]
    },
    {
      "id": "trigger-analysis",
      "title": "CRISIS TRIGGERS",
      "subtitle": "The 8 major trigger categories that can initiate or amplify economic crises.",
      "entryIds": [
        18,
        19,
        20,
        21,
        22,
        23,
        24,
        25,
        26,
        68,
        69,
        70,
        71,
        72,
        73,
        74,
        75,
        76,
        77,
        99,
        122,
        129,
        137,
        138
      ],
      "keyStats": [
        {
          "stat": "8",
          "label": "Major trigger categories tracked"
        },
        {
          "stat": "14-20x",
          "label": "AI capex-to-revenue gap"
        },
        {
          "stat": "$700B",
          "label": "AI infrastructure investment"
        }
      ]
    },
    {
      "id": "historical-precedents",
      "title": "HISTORICAL PRECEDENTS",
      "subtitle": "Past crises analyzed for patterns, triggers, and recovery dynamics.",
      "entryIds": [
        27,
        28,
        29,
        30,
        31,
        93,
        94,
        95,
        96,
        97,
        98,
        101,
        102,
        103,
        104,
        105,
        106,
        107,
        108,
        109,
        110,
        111,
        112,
        113,
        114,
        117,
        118,
        119,
        120
      ],
      "keyStats": [
        {
          "stat": "-78%",
          "label": "NASDAQ decline in dot-com crash"
        },
        {
          "stat": "4+ years",
          "label": "Typical recovery to pre-crisis per-capita income"
        },
        {
          "stat": "$107 to $7 to $3000+",
          "label": "Amazon through the dot-com crash"
        }
      ]
    },
    {
      "id": "compound-interactions",
      "title": "COMPOUND INTERACTIONS",
      "subtitle": "How multiple triggers combine to create systemic crises — the most dangerous scenarios.",
      "entryIds": [
        32,
        33,
        34,
        35,
        36,
        45,
        46,
        47,
        63,
        64,
        65,
        100,
        123,
        130,
        135,
        136,
        139,
        151,
        153
      ],
      "keyStats": [
        {
          "stat": "Multiplicative",
          "label": "Combined trigger effects are not additive"
        },
        {
          "stat": "85%",
          "label": "Correlations spike during stress"
        },
        {
          "stat": "ECB insight",
          "label": "Multi-segment stress is qualitatively different"
        }
      ]
    },
    {
      "id": "institutional-approaches",
      "title": "INSTITUTIONAL INVESTOR APPROACHES",
      "subtitle": "How Bridgewater, Universa, AQR, and GMO systematically monitor and position for crisis risk.",
      "entryIds": [
        37,
        38,
        39,
        40
      ],
      "keyStats": [
        {
          "stat": "+26.2%",
          "label": "Bridgewater Pure Alpha YTD Sep 2025"
        },
        {
          "stat": "300+",
          "label": "Decision rules in Bridgewater's systematic approach"
        },
        {
          "stat": "20%",
          "label": "Universa's tail hedge allocation when VIX inverts"
        }
      ]
    },
    {
      "id": "emerging-methods",
      "title": "EMERGING METHODS",
      "subtitle": "AI, machine learning, and alternative data approaches to crisis prediction.",
      "entryIds": [
        41,
        42,
        43,
        44
      ],
      "keyStats": [
        {
          "stat": "1-3mo",
          "label": "High-frequency nowcasting lead on official projections"
        },
        {
          "stat": "90%+",
          "label": "EWS papers published after 2006"
        },
        {
          "stat": "Moral hazard",
          "label": "Key risk of ML-based prediction systems"
        }
      ]
    }
  ],
  "entries": [
    {
      "id": 0,
      "sourceId": "cm-001",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "credit-debt",
      "title": "Minsky's Financial Instability Hypothesis",
      "content": "Hyman Minsky's Financial Instability Hypothesis (FIH) is the foundational framework for understanding how financial crises emerge endogenously from periods of stability. The core insight is counterintuitive: stability itself breeds instability. During prolonged economic calm, borrowers, lenders, and regulators progressively take on more risk because recent experience tells them risk is low. This is not irrational behavior — it is a natural, predictable response to a stable environment.\n\nMinsky categorized financing positions into three regimes. Hedge financing: income covers both principal and interest payments. Speculative financing: income covers interest but not principal, requiring debt rollover. Ponzi financing: income covers neither — the borrower depends on asset appreciation to refinance or sell. The key dynamic is that economies naturally migrate from hedge-dominant to Ponzi-dominant during expansions. Each successful year of returns validates more aggressive positioning.\n\nThe 'Minsky Moment' occurs when Ponzi-dependent units can no longer refinance. Asset prices that were sustained by leverage-fueled buying suddenly face forced selling. The resulting cascade is self-reinforcing: falling prices trigger margin calls, which force more selling, which drives prices lower. This is why the transition from boom to crisis is always faster than the buildup — leverage amplifies on the way down just as it did on the way up.\n\nMinsky's framework is particularly relevant for monitoring the AI capex cycle. When companies borrow heavily to build AI infrastructure based on projected future revenues that may not materialize, they are engaging in speculative or Ponzi financing. The $700B+ annual AI capex against $35-50B in actual AI revenue represents a classic Minsky dynamic: the gap between investment and income is being bridged by expectations of future growth, not current cash flows.",
      "dataSource": "Academic literature, BIS working papers, Levy Economics Institute archives",
      "leadTime": "Framework-level — identifies regime positioning rather than timing specific events",
      "reliability": "High conceptual reliability. Validated across multiple crisis cycles (2000 dot-com, 2008 GFC, 1997 Asian crisis). Limitation: does not predict timing — only identifies when conditions are ripe for a Minsky Moment.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis",
        "Dot-Com Bubble 2000-2002",
        "1997 Asian Financial Crisis",
        "1929 Great Depression",
        "Savings & Loan Crisis 1980s"
      ],
      "compoundsWith": [
        "asset-bubble",
        "credit-debt",
        "systemic-banking",
        "ai-specific"
      ],
      "tags": [
        "minsky",
        "financial-instability",
        "ponzi-finance",
        "endogenous-crisis",
        "leverage-cycle",
        "foundational-framework"
      ]
    },
    {
      "id": 1,
      "sourceId": "cm-002",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "credit-debt",
      "title": "Dalio's Debt Cycle Framework",
      "content": "Ray Dalio's debt cycle framework, developed through decades of research at Bridgewater Associates, identifies two overlapping cycles that govern economic dynamics. The short-term debt cycle runs 5-8 years, driven by central bank interest rate adjustments and credit expansion/contraction. The long-term debt cycle runs 75-100 years, encompassing the full arc from low debt levels through maximum leverage to eventual deleveraging. We are currently deep into a long-term debt cycle that arguably began after World War II.\n\nDalio distinguishes between 'beautiful' and 'ugly' deleveraging. A beautiful deleveraging balances four levers: austerity (spending cuts), debt restructuring/defaults, money printing/monetization, and wealth transfers (taxation). When these are calibrated correctly, the economy deleverages without depression. An ugly deleveraging overweights one lever — typically either deflationary austerity (1930s) or inflationary money printing (Weimar Germany, Zimbabwe). The skill of policymakers in balancing these levers determines whether deleveraging is manageable or catastrophic.\n\nA critical insight from Dalio's framework is that each short-term debt cycle peaks at a higher debt level than the last. Central banks lower rates to end each recession, which enables more borrowing in the next expansion. Over decades, this ratchet effect pushes total debt-to-GDP to levels where further monetary stimulus becomes ineffective — the 'pushing on a string' problem. When interest rates hit zero (or go negative), the primary tool for managing short-term cycles is exhausted.\n\nFor crisis monitoring, Dalio's framework provides specific metrics to track: total debt-to-GDP ratios, the gap between debt service costs and income growth, central bank capacity (how much room exists to cut rates), and the distribution of debt across hedge/speculative/Ponzi categories. The framework also emphasizes that the biggest risk is not the debt level itself but the rate of change — rapid credit expansion followed by sudden contraction is more dangerous than high but stable debt levels.",
      "dataSource": "Bridgewater Associates research, Dalio's 'Principles for Navigating Big Debt Crises' (2018), 'The Changing World Order' (2021)",
      "leadTime": "Short cycle: 1-3 years for identifying late-cycle positioning. Long cycle: decades-level framework for understanding structural vulnerability.",
      "reliability": "High. Bridgewater's systematic application of this framework contributed to navigating 2008 successfully (Pure Alpha gained while markets crashed). The framework has been backtested across 48 debt crises spanning 500 years of economic history.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis",
        "1930s Great Depression",
        "1970s Stagflation",
        "Japan Lost Decades 1990-present",
        "European Sovereign Debt Crisis 2010-2012"
      ],
      "compoundsWith": [
        "credit-debt",
        "currency-inflation",
        "yield-curve",
        "systemic-banking"
      ],
      "tags": [
        "dalio",
        "debt-cycle",
        "deleveraging",
        "bridgewater",
        "long-term-debt-cycle",
        "short-term-debt-cycle",
        "monetary-policy"
      ]
    },
    {
      "id": 2,
      "sourceId": "cm-003",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "credit-debt",
      "title": "Reinhart-Rogoff Great Credit Contraction",
      "content": "Carmen Reinhart and Kenneth Rogoff's research, synthesized in 'This Time Is Different' (2009) and subsequent papers, reframed how we understand financial crises. Their central argument is that financial crises should be understood as credit contractions rather than recessions. The distinction matters: a recession is a decline in output that typically recovers within 1-2 years. A credit contraction involves the destruction of the credit system itself — borrowers defaulting, lenders retrenching, collateral values collapsing — and recovery takes 4+ years for per capita income to return to pre-crisis levels.\n\nTheir dataset spans 800 years of financial crises across 66 countries, making it the most comprehensive empirical study of crisis dynamics ever assembled. Key findings include: banking crises are followed by an average 35% decline in real housing prices over 6 years, equity prices fall 56% over 3.5 years, unemployment rises an average of 7 percentage points over 4 years, and government debt rises an average of 86% in the three years following the crisis onset. These are not worst-case scenarios — they are averages.\n\nThe 'debt overhang' concept is central to Reinhart-Rogoff. When public debt exceeds 90% of GDP, median growth rates fall by about 1 percentage point and average growth falls considerably more. While the specific 90% threshold was debated (the Herndon-Ash-Pollin critique identified spreadsheet errors), the broader finding — that very high debt levels are associated with lower growth — has held up across multiple studies. The mechanism is straightforward: high debt requires more income to service, leaving less for investment and consumption.\n\nFor crisis monitoring, Reinhart-Rogoff provide two critical insights. First, the phrase 'this time is different' is the most dangerous sentence in finance — every credit boom features sophisticated arguments for why historical patterns don't apply (new technology, new financial instruments, new regulatory framework). Second, the best predictor of a financial crisis is the rate of credit expansion relative to GDP growth. When credit grows significantly faster than the economy for sustained periods, a crisis becomes not a question of if but when.",
      "dataSource": "Reinhart & Rogoff 'This Time Is Different' (2009), 'The Liquidation of Government Debt' (2015), NBER working papers, IMF Global Financial Stability Reports",
      "leadTime": "2-5 years using credit-to-GDP deviation as primary signal. Post-crisis, provides timeline expectations for recovery duration.",
      "reliability": "Strong empirical basis across 800 years of data. Correctly predicted the extended recovery timeline after 2008. Criticism: the 90% debt threshold is less precise than originally claimed, and historical analogies can miss novel crisis mechanisms.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis",
        "1930s Great Depression",
        "1997 Asian Financial Crisis",
        "Latin American Debt Crises 1980s",
        "European Sovereign Debt Crisis 2010-2012",
        "Japanese Lost Decades"
      ],
      "compoundsWith": [
        "credit-debt",
        "systemic-banking",
        "currency-inflation"
      ],
      "tags": [
        "reinhart-rogoff",
        "credit-contraction",
        "debt-overhang",
        "this-time-is-different",
        "empirical-crisis-history",
        "recovery-timeline"
      ]
    },
    {
      "id": 3,
      "sourceId": "cm-004",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "systemic-banking",
      "title": "Network Contagion Models",
      "content": "Network contagion models represent a paradigm shift in crisis analysis, moving from the study of individual institutions to the study of interconnections between them. The financial system is modeled as a network where nodes are institutions (banks, funds, insurers, clearinghouses) and edges are financial obligations (loans, derivatives contracts, repo agreements, common asset holdings). The key insight is that the structure of this network determines how shocks propagate — a shock to one node can cascade through the network in ways that are disproportionate to the initial trigger.\n\nThe most important finding from network theory applied to finance is the 'robust-yet-fragile' property. Highly interconnected financial networks are robust to random shocks — the failure of an arbitrary institution is absorbed by the network. However, they are extremely fragile to targeted shocks — the failure of a highly connected hub institution can cascade through the entire system. This property explains why regulators focus on 'systemically important financial institutions' (SIFIs): it is not their size per se that matters, but their centrality in the network of financial obligations.\n\nThe OFR (Office of Financial Research) Contagion Index operationalizes this framework. It measures the fraction of the banking system's total assets that would be wiped out by the worst-case contagion scenario, considering both direct exposures (counterparty default) and indirect exposures (common asset holdings, fire sale spillovers). The formula accounts for three channels: bilateral credit exposures between institutions, shared portfolio overlaps that create correlated fire sale pressure, and funding market interconnections where one institution's distress withdraws liquidity from others.\n\nNetwork models have revealed that common asset shocks — where multiple institutions hold similar portfolios — can be more dangerous than direct counterparty defaults. When Institution A sells an asset under stress, it drives down the price for Institutions B, C, and D who hold the same asset, potentially triggering their own forced sales. This 'fire sale externality' is invisible in traditional institution-by-institution analysis and only becomes apparent when viewing the system as a network. The 2008 crisis demonstrated this vividly: the correlation of mortgage-backed securities across bank balance sheets created a system-wide vulnerability that no single institution's risk management could detect.",
      "dataSource": "OFR Contagion Index (quarterly), BIS research papers, Federal Reserve network analysis, academic literature (Acemoglu et al., Allen & Gale, Battiston et al.)",
      "leadTime": "Structural vulnerability assessment rather than timing prediction. Network fragility can persist for years before a trigger event activates the contagion channel.",
      "reliability": "Strong theoretical foundation validated by 2008 GFC analysis. Limitations: network data is incomplete (OTC derivatives are poorly mapped), models require assumptions about fire sale dynamics, and novel connection types (e.g., crypto-TradFi bridges) may not be captured.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (Lehman cascade)",
        "1998 LTCM (hedge fund interconnection)",
        "2023 SVB/Credit Suisse (bank run contagion)",
        "2020 March COVID liquidity crisis"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "carry-trade",
        "asset-bubble"
      ],
      "tags": [
        "network-theory",
        "contagion",
        "robust-yet-fragile",
        "ofr-contagion-index",
        "systemic-risk",
        "fire-sale-externality",
        "common-asset-shocks"
      ]
    },
    {
      "id": 4,
      "sourceId": "cm-005",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "general",
      "title": "Input-Output Cascade Models",
      "content": "Input-output cascade models, rooted in Wassily Leontief's Nobel Prize-winning work, analyze how economic shocks propagate through production networks. Unlike financial contagion models that track money and credit flows, input-output models track the physical flow of goods and services between industries. When Industry A supplies a critical input to Industries B, C, and D, a disruption to A creates downstream constraints that cascade through the entire production chain.\n\nThe key insight for crisis monitoring is that supply-side shocks transmit differently than demand-side shocks. A demand shock (consumers stop buying) is dispersed across many suppliers and attenuates as it moves upstream. A supply shock (a critical input becomes unavailable) concentrates as it moves downstream — every industry that depends on the disrupted input is simultaneously constrained. This asymmetry means that supply-side crises can be far more severe than demand-side crises of equivalent initial magnitude.\n\nModern input-output analysis has been enhanced by granular supply chain data and computational power. Researchers can now model cascades at the firm level rather than the industry level, revealing 'granular origins' of aggregate fluctuations — where a shock to a single large firm (e.g., TSMC, a major oil producer, a critical rare earth mine) can create measurable GDP-level effects. This finding, documented by Acemoglu, Carvalho, Ozdaglar, and Tahbaz-Salehi, overturns the traditional macroeconomic assumption that firm-level shocks average out in aggregate.\n\nFor crisis monitoring, input-output models identify which industries and firms are 'systemically important' from a production perspective — analogous to SIFIs in finance. Critical chokepoints include semiconductor fabrication (concentrated in Taiwan and South Korea), rare earth processing (dominated by China), energy infrastructure, and increasingly, cloud computing capacity. A geopolitical or natural disaster event affecting these chokepoints would trigger cascading production failures across multiple industries simultaneously, with each downstream industry amplifying the original shock rather than absorbing it.",
      "dataSource": "BEA Input-Output Accounts, OECD Inter-Country Input-Output tables, World Input-Output Database (WIOD), academic literature (Acemoglu et al. 2012)",
      "leadTime": "Not a timing predictor — identifies structural vulnerabilities and amplification channels. Combined with geopolitical monitoring, can identify which supply chains are most at risk.",
      "reliability": "Strong empirical validation during COVID-19 supply chain disruptions (2020-2022) and semiconductor shortage. Models correctly predicted which industries would face cascading constraints. Limitation: real-time supply chain data is sparse and often proprietary.",
      "historicalPrecedents": [
        "COVID-19 Supply Chain Crisis 2020-2022",
        "2011 Tohoku Earthquake (Japanese auto supply chain)",
        "2021 Suez Canal Blockage",
        "2022 Semiconductor Shortage"
      ],
      "compoundsWith": [
        "geopolitical",
        "energy-shock",
        "ai-specific",
        "general"
      ],
      "tags": [
        "input-output",
        "supply-chain",
        "production-networks",
        "cascade",
        "chokepoints",
        "leontief",
        "granular-origins"
      ]
    },
    {
      "id": 5,
      "sourceId": "cm-006",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "general",
      "title": "OFR Financial Stress Index (OFR FSI)",
      "content": "The Office of Financial Research Financial Stress Index is among the most comprehensive real-time measures of financial system stress available. Updated daily, it synthesizes 33 financial market variables across five categories: credit (corporate bond spreads, municipal spreads, ABS spreads), equity valuation (S&P 500 price-earnings, market volatility), funding (TED spread, commercial paper-Treasury spread, cross-currency basis), safe assets (on-the-run/off-the-run Treasury spread, Treasury bid-ask), and volatility (VIX, MOVE index, swaption volatility).\n\nThe OFR FSI is constructed using a dynamic factor model that extracts a common stress signal from the 33 underlying variables. Each variable is normalized relative to its own history, so the index captures deviations from normal conditions rather than absolute levels. A reading of zero represents median stress conditions. Positive values indicate above-average stress, with readings above 5 historically associated with significant market dislocations. The index reached approximately 25 during the peak of the 2008 crisis and spiked above 15 during the March 2020 COVID liquidity crisis.\n\nWhat makes the OFR FSI particularly valuable for crisis monitoring is its daily frequency and its decomposition capability. Analysts can see which of the five categories is contributing most to overall stress at any given moment. This decomposition reveals the nature of the stress: credit-driven stress has different implications than funding-driven stress. Credit stress suggests concerns about borrower solvency (slow-moving, typically preceded by deteriorating fundamentals). Funding stress suggests concerns about liquidity and counterparty risk (fast-moving, can cascade quickly through the interbank market). The pattern of category contributions often reveals the crisis mechanism before it is widely recognized.\n\nThe OFR FSI also provides regional breakdowns (US, Europe, other advanced economies, emerging markets), enabling monitoring of contagion patterns. A crisis that begins with elevated stress in one region and then spreads to others follows a different trajectory than synchronized global stress. The 1997 Asian crisis showed a clear contagion pattern from EM to developed markets, while the 2008 GFC was more synchronized, and the March 2026 Iran tensions showed differentiated EM vs US responses.",
      "dataSource": "Office of Financial Research (US Treasury), updated daily, available at https://www.financialresearch.gov/financial-stress-index/",
      "leadTime": "Contemporaneous to very short-term. Primarily a real-time stress measure rather than a leading indicator. However, sustained elevated readings (weeks above 3-4) often precede more severe market dislocations.",
      "reliability": "High reliability as a stress measure. Captured all major stress episodes since its 2000 start date. Limitation: as a daily market-based measure, it reflects financial market stress rather than real economy stress — markets can be stressed while the economy is fine, and vice versa.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (peak ~25)",
        "March 2020 COVID (peak ~15)",
        "2011 European Sovereign Debt (peak ~8)",
        "2023 SVB/Banking stress (peak ~5)",
        "March 2026 Iran tensions"
      ],
      "compoundsWith": [
        "credit-debt",
        "systemic-banking",
        "carry-trade",
        "currency-inflation",
        "asset-bubble"
      ],
      "tags": [
        "ofr",
        "financial-stress-index",
        "daily",
        "33-variables",
        "5-categories",
        "real-time-monitoring",
        "stress-decomposition"
      ]
    },
    {
      "id": 6,
      "sourceId": "cm-007",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "general",
      "title": "Cleveland Fed Composite Financial Stress Index (CFSI)",
      "content": "The Cleveland Federal Reserve's Composite Financial Stress Index tracks 11 components spanning four financial market segments: credit markets (corporate bond spreads, commercial paper-Treasury spread), equity markets (stock market returns, stock-bond correlation), interbank markets (TED spread, interbank liquidity spread), and foreign exchange markets (weighted dollar index). The CFSI is updated monthly, providing a slower but more stable signal than daily indices.\n\nThe CFSI uses a 5-level severity classification system based on historical percentile rankings. Grade I (Below Normal Stress) indicates benign financial conditions where credit is flowing freely and market functioning is normal. Grade II (Normal Stress) represents typical financial market conditions with no elevated concerns. Grade III (Moderate Stress) flags conditions that warrant attention — credit spreads are widening, volatility is above average, or interbank lending is showing friction. Grade IV (Significant Stress) indicates serious financial market strain with elevated probability of broader economic impact. Grade V (Extreme Stress) signals systemic crisis conditions, historically associated with banking panics and severe recessions.\n\nThe transition dynamics between grades are as important as the current grade itself. A rapid jump from Grade I to Grade III suggests a shock event that overwhelmed normal market resilience. A gradual drift from Grade II to Grade III to Grade IV suggests building systemic pressure that the market cannot self-correct. Historically, the most dangerous pattern is a prolonged period at Grade I (complacency, compressed risk premiums) followed by a rapid escalation to Grade IV or V. This pattern — the 'Minsky progression' — was visible in the 2006-2008 transition and the 2019-2020 COVID shock.\n\nFor crisis monitoring, the CFSI's monthly frequency makes it better suited as a trend indicator than a real-time alarm. It should be used alongside daily indices (OFR FSI, VIX) to distinguish between temporary market spikes and sustained deterioration. When the CFSI confirms a stress reading that daily indices have been showing for weeks, the signal is significantly more credible than either alone.",
      "dataSource": "Federal Reserve Bank of Cleveland, updated monthly, available through FRED database (series: CFSI)",
      "leadTime": "Contemporaneous to 1-3 months leading. Grade transitions from low to high have historically preceded economic downturns by 1-6 months.",
      "reliability": "Good track record of capturing major stress episodes. The 5-grade system provides intuitive severity assessment. Limitation: monthly frequency means it can lag fast-moving crises by 2-4 weeks.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (Grade V)",
        "2011 European Sovereign Crisis (Grade IV)",
        "2020 COVID Crisis (Grade V briefly)",
        "2023 Banking Stress (Grade III)"
      ],
      "compoundsWith": [
        "credit-debt",
        "systemic-banking",
        "currency-inflation"
      ],
      "tags": [
        "cleveland-fed",
        "cfsi",
        "monthly",
        "5-grade-severity",
        "11-components",
        "trend-indicator"
      ]
    },
    {
      "id": 7,
      "sourceId": "cm-008",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "general",
      "title": "St. Louis Fed Financial Stress Index (STLFSI4)",
      "content": "The St. Louis Federal Reserve's Financial Stress Index (STLFSI4, the fourth revision) is a weekly composite of 18 financial market series organized into three groups: 7 interest rate series (effective federal funds rate, 2-year Treasury, 10-year Treasury, 30-year Treasury, Baa corporate bond yield, Merrill Lynch high-yield bond index, 30-year fixed mortgage rate), 6 yield spread series (TED spread, corporate Baa-Treasury spread, high-yield spread, 10-year TIPS spread, 3-month Treasury-federal funds spread, 10-year minus 3-month Treasury spread), and 5 other indicators (J.P. Morgan Emerging Market Bond Index, Chicago Board Options Exchange volatility index, Merrill Lynch bond volatility index, 10-year Treasury futures implied volatility, S&P 500 financial sector index).\n\nThe STLFSI4 is constructed using principal component analysis (PCA), extracting the first principal component from the 18 series as the common factor representing financial stress. Each series is normalized to have zero mean and unit standard deviation, so the index value represents standard deviations from the average. A reading of zero is average stress. Below zero indicates below-average stress (calm markets, tight spreads, low volatility). Above zero indicates above-average stress, with readings above 1.5 historically associated with significant market dislocations and readings above 3.0 indicating extreme systemic stress.\n\nThe weekly frequency positions the STLFSI4 between daily indices (OFR FSI) and monthly indices (Cleveland CFSI) in terms of responsiveness vs. stability. It is fast enough to capture rapidly evolving stress events within a week or two, while smoothing out the daily noise that can generate false signals. This makes it particularly useful for tracking the trajectory of an evolving crisis — is stress building, plateauing, or receding?\n\nA key feature of the STLFSI4 is its long history — the index can be calculated back to the early 1990s, providing a rich historical context for interpreting current readings. This long history shows that extended periods of very low stress (below -1.0) have reliably preceded major stress episodes. The interpretation is Minsky-consistent: prolonged calm creates complacency, risk-taking, and compressed premiums that set the stage for the next dislocation.",
      "dataSource": "Federal Reserve Bank of St. Louis, updated weekly (Fridays), available through FRED database (series: STLFSI4)",
      "leadTime": "Primarily contemporaneous. Extended low readings (below -1.0 for 12+ months) serve as a contrarian leading indicator — the longer stress is suppressed, the larger the eventual correction.",
      "reliability": "Strong track record spanning 30+ years. PCA methodology is transparent and reproducible. Limitation: as a weekly index, it can miss intra-week stress spikes. Also, PCA weights are data-driven and can shift subtly with revisions.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (peak ~5.5)",
        "2020 COVID Crisis (peak ~9.0, briefly)",
        "2011 European Crisis (~1.5)",
        "Late 2006 pre-GFC (below -1.0 for extended period)"
      ],
      "compoundsWith": [
        "credit-debt",
        "yield-curve",
        "systemic-banking",
        "asset-bubble"
      ],
      "tags": [
        "st-louis-fed",
        "stlfsi4",
        "weekly",
        "18-series",
        "pca",
        "interest-rates",
        "yield-spreads"
      ]
    },
    {
      "id": 8,
      "sourceId": "cm-009",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "general",
      "title": "ECB Composite Indicator of Systemic Stress (CISS)",
      "content": "The European Central Bank's Composite Indicator of Systemic Stress (CISS) is methodologically distinct from US stress indices and captures a phenomenon they miss: the interaction effect between market segments. While US indices typically aggregate stress across categories using fixed or PCA-derived weights, the CISS uses portfolio-theoretical aggregation with time-varying cross-correlations between five market segments: money market, bond market, equity market, financial intermediaries, and foreign exchange.\n\nThe critical innovation is that the CISS is designed to peak not when any single market is under stress, but when multiple segments are simultaneously stressed AND their stress is correlated. This reflects the ECB's research finding that systemic risk is fundamentally about the interaction between segments rather than the stress level of any individual segment. A spike in equity volatility alone is a market correction. A simultaneous spike in equity volatility, credit spreads, and interbank funding costs, with rising correlation between them, is a systemic crisis.\n\nThe portfolio-theoretical weighting means the CISS increases non-linearly when cross-segment correlations rise. During normal times, stress in one segment may be offset by calm in others (diversification effect). During systemic crises, all segments move together (correlation approaching 1), and the portfolio 'diversification' disappears — the CISS spikes dramatically. This mathematical property makes the CISS particularly good at distinguishing between isolated market stress and true systemic events.\n\nThe CISS ranges from 0 to 1, with readings above 0.375 historically associated with periods of identifiable systemic stress. The index captured the 2008 GFC peak (near 0.8), the 2011-2012 European sovereign debt crisis (above 0.5), and the March 2020 COVID shock (above 0.5). Importantly, it did NOT trigger during isolated market events like the 2015 Chinese equity selloff or the 2018 Q4 correction, correctly identifying those as non-systemic episodes. This selectivity — fewer false positives at the cost of slightly slower detection — makes it a high-confidence signal.",
      "dataSource": "European Central Bank, updated weekly, available through ECB Statistical Data Warehouse",
      "leadTime": "Contemporaneous — designed to identify systemic stress in real time rather than predict it. However, a rising CISS from low levels (0.1 to 0.2 to 0.3) provides 2-4 weeks of early warning before threshold breaches.",
      "reliability": "Excellent for distinguishing systemic from non-systemic stress — very few false positives. Limitation: Euro-area focused, may miss US-centric or EM-centric stress that has not yet transmitted to European markets.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (peak ~0.8)",
        "2011-2012 European Sovereign Debt Crisis (peak ~0.55)",
        "2020 COVID Crisis (peak ~0.55)",
        "2015 China selloff (did NOT breach threshold — correct non-signal)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "currency-inflation",
        "general"
      ],
      "tags": [
        "ecb",
        "ciss",
        "systemic-stress",
        "cross-correlation",
        "portfolio-theory",
        "europe",
        "interaction-effect",
        "few-false-positives"
      ]
    },
    {
      "id": 9,
      "sourceId": "cm-010",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "systemic-banking",
      "title": "Cleveland Fed Systemic Risk Indicator (SRI)",
      "content": "The Cleveland Federal Reserve's Systemic Risk Indicator (SRI) takes a fundamentally different approach from market-based stress indices. Instead of measuring financial market conditions (spreads, volatility, correlations), it measures the health of the banking system directly by calculating the distance-to-default for approximately 100 large US bank holding companies.\n\nDistance-to-default is a Merton model concept: it measures how far a bank's asset value is from the point where its liabilities exceed its assets (the default boundary). A high distance-to-default means the bank has a large cushion — its assets substantially exceed its liabilities, and it would take a major shock to push it toward insolvency. A low distance-to-default means the bank is operating near the edge — a moderate decline in asset values could trigger insolvency concerns. The SRI aggregates individual bank distance-to-default measures into a system-wide indicator.\n\nThe SRI's threshold levels are critical for crisis monitoring. When the aggregate SRI falls below 0.1, it historically signals major banking system stress — the entire banking sector is operating dangerously close to its collective default boundary. This threshold was breached during the 2008-2009 crisis and briefly approached during the 2020 COVID shock. Readings between 0.1 and 0.3 indicate moderate banking stress. Readings above 0.5 indicate a healthy banking system with substantial buffers.\n\nWhat makes the SRI uniquely valuable is that it captures banking system vulnerability before market stress indices react. Bank balance sheet deterioration — rising non-performing loans, declining asset values, eroding capital ratios — shows up in distance-to-default calculations before it manifests as wider credit spreads or interbank funding stress. The SRI provided meaningful warning signals in late 2007, several months before the acute phase of the GFC. It also captured the slow deterioration of regional bank health that preceded the March 2023 SVB/Signature/First Republic failures.",
      "dataSource": "Federal Reserve Bank of Cleveland, updated daily, derived from equity market data (stock prices, volatility) combined with balance sheet data for ~100 bank holding companies",
      "leadTime": "1-6 months ahead of acute banking crises. The gradual decline in aggregate distance-to-default provides a building-pressure signal that precedes sudden failure events.",
      "reliability": "Good track record for banking-specific stress. Limitation: uses equity market data, which can be volatile and may overreact to sentiment. Also, it measures publicly traded banks — shadow banking and non-bank financial institutions are not captured.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (below 0.1 threshold)",
        "2023 SVB/Signature/First Republic (declining trend preceded failures)",
        "2020 COVID (sharp decline, rapid recovery)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "asset-bubble"
      ],
      "tags": [
        "cleveland-fed",
        "sri",
        "distance-to-default",
        "banking-system",
        "merton-model",
        "daily",
        "bank-health"
      ]
    },
    {
      "id": 10,
      "sourceId": "cm-011",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "systemic-banking",
      "title": "Federal Reserve Banking System Vulnerability Monitor",
      "content": "The Federal Reserve's Banking System Vulnerability Monitor assesses US banking system resilience across four distinct dimensions, each capturing a different channel through which the banking system can fail. This multi-dimensional approach recognizes that banking crises can originate from different sources and manifest through different mechanisms, so no single metric is sufficient.\n\nThe four dimensions are: (1) Capital Adequacy — measures the ratio of bank equity to total assets and risk-weighted assets across the banking system. When aggregate capital ratios decline, the system has less buffer to absorb losses. (2) Fire Sale Vulnerability — models the systemic impact if banks were forced to sell assets simultaneously. This dimension captures the common asset exposure problem: if many banks hold similar assets, forced selling by one creates mark-to-market losses for all others. (3) Liquidity Risk — measures the mismatch between liquid assets and short-term liabilities across the banking system. Banks that fund long-term assets with short-term wholesale funding are vulnerable to runs. (4) Run Vulnerability — assesses the proportion of bank funding that comes from uninsured deposits and other 'runnable' sources, and models the speed at which a confidence loss could drain funding.\n\nEach dimension is reported as a percentile relative to historical distribution, updated quarterly. The framework gained significant credibility after the March 2023 banking stress: the Run Vulnerability dimension had been flagging elevated risk for several quarters before SVB's failure, as the shift to higher interest rates had created large unrealized losses on bank bond portfolios while simultaneously increasing the incentive for depositors to move money to higher-yielding alternatives.\n\nThe interaction between dimensions matters more than any individual reading. The most dangerous combination is high Fire Sale Vulnerability AND high Run Vulnerability simultaneously — this creates a self-reinforcing loop where runs force fire sales, which create losses, which accelerate runs. The 2008 GFC featured this deadly combination. The Fed's quarterly reports now explicitly analyze these interaction effects, making this one of the most theoretically grounded institutional systemic risk assessments available.",
      "dataSource": "Federal Reserve Board, Financial Stability Report (published twice yearly), with quarterly data updates. Based on confidential supervisory data (Call Reports, FR Y-9C filings).",
      "leadTime": "1-4 quarters leading. Dimension deterioration provides early warning of building vulnerability, though it cannot predict the timing of the trigger event.",
      "reliability": "High credibility — based on confidential supervisory data that is more detailed and accurate than market-based measures. Correctly identified SVB-type vulnerability pre-2023. Limitation: quarterly frequency means it can miss rapidly evolving situations. Also, non-bank financial institutions are not covered.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (all four dimensions elevated)",
        "2023 SVB/Banking stress (Run Vulnerability elevated)",
        "2020 COVID (Liquidity dimension briefly stressed)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "carry-trade"
      ],
      "tags": [
        "federal-reserve",
        "banking-vulnerability",
        "capital-adequacy",
        "fire-sale",
        "liquidity-risk",
        "run-vulnerability",
        "quarterly",
        "supervisory-data"
      ]
    },
    {
      "id": 11,
      "sourceId": "cm-012",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "yield-curve",
      "title": "Yield Curve Inversion: 10-Year minus 3-Month Treasury Spread",
      "content": "The 10-year minus 3-month Treasury yield spread is the single most reliable recession predictor in the macroeconomic toolkit, backed by decades of empirical evidence and multiple independent validation studies. When this spread turns negative (inverts) — meaning short-term rates exceed long-term rates — it signals that the bond market expects future economic weakness severe enough to force the Federal Reserve to cut short-term rates significantly.\n\nThe mechanism is straightforward. Normally, longer-term bonds yield more than shorter-term bonds to compensate investors for duration risk (the term premium). An inversion means investors are so concerned about future economic conditions that they are willing to accept lower yields on long-term bonds to lock in safe returns, while short-term rates remain elevated due to current Fed policy. The inversion reflects the market's collective assessment that current monetary policy is too tight for the economy's trajectory.\n\nThe track record is remarkable. The 10y-3m spread has inverted before every US recession since 1969, with lead times ranging from 10 months to 3 years. The average lead time is approximately 14 months. Critically, the 10y-3m spread has produced only one significant false positive in this period (a brief inversion in 1966 that was followed by a significant slowdown but not an official recession by NBER criteria). This gives the indicator a success rate of approximately 90-95% depending on how you classify the 1966 episode.\n\nRecent research has shown that the signal is robust across different monetary policy regimes, including unconventional policy environments (QE, zero lower bound). The New York Fed maintains a recession probability model based on this spread that is widely followed. However, the indicator says nothing about the severity of the coming recession — a mild inversion can precede a severe recession, and vice versa. It also does not account for non-recessionary financial crises (e.g., market crashes that do not coincide with GDP contractions). The spread inverted in 2022-2023 and remained inverted for the longest period on record, creating an ongoing test case for the indicator's predictive power.",
      "dataSource": "FRED (Federal Reserve Economic Data): T10Y3M series, updated daily. New York Fed Recession Probability model, updated monthly.",
      "leadTime": "10 months to 3 years, with average of ~14 months from first inversion to recession onset. The un-inversion (steepening) often occurs 3-6 months before recession, as the Fed begins cutting rates.",
      "reliability": "Highest of any single recession indicator. ~90-95% accuracy across 55+ years. One possible false positive (1966). Widely validated across academic studies and monetary policy regimes.",
      "historicalPrecedents": [
        "1969-70 Recession",
        "1973-75 Recession",
        "1980 Recession",
        "1981-82 Recession",
        "1990-91 Recession",
        "2001 Recession",
        "2007-09 Great Recession",
        "2020 COVID Recession",
        "2022-2023 prolonged inversion"
      ],
      "compoundsWith": [
        "yield-curve",
        "credit-debt",
        "currency-inflation"
      ],
      "tags": [
        "yield-curve",
        "10y-3m",
        "recession-predictor",
        "treasury-spread",
        "most-reliable",
        "14-month-lead",
        "new-york-fed"
      ]
    },
    {
      "id": 12,
      "sourceId": "cm-013",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "yield-curve",
      "title": "Yield Curve Inversion: 10-Year minus 2-Year Treasury Spread",
      "content": "The 10-year minus 2-year Treasury yield spread is the most widely watched and media-reported yield curve measure, though it is slightly less reliable than the 10y-3m spread as a recession predictor. Its popularity stems from its simplicity and the fact that the 2-year yield is often considered a proxy for expected Fed policy, making the 10y-2y spread a natural measure of term premium and economic expectations.\n\nThe 10y-2y spread inverted before the last several recessions, but with more variability in lead times and a few more ambiguous signals than the 10y-3m spread. Research by the Federal Reserve Bank of San Francisco found that the 10y-3m spread outperforms the 10y-2y spread in formal recession prediction models, though the difference is modest. The 10y-2y spread's advantage is that it tends to invert earlier than the 10y-3m spread (because the 2-year yield responds to expected policy changes faster than the 3-month yield), but this earlier signal also generates more noise.\n\nThe 10y-2y spread has a well-documented behavioral dynamic: its inversion generates significant media attention, which itself can tighten financial conditions through confidence effects. When headlines declare 'the yield curve has inverted,' businesses may become more cautious about investment and hiring, consumers may reduce spending, and lenders may tighten credit standards. This creates a partial self-fulfilling prophecy channel that may amplify the indicator's predictive power beyond what pure economic fundamentals would warrant.\n\nFor crisis monitoring, the 10y-2y spread should be tracked alongside the 10y-3m spread. When both are inverted simultaneously, the recession signal is strongest. When only the 10y-2y is inverted but the 10y-3m is not, the signal is more ambiguous and may reflect term premium distortions rather than genuine recession expectations. The un-inversion pattern is equally important: when a deeply inverted curve begins to steepen rapidly (short-term rates dropping faster than long-term rates), it often signals that the recession is imminent — the bond market is pricing in emergency Fed rate cuts.",
      "dataSource": "FRED (Federal Reserve Economic Data): T10Y2Y series, updated daily",
      "leadTime": "6 months to 2 years from inversion to recession onset. Generally signals earlier than 10y-3m but with more variability.",
      "reliability": "Good but slightly inferior to 10y-3m spread. More prone to brief inversions that may not signal recession. Widely followed, which creates reflexive confidence effects.",
      "historicalPrecedents": [
        "1978-80 Recession",
        "1989-90 Pre-recession",
        "2000 Pre-dot-com recession",
        "2006-07 Pre-GFC",
        "2019 Pre-COVID",
        "2022 deep inversion"
      ],
      "compoundsWith": [
        "yield-curve",
        "credit-debt",
        "currency-inflation"
      ],
      "tags": [
        "yield-curve",
        "10y-2y",
        "recession-predictor",
        "media-watched",
        "self-fulfilling",
        "term-premium"
      ]
    },
    {
      "id": 13,
      "sourceId": "cm-014",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "Credit-to-GDP Gap (Basel III Reference Indicator)",
      "content": "The credit-to-GDP gap is the BIS's and Basel Committee's primary reference indicator for the countercyclical capital buffer (CCyB), making it the only crisis indicator formally embedded in global banking regulation. It measures the deviation of the credit-to-GDP ratio from its long-term trend, calculated using a one-sided Hodrick-Prescott filter. When private sector credit is growing significantly faster than GDP — indicating that the economy is becoming overleveraged — the gap turns positive, signaling the buildup of systemic risk.\n\nThe theoretical foundation is straightforward: credit growth that outpaces economic growth must eventually reverse, either through gradual deleveraging (benign) or sudden credit contraction (crisis). The gap's power comes from measuring not the level of credit but its deviation from trend — a country with high but stable credit-to-GDP is less at risk than a country with lower but rapidly rising credit-to-GDP. This captures the Minsky dynamic where stability and gradually escalating leverage create the conditions for sudden reversal.\n\nThe Basel Committee's research found that the credit-to-GDP gap provides the best early warning of banking crises among a wide range of potential indicators, with an optimal threshold of approximately 10 percentage points (when the gap exceeds 10pp, banking crisis probability rises substantially). The lead time is impressive: 2-5 years before crisis onset, making it one of the longest-range leading indicators available. This long lead time allows regulators to activate countercyclical buffers — requiring banks to hold extra capital during boom periods — years before the crisis actually hits.\n\nHowever, the credit-to-GDP gap has significant practical limitations. The noise-to-signal ratio is high, estimated at 75-150% depending on the methodology and country sample. This means that for every correct signal, there are 0.75 to 1.5 false signals. The one-sided HP filter produces significant end-of-sample revisions, meaning real-time estimates of the gap can differ substantially from final estimates. Additionally, the gap performs poorly in countries that have experienced structural shifts in financial depth (e.g., financial liberalization that permanently raises the credit-to-GDP ratio). Despite these limitations, no alternative indicator has been found that provides comparable long-range warning with acceptable accuracy.",
      "dataSource": "BIS Statistics (quarterly), national central banks, FRED. The BIS publishes credit-to-GDP gaps for 43 countries.",
      "leadTime": "2-5 years — longest lead time of any widely validated indicator. Optimal for macro-prudential policy activation.",
      "reliability": "Best long-range banking crisis predictor per Basel Committee research. However, high noise-to-signal ratio (75-150%) makes it unreliable as a standalone timing indicator. Should be combined with shorter-range indicators for actionable signals.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (gap exceeded 10pp in US by 2006)",
        "1997 Asian Financial Crisis (extremely elevated in Thailand, Korea)",
        "1990 Japan bubble (gap >30pp)",
        "Nordic banking crises 1990-92"
      ],
      "compoundsWith": [
        "credit-debt",
        "asset-bubble",
        "systemic-banking"
      ],
      "tags": [
        "credit-to-gdp-gap",
        "bis",
        "basel-iii",
        "countercyclical-buffer",
        "long-range",
        "hp-filter",
        "macro-prudential",
        "noise-to-signal"
      ]
    },
    {
      "id": 14,
      "sourceId": "cm-015",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "CAPE / Shiller PE Ratio",
      "content": "The Cyclically Adjusted Price-to-Earnings ratio (CAPE), developed by Nobel laureate Robert Shiller, divides the S&P 500 price by the average of 10 years of inflation-adjusted earnings. By using a 10-year earnings average, it smooths out the cyclical fluctuations in corporate profits that make the standard PE ratio misleading at cycle extremes (earnings are high at peaks, making PE look cheap; earnings are low at troughs, making PE look expensive).\n\nThe CAPE's primary value is as a 10-20 year return predictor rather than a short-term crash indicator. High CAPE readings (above 25-30) are associated with lower subsequent 10-year returns, while low CAPE readings (below 15) are associated with higher subsequent returns. The 20th-century average CAPE was approximately 15.21. The CAPE reached 44 before the dot-com crash in 2000, approximately 27 before the 2008 GFC, and has been elevated above 30 for much of the 2020s, reaching above 35 in late 2024 and 2025.\n\nAs a crisis indicator, the CAPE has significant limitations. It can remain elevated for years — even decades — before reverting to the mean. Japan's CAPE was extremely high throughout the late 1980s, years before the 1990 crash. The US CAPE was above 25 for most of the late 1990s, but the crash did not come until 2000. This makes the CAPE a poor timing tool: it tells you that returns will be low and that the market is vulnerable to a shock, but it does not tell you when the shock will occur.\n\nHowever, the CAPE is extremely valuable in compound analysis. A high CAPE combined with tightening financial conditions, rising credit stress, or geopolitical shocks creates a much more dangerous cocktail than the same shocks at average or low valuations. When markets are expensive, there is less margin for error — any negative surprise triggers larger corrections because there is more 'air' under prices. The CAPE should be interpreted as a measure of market fragility: at CAPE 15, the market can absorb shocks; at CAPE 35, the same shocks cause outsized damage.",
      "dataSource": "Robert Shiller's online dataset (updated monthly), available at Yale: http://www.econ.yale.edu/~shiller/data.htm. Also available through Barclays and multpl.com.",
      "leadTime": "10-20 years for return prediction. As a crash indicator, no specific lead time — elevated CAPE indicates vulnerability but not timing.",
      "reliability": "Extremely high reliability as a 10-20 year return predictor (R-squared ~40% for subsequent 10-year returns). Poor reliability as a crash timing indicator. Cannot distinguish between 'expensive but going higher' and 'expensive and about to crash.'",
      "historicalPrecedents": [
        "1929 crash (CAPE ~33)",
        "2000 dot-com peak (CAPE ~44)",
        "2007 pre-GFC (CAPE ~27)",
        "1982 secular bottom (CAPE ~8)",
        "2009 post-GFC bottom (CAPE ~13)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "credit-debt",
        "yield-curve",
        "ai-specific"
      ],
      "tags": [
        "cape",
        "shiller-pe",
        "valuation",
        "long-term-returns",
        "mean-reversion",
        "market-fragility",
        "10-year-earnings"
      ]
    },
    {
      "id": 15,
      "sourceId": "cm-016",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "Mispricing Risk Indicator",
      "content": "Mispricing risk, as identified in BIS and academic research, is measured by the combination of rapid price growth with anomalously low volatility. It outperforms traditional credit metrics as a crisis predictor in several empirical studies — a finding that surprised many researchers who expected credit conditions to dominate. The indicator captures a fundamentally different phenomenon than credit excess: it detects complacency, the market condition where participants systematically underestimate risk.\n\nThe mechanism is rooted in behavioral finance and market microstructure. When asset prices are rising steadily and volatility is compressed, several self-reinforcing dynamics take hold. Portfolio insurance and value-at-risk models reduce required hedging (low vol = low risk = less hedging needed). Options sellers offer cheaper protection (low implied volatility = cheap puts). Leveraged strategies that profit from low volatility (carry trades, volatility selling, risk parity) attract more capital, which further suppresses volatility. The result is a market that appears calm and well-functioning but is actually increasingly brittle — positioned for a large move that the market's own risk models say is extremely unlikely.\n\nThe lead time for mispricing risk is typically 2-3 years, placing it between short-range indicators (VIX, stress indices) and long-range indicators (credit-to-GDP gap). This lead time is useful for strategic positioning: it provides enough warning to reduce exposure to overpriced assets without requiring the precise timing that shorter-range indicators demand. The indicator flagged elevated risk in 2005-2006 (2-3 years before the GFC), in 1998-1999 (1-2 years before the dot-com peak), and in 2017-2018 (ahead of the late 2018 volatility event, though not a full crisis).\n\nFor the current AI investment cycle, mispricing risk is particularly relevant. AI-related stocks experienced rapid price appreciation with periods of unusually low volatility through 2024 and into 2025 — exactly the pattern the indicator flags. The concentration of market gains in a small number of mega-cap tech stocks (the 'Magnificent 7' phenomenon) amplifies the risk because portfolio-level volatility can be low even while single-stock concentration risk is extreme. When a correction does come, the concentration means the index-level impact is far larger than the volatility models predicted.",
      "dataSource": "BIS working papers, academic research (Borio and Drehmann, Aldasoro et al.), calculated from price series and realized volatility data",
      "leadTime": "2-3 years — medium-range. Useful for strategic positioning and portfolio adjustment rather than tactical trading.",
      "reliability": "Outperforms credit-to-GDP gap in some studies for medium-term crisis prediction. Lower noise-to-signal ratio than credit gap. Limitation: the indicator is less standardized than other metrics — different researchers use different specifications for 'rapid price growth' and 'low volatility.'",
      "historicalPrecedents": [
        "2005-2006 pre-GFC (elevated)",
        "1998-1999 pre-dot-com (elevated)",
        "2017-2018 pre-volatility event (moderately elevated)",
        "2024-2025 AI stock concentration (potentially elevated)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "carry-trade",
        "credit-debt"
      ],
      "tags": [
        "mispricing-risk",
        "complacency",
        "low-volatility",
        "bis",
        "outperforms-credit",
        "2-3-year-lead",
        "behavioral-finance"
      ]
    },
    {
      "id": 16,
      "sourceId": "cm-017",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "Debt Service Ratio (DSR)",
      "content": "The Debt Service Ratio measures the share of income devoted to servicing debt — both principal and interest payments. Published by the BIS for the private non-financial sector of 32 economies, it captures how stretched borrowers are at any given moment. Unlike the credit-to-GDP gap (which measures the stock of debt relative to economic output), the DSR measures the flow burden of debt (what borrowers actually have to pay each period). This makes it directly sensitive to interest rate changes: even if debt levels are stable, a rate increase raises the DSR by increasing interest payments.\n\nThe DSR has a lead time of up to 2 years for banking crises and recessions, based on BIS research across multiple countries and crisis episodes. Its predictive power comes from a simple mechanism: when borrowers are spending an unusually large share of income on debt service, they have less capacity to absorb negative shocks (income loss, rate increases, asset price declines). At high DSR levels, even a moderate shock can push marginal borrowers into default, triggering a cascade of credit losses through the banking system.\n\nThe DSR's sensitivity to interest rates makes it particularly relevant during monetary tightening cycles. The 2022-2024 rate hiking cycle by the Federal Reserve and other central banks mechanically increased DSRs across the developed world. The full impact of rate increases on DSR operates with a lag — existing fixed-rate loans are unaffected until they mature and must be refinanced. For example, the UK, where mortgage terms are typically 2-5 year fixed rates, experienced a rolling DSR increase as mortgages repriced over 2023-2025. The US, with 30-year fixed mortgage dominance, has a much longer transmission lag for household DSR but corporate DSR (which relies more on floating-rate and shorter-term debt) adjusted more quickly.\n\nFor crisis monitoring, the DSR is most powerful when combined with income growth projections. A high DSR in an economy with rising incomes is manageable — borrowers are growing into their debt. A high DSR in an economy with stagnant or declining incomes is a crisis precursor — borrowers are drowning. The combination of elevated DSR + decelerating income growth + continued rate elevation is the canonical pre-crisis signature for credit-driven recessions.",
      "dataSource": "BIS Statistics (quarterly), published for 32 countries. Available at https://www.bis.org/statistics/dsr.htm",
      "leadTime": "Up to 2 years for banking crises. Shorter lead time (6-12 months) for consumer spending slowdowns.",
      "reliability": "Strong empirical validation across countries and crisis types. Limitation: the BIS DSR uses aggregated data — it cannot capture distributional stress (e.g., if 80% of borrowers are fine but 20% are severely stressed, the aggregate DSR may look moderate).",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (US household DSR peaked ~18.5% in 2007)",
        "1990 UK recession (mortgage DSR spike after rate increases)",
        "Nordic banking crises 1990-92 (elevated DSR from rate shocks)"
      ],
      "compoundsWith": [
        "credit-debt",
        "yield-curve",
        "currency-inflation",
        "systemic-banking"
      ],
      "tags": [
        "debt-service-ratio",
        "bis",
        "flow-measure",
        "rate-sensitivity",
        "income-coverage",
        "mortgage-repricing",
        "2-year-lead"
      ]
    },
    {
      "id": 17,
      "sourceId": "cm-018",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "VIX Term Structure Inversion",
      "content": "The VIX term structure — the relationship between near-term and longer-term VIX futures — provides a rapid-response crash signal that operates on a timescale of days to weeks. Under normal conditions, the VIX term structure is in 'contango': longer-dated VIX futures trade at a premium to shorter-dated futures, reflecting the uncertainty premium of more distant time horizons. When the term structure inverts into 'backwardation' — short-term VIX exceeds medium-term VIX — it signals that the market expects imminent stress that is more severe than the baseline expectation for the coming months.\n\nThe signal is generated by options market microstructure. When institutional investors urgently need portfolio protection, they bid up short-term put options, which increases near-term implied volatility (the VIX). If the demand for protection is intense and concentrated in the near term, it pushes short-term implied vol above longer-term implied vol, inverting the term structure. This pattern reflects not just increased fear but increased urgency — the market expects something bad to happen soon, not just eventually.\n\nBackwardation events are relatively rare — the VIX term structure spends the vast majority of time in contango. When backwardation occurs and persists for more than 1-2 days, it has historically coincided with or slightly preceded major market dislocations. The VIX term structure inverted during the October 2008 crash, the August 2015 China-driven selloff, the February 2018 'Volmageddon' event, the March 2020 COVID crash, and several other acute stress episodes.\n\nThe indicator's value lies in its speed and specificity. Unlike most crisis indicators that operate on monthly or quarterly timescales, the VIX term structure provides near-real-time information about institutional stress positioning. However, its short lead time (days to weeks at most) makes it useful for tactical risk management rather than strategic positioning. It answers the question 'is a crash happening right now or about to happen?' rather than 'is a crisis building over the coming year?' For a comprehensive monitoring system, VIX term structure serves as the canary in the coal mine — it fires last but fires first when the crisis actually arrives.",
      "dataSource": "CBOE VIX Index and VIX futures curves, real-time data available through CBOE and major data providers. Calculated as difference between front-month and second-month VIX futures.",
      "leadTime": "Days to weeks. Extremely short-range. Primarily contemporaneous with acute stress events.",
      "reliability": "High specificity for acute market stress events — when backwardation persists, a significant market dislocation is almost always occurring or imminent. Limitation: brief (intra-day) backwardation events can be noise. Also, the signal only captures equity market stress — crises that manifest first in credit or FX markets may not trigger VIX inversion until later.",
      "historicalPrecedents": [
        "October 2008 crash (deep backwardation)",
        "August 2015 China selloff (brief backwardation)",
        "February 2018 Volmageddon (inverted before VIX spike)",
        "March 2020 COVID crash (deep sustained backwardation)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "carry-trade",
        "geopolitical",
        "systemic-banking"
      ],
      "tags": [
        "vix",
        "term-structure",
        "backwardation",
        "contango",
        "crash-signal",
        "real-time",
        "options-market",
        "short-range"
      ]
    },
    {
      "id": 18,
      "sourceId": "cm-019",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "credit-debt",
      "title": "Credit/Debt Crisis Trigger",
      "content": "The credit/debt trigger encompasses the broad category of crisis events originating from excessive borrowing, lending deterioration, or sudden credit contraction. This is historically the most common origin of major financial crises — Reinhart and Rogoff's research shows that the majority of systemic banking crises are preceded by credit booms, making credit conditions the single most important domain for crisis monitoring.\n\nKey metrics to track include: corporate debt-to-earnings ratios (how leveraged companies are relative to their income), consumer credit growth rates (accelerating consumer borrowing signals late-cycle behavior), sovereign debt-to-GDP (government borrowing capacity), credit-to-GDP gap (Basel III reference, see cm-014), high-yield bond spreads (the price the market charges for lending to risky borrowers), and lending standards surveys (the Fed's Senior Loan Officer Opinion Survey, or SLOOS, captures whether banks are tightening or loosening credit).\n\nThe credit trigger activates through several channels. Gradual deterioration: credit quality slowly worsens as borrowers take on more risk during expansions, eventually reaching a tipping point where defaults spike. Sudden repricing: an external event (rate shock, geopolitical crisis, accounting scandal) causes the market to suddenly reassess credit risk, leading to a jump in spreads and a freeze in new issuance. Rollover crisis: borrowers who depend on refinancing maturing debt find they cannot refinance at affordable rates, triggering defaults even among borrowers whose businesses are viable. Credit channel freeze: banks or capital markets stop lending entirely, cutting off credit to the real economy and causing a sharp contraction.\n\nThe current (2026) credit environment features elevated corporate debt, particularly in the BBB-rated segment (the lowest investment-grade category). A significant portion of corporate debt was issued during the low-rate 2020-2021 period and is approaching refinancing dates at substantially higher rates. The 'maturity wall' of corporate debt refinancing in 2025-2027 represents a concrete rollover risk. Additionally, private credit markets have grown enormously ($1.7T+ by 2025 estimates), and their opacity makes systemic risk assessment more difficult than for publicly traded debt.",
      "dataSource": "FRED (credit spreads, SLOOS), BIS (credit-to-GDP), Bloomberg/ICE (high-yield indices), S&P/Moody's (default rates), Federal Reserve Financial Stability Report",
      "leadTime": "Variable: credit-to-GDP gap provides 2-5 years. Spread widening provides 3-12 months. SLOOS tightening provides 6-12 months. Rollover/maturity wall analysis provides 1-2 years.",
      "reliability": "Credit excesses precede the majority of banking crises. However, false positives are common — credit conditions can deteriorate and then stabilize without crisis. The key is the combination of elevated credit with a trigger event.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis",
        "1990-91 S&L Crisis",
        "1997 Asian Financial Crisis",
        "2001 Dot-com/corporate credit (Enron, WorldCom)",
        "European Sovereign Debt Crisis 2010-2012"
      ],
      "compoundsWith": [
        "yield-curve",
        "asset-bubble",
        "systemic-banking",
        "currency-inflation",
        "carry-trade"
      ],
      "tags": [
        "credit-trigger",
        "corporate-debt",
        "consumer-credit",
        "sovereign-debt",
        "spreads",
        "sloos",
        "maturity-wall",
        "private-credit"
      ]
    },
    {
      "id": 19,
      "sourceId": "cm-020",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "asset-bubble",
      "title": "Asset Bubble Crisis Trigger",
      "content": "Asset bubble triggers encompass crisis events originating from the deflation of overvalued asset markets — equities, real estate, crypto, or other asset classes where prices have departed substantially from fundamental valuations. The danger is not the overvaluation itself but the leverage, speculation, and behavioral dynamics that accompany it. When prices reverse, the unwinding of leveraged positions and the collapse of speculative confidence can cascade through the financial system.\n\nKey metrics for bubble monitoring include: CAPE/Shiller PE ratio relative to historical norms (current levels above 35 indicate extreme overvaluation by historical standards), market concentration risk (the weight of the top 10 stocks in the S&P 500, which has exceeded 35% — a level not seen since the early 1970s), margin debt levels (NYSE margin debt as a percentage of GDP), IPO/SPAC activity (frothy new issuance indicates late-cycle speculation), and sector-specific metrics like the AI capex-to-revenue gap.\n\nThe 'Magnificent 7' concentration phenomenon of 2023-2025 represents a specific form of asset bubble risk. When a small number of stocks drive the majority of index returns, portfolios that appear diversified (because they hold 'the index') are actually highly concentrated bets on a handful of companies. This concentration creates a vulnerability: negative news affecting even one or two of these companies can move the entire market, because they represent such a large share of the index. The NVIDIA single-stock risk to the S&P 500 in 2024-2025 was a concrete example — a company whose individual news events could move the entire index by 1-2%.\n\nBubble deflation can be gradual or sudden. Gradual deflation (the 'slowly leaking balloon') occurs when fundamental disappointment accumulates over quarters — companies fail to deliver the growth that justified elevated valuations, and prices drift lower. Sudden deflation (the 'pop') occurs when a catalyst — margin call cascade, fraud revelation, external shock — triggers panic selling. The AI investment cycle is at risk of both: gradual deflation as AI revenue growth disappoints capex levels, and sudden deflation if a broader financial shock forces leveraged AI investors to sell simultaneously.",
      "dataSource": "Shiller CAPE (monthly), S&P 500 concentration data (daily), NYSE margin statistics (monthly), FINRA margin debt, Bloomberg PE and valuation multiples",
      "leadTime": "CAPE provides 10-20 year return forecasting but no timing. Concentration risk is a structural vulnerability measure. Margin debt peaks typically 3-12 months before corrections.",
      "reliability": "Bubble identification in retrospect has 100% accuracy; bubble identification in real-time is notoriously difficult ('the market can stay irrational longer than you can stay solvent'). Valuation extremes correctly identify vulnerability but provide no timing information.",
      "historicalPrecedents": [
        "1929 Great Crash",
        "1989 Japan Nikkei peak",
        "2000 Dot-com bubble",
        "2007-08 Housing bubble",
        "2021-22 Crypto/SPAC/Meme bubble"
      ],
      "compoundsWith": [
        "credit-debt",
        "ai-specific",
        "carry-trade",
        "systemic-banking"
      ],
      "tags": [
        "asset-bubble",
        "concentration-risk",
        "cape",
        "margin-debt",
        "mag-7",
        "speculation",
        "leverage",
        "bubble-deflation"
      ]
    },
    {
      "id": 20,
      "sourceId": "cm-021",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "geopolitical",
      "title": "Geopolitical Crisis Trigger",
      "content": "Geopolitical triggers encompass crisis events originating from international conflicts, trade wars, sanctions, regime changes, and other political shocks that disrupt economic activity and market confidence. Unlike credit or bubble triggers that build slowly and predictably, geopolitical triggers can materialize rapidly with little warning, making them the most difficult category to forecast but also the most important to monitor in real-time.\n\nThe financial market impact of geopolitical events depends on several factors: geographic proximity to critical economic infrastructure (oil fields, shipping lanes, semiconductor fabs), involvement of major economic powers (US, China, EU), impact on commodity prices (especially energy), and duration/escalation potential. Research by Dario Caldara and Matteo Iacoviello at the Federal Reserve Board produced the Geopolitical Risk (GPR) Index, which quantifies geopolitical threats and events using news-based text analysis. Their research shows that elevated GPR is associated with reduced investment, lower equity returns, and increased capital flows to safe havens.\n\nThe current (2026) geopolitical landscape features several active or potential trigger points: US-China technology competition and semiconductor export controls, the ongoing reorganization of global supply chains ('friend-shoring'), active conflicts in multiple regions, and trade policy uncertainty including tariff regimes. Each of these has the potential to escalate into a market-moving event. The March 2026 Iran tensions demonstrated how quickly geopolitical shocks can propagate through financial markets — oil prices spiked, emerging market currencies sold off, and safe-haven assets rallied within hours.\n\nFor crisis monitoring, geopolitical triggers are most dangerous when they compound with existing financial vulnerabilities. A geopolitical shock in a low-leverage, well-capitalized financial system causes temporary market volatility that reverses within weeks. The same shock in a high-leverage, asset-bubble environment can trigger margin calls, forced selling, and systemic cascades. This is why geopolitical monitoring must be integrated with financial vulnerability assessment — the question is not just 'what might happen geopolitically?' but 'how would the current financial system absorb it?'",
      "dataSource": "Caldara-Iacoviello Geopolitical Risk Index (monthly), news monitoring, government policy announcements, military intelligence assessments (open source)",
      "leadTime": "Highly variable. Some geopolitical events (trade wars, sanctions escalation) build over months with visible warning signs. Others (military strikes, regime changes) can occur with hours of warning. Intelligence-based early warning systems may provide days to weeks of lead time for military events.",
      "reliability": "Geopolitical events are inherently unpredictable. The GPR Index captures market-relevant geopolitical risk but cannot predict specific events. Historical base rates: major geopolitical shocks that move markets >5% occur approximately once every 2-3 years.",
      "historicalPrecedents": [
        "1973 Oil Embargo",
        "1990 Gulf War",
        "2001 September 11 attacks",
        "2014 Crimea annexation/sanctions",
        "2022 Russia-Ukraine war",
        "March 2026 Iran tensions"
      ],
      "compoundsWith": [
        "energy-shock",
        "currency-inflation",
        "asset-bubble",
        "carry-trade",
        "general"
      ],
      "tags": [
        "geopolitical",
        "conflict",
        "sanctions",
        "trade-war",
        "gpr-index",
        "supply-chain",
        "safe-haven",
        "rapid-onset"
      ]
    },
    {
      "id": 21,
      "sourceId": "cm-022",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "energy-shock",
      "title": "Energy Shock Crisis Trigger",
      "content": "Energy shock triggers encompass crisis events originating from sudden disruptions to energy supply or extreme price movements. Oil remains the most systemically important commodity despite the energy transition, because it is embedded in virtually every supply chain (transportation, petrochemicals, agriculture, manufacturing). A significant oil price spike acts as a tax on the global economy, transferring wealth from consumers and oil-importing nations to oil producers, reducing disposable income, and raising input costs across industries.\n\nKey metrics for energy shock monitoring include: Brent/WTI crude price deviation from 12-month average (spikes above 50% deviation historically correlate with recessions), OPEC spare capacity (the buffer available to absorb supply disruptions — lower spare capacity = higher shock vulnerability), strategic petroleum reserve levels, oil inventory data (OECD commercial inventories), and the energy intensity of GDP (which determines how sensitive the economy is to price changes — this has declined over decades but remains significant).\n\nThe transmission mechanism from energy shock to financial crisis operates through several channels. Direct cost impact: higher energy prices raise costs for energy-intensive industries (transport, chemicals, agriculture, manufacturing), compressing margins and potentially triggering defaults in leveraged companies. Inflationary pressure: energy price spikes feed through to CPI, potentially forcing central banks to tighten monetary policy even during economic weakness (the stagflation trap). Consumer spending compression: higher gas and heating costs reduce discretionary spending, hitting retail, hospitality, and entertainment. Supply chain disruption: energy shortages or extreme prices can physically disrupt production and transportation, creating cascading input-output effects.\n\nHistorical energy shocks have been responsible for or contributed to several major economic crises: the 1973 OPEC embargo triggered a deep global recession, the 1979 Iranian Revolution oil shock contributed to the early 1980s recession, and the 2022 Russia-Ukraine energy crisis pushed Europe toward recession. The interaction between energy shocks and monetary policy is critical: if inflation is already elevated when an energy shock hits, central banks face an impossible choice between fighting inflation (raise rates, worsen recession) and supporting the economy (cut rates, risk inflation spiral).",
      "dataSource": "EIA (Energy Information Administration) weekly reports, IEA (International Energy Agency) monthly oil market reports, OPEC monthly oil market reports, Baker Hughes rig count, CFTC Commitments of Traders (oil futures positioning)",
      "leadTime": "Geopolitical energy shocks: hours to days. Supply/demand imbalance buildups: months. OPEC policy shifts: weeks. Overall, energy shocks are among the hardest to predict with useful lead times.",
      "reliability": "The causal link between large energy price spikes and economic slowdowns is well-established. However, the economy's sensitivity to energy shocks has declined over time (lower energy intensity, more diverse energy mix, better hedging). A $20/barrel spike in 2026 has less impact than a $20 spike in 1979.",
      "historicalPrecedents": [
        "1973 OPEC Oil Embargo",
        "1979 Iranian Revolution oil shock",
        "1990 Gulf War oil spike",
        "2008 oil spike to $147/barrel",
        "2022 Russia-Ukraine energy crisis"
      ],
      "compoundsWith": [
        "geopolitical",
        "currency-inflation",
        "credit-debt",
        "general"
      ],
      "tags": [
        "energy-shock",
        "oil",
        "opec",
        "supply-disruption",
        "stagflation",
        "inflation-transmission",
        "spare-capacity"
      ]
    },
    {
      "id": 22,
      "sourceId": "cm-023",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "currency-inflation",
      "title": "Currency/Inflation Crisis Trigger",
      "content": "Currency and inflation triggers encompass crisis events originating from rapid currency depreciation, persistent above-target inflation, loss of central bank credibility, or sudden shifts in capital flows. These triggers are particularly dangerous for emerging markets (where currency crises can spiral into sovereign defaults) but also affect developed economies when inflation becomes entrenched or policy errors undermine confidence.\n\nKey metrics include: DXY (US Dollar Index) — dollar strengthening tightens global financial conditions because so much international debt is denominated in dollars; CPI and PCE inflation readings relative to central bank targets; inflation expectations (breakeven rates, University of Michigan survey) — if expectations become 'de-anchored,' inflation becomes self-fulfilling; real interest rates (nominal rates minus inflation) — deeply negative real rates indicate eroding purchasing power and potential currency pressure; and cross-currency basis swaps, which measure the premium for obtaining dollars in international markets.\n\nThe inflation-policy trap is the most dangerous dynamic in this category. When inflation is elevated and the central bank is tightening monetary policy, any additional negative shock (geopolitical event, energy spike, financial stress) creates an impossible trilemma: the central bank must choose between continuing to fight inflation (risking a deeper downturn), pausing to support the economy (risking inflation re-acceleration), or intervening in financial markets (risking credibility). The 2022-2023 period demonstrated this tension when the Fed continued hiking rates even as SVB and regional banks came under stress.\n\nCurrency crises in emerging markets follow a well-documented pattern: capital inflows during good times fund current account deficits and asset appreciation. When global conditions tighten (typically driven by Fed rate hikes or risk-off sentiment), capital flows reverse. The local currency depreciates, raising the burden of dollar-denominated debt. Companies and governments that borrowed in dollars face margin calls and rollover difficulties. The central bank must choose between defending the currency (by raising rates, which crushes the domestic economy) or letting it depreciate (which worsens the debt burden). This dynamic was central to the 1997 Asian crisis, the 2018 EM selloff, and remains a live risk for several economies in 2026.",
      "dataSource": "FRED (CPI, PCE, DXY, breakeven rates), BIS (effective exchange rates, cross-border claims), IMF (reserve adequacy), central bank communications",
      "leadTime": "Inflation trends: 6-18 months (inflation is a lagging indicator, but policy responses have forward-looking impact). Currency crises: days to weeks once capital flight begins, but balance of payments vulnerabilities can be identified months in advance.",
      "reliability": "Inflation persistence is well-forecasted by existing models. Currency crisis timing is notoriously difficult — vulnerabilities can persist for years before a trigger event forces the crisis. The combination of high external debt + current account deficit + strong dollar is a reliable vulnerability indicator, but the timing of the crisis depends on unpredictable trigger events.",
      "historicalPrecedents": [
        "1997 Asian Financial Crisis (currency cascade)",
        "2018 Turkey/Argentina currency crises",
        "1970s US stagflation",
        "2022-2023 global inflation surge",
        "1992 ERM crisis (GBP)"
      ],
      "compoundsWith": [
        "credit-debt",
        "energy-shock",
        "geopolitical",
        "carry-trade",
        "yield-curve"
      ],
      "tags": [
        "currency-crisis",
        "inflation",
        "dxy",
        "dollar-strength",
        "de-anchoring",
        "em-vulnerability",
        "capital-flight",
        "policy-trilemma"
      ]
    },
    {
      "id": 23,
      "sourceId": "cm-024",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "systemic-banking",
      "title": "Systemic Banking Crisis Trigger",
      "content": "Systemic banking triggers encompass crisis events originating from distress within the banking system itself — bank failures, credit freezes, interbank market dysfunction, or runs on deposits. Banking crises are distinct from other financial crises because banks are the transmission mechanism for credit to the real economy. When banks cannot or will not lend, even healthy businesses lose access to working capital, and economic activity contracts sharply.\n\nKey metrics include: distance-to-default aggregate (Cleveland Fed SRI, see cm-010), bank CDS spreads (the cost of insuring against bank default), interbank lending rates vs. risk-free rates (TED spread, OIS-Libor spread), commercial real estate (CRE) exposure as a percentage of bank capital (CRE concentration has been a key vulnerability since 2022), unrealized losses on bank securities portfolios, deposit flow data (large outflows from specific banks or bank categories), and Federal Reserve discount window borrowing (usage spikes indicate banks cannot fund themselves in private markets).\n\nThe March 2023 banking stress (SVB, Signature Bank, First Republic) revealed a vulnerability that was hiding in plain sight: banks that had purchased long-duration bonds during the low-rate era were sitting on massive unrealized losses. As long as these bonds were held to maturity, the losses were theoretical. But when deposit outflows forced banks to sell bonds at market prices, the unrealized losses became real, destroying capital. This 'duration risk' vulnerability was identifiable from publicly available data but was not widely flagged by the monitoring community until after the crisis.\n\nCommercial real estate represents the most visible current banking vulnerability. Office vacancy rates have remained elevated post-pandemic, and many CRE loans originated in the 2019-2021 period are approaching maturity with property values significantly below original underwriting assumptions. Regional and community banks are disproportionately exposed to CRE (in some cases, CRE loans represent 300%+ of bank capital). The CRE maturity wall of 2025-2027, combined with persistently high vacancy rates, represents a concrete, identifiable trigger for banking sector stress. The open question is whether the problem is concentrated enough in individual banks to trigger failures, or distributed broadly enough to be absorbed by the system.",
      "dataSource": "FDIC Quarterly Banking Profile, Federal Reserve H.8 statistical release (bank balance sheets), Fed discount window data, bank call reports, CoStar/CBRE (CRE data), Bloomberg (bank CDS spreads)",
      "leadTime": "Distance-to-default: 1-6 months. CDS spread widening: weeks to months. Deposit flows: days to weeks (but data is lagged by weeks). CRE maturity wall: identifiable 1-2 years ahead.",
      "reliability": "Banking stress indicators have a good track record but tend to fire late — by the time bank CDS spreads spike or discount window usage surges, the crisis is already underway. CRE concentration analysis is more forward-looking but the timing of CRE-driven bank stress depends on loan-by-loan maturity schedules.",
      "historicalPrecedents": [
        "2008 GFC (Lehman, Bear Stearns, WaMu)",
        "2023 SVB/Signature/First Republic",
        "1980s S&L Crisis",
        "2010-12 European banking stress",
        "1930s Great Depression banking collapse"
      ],
      "compoundsWith": [
        "credit-debt",
        "asset-bubble",
        "carry-trade",
        "currency-inflation"
      ],
      "tags": [
        "banking-crisis",
        "bank-runs",
        "cre-exposure",
        "deposit-flight",
        "duration-risk",
        "unrealized-losses",
        "discount-window",
        "too-big-to-fail"
      ]
    },
    {
      "id": 24,
      "sourceId": "cm-025",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "ai-specific",
      "title": "AI Bubble / Capex Overinvestment Trigger",
      "content": "The AI-specific crisis trigger tracks the risk of a major correction in the artificial intelligence investment cycle. As of early 2026, cumulative AI infrastructure capital expenditure by major tech companies has approached or exceeded $700 billion annually, while actual AI-generated revenue across the industry is estimated at $35-50 billion — a 14-20x gap between investment and revenue. This gap must close either through revenue catching up to investment (the bull case) or investment correcting down to match reality (the bear case). History strongly suggests that some combination of both occurs.\n\nThe structural parallels to previous technology bubbles are striking. The dot-com era featured 'vendor financing' — telecom equipment makers (Cisco, Lucent, Nortel) lent money to their own customers to buy their products, creating circular revenue that collapsed when the financing dried up. The current AI cycle features hyperscalers (Microsoft, Google, Amazon, Meta) spending massively on NVIDIA GPUs, partially financed by the cash flows from cloud services that themselves benefit from AI hype. If cloud spending slows, GPU purchases slow, NVIDIA revenue declines, and the AI narrative reverses — a similar self-referential dynamic.\n\nGPU depreciation represents an under-appreciated risk factor. Current-generation GPUs (H100, B100 series) are depreciating assets that lose value as next-generation chips arrive. Companies that invested billions in GPU clusters face writedowns if (a) the expected revenue to justify the investment does not materialize, or (b) newer, more efficient chips make existing inventory obsolete before it has been fully amortized. The rapid pace of AI chip development — roughly annual generational improvements — means the amortization window is short and the obsolescence risk is high.\n\nThe commoditization risk adds another layer. As open-source AI models approach the performance of proprietary models, and as inference costs decline dramatically, the pricing power that justified massive AI infrastructure investment may erode. If the 'moat' of proprietary AI narrows, the companies that spent hundreds of billions building infrastructure find they are competing on price rather than capability — a classic commoditization trap that compresses margins and destroys the ROI assumptions underlying the original investment. For crisis monitoring, the key metrics are: actual AI revenue growth rates vs. capex growth rates, GPU utilization rates, cloud revenue growth deceleration, and any signs of AI project cancellations or spending pullbacks by enterprise customers.",
      "dataSource": "Company earnings reports (capex disclosures), semiconductor industry data (SIA, SEMI), AI revenue estimates (Gartner, IDC), GPU utilization data (limited availability), cloud revenue growth rates",
      "leadTime": "The vulnerability is identifiable now; the timing of correction is unknown. Potential catalysts include: a major AI project failure/cancellation, significant earnings misses by AI-adjacent companies, or a broader market correction that forces reassessment of growth projections. Lead time from first visible signs of AI revenue disappointment to full correction could be 6-18 months.",
      "reliability": "High confidence that the capex-revenue gap must close. Low confidence in timing and mechanism. The gap could close gradually (revenue catches up partially, capex moderates) rather than through a dramatic crash. However, the magnitude of the gap and the leverage embedded in AI-related stocks suggest that the correction, when it comes, will be significant.",
      "historicalPrecedents": [
        "2000-2002 Dot-com crash (telecom capex bubble)",
        "2022 Meta capex crash (75% stock decline on metaverse spending)",
        "1980s biotech hype cycle",
        "2017-2018 crypto infrastructure buildout and crash"
      ],
      "compoundsWith": [
        "asset-bubble",
        "credit-debt",
        "carry-trade",
        "general"
      ],
      "tags": [
        "ai-bubble",
        "capex-revenue-gap",
        "gpu-depreciation",
        "commoditization",
        "nvidia",
        "hyperscaler",
        "vendor-financing",
        "dot-com-parallel"
      ]
    },
    {
      "id": 25,
      "sourceId": "cm-026",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "carry-trade",
      "title": "Carry Trade Unwinding Trigger",
      "content": "Carry trades — borrowing in low-interest-rate currencies to invest in higher-yielding assets — are a major source of systemic risk because they create hidden leverage that is invisible to most market participants and regulators. The classic carry trade borrows in Japanese yen (historically near-zero rates) or Swiss franc and invests in emerging market bonds, high-yield corporate debt, or equity markets. The trade is profitable as long as (a) the interest rate differential persists and (b) the funding currency does not appreciate. When either condition changes, the unwind can be devastating.\n\nThe carry trade follows a characteristic pattern described as 'stairs up, elevator down.' Profits accumulate slowly over weeks and months as the interest rate differential generates steady income. Losses materialize in hours or days when the funding currency suddenly strengthens (typically during a risk-off event), forcing traders to cover their positions by buying back the funding currency — which pushes it higher, causing more losses, forcing more covering. This self-reinforcing dynamic creates the extreme speed and violence of carry trade unwinds.\n\nThe aggregate size of global carry trades is nearly impossible to measure precisely, because much of the positioning occurs through OTC derivatives, FX forwards, and cross-border banking flows that are not reported in real-time. Estimates range from $500 billion to $2 trillion in yen carry trades alone, with significant additional positioning in other currency pairs. This opacity means that carry trade unwinds can surprise even sophisticated market participants — the trigger event may seem minor relative to the market reaction, because the hidden leverage amplifies the impact.\n\nThe Bank of Japan's monetary policy normalization creates a specific carry trade risk for 2025-2026. As Japan moves away from ultra-loose monetary policy and the yen strengthens, the incentive to maintain yen-funded carry trades diminishes. A sudden yen strengthening event — perhaps triggered by an unexpected BOJ rate hike or a global risk-off event — could trigger a rapid unwind. The July 2024 mini-crisis, when a modest BOJ policy shift triggered a sharp yen rally and global equity selloff, was a preview of this dynamic at a relatively small scale. A larger, more sustained yen strengthening event would have proportionally larger effects, with contagion spreading from FX markets to fixed income, equities, and emerging markets simultaneously.",
      "dataSource": "BIS Triennial FX Survey, CFTC Commitments of Traders (currency futures positioning), TFX (Tokyo Financial Exchange) positioning data, cross-border banking statistics",
      "leadTime": "Structural vulnerability can be identified months ahead (elevated carry positioning, narrowing rate differentials, BOJ policy signals). The unwind itself typically occurs over 1-5 days with minimal advance warning.",
      "reliability": "Carry trade unwinds are well-documented and mechanistically understood. The difficulty is identifying the trigger event and predicting the magnitude of unwinding. Positioning data is incomplete, making it hard to estimate the potential size of an unwind event.",
      "historicalPrecedents": [
        "1998 LTCM/Yen carry unwind",
        "2007-2008 carry trade unwind (yen appreciated 25%+ during GFC)",
        "2015 SNB peg removal (EUR/CHF)",
        "July 2024 BOJ mini-crisis",
        "1997 Asian Financial Crisis (carry-funded EM investments)"
      ],
      "compoundsWith": [
        "currency-inflation",
        "asset-bubble",
        "geopolitical",
        "systemic-banking"
      ],
      "tags": [
        "carry-trade",
        "yen",
        "boj",
        "hidden-leverage",
        "stairs-up-elevator-down",
        "fx",
        "unwind",
        "otc-opacity"
      ]
    },
    {
      "id": 26,
      "sourceId": "cm-027",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "general",
      "title": "Correlation Breakdown / Diversification Failure",
      "content": "Correlation breakdown describes the phenomenon where asset correlations spike during crises, causing diversified portfolios to suffer losses far exceeding what their normal-condition risk models predicted. Under normal market conditions, stocks, bonds, commodities, and alternative assets exhibit moderate or negative correlations — losses in one asset class are partially offset by gains in another. During crises, correlations across asset classes converge toward 1.0, and diversification — the fundamental risk management tool — fails precisely when it is needed most.\n\nThe mechanism is driven by forced selling and risk reduction. When a major shock hits and portfolio losses trigger margin calls or value-at-risk breaches, investors sell whatever is liquid — not what is fundamentally impaired. This means that safe-haven assets, high-quality stocks, and unrelated asset classes get sold alongside the originally impacted assets. The forced selling creates correlated price declines across fundamentally uncorrelated assets. The March 2020 COVID crisis demonstrated this dramatically: in the first two weeks of the selloff, even US Treasuries — the ultimate safe haven — sold off alongside equities, as investors liquidated everything for cash.\n\nValue-at-Risk (VaR) models and risk parity strategies are particularly vulnerable to correlation breakdown. VaR models typically use trailing correlation estimates, which are low during calm periods. When correlations spike, the model's risk estimate is suddenly too low, forcing additional selling to meet risk limits — which further increases correlations. Risk parity strategies, which allocate based on the risk contribution of each asset class (assuming diversification benefits from low correlations), are similarly caught: when correlations spike, the strategy is suddenly over-leveraged and must reduce positions across all assets simultaneously.\n\nFor crisis monitoring, the key metrics are: realized cross-asset correlations (rolling 30-day and 90-day), the dispersion of returns across sectors and asset classes (low dispersion = high correlation = dangerous), the VIX-correlation relationship (historically, VIX above 30 is associated with significantly elevated cross-asset correlations), and the performance of traditional safe havens during stress events (if Treasuries and gold fail to rally during equity selloffs, it signals correlation breakdown). The correlation structure itself is a measure of crisis severity — when it breaks, the crisis is more severe than the initial trigger would suggest, because the amplification mechanism of forced selling is engaged.",
      "dataSource": "Bloomberg (cross-asset correlation matrices), CBOE (implied correlation indices), portfolio analytics platforms, academic research (Longin and Solnik, Forbes and Rigobon)",
      "leadTime": "Near-zero lead time — correlation breakdown is a crisis amplification mechanism, not a predictor. However, indicators of vulnerability can be tracked: extremely low trailing correlations (indicating complacency), high portfolio leverage conditioned on low-correlation assumptions, and large assets under management in risk parity / VaR-constrained strategies.",
      "reliability": "Correlation breakdown during severe stress events is highly reliable — it happens in virtually every major crisis. The key uncertainty is threshold: moderate stress events may not trigger full correlation breakdown, while severe events consistently do. VIX above 35-40 has historically been a reliable indicator that correlation breakdown is occurring.",
      "historicalPrecedents": [
        "2008 GFC (everything sold simultaneously)",
        "March 2020 COVID (Treasuries sold alongside equities)",
        "1998 LTCM (convergence trades diverged simultaneously)",
        "2022 (stocks AND bonds down simultaneously for first time in decades)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "carry-trade",
        "systemic-banking",
        "credit-debt",
        "general"
      ],
      "tags": [
        "correlation-breakdown",
        "diversification-failure",
        "forced-selling",
        "var-breach",
        "risk-parity",
        "amplification",
        "cross-asset",
        "safe-haven-failure"
      ]
    },
    {
      "id": 27,
      "sourceId": "cm-028",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "asset-bubble",
      "title": "Dot-Com Bubble and Crash (2000-2002)",
      "content": "The dot-com bubble and subsequent crash is the most directly relevant historical precedent for the current AI investment cycle. Between 1995 and 2000, the NASDAQ Composite rose approximately 400%, driven by euphoric investment in internet-related companies. The crash from March 2000 to October 2002 destroyed $5 trillion in market value, with the NASDAQ declining 78% from peak to trough. Many prominent companies lost 90-99% of their value, and hundreds went bankrupt.\n\nThe parallels to the current AI cycle are striking and specific. Telecom companies in the late 1990s engaged in massive infrastructure buildouts — laying fiber optic cable, building data centers, deploying network equipment — based on exponential growth projections for internet traffic. The traffic projections were largely correct (internet usage did grow exponentially), but the revenue projections were wildly wrong (competition and commoditization meant that transmitting data became nearly free). Companies like WorldCom, Global Crossing, and Qwest spent tens of billions on infrastructure that generated far less revenue than projected. The current AI cycle features similar dynamics: the demand for compute is real, but the revenue per unit of compute may decline rapidly as models commoditize.\n\nVendor financing was a critical accelerant. Cisco, Nortel, and Lucent lent money to their customers to buy their equipment, creating circular revenue: the equipment makers booked revenue from sales financed by their own balance sheets. When the customers could not service the debt, the equipment makers faced simultaneous revenue collapse and loan losses. In the AI cycle, the hyperscaler dynamic is similar though not identical: Microsoft, Google, and Amazon are both the buyers of NVIDIA GPUs and the providers of cloud services that AI startups use to consume that compute. The self-referential revenue loop creates vulnerability.\n\nCritically, the dot-com crash did not destroy the internet — it destroyed the overvaluation and overcapacity. Amazon went from $107 to $7 per share and then eventually to over $3,000. The technology was real; the valuations were not. This precedent is essential for AI crisis analysis: a major AI market correction would not mean AI is a failure, any more than the dot-com crash meant the internet was a failure. It would mean that the financial market had overestimated the speed and magnitude of AI monetization, and that the resulting correction would be severe for investors but ultimately healthy for the technology's long-term development.",
      "dataSource": "Market data (NASDAQ historical), SEC filings, academic studies of dot-com crash (Ofek and Richardson, Brunnermeier and Nagel), journalistic accounts",
      "leadTime": "N/A — historical precedent. The precedent informs the AI bubble trigger analysis (cm-025) with specific parallels.",
      "reliability": "Well-documented historical event with extensive academic and journalistic analysis. The specific parallels (capex bubble, vendor financing, technology adoption curve vs. revenue curve) are genuine and analytically useful, though every cycle has unique elements.",
      "historicalPrecedents": [
        "1990s railroad bubble (infrastructure buildout ahead of revenue)",
        "South Sea Bubble (1720)",
        "1920s Radio/Electricity bubble"
      ],
      "compoundsWith": [
        "ai-specific",
        "asset-bubble",
        "credit-debt"
      ],
      "tags": [
        "dot-com",
        "nasdaq",
        "telecom-capex",
        "vendor-financing",
        "amazon-97-percent-drop",
        "ai-parallel",
        "infrastructure-overbuilt",
        "commoditization"
      ]
    },
    {
      "id": 28,
      "sourceId": "cm-029",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "credit-debt",
      "title": "2008 Global Financial Crisis",
      "content": "The 2008 Global Financial Crisis remains the defining crisis event of the modern era and the primary reference case for systemic risk analysis. The crisis demonstrated how the interaction of multiple risk factors — none of which was individually catastrophic — could produce a systemic meltdown when they occurred simultaneously in a highly leveraged, interconnected financial system.\n\nThe crisis had three interacting components. First, a housing bubble: US home prices had risen approximately 124% from 1997 to 2006, fueled by loose lending standards, securitization that divorced loan origination from credit risk, and a widespread belief that national home prices could not decline simultaneously. Second, excessive leverage: banks, investment banks, and shadow banking entities (SIVs, CDOs, money market funds) operated with leverage ratios of 30:1 to 50:1, meaning a 2-3% decline in asset values could wipe out equity. Third, systemic interconnection: the securitization chain connected mortgage borrowers to global banks through complex, opaque webs of derivatives (CDOs, CDO-squareds, CDS), meaning that losses propagated instantaneously across the entire financial system.\n\nThe crisis unfolded in a characteristic sequence that provides a template for future systemic crises. Phase 1 (early warnings, 2006-2007): subprime mortgage delinquencies rose, housing starts declined, and the ABX index (which tracked subprime mortgage-backed securities) began falling. These signals were dismissed by most market participants as 'contained.' Phase 2 (stress transmission, 2007-early 2008): BNP Paribas froze three funds, Bear Stearns collapsed, and interbank lending markets seized. The crisis was clearly no longer contained but was still viewed as a liquidity problem. Phase 3 (systemic collapse, September-October 2008): Lehman Brothers' bankruptcy triggered a global panic, money market funds 'broke the buck,' credit markets froze completely, and governments intervened with emergency bailouts and guarantees.\n\nThe recovery timeline validated Reinhart-Rogoff's predictions: US per capita GDP did not return to 2007 levels until 2013, a 6-year recovery. Government debt surged as fiscal stimulus replaced collapsed private spending. The policy response — near-zero interest rates, quantitative easing, bank bailouts — successfully prevented a depression but created new vulnerabilities (asset price inflation, moral hazard, zombie companies) that shaped the subsequent decade's crisis landscape. Every post-2008 crisis analysis tool and institutional framework was influenced by this event.",
      "dataSource": "FCIC (Financial Crisis Inquiry Commission) Report (2011), Bernanke/Geithner/Paulson accounts, academic literature (Gorton, Brunnermeier, Acharya), market data",
      "leadTime": "N/A — historical precedent. Key lesson: early warnings were present 18-24 months before the acute phase but were dismissed as 'contained.'",
      "reliability": "The most thoroughly documented and studied financial crisis in history. Lessons are well-established, though the specific crisis mechanism (securitized mortgage chain) is unlikely to repeat in the same form.",
      "historicalPrecedents": [
        "1930s Great Depression (closest precedent for 2008)",
        "1990s Japan bubble collapse",
        "1997 Asian Financial Crisis"
      ],
      "compoundsWith": [
        "credit-debt",
        "asset-bubble",
        "systemic-banking",
        "carry-trade"
      ],
      "tags": [
        "gfc",
        "2008",
        "lehman",
        "subprime",
        "securitization",
        "leverage",
        "systemic-collapse",
        "too-big-to-fail",
        "credit-contraction"
      ]
    },
    {
      "id": 29,
      "sourceId": "cm-030",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "carry-trade",
      "title": "1997 Asian Financial Crisis",
      "content": "The 1997 Asian Financial Crisis is the canonical example of how carry trade dynamics, currency mismatches, and contagion can transform a localized crisis into a regional or global event. The crisis began in Thailand in July 1997 when the baht's peg to the US dollar collapsed after the government exhausted its foreign exchange reserves defending it. Within months, the crisis spread to Indonesia, South Korea, Malaysia, and the Philippines, eventually triggering the 1998 Russian default and the LTCM collapse.\n\nThe pre-crisis setup featured the classic carry trade vulnerability: Southeast Asian economies offered high interest rates and strong growth prospects, attracting large capital inflows from international investors. Much of this capital was short-term and denominated in foreign currencies (primarily US dollars and Japanese yen). Local banks borrowed in dollars at lower rates and lent in local currency at higher rates, earning the spread. Corporations borrowed in dollars to fund domestic expansion. The currency peg regimes made this appear safe — with stable exchange rates, the currency mismatch seemed like free money.\n\nWhen confidence cracked, the self-reinforcing unwind was devastating. Capital outflows forced currency devaluations, which dramatically increased the local-currency cost of dollar-denominated debt. Companies and banks that had appeared solvent at the pegged exchange rate became insolvent at the depreciated rate. The desperate attempt to defend currencies by raising interest rates (recommended by the IMF at the time) crushed domestic economies without stopping the capital flight. Indonesia's GDP contracted 13.1% in 1998. Thailand and South Korea experienced 7-10% contractions.\n\nThe contagion mechanism is critical for modern crisis monitoring. The crisis spread not because Thailand's problems directly caused Indonesia's problems, but because (a) investors reassessed the entire region as sharing similar vulnerabilities (currency pegs, current account deficits, short-term foreign borrowing), and (b) losses in one country forced investors to sell positions in others to meet margin calls and redemptions. This 'common creditor' channel — where the same investors are exposed to multiple countries — remains the primary mechanism for crisis contagion and is directly applicable to today's interconnected global financial system.",
      "dataSource": "IMF crisis reports, BIS capital flow data, academic literature (Radelet and Sachs, Corsetti, Pesenti and Roubini), country-level economic data",
      "leadTime": "N/A — historical precedent. Key lesson: the vulnerabilities (current account deficits, short-term foreign borrowing, currency pegs) were identifiable years before the crisis but were dismissed because growth was strong.",
      "reliability": "Well-documented and extensively studied. The crisis clearly demonstrated carry trade risks, contagion dynamics, and the dangers of currency mismatches. Policy lessons influenced subsequent EM reserve accumulation and flexible exchange rate adoption.",
      "historicalPrecedents": [
        "1994 Mexican Tequila Crisis (precursor pattern)",
        "1992 ERM crisis (currency peg collapse)",
        "1998 Russian default (contagion from Asia)"
      ],
      "compoundsWith": [
        "carry-trade",
        "currency-inflation",
        "systemic-banking",
        "geopolitical"
      ],
      "tags": [
        "asian-crisis",
        "1997",
        "contagion",
        "carry-trade",
        "currency-peg",
        "imf",
        "capital-flight",
        "common-creditor"
      ]
    },
    {
      "id": 30,
      "sourceId": "cm-031",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "geopolitical",
      "title": "March 2026 Iran Geopolitical Shock",
      "content": "The March 2026 Iran tensions serve as a real-time case study of how modern monitoring systems track and differentiate the financial impact of geopolitical shocks. The event demonstrated several important principles for crisis monitoring: the speed at which geopolitical shocks transmit to financial markets, the differentiated impact across market segments, and the importance of compound analysis (what else was happening in markets when the shock hit).\n\nThe immediate market response followed the classic geopolitical shock pattern: oil prices spiked (directly reflecting supply disruption risk), emerging market currencies sold off (reflecting risk-off capital flows and dollar strengthening), equities declined (reflecting increased uncertainty and risk premiums), and traditional safe-haven assets rallied (US Treasuries, gold, Swiss franc). The speed of transmission was measured in minutes — algorithmic trading and 24-hour global markets ensure that geopolitical news is priced almost instantly.\n\nHowever, the differentiated impact revealed important structural information. Emerging markets with direct exposure to Middle Eastern trade or energy imports experienced significantly larger currency and equity declines than those with less exposure. US markets initially sold off but showed more resilience, partly because the US is a net energy producer (benefiting from higher oil prices in some sectors) and partly because the dollar's safe-haven status attracted capital flows. This differentiated response is analytically valuable — it reveals which markets and economies have structural vulnerabilities to specific geopolitical scenarios.\n\nThe most important lesson from the March 2026 event for the crisis monitoring framework is the interaction between geopolitical shocks and pre-existing financial conditions. The market impact was moderated by the fact that US equities had already experienced some correction earlier in 2026, reducing the 'air under prices.' Had the same geopolitical shock hit during a period of extreme overvaluation and complacency (say, at late-2024 levels), the market reaction might have been significantly larger because it would have punctured both the geopolitical assumption AND the valuation assumption simultaneously. This compound effect is why geopolitical monitoring must always be contextualized within the broader financial vulnerability landscape.",
      "dataSource": "Real-time market data, OFR FSI readings, CBOE VIX, oil futures, FX rates, news analysis",
      "leadTime": "N/A — real-time case study. Key lesson: geopolitical shocks are inherently unpredictable in timing but their market transmission patterns are predictable and informative.",
      "reliability": "Demonstrates that geopolitical shock impact is highly dependent on pre-existing financial conditions. The same shock has different severity depending on market positioning, leverage, and valuation levels at the time of impact.",
      "historicalPrecedents": [
        "1990 Gulf War oil spike",
        "2022 Russia-Ukraine (energy crisis, sanctions)",
        "2011 Arab Spring (EM contagion)"
      ],
      "compoundsWith": [
        "geopolitical",
        "energy-shock",
        "currency-inflation",
        "asset-bubble"
      ],
      "tags": [
        "iran-2026",
        "geopolitical-shock",
        "oil-spike",
        "differentiated-impact",
        "em-vs-us",
        "compound-analysis",
        "real-time-monitoring"
      ]
    },
    {
      "id": 31,
      "sourceId": "cm-032",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "asset-bubble",
      "title": "Meta 2022 Crash and Recovery Pattern",
      "content": "Meta Platforms' stock price trajectory from 2021 to 2024 — peaking above $380, crashing 75% to below $90, then recovering to all-time highs above $600 — provides a critical template for understanding how technology company bubbles and crashes play out at the individual stock level, with direct implications for how AI-related stocks might behave in a correction.\n\nThe crash was driven by a convergence of factors: slowing revenue growth (Apple's iOS privacy changes reduced ad targeting effectiveness), massive capex increases (the 'metaverse' pivot consumed $15B+ annually with no visible return), and a broader tech sector rotation as interest rates rose. At the bottom, the market was pricing Meta as if the metaverse investment would destroy the company — a thesis that proved completely wrong as the company's core advertising business remained enormously profitable.\n\nThe recovery pattern is the key lesson. Meta's leadership cut costs aggressively (the 'year of efficiency'), refocused on core advertising products, and the AI narrative actually benefited the company (Meta's open-source Llama models and AI ad optimization became market positives). The stock not only recovered but exceeded its previous highs. Investors who panic-sold at the bottom locked in 75% losses on what turned out to be a temporary — albeit severe — correction in a fundamentally strong business.\n\nThis pattern — panic sell the real thing, miss the enormous recovery — has profound implications for AI crisis monitoring. If and when an AI-related correction hits NVIDIA, Microsoft, Google, Amazon, or other AI leaders, the critical analytical question is: 'Is this a Meta 2022 pattern (temporary overreaction to real-but-manageable problems) or a Cisco 2000 pattern (permanent impairment of the business model)?' The answer determines whether the crash is a buying opportunity or the beginning of a lost decade. For investors in the AI-Rev audience, this historical precedent is essential context — it provides both the bear case (75% crashes happen to real companies) and the bull case (recovery to new highs is possible within 2 years if the underlying business is sound).",
      "dataSource": "Meta Platforms SEC filings, market data, company earnings calls, analyst reports",
      "leadTime": "N/A — historical precedent. Key lesson: the magnitude of tech stock corrections can be extreme (-75%) even for fundamentally sound businesses, and recovery can be equally dramatic.",
      "reliability": "Well-documented single-stock case study. Limitation: one company's experience does not generalize — for every Meta recovery, there is a Cisco, Intel, or Nokia that never recovered to its peak.",
      "historicalPrecedents": [
        "Amazon 2000-2001 ($107 to $7, eventual recovery to $3000+)",
        "Apple 1997 (near bankruptcy, eventual world's most valuable company)",
        "Cisco 2000 (crashed and never recovered to peak — counter-example)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific"
      ],
      "tags": [
        "meta",
        "75-percent-crash",
        "recovery",
        "panic-selling",
        "year-of-efficiency",
        "buying-opportunity-vs-permanent-impairment",
        "ai-correction-template"
      ]
    },
    {
      "id": 32,
      "sourceId": "cm-033",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "energy-shock",
      "title": "Oil Shock + Inflation + Policy Tightening Compound",
      "content": "The interaction of energy price shocks, elevated inflation, and restrictive monetary policy creates one of the most dangerous compound crisis scenarios — historically associated with the deepest and most prolonged economic downturns. Each component individually is manageable, but when all three coincide, policymakers face an impossible trilemma that eliminates the normal tools for crisis response.\n\nThe mechanism works as follows. An oil price shock directly raises CPI through higher gasoline, heating, and transportation costs, and indirectly raises costs across the economy through higher input prices for manufacturing, agriculture, and services. If inflation is already elevated before the oil shock (as it was in 1979 and partially in 2022), the combined effect pushes inflation well above the central bank's comfort zone. The central bank, committed to its inflation mandate, must either (a) continue tightening monetary policy, which crushes economic activity on top of the oil shock's contractionary effect, or (b) pause or ease policy, which risks de-anchoring inflation expectations and creating a wage-price spiral.\n\nOption (a) produces stagflation: high inflation + recession simultaneously. The 1979-1982 period is the canonical example — Volcker's Fed raised the federal funds rate to 20% to break inflation, causing the deepest recession since the 1930s but ultimately restoring price stability. Option (b) produces persistent inflation: the 1970s Fed (under Burns and Miller) tried to accommodate oil shocks with loose policy, resulting in a decade of elevated inflation that eroded living standards and required even more severe tightening to eventually resolve.\n\nFor crisis monitoring in 2026, this compound scenario is particularly relevant because: (a) inflation has proven stickier than expected in many developed economies, with services inflation remaining elevated; (b) geopolitical tensions (Iran, Russia-Ukraine, Middle East instability) create ongoing energy supply disruption risk; and (c) central banks have less room to cut rates than in pre-2022 cycles because they have already used much of their policy space. A major oil shock in this environment would face the worst-case compound: the Fed unable to cut because inflation is too high, unable to hike because the economy is already slowing, and unable to do nothing because markets would interpret inaction as policy paralysis.",
      "dataSource": "Historical analysis (1970s, 2008, 2022), FRED (oil prices, CPI, Fed funds rate), IMF World Economic Outlook scenarios",
      "leadTime": "Each component can be tracked individually: oil supply risk (days-weeks), inflation trends (months), policy trajectory (months). The compound risk can be assessed continuously by monitoring the overlap.",
      "reliability": "The compound interaction is historically proven — every instance of simultaneous energy shock + high inflation + tight policy has produced severe economic outcomes. The key uncertainty is whether the compound will actually occur (geopolitical events are unpredictable).",
      "historicalPrecedents": [
        "1973-75 oil embargo + inflation + Nixon-era tightening",
        "1979-82 Iranian Revolution + inflation + Volcker shock",
        "2022 Russia-Ukraine + inflation + Fed hiking cycle"
      ],
      "compoundsWith": [
        "energy-shock",
        "currency-inflation",
        "yield-curve",
        "credit-debt",
        "geopolitical"
      ],
      "tags": [
        "compound-crisis",
        "oil-shock",
        "stagflation",
        "policy-trilemma",
        "volcker",
        "impossible-choice",
        "services-inflation"
      ]
    },
    {
      "id": 33,
      "sourceId": "cm-034",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "credit-debt",
      "title": "Credit Freeze + Asset Bubble + Leverage Compound",
      "content": "The simultaneous occurrence of a credit market freeze, asset bubble deflation, and high leverage is the cocktail that produced the 2008 Global Financial Crisis — the most severe financial crisis since the Great Depression. Each component reinforces the others in a self-amplifying feedback loop that can overwhelm all normal market circuit breakers and stabilization mechanisms.\n\nThe feedback loop operates as follows. Asset bubble deflation (e.g., falling home prices in 2007-08, falling tech stock prices in 2000-01) destroys the collateral backing leveraged positions. When collateral values decline, leveraged investors face margin calls: they must either post additional collateral or sell assets to reduce leverage. The forced selling drives asset prices lower, destroying more collateral, triggering more margin calls — a classic doom loop. Simultaneously, the credit market freezes because lenders, observing falling collateral values and rising default rates, pull back from lending. New credit issuance drops to near zero, and existing credit lines are not renewed. This credit freeze prevents the normal market stabilization mechanism (bargain hunters buying cheap assets with borrowed money) from operating.\n\nThe leverage component is the amplifier that turns a manageable correction into a systemic crisis. At 10:1 leverage, a 10% asset decline wipes out equity. At 30:1 leverage (common among investment banks pre-2008), a 3.3% asset decline wipes out equity. The higher the leverage, the smaller the initial shock needed to trigger the doom loop. Pre-2008, many financial institutions were operating at 30:1 or higher leverage, meaning the financial system was essentially a loaded spring — any trigger, even a modest one, could set off the cascade.\n\nFor current crisis monitoring, the relevant metrics are: (a) aggregate leverage in the financial system (including shadow banking, which is estimated at $63T+ globally as of 2025); (b) asset valuations relative to fundamentals (CAPE, credit spreads, real estate cap rates); and (c) credit market functioning indicators (new issuance volumes, bid-ask spreads, repo market rates). When all three metrics flash warning — high leverage, overvalued assets, and tightening credit conditions — the compound risk is elevated regardless of what the specific trigger might be.",
      "dataSource": "Financial Stability Board (leverage data), FRED (credit spreads), BIS (shadow banking monitor), SEC (broker-dealer leverage), Bloomberg (credit issuance volumes)",
      "leadTime": "Leverage and valuation metrics are observable months to years in advance. Credit freeze onset is rapid (days to weeks). The compound risk can be monitored continuously.",
      "reliability": "The 2008 crisis provided definitive validation of this compound mechanism. However, post-2008 regulation (Dodd-Frank, Basel III) has reduced banking sector leverage significantly. The risk has partially migrated to non-bank financial institutions (private credit, hedge funds) where regulation is lighter.",
      "historicalPrecedents": [
        "2008 Global Financial Crisis (definitive example)",
        "1929-33 Great Depression (leverage + margin calls + bank failures)",
        "1998 LTCM (leverage + spread blowout + liquidity freeze)"
      ],
      "compoundsWith": [
        "credit-debt",
        "asset-bubble",
        "systemic-banking",
        "carry-trade"
      ],
      "tags": [
        "compound-crisis",
        "credit-freeze",
        "leverage",
        "doom-loop",
        "margin-calls",
        "forced-selling",
        "shadow-banking",
        "collateral-destruction"
      ]
    },
    {
      "id": 34,
      "sourceId": "cm-035",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "general",
      "title": "Multiple Trigger Simultaneous Activation",
      "content": "The ECB's research on the CISS indicator (cm-009) formalized an insight that is central to crisis monitoring: systemic stress is greatest not when any single market segment is under severe stress, but when stress prevails across multiple segments simultaneously and those segments are correlated. This principle, which the ECB captures through portfolio-theoretical weighting with time-varying cross-correlations, applies broadly to the interaction between crisis trigger categories.\n\nConsider the difference between a single-trigger and multi-trigger crisis. A single-trigger event — say, a geopolitical shock that raises oil prices — creates a one-dimensional stress pattern. Oil-sensitive sectors suffer, defense stocks rally, and the overall market impact is moderate because the shock is concentrated. A multi-trigger event — say, a geopolitical shock that raises oil prices WHILE the yield curve is inverted AND credit spreads are widening AND AI stocks are correcting — creates a multi-dimensional stress pattern where each trigger amplifies the others. Higher oil prices worsen the inflation outlook, making the inverted yield curve more ominous. Wider credit spreads indicate that the credit market is already fragile, so the additional oil shock may push marginal borrowers over the edge. AI stock correction reduces the equity market's ability to absorb broader losses.\n\nThe key insight is that the relationship between triggers is non-linear. Two simultaneous triggers are not twice as dangerous as one — they may be 3-5x as dangerous because each trigger impairs the system's ability to absorb the others. Three simultaneous triggers may be 10x as dangerous. This non-linearity is why the ECB CISS uses portfolio-theoretical weighting: the index spikes not when individual segments are stressed but when cross-segment correlations rise, because rising correlations indicate that the system's ability to absorb shocks through diversification is collapsing.\n\nFor practical crisis monitoring, this means tracking the trigger dashboard as a system rather than monitoring each trigger independently. The total number of triggers in 'elevated' or 'critical' status, and especially the correlation between trigger movements, provides a meta-indicator of systemic risk that is more informative than any individual trigger reading. A monitoring system should flag when three or more trigger categories are simultaneously above historical norms, even if none individually has reached crisis levels.",
      "dataSource": "Composite analysis across all trigger categories, ECB CISS methodology papers, BIS systemic risk research",
      "leadTime": "Depends on the specific triggers involved. The compound analysis adds no additional lead time but significantly increases the confidence and severity assessment of existing signals.",
      "reliability": "The principle is empirically validated across every major systemic crisis. The key finding — that crises are most severe when multiple triggers activate simultaneously — is essentially unanimous in the academic and institutional literature. Limitation: determining exact thresholds for 'elevated' across diverse trigger categories requires judgment.",
      "historicalPrecedents": [
        "2008 GFC (credit + housing + leverage + banking simultaneously)",
        "1997 Asian Crisis (currency + carry trade + banking simultaneously)",
        "1979-82 (oil + inflation + tight money simultaneously)"
      ],
      "compoundsWith": [
        "general",
        "credit-debt",
        "asset-bubble",
        "geopolitical",
        "energy-shock",
        "currency-inflation",
        "systemic-banking",
        "ai-specific",
        "carry-trade"
      ],
      "tags": [
        "compound-crisis",
        "non-linear",
        "multi-trigger",
        "ecb-ciss",
        "systemic",
        "cross-correlation",
        "diversification-failure",
        "meta-indicator"
      ]
    },
    {
      "id": 35,
      "sourceId": "cm-036",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "general",
      "title": "Supply Chain Disruption Cascade",
      "content": "Supply chain disruption cascades represent a compound crisis mechanism where a shock to one part of the production network propagates through input-output linkages to create widespread economic disruption. Unlike financial contagion (which spreads through credit and capital markets), supply chain cascades spread through the physical production of goods and services, making them resistant to financial policy interventions.\n\nThe COVID-19 pandemic (2020-2022) provided the most comprehensive real-world demonstration of supply chain cascade dynamics. Initial lockdowns disrupted production in specific sectors (auto manufacturing, electronics assembly, container shipping), but the effects cascaded rapidly. Auto manufacturers could not produce cars because semiconductor fabs in Asia had shut down. Semiconductor shortages then spread to every industry that uses chips — medical devices, appliances, industrial equipment, telecommunications. Container shipping bottlenecks created backlogs that raised costs across all imported goods. Each disruption created downstream effects that amplified the original shock.\n\nThe critical feature of supply chain cascades, as modeled by input-output analysis (cm-005), is that they concentrate rather than disperse as they move downstream. A 20% reduction in semiconductor production does not cause a 20% reduction in auto production — it causes a 100% stoppage for any model that requires the specific chip that is unavailable, while other models are unaffected. This 'all-or-nothing' property of production (you cannot make 80% of a car) means that even moderate supply disruptions can have disproportionate output effects.\n\nFor crisis monitoring, the key supply chain chokepoints to track include: semiconductor fabrication (TSMC in Taiwan produces ~55% of global semiconductor revenue and ~90% of advanced chips), rare earth processing (China controls ~60% of mining and ~90% of processing), energy infrastructure (concentrated pipeline routes, LNG terminals), and increasingly, cloud computing and AI training capacity (concentrated in a handful of hyperscaler data centers). A disruption to any of these chokepoints — whether from geopolitical events, natural disasters, or cyberattacks — would create cascading effects through the global production network.",
      "dataSource": "BEA Input-Output tables, World Input-Output Database, company supply chain disclosures, semiconductor industry data (SEMI, SIA), logistics data (Freightos Baltic Index, container ship tracking)",
      "leadTime": "Chokepoint vulnerability can be identified years in advance. Specific disruption triggers (geopolitical, natural disaster) may provide days to months of warning depending on the event type.",
      "reliability": "Supply chain cascade dynamics were dramatically validated during COVID-19. The mechanisms are well-understood. The key uncertainty is which chokepoints will be disrupted and when.",
      "historicalPrecedents": [
        "COVID-19 supply chain crisis (2020-2022)",
        "2011 Tohoku earthquake (auto/electronics supply chains)",
        "2021 Suez Canal blockage",
        "2022 semiconductor shortage"
      ],
      "compoundsWith": [
        "geopolitical",
        "energy-shock",
        "ai-specific",
        "general"
      ],
      "tags": [
        "supply-chain",
        "cascade",
        "chokepoints",
        "semiconductor",
        "tsmc",
        "rare-earth",
        "input-output",
        "all-or-nothing",
        "covid-validated"
      ]
    },
    {
      "id": 36,
      "sourceId": "cm-037",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "ai-specific",
      "title": "AI Bubble Deflation + Broader Market Stress Compound",
      "content": "The most dangerous scenario for the current market environment is the deflation of the AI investment bubble occurring simultaneously with broader financial market stress. This compound scenario would combine the sector-specific dynamics of AI correction with system-wide risk amplification, creating a crisis significantly larger than either component alone.\n\nThe AI bubble exists within the broader equity market, not separate from it. The 'Magnificent 7' (or their 2026 equivalent) represent 25-35% of the S&P 500 by weight. A significant correction in these stocks — say, 30-50% declines driven by AI revenue disappointment — would mechanically drag the entire S&P 500 down by 8-17% even if every other stock was unchanged. Index funds, ETFs, and passive strategies would be forced to sell precisely because the AI stocks are falling, regardless of the fundamental merits of the other 493 stocks. This mechanical index concentration risk means an AI correction is automatically a broad market event.\n\nNow layer on broader financial stress: rising credit spreads, an inverted yield curve, a banking sector weakened by CRE exposure, or a geopolitical shock. In this environment, the AI correction does not just cause a market decline — it triggers the compound mechanisms described in other entries: margin calls (cm-034), correlation breakdown (cm-027), carry trade unwinding (cm-026), and potential credit freezes. Investors who were using appreciated AI stocks as collateral for other positions face margin calls. Risk parity strategies that assumed low correlation between tech and other sectors find their assumptions breaking. Corporate borrowers in the AI ecosystem (startups, data center builders, GPU financing) face rollover difficulties as lenders reassess the sector.\n\nThe timing intersection is critical. If AI revenue disappointment becomes apparent during a period of otherwise healthy markets and low stress, the correction is likely to be a sector rotation — painful for AI investors but absorbed by the broader market. If AI disappointment coincides with or triggers broader financial stress (because the market had been relying on AI exceptionalism to justify overall valuations), the correction becomes systemic. Monitoring for this compound requires tracking both AI-specific indicators (capex-to-revenue gap, GPU utilization, AI stock concentration) AND broader financial system indicators (stress indices, credit conditions, leverage) simultaneously.",
      "dataSource": "Cross-referencing AI-specific data (cm-025) with composite stress indices (cm-006 through cm-011) and trigger monitors (cm-019 through cm-027)",
      "leadTime": "The structural vulnerability is identifiable now. The timing of the compound event depends on when AI revenue disappoints AND when broader stress appears. Either could come first; the crisis risk peaks when both coincide.",
      "reliability": "High confidence that the compound scenario would be significantly worse than either component alone. This is supported by the historical pattern of bubble deflation + macro stress compounds (dot-com crash during 2001 recession, housing crash during 2008 credit crisis). Low confidence in timing.",
      "historicalPrecedents": [
        "2000-02 dot-com crash + recession (tech bubble + macro weakness)",
        "2008 housing crash + credit crisis (asset bubble + financial system stress)",
        "1929 stock market crash + bank failures (market bubble + banking crisis)"
      ],
      "compoundsWith": [
        "ai-specific",
        "asset-bubble",
        "credit-debt",
        "systemic-banking",
        "carry-trade",
        "general"
      ],
      "tags": [
        "ai-compound",
        "mag-7-concentration",
        "index-mechanical-risk",
        "simultaneous-stress",
        "sector-rotation-vs-systemic",
        "worst-case-scenario"
      ]
    },
    {
      "id": 37,
      "sourceId": "cm-038",
      "section": "institutional-approaches",
      "category": "institutional-investor",
      "subcategory": "general",
      "title": "Bridgewater Associates: Debt Cycle / Risk Parity Approach",
      "content": "Bridgewater Associates, founded by Ray Dalio, manages approximately $150 billion and has developed the most systematized approach to crisis prediction and navigation among institutional investors. Their methodology centers on two pillars: the debt cycle framework (see cm-002) applied through a rules-based system, and the All Weather / Risk Parity portfolio construction methodology designed to perform across all economic environments.\n\nBridgewater's Pure Alpha fund makes systematic bets across global markets based on Dalio's economic machine model. The system tracks over 300 decision rules that monitor credit conditions, monetary policy, growth indicators, and balance of payments data across 20+ countries. Each rule captures a specific economic relationship (e.g., 'when real interest rates exceed nominal GDP growth for X quarters, reduce credit exposure'). The system's track record validates the approach: Pure Alpha gained during the 2008 GFC and reportedly returned +26.2% year-to-date through September 2025, demonstrating its ability to navigate both crises and complex macro environments.\n\nThe alpha-beta separation concept is Bridgewater's key contribution to institutional portfolio construction. Beta is the return from passive market exposure (e.g., being long stocks or bonds). Alpha is the return from active decision-making (e.g., timing credit cycles, trading relative value). Bridgewater argues that most institutional portfolios are dominated by a single beta (US equities) disguised as diversification, and that true risk management requires explicitly separating and balancing beta exposures across economic environments (growth rising/falling, inflation rising/falling).\n\nFor crisis monitoring, Bridgewater's approach provides several actionable insights. Their published research (available through bridgewater.com) regularly identifies late-cycle indicators and vulnerability assessments. Their framework's emphasis on the 'long-term debt cycle' suggests we are in the late stages of a multi-decade credit expansion, where each recession requires increasingly extraordinary policy responses (lower rates, more QE, larger fiscal deficits), and the policy toolkit is becoming exhausted. Their All Weather framework also implies that the traditional 60/40 stock/bond portfolio provides much less crisis protection than investors believe, because both stocks and bonds can decline simultaneously in inflationary environments (as demonstrated in 2022).",
      "dataSource": "Bridgewater Associates published research, Dalio's books and LinkedIn posts, investor letters (some publicly available), 13F filings (SEC)",
      "leadTime": "The systematized approach provides continuous forward-looking assessment. Bridgewater's framework is designed to anticipate regime changes months to quarters in advance.",
      "reliability": "Among the strongest track records in the hedge fund industry, particularly during crisis periods. Limitation: specific positions and signals are proprietary and not available in real-time. Published research provides framework-level insights but not trading signals.",
      "historicalPrecedents": [
        "2008 GFC (Pure Alpha gained while markets crashed)",
        "2010-12 European crisis (successfully navigated)",
        "2020 COVID (mixed early performance, strong recovery)",
        "2025 (+26.2% YTD through Sep)"
      ],
      "compoundsWith": [
        "credit-debt",
        "yield-curve",
        "currency-inflation",
        "general"
      ],
      "tags": [
        "bridgewater",
        "dalio",
        "pure-alpha",
        "risk-parity",
        "all-weather",
        "300-decision-rules",
        "alpha-beta-separation",
        "debt-cycle"
      ]
    },
    {
      "id": 38,
      "sourceId": "cm-039",
      "section": "institutional-approaches",
      "category": "institutional-investor",
      "subcategory": "general",
      "title": "Universa Investments: Tail Risk Hedging",
      "content": "Universa Investments, founded by Mark Spitznagel and advised by Nassim Nicholas Taleb, practices a pure tail-risk hedging strategy that is designed to profit dramatically during market crashes while accepting small steady losses during normal markets. Their approach is philosophically opposed to continuous risk measurement (VaR, stress indices) and instead focuses on positioning for extreme events — the 'black swans' that conventional risk models assign negligible probability.\n\nUniversa's strategy centers on maintaining a portfolio of deep out-of-the-money put options and other instruments that are cheap during calm markets (because the probability of extreme events is priced as very low) but become enormously valuable during crashes. The VIX term structure (cm-018) is a key monitoring input: when the VIX term structure is in steep contango (short-term vol much lower than long-term vol), puts are relatively cheap and the hedge is easy to maintain. When the term structure flattens or inverts, it signals that near-term crash risk is already being priced, and the hedge is more expensive to maintain.\n\nThe strategy's reported returns are extraordinary during crisis periods. Universa's tail hedge program reportedly gained over 3,600% during the March 2020 COVID crash — a return that, even at a small portfolio allocation (1-3%), would have offset significant losses on the remaining portfolio. The key insight is that tail risk hedging does not try to predict crashes — it simply ensures that the portfolio benefits enormously when crashes occur, regardless of the cause.\n\nFor the crisis monitoring framework, Universa's approach offers several principles. First, the most important variable is not 'what will the crisis be?' but 'how convex is my exposure to crisis events?' Traditional monitoring tries to predict the specific crisis; Universa's approach makes prediction unnecessary by maintaining asymmetric exposure. Second, the cost of protection varies dramatically over time — protection is cheapest precisely when complacency is highest (low VIX, steep contango), which is when most investors feel they least need it. Third, the strategy validates the view that extreme events are more common than Gaussian models predict — the 'fat tails' documented by Mandelbrot and Taleb are a persistent feature of financial markets, not an anomaly.",
      "dataSource": "Universa investor letters (limited public availability), Taleb's published works, CBOE options data, VIX term structure data",
      "leadTime": "Universa's approach explicitly does not attempt to predict crisis timing. The framework is about maintaining persistent hedge exposure rather than timing entry/exit.",
      "reliability": "Extremely effective during severe crashes (March 2020 is the prime example). The strategy's limitation is the steady cost of carry during non-crisis periods — puts expire worthless during calm markets. The net portfolio return depends on how frequently crashes occur and how severe they are. Over a full cycle, the strategy can be highly effective or costly depending on crash frequency.",
      "historicalPrecedents": [
        "March 2020 COVID crash (+3,600% on hedge)",
        "2008 GFC (Universa predecessors profited)",
        "2018 Volmageddon (February VIX spike)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "carry-trade",
        "general"
      ],
      "tags": [
        "universa",
        "tail-risk",
        "black-swan",
        "taleb",
        "spitznagel",
        "put-options",
        "convexity",
        "vix-term-structure",
        "crash-protection"
      ]
    },
    {
      "id": 39,
      "sourceId": "cm-040",
      "section": "institutional-approaches",
      "category": "institutional-investor",
      "subcategory": "general",
      "title": "AQR Capital Management: Risk Decomposition Approach",
      "content": "AQR Capital Management, founded by Cliff Asness, applies systematic quantitative analysis to decompose investment returns and risks into their component parts. Their research has produced several insights directly relevant to crisis monitoring: the distinction between genuine alpha and disguised beta, the identification of hidden risk factors in seemingly diversified portfolios, and the study of momentum crashes.\n\nAQR's risk decomposition framework identifies several common sources of 'hidden' risk in institutional portfolios. Option writing risk: many strategies that appear to generate steady alpha are actually selling volatility (collecting premiums by writing options that are rarely exercised). These strategies perform well in 95% of environments but suffer catastrophic losses during the 5% of extreme events — essentially, they are short tail risk. Leverage amplification: some strategies achieve their returns through high leverage on low-volatility assets. When volatility unexpectedly spikes, the leveraged position generates outsized losses. Carry risk: earning income from interest rate differentials, credit spreads, or volatility premiums appears as alpha but is actually compensation for bearing specific crash risks.\n\nAQR's research on momentum crashes is particularly relevant for crisis monitoring. Momentum strategies (buying recent winners, selling recent losers) are one of the most persistent and well-documented return factors in finance. However, momentum is vulnerable to sudden 'crashes' where recent winners plummet and recent losers surge — typically during sharp market reversals or regime changes. The March 2009 market bottom was accompanied by a devastating momentum crash, as beaten-down financial stocks surged while momentum-favored defensive stocks lagged. These momentum crashes are both a consequence of crisis (forced selling of winners to meet margin calls) and an amplifier of crisis (momentum-following strategies dumping positions accelerate price moves).\n\nFor the crisis monitoring framework, AQR's contribution is the discipline of asking 'what risk is this return compensating me for?' When any asset or strategy is performing unusually well, AQR's framework asks whether the return reflects genuine skill or simply the accumulation of hidden risk that will be realized during a crisis. Applied to the current AI investment environment: are AI stock returns reflecting genuine value creation, or are they partly the result of carry-trade-like dynamics (borrowing at low rates to fund capex) and momentum (buying because price is rising) that will reverse violently when conditions change?",
      "dataSource": "AQR published research (aqr.com/insights), academic papers by Asness et al., 13F filings, Journal of Portfolio Management articles",
      "leadTime": "AQR's risk decomposition is a structural analysis tool rather than a timing tool. It identifies vulnerabilities that may persist for extended periods before materializing as losses.",
      "reliability": "AQR's academic rigor is among the highest in the hedge fund industry. Their risk decomposition framework has been validated across multiple crisis periods. Limitation: knowing that a strategy carries hidden crash risk does not tell you when the crash will occur.",
      "historicalPrecedents": [
        "March 2009 momentum crash (one of the largest in history)",
        "2020 'value crash' (value factors suffered extreme drawdown)",
        "2007 quant meltdown (factor crowding + deleveraging)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "carry-trade",
        "credit-debt",
        "general"
      ],
      "tags": [
        "aqr",
        "asness",
        "risk-decomposition",
        "alpha-vs-beta",
        "momentum-crash",
        "hidden-risk",
        "volatility-selling",
        "leverage-amplification"
      ]
    },
    {
      "id": 40,
      "sourceId": "cm-041",
      "section": "institutional-approaches",
      "category": "institutional-investor",
      "subcategory": "asset-bubble",
      "title": "GMO: Long-Term Valuation Reversion Approach",
      "content": "GMO (Grantham, Mayo, Van Otterloo), co-founded by Jeremy Grantham, has built its reputation on identifying bubbles and forecasting long-term asset class returns. Grantham has publicly called major bubbles — including the dot-com bubble (1999), the housing bubble (2007), and what he termed the 'superbubble' in 2022 — with a track record that, while not perfect on timing, has been remarkably accurate on direction and magnitude.\n\nGMO's methodology centers on one core principle: asset prices revert to fair value over 7-year periods. They define fair value using long-term averages of fundamental metrics (earnings yield, dividend yield, profit margins) adjusted for the current economic environment. Their 7-Year Asset Class Return Forecast, published quarterly, projects annualized real returns for major asset classes based on how far current prices deviate from fair value. When the forecast shows deeply negative expected returns for an asset class, it indicates overvaluation that GMO expects to correct over the coming 7 years.\n\nGMO's framework has produced specific insights relevant to the current environment. Their research on 'quality' stocks (companies with high and stable profitability, strong balance sheets, low investment needs) shows that these stocks have historically offered both higher returns and lower crash exposure than the broader market — providing a form of natural crisis protection. Their analysis of profit margin sustainability argues that the historically elevated profit margins of the 2010s-2020s are partially temporary (driven by low rates, low taxes, globalization) and will face mean-reversion pressure, which would reduce the fair value of equity markets.\n\nThe patient capital philosophy is itself a crisis insight. GMO has historically been early — sometimes years early — in identifying overvaluation and positioning conservatively. This means they underperform during the late stages of bubbles (when the last 30-50% of gains often occur) but dramatically outperform during and after crashes. For individual investors, the GMO lesson is that being early is not the same as being wrong, and that the opportunity cost of early caution (missing bubble gains) is much lower than the cost of late enthusiasm (riding the bubble into the crash). For the crisis monitoring framework, GMO's 7-Year Forecast provides a structural valuation assessment that complements shorter-range indicators.",
      "dataSource": "GMO quarterly forecasts (publicly available at gmo.com), Grantham's quarterly letters, GMO research papers, academic work on long-term return prediction",
      "leadTime": "7-year forecast horizon. GMO's framework explicitly does not predict timing — it predicts the direction and approximate magnitude of 7-year forward returns.",
      "reliability": "Strong long-term track record on directional calls and return forecasting. The 7-Year Forecast has correctly identified periods of overvaluation and undervaluation across multiple asset classes. Limitation: the 7-year horizon means you can be 'right' but suffer significant interim losses while waiting for reversion.",
      "historicalPrecedents": [
        "1999-2000 dot-com call (correct, early by ~18 months)",
        "2007 housing/credit call (correct, early by ~12 months)",
        "2022 'superbubble' call (partially realized, ongoing)",
        "2009 EM value call (correct, EM outperformed)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "credit-debt",
        "ai-specific"
      ],
      "tags": [
        "gmo",
        "grantham",
        "valuation-reversion",
        "7-year-forecast",
        "patient-capital",
        "quality-stocks",
        "mean-reversion",
        "bubble-calling"
      ]
    },
    {
      "id": 41,
      "sourceId": "cm-042",
      "section": "emerging-methods",
      "category": "emerging-method",
      "subcategory": "general",
      "title": "Machine Learning for Crisis Prediction",
      "content": "Machine learning approaches to crisis prediction represent a significant methodological frontier, offering the potential to capture non-linear relationships and complex interaction effects that traditional econometric models miss. ML models for financial crisis prediction typically use supervised learning: training on historical crisis episodes to identify patterns in pre-crisis data that precede systemic events. The input features span the full range of indicators tracked in this knowledge base — yield curves, credit spreads, asset valuations, capital flows, sentiment measures — and the models learn which combinations of feature values are most predictive of crisis onset.\n\nThe primary advantage of ML models is their ability to capture non-linear and interaction effects. Traditional models (logistic regression, probit models) assume specific functional forms for how variables relate to crisis probability. ML models (random forests, gradient boosting, neural networks) make no such assumptions and can discover complex, non-obvious patterns. For example, an ML model might learn that yield curve inversion is only predictive of crisis when combined with elevated credit-to-GDP gap AND low VIX — a three-way interaction that would be difficult to specify in a traditional model.\n\nHowever, ML approaches face significant challenges in this domain. The fundamental problem is small sample size: there have been approximately 15-20 systemic financial crises in developed markets since 1970. This is far too few observations to train complex models without severe overfitting risk. Models may learn the specific features of past crises rather than the general principles underlying them, and then fail on novel crisis types. The moral hazard risk is also real: if a widely adopted ML model identifies specific conditions as 'safe,' market participants may take more risk during those conditions, creating the very vulnerability the model says does not exist (a financial version of Goodhart's Law).\n\nThe most promising ML applications for crisis monitoring are not pure prediction but augmented analysis: using ML to identify anomalous patterns in high-dimensional data that warrant human investigation, to improve the speed and accuracy of nowcasting (estimating current conditions from partial data), and to stress-test portfolios against a wider range of scenarios than human analysts might consider. These applications leverage ML's strengths (pattern recognition, high-dimensional data processing) while mitigating its weaknesses (small samples, inability to anticipate truly novel events).",
      "dataSource": "Academic research (Journal of Financial Economics, Journal of Banking & Finance, IMF working papers), BIS Innovation Hub, OFR research, fintech company publications",
      "leadTime": "Depends on the specific model. ML models trained on monthly data typically produce 3-12 month forecasts. Higher-frequency models (daily/weekly) produce shorter-range forecasts. Reported lead times in academic studies range from 1 month to 2 years.",
      "reliability": "In backtesting, ML models often outperform traditional models. Out-of-sample performance is less clear, as few models have been tested through multiple crisis cycles. The field is actively developing and no consensus 'best model' has emerged.",
      "historicalPrecedents": [
        "N/A — emerging method. Some models have been backtested against 2008 GFC, 2020 COVID, and European sovereign crisis with promising but not definitive results."
      ],
      "compoundsWith": [
        "general"
      ],
      "tags": [
        "machine-learning",
        "ai-prediction",
        "non-linear",
        "overfitting-risk",
        "small-sample",
        "goodharts-law",
        "augmented-analysis",
        "emerging"
      ]
    },
    {
      "id": 42,
      "sourceId": "cm-043",
      "section": "emerging-methods",
      "category": "emerging-method",
      "subcategory": "general",
      "title": "High-Frequency Nowcasting",
      "content": "High-frequency nowcasting uses weekly or even daily data to produce real-time estimates of economic conditions, dramatically outperforming the monthly or quarterly data that official statistics agencies provide. During crises — when conditions change rapidly and policy decisions must be made with urgency — this speed advantage becomes enormously valuable. Traditional GDP data is released with a 4-8 week lag; nowcasting models can estimate current-quarter GDP in real-time.\n\nThe methodology combines high-frequency data sources (weekly unemployment claims, daily credit card transaction volumes, real-time freight shipping data, electricity consumption, web search trends) with econometric models (dynamic factor models, mixed-frequency VAR) that relate these timely indicators to slower-moving official statistics. The key innovation is bridging the frequency gap: using data that is available daily or weekly to infer the current state of monthly or quarterly variables.\n\nDuring the early stages of the COVID-19 crisis, high-frequency nowcasting models detected the economic collapse 1-3 months before official statistics confirmed it. Traditional GDP showed a small decline; credit card spending data showed a 30%+ collapse in restaurant and travel spending within days. Weekly unemployment claims spiked to levels that monthly employment surveys would not capture for weeks. This speed advantage allowed policymakers and investors who were monitoring high-frequency data to respond much faster than those relying on official statistics.\n\nFor crisis monitoring, the practical implication is that weekly data should be weighted heavily during stress periods. The New York Fed's Weekly Economic Index, the Federal Reserve Bank of Cleveland's nowcasting model, and the Atlanta Fed's GDPNow are all publicly available high-frequency nowcasting tools. During normal economic conditions, these tools provide modest informational advantage over monthly data. During crises, they provide a massive informational advantage — the difference between knowing the economy is collapsing in real-time versus learning about it 6 weeks later. The crisis monitoring system should automatically increase its weighting of high-frequency data when stress indices cross into elevated territory.",
      "dataSource": "New York Fed Weekly Economic Index, Atlanta Fed GDPNow, Cleveland Fed nowcasting model, Google Trends, credit card transaction data (anonymized), freight data (Cass Freight Index), electricity consumption (EIA weekly)",
      "leadTime": "1-3 months ahead of official statistics. This is not prediction of the future — it is faster measurement of the present, which effectively functions as a 1-3 month lead on official data releases.",
      "reliability": "Demonstrated strong performance during COVID-19 crisis (the most severe rapid-onset economic event in modern history). Less tested during slower-developing crises where weekly data may not provide as much advantage. Noise is higher than monthly/quarterly official data.",
      "historicalPrecedents": [
        "COVID-19 March 2020 (nowcasting detected collapse weeks before official data)",
        "2008 GFC (less dramatic advantage because the crisis developed over months rather than weeks)"
      ],
      "compoundsWith": [
        "general"
      ],
      "tags": [
        "nowcasting",
        "high-frequency",
        "weekly-data",
        "real-time",
        "gdpnow",
        "wei",
        "speed-advantage",
        "covid-validated"
      ]
    },
    {
      "id": 43,
      "sourceId": "cm-044",
      "section": "emerging-methods",
      "category": "emerging-method",
      "subcategory": "systemic-banking",
      "title": "Temporal Graph Learning for Default Prediction",
      "content": "Temporal graph learning represents the cutting edge of network-based financial risk analysis, combining the network contagion models (cm-004) with modern deep learning techniques to create dynamic, evolving representations of financial system risk. Unlike static network models that analyze a snapshot of financial interconnections, temporal graph learning models the evolution of the network over time — how connections form, strengthen, weaken, and break — and uses this dynamic structure to predict future defaults and systemic events.\n\nThe methodology works by representing the financial system as a graph (nodes = institutions, edges = financial relationships) and then applying graph neural network (GNN) architectures that can learn from both the network structure and the temporal evolution of node and edge features. Each institution-node carries features (balance sheet data, stock price, CDS spreads, credit ratings) that change over time. Each relationship-edge carries features (exposure size, collateral quality, maturity profile) that also evolve. The model learns to identify network patterns that precede default events — for example, a cluster of institutions whose mutual exposures are increasing while their individual credit metrics are deteriorating.\n\nThe key innovation over static network models is the ability to capture temporal dynamics: not just 'how connected is the system?' but 'how is the connectivity changing?' A financial system where interconnections are rapidly increasing (institutions taking on more counterparty exposure) while individual institution health is deteriorating is much more dangerous than a system with stable interconnections and stable health — but a static model cannot distinguish between the two states. Temporal graph learning can.\n\nResearch in this area, published in venues like NeurIPS, ICML, and the Journal of Financial Economics, has shown promising results in default prediction tasks. Models trained on historical corporate and bank default data achieve significantly better prediction accuracy than traditional credit scoring models, particularly for predicting cascading defaults (where one institution's failure triggers others). The limitation is data availability: detailed counterparty exposure data is largely unavailable to researchers (it is held by central banks and regulators under confidentiality). Models trained on publicly available data (CDS spreads, stock returns, credit ratings) capture some but not all of the network dynamics.",
      "dataSource": "Academic research papers, OFR working papers, central bank research departments, fintech company publications. Underlying data: publicly available market data (limited) or confidential supervisory data (accessible only to regulators).",
      "leadTime": "Research models have demonstrated 1-6 month lead times for individual default prediction. System-wide crisis prediction lead times are less well-established.",
      "reliability": "Promising but unproven at scale. Out-of-sample backtesting shows improvement over traditional models, but no temporal graph learning model has been tested through a real-time systemic crisis. The field is approximately 3-5 years from potential production deployment at major institutions.",
      "historicalPrecedents": [
        "N/A — emerging method. Backtested against 2008 GFC and 2023 banking stress with encouraging but preliminary results."
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt"
      ],
      "tags": [
        "temporal-graph-learning",
        "gnn",
        "network-dynamics",
        "deep-learning",
        "default-prediction",
        "cascading-default",
        "emerging",
        "cutting-edge"
      ]
    },
    {
      "id": 44,
      "sourceId": "cm-045",
      "section": "emerging-methods",
      "category": "emerging-method",
      "subcategory": "general",
      "title": "Alternative Data Sources for Crisis Detection",
      "content": "Alternative data — information sources outside traditional financial and economic data — represents a growing frontier in crisis detection and monitoring. These sources include social media sentiment analysis, news text analytics, satellite imagery, real-time transaction data, web scraping, shipping and logistics data, and environmental/weather data. The appeal is that alternative data can provide signals faster and with different informational content than traditional data sources.\n\nSocial media and news sentiment analysis uses natural language processing (NLP) to quantify the tone and intensity of financial commentary across platforms (Twitter/X, Reddit, financial news, analyst reports). Research has shown that aggregate sentiment shifts — particularly sudden increases in negative sentiment or uncertainty language — can precede market moves by hours to days. The Caldara-Iacoviello Geopolitical Risk Index (used in cm-021) is essentially a news sentiment analysis tool applied to geopolitical content. More granular sentiment analysis can track fear/uncertainty around specific topics (e.g., 'bank run' mentions, 'credit freeze' language, 'AI bubble' discussion).\n\nSatellite imagery and physical data provide economic signals that are harder to manipulate than self-reported data. Satellite tracking of oil tanker movements reveals supply flows before official reports. Nighttime light intensity correlates with economic activity, particularly in countries with unreliable official statistics. Retail parking lot occupancy tracks consumer spending. Construction activity visible from space tracks infrastructure investment. These sources are particularly valuable for monitoring emerging markets where official statistics may be delayed, inaccurate, or manipulated.\n\nReal-time transaction data, primarily from credit card processors and payment networks, provides the most granular available measure of economic activity. During COVID-19, transaction data showed the spending collapse within days, while traditional economic indicators lagged by weeks. For crisis monitoring, sudden shifts in spending patterns — particularly sharp declines in discretionary categories (restaurants, travel, entertainment) — provide an early signal of consumer confidence deterioration that precedes traditional survey data.\n\nThe limitations of alternative data are significant. Much of it is proprietary and expensive, available only to institutional investors who can afford the data feeds. The signal-to-noise ratio is often low, requiring sophisticated processing to extract meaningful signals. Historical backtesting is limited because most alternative data sources only became available in the 2010s. And there is a real risk of 'data mining' — finding spurious correlations in high-dimensional alternative data that do not hold out of sample. Despite these limitations, alternative data is increasingly integrated into institutional crisis monitoring systems as a complement to traditional indicators.",
      "dataSource": "Twitter/X API, Reddit API, Google Trends, news aggregators (GDELT, EventRegistry), satellite imagery providers (Planet Labs, Maxar), credit card transaction data (aggregated/anonymized), AIS ship tracking data",
      "leadTime": "Hours to weeks ahead of traditional data, depending on the source. Social sentiment: hours to days. Transaction data: days. Satellite: days to weeks. The advantage is speed and granularity rather than long-range forecasting.",
      "reliability": "Nascent and highly variable. Some alternative data sources have demonstrated clear predictive value (transaction data, shipping data). Others (social media sentiment) have mixed track records. The field is rapidly evolving.",
      "historicalPrecedents": [
        "COVID-19 (credit card data and mobility data detected economic collapse before official statistics)",
        "2022 Russia-Ukraine (satellite imagery tracked military buildup before invasion)",
        "Various EM crises (nighttime light data revealed economic deterioration before official GDP revisions)"
      ],
      "compoundsWith": [
        "general",
        "geopolitical"
      ],
      "tags": [
        "alternative-data",
        "satellite",
        "sentiment-analysis",
        "nlp",
        "transaction-data",
        "google-trends",
        "ship-tracking",
        "emerging",
        "signal-to-noise"
      ]
    },
    {
      "id": 45,
      "sourceId": "cm-046",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "general",
      "title": "Industries Most Vulnerable to Economic Crisis",
      "content": "Analysis drawn from the AI-Rev engine's Global Economic Crisis scenario reveals which of the 28 tracked industries are most vulnerable to a compound crisis event (simultaneous market crash + inflation spike + oil shock + geopolitical conflict + credit freeze). The vulnerability ranking differs significantly from normal economic sensitivity because a compound crisis attacks multiple channels simultaneously — demand destruction, supply disruption, credit withdrawal, and confidence collapse all hit at once.\n\nThe most vulnerable industries fall into several categories. Cyclical/discretionary industries: retail, entertainment/media, real estate, and consumer goods suffer immediately as consumers cut spending. These industries have thin margins, high operating leverage, and revenue that correlates directly with consumer confidence — all of which collapse during a compound crisis. High-leverage industries: telecommunications, airlines/logistics, and utilities carry heavy debt loads; during a credit freeze, their inability to roll over debt creates solvency risk even if the underlying business is viable. Capital-intensive industries: semiconductors, automotive/transport, and manufacturing face the double hit of demand collapse + supply chain disruption + inability to fund operations.\n\nThe AI-adjacent industries face a unique compound vulnerability. Cloud computing and data center operators have committed to massive capex programs funded by expected future revenue growth. In a compound crisis, enterprise customers cut IT spending (demand destruction), credit markets tighten (funding the capex becomes harder), and the AI revenue growth that justified the investment decelerates or reverses. Semiconductor companies face the same dynamic one step upstream. The NVIDIA supply chain (TSMC, ASML, memory makers) is particularly exposed because AI chips are high-ASP, discretionary capex items — the first thing enterprises cut in a downturn.\n\nFinancial services faces disproportionate compound crisis risk because it sits at the intersection of credit freeze (direct operational impact), asset bubble deflation (portfolio losses), and contagion dynamics (counterparty and common-asset exposure). The March 2023 banking stress showed how quickly confidence can evaporate in the financial sector. In a full compound crisis, the banking system's role as credit transmission mechanism means its dysfunction amplifies the crisis across all other industries.",
      "dataSource": "AI-Rev engine crisis scenario analysis (28-industry model), historical sector performance during crises, BEA input-output tables, Bloomberg sector returns during stress events",
      "leadTime": "Vulnerability is structural and identifiable now. The timing depends on when a compound crisis actually materializes.",
      "reliability": "Historical sector performance during crises is well-documented and the patterns are consistent. The AI-Rev engine's scenario analysis is based on these historical patterns plus cross-industry effect modeling. Limitation: each crisis is different, and sector vulnerability rankings can shift based on the specific mix of triggers.",
      "historicalPrecedents": [
        "2008 GFC (financials, real estate, autos most impacted)",
        "2020 COVID (travel, hospitality, entertainment most impacted)",
        "2001 dot-com (tech, telecom most impacted)",
        "1973-75 oil crisis (energy-intensive industries most impacted)"
      ],
      "compoundsWith": [
        "credit-debt",
        "asset-bubble",
        "energy-shock",
        "systemic-banking",
        "ai-specific"
      ],
      "tags": [
        "industry-vulnerability",
        "ai-rev-engine",
        "cyclical",
        "leverage",
        "capital-intensive",
        "compound-crisis",
        "sector-analysis",
        "28-industries"
      ]
    },
    {
      "id": 46,
      "sourceId": "cm-047",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "general",
      "title": "Industries Most Resilient to Economic Crisis",
      "content": "Counter to the vulnerability analysis (cm-046), certain industries historically maintain stability or even benefit during economic crises. Understanding resilience patterns is critical for crisis monitoring because it identifies where capital flows during stress events (the 'flight to safety' at the sector level) and which parts of the economy continue to function as anchors.\n\nDefense and aerospace has historically been among the most crisis-resilient industries, particularly during geopolitically-driven crises. Government defense spending is budgeted years in advance, insulated from short-term economic cycles, and often increases during geopolitical conflicts — the very events that trigger crises in other sectors. The defense-industrial complex operates on long-term contracts with guaranteed revenue, making it effectively counter-cyclical during geopolitical crises.\n\nHealthcare and pharmaceuticals show strong crisis resilience because demand is non-discretionary — people continue to need medical care and medications regardless of economic conditions. During the 2008 GFC, the healthcare sector outperformed the S&P 500 by approximately 20 percentage points. However, healthcare is not immune to all crisis types: a severe credit freeze can impair hospital operations (which are often debt-financed), and insurance/coverage disruptions during mass unemployment can reduce healthcare demand at the margins.\n\nUtilities provide essential services (electricity, water, gas) with regulated revenue streams and relatively inelastic demand. They typically outperform during recessions and market crashes because investors seek their stable dividends and predictable cash flows. Consumer staples (food, household products, basic personal care) follow a similar pattern — demand holds up because these products are necessities. During the 2008 GFC, consumer staples outperformed the broad market by approximately 25 percentage points.\n\nThe energy sector shows a mixed resilience pattern that depends on the crisis type. During demand-driven recessions, energy suffers (2008-2009, 2020). During supply-driven geopolitical crises, energy benefits (1973, 2022). This makes energy an effective hedge against geopolitical/inflation crises but a vulnerability during credit/demand crises. For portfolio construction in a crisis monitoring context, the key insight is that no single sector is resilient to all crisis types — diversification across resilience factors (non-discretionary demand, government spending, regulated revenue, commodity exposure) provides broader protection than concentration in any single 'defensive' sector.",
      "dataSource": "AI-Rev engine crisis scenario analysis, S&P 500 sector returns during historical crises, Bloomberg sector analytics, academic research on defensive sectors",
      "leadTime": "Resilience patterns are structural and identifiable in advance. Useful for portfolio positioning before crises materialize.",
      "reliability": "Historical resilience patterns are well-documented and persistent across crisis types. The key caveat is that 'resilient' does not mean 'unaffected' — even defensive sectors decline during severe crises, just less than cyclical sectors.",
      "historicalPrecedents": [
        "2008 GFC (healthcare, utilities, staples outperformed)",
        "2020 COVID (healthcare, defense, tech outperformed)",
        "2022 inflation/rate shock (energy outperformed)",
        "1973 oil crisis (defense, energy outperformed)"
      ],
      "compoundsWith": [
        "general",
        "geopolitical",
        "energy-shock"
      ],
      "tags": [
        "crisis-resilience",
        "defensive-sectors",
        "defense",
        "healthcare",
        "utilities",
        "consumer-staples",
        "flight-to-safety",
        "counter-cyclical"
      ]
    },
    {
      "id": 47,
      "sourceId": "cm-048",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "general",
      "title": "Cross-Industry Cascade During Crisis vs. Normal AI Transformation",
      "content": "The AI-Rev engine models cross-industry effects under both normal AI transformation scenarios and crisis scenarios, and the comparison reveals a critical insight: the direction, magnitude, and speed of cross-industry effects differ dramatically between these two regimes. Understanding this difference is essential for crisis monitoring because it means that cross-industry models calibrated during normal times will systematically underestimate crisis-period effects.\n\nDuring normal AI transformation, cross-industry effects are primarily positive-sum: AI adoption in one industry creates opportunities for adjacent industries. Cloud computing growth drives semiconductor demand. AI-enhanced manufacturing improves supply chain efficiency for retailers. Financial services AI creates demand for data analytics tools. These effects propagate slowly (quarters to years), are generally amplifying (positive effects begetting more positive effects), and allow industries time to adapt. The AI-Rev engine's 167 mapped cross-effects primarily capture these normal-regime dynamics.\n\nDuring a crisis, cross-industry effects reverse character. Instead of positive-sum opportunity creation, the effects become negative-sum cascading constraints. A credit freeze in financial services does not just reduce lending — it chokes working capital for every other industry. An AI capex pullback does not just affect semiconductor companies — it reduces cloud revenue, which reduces data center construction, which reduces equipment manufacturing, which reduces raw material demand. These crisis-regime effects propagate fast (weeks to months, not quarters to years), are amplifying in the negative direction, and overwhelm industries' ability to adapt.\n\nThe most dangerous cross-industry crisis dynamic is when an industry that was a growth engine during the expansion becomes a drag during the contraction. The AI/tech sector currently functions as the primary growth engine for the US economy and stock market — its capex drives employment, its revenue growth drives earnings expectations, and its stock performance drives index returns. If this sector reverses from engine to anchor (as happened to financials in 2008, to telecom in 2001, and to energy in 2014-2016), the cross-industry cascade is amplified because the rest of the economy had come to depend on it. The concentration of growth drivers in a small number of AI-adjacent companies makes this 'engine reversal' particularly impactful — the economy has fewer alternative growth engines to compensate.\n\nFor crisis monitoring, this means that the AI-Rev engine's cross-effect model should be run in both regimes: normal-expansion mode (positive-sum, slow propagation) and crisis mode (negative-sum, fast propagation, engine reversal). Comparing the outputs reveals which industries are most exposed to the regime switch and where the largest amplification effects occur.",
      "dataSource": "AI-Rev engine cross-industry effect model (167 mapped effects), BEA input-output tables, historical sector correlation data during crises vs. normal periods",
      "leadTime": "The regime-dependent nature of cross-industry effects is a structural insight. It does not predict crisis timing but does predict how the crisis will propagate once triggered.",
      "reliability": "The principle that cross-industry effects change character during crises is well-supported by historical evidence. The specific magnitude estimates depend on the AI-Rev engine's modeling assumptions and are less certain.",
      "historicalPrecedents": [
        "2008 GFC (financial engine reversal cascaded through all sectors)",
        "2001 dot-com (tech/telecom engine reversal)",
        "2020 COVID (services/travel engine reversal, but tech became an engine — unusual)"
      ],
      "compoundsWith": [
        "ai-specific",
        "credit-debt",
        "asset-bubble",
        "systemic-banking",
        "general"
      ],
      "tags": [
        "cross-industry-cascade",
        "ai-rev-engine",
        "regime-dependent",
        "engine-reversal",
        "positive-sum-vs-negative-sum",
        "concentration-risk",
        "167-cross-effects"
      ]
    },
    {
      "id": 48,
      "sourceId": "cm-049",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "yield-curve",
      "title": "10Y-3M Yield Curve - Current Reading (March 2026)",
      "content": "The 10-Year minus 3-Month Treasury spread currently stands at +0.71%, meaning the yield curve is no longer inverted after one of the longest and deepest inversions in modern history. The persistent inversion that began in late 2022 ended approximately in September 2024, placing us roughly 18 months into the post-inversion phase. While many commentators treated the end of inversion as an 'all clear' signal, the historical record tells the opposite story: the most dangerous phase of the yield curve cycle is not the inversion itself, but the rapid un-inversion that follows.\n\nThe mechanism is straightforward but widely misunderstood. Yield curve inversions signal that bond markets expect economic weakness ahead — short-term rates exceed long-term rates because traders anticipate the Fed will be forced to cut rates in response to a deteriorating economy. The inversion does not cause the recession; it reflects the market's forward-looking assessment. The un-inversion occurs when the anticipated deterioration begins to materialize: either the Fed starts cutting short-term rates (pulling down the short end) or long-term inflation expectations spike (pushing up the long end), or both. In the current case, the Fed has cut to 3.50-3.75% while 10-year yields remain elevated at 4.38%, producing a positive spread through a combination of rate cuts and sticky long-term inflation expectations.\n\nThe historical precedent is stark and consistent. The yield curve un-inverted in late 2000, and the recession began in March 2001. The curve un-inverted in mid-2007, and the Great Recession began in December 2007. In both cases, the lag from un-inversion to recession onset was 6-12 months. We are now approximately 18 months past un-inversion without a declared recession, which either means the pattern is breaking or the lag is longer this cycle due to the unprecedented fiscal stimulus and consumer balance sheet strength from locked-in low mortgage rates. The current +0.71% spread with an actively steepening trajectory matches the 2000 and 2007 pre-recession configurations almost exactly.",
      "dataSource": "FRED Series T10Y3M",
      "leadTime": "Historically 6-18 months from inversion to recession, with recession often beginning shortly after un-inversion. Currently 18 months post-inversion.",
      "reliability": "The 10Y-3M spread has predicted every US recession since the 1960s with only one false positive (1966). The un-inversion-to-recession pattern held in 2001 and 2007-2008.",
      "historicalPrecedents": [
        "Late 2000 un-inversion preceded March 2001 recession",
        "Mid-2007 un-inversion preceded December 2007 Great Recession",
        "1989 un-inversion preceded July 1990 recession"
      ],
      "compoundsWith": [
        "currency-inflation",
        "systemic-banking",
        "energy-shock",
        "asset-bubble"
      ],
      "tags": [
        "yield-curve",
        "recession-signal",
        "un-inversion",
        "FRED",
        "10Y-3M",
        "current-reading"
      ]
    },
    {
      "id": 49,
      "sourceId": "cm-050",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "yield-curve",
      "title": "10Y-2Y Yield Curve - Current Reading (March 2026)",
      "content": "The 10-Year minus 2-Year Treasury spread currently sits at +0.65%, with the 10-year yield at 4.38% and the 2-year yield at 3.73%. Like the 10Y-3M spread, this measure has fully un-inverted after a prolonged period of negative readings. However, the manner of un-inversion is critically important: this is a 'bear steepening' pattern, where the spread is widening primarily because longer-term yields are rising (or staying elevated) while shorter-term yields decline. This is the dangerous variant of steepening.\n\nBear steepening occurs when the bond market simultaneously prices in two things: near-term economic weakness (pulling down shorter-term yields as rate cuts are expected) and longer-term inflation or fiscal risk (keeping long-term yields elevated). The result is a positive spread that looks healthy on the surface but actually reflects deep market anxiety about the economic trajectory. In contrast, 'bull steepening' occurs when short rates fall faster than long rates in a low-inflation environment — that is the benign variant associated with the early phase of economic recovery. The current configuration is unambiguously bear steepening: long-term yields remain above 4.3% despite Fed cuts, driven by persistent inflation expectations, massive fiscal deficits, and now the oil price shock from Hormuz disruptions.\n\nThe bear steepening pattern in late 2000 and summer 2007 preceded two of the worst economic downturns in modern American history. In both cases, the pattern persisted for 3-9 months before the recession officially began. The mechanism is that bear steepening reflects the bond market's judgment that the central bank will be unable to engineer a soft landing — that the economy is weakening but inflationary pressures prevent the aggressive rate cuts needed to cushion the blow. This is precisely the Fed's current dilemma: softening labor markets and slowing growth argue for rate cuts, while $112 oil and sticky services inflation argue against them. The bond market is pricing in this trap, and the bear steepening is the visible manifestation.",
      "dataSource": "FRED Series T10Y2Y",
      "leadTime": "Bear steepening phases have historically preceded recessions by 3-12 months.",
      "reliability": "The 10Y-2Y spread is the most widely followed yield curve measure. Bear steepening specifically has a strong track record as a pre-recession signal, though it has occasionally occurred without subsequent recession.",
      "historicalPrecedents": [
        "Late 2000 bear steepening preceded 2001 recession",
        "Summer 2007 bear steepening preceded Great Recession",
        "1989-1990 bear steepening preceded 1990 recession"
      ],
      "compoundsWith": [
        "currency-inflation",
        "energy-shock",
        "systemic-banking",
        "asset-bubble"
      ],
      "tags": [
        "yield-curve",
        "bear-steepening",
        "recession-signal",
        "10Y-2Y",
        "current-reading"
      ]
    },
    {
      "id": 50,
      "sourceId": "cm-051",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "Credit-to-GDP Gap - Current Reading (March 2026)",
      "content": "The credit-to-GDP gap — the deviation of the private credit-to-GDP ratio from its long-term trend — currently reads slightly negative, meaning private sector credit is growing more slowly than its historical trend relative to GDP. This places the gap comfortably below the Bank for International Settlements' 2.0% danger threshold, which has historically served as a reliable early warning signal for banking crises. The negative reading reflects post-pandemic deleveraging dynamics: the massive fiscal and monetary stimulus of 2020-2021 boosted GDP faster than private credit expanded, and subsequent tightening has restrained new credit creation.\n\nThe BIS credit-to-GDP gap is one of the most rigorously backtested crisis indicators in the institutional risk monitoring toolkit. Research published by the BIS across multiple working papers demonstrates that when this gap exceeds 10%, a banking crisis follows within 1-3 years with remarkable consistency. The mechanism is intuitive: when private credit grows much faster than the economy, it means borrowers are taking on debt that the economy's productive capacity cannot support. The gap above 2% is the early warning zone; above 10% is the danger zone. The gap reached approximately 12% in the United States before the 2007-2008 crisis, providing a clear signal several years in advance.\n\nThe current negative reading is genuinely reassuring for this specific indicator, but it comes with important caveats. First, the gap measures private sector credit — it does not capture the massive expansion of public sector debt, which has surged to 128% of GDP. The risk has shifted from private balance sheets to the sovereign balance sheet, which the credit-to-GDP gap does not measure. Second, the gap can turn quickly: if the Fed's rate cuts stimulate a new credit expansion cycle while GDP growth stalls (stagflation scenario), the gap could swing from negative to positive within 12-18 months. Third, some analysts argue that the gap understates risk because it compares to a trend that was itself elevated during the pre-2008 credit boom. Despite these caveats, the credit-to-GDP gap's current reading provides one of the few genuinely positive signals in the current risk landscape.",
      "dataSource": "BIS Credit-to-GDP gaps dataset",
      "leadTime": "2-5 years when gap exceeds 10%. Early warning (above 2%) provides 1-3 years lead time.",
      "reliability": "Among the most reliable crisis indicators when elevated. The BIS considers it the single best early warning indicator for banking crises. Current negative reading is a genuine positive signal, though it does not capture public debt risk.",
      "historicalPrecedents": [
        "2005-2007 gap above 10% preceded 2008 GFC",
        "1997 Asian economies showed elevated gaps before crisis",
        "Spain/Ireland gaps above 25% before 2010 crisis"
      ],
      "compoundsWith": [
        "systemic-banking",
        "asset-bubble",
        "currency-inflation"
      ],
      "tags": [
        "credit-gap",
        "BIS",
        "deleveraging",
        "private-credit",
        "early-warning",
        "current-reading"
      ]
    },
    {
      "id": 51,
      "sourceId": "cm-052",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "Shiller CAPE Ratio - Current Reading (March 2026)",
      "content": "The Cyclically Adjusted Price-to-Earnings ratio (CAPE), developed by Nobel laureate Robert Shiller, currently stands at 38.93 — more than 2.56 times the historical average of 15.21 dating back to 1871. This ratio smooths earnings over a 10-year inflation-adjusted period to eliminate cyclical distortions, providing the clearest long-term valuation signal available for the US equity market. At nearly 39, the current CAPE exceeds every historical peak except the absolute apex of the dot-com bubble in early 2000, when it briefly touched 44.19.\n\nThe CAPE ratio above 30.0 is widely considered extreme overvaluation territory by institutional investors, and the historical record supports this assessment with brutal clarity. There have been only three sustained periods with CAPE above 30: the late 1920s (peaked at ~33, followed by the 1929 crash and Great Depression), the late 1990s through early 2000 (peaked at ~44, followed by the dot-com crash and 50%+ decline in the S&P 500), and the current period beginning in 2020. In every previous instance, CAPE above 30 was followed by a period of dramatically below-average returns over the subsequent 10-year period. Shiller's own research demonstrates a strong negative correlation between CAPE levels and subsequent 10-year real returns — at current levels, the expected 10-year real annualized return is approximately 0-2%, compared to the historical average of 6.5%.\n\nCritics of CAPE argue that structural changes justify elevated valuations: higher corporate profit margins due to technology and globalization, the shift from capital-intensive to asset-light business models, low real interest rates for much of the post-2008 era, and the dominance of high-margin tech companies in the index. These arguments have some validity — the current era's profit margins are genuinely higher than the 1950-2000 average. However, the magnitude of the deviation is extreme. Even adjusting for elevated margins, CAPE is arguably 50-70% above fair value. Furthermore, the 'this time is different' arguments mirror almost exactly the language used to justify valuations in 1999 ('new economy,' 'paradigm shift,' 'old metrics don't apply'). The CAPE at 38.93 is not a timing signal — markets can remain overvalued for years — but it is a screaming signal about the magnitude of potential downside risk if any catalyst triggers a revaluation.",
      "dataSource": "Robert Shiller's Yale website, YCharts",
      "leadTime": "Not a timing indicator — CAPE can remain elevated for years. Strong predictor of 10-year forward returns. Extreme readings (above 30) increase the severity of any correction that does occur.",
      "reliability": "Among the most reliable long-term valuation indicators. Correctly identified overvaluation before every major crash. Limitation: poor timing signal — markets can remain expensive for extended periods.",
      "historicalPrecedents": [
        "1929 CAPE ~33 preceded Great Depression",
        "2000 CAPE ~44 preceded dot-com crash",
        "2007 CAPE ~28 preceded GFC (below current level)"
      ],
      "compoundsWith": [
        "yield-curve",
        "credit-debt",
        "ai-specific",
        "energy-shock"
      ],
      "tags": [
        "CAPE",
        "overvaluation",
        "Shiller",
        "dot-com-parallel",
        "long-term-returns",
        "current-reading"
      ]
    },
    {
      "id": 52,
      "sourceId": "cm-053",
      "section": "composite-indices",
      "category": "composite-index",
      "subcategory": "general",
      "title": "OFR Financial Stress Index - Current Reading (March 2026)",
      "content": "The Office of Financial Research Financial Stress Index — the most comprehensive daily measure of financial system stress, synthesizing 33 market variables across five categories — currently reads -0.56. Since the index baseline is 0.0 (representing median historical stress), a reading of -0.56 indicates financial stress is modestly below average. On its face, this is a reassuring signal: markets are not in panic mode, credit is flowing, and interbank funding is functioning normally. However, the decomposition by category reveals a more nuanced and concerning picture.\n\nOf the five OFR FSI categories (credit, equity valuation, funding, safe assets, and volatility), the volatility component is notably elevated relative to the others. This divergence reflects a market that is calm in terms of credit availability and funding conditions but increasingly anxious about tail risks — geopolitical instability (Iran-US conflict, Hormuz disruption), energy price shocks, and policy uncertainty around the Fed's rate path. The pattern of low overall stress with elevated volatility is characteristic of a market in 'pre-crisis' positioning: institutions have not yet begun to deleverage or pull credit lines (which would elevate the credit and funding components), but they are actively hedging against the possibility that conditions could deteriorate rapidly. This creates a surface-level calm that masks growing tail-risk anxiety.\n\nThe last time the OFR FSI showed a significant sustained reading above 0.0 was March 2020, when the COVID-induced global liquidity freeze pushed the index above 15 in a matter of days — one of the fastest spikes in the index's history. The speed of that transition from sub-zero to double-digits illustrates a critical property of stress indices: they tend to be non-linear. Stress can remain low for extended periods as risks build, then spike violently when a trigger event overwhelms market circuit breakers. The current -0.56 reading should therefore not be interpreted as 'all clear' but rather as 'calm before potential storm' — particularly given that the volatility component is already diverging from the headline number. Historical analysis shows that volatility-led divergences have preceded broader stress spikes by 2-8 weeks in multiple prior episodes.",
      "dataSource": "OFR Financial Stress Index data portal",
      "leadTime": "Contemporaneous to very short-term. Volatility component divergence can lead broader stress by 2-8 weeks.",
      "reliability": "High reliability as a stress measure. The -0.56 headline is genuine but the volatility divergence warrants monitoring.",
      "historicalPrecedents": [
        "March 2020 spike from sub-zero to 15+ in days",
        "2008 GFC peak near 25",
        "2023 SVB stress peak near 5"
      ],
      "compoundsWith": [
        "credit-debt",
        "systemic-banking",
        "energy-shock",
        "currency-inflation",
        "asset-bubble"
      ],
      "tags": [
        "OFR-FSI",
        "volatility",
        "stress-index",
        "divergence",
        "pre-crisis-positioning",
        "current-reading"
      ]
    },
    {
      "id": 53,
      "sourceId": "cm-054",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "VIX Level and Term Structure - Current Reading (March 2026)",
      "content": "The CBOE Volatility Index (VIX) currently stands at 31.05, with its futures term structure in pronounced backwardation — front-month futures trading at a -6.89% premium relative to longer-dated contracts. This combination of elevated spot VIX plus inverted term structure represents one of the most unambiguous near-term danger signals in the financial markets. It means that options traders are paying significantly more for short-term downside protection than for longer-term protection, which only happens when the market perceives an imminent threat rather than a slow-building risk.\n\nThe VIX above 25 is the first danger threshold, indicating that institutional investors are actively hedging against significant near-term declines. The VIX above 30 — where we currently sit — escalates this signal considerably: it corresponds roughly to the market pricing in daily moves of approximately 2% as 'normal,' which is characteristic of acute stress periods rather than typical market fluctuations. For context, the VIX's long-run average is approximately 19-20, and it spends most of its time between 12 and 25. Sustained readings above 30 have historically been associated with active market crises: the 2008 GFC (VIX peaked above 80), the 2020 COVID crash (VIX peaked above 82), and the 2022 Russia-Ukraine invasion and rapid Fed hiking cycle (VIX peaked above 36).\n\nThe inverted term structure (backwardation) adds a critical dimension that the VIX level alone does not capture. In normal markets, VIX futures trade in contango — longer-dated futures cost more than shorter-dated ones, reflecting the natural 'insurance premium' for longer time horizons. When this structure inverts, it means traders are paying a higher premium for near-term protection than for longer-term protection, which implies they believe the threat is immediate and specific rather than diffuse and gradual. The -6.89% front-month premium is deep by historical standards. The last prolonged period of deep VIX backwardation was during the 2022 Russia-Ukraine invasion and Fed tightening shock. The current backwardation is almost certainly driven by the Hormuz crisis, tariff escalation fears, and uncertainty about how the oil price shock will transmit through inflation expectations and Fed policy. Taken together, VIX at 31 with -6.89% backwardation signals that institutional money is positioned for an imminent acute event.",
      "dataSource": "CBOE, VIXCentral",
      "leadTime": "Very short-term — days to weeks. VIX backwardation signals an imminent event rather than a distant risk.",
      "reliability": "High for near-term directional signal. VIX above 30 plus backwardation has historically preceded significant market declines within 1-4 weeks. Limitation: VIX can spike and normalize quickly without broader crisis materializing.",
      "historicalPrecedents": [
        "2022 Russia-Ukraine invasion (VIX >36 with backwardation)",
        "March 2020 COVID (VIX >80 with extreme backwardation)",
        "2008 GFC (VIX >80 with sustained backwardation)",
        "August 2015 China devaluation (brief VIX spike >40)"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "yield-curve",
        "asset-bubble",
        "currency-inflation"
      ],
      "tags": [
        "VIX",
        "backwardation",
        "volatility",
        "panic-signal",
        "term-structure",
        "options-market",
        "current-reading"
      ]
    },
    {
      "id": 54,
      "sourceId": "cm-055",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "US Corporate Debt-to-GDP - Current Reading (March 2026)",
      "content": "US non-financial corporate debt currently stands at approximately 49.6% of GDP, as measured by the Federal Reserve's Z.1 Financial Accounts (Flow of Funds). This places corporate leverage near a critical threshold: sustained readings above 50% of GDP historically represent a zone where corporate debt servicing becomes highly vulnerable to interest rate increases, revenue slowdowns, or both simultaneously. For context, the pre-2008 crisis peak was approximately 45% of GDP, meaning current corporate leverage already exceeds the level that preceded the worst financial crisis in 80 years.\n\nThe trajectory of corporate debt tells an important story about post-pandemic dynamics. During the pandemic response in 2020, corporate debt briefly spiked above 55% of GDP as companies drew down credit lines and issued emergency debt while GDP simultaneously contracted — creating a double effect on the ratio. Federal Reserve emergency interventions (including the unprecedented step of buying corporate bonds) prevented this spike from triggering a cascading default cycle. Since then, strong nominal GDP growth (partly inflation-driven) has improved the ratio, but underlying debt levels have not materially declined. Companies refinanced pandemic-era debt at higher rates, and the maturity wall of corporate bonds coming due in 2025-2027 represents a significant vulnerability: approximately $2.5 trillion in investment-grade and high-yield bonds will need to be rolled over at interest rates 200-400 basis points higher than their original issuance.\n\nThe danger threshold of 50% is not arbitrary — it reflects the point at which a meaningful interest rate increase can push aggregate corporate interest expense above the growth rate of corporate revenue, creating a net cash flow squeeze across the entire corporate sector. At 49.6%, we are essentially at this threshold. The current Fed Funds rate of 3.50-3.75% means companies are refinancing into a fundamentally different rate environment than the near-zero regime under which much of this debt was issued. The combination of elevated leverage near 50% of GDP, a massive refinancing wall, and an interest rate environment 300+ basis points above the issuance rate creates a slow-motion stress test on corporate balance sheets. Any additional shock — revenue slowdown from recession, further rate increases from oil-driven inflation — would tip this indicator into unambiguous danger territory.",
      "dataSource": "Federal Reserve Z.1 Financial Accounts",
      "leadTime": "Slow-moving indicator. Elevated levels create vulnerability that persists for quarters to years, with the actual crisis triggered by rate changes, revenue shocks, or refinancing failures.",
      "reliability": "Strong empirical relationship between elevated corporate debt-to-GDP and subsequent financial stress. The 2008 GFC occurred at a lower ratio than today's reading.",
      "historicalPrecedents": [
        "Pre-2008 GFC peak at ~45% preceded worst financial crisis in 80 years",
        "2020 COVID spike above 55% required unprecedented Fed intervention",
        "Early 2000s ratio at ~44% preceded corporate default wave"
      ],
      "compoundsWith": [
        "yield-curve",
        "systemic-banking",
        "currency-inflation",
        "ai-specific"
      ],
      "tags": [
        "corporate-debt",
        "leverage",
        "Z1",
        "refinancing-wall",
        "interest-rate-vulnerability",
        "current-reading"
      ]
    },
    {
      "id": 55,
      "sourceId": "cm-056",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "systemic-banking",
      "title": "Commercial Real Estate Delinquency - Current Reading (March 2026)",
      "content": "The overall past-due and non-accrual (PDNA) rate for commercial real estate loans stands at 1.56%, according to the most recent FDIC Quarterly Banking Profile. This headline number remains below the pre-pandemic average of approximately 1.94%, which on its surface appears manageable. However, the aggregate figure masks severe divergence across property types that represents one of the most significant risks to the US regional banking system. Office and multifamily property segments are performing substantially worse than the overall figure suggests, with delinquency rates well above their 2019 baselines and deteriorating.\n\nThe office sector is the epicenter of CRE stress. The post-pandemic structural shift toward remote and hybrid work has permanently reduced demand for office space in most major metropolitan areas. Office vacancy rates nationally have climbed above 18%, and in some urban cores they exceed 25%. Property values have declined 30-40% from 2019 peaks in the most affected markets. Loans written against these properties at 2019-2021 valuations are now significantly underwater, and as they come up for refinancing, borrowers face the choice of injecting substantial equity (often infeasible) or defaulting. The multifamily segment, while less severely impacted, is experiencing stress from a combination of new supply deliveries (projects started during the 2021-2022 building boom) and softening rent growth, particularly in Sun Belt markets that saw the most aggressive development.\n\nThe systemic danger threshold for CRE delinquencies is a rapid acceleration past 3.0%, which would signal that the manageable slow bleed has transformed into an acute credit event. During the 2009-2010 peak, CRE delinquency rates exceeded 8%, contributing to hundreds of regional bank failures and necessitating FDIC intervention on a massive scale. The current trajectory is concerning but not yet critical: the pace of deterioration is gradual rather than sudden, and the largest banks have been proactively increasing loss reserves for CRE exposure. The primary risk is concentrated in regional and community banks, which hold a disproportionate share of CRE loans relative to their total assets — for some smaller banks, CRE represents 300%+ of their tier 1 capital. If the office sector deterioration accelerates or spreads to other property types, these banks are the most vulnerable link in the chain.",
      "dataSource": "FDIC Quarterly Banking Profile",
      "leadTime": "Slow-moving. CRE stress typically builds over 12-24 months before reaching systemic levels. The office sector has been deteriorating since 2022.",
      "reliability": "FDIC data is authoritative and comprehensive. CRE stress is well-documented and actively monitored by regulators. The property-type divergence is real and concerning.",
      "historicalPrecedents": [
        "2009-2010 CRE delinquencies exceeded 8%, hundreds of bank failures",
        "1990-1991 S&L crisis had major CRE component",
        "2023 regional bank stress partially driven by CRE concerns"
      ],
      "compoundsWith": [
        "credit-debt",
        "yield-curve",
        "currency-inflation"
      ],
      "tags": [
        "CRE",
        "office",
        "delinquency",
        "regional-banks",
        "multifamily",
        "vacancy-rates",
        "current-reading"
      ]
    },
    {
      "id": 56,
      "sourceId": "cm-057",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "Household Debt Service Ratio - Current Reading (March 2026)",
      "content": "The household debt service ratio (DSR) — the percentage of disposable personal income devoted to debt payments — currently stands at 11.32%, according to the Federal Reserve's quarterly estimate (FRED Series TDSP). This is meaningfully below the historical average of approximately 12.1% observed during the 2010-2019 period, and dramatically below the pre-crisis peak of nearly 16.0% reached in late 2007. The danger threshold for this indicator is generally considered to be 13-14%, the zone where debt servicing begins to crowd out consumption spending and household financial fragility increases sharply. At 11.32%, American households are, in aggregate, in a surprisingly healthy position relative to their debt obligations.\n\nThe primary explanation for this resilience is the mortgage rate lock-in effect. During the 2020-2021 period of near-zero interest rates, approximately 60% of outstanding mortgages were refinanced at rates below 4%, and many new purchases were financed at rates between 2.5% and 3.5%. Since mortgages represent the largest component of household debt service by a wide margin, this mass refinancing event effectively locked in low debt service costs for the majority of American homeowners for 15-30 years. Even though the current 30-year mortgage rate exceeds 7%, this only affects new borrowers — the existing stock of mortgages remains at historically low rates. This creates an unusual situation where the household sector is substantially insulated from the Fed's interest rate increases, which is precisely why consumer spending has remained more resilient than most economists expected.\n\nThis is one of the few indicators in the current risk landscape that is genuinely positive. However, it comes with important caveats. First, the aggregate figure masks distributional stress: renters, younger households without fixed-rate mortgages, and households with significant credit card or auto loan debt are experiencing much higher effective debt service ratios than the 11.32% average suggests. Credit card delinquencies have been rising steadily, and auto loan delinquencies are above pre-pandemic levels for subprime borrowers. Second, the household sector's resilience is partly what is preventing the recession that other indicators are signaling — but this resilience also gives the Fed less urgency to cut rates aggressively, which in turn maintains pressure on the corporate sector and commercial real estate. The household DSR may be inadvertently contributing to a longer, slower-building crisis by masking the pain that would otherwise force faster policy intervention.",
      "dataSource": "FRED Series TDSP",
      "leadTime": "Slow-moving. Household balance sheets change gradually. Current strength provides a buffer that extends the timeline before any potential crisis.",
      "reliability": "Fed data is authoritative. The 11.32% reading is a genuine positive signal for household resilience, though it does not capture distributional stress.",
      "historicalPrecedents": [
        "Late 2007 peak at ~16% preceded wave of mortgage defaults",
        "2010-2019 average of 12.1% represented post-crisis normal",
        "Current 11.32% is among the lowest readings in 20+ years"
      ],
      "compoundsWith": [
        "systemic-banking",
        "asset-bubble",
        "currency-inflation"
      ],
      "tags": [
        "household-debt",
        "consumer-health",
        "mortgage-rates",
        "rate-lock-in",
        "debt-service-ratio",
        "current-reading"
      ]
    },
    {
      "id": 57,
      "sourceId": "cm-058",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "Mag 7 Concentration Risk - Current Reading (March 2026)",
      "content": "The Magnificent 7 — Apple, Microsoft, Alphabet, Amazon, NVIDIA, Meta, and Tesla — currently account for approximately 33.3% of the entire S&P 500 market capitalization. This level of index concentration is unprecedented in the modern era and far exceeds the 25% threshold that portfolio strategists and index providers have identified as the danger zone for concentration risk. To put this in perspective, seven companies out of 503 constituents control one-third of the index, meaning the other 496 companies collectively account for two-thirds. The S&P 500 has effectively become a Magnificent 7 tracking fund with a diversified tail.\n\nThe danger of extreme index concentration is not primarily about the quality of the underlying companies — several of the Mag 7 are genuinely exceptional businesses with strong cash flows and competitive moats. The danger is structural: when a small number of stocks drive index performance, the entire index becomes fragile to sector-specific or company-specific shocks that would otherwise be diversified away. If NVIDIA faces an AI capex slowdown, or if antitrust action forces Alphabet to restructure, or if Tesla's margins compress further, the impact on the S&P 500 is disproportionately large relative to the economic significance of the event. This fragility means that 'the market' as measured by cap-weighted indices is not a diversified reflection of the economy — it is a concentrated bet on the continued outperformance of AI-adjacent mega-caps.\n\nHistorical precedents for this level of concentration are sobering. The Nifty Fifty of the early 1970s — a group of approximately 50 large-cap 'one decision' stocks that dominated institutional portfolios — saw dramatic underperformance during the 1973-1974 bear market, with many of these stocks losing 60-80% of their value despite being fundamentally sound businesses. The dot-com era saw similar dynamics: the five largest stocks (Microsoft, GE, Cisco, Intel, Walmart) represented approximately 18% of the S&P 500 at the 2000 peak — significantly less than today's 33.3% — and their subsequent underperformance dragged the index down for years. The current concentration exceeds even those historical extremes, suggesting that any rotation away from mega-cap tech would produce an unusually painful 'catch-down' in cap-weighted indices, even if the broader economy and equal-weighted indices hold up relatively well.",
      "dataSource": "S&P Dow Jones Indices, Goldman Sachs Global Investment Research",
      "leadTime": "Concentration risk is a structural vulnerability, not a timing signal. It amplifies the impact of any sector-specific catalyst.",
      "reliability": "Index concentration data is objective and well-measured. The relationship between extreme concentration and subsequent cap-weighted underperformance is historically consistent.",
      "historicalPrecedents": [
        "Nifty Fifty 1970s concentration preceded 1973-74 crash",
        "2000 dot-com top-5 at ~18% preceded years of underperformance",
        "Current 33.3% exceeds all modern precedents"
      ],
      "compoundsWith": [
        "ai-specific",
        "yield-curve",
        "credit-debt",
        "geopolitical"
      ],
      "tags": [
        "concentration-risk",
        "Mag-7",
        "index-fragility",
        "tech-bubble",
        "Nifty-Fifty",
        "cap-weighted",
        "current-reading"
      ]
    },
    {
      "id": 58,
      "sourceId": "cm-059",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "Global Sovereign Debt Levels - Current Reading (March 2026)",
      "content": "Global sovereign debt has reached levels that would have been considered unthinkable a generation ago. The United States carries approximately 128% debt-to-GDP, with Congressional Budget Office projections showing a trajectory toward 133% by 2030 under current fiscal policies. Japan continues to lead advanced economies at approximately 230% debt-to-GDP, a level sustained only because the Bank of Japan is the dominant buyer of its own government's bonds through yield curve control. The European Union averages 85-90% debt-to-GDP, though this masks dramatic variance: Italy at approximately 140%, France at 110%, Germany at 65%. The traditional danger threshold from the Reinhart-Rogoff framework is 90-100% of GDP, above which economic growth rates have historically declined.\n\nThe US fiscal trajectory is particularly concerning because unlike Japan — which finances its debt almost entirely through domestic savings — the US depends significantly on foreign buyers, particularly China, Japan, and sovereign wealth funds in the Middle East and Asia. When these buyers reduce their appetite for Treasuries (as has been occurring gradually over the past several years), the Treasury must offer higher yields to attract marginal buyers, which increases the interest burden, which increases the deficit, which increases debt issuance — a self-reinforcing fiscal doom loop. The federal government's net interest payments already exceed $800 billion annually, surpassing the defense budget, and they are projected to exceed $1 trillion within the next 12-18 months at current rate levels.\n\nThe practical danger from elevated sovereign debt materializes through several channels. First, the fiscal space to respond to the next crisis is constrained — the $5 trillion in pandemic stimulus was financed by debt markets that expected rates to stay low; a similar response at current rates would be dramatically more expensive and potentially market-destabilizing. Second, high sovereign debt crowds out private investment as government borrowing absorbs savings that would otherwise fund private enterprise. Third, there is a reflexive relationship between sovereign debt levels and market confidence: when debt-to-GDP crosses certain psychological thresholds, bond markets can suddenly demand a risk premium (as happened to Italy in 2011, Greece in 2010, and several emerging markets repeatedly). The US has thus far avoided this dynamic due to the dollar's reserve currency status, but that privilege is not unlimited. Japan's ability to sustain 230% debt-to-GDP is frequently cited as evidence that high debt is manageable, but Japan's domestic financing structure, persistent deflation, and current account surplus make it a poor template for the US.",
      "dataSource": "IMF World Economic Outlook, OECD",
      "leadTime": "Sovereign debt crises typically build over years and trigger suddenly when bond markets lose confidence. The fiscal doom loop can accelerate rapidly once it begins.",
      "reliability": "Debt-to-GDP ratios are well-measured and widely available. The relationship between very high debt levels and constrained fiscal space is established. The specific tipping point varies by country.",
      "historicalPrecedents": [
        "Greece 2010 (debt crisis at ~130% GDP)",
        "Italy 2011 (bond market revolt at ~120% GDP)",
        "Japan sustained 230%+ through domestic financing",
        "US trajectory toward 133% by 2030"
      ],
      "compoundsWith": [
        "currency-inflation",
        "yield-curve",
        "systemic-banking",
        "geopolitical"
      ],
      "tags": [
        "sovereign-debt",
        "fiscal",
        "IMF",
        "debt-sustainability",
        "doom-loop",
        "interest-burden",
        "current-reading"
      ]
    },
    {
      "id": 59,
      "sourceId": "cm-060",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "ai-specific",
      "title": "AI Capex vs Revenue Gap - Current Reading (March 2026)",
      "content": "The global AI infrastructure investment cycle has now committed over $600 billion in capital expenditure across hyperscalers (Microsoft, Google, Amazon, Meta, Oracle), chip manufacturers (NVIDIA, AMD, TSMC, Intel), and data center developers. This figure encompasses announced and in-progress spending on GPU clusters, data center construction, power infrastructure, and networking equipment. Meanwhile, actual AI-specific end-user software revenue — money paid by businesses and consumers for AI tools, services, and applications — remains a fraction of this investment. Estimates vary, but credible analyses from Sequoia Capital and Goldman Sachs place AI end-user revenue in the $50-80 billion range, meaning the capex-to-revenue ratio is approximately 8:1 to 12:1.\n\nThis gap between infrastructure investment and revenue generation has a precise historical parallel: the telecom and fiber-optic boom of 1998-2001. During that period, telecom companies invested over $500 billion in fiber-optic networks based on projections that internet traffic would grow exponentially. The traffic did grow — but it grew into infrastructure that was dramatically overbuilt relative to the revenue it generated. The result was a cascade of corporate bankruptcies (WorldCom, Global Crossing, JDS Uniphase), massive write-downs, and a 5+ year capex hangover where new investment dried up. Critically, the technology itself was genuinely transformative — the internet did change everything, just as AI proponents claim — but the investment timeline was mismatched with the revenue timeline. Being right about the technology while being wrong about the timeline is financially equivalent to being wrong.\n\nThe danger threshold for the AI capex-revenue gap is the point at which infrastructure overbuild becomes apparent and companies begin to freeze or cut capex in response. This 'capex cliff' is the most dangerous moment because the entire AI supply chain — from chip manufacturers to data center construction firms to power utilities — has been sized for continued growth. When growth deceleration hits, it propagates backward through the supply chain with amplification at each stage (the bullwhip effect). Sequoia Capital's widely circulated '$600B question' analysis explicitly flagged this risk, and Goldman Sachs research has similarly questioned whether the return on AI investment justifies the scale of spending. The gap is not yet triggering alarm bells in equity markets because the hyperscalers generating the capex are enormously profitable companies with the balance sheets to sustain investment losses — unlike the debt-funded telecom companies of 2000. This difference buys time but does not eliminate the risk: if AI revenue does not grow into the infrastructure investment within 2-3 years, the capex freeze will arrive regardless of balance sheet strength.",
      "dataSource": "Sequoia Capital publications, Goldman Sachs macro research",
      "leadTime": "The capex-revenue gap typically persists for 2-4 years before triggering a capex freeze. The telecom parallel suggests the correction comes 3-5 years after peak overbuilding begins.",
      "reliability": "The existence of the gap is well-documented. The timing of any resulting correction is uncertain. Key difference from telecom 2000: current AI spenders are more financially resilient.",
      "historicalPrecedents": [
        "Telecom/fiber-optic boom 1998-2001 ($500B+ invested, mass bankruptcies)",
        "Railroad boom 1840s-1850s (massive overbuilding, eventual consolidation)",
        "Japanese semiconductor overcapacity 1990s"
      ],
      "compoundsWith": [
        "asset-bubble",
        "credit-debt",
        "yield-curve"
      ],
      "tags": [
        "AI-capex",
        "revenue-gap",
        "infrastructure-overbuild",
        "telecom-parallel",
        "Sequoia",
        "Goldman-Sachs",
        "current-reading"
      ]
    },
    {
      "id": 60,
      "sourceId": "cm-061",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "energy-shock",
      "title": "Crude Oil Price - Current Reading (March 2026)",
      "content": "Brent crude oil currently trades at approximately $112.57 per barrel, with West Texas Intermediate (WTI) crossing the $100 threshold. This dramatic surge is directly attributable to the effective closure of the Strait of Hormuz following Operation Epic Fury — the US military campaign targeting Iranian military infrastructure that began in late February 2026. The Strait of Hormuz is the most critical chokepoint in global energy infrastructure: approximately 17.8 million barrels per day of crude oil transit through the strait, representing roughly 20% of total global oil demand. Even a partial disruption to Hormuz traffic creates immediate and severe supply-side pressure on global oil markets.\n\nThe $120 per barrel threshold is widely considered the level at which oil prices begin to function as a massive consumer tax, reducing discretionary spending, compressing corporate margins outside the energy sector, and tipping inflation dynamics from 'manageable' to 'acute.' At $112.57, Brent is rapidly approaching this threshold. The last comparable oil price shock occurred in early 2022 following Russia's invasion of Ukraine, when Brent briefly topped $130. That spike triggered a wave of global inflation that forced central banks worldwide into aggressive tightening cycles, collapsing equity and bond markets simultaneously. Before that, the 2007-2008 oil spike to $147 per barrel coincided with the early stages of the Global Financial Crisis, acting as an accelerant on consumer distress that was already building from the subprime mortgage crisis.\n\nThe current oil price surge is particularly dangerous because it is occurring against a backdrop of already-elevated inflationary concerns, an un-inverting yield curve signaling recession risk, and a Federal Reserve trapped between conflicting mandates. The Fed cannot ignore oil-driven inflation resurgence (which would argue for higher rates or at minimum holding steady), but it also cannot ignore the economic drag from $112 oil (which would argue for rate cuts to cushion the blow). This policy trap is precisely the mechanism that produces stagflation — simultaneous economic contraction and rising prices — which is the most destructive macroeconomic outcome because it eliminates the standard policy responses to both problems. The speed of the current price surge (Brent up approximately 51% in one month) adds to the danger, as rapid price shocks are more economically damaging than gradual price increases of the same magnitude because households and businesses cannot adjust spending patterns quickly enough.",
      "dataSource": "ICE (Brent), CME Group (WTI)",
      "leadTime": "Oil price shocks transmit to consumer prices within 1-3 months and to broader economic activity within 3-6 months.",
      "reliability": "Oil prices are directly observable. The relationship between oil above $120 and recession/inflation has strong historical precedent.",
      "historicalPrecedents": [
        "2022 Russia-Ukraine spike to $130 triggered global inflation wave",
        "2007-2008 spike to $147 coincided with GFC onset",
        "1973 oil embargo quadrupled prices, triggered stagflation",
        "1979 Iranian Revolution doubled prices, triggered deep recession"
      ],
      "compoundsWith": [
        "currency-inflation",
        "yield-curve",
        "geopolitical",
        "systemic-banking"
      ],
      "tags": [
        "oil-shock",
        "Hormuz",
        "Iran",
        "energy-inflation",
        "Operation-Epic-Fury",
        "stagflation-risk",
        "current-reading"
      ]
    },
    {
      "id": 61,
      "sourceId": "cm-062",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "systemic-banking",
      "title": "US Bank Unrealized Losses - Current Reading (March 2026)",
      "content": "US banks held approximately $306.1 billion in unrealized losses on their available-for-sale (AFS) and held-to-maturity (HTM) securities portfolios as of Q4 2025, according to the FDIC Quarterly Banking Profile. This figure represents a meaningful improvement from the SVB-era peak of over $500 billion in late 2022 and early 2023, when the fastest interest rate hiking cycle in four decades caused the market value of banks' bond portfolios to plummet. The improvement reflects a combination of bonds maturing (returning to par value), some portfolio repositioning by more proactive banks, and the modest decline in long-term yields from their October 2023 peaks.\n\nThe $306 billion figure must be understood in the context of the mechanism that collapsed Silicon Valley Bank, Signature Bank, and First Republic Bank in March 2023. These failures were not caused by unrealized losses alone — they were caused by the deadly combination of large unrealized losses AND rapid deposit outflows. When depositors flee, banks must sell securities to meet withdrawal demands. If those securities are held at a loss, the sale crystallizes the paper loss into a real loss, which further erodes confidence, which accelerates deposit flight, which forces more sales — a self-reinforcing doom loop. The critical variable is not the absolute level of unrealized losses but the ratio of unrealized losses to tier 1 capital combined with the proportion of funding from uninsured (runnable) deposits. Banks with large unrealized losses but stable, insured deposit bases are not at imminent risk. Banks with large unrealized losses AND heavy reliance on uninsured deposits are vulnerable to the same dynamic that destroyed SVB.\n\nThe improving trend from $500 billion to $306 billion is genuinely positive, but the current oil-driven rise in inflation expectations could reverse this progress. If long-term yields rise from their current 4.38% back toward the 5.0% peak of October 2023 — a plausible scenario if oil prices remain above $110 and the Fed signals that rate cuts are off the table — unrealized losses would expand again, potentially back toward $400 billion or higher. The FDIC has implemented enhanced monitoring and reporting requirements since the SVB debacle, and the largest banks have stress-tested against further rate increases. However, the regional and community bank sector — which collectively holds a disproportionate share of long-duration securities relative to their capital — remains the most vulnerable segment. A renewed surge in unrealized losses coinciding with deposit competition from money market funds offering 4%+ yields could recreate the conditions that produced the 2023 banking crisis.",
      "dataSource": "FDIC Quarterly Banking Profile",
      "leadTime": "Unrealized losses are a vulnerability that can persist for years or crystallize in days, depending on whether a deposit flight trigger occurs.",
      "reliability": "FDIC data is authoritative. The SVB precedent demonstrated exactly how unrealized losses translate to bank failures. Current $306B is an improvement but remains historically elevated.",
      "historicalPrecedents": [
        "2023 SVB/Signature/First Republic failures (unrealized losses + deposit flight)",
        "2022-2023 peak unrealized losses above $500B",
        "S&L crisis 1980s (interest rate mismatch on bank portfolios)"
      ],
      "compoundsWith": [
        "yield-curve",
        "currency-inflation",
        "credit-debt"
      ],
      "tags": [
        "unrealized-losses",
        "SVB",
        "deposit-flight",
        "FDIC",
        "banking-fragility",
        "HTM",
        "AFS",
        "current-reading"
      ]
    },
    {
      "id": 62,
      "sourceId": "cm-063",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "currency-inflation",
      "title": "Fed Funds Rate - Current Reading (March 2026)",
      "content": "The Federal Reserve's target range for the federal funds rate stands at 3.50-3.75% following the March 2026 FOMC meeting. This represents a cumulative 175 basis points of cuts from the cycle peak of 5.25-5.50%, reflecting the Fed's response to a gradually softening labor market and moderating core inflation through much of 2025. However, the pace of cuts has slowed dramatically: after cutting 50 basis points in September 2024 and 25 basis points each in November and December 2024, followed by further reductions through 2025, the Fed is now widely expected to deliver only one additional cut in 2026, bringing the rate to approximately 3.25-3.50% by year-end.\n\nThe Fed finds itself in what traders call a 'policy trap' — caught between two conflicting mandates with no clean escape. On one side, the labor market is softening, business investment surveys are weakening, and yield curve signals suggest recession risk is elevated. These factors argue forcefully for faster rate cuts to cushion the economy. On the other side, the oil price shock from the Hormuz crisis has injected a massive inflationary impulse into the economy, consumer inflation expectations are rising, and the experience of the 2021-2022 inflation surprise has made the Fed extremely cautious about declaring victory prematurely. Cutting rates aggressively into an oil-driven inflation surge risks a repeat of the 1970s error when Arthur Burns' premature easing allowed inflation to become entrenched, ultimately requiring Paul Volcker's devastating 20% rate to break it.\n\nThe current 3.50-3.75% rate provides approximately 350 basis points of room to cut before hitting the zero lower bound — substantially more policy ammunition than existed in early 2020 when the Fed was forced to cut from 1.50-1.75% to near zero. This is a meaningful buffer that gives the Fed more room to respond to a crisis than it had going into the pandemic. However, the operational constraint is not the zero bound but the inflation constraint: if oil-driven inflation pushes headline CPI back above 4%, the Fed may be unable to cut at all without losing credibility and de-anchoring inflation expectations. Markets had priced in an aggressive cutting cycle through 2025 and early 2026; the violent recalibration to 'only one cut in 2026' has itself created financial stress, as asset valuations, corporate financing plans, and consumer borrowing decisions were all premised on substantially lower rates than are now expected.",
      "dataSource": "CME FedWatch Tool, Federal Reserve Board of Governors",
      "leadTime": "Fed policy changes transmit to the broader economy with a 6-18 month lag. Current rate level is constraining economic activity from decisions made 12+ months ago.",
      "reliability": "The fed funds rate is directly observable. CME FedWatch probabilities provide reliable near-term guidance. The policy trap dynamic is clearly visible in Fed communications and market pricing.",
      "historicalPrecedents": [
        "1970s Burns Fed cut too early, inflation re-accelerated",
        "2020 emergency cut from 1.50% to 0% in response to COVID",
        "2007-2008 Fed cut from 5.25% to near-zero as crisis unfolded"
      ],
      "compoundsWith": [
        "energy-shock",
        "yield-curve",
        "credit-debt",
        "systemic-banking",
        "asset-bubble"
      ],
      "tags": [
        "Fed-funds",
        "rate-cuts",
        "higher-for-longer",
        "monetary-policy",
        "policy-trap",
        "inflation-constraint",
        "current-reading"
      ]
    },
    {
      "id": 63,
      "sourceId": "cm-064",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "energy-shock",
      "title": "Compound Scenario — Geopolitical Stagflation Shock (March 2026)",
      "content": "A compound crisis scenario is actively building in March 2026 that directly parallels the two most devastating stagflation episodes in modern economic history: the 1973 oil embargo and the 1979 Iranian Revolution. The core mechanism is the simultaneous collision of an energy supply shock (oil past $110 due to Hormuz disruption) with a yield curve un-inversion signal (both 10Y-3M and 10Y-2Y spreads positive after prolonged inversion). When energy-driven inflation spikes just as the yield curve completes its classic pre-recession un-inversion pattern, central banks face an impossible policy choice: they cannot cut rates to support a weakening economy because doing so would pour fuel on inflation that is already being stoked by energy prices.\n\nThe 1973 parallel is instructive. The Arab oil embargo quadrupled oil prices virtually overnight, injecting a massive inflationary shock into an economy that was already showing signs of cyclical weakness. The Fed, under Arthur Burns, attempted to thread the needle by maintaining accommodative policy to avoid recession while hoping oil prices would normalize. Instead, inflation became entrenched, the economy entered a deep recession anyway, and the resulting 'stagflation' — a word that had to be invented because economists previously believed simultaneous inflation and recession were theoretically impossible — destroyed wealth across every asset class. Neither stocks (which fell 48% peak-to-trough in 1973-74), bonds (which lost value as inflation eroded purchasing power), nor cash (which lost purchasing power even faster) provided shelter. The 1979 sequel was equally devastating: the Iranian Revolution disrupted oil supply, inflation re-accelerated to 14.8%, and the economy required Volcker's aggressive 20% interest rate to break the cycle — at the cost of the deepest recession since the Great Depression.\n\nThe current configuration maps disturbingly well onto this template. Oil has surged 51% in one month to $112 due to the Hormuz disruption — a supply-side shock with no quick resolution given the military nature of the disruption. Both yield curve measures have un-inverted in the classic pre-recession pattern. The Fed is trapped at 3.50-3.75% with only one expected cut in 2026, unable to provide the aggressive stimulus the weakening economy needs because inflation is resurging. The VIX at 31 with an inverted term structure confirms that institutional investors see imminent danger. The key question is whether the current setup will follow the full 1973/1979 stagflation template or whether structural differences — household balance sheet resilience, reduced US oil import dependence, modern supply chain flexibility — provide enough cushion to avoid the worst outcome. But the compound indicator pattern is active and it is the single most dangerous configuration in the current risk landscape.",
      "dataSource": "Synthesis of cm-049, cm-050, cm-054, cm-061, cm-063",
      "leadTime": "Stagflation scenarios typically take 3-9 months to fully manifest after the energy shock trigger, as oil price increases transmit through supply chains to consumer prices.",
      "reliability": "The individual indicators are reliable. The compound interaction is supported by strong historical precedent (1973, 1979). The key uncertainty is whether structural differences from the 1970s provide sufficient protection.",
      "historicalPrecedents": [
        "1973 oil embargo + recession = stagflation",
        "1979 Iranian Revolution + recession = stagflation sequel",
        "2022 Russia-Ukraine oil spike + tightening cycle (milder version)"
      ],
      "compoundsWith": [
        "yield-curve",
        "currency-inflation",
        "geopolitical",
        "systemic-banking",
        "credit-debt"
      ],
      "tags": [
        "stagflation",
        "oil-shock",
        "yield-curve",
        "Fed-trap",
        "1973-parallel",
        "1979-parallel",
        "compound-scenario",
        "current-reading"
      ]
    },
    {
      "id": 64,
      "sourceId": "cm-065",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "yield-curve",
      "title": "Compound Scenario — Yield Curve Un-Inversion Trap (March 2026)",
      "content": "A classic pre-recession compound configuration is now firmly in place: both major yield curve measures — the 10Y-3M spread at +0.71% and the 10Y-2Y spread at +0.65% — have un-inverted from their prolonged inversion, while the VIX term structure has simultaneously flipped into deep backwardation at -6.89%. This three-indicator convergence is one of the most historically reliable pre-recession signals because each component captures a different dimension of the same underlying dynamic: the economy is transitioning from late-cycle stress to active contraction.\n\nThe critical insight that most market participants miss is that recessions do not begin while the yield curve is inverted — they begin after it un-inverts. The inversion is the warning shot; the un-inversion is the execution. This is because the un-inversion typically occurs when central banks begin cutting short-term rates in response to deteriorating economic data, which steepens the curve back to positive territory. The rate cuts are not preventive — they are reactive, arriving after the damage is already embedded in the economic pipeline. Simultaneously, when VIX futures flip to backwardation, it means institutional investors are paying premium prices for immediate downside protection — they see the threat as imminent rather than distant. The combination of 'the economy is weakening' (yield curve) and 'something bad is about to happen' (VIX backwardation) creates a self-reinforcing dynamic where defensive positioning itself contributes to the tightening of financial conditions.\n\nThe direct historical parallels are late 2000 and summer 2007, both of which preceded devastating economic downturns. In late 2000, both yield curve measures un-inverted as the Fed began cutting rates in response to the tech bust, while VIX entered backwardation as options traders positioned for the fallout. The recession officially began in March 2001. In summer 2007, the same pattern played out: yield curves steepened positive, VIX backwardation appeared, and the Great Recession began in December 2007. In both cases, the compound signal preceded the recession by 3-9 months, though in both cases the equity market had already begun declining by the time the compound configuration was fully formed. The current setup has been building since late 2025, and March 2026 marks the first month where all three components are simultaneously in the danger zone. Whether this leads to a 2001-style moderate recession or a 2007-style severe contraction depends heavily on whether the oil price shock (cm-061) amplifies the downturn into full stagflation (cm-064).",
      "dataSource": "Synthesis of cm-049, cm-050, cm-054",
      "leadTime": "3-9 months from full compound configuration to recession onset, based on 2000 and 2007 precedents.",
      "reliability": "High — the un-inversion + VIX backwardation combination has preceded every major recession in the data period. False positive risk exists but is low when both yield curve measures and VIX backwardation are simultaneously signaling.",
      "historicalPrecedents": [
        "Late 2000: un-inversion + VIX backwardation preceded March 2001 recession",
        "Summer 2007: same pattern preceded December 2007 Great Recession",
        "1989: similar configuration preceded July 1990 recession"
      ],
      "compoundsWith": [
        "energy-shock",
        "currency-inflation",
        "asset-bubble",
        "systemic-banking",
        "credit-debt"
      ],
      "tags": [
        "un-inversion",
        "VIX-backwardation",
        "recession-timing",
        "2000-parallel",
        "2007-parallel",
        "compound-scenario",
        "current-reading"
      ]
    },
    {
      "id": 65,
      "sourceId": "cm-066",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "asset-bubble",
      "title": "Compound Scenario — Mega-Cap Bubble Deflation (March 2026)",
      "content": "Three powerful indicators are converging to create what may be the most concentrated equity market vulnerability since March 2000. The Magnificent 7 now represent 33.3% of the S&P 500 market capitalization — a concentration that exceeds any modern precedent. Simultaneously, the Shiller CAPE ratio stands at 38.93, more than double the historical average and within striking distance of the dot-com era peak of 44. And underpinning the valuation of the largest Mag 7 members is an AI capital expenditure cycle that has committed over $600 billion against a fraction of that in actual AI revenue — a capex-to-revenue gap that directly mirrors the telecom infrastructure overbuild of 1998-2001.\n\nThe danger of this triple convergence is that each component reinforces the others. The Mag 7's extreme concentration means that the S&P 500 is essentially a leveraged bet on AI-adjacent mega-caps. The elevated CAPE ratio means that this bet is priced for perfection — any disappointment in earnings, growth, or the AI narrative will be punished disproportionately because there is no valuation cushion. And the AI capex-revenue gap means that the fundamental justification for these valuations depends on a revenue ramp that has not yet materialized. If AI revenue growth disappoints expectations — not AI being a failure, just AI revenue growing at 30% instead of the 100%+ that justifies current multiples — the capex freeze would cascade through the supply chain, hitting NVIDIA's revenue, which would hit the Mag 7's collective earnings, which would hit the S&P 500's cap-weighted performance, which would hit every index fund and ETF in the world.\n\nThe March 2000 parallel is almost structurally identical. In early 2000, the top five stocks (Microsoft, GE, Cisco, Intel, Walmart) dominated index performance, the CAPE was at 44, and the telecom/internet infrastructure buildout had committed hundreds of billions against a fraction of that in end-user revenue. The correction that followed was prolonged and painful: the S&P 500 fell 49% from peak to trough, the NASDAQ fell 78%, and it took the S&P 500 seven years to recover its 2000 peak in nominal terms. Critically, the technology itself was not wrong — the internet did transform the economy — but the valuation overshoot took years to correct. The current setup differs in one important respect: the Mag 7 companies are far more profitable than the dot-com era darlings, with massive cash flows and strong balance sheets. This financial resilience means the correction, if it comes, would likely be a slow deflation (multiple compression, rotation to equal-weight) rather than a rapid crash. But for cap-weighted index investors — which includes virtually every passive retirement fund in America — the impact would be substantial regardless of the speed.",
      "dataSource": "Synthesis of cm-052, cm-058, cm-060",
      "leadTime": "Bubble deflation timing is inherently unpredictable. The 2000 bubble took 6 months from first crack (March 2000) to confirmed bear market. Slow deflation scenarios can take 12-24 months.",
      "reliability": "The individual indicators are reliable. The compound interaction is well-supported by the 2000 precedent. The key uncertainty is timing and whether the correction is sharp or gradual.",
      "historicalPrecedents": [
        "March 2000 dot-com bubble (CAPE 44, extreme concentration, telecom capex gap)",
        "Nifty Fifty deflation 1973-74 (concentration + overvaluation)",
        "Japan Nikkei 1989-90 (extreme valuations + concentrated gains)"
      ],
      "compoundsWith": [
        "ai-specific",
        "yield-curve",
        "credit-debt",
        "currency-inflation"
      ],
      "tags": [
        "Mag-7",
        "CAPE",
        "AI-bubble",
        "rotation",
        "dot-com-parallel",
        "cap-weighted-risk",
        "compound-scenario",
        "current-reading"
      ]
    },
    {
      "id": 66,
      "sourceId": "cm-067",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "energy-shock",
      "title": "90-Day Change — Oil Price Breakout (March 2026)",
      "content": "In the span of approximately 30 days following the initiation of strikes on Iranian military infrastructure in late February 2026, Brent crude oil prices exploded by approximately 51%, surging from the mid-$70s to cross $112 per barrel. This is the largest single-indicator percentage movement in the crisis monitor's tracked universe over the past 90 days, and it represents the kind of rapid exogenous shock that cascades through the entire economic system in ways that slower-building risks cannot. The speed of the move matters as much as the magnitude: a 51% increase over 12 months allows businesses and consumers to adjust gradually; a 51% increase in one month is a body blow that lands before any adaptation is possible.\n\nThe transmission channels from a rapid oil price surge to broader economic damage are well-documented and begin affecting the economy within weeks. First, gasoline and diesel prices spike immediately, functioning as a regressive tax that disproportionately impacts lower-income households and transportation-dependent businesses. Second, jet fuel and shipping costs jump, increasing the cost of everything that moves by air, rail, or road — which is effectively everything. Third, petrochemical feedstock costs rise, impacting manufacturing inputs from plastics to fertilizers. Fourth, and most critically for financial markets, inflation expectations re-anchor higher, which constrains the Federal Reserve's ability to provide the rate cuts that the weakening economy needs. Each of these channels operates on a different timeline: gasoline hits immediately, transportation costs within 2-4 weeks, manufacturing inputs within 1-3 months, and inflation expectations take 3-6 months to fully recalibrate.\n\nThe geopolitical dimension adds a layer of persistence that distinguishes this from a typical demand-driven or OPEC-driven price spike. Demand-driven spikes (like the 2008 surge) self-correct as high prices destroy demand. OPEC-driven spikes can be resolved through production decisions. But a military conflict affecting the Strait of Hormuz has no market-based self-correction mechanism: the strait remains disrupted until the military situation resolves, regardless of what oil prices do. This means the standard assumption that 'high prices cure high prices' does not apply in the current scenario. Insurance premiums for tanker traffic through the region have skyrocketed, further adding to the effective cost of oil even for cargoes that successfully transit alternative routes. The combination of speed, magnitude, and geopolitical persistence makes this oil price breakout the single most destabilizing near-term development in the global economic landscape.",
      "dataSource": "ICE (Brent), CME Group (WTI), shipping insurance market data",
      "leadTime": "Oil price shocks transmit within weeks (gasoline) to months (manufacturing costs, inflation expectations).",
      "reliability": "Price data is directly observable. The 51% surge and its geopolitical driver are facts, not estimates. The persistence of geopolitically-driven supply disruptions is historically well-documented.",
      "historicalPrecedents": [
        "2022 Russia-Ukraine invasion (Brent spiked ~60% over 2 months)",
        "1990 Gulf War (oil doubled in 3 months)",
        "1979 Iranian Revolution (oil doubled over 6 months)",
        "1973 oil embargo (oil quadrupled in 3 months)"
      ],
      "compoundsWith": [
        "currency-inflation",
        "yield-curve",
        "geopolitical",
        "systemic-banking",
        "credit-debt"
      ],
      "tags": [
        "oil-shock",
        "Iran",
        "supply-chain-inflation",
        "rapid-change",
        "Hormuz",
        "transmission-channels",
        "90-day-change",
        "current-reading"
      ]
    },
    {
      "id": 67,
      "sourceId": "cm-068",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "currency-inflation",
      "title": "90-Day Change — Higher For Longer Re-pricing (March 2026)",
      "content": "Over the past 90 days, one of the most significant shifts in the financial landscape has been the violent recalibration of interest rate expectations. In January 2026, fed funds futures markets were pricing in an aggressive rate-cutting cycle, with expectations for 3-4 additional cuts in 2026 that would bring the federal funds rate below 3.0% by year-end. Over the subsequent 90 days, this expectation has been systematically demolished. As of late March 2026, the market now expects only one additional cut in 2026, bringing the rate to approximately 3.25-3.50%, with the Fed holding at 3.50-3.75% for the foreseeable future. The 10-year Treasury yield has risen back to 4.38%, reflecting the market's acknowledgment that the 'higher for longer' narrative has returned with a vengeance.\n\nThe catalysts for this violent repricing are multiple and self-reinforcing. The Hormuz oil shock injected a fresh inflationary impulse that makes further rate cuts politically and economically untenable for the Fed. Services inflation, which had been gradually moderating through 2025, showed signs of re-accelerating in early 2026 data. The labor market, while softening, has not deteriorated to the degree that would give the Fed cover for aggressive easing. And the Fed's own communications have shifted noticeably more hawkish, with several FOMC members publicly emphasizing the risks of cutting too aggressively into a resurgent inflation environment. The collective effect has been a repricing of the entire yield curve: short-term rates stayed higher than expected, long-term rates rose on inflation expectations, and the gap between what markets wanted (aggressive cuts) and what they are getting (perhaps one cut) has created a financial conditions tightening that amplifies the economic drag.\n\nThe 'higher for longer' repricing has cascading effects across virtually every asset class and financial decision. Corporate borrowers who had deferred refinancing in anticipation of lower rates are now facing the reality that rates may not decline meaningfully, forcing either refinancing at current elevated levels or extending maturity at unfavorable terms. Homebuyers who had paused purchases waiting for lower mortgage rates face the prospect that 7%+ mortgage rates persist through 2026. Equity valuations that were premised on a return to lower discount rates must be re-evaluated. Commercial real estate borrowers facing maturing loans had been counting on lower rates to make refinancing viable — the 'higher for longer' repricing eliminates that escape route. Perhaps most importantly, the repricing itself feeds back into financial conditions: higher-than-expected rates tighten financial conditions, which slows the economy, which increases recession risk, which the Fed cannot offset with cuts because of inflation — completing the policy trap described in cm-063.",
      "dataSource": "CME FedWatch Tool, Federal Reserve Board of Governors, Treasury yield data",
      "leadTime": "Rate expectation repricing transmits to the real economy over 3-12 months as financial decisions (refinancing, investment, borrowing) adjust to the new rate path.",
      "reliability": "Market pricing data is directly observable. The shift from 3-4 expected cuts to 1 expected cut is a factual observation. The cascading effects on corporate refinancing, housing, and CRE are well-understood transmission mechanisms.",
      "historicalPrecedents": [
        "Late 2023 'higher for longer' repricing (10Y yield touched 5%)",
        "2013 Taper Tantrum (violent rate repricing on unexpected Fed hawkishness)",
        "1994 bond market massacre (unexpected tightening destroyed bond portfolios)"
      ],
      "compoundsWith": [
        "energy-shock",
        "yield-curve",
        "credit-debt",
        "systemic-banking",
        "asset-bubble"
      ],
      "tags": [
        "rate-expectations",
        "higher-for-longer",
        "Fed",
        "bond-market",
        "repricing",
        "financial-conditions",
        "90-day-change",
        "current-reading"
      ]
    },
    {
      "id": 68,
      "sourceId": "cm-069",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "systemic-banking",
      "title": "US Treasury Basis Trade (Hedge Fund Repo Leverage)",
      "content": "Hedge funds have built an enormous leveraged position exploiting tiny price discrepancies between cash US Treasury bonds and Treasury futures contracts — the so-called 'basis trade.' The mechanics are deceptively simple: buy the slightly cheaper cash Treasury, sell the slightly richer futures contract, and pocket the spread at convergence. The problem is that the spread is measured in basis points, so generating meaningful returns requires extreme leverage. Current estimates place the notional exposure at $1.85-2 trillion, largely domiciled in Cayman Islands entities that fall outside direct US regulatory oversight. Gross leverage ratios average over 600%, meaning these funds are borrowing six dollars for every dollar of equity. The funding comes overwhelmingly from overnight and short-term repurchase agreements (repo), creating a maturity mismatch of staggering proportions: long-dated Treasury positions funded by borrowing that must be rolled every single day.\n\nThe systemic danger emerges when volatility spikes. With the VIX at 31, Treasury market volatility (measured by the MOVE index) is already elevated. If volatility surges further — from a geopolitical shock, a failed Treasury auction, or a sudden shift in Fed expectations — repo counterparties will immediately increase haircuts (the collateral margin required on each loan). A haircut increase from 2% to 4% on a $2 trillion position would require $40 billion in additional collateral overnight. Funds that cannot meet these calls must liquidate Treasury positions, but they are all holding the same trade. The result is synchronized selling of the most important collateral asset in the global financial system. This is not a theoretical risk: in March 2020, basis trade unwinds contributed directly to Treasury market dysfunction so severe that the Federal Reserve was forced to purchase $1.6 trillion in Treasuries in three weeks to restore market functioning.\n\nThe LTCM parallel from 1998 is instructive but understates current risk. LTCM's notional exposure was approximately $125 billion — roughly 1/15th of today's basis trade. LTCM was one fund; today's basis trade is distributed across dozens of multi-strategy hedge funds, creating a coordination problem that is arguably worse: no single entity can be bailed out to stop the cascade. The Cayman domiciliation further complicates matters, as US regulators have limited visibility into the true scale of positions and limited authority to impose orderly wind-downs. The Office of Financial Research (OFR) and the Financial Stability Oversight Council (FSOC) have both flagged this risk, but regulatory action has been slow. Meanwhile, the trade continues to grow because in calm markets it generates reliable returns — the classic Minsky dynamic of stability breeding instability.\n\nThe trigger mechanism is a sudden Treasury volatility jump or repo funding cost spike that triggers automatic margin calls across multiple prime brokers simultaneously. Because the trade is crowded and the positions are similar, the forced selling would be one-directional, potentially freezing the world's primary collateral market and cascading into every asset class that relies on Treasury collateral for funding — which is effectively all of them.",
      "dataSource": "OFR Repo Market Data, CFTC Traders in Financial Futures (leveraged fund short open interest), Federal Reserve Financial Stability Report, BIS Quarterly Review",
      "leadTime": "Hours to days. Repo funding stress can emerge overnight. The March 2020 episode went from calm to crisis in approximately 5 trading days.",
      "reliability": "High. Position sizes are observable through CFTC data and repo market volumes. The mechanical trigger (haircut increase → forced selling) is well-understood. The 2020 precedent validates the transmission mechanism.",
      "historicalPrecedents": [
        "LTCM 1998 ($125B notional, required Fed-coordinated bailout)",
        "March 2020 Treasury market dysfunction ($1.6T Fed intervention)",
        "1998 Russian default triggering leveraged carry trade unwinds"
      ],
      "compoundsWith": [
        "systemic-banking",
        "yield-curve",
        "asset-bubble"
      ],
      "tags": [
        "basis-trade",
        "repo",
        "hedge-fund-leverage",
        "LTCM-parallel",
        "Treasury-market",
        "collateral"
      ]
    },
    {
      "id": 69,
      "sourceId": "cm-070",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "credit-debt",
      "title": "2026 Corporate Debt Maturity Wall",
      "content": "The global corporate debt market faces its most concentrated maturity wall since the Global Financial Crisis. Approximately $1.8 trillion in US corporate debt is scheduled to mature across 2025-2026, with the global figure reaching $2.78 trillion. The critical context is that the vast majority of this debt was issued during the zero-rate era of 2020-2021, when the Federal Reserve held rates near zero and corporate bond markets were awash in liquidity. Companies that locked in 3-4% coupon rates must now refinance at 6-8% or higher — a 2-3x increase in borrowing costs that flows directly to the bottom line. For companies already operating on thin margins, this refinancing cost increase is the difference between solvency and distress.\n\nThe oil price surge to $112 per barrel has dramatically worsened this dynamic by anchoring the 'higher for longer' rate environment. With energy-driven inflation pressures re-emerging, the Federal Reserve has minimal room to cut rates, eliminating the escape valve that many corporate treasurers had been counting on. The combination of higher operating costs (inflation, energy) and higher financing costs (refinancing at elevated rates) creates a simultaneous margin compression that is particularly lethal for BBB-rated issuers — the lowest rung of investment-grade debt. BBB-rated corporate bonds represent approximately $3.5 trillion of the US investment-grade market, and any wave of downgrades from BBB to BB (junk) would trigger forced selling by the enormous pool of institutional investors — pension funds, insurance companies, mutual funds — that are mandated to hold only investment-grade securities. This 'fallen angel' cascade was a major amplifier during March 2020 and is an even larger risk today given the growth of the BBB cohort.\n\nThe historical parallel to the 1989-1990 junk bond crash is apt but insufficient. That episode involved approximately $200 billion in high-yield debt and was concentrated in a handful of issuers associated with leveraged buyout excess. Today's maturity wall is an order of magnitude larger, spans thousands of issuers across every sector, and coincides with multiple other stress vectors (oil shock, geopolitical uncertainty, elevated equity valuations). The CLO (Collateralized Loan Obligation) market adds another layer of vulnerability: CLOs hold approximately $1 trillion in leveraged loans, many of which are floating-rate instruments whose borrowers are already struggling with elevated interest payments. A wave of defaults in the leveraged loan market would cascade through CLO tranches, potentially impairing even senior tranches that were previously considered safe.\n\nThe trigger is not a single event but a grinding process: cash flow exhaustion as operating costs (inflation-driven) and refinancing costs (rate-driven) bite simultaneously. Companies that could survive one headwind — either higher rates or higher input costs — cannot survive both. The most vulnerable are 'zombie companies' that have been surviving only by rolling over cheap debt: estimates suggest 10-15% of publicly traded US companies qualify as zombies by the standard definition (interest coverage ratio below 1.0x for three consecutive years).",
      "dataSource": "S&P Global Ratings, Moody's default rate tracker, ICE BofA US High Yield Option-Adjusted Spread (OAS), Federal Reserve Financial Stability Report",
      "leadTime": "3-12 months. Maturity dates are known in advance. Default rates lag economic deterioration by 6-9 months. The BBB downgrade cascade, once triggered, can accelerate rapidly (weeks).",
      "reliability": "High. Maturity schedules are publicly available and precise. The mechanical relationship between higher rates and refinancing costs is arithmetic, not speculative. The BBB fallen-angel risk is well-documented by rating agencies.",
      "historicalPrecedents": [
        "1989-1990 Junk Bond Crash (Drexel Burnham collapse)",
        "2001-2002 post-dot-com corporate default wave (WorldCom, Enron)",
        "March 2020 corporate bond market freeze (Fed created emergency facilities)",
        "2008-2009 corporate default spike (14.7% high-yield default rate)"
      ],
      "compoundsWith": [
        "credit-debt",
        "currency-inflation",
        "energy-shock"
      ],
      "tags": [
        "maturity-wall",
        "refinancing",
        "junk-bonds",
        "CLOs",
        "zombie-companies",
        "downgrade-cascade"
      ]
    },
    {
      "id": 70,
      "sourceId": "cm-071",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "systemic-banking",
      "title": "Shadow Insurance & PE Annuity Capture",
      "content": "Over the past decade, private equity firms have quietly executed one of the most consequential financial engineering operations in modern history: the acquisition and transformation of the US life insurance industry. Firms including Apollo, KKR, Blackstone, and Brookfield have acquired or affiliated with life insurance companies managing hundreds of billions in policyholder assets. The strategy is straightforward: use the steady stream of insurance premiums and annuity payments as a captive source of long-duration funding, then invest that capital into higher-yielding but less liquid assets — private credit, CLOs, structured products, and direct lending — that generate the spread needed to justify the PE firm's returns. As of the latest available data, approximately $849 billion in private placements and illiquid credit instruments sit on life insurer balance sheets, a dramatic shift from the traditional insurance investment model of high-quality bonds and mortgages.\n\nThe regulatory arbitrage is achieved through a mechanism known as 'shadow reinsurance.' PE-affiliated insurers cede their riskiest liabilities to affiliated or captive reinsurance entities domiciled in Bermuda, the Cayman Islands, or other offshore jurisdictions with lighter capital requirements. This allows the US-domiciled insurer to report strong capital ratios while the actual risk has been transferred to entities with less regulatory oversight, weaker capital buffers, and limited transparency. The National Association of Insurance Commissioners (NAIC) has raised concerns about this practice, but the offshore nature of the reinsurance entities places them largely beyond US regulatory reach. The result is a system where the risks borne by policyholders — retirees depending on annuity payments for their living expenses — are backed by assets and capital structures that those policyholders cannot see or evaluate.\n\nThe trigger scenario is a dual shock: credit defaults in the illiquid asset portfolio coinciding with a surge in policyholder annuity surrenders driven by the cost of living crisis. When policyholders surrender annuities (withdraw their money), insurers must liquidate assets to meet the redemptions. But the illiquid private credit and structured products that now dominate these portfolios cannot be sold quickly without severe price concessions. This liquidity mismatch — liquid liabilities backed by illiquid assets — is the classic precondition for a run. The historical parallel is the Executive Life Insurance Company collapse of 1991, which was triggered by exactly this dynamic: an insurer that had invested heavily in Drexel Burnham junk bonds faced a wave of policyholder surrenders when Drexel collapsed, could not liquidate its illiquid portfolio fast enough, and was seized by California regulators.\n\nThe scale today dwarfs the Executive Life episode. The $849 billion in illiquid assets on life insurer balance sheets is backstopped by capital structures that have been optimized for return, not resilience. If a credit cycle turns — and the maturity wall, rising rates, and oil shock are all pushing in that direction — the PE-affiliated insurers would face simultaneous pressure from asset deterioration and liability withdrawals. The CLO market, which is a significant component of these portfolios, would face forced selling that cascades into the broader credit market. The political dimension adds fuel: millions of retirees discovering that their annuity payments are backed by Bermuda-domiciled entities holding distressed private credit would create regulatory and political firestorm.",
      "dataSource": "NAIC Capital Adequacy reports, Federal Insurance Office (FIO) systemic risk publications, AM Best insurer financial strength ratings, Bermuda Monetary Authority disclosures",
      "leadTime": "6-18 months. Illiquid asset deterioration is slow-moving but surrender waves can accelerate rapidly once confidence erodes. Regulatory intervention adds unpredictable timing.",
      "reliability": "Moderate-high. Asset allocation data is available through NAIC filings. The liquidity mismatch is structural and measurable. The trigger (credit cycle turn + surrender wave) depends on multiple conditions aligning.",
      "historicalPrecedents": [
        "Executive Life Insurance Company collapse 1991 (junk bond portfolio + surrender wave)",
        "AIG 2008 (insurance entity with concentrated financial risk, required $182B bailout)",
        "Equitable Life (UK) 2000 (guaranteed annuity obligations exceeded asset values)"
      ],
      "compoundsWith": [
        "credit-debt",
        "systemic-banking"
      ],
      "tags": [
        "shadow-insurance",
        "private-equity",
        "annuities",
        "Bermuda-reinsurance",
        "illiquid-assets",
        "regulatory-arbitrage"
      ]
    },
    {
      "id": 71,
      "sourceId": "cm-072",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "asset-bubble",
      "title": "0DTE Options Gamma Squeeze Risk",
      "content": "Zero-days-to-expiration (0DTE) options — contracts that expire on the same day they are traded — have exploded to represent 56% of all S&P 500 options volume, a structural transformation of the derivatives market that has occurred in barely two years. The daily notional exposure is measured in trillions of dollars. These instruments are used by retail traders seeking leveraged directional bets, institutional investors managing intraday hedging, and volatility-targeting strategies that algorithmically adjust exposure based on realized volatility. The market ecosystem has reorganized around this massive flow: options market makers — the firms that provide liquidity by taking the other side of these trades — now hold enormous negative gamma positions that mechanically force them to trade in the same direction as the market move, amplifying rather than dampening volatility.\n\nThe mechanics of the gamma squeeze are precise and mechanical. When market makers sell options to buyers, they hedge their exposure by buying or selling the underlying S&P 500 futures. Negative gamma means that as the market falls, their hedging algorithms require them to sell more futures — pushing the market further down — and as the market rises, they must buy more futures — pushing the market further up. In normal conditions, this creates manageable oscillations around key 'gamma wall' strike prices where large option open interest is concentrated. But when VIX is elevated (currently 31), option premiums are high, gamma exposure is extreme, and market liquidity is thin. If an intraday shock — a surprise economic data release, a geopolitical headline, or a large institutional liquidation — pushes the market through a gamma wall strike price, the mechanical hedging response becomes self-reinforcing: dealers sell into a falling market, which pushes it through the next gamma level, triggering more selling, in a cascade that can produce flash-crash dynamics within minutes.\n\nThe historical parallels illuminate the risk but understate the current scale. The February 2018 'Volmageddon' event saw the XIV (inverse volatility ETN) collapse 96% in a single day as volatility-targeting strategies created a self-reinforcing selling cascade. The 1987 crash was amplified by 'portfolio insurance' — a strategy that, like today's 0DTE gamma hedging, required selling into a falling market. Both episodes demonstrated that strategies which individually appear to reduce risk can collectively create systemic risk when they all attempt to execute the same hedging trade simultaneously. Today's 0DTE market is orders of magnitude larger than either precedent, and the speed of algorithmic execution means that the cascade can develop faster than any human intervention can contain it.\n\nThe trigger scenario requires an intraday macro shock while VIX is elevated and S&P 500 futures liquidity is thin — conditions that already exist. A significant portion of 0DTE volume is concentrated at round-number strike prices (e.g., 5000, 5100, 5200 on the S&P 500), creating identifiable 'break points' where a breach would trigger the most violent dealer hedging flows. SpotGamma and similar analytics providers track these levels in real-time, but the information is widely available, meaning that traders front-running an anticipated gamma cascade could accelerate the very event they are trying to profit from.",
      "dataSource": "CBOE 0DTE volume data, SpotGamma GEX models, Tier1Alpha gamma analysis, CME S&P 500 futures liquidity data",
      "leadTime": "Minutes to hours. 0DTE gamma cascades are intraday events. The structural conditions (high VIX, thin liquidity, concentrated gamma) can persist for weeks before a catalyst triggers the cascade.",
      "reliability": "High for structural assessment. The gamma mechanics are mathematical, not speculative. The specific trigger timing is unpredictable. The precedents (1987, 2018) validate the self-reinforcing selling mechanism.",
      "historicalPrecedents": [
        "Volmageddon February 2018 (XIV collapse, vol-targeting cascade)",
        "October 1987 crash (portfolio insurance amplification)",
        "August 2015 flash crash (ETF-underlying disconnect + algorithmic selling)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "systemic-banking"
      ],
      "tags": [
        "0DTE",
        "gamma-squeeze",
        "options",
        "flash-crash",
        "Volmageddon-parallel",
        "market-makers",
        "procyclical"
      ]
    },
    {
      "id": 72,
      "sourceId": "cm-073",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "systemic-banking",
      "title": "CCP Margin Procyclicality (Clearinghouse Cascade)",
      "content": "Central Counterparty Clearinghouses (CCPs) were designed to reduce systemic risk by standing between derivative buyers and sellers, ensuring that a default by one party does not cascade to the other. In the aftermath of the 2008 crisis, regulators mandated that the vast majority of over-the-counter (OTC) derivatives be centrally cleared through CCPs — a well-intentioned reform that has inadvertently concentrated enormous risk in a handful of nodes. Over 60% of global OTC derivatives are now centrally cleared, with tens of trillions in notional value concentrated primarily in LCH (London), CME Clearing (Chicago), and ICE Clear (Atlanta/London). These entities are now the most systemically important financial institutions in the world, yet they operate with relatively thin capital buffers relative to the exposures they intermediate.\n\nThe procyclicality problem is structural and well-documented by the BIS and CPMI-IOSCO. CCPs set Initial Margin (IM) requirements using volatility-sensitive models that automatically demand more collateral when markets become turbulent. During the March 2020 crisis, CCPs issued margin calls totaling hundreds of billions of dollars in a matter of days, forcing banks and funds to liquidate performing assets to raise cash — the 'dash for cash' that froze Treasury, corporate bond, and money markets simultaneously. The UK's LDI (Liability-Driven Investment) pension crisis of September 2022 provided another vivid demonstration: as gilt yields spiked, LCH's margin calls on interest rate swaps forced pension funds to dump gilts to meet the calls, which pushed yields higher, which triggered more margin calls, in a spiral that required Bank of England emergency intervention to halt.\n\nThe nightmare scenario for CCPs is a sustained multi-day, multi-asset volatility spike — equities, bonds, oil, and currencies all moving violently simultaneously. In this scenario, the margin models for equity derivatives, interest rate swaps, commodity futures, and FX forwards would all demand significantly higher collateral at the same time, from the same set of clearing members (the Global Systemically Important Banks — GSIBs). The aggregate margin call would be unmeetable through normal liquidity channels, forcing banks to choose between meeting CCP margin calls and meeting their own obligations to depositors and counterparties. A clearing member default at a major CCP would trigger the CCP's 'default waterfall' — a sequence of loss-absorption mechanisms that includes haircutting the gains of non-defaulting members, a mechanism that could spread the crisis to solvent institutions.\n\nThe regulatory response to CCP risk has been slow relative to the growth of centrally cleared exposures. Stress testing of CCPs is less rigorous than bank stress testing, and the interconnections between CCPs (a default at one could cascade to others through shared clearing members) are poorly understood. The CPMI-IOSCO Public Quantitative Disclosures provide some transparency, but the disclosures are quarterly and backward-looking, offering little real-time insight into stress conditions. The fundamental tension is that the post-2008 regulatory framework assumed CCPs would reduce systemic risk by mutualizing and managing it — but it may have simply concentrated that risk into fewer, larger, and more interconnected nodes.",
      "dataSource": "CPMI-IOSCO Public Quantitative Disclosures for CCPs, BIS quarterly reviews, Bank of England Financial Stability Reports, Federal Reserve Financial Stability Report",
      "leadTime": "Days. CCP margin calls are issued daily (or intraday in extreme conditions). The LDI crisis went from stable to emergency intervention in approximately 5 trading days. A multi-asset scenario could develop even faster.",
      "reliability": "High for structural assessment. The procyclicality mechanism is documented and has been demonstrated in 2020 and 2022. The multi-asset simultaneous stress scenario is the key uncertainty — it has not been tested at scale.",
      "historicalPrecedents": [
        "UK LDI Pension Crisis September 2022 (LCH margin calls triggered gilt death spiral)",
        "March 2020 'Dash for Cash' (CCP margin calls contributed to Treasury market freeze)",
        "MF Global 2011 (clearing member default, client asset segregation failure)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "energy-shock",
        "yield-curve",
        "asset-bubble"
      ],
      "tags": [
        "CCP",
        "margin-calls",
        "clearinghouse",
        "procyclical",
        "LDI-parallel",
        "dash-for-cash",
        "derivatives"
      ]
    },
    {
      "id": 73,
      "sourceId": "cm-074",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "credit-debt",
      "title": "Private Credit 'Amend and Pretend'",
      "content": "The private credit market has grown from a niche corner of institutional finance into a $1.7-3.5 trillion global asset class in barely a decade, fueled by institutional investors — pension funds, endowments, sovereign wealth funds — chasing yield in a low-rate world. The core product is floating-rate loans to middle-market companies (typically $10-500M in revenue) that are too small or too leveraged for traditional bank lending or public bond markets. Private credit funds, including Business Development Companies (BDCs), direct lending funds, and PE-affiliated credit platforms, have become the primary lenders to this segment. The floating-rate structure means that as the Fed raised rates, borrowers' interest payments increased automatically — a feature that was marketed as 'rate protection' for lenders but has become a vice grip on borrowers.\n\nThe 'amend and pretend' dynamic is now widespread. When a borrower cannot cover its interest payments, the private credit lender has a choice: declare a default (which forces a loss recognition and looks bad on the fund's track record) or restructure the loan to avoid a technical default. Increasingly, lenders are choosing the latter, converting cash interest payments to PIK (Payment-In-Kind) — a mechanism where unpaid interest is simply added to the loan principal. The borrower avoids default, the lender avoids marking a loss, and both parties pretend the situation is sustainable. The fund continues to report the loan at par (or near par) in its quarterly valuations because there has been no 'credit event.' But the economic reality is that the borrower is sinking deeper into debt, the lender's actual recovery in a liquidation scenario is deteriorating, and the fund's reported returns are increasingly fictional.\n\nThe opacity of private credit is the critical amplifier. Unlike public bonds, which trade on exchanges with transparent pricing, private credit instruments are valued using internal models controlled by the fund manager — the same entity that has an economic incentive to report stable valuations (to avoid investor redemptions and maintain their fee income). There is no public secondary market to discipline these valuations. An independent analysis by the Cliffwater Direct Lending Index (CDLI) provides some aggregate transparency, but individual fund exposures are opaque. This is structurally analogous to the pre-2008 subprime CDO market, where the complexity and illiquidity of the instruments allowed originators to maintain mark-to-model valuations that bore little relationship to actual default risk.\n\nThe trigger is an economic contraction that hits middle-market revenues, converting the hidden PIK stress into outright bankruptcies that can no longer be papered over. When borrowers' revenues decline, even the PIK restructuring becomes insufficient — the company simply cannot service any form of debt. At that point, the fund must recognize losses, mark down the portfolio, and report actual (rather than fictional) returns to LPs. Pension funds and endowments that allocated to private credit as a 'stable' alternative to public bonds would discover that the stability was an accounting illusion. The capital trapped in locked-up fund structures cannot be redeemed on demand, creating the potential for a confidence crisis among institutional investors that cascades to other alternative asset classes.",
      "dataSource": "Cliffwater Direct Lending Index (CDLI), Alternative Credit Council non-accrual disclosures, Preqin private credit AUM data, SEC Division of Investment Management publications",
      "leadTime": "6-18 months. Private credit deterioration is slow-moving due to the amend-and-pretend dynamic. The transition from hidden stress to acknowledged defaults typically requires a macroeconomic recession to force the issue.",
      "reliability": "Moderate. The structural vulnerabilities are well-documented. The specific timing and magnitude of loss recognition is uncertain because the opacity that is the core problem also limits monitoring. PIK rates and non-accrual rates provide the best available signals.",
      "historicalPrecedents": [
        "2007-2008 Subprime CDO crisis (opaque instruments, mark-to-model valuations, concentrated risk)",
        "2001 Enron/WorldCom (financial engineering masking deterioration until sudden collapse)",
        "1989-1990 S&L Crisis (extend-and-pretend on commercial real estate loans)"
      ],
      "compoundsWith": [
        "credit-debt",
        "systemic-banking"
      ],
      "tags": [
        "private-credit",
        "PIK",
        "amend-and-pretend",
        "opaque-lending",
        "BDCs",
        "subprime-parallel",
        "floating-rate"
      ]
    },
    {
      "id": 74,
      "sourceId": "cm-075",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "asset-bubble",
      "title": "Passive Index Liquidity Illusion",
      "content": "Over $10 trillion is now benchmarked to passive US equity index funds and ETFs, representing one of the most significant structural transformations in financial market history. The core promise of passive investing is simple and compelling: buy the entire market at near-zero cost and earn the market return. But this promise contains a hidden assumption that breaks down in a crisis — the assumption that the index is liquid because its components are liquid. In reality, the Magnificent 7 stocks represent approximately 33% of the S&P 500 by market capitalization, but the remaining 493 stocks account for the other 67% of the weight with dramatically less daily trading volume. Passive funds must hold all 500 stocks in proportion to their index weight, creating a situation where the ETF wrapper is significantly more liquid than its underlying basket of securities.\n\nThe mechanics of ETF creation and redemption expose this fragility. In normal markets, Authorized Participants (APs) — typically large banks — arbitrage away any gap between an ETF's price and its net asset value by creating or redeeming ETF shares against baskets of the underlying stocks. When there are large outflows from SPY or VOO, APs must sell the ENTIRE basket of 500 stocks in proportion to the index weight. For Apple or Microsoft, this selling is easily absorbed by deep, liquid markets. But for the 300th or 400th stock in the S&P 500 — a mid-cap company with average daily volume of perhaps $50-100 million — a $500 million AP redemption basket could represent multiple days of normal trading volume. The result is that passive outflows create disproportionate selling pressure on the less liquid components of the index, crashing their prices and widening the discount between the ETF price and the underlying portfolio value.\n\nThe historical parallel is the 1929 investment trust collapse, which represents the last time a popular financial wrapper was more liquid than its underlying assets. Investment trusts of the 1920s pooled investor money into diversified portfolios, traded on exchanges at premiums to their net asset values, and were marketed to retail investors as a safe, easy way to participate in the stock market boom. When the market turned, investors rushed to sell their trust shares, but the trusts could not liquidate their underlying holdings fast enough to meet redemptions. The trusts' share prices collapsed to deep discounts, and the forced selling of underlying stocks amplified the broader market decline. The structural analogy to today's passive ETF ecosystem is uncomfortably precise.\n\nThe concentration risk amplifies the illusion. With 33% of the index in seven stocks, and those seven stocks trading at elevated multiples justified partly by passive inflows themselves (a reflexive dynamic where buying begets buying), any sustained risk-off event that triggers passive outflows would disproportionately impact the very stocks that have driven index performance. The spread between the S&P 500 cap-weighted index (SPY) and the equal-weighted index (RSP) serves as a real-time monitor of this concentration risk. A widening spread indicates increasing reliance on a narrowing base of mega-cap performers — the classic 'generals leading while the troops retreat' pattern that has preceded every major market top. Currently, this spread is near historically extreme levels, suggesting that the passive liquidity illusion is more dangerous than at any prior point.",
      "dataSource": "ETF net flow data (ICI, Bloomberg), SPY/RSP spread data, AP redemption basket analysis, SEC Office of Analytics publications",
      "leadTime": "Days to weeks. ETF outflow pressure builds over days as institutional investors rebalance. The August 2015 flash crash demonstrated that ETF-underlying disconnects can emerge within minutes during acute stress.",
      "reliability": "High for structural assessment. The mechanics of AP redemption and the liquidity mismatch between the ETF and its underlying basket are well-documented. The trigger (synchronized risk-off causing massive outflows) depends on a catalyst.",
      "historicalPrecedents": [
        "1929 Investment Trust collapse (liquid wrapper over illiquid assets)",
        "August 2015 flash crash (ETF prices disconnected from underlying NAVs by up to 20%)",
        "March 2020 corporate bond ETF discount to NAV (LQD, HYG traded at 5%+ discounts)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "systemic-banking"
      ],
      "tags": [
        "passive-investing",
        "ETF-liquidity",
        "concentration-risk",
        "Mag-7",
        "SPY-RSP-spread",
        "1929-parallel"
      ]
    },
    {
      "id": 75,
      "sourceId": "cm-076",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "carry-trade",
      "title": "Yen Carry Trade Unwind Phase 2",
      "content": "For decades, the yen carry trade has been one of the largest and most consequential macro trades in global finance. The mechanics are straightforward: borrow Japanese yen at near-zero interest rates, convert to dollars or other higher-yielding currencies, and invest in assets that generate returns exceeding the yen borrowing cost. The spread between Japanese rates (effectively zero for most of the past 25 years) and US rates (currently 3.50-3.75%) makes this trade enormously profitable in calm markets. The total notional size is difficult to estimate precisely because much of it is conducted through FX swaps, cross-currency basis swaps, and structured products that obscure the underlying yen funding — but credible estimates place it in the trillions of dollars. Japanese institutional investors alone (life insurers, pension funds, banks) hold approximately $4 trillion in foreign assets, much of it effectively constituting a carry trade.\n\nThe Bank of Japan's normalization of monetary policy represents a structural regime change that threatens the entire edifice. After decades of zero or negative interest rates, the BOJ has begun raising rates, and Japanese government bond yields have responded dramatically — the 40-year JGB yield has reached 4.24%, a level that would have been inconceivable just two years ago. This yield rise makes domestic Japanese assets increasingly attractive relative to foreign holdings, creating a structural incentive for Japanese capital to repatriate. Every basis point of JGB yield increase reduces the incentive to hold US Treasuries, US equities, or other foreign assets funded by yen borrowing. The repatriation of even a fraction of Japan's $4 trillion in foreign holdings would create significant selling pressure across US asset markets.\n\nThe Phase 1 unwind occurred in August 2024, when a modest BOJ rate hike triggered a violent yen appreciation and a global equity selloff — the Nikkei fell 12% in a single day, and US markets experienced their sharpest decline in months. That episode was a tremor, not the earthquake. The carry trade positions rebuilt partially as the yen weakened again, but the structural dynamics have continued to evolve in the direction of further unwinding. BOJ Governor Ueda has signaled willingness to continue normalizing, and domestic Japanese inflation — dormant for decades — is now running above 2%, giving the BOJ both justification and pressure to raise rates further.\n\nThe trigger for a Phase 2 unwind would be further BOJ rate hikes, a decline in US yields (which compresses the carry), or currency intervention by Japanese authorities to support the yen. The 1998 parallel is directly relevant: the yen carry trade unwind of October 1998 was catalyzed by Russia's debt default and the resulting flight to quality, which caused rapid yen appreciation as carry trades were unwound. The yen strengthened roughly 15% in a matter of weeks, contributing to the cascade that brought down LTCM. A similar dynamic today — yen appreciation triggering stop-losses on leveraged carry positions — would force selling of US Treasuries (reducing demand in a market already stressed by basis trade unwinding and supply concerns), US equities (particularly Mag 7 names held by Japanese institutions), and other dollar-denominated assets, creating a global deleveraging impulse originating from the world's third-largest economy.",
      "dataSource": "USD/JPY exchange rate, JGB yield curve (Ministry of Finance Japan), BOJ monetary policy statements, US Treasury International Capital (TIC) data on Japanese holdings",
      "leadTime": "Days to weeks for acute episodes. The 1998 unwind took approximately 3 weeks from catalyst to crisis. The August 2024 Phase 1 episode developed over approximately 5 trading days. Structural repatriation is a multi-year process.",
      "reliability": "High. The carry trade mechanics are well-understood. BOJ policy direction is observable. The August 2024 episode validated the transmission mechanism. The key uncertainty is timing and whether the unwind is gradual (manageable) or sudden (destabilizing).",
      "historicalPrecedents": [
        "1998 Yen carry trade unwind (catalyzed LTCM collapse)",
        "August 2024 Phase 1 unwind (Nikkei -12% in one day, global equity selloff)",
        "2007 Yen carry trade partial unwind (February-March volatility spike)"
      ],
      "compoundsWith": [
        "carry-trade",
        "yield-curve",
        "asset-bubble"
      ],
      "tags": [
        "yen-carry-trade",
        "BOJ",
        "JGB",
        "LTCM-parallel",
        "repatriation",
        "USD-JPY"
      ]
    },
    {
      "id": 76,
      "sourceId": "cm-077",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "systemic-banking",
      "title": "CRE & Regional Bank Doom Loop",
      "content": "The commercial real estate (CRE) crisis and regional bank vulnerability have become locked in a mutually reinforcing doom loop that represents one of the most dangerous slow-motion financial risks in the current environment. Regional and community banks hold approximately 70% of all commercial real estate loans in the United States — a concentration that makes the banking sector's health inextricable from CRE valuations. Office and retail properties have been permanently impaired by post-COVID behavioral shifts: office occupancy rates in major metros remain 40-50% below pre-pandemic levels, and the shift to remote and hybrid work shows no signs of reversing. Retail properties continue to face secular headwinds from e-commerce penetration. The Trepp CMBS delinquency rate has reached 7.30%, its highest level since the aftermath of the 2008 crisis, and the trajectory is worsening.\n\nThe maturity wall compounds the structural impairment. Approximately $936 billion in CRE loans are scheduled to mature in 2026, with an additional $1.26 trillion maturing in 2027. These loans were originated when interest rates were 2-4% and property valuations reflected pre-COVID usage patterns. They must now refinance into an environment where rates are 6-8%, property values have declined 25-40% for office (and more in some markets), and lenders are demanding lower loan-to-value ratios. For many properties, the math is simply impossible: a building that was worth $100 million and had a $70 million loan is now worth $60 million, meaning the loan exceeds the property's value. No rational lender will refinance this at any interest rate. The borrower's only options are to inject significant new equity (unlikely for impaired properties), negotiate a distressed sale, or default.\n\nBanks have responded to this reality with 'extend and pretend' — extending loan maturities, modifying terms, and deferring loss recognition rather than foreclosing and realizing losses that would impair their capital ratios. Regulators have tacitly allowed this approach to avoid triggering a wave of bank failures, but it creates a zombie dynamic where impaired loans remain on bank books at inflated values, consuming capital and lending capacity that could otherwise support new economic activity. The longer extend-and-pretend continues, the larger the eventual loss recognition when it finally occurs — either because a regulator forces the issue, a bank's capital position deteriorates past the point of no return, or an external event (economic recession, further rate increases) makes the fiction unsustainable.\n\nThe doom loop operates as follows: CRE defaults erode bank capital → weakened banks restrict lending → restricted lending reduces economic activity → reduced economic activity further depresses CRE values → further CRE defaults erode bank capital. This is the same dynamic that drove the 1980s-1990s Savings & Loan Crisis, which ultimately required a $160 billion taxpayer bailout (approximately $400 billion in today's dollars) and resulted in the failure of over 1,000 financial institutions. The current situation differs in one important respect: the S&L crisis was concentrated in a single property type (residential/multifamily) in a few states (Texas, California), while today's CRE crisis spans office, retail, and mixed-use properties across every major metro. The geographic and property-type diversification that would normally limit contagion does not apply when the impairment is driven by structural behavioral change (remote work, e-commerce) rather than a localized economic shock.",
      "dataSource": "Trepp CMBS Delinquency Rates, Federal Reserve H.8 Data (commercial bank balance sheets), FDIC Quarterly Banking Profile, CoStar/Green Street CRE valuation indices",
      "leadTime": "12-24 months for the doom loop to fully develop. Individual bank failures can be sudden (SVB collapsed in 48 hours), but the broader CRE-bank feedback loop operates over quarters as maturity deadlines arrive and extend-and-pretend options expire.",
      "reliability": "High. CRE maturity schedules are public. Bank CRE exposure concentrations are reported in regulatory filings. Trepp delinquency data is observable. The doom loop dynamic is historically validated by the S&L crisis. The key uncertainty is regulatory response — will regulators force loss recognition or continue to allow extend-and-pretend?",
      "historicalPrecedents": [
        "1980s-1990s S&L Crisis (CRE concentration + extend-and-pretend + eventual mass failures)",
        "2008-2010 regional bank failures (292 FDIC-insured institutions failed, many from CRE exposure)",
        "Japan 1990s 'zombie lending' (banks kept extending to insolvent borrowers, prolonging stagnation for a decade)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "energy-shock"
      ],
      "tags": [
        "CRE",
        "regional-banks",
        "doom-loop",
        "extend-and-pretend",
        "S&L-parallel",
        "office-collapse",
        "Trepp"
      ]
    },
    {
      "id": 77,
      "sourceId": "cm-078",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "currency-inflation",
      "title": "Eurodollar / Offshore Dollar Collateral Squeeze",
      "content": "The Eurodollar system — the vast network of US dollar-denominated deposits, loans, and liabilities held outside the United States — represents the hidden plumbing of global finance. With over $13 trillion in offshore dollar liabilities, the system operates largely outside the Federal Reserve's direct jurisdiction and visibility. Foreign banks, corporations, and sovereigns need dollars to service their dollar-denominated debts, fund cross-border trade, and settle international transactions. This demand for dollars is structural and persistent: roughly 60% of global central bank reserves, 50% of international trade invoicing, and 60% of international banking claims are denominated in US dollars. The system works smoothly when dollar liquidity is abundant and collateral (primarily US Treasuries) is plentiful and pristine. It breaks catastrophically when either condition fails.\n\nThe current environment is creating a collateral squeeze through multiple channels. High US interest rates make dollar funding expensive for foreign borrowers. Elevated Treasury market volatility (driven by basis trade activity, supply concerns, and geopolitical uncertainty) reduces the collateral value of Treasuries — lenders demand larger haircuts, which means each Treasury bond secures less borrowing. The US fiscal deficit is flooding the market with new Treasury supply, which paradoxically can worsen collateral quality by increasing duration risk and market volatility. And geopolitical tensions — particularly between the US and China — create the risk that dollar assets held by certain foreign entities could be frozen or sanctioned, introducing a political risk premium into what should be the world's safest collateral. The net effect is that the $13 trillion in offshore dollar liabilities is backed by a collateral base that is simultaneously becoming more expensive, more volatile, and politically less certain.\n\nThe trigger scenario is a geopolitical shock or sudden regulatory tightening that causes foreign institutions to hoard dollar collateral rather than lending it out. This hoarding instinct — rational for each individual institution — creates a collective shortage that freezes cross-border trade finance and dollar funding markets globally. The mechanism was vividly demonstrated in August 2007, when European banks that had loaded up on US subprime-linked assets suddenly lost access to dollar funding in the interbank market. The freeze spread within days from a few French money market funds (BNP Paribas) to the entire European banking system, and it was this event — not the later Lehman collapse — that marked the true beginning of the Global Financial Crisis. The Federal Reserve was forced to establish emergency swap lines with foreign central banks to provide dollar liquidity that the private market could no longer supply.\n\nThe vulnerability today is arguably greater than in 2007. Offshore dollar liabilities have grown substantially. The geopolitical landscape is more fractured, with active discussions about 'de-dollarization' and the weaponization of the dollar system through sanctions. Emerging market sovereigns and corporations have accumulated enormous dollar-denominated debt that they can only service through dollar earnings or dollar borrowing — a structural dependency that becomes a death trap when dollar funding tightens. The cross-currency basis swap spread and the FRA-OIS spread serve as the primary early-warning indicators: when these spreads widen significantly, it signals that foreign institutions are having difficulty accessing dollar funding and are willing to pay a premium for it. Both spreads have been elevated relative to historical norms, reflecting the underlying stress in the offshore dollar system.",
      "dataSource": "Cross-Currency Basis Swap spreads, FRA-OIS spread, US Treasury International Capital (TIC) data, BIS International Banking Statistics, Federal Reserve swap line usage data",
      "leadTime": "Days to weeks. The August 2007 Eurodollar freeze developed over approximately 10 days from first signs of stress to full-blown crisis. The speed depends on the nature of the triggering event — a geopolitical shock (sanctions, asset freezes) can cause immediate hoarding, while a gradual tightening of dollar supply creates a slower squeeze.",
      "reliability": "High for structural assessment. The $13T in offshore liabilities is documented by BIS. The collateral squeeze mechanics are well-understood from 2007-2008. The cross-currency basis swap and FRA-OIS spreads provide reliable real-time monitoring. The specific trigger (geopolitical shock, regulatory tightening) is unpredictable.",
      "historicalPrecedents": [
        "August 2007 European interbank market freeze (BNP Paribas, true start of GFC)",
        "March 2020 global dollar shortage (Fed expanded swap lines to 14 central banks)",
        "1997-1998 Asian Financial Crisis (dollar funding withdrawal from EM economies)",
        "2022 UK gilt crisis (offshore institutions scrambling for dollar and sterling liquidity)"
      ],
      "compoundsWith": [
        "currency-inflation",
        "systemic-banking",
        "geopolitical"
      ],
      "tags": [
        "Eurodollar",
        "offshore-dollar",
        "collateral-squeeze",
        "FRA-OIS",
        "shadow-banking",
        "EM-vulnerability",
        "2007-parallel"
      ]
    },
    {
      "id": 78,
      "sourceId": "cm-079",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "asset-bubble",
      "title": "S&P 500 - Current Reading",
      "content": "The S&P 500 stands at approximately 6,369 as of late March 2026, representing a year-to-date decline of roughly 6.8% from the January 2026 all-time high of 7,002.58. The pullback has been driven primarily by a stagflationary shock — the combination of renewed inflation fears (Core PCE re-accelerating to 3.1%) and slowing economic momentum visible in consumer confidence and China manufacturing data. A significant tech rotation has contributed, with mega-cap AI plays giving back some of their 2024-2025 gains as investors question the timeline for AI capex to translate into revenue. The Magnificent 7 concentration risk that defined the 2024 rally has become a vulnerability as sector rotation accelerates.\n\nDespite the pullback, the index remains historically elevated by virtually any valuation measure. The Shiller CAPE ratio, while compressed from its January peak, still sits well above its long-term average of approximately 17. The decline from ATH is notable but modest compared to prior bear markets — a 6.8% drawdown would barely register as a correction by historical standards. The key question is whether this represents healthy consolidation within a secular bull market or the opening phase of a larger repricing driven by the compound interaction of tariff uncertainty, sticky inflation, and AI capex disappointment. The fact that the decline has been orderly rather than panicked suggests institutional positioning remains relatively complacent, which paradoxically increases the risk of a sharper move if multiple triggers activate simultaneously.",
      "dataSource": "YCharts, financial platforms",
      "leadTime": "Real-time. Equity indices are lagging indicators of economic health but leading indicators of financial conditions and wealth effects.",
      "reliability": "High as a current reading. The index value and YTD performance are factual. Interpretation of whether current levels represent fair value, overvaluation, or a buying opportunity depends on forward earnings assumptions that are inherently uncertain.",
      "historicalPrecedents": [
        "2000 Dot-Com peak and subsequent 49% decline",
        "2007 pre-GFC peak and 57% decline",
        "2022 inflation-driven 25% correction",
        "1973-74 stagflationary bear market (-48%)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific"
      ],
      "tags": [
        "S&P-500",
        "YTD",
        "tech-rotation",
        "stagflation"
      ]
    },
    {
      "id": 79,
      "sourceId": "cm-080",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "US Unemployment Rate - Current Reading",
      "content": "The US unemployment rate stands at 4.4% as of the February 2026 BLS Employment Situation release (published March 6, 2026), representing a four-year high for the post-pandemic labor market cycle. The rate has been creeping upward from 4.3% in January and from the cycle low of 3.4% reached in early 2023. While 4.4% is still historically moderate — well below the 6.0% long-term average and far below recessionary peaks of 10%+ — the direction of travel is more concerning than the level. The labor market has entered what economists describe as a 'low hiring, low firing' equilibrium: employers are not aggressively laying off workers, but they are also not creating new positions at the pace needed to absorb population growth and labor force re-entrants.\n\nThis dynamic creates a deceptively calm surface over deteriorating fundamentals. Initial jobless claims remain historically low (210,000), which suggests the firing side of the equation is stable. But the hiring rate — measured by the JOLTS survey — has declined to levels not seen since 2019, meaning that workers who do lose jobs face longer periods of unemployment before finding new ones. The Sahm Rule, which triggers a recession signal when the 3-month moving average of unemployment rises 0.5 percentage points above its 12-month low, has been flirting with activation for several months. At 4.4% vs. a 12-month low of approximately 3.9%, the Sahm indicator is uncomfortably close to its threshold. The labor market's deterioration, while gradual, compounds with sticky inflation to create a stagflationary dynamic where the Fed faces the impossible choice of cutting rates to support employment (risking further inflation) or holding rates to fight inflation (risking accelerating job losses).",
      "dataSource": "BLS",
      "leadTime": "Monthly, with approximately 5-week lag. The unemployment rate is a lagging indicator — by the time it rises significantly, recession is often already underway or imminent.",
      "reliability": "High. BLS methodology is well-established and consistently applied. The headline rate (U-3) understates true labor market slack because it excludes discouraged workers and involuntary part-time workers (U-6 at approximately 8.0%).",
      "historicalPrecedents": [
        "2007-2009 unemployment rose from 4.4% to 10.0%",
        "2001 recession: unemployment rose from 3.9% to 6.3%",
        "1990-91 recession: 5.2% to 7.8%",
        "Sahm Rule has accurately identified every recession since 1970"
      ],
      "compoundsWith": [
        "credit-debt",
        "currency-inflation"
      ],
      "tags": [
        "unemployment",
        "BLS",
        "labor-market"
      ]
    },
    {
      "id": 80,
      "sourceId": "cm-081",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "currency-inflation",
      "title": "US CPI Year-over-Year - Current Reading",
      "content": "Headline CPI inflation stands at 2.4% year-over-year and Core CPI (excluding food and energy) at 2.5% as of the February 2026 BLS release published March 11, 2026. Both readings are essentially flat versus January, suggesting inflation is stabilizing but refusing to complete the 'last mile' descent to the Fed's 2% target. The headline figure is heavily influenced by volatile energy prices, which have been bouncing between relief (increased US production) and spikes (Middle East escalation, OPEC+ supply management). Services inflation remains the stubborn component, running well above 3% and reflecting persistent wage growth in labor-intensive sectors like healthcare, education, and housing.\n\nThe CPI readings present an optimistic surface that masks a more concerning reality visible in the Core PCE data (see cm-082). The divergence between CPI (2.4-2.5%) and Core PCE (3.1%) is unusually large and reflects methodological differences in how the two measures weight housing, healthcare, and portfolio management fees. The Fed explicitly targets PCE, not CPI, which means the more benign CPI headline is misleading as a guide to monetary policy. For crisis monitoring purposes, the CPI reading suggests inflation is normalizing — but the PCE reading contradicts this interpretation. The sticky services component in both measures reflects structural factors (aging population driving healthcare demand, housing supply constraints, tight labor market supporting wage growth) that are unlikely to resolve through monetary policy alone. This creates the conditions for a prolonged period of 'above target but not crisis-level' inflation that erodes real purchasing power and constrains the Fed's ability to cut rates in response to economic weakness.",
      "dataSource": "BLS CPI",
      "leadTime": "Monthly, with approximately 2-week lag. CPI is a coincident-to-lagging indicator of inflationary pressures.",
      "reliability": "High as a statistical measure. The BLS methodology is transparent and consistently applied. However, CPI's fixed-basket approach can overstate or understate lived inflation depending on consumption patterns. The divergence from PCE creates ambiguity about the true inflation picture.",
      "historicalPrecedents": [
        "1970s double-dip inflation (CPI appeared to normalize then re-accelerated)",
        "2021-2023 post-COVID inflation surge (CPI peaked at 9.1%)",
        "2018 'transitory' inflation scare that proved short-lived"
      ],
      "compoundsWith": [
        "currency-inflation",
        "energy-shock"
      ],
      "tags": [
        "CPI",
        "inflation",
        "services-inflation"
      ]
    },
    {
      "id": 81,
      "sourceId": "cm-082",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "currency-inflation",
      "title": "Core PCE Year-over-Year - Current Reading",
      "content": "Core PCE (Personal Consumption Expenditures excluding food and energy) stands at 3.1% year-over-year as of the latest BEA release in mid-March 2026, marking a massive and alarming divergence from the more benign CPI reading of 2.5%. This is the worst Core PCE reading in nearly two years, and the trajectory has been unmistakably upward over the past quarter — re-accelerating from approximately 2.6% in late 2025 after what appeared to be a steady disinflationary trend. The re-acceleration is being driven almost entirely by core services, particularly healthcare, housing (owners' equivalent rent), and financial services fees. The PCE measure captures these components with different weights than CPI, and those weights more accurately reflect actual consumer spending patterns, which is why the Fed explicitly targets PCE as its preferred inflation gauge.\n\nThe significance of this reading cannot be overstated for crisis monitoring. A 3.1% Core PCE effectively eliminates any possibility of Fed rate cuts in the near term, regardless of what happens in the labor market or equity markets. The Fed has been explicit that it needs to see sustained progress toward 2% before easing, and the direction of travel is now backwards. This creates a monetary policy trap: if the economy weakens (unemployment rising, consumer confidence plunging, China contracting), the Fed would normally cut rates to provide support — but with Core PCE at 3.1% and re-accelerating, cutting rates risks embedding inflation expectations above target. This is the textbook definition of stagflation risk: the economy needs easier monetary policy but inflation prohibits it. The 1970s parallel is directly relevant — the Fed's premature easing under Arthur Burns after the 1973-74 recession allowed inflation to re-accelerate, ultimately requiring the Volcker shock of 20% interest rates to break the cycle. The current Fed is acutely aware of this history, which makes them more likely to hold rates higher for longer even at the cost of economic pain.",
      "dataSource": "BEA",
      "leadTime": "Monthly, with approximately 4-week lag. PCE is released later than CPI but is considered the more comprehensive and accurate measure.",
      "reliability": "High. The BEA's methodology is rigorous and the PCE measure's chain-weighted approach adjusts for substitution effects that CPI misses. The Fed's explicit targeting of PCE makes this the single most important inflation reading for monetary policy.",
      "historicalPrecedents": [
        "1970s double-dip: Core PCE appeared to normalize then re-accelerated violently",
        "2022-2023: Core PCE peaked above 5% before beginning descent",
        "1998: Core PCE fell below 1.5%, enabling Greenspan to ease into the dot-com bubble"
      ],
      "compoundsWith": [
        "currency-inflation",
        "yield-curve"
      ],
      "tags": [
        "Core-PCE",
        "Fed-preferred",
        "services-inflation",
        "re-acceleration"
      ]
    },
    {
      "id": 82,
      "sourceId": "cm-083",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "ISM Manufacturing PMI - Current Reading",
      "content": "The ISM Manufacturing PMI registered 52.4 in February 2026 (released March 2), representing the second consecutive month of expansion above the 50 threshold. This is a notable development: prior to the January 2026 reading of 52.6, US manufacturing had been in contraction for 37 of the preceding 40 months — one of the longest manufacturing downturns in post-war history. The return to expansion reflects a combination of factors including tariff-driven reshoring activity, inventory restocking after a prolonged destocking cycle, and some boost from front-loaded orders as businesses anticipated further tariff escalation. New orders and production sub-indices both remained in expansion territory, though the employment sub-index lagged, suggesting manufacturers are meeting demand through productivity improvements rather than hiring.\n\nThe slight deceleration from January's 52.6 to February's 52.4 is not concerning on its own but warrants monitoring in context. Manufacturing expansion during a period of elevated interest rates, consumer uncertainty, and global trade disruption suggests the sector may be responding to artificial demand drivers (tariff front-running, government incentive programs) rather than organic economic strength. The prices-paid sub-index has been elevated, reflecting tariff-driven input cost increases that manufacturers are struggling to pass through to customers amid competitive pressure. If manufacturing expansion is being sustained by one-time factors rather than durable demand growth, the reversal — when it comes — could be abrupt. The contrast with China's manufacturing PMI (49.0, contraction) is also notable: US manufacturing is expanding while the world's largest manufacturing economy is contracting, a divergence that historically does not persist for long.",
      "dataSource": "ISM",
      "leadTime": "Monthly, released on first business day. PMI is a leading indicator — readings below 50 for extended periods have preceded every recession since 1948.",
      "reliability": "High. The ISM survey methodology has been consistent since 1948, providing one of the longest and most reliable economic time series available. Individual monthly readings can be noisy, but 3-month trends are highly informative.",
      "historicalPrecedents": [
        "2001: PMI fell below 50 six months before recession was officially declared",
        "2008: PMI plunged to 32.4 during GFC — lowest since 1980",
        "2019-2020: Extended sub-50 readings preceded COVID recession"
      ],
      "compoundsWith": [
        "general"
      ],
      "tags": [
        "ISM",
        "manufacturing",
        "PMI",
        "expansion"
      ]
    },
    {
      "id": 83,
      "sourceId": "cm-084",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "ISM Services PMI - Current Reading",
      "content": "The ISM Services PMI surged to 56.1 in February 2026 (released March 4), its highest level since August 2022 and a significant jump from January's 53.8. The services sector — which accounts for roughly 77% of US GDP — is showing robust expansion even as consumer confidence collapses and manufacturing barely holds above water. Business activity, new orders, and supplier deliveries all expanded at accelerating rates. The breadth of expansion was notable: 14 of 18 services industries reported growth, suggesting this is not a narrow phenomenon but a broad-based services boom.\n\nHowever, the services strength is a double-edged sword for the macro outlook and crisis monitoring specifically. The primary reason services inflation (healthcare, housing, financial services, education) remains sticky above 3% is precisely because services demand remains strong. A 56.1 services PMI is inconsistent with the Fed achieving its 2% inflation target — you cannot have booming services demand without upward wage pressure and services price increases. This creates the perverse dynamic where good economic news (strong services) is bad monetary policy news (inflation won't come down). The late-cycle divergence between strong services and tepid manufacturing is a classic pattern: in 2006-2007, services PMI remained above 55 well into the credit crisis because services activity tends to lag the deterioration in goods-producing, credit-sensitive sectors. By the time services PMI finally broke below 50 in late 2008, the recession was already a year old. The current divergence — ISM Services at 56.1 vs. Manufacturing barely above 50 — may be masking underlying fragility in the economy's most interest-rate-sensitive sectors.",
      "dataSource": "ISM",
      "leadTime": "Monthly, released on third business day. Services PMI is a coincident-to-lagging indicator — services activity typically holds up longer than manufacturing in downturns.",
      "reliability": "High. ISM Services survey covers a representative sample of the US services economy. The series is shorter than manufacturing (began 1997 vs. 1948) but has proven reliable through multiple cycles.",
      "historicalPrecedents": [
        "2006-2007: Services PMI stayed above 50 well into credit crisis",
        "2008: Services PMI collapsed to 37.8 — worst reading in series history",
        "2020: COVID plunged services to 41.8 before rapid recovery"
      ],
      "compoundsWith": [
        "currency-inflation"
      ],
      "tags": [
        "ISM-Services",
        "PMI",
        "services-boom",
        "late-cycle"
      ]
    },
    {
      "id": 84,
      "sourceId": "cm-085",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "Consumer Confidence - Current Reading",
      "content": "Consumer confidence has collapsed to levels not seen in over a decade, with the Conference Board Consumer Confidence Index plummeting to 84.5 in January 2026 — a 12-year low — before recovering marginally to 91.2 in February. The University of Michigan Consumer Sentiment Index has followed a similar trajectory, plunging an additional 6% in late March 2026 to reach levels associated with prior recessions. The destruction in consumer psychology has been driven by a confluence of shocks: tariff-driven price increase fears, renewed geopolitical volatility (Middle East escalation, US-China tensions), persistent inflation that has eroded real purchasing power for three consecutive years, and equity market weakness that has dented household wealth effects. The expectations component — which measures how consumers feel about the future — has deteriorated even more sharply than present conditions, suggesting consumers are bracing for worse to come.\n\nThe severity of the confidence collapse matters enormously for crisis monitoring because consumer spending constitutes approximately 68% of US GDP. When consumers become pessimistic, they pull back on discretionary spending, delay major purchases (homes, cars, appliances), and increase savings rates — all of which create a self-fulfilling economic slowdown. The confidence-spending transmission mechanism typically operates with a 2-4 month lag: confidence collapses first, then spending follows. The current readings are consistent with a significant spending pullback materializing by mid-2026. Critically, consumer confidence is collapsing simultaneously with rising unemployment (4.4%), sticky inflation (Core PCE 3.1%), and equity market weakness — a combination that creates compounding negative wealth and income effects. When consumers feel poorer (stocks down), face higher prices (inflation), worry about job security (rising unemployment), and expect things to get worse (plunging expectations), the resulting demand destruction can tip a slowing economy into outright recession.",
      "dataSource": "Conference Board, U of Michigan",
      "leadTime": "Monthly. Consumer confidence is a leading indicator with 2-4 month transmission to spending. Extreme readings have preceded every recession in the survey's history.",
      "reliability": "Moderate-to-high. Both surveys have long track records of identifying turning points. However, confidence can be volatile and occasionally gives false recession signals (e.g., 2011 debt ceiling crisis crashed confidence but no recession followed). The current decline's breadth across both surveys increases reliability.",
      "historicalPrecedents": [
        "2008: Conference Board fell to 25.3, its all-time low, during GFC",
        "2011: Debt ceiling crisis crashed confidence to 40s without recession",
        "1990: Confidence collapse preceded Gulf War recession",
        "2020: COVID plunged confidence but recovery was rapid due to fiscal stimulus"
      ],
      "compoundsWith": [
        "geopolitical",
        "energy-shock",
        "currency-inflation"
      ],
      "tags": [
        "consumer-confidence",
        "sentiment",
        "demand-destruction"
      ]
    },
    {
      "id": 85,
      "sourceId": "cm-086",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "Initial Jobless Claims - Current Reading",
      "content": "Initial jobless claims stand at 210,000 for the week ending March 21, 2026 (released March 26 by the Department of Labor), remaining at historically low levels that signal no acute deterioration in the firing side of the labor market equation. Continuing claims have fallen to near 2-year lows, indicating that workers who are laid off are either finding new employment relatively quickly or exiting the labor force. The four-week moving average — which smooths out weekly volatility from holidays, weather events, and seasonal adjustment quirks — has remained in the 210,000-225,000 range for several months, a remarkably stable and benign reading.\n\nHowever, the low claims data must be interpreted alongside the broader labor market context rather than in isolation. The 'low hiring, low firing' dynamic means that initial claims can remain low even as the labor market fundamentally weakens — employers are not firing existing workers, but they are not creating new positions either. This was precisely the pattern observed in 2006-2007: initial claims remained below 350,000 (then considered very low) until October 2008, well after the recession had begun in December 2007. The initial claims series is most useful as a crisis indicator when it spikes suddenly rather than drifts gradually — a spike above 300,000 sustained for multiple weeks would be a clear recession signal. The current absence of such a spike is reassuring for the near term but does not preclude a rapid deterioration if a triggering event (financial shock, geopolitical escalation, tariff retaliation) causes businesses to shift from 'pause hiring' to 'begin layoffs.' The speed of that transition can be extremely fast, as demonstrated in March 2020 when claims went from 211,000 to 6.9 million in two weeks.",
      "dataSource": "US Department of Labor",
      "leadTime": "Weekly, with 5-day lag. Initial claims are the highest-frequency labor market indicator available. Spikes provide early warning of recession with 1-3 month lead time.",
      "reliability": "High for detecting acute deterioration. Very reliable when claims spike suddenly. Less useful for detecting gradual labor market weakening (the 'low hiring, low firing' scenario). State-level data can be noisy due to processing backlogs and seasonal adjustment issues.",
      "historicalPrecedents": [
        "March 2020: Claims exploded from 211K to 6.9M in two weeks",
        "2008: Claims remained below 350K until October 2008 despite recession starting December 2007",
        "2001: Claims crossed 400K threshold in September 2001, confirming recession"
      ],
      "compoundsWith": [
        "general"
      ],
      "tags": [
        "jobless-claims",
        "labor-market",
        "stability"
      ]
    },
    {
      "id": 86,
      "sourceId": "cm-087",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "Baltic Dry Index - Current Reading",
      "content": "The Baltic Dry Index (BDI) stands at 2,031 as of March 27, 2026, reflecting a 7% decline over the past month but a 26% increase year-over-year. The index — which tracks the cost of shipping dry bulk commodities (iron ore, coal, grain, cement) on major maritime routes — serves as a real-time proxy for global trade activity and industrial demand. Unlike financial market indicators that can be distorted by speculation and leverage, the BDI reflects actual physical commerce: ships are either booked to carry cargo or they are not. The current reading of approximately 2,000 is in the normal-to-elevated range, well above the crisis lows of 290 hit in February 2016 (when global trade nearly froze) but below the speculative peaks above 10,000 seen in 2008.\n\nThe month-over-month decline warrants attention in context. Shipping costs are being supported by several structural factors: Houthi attacks in the Red Sea continue to force longer routing around the Cape of Good Hope, adding transit time and effectively reducing available shipping capacity. Energy costs remain elevated, feeding through to bunker fuel prices. And seasonal patterns in Chinese iron ore imports and grain shipments create natural fluctuations. The year-over-year strength suggests that despite headline economic uncertainty, physical trade flows remain robust — goods are still moving across oceans in volume. However, the BDI's usefulness as a crisis indicator operates with a lag: the index typically reflects orders placed weeks or months ago, meaning current readings may not yet capture a demand pullback from the consumer confidence collapse and China manufacturing contraction visible in other indicators. A sustained decline below 1,500 would signal genuine trade contraction and warrant escalation in crisis monitoring.",
      "dataSource": "Baltic Exchange",
      "leadTime": "Daily. The BDI is a leading indicator for industrial production and global trade with approximately 2-3 month lead time. Sharp declines have preceded global recessions.",
      "reliability": "Moderate-to-high. The index is based on actual shipping contracts, making it resistant to financial speculation. However, it can be distorted by supply-side factors (ship scrapping, new builds, route disruptions) that have nothing to do with demand. The Houthi disruptions currently inflate the reading relative to actual trade volume.",
      "historicalPrecedents": [
        "2008: BDI collapsed 94% from 11,793 to 663, presaging global trade freeze",
        "2016: BDI hit all-time low of 290 amid China slowdown fears",
        "2020: COVID disrupted shipping but recovery was rapid as stimulus fueled goods demand"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical"
      ],
      "tags": [
        "BDI",
        "shipping",
        "global-trade"
      ]
    },
    {
      "id": 87,
      "sourceId": "cm-088",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "geopolitical",
      "title": "China Manufacturing PMI - Current Reading",
      "content": "China's official NBS Manufacturing PMI came in at 49.0 for February 2026 (released March 4), marking the second consecutive month of contraction below the 50 threshold. The reading declined from 49.3 in January, with new orders, new export orders, and employment sub-indices all firmly in contraction territory. While some of the weakness is attributable to the Lunar New Year holiday distortion (the Spring Festival fell in late January/early February, disrupting factory schedules), the underlying trend is clearly downward. Export orders were particularly weak, reflecting the compound impact of US tariff escalation, European demand softening, and the ongoing shift of manufacturing supply chains away from China toward Vietnam, India, and Mexico.\n\nChina's manufacturing contraction has profound implications for the global crisis monitor. As the world's largest manufacturing economy and the primary driver of commodity demand, a sustained Chinese slowdown transmits through multiple channels: reduced demand for raw materials (iron ore, copper, oil) pressures commodity-exporting economies (Australia, Brazil, Middle East); weaker Chinese imports reduce export revenues for European and Asian economies; and domestic economic weakness increases the risk of competitive currency devaluation (yuan weakening) that could trigger retaliatory devaluations across Asia. The Chinese property sector crisis — with Evergrande and Country Garden in various stages of restructuring — continues to act as a structural drag on domestic demand and consumer confidence. Beijing has deployed numerous stimulus measures (rate cuts, reserve requirement reductions, property market support) but the 'pushing on a string' dynamic is evident: monetary stimulus is not translating into credit demand because households and businesses are focused on deleveraging rather than borrowing. The divergence between US manufacturing (expanding at 52.4) and Chinese manufacturing (contracting at 49.0) is unusual and historically unstable — global manufacturing tends to converge, and the resolution of this divergence will significantly impact the crisis outlook.",
      "dataSource": "China NBS",
      "leadTime": "Monthly, released on last day of month. China PMI is a leading indicator for global industrial demand and commodity prices with 1-2 month lead time.",
      "reliability": "Moderate. The official NBS PMI is widely suspected of being smoothed by Chinese statistical authorities, with the private Caixin PMI sometimes telling a different story. However, readings below 50 from the official survey are particularly informative because they represent contraction despite any upward bias in the methodology.",
      "historicalPrecedents": [
        "2015-2016: China PMI contraction triggered global growth scare and commodity crash",
        "2008: China PMI fell to 38.8 during GFC before massive stimulus response",
        "2020: COVID plunged China PMI to 35.7 before rapid V-shaped recovery"
      ],
      "compoundsWith": [
        "geopolitical",
        "general"
      ],
      "tags": [
        "China-PMI",
        "contraction",
        "global-demand",
        "export-weakness"
      ]
    },
    {
      "id": 88,
      "sourceId": "cm-089",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "Gold Price - Current Reading",
      "content": "Gold is trading in the $4,433-$4,490/oz range as of late March 2026, having corrected approximately 15-20% from a staggering all-time high of roughly $5,600/oz reached in late January 2026. Even after this significant correction, the current price represents a massive super-cycle move that has no precedent in the modern era of gold trading. Gold has roughly doubled from its $2,000/oz levels of early 2024, driven by a convergence of structural forces: central bank de-dollarization (China, Russia, India, and Middle Eastern sovereigns have been aggressive buyers), safe-haven demand amid escalating geopolitical tensions (Ukraine, Middle East, US-China), inflation hedging as Core PCE re-accelerates to 3.1%, and retail/ETF inflows as gold's momentum attracts trend-following capital.\n\nThe gold price at $4,400+ is a crisis indicator in itself — it reflects the collective judgment of central banks, institutional investors, and sovereign wealth funds that the current geopolitical and monetary environment carries exceptional risk. Central bank gold purchases have been running at record levels for three consecutive years, with the People's Bank of China alone adding hundreds of tonnes to its reserves. This is not speculative froth but strategic reserve diversification by entities with the deepest information advantages about geopolitical risk. The 15-20% correction from the $5,600 peak is notable: it may represent profit-taking after an unsustainable parabolic move, or it may represent a healthy consolidation before the next leg higher if any of the structural drivers (de-dollarization, inflation, geopolitical escalation) intensify further. For crisis monitoring, gold above $4,000 should be treated as a persistent amber warning signal — the market is pricing in an elevated probability of monetary system stress, geopolitical escalation, or inflation regime change that would justify continued flight from fiat currencies into hard assets.",
      "dataSource": "CBS News, FXLeaders",
      "leadTime": "Real-time. Gold typically leads other safe-haven moves by days to weeks. Sharp gold rallies have preceded or coincided with every major financial crisis since 1971.",
      "reliability": "High as a sentiment and positioning indicator. Gold's price reflects the aggregate risk assessment of the most sophisticated global actors (central banks, sovereign wealth funds). The signal is less reliable for timing specific events — gold can remain elevated for extended periods without a crisis materializing.",
      "historicalPrecedents": [
        "1979-1980: Gold surged to $850 amid inflation, Iran hostage crisis, Soviet invasion of Afghanistan",
        "2008-2011: Gold rose from $700 to $1,920 during and after GFC",
        "2020: Gold hit $2,075 on COVID monetary stimulus fears",
        "2024-2026: Current super-cycle driven by de-dollarization and geopolitical risk"
      ],
      "compoundsWith": [
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "gold",
        "safe-haven",
        "de-dollarization",
        "super-cycle"
      ]
    },
    {
      "id": 89,
      "sourceId": "cm-090",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "Bitcoin Price - Current Reading",
      "content": "Bitcoin is trading at approximately $66,774 as of late March 2026, having declined sharply from its all-time high of $126,080 set earlier in 2026 as the broader risk-off rotation accelerates. The decline through March reflects a classic pattern in Bitcoin's evolving relationship with traditional risk assets: despite the narrative of Bitcoin as 'digital gold' and an inflation hedge, its actual correlation with risk assets (particularly Nasdaq and high-growth equities) has remained stubbornly positive during periods of genuine market stress. When investors liquidate speculative positions to raise cash or reduce portfolio risk, Bitcoin tends to sell off alongside other risk assets rather than providing the safe-haven diversification that gold offers. The current selloff is consistent with this pattern — as equity markets decline, inflation fears rise, and the probability of Fed rate cuts evaporates, speculative capital is rotating out of crypto and into cash, short-duration Treasuries, and gold.\n\nFor crisis monitoring, Bitcoin's behavior is informative primarily as a gauge of speculative appetite and risk tolerance rather than as a fundamental economic indicator. A declining Bitcoin amid rising gold is a classic risk-off signal: capital is flowing from speculative assets toward traditional safe havens. The $66,000 level remains historically elevated — Bitcoin was below $30,000 as recently as mid-2023 — which means that despite the recent decline, significant speculative capital remains deployed in crypto that could face liquidation pressure in a genuine crisis scenario. Leveraged Bitcoin positions (perpetual futures, DeFi lending protocols, margin trading) create the potential for cascade liquidations if key support levels break, with $60,000 representing a psychologically and technically important threshold. A break below $60,000 could trigger a deleveraging event that compounds broader market stress by forcing crypto-exposed funds to sell other assets to meet margin requirements.",
      "dataSource": "Major crypto exchanges",
      "leadTime": "Real-time. Bitcoin is a coincident indicator of risk appetite. Its 24/7 trading makes it the first asset to react to weekend geopolitical events, sometimes providing an early signal for Monday equity market opens.",
      "reliability": "Moderate. Bitcoin's price is influenced by crypto-specific factors (regulatory news, exchange collapses, halving cycles) that can overwhelm macro signals. Most useful as a risk-appetite gauge when interpreted alongside gold, VIX, and credit spreads rather than in isolation.",
      "historicalPrecedents": [
        "2022: Bitcoin fell 77% from $69K to $15.5K alongside equity bear market and crypto contagion (Luna, FTX)",
        "March 2020: Bitcoin crashed 50% in two days alongside COVID equity selloff",
        "2018: Bitcoin fell 84% during crypto winter despite no broader macro crisis"
      ],
      "compoundsWith": [
        "asset-bubble"
      ],
      "tags": [
        "Bitcoin",
        "crypto",
        "risk-off",
        "speculative-assets"
      ]
    },
    {
      "id": 90,
      "sourceId": "cm-091",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "currency-inflation",
      "title": "US Dollar Index (DXY) - Current Reading",
      "content": "The US Dollar Index (DXY) has surged back above the psychologically significant 100 level, trading at 100.15 as of late March 2026 — its first sustained break above 100 in 2026. The rally has been driven primarily by safe-haven flows as investors seek dollar-denominated assets amid escalating Middle East tensions, renewed oil price volatility, and the re-acceleration of US inflation (Core PCE at 3.1%) which has pushed rate cut expectations off the table. The dollar's strength is being amplified by relative weakness in other major currencies: the euro faces headwinds from European economic stagnation and political uncertainty, the yen remains under pressure from the still-vast interest rate differential with the US (despite BOJ tightening), and the British pound is weighed down by sluggish UK growth.\n\nA strong dollar creates a complex web of cross-border stresses that compounds other crisis indicators. For emerging market economies, dollar strength is particularly punishing: countries with dollar-denominated debt (estimated at $4+ trillion for EM sovereigns and corporates) face higher debt servicing costs in local currency terms, effectively tightening financial conditions without any change in domestic monetary policy. For EM central banks, defending their currencies against dollar strength requires either spending foreign exchange reserves (depleting their crisis buffer) or raising domestic interest rates (slowing their own economies). US exporters face competitive headwinds as their products become more expensive in foreign currency terms. And for the global Eurodollar system (see cm-078), dollar strength increases the cost of offshore dollar funding, widening cross-currency basis swap spreads and tightening the financial plumbing that underpins international trade and capital flows. The DXY above 100 is not a crisis level — the index reached 114 in September 2022 — but the directional move higher, combined with the structural forces driving it (safe-haven demand, rate differentials, geopolitical risk), suggests the dollar wrecking ball dynamic that caused significant EM stress in 2022 may be reactivating.",
      "dataSource": "ICE Data",
      "leadTime": "Real-time. Dollar strength transmits to EM economies with a 1-3 month lag as hedging positions expire and debt servicing dates arrive. The impact on US corporate earnings shows up with a 1-2 quarter lag.",
      "reliability": "High as a current reading. The DXY is a well-defined, transparent index weighted against six major currencies (EUR 57.6%, JPY 13.6%, GBP 11.9%, CAD 9.1%, SEK 4.2%, CHF 3.6%). Its limitation is that it does not capture the dollar's relationship with emerging market currencies, which is where the most acute stress typically occurs.",
      "historicalPrecedents": [
        "2022: DXY surged to 114, creating EM debt crises and forcing BOJ intervention",
        "2014-2015: Strong dollar contributed to commodity collapse and EM stress",
        "1997-1998: Dollar strength was a key factor in the Asian Financial Crisis",
        "1985: Plaza Accord was required to reverse destabilizing dollar strength"
      ],
      "compoundsWith": [
        "currency-inflation",
        "geopolitical",
        "carry-trade"
      ],
      "tags": [
        "DXY",
        "dollar-strength",
        "safe-haven",
        "EM-pressure"
      ]
    },
    {
      "id": 91,
      "sourceId": "cm-092",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "credit-debt",
      "title": "30-Year Fixed Mortgage Rate - Current Reading",
      "content": "The 30-year fixed mortgage rate has jumped to 6.38% as of Freddie Mac's Primary Mortgage Market Survey released March 26, 2026 — the highest reading in over six months and a sharp 16 basis point increase in a single week. The spike reflects the transmission of renewed inflation fears into the long end of the Treasury yield curve: mortgage rates are primarily determined by the 10-year Treasury yield plus a spread for credit and prepayment risk, and the 10-year yield has surged on the back of Core PCE re-accelerating to 3.1% and the resulting evaporation of rate cut expectations. The mortgage spread itself has also been elevated relative to historical norms, reflecting ongoing uncertainty in the agency MBS market and the Federal Reserve's continued balance sheet runoff (quantitative tightening) which removes a major buyer of mortgage-backed securities.\n\nThe housing market implications are severe and compound with broader economic weakness. At 6.38%, a 30-year mortgage on a median-priced US home (approximately $410,000) requires a monthly payment of roughly $2,560 — nearly double what the same home would have cost at 2021's 2.65% rate ($1,650). This affordability crisis has frozen the housing market in a classic 'lock-in' effect: existing homeowners with sub-4% mortgages have no incentive to sell and take on a 6%+ rate, which constrains supply and prevents price correction. New buyers face qualification barriers as debt-to-income ratios are stretched. Housing starts and existing home sales have been running at multi-decade lows for this stage of the economic cycle. The 16bps single-week spike is particularly concerning because it signals that the market is repricing for a higher-for-longer rate regime rather than the gradual normalization previously expected. Each 100bps increase in mortgage rates reduces purchasing power by approximately 10%, further eroding affordability and consumer confidence (already at 12-year lows) while putting upward pressure on the owners' equivalent rent component of inflation — creating a vicious feedback loop between rates, housing costs, and the inflation measures that determine rates.",
      "dataSource": "Freddie Mac PMMS",
      "leadTime": "Weekly. Mortgage rates are a coincident indicator of financial conditions and a leading indicator of housing activity with 1-3 month transmission. Housing activity is a leading indicator of broader economic activity with 2-4 quarter lead time.",
      "reliability": "High. Freddie Mac's survey has been the benchmark mortgage rate series since 1971. The rate directly determines housing affordability and purchasing behavior. The relationship between mortgage rates, housing activity, and economic growth is one of the most robust and well-documented in economics.",
      "historicalPrecedents": [
        "1981: Mortgage rates reached 18.63%, devastating housing and contributing to deep recession",
        "2006-2007: Rates above 6% contributed to housing bubble burst",
        "2022-2023: Rates surged from 3% to 7.79%, freezing housing market",
        "2019: Rate decline from 5% to 3.5% fueled housing recovery"
      ],
      "compoundsWith": [
        "credit-debt",
        "currency-inflation",
        "yield-curve"
      ],
      "tags": [
        "mortgage-rates",
        "housing",
        "affordability",
        "Freddie-Mac"
      ]
    },
    {
      "id": 92,
      "sourceId": "cm-093",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "general",
      "title": "Copper Price (Dr. Copper) - Current Reading",
      "content": "Copper — often called 'Dr. Copper' for its supposed PhD in economics due to its historical correlation with global economic health — is trading at approximately $5.47/lb ($12,075/metric tonne) as of late March 2026. The price has corrected from a stunning all-time high of $6.58/lb reached in January 2026 but remains dramatically elevated by historical standards, significantly higher year-over-year. The copper market is experiencing a fundamental structural transformation that is breaking the traditional correlation between copper prices and economic cyclicality. Historically, copper declined during economic slowdowns because industrial demand weakened. Today, copper demand is being driven by structural forces that operate independently of — and often contrary to — the business cycle: AI data center construction (each hyperscale data center requires 30,000-60,000 tonnes of copper for electrical systems, cooling infrastructure, and networking), renewable energy installation (wind turbines, solar panels, and grid infrastructure are 4-12x more copper-intensive than fossil fuel equivalents), electric vehicle manufacturing (EVs use 3-4x more copper than ICE vehicles), and grid modernization to support electrification.\n\nThe structural supply-demand imbalance is the defining feature of the current copper market and explains why prices remain elevated despite economic headwinds. Mine supply growth has been constrained by years of underinvestment during the 2015-2020 commodity bear market, declining ore grades at existing mines, lengthy permitting timelines for new projects (10-15 years from discovery to production), and increasing ESG and community opposition to new mining. The International Copper Study Group and major producers estimate a cumulative supply deficit of 5-10 million tonnes by 2030 — a shortfall that cannot be closed by recycling or substitution alone. For crisis monitoring, copper's behavior diverges from traditional interpretation: elevated copper prices during an economic slowdown signal structural demand strength (AI buildout, energy transition) rather than economic health. The fact that copper has only corrected 17% from its ATH despite China manufacturing contraction, consumer confidence collapse, and equity market weakness is remarkable — it suggests that structural demand from AI and energy transition is absorbing supply that would otherwise depress prices in a typical slowdown. If copper breaks below $4.50/lb, it would signal that cyclical weakness is overwhelming structural demand, which would be a significant negative signal for the AI capex narrative and global industrial activity.",
      "dataSource": "COMEX, LME via Trading Economics",
      "leadTime": "Real-time. Copper has historically led economic turning points by 1-3 months. However, the structural demand shift from AI and energy transition may reduce its reliability as a cyclical indicator going forward.",
      "reliability": "Moderate — currently transitioning. Copper's traditional PhD-in-economics reliability is being complicated by structural demand forces that decouple price from cyclical economic activity. The indicator remains useful but requires contextualization: elevated prices may reflect structural deficit rather than economic strength.",
      "historicalPrecedents": [
        "2008: Copper crashed 66% from $4.08 to $1.38 during GFC, accurately signaling global recession",
        "2011: Copper peaked at $4.62 and declined for 5 years as China investment slowed",
        "2020: COVID crashed copper to $2.12 before stimulus-driven recovery to new highs",
        "2024-2026: AI/energy transition structural demand has created new price regime"
      ],
      "compoundsWith": [
        "ai-specific",
        "energy-shock"
      ],
      "tags": [
        "copper",
        "Dr-Copper",
        "structural-deficit",
        "AI-demand",
        "energy-transition"
      ]
    },
    {
      "id": 93,
      "sourceId": "cm-094",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "energy-shock",
      "title": "1973 Oil Crisis — Trigger Sequence Analysis",
      "content": "The 1973 oil crisis is the canonical example of a geopolitical supply shock detonating an already-fragile monetary and inflationary regime. The pre-conditions were established years before the embargo itself. In August 1971, Nixon closed the gold window, ending the Bretton Woods system of fixed exchange rates and removing the anchor that had constrained US monetary expansion. Federal Reserve Chairman Arthur Burns, under intense political pressure from the White House ahead of the 1972 election, pursued aggressively expansionary monetary policy — M1 money supply was growing at over 8% year-over-year through 1972-73. This monetary expansion was already generating significant inflation before any oil shock: CPI was running at approximately 6% by mid-1973, driven by food prices, commodity speculation, and the devaluation of the dollar. Global manufacturing capacity was operating at or near maximum utilization rates across the industrialized world, leaving no slack to absorb further cost pressures. The economy was, in Minsky's framework, already in a speculative-to-Ponzi financing regime — leveraged, overheated, and dependent on continued expansion to service accumulated debts.\n\nThe trigger sequence unfolded with a precision that reveals how financial markets often lead geopolitical catalysts. In January 1973, the Dow Jones Industrial Average peaked at approximately 1,050 and began a secular decline that would ultimately take it to 577 by December 1974 — a 45% drawdown. The equity market was responding to tightening monetary conditions and rising inflation expectations months before any oil embargo. By June 1973, the yield curve (10-year Treasury minus 3-month T-bill) inverted — the bond market's most reliable recession signal. This inversion preceded the geopolitical trigger by four full months. On October 6, 1973, Egypt and Syria launched a coordinated surprise attack on Israel, beginning the Yom Kippur War. On October 17, the Organization of Arab Petroleum Exporting Countries (OAPEC) announced an oil embargo against nations supporting Israel, including the United States. Crude oil prices surged from approximately $3 per barrel to $11.65 per barrel by January 1974 — a nearly 300% increase in three months.\n\nThe compound interaction between fiat currency devaluation and a geopolitical supply shock produced the phenomenon that economists had believed was theoretically impossible: stagflation — simultaneous high inflation and economic contraction. The oil price surge crushed corporate profit margins across virtually every sector of the economy. Transportation, manufacturing, agriculture, chemicals, and utilities all faced input cost spikes they could not immediately pass through. The result was layoffs and production cuts (the 'stag' component). But the inflationary dynamics were self-reinforcing through a mechanism that has no modern parallel: unionized cost-of-living adjustment (COLA) clauses. Approximately 60% of the US workforce in 1973 was covered by contracts with automatic wage escalators tied to CPI. As oil pushed CPI higher, wages automatically ratcheted up, which raised corporate costs further, which raised prices further — the textbook wage-price spiral. This COLA-driven feedback loop paralyzed the Federal Reserve: cutting rates to fight recession would accelerate inflation, but raising rates to fight inflation would deepen the recession. Burns chose the worst of both worlds, vacillating between tight and loose policy, which prolonged the stagflation for years.\n\nThe market indicators told the story with brutal clarity. The yield curve, which had been normal through 1972, inverted in June 1973 and deepened through early 1974 as short-term rates spiked on inflation expectations while long-term rates reflected recession probability. The S&P 500's 30-day realized volatility surged from approximately 12% in 1972 to over 30% by mid-1974. Baa corporate bond spreads widened from approximately 1.5% to over 2.5%, reflecting credit stress as margins collapsed. The S&P 500 took 93 months — nearly eight years — to recover its nominal peak, finally surpassing it in July 1980. In inflation-adjusted terms, the recovery took even longer: real returns did not surpass the 1973 peak until 1993, a full 20-year round trip. The earliest actionable warning was the yield curve inversion in June 1973, which gave four months of lead time before the geopolitical trigger that most analysts retrospectively blame for the crisis. The inversion was the market's verdict that the monetary regime was already unsustainable — the oil embargo merely determined the specific form the crisis would take, not whether a crisis would occur.",
      "dataSource": "Federal Reserve Economic Data (FRED), Bureau of Labor Statistics CPI archives, OPEC historical pricing, NBER business cycle dating, academic literature (Blinder, Hamilton, Barsky & Kilian)",
      "leadTime": "Yield curve inversion gave 4 months lead before geopolitical trigger. Equity market peaked 9 months before the embargo. CPI acceleration gave 12+ months warning that the monetary regime was unsustainable.",
      "reliability": "Very high — one of the most thoroughly studied crisis episodes in economic history. The trigger sequence (monetary excess → inflation → geopolitical shock → stagflation) is well-documented and the indicator readings are precisely known.",
      "historicalPrecedents": [
        "1979 Second Oil Crisis (similar dynamics, tighter starting conditions)",
        "2022 Russia-Ukraine energy shock (European gas dependence parallel)",
        "1971 Nixon Shock (precipitating monetary event)"
      ],
      "compoundsWith": [
        "energy-shock",
        "currency-inflation",
        "yield-curve",
        "geopolitical"
      ],
      "tags": [
        "1973",
        "oil-crisis",
        "stagflation",
        "Yom-Kippur",
        "wage-price-spiral",
        "Bretton-Woods",
        "OAPEC",
        "COLA",
        "Arthur-Burns",
        "yield-curve-inversion"
      ]
    },
    {
      "id": 94,
      "sourceId": "cm-095",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "carry-trade",
      "title": "1997 Asian Financial Crisis — Trigger Sequence Analysis",
      "content": "The 1997 Asian Financial Crisis is the defining case study of how currency peg regimes, current account imbalances, and short-term foreign-currency borrowing can interact to produce a self-reinforcing doom loop that transforms localized financial stress into a regional and ultimately global crisis. The pre-conditions were structural and had been building for nearly a decade. Thailand, Indonesia, South Korea, Malaysia, and the Philippines had all pursued export-led growth strategies financed by massive inflows of foreign capital — much of it short-term, dollar-denominated debt. Thailand's current account deficit exceeded 8% of GDP by 1996, an extreme level that academic research (Milesi-Ferretti and Razin) had identified as the danger threshold beyond which sudden stops become likely. Critically, these countries maintained fixed or semi-fixed exchange rate pegs to the US dollar, which gave foreign lenders the illusion of zero currency risk and domestic borrowers no incentive to hedge their dollar-denominated obligations. The peg created a moral hazard trap: it encouraged precisely the behavior (unhedged foreign borrowing) that would prove catastrophic when the peg broke.\n\nThe funding structure was extraordinarily fragile. Thai finance companies and banks were borrowing short-term in US dollars (3-6 month maturities) through the Bangkok International Banking Facility (BIBF) and relending long-term in Thai baht into illiquid real estate developments and speculative projects with 5-10 year time horizons. This maturity mismatch — the essence of banking fragility — was compounded by a currency mismatch: liabilities in dollars, assets in baht. As long as the peg held and dollar funding remained available, the spread was enormously profitable. The moment either condition failed, the entire structure would collapse. The trigger sequence began not with a dramatic event but with a gradual deterioration. In mid-1996, Thai export growth slowed sharply due to a global semiconductor downturn and competition from a weakening Japanese yen (which made Japanese exports cheaper relative to Thai exports). This export weakness widened the current account deficit further and raised questions about the sustainability of the peg.\n\nThe crisis accelerated through a series of increasingly severe shocks. In February-March 1997, Finance One — Thailand's largest finance company — collapsed under the weight of bad real estate loans, signaling that the property bubble had burst. By May 1997, international hedge funds, led by speculative attacks from macro traders who recognized the fundamental unsustainability of the peg, were aggressively shorting the baht in the forward market. The Bank of Thailand defended the peg by selling foreign exchange reserves and entering into forward contracts — a strategy that appeared to work temporarily but was actually burning through reserves at an unsustainable rate. By late June, Thailand's usable reserves were nearly exhausted (committed through forward obligations), and on July 2, 1997, the government was forced to abandon the peg. The baht dropped 20% instantly on the announcement and continued falling, eventually losing approximately 50% of its value against the dollar.\n\nWhat made the Asian crisis uniquely devastating was the 'twin crisis' dynamic — the simultaneous occurrence of a currency crisis and a banking crisis, each reinforcing the other in a doom loop. When the baht devalued 50%, every Thai company and bank with unhedged dollar-denominated debt saw that debt double in local currency terms overnight. Companies that were profitable and solvent on July 1 were technically bankrupt on July 3 — not because their operations had changed but because their debt burden had doubled in baht terms. This instant insolvency triggered bank losses (as borrowers defaulted), which eroded bank capital, which caused banks to call in loans and restrict credit, which deepened the recession, which further impaired asset values, which caused more defaults. The doom loop was compounded by capital flight: as foreign investors saw the devaluation and banking collapse, they withdrew capital as fast as possible, putting further downward pressure on the currency, which further increased the dollar debt burden. Thai overnight interbank rates spiked to over 20% in May 1997 as the Bank of Thailand desperately tried to make shorting expensive. The US VIX remained remarkably muted at approximately 20 through the summer of 1997 — the contagion had not yet reached US markets — before spiking to 38 in October when the crisis spread to South Korea and threatened the global financial system. EMBI (Emerging Market Bond Index) spreads exploded from approximately 400 basis points to over 1,000 basis points. The Thai SET index and Korean KOSPI took 6-10 years to recover to their pre-crisis levels in US dollar terms. The earliest actionable warning was the current account deficit exceeding 8% of GDP combined with a surging US dollar, which gave 12-18 months of lead time before the acute crisis.",
      "dataSource": "IMF post-crisis assessments, BIS working papers on twin crises, Bank of Thailand data, Radelet & Sachs analysis, Corsetti-Pesenti-Roubini academic literature, EMBI historical spreads",
      "leadTime": "Current account deficit exceeding 8% + USD appreciation gave 12-18 months lead. Finance One collapse (Feb 1997) gave 4 months warning before the peg broke. The doom loop from peg break to systemic crisis took approximately 3 months.",
      "reliability": "Very high. The twin crisis mechanism (currency devaluation → debt burden increase → banking collapse → capital flight → further devaluation) is thoroughly documented and has repeated in multiple emerging market episodes. The key lesson is that currency pegs create hidden fragility by encouraging unhedged borrowing.",
      "historicalPrecedents": [
        "1994 Mexican Tequila Crisis (similar peg collapse + capital flight)",
        "1998 Russian debt default (contagion from Asian crisis)",
        "2001 Argentine crisis (currency board collapse)",
        "2018 Turkish lira crisis (unhedged corporate dollar debt)"
      ],
      "compoundsWith": [
        "carry-trade",
        "currency-inflation",
        "systemic-banking",
        "credit-debt"
      ],
      "tags": [
        "1997",
        "Asian-crisis",
        "Thailand",
        "twin-crisis",
        "currency-peg",
        "doom-loop",
        "EMBI",
        "carry-trade",
        "BIBF",
        "Finance-One",
        "current-account"
      ]
    },
    {
      "id": 95,
      "sourceId": "cm-096",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "asset-bubble",
      "title": "2000 Dot-Com Crash — Trigger Sequence Analysis",
      "content": "The dot-com crash is the most granular and instructive trigger sequence analysis available for the current AI investment cycle, because the specific mechanism — an equity funding squeeze triggered by liquidity withdrawal — has a precise modern parallel in AI capex funded by hyperscaler balance sheets. The pre-conditions were extreme by any historical measure. The Shiller CAPE ratio reached 44.19 in December 1999, the highest level in the 130-year history of the dataset — surpassing even the 1929 peak of approximately 33. The NASDAQ Composite's aggregate price-to-earnings ratio was effectively infinite because the index's aggregate earnings were negative: the combined losses of unprofitable dot-com companies exceeded the combined profits of the profitable ones. Retail speculation had reached mania levels, with day-trading accounts proliferating and initial public offerings routinely doubling or tripling on their first day of trading regardless of business fundamentals. The corporate model for the majority of internet companies was explicitly premised on continuous equity issuance: burn cash to grow, issue more stock to fund the burn, repeat. This model depended entirely on a rising stock market — the moment share prices stopped rising, the funding mechanism would seize.\n\nThe trigger sequence reveals how multiple seemingly unrelated events can converge to pop a bubble. In June 1999, the Federal Reserve began raising interest rates, tightening financial conditions into what was already an overheated speculative environment. The rate hikes did not immediately affect equities — the NASDAQ continued rising through January 2000, seemingly impervious to tighter monetary policy. The hidden catalyst was Y2K. In the second half of 1999, the Federal Reserve had injected substantial emergency liquidity into the banking system through expanded repo operations, ensuring that the Y2K date transition would not cause a financial system failure. This additional liquidity — which was never about the stock market — provided an inadvertent tailwind for risk assets. When January 2000 passed without Y2K disruptions, the Fed began withdrawing the emergency repo liquidity. This withdrawal, combined with the rate hikes already in progress, created a double tightening impulse that the market absorbed with a delay. In February 2000, the yield curve (10-year minus 2-year Treasury) inverted — the bond market's canonical recession warning. NASDAQ peaked on March 10, 2000, at 5,048.62.\n\nThe crash was catalyzed by a specific piece of journalism that crystallized what sophisticated investors already knew. On March 20, 2000, Barron's published its landmark 'Burning Up' cover story, which systematically analyzed the cash burn rates and remaining cash positions of 51 prominent internet companies. The article's conclusion was devastating: at their current burn rates, the majority of these companies would exhaust their cash within 12 months. This was not a speculative opinion — it was basic arithmetic applied to public financial statements. The article did not cause the crash, but it provided the focal point that coordinated the shift from 'we all know this is overvalued but the trend is our friend' to 'the music has stopped.' The NASDAQ fell 9% in the week following publication and never recovered.\n\nThe compound interaction that made the dot-com crash so severe was the Equity Funding Squeeze. Federal Reserve rate hikes plus Y2K liquidity withdrawal compressed valuations across the board. But dot-com business models were uniquely vulnerable because they relied on issuing equity to fund operating losses — a mechanism that required continuously rising share prices. When share prices fell, companies could no longer issue stock at acceptable dilution levels, which meant they could no longer fund operations, which meant they would run out of cash, which further depressed share prices, creating a doom loop. Critically, the contagion spread beyond unprofitable dot-coms: when the internet companies failed, they cancelled orders for telecommunications and networking equipment from profitable companies like Cisco, Nortel, and Lucent. These companies had built massive manufacturing capacity and inventory for dot-com orders that evaporated overnight, resulting in enormous writedowns and layoffs in what had been the 'real economy' backbone of the tech sector. The indicator readings were stark: the 10Y-2Y inverted in February 2000 and then un-inverted in late 2000 as the Fed panicked and began cutting rates. The VIX hovered at 20-25 during the melt-up phase (a remarkably low level of fear during a mania), then surged above 30 in April 2000 and remained elevated for years. High-yield credit spreads widened from approximately 400 basis points to over 500 basis points. The NASDAQ took 180 months — 15 full years — to reclaim the 5,000 level, finally reaching it in April 2015. The earliest actionable warning was the divergence between high-yield credit spreads and equity prices in mid-1999: credit markets were pricing in increasing risk while the NASDAQ went parabolic, a divergence that gave approximately 9 months of lead time before the peak.",
      "dataSource": "NASDAQ historical data, Shiller CAPE dataset, Federal Reserve repo operations data, Barron's archives, SEC filing analysis (cash burn rates), NBER dating, academic literature (Ofek & Richardson, Brunnermeier)",
      "leadTime": "HY credit spread divergence gave 9 months lead (mid-1999). Yield curve inversion gave 1 month lead (Feb 2000). Barron's 'Burning Up' (Mar 20, 2000) was effectively a coincident indicator at the top.",
      "reliability": "Very high — exhaustively documented by both academic research and contemporaneous journalism. The trigger sequence (rate hikes → liquidity withdrawal → valuation compression → equity funding squeeze → doom loop) is precisely traceable. The 15-year recovery timeline is a crucial data point for calibrating expectations about the current AI cycle.",
      "historicalPrecedents": [
        "1929 stock market crash (similar retail mania, margin-driven leverage)",
        "1720 South Sea Bubble (subscription model parallels equity issuance dependency)",
        "1990s Japan Nikkei collapse (asset bubble from excess liquidity)"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "yield-curve",
        "credit-debt"
      ],
      "tags": [
        "dot-com",
        "NASDAQ",
        "CAPE-44",
        "equity-funding-squeeze",
        "Y2K-liquidity",
        "Barrons",
        "15-year-recovery",
        "Burning-Up",
        "rate-hike-trigger",
        "concentration-risk"
      ]
    },
    {
      "id": 96,
      "sourceId": "cm-097",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "systemic-banking",
      "title": "2008 Global Financial Crisis — Trigger Sequence Analysis",
      "content": "The 2008 Global Financial Crisis is the most systemically complex crisis in modern history, and a detailed trigger sequence analysis reveals how a localized housing problem in a few US states became a global financial meltdown through a specific chain of collateral contagion that has direct parallels to current market risks. The pre-conditions were an interlocking set of structural vulnerabilities, none of which was individually catastrophic. NINJA mortgages (No Income, No Job, No Assets) represented the extreme end of underwriting degradation, but the real systemic risk lay not in the mortgages themselves but in the derivative structures built on top of them. Mortgage-backed securities (MBS), collateralized debt obligations (CDOs), and synthetic CDOs created a multiplicative chain where a single pool of subprime mortgages could generate notional exposure many times its face value. Credit rating agencies — Moody's, S&P, Fitch — assigned AAA ratings to senior tranches of these structures using Gaussian copula models that assumed housing price declines could not be correlated across geographies, a catastrophically wrong assumption. The shadow banking system — SIVs, conduits, money market funds, broker-dealers — funded itself through short-term repurchase agreements (repos) using these AAA-rated MBS and CDOs as collateral. This created a maturity mismatch of enormous scale: long-term, illiquid mortgage risk funded by overnight borrowing that could be pulled at any moment.\n\nThe trigger sequence stretched over approximately 30 months from first warning to systemic collapse, making it one of the longest and most clearly signaled crises in financial history. In February 2006, the yield curve (10-year minus 2-year Treasury) inverted — an event that had preceded every US recession since 1955 with only one false signal (1966). In June 2006, US housing prices peaked as measured by the S&P/Case-Shiller National Home Price Index, beginning a decline that would ultimately reach 27% nationally (and 50%+ in markets like Las Vegas, Phoenix, and Miami). In February 2007, the ABX-HE index — which tracked the price of credit default swaps on subprime mortgage-backed securities — crashed, providing the first precise market-based signal that the subprime mortgage pool was deteriorating rapidly. In June 2007, Bear Stearns was forced to liquidate two hedge funds that had been heavily invested in subprime CDOs, revealing that the losses were large enough to destroy dedicated investment vehicles. Then came the event that, in retrospect, marked the true beginning of the crisis: on August 9, 2007, BNP Paribas announced it was freezing three investment funds, citing 'the complete evaporation of liquidity in certain market segments of the US securitization market.' This announcement was the silent start of the interbank crunch — European banks that had loaded up on US subprime exposure suddenly could not value their assets, which meant they could not determine their counterparties' solvency, which meant they stopped lending to each other.\n\nThe compound interaction that made 2008 uniquely devastating was what can be termed 'collateral contagion' — a mechanism with direct relevance to current risks around the Treasury basis trade (cm-069) and private credit (cm-074). The chain operated as follows: falling home prices caused mortgage defaults, which caused CDO downgrades from AAA to junk (sometimes skipping intermediate grades entirely). These CDOs were being used as collateral in the repo market — banks and broker-dealers had pledged them to borrow short-term cash for their daily operations. When the CDOs were downgraded, they became ineligible as repo collateral, triggering margin calls. But the collateral was illiquid — there were no buyers for downgraded CDOs at any price, because no one could determine the actual loss content. Banks that faced margin calls could not sell the collateral and could not substitute other collateral (because their entire balance sheet was contaminated by the same exposure). The only option was to hoard cash, which meant refusing to lend to other banks, which caused LIBOR (the interbank lending rate) to spike. The LIBOR spike transmitted the crisis from structured credit markets to every business and household with a floating-rate loan or credit facility tied to LIBOR — which was essentially the entire economy. What had been a localized housing problem in a few overheated markets became a systemic global liquidity freeze because the collateral chain connected mortgage borrowers in Nevada to banks in Germany through a web of derivatives that regulators had never mapped and market participants had never stress-tested.\n\nThe indicator readings document the progression with devastating clarity. The 10Y-2Y yield curve was deeply inverted through mid-2006 and stayed inverted for a full year — one of the longest inversions in the modern era. It then bull-steepened dramatically in late 2007 as the Fed began emergency rate cuts, signaling that the bond market had shifted from pricing inflation risk to pricing systemic collapse. The VIX bottomed at approximately 10 in early 2007 (extreme complacency), stepped up to the 20s in August 2007 (first stress), and then exploded to an all-time closing high of 80.86 on October 27, 2008. The TED spread (the difference between 3-month LIBOR and 3-month T-bills) and the LIBOR-OIS spread — the most precise measures of interbank stress — blew out in August 2007, providing the earliest market-based signal that the banking system was seizing. High-yield bond spreads widened from approximately 240 basis points in June 2007 to over 2,000 basis points by late 2008 — an eightfold increase that priced in widespread corporate default. The S&P 500 took 65 months — over five years — to surpass its October 2007 high, finally doing so in March 2013. The earliest actionable warning was the yield curve inversion in February 2006, which gave approximately 18 months of lead time before the acute phase. The ABX crash in February 2007 provided a more targeted, higher-confidence signal with approximately 6 months of lead time before the interbank freeze.",
      "dataSource": "FCIC Report, Federal Reserve H.15 (LIBOR/TED spread), S&P/Case-Shiller, ABX-HE indices (Markit), VIX historical data, BIS banking statistics, Bernanke/Geithner memoirs",
      "leadTime": "Yield curve inversion (Feb 2006) gave 18 months lead. ABX crash (Feb 2007) gave 6 months precision lead to interbank freeze. Bear Stearns funds (Jun 2007) gave 2 months lead. BNP Paribas (Aug 2007) was effectively the start of the crisis itself.",
      "reliability": "Highest possible — the most documented financial crisis in history, with a congressional investigation (FCIC), thousands of academic papers, and detailed memoirs from key participants. The collateral contagion mechanism is precisely understood and directly applicable to current risks.",
      "historicalPrecedents": [
        "1930s Great Depression (systemic bank failure cascade)",
        "1990s Japan (real estate bubble → zombie banking → lost decade)",
        "2023 SVB/regional bank crisis (modern echo of confidence-driven bank runs)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "asset-bubble",
        "yield-curve"
      ],
      "tags": [
        "2008-GFC",
        "Lehman",
        "subprime",
        "CDO",
        "ABX",
        "collateral-contagion",
        "LIBOR",
        "repo",
        "shadow-banking",
        "BNP-Paribas",
        "TED-spread",
        "margin-calls"
      ]
    },
    {
      "id": 97,
      "sourceId": "cm-098",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "general",
      "title": "2020 COVID Crash — Trigger Sequence Analysis",
      "content": "The March 2020 COVID crash is unique among financial crises in that it combined an exogenous biological shock with pre-existing financial fragilities to produce the fastest equity market decline in history, followed by the fastest recovery in history — a V-shaped pattern that was entirely a function of unprecedented policy intervention rather than organic economic healing. The pre-conditions revealed that the financial system's apparent resilience during the 2010s expansion masked profound structural fragilities. Corporate leverage had reached historic extremes: the BBB-rated segment of the investment-grade bond market — the lowest tier before junk status — had expanded from approximately $0.7 trillion in 2008 to over $3.5 trillion by 2019, creating a massive pool of debt one downgrade away from triggering forced selling by investment-grade-only mandates. Just-in-time supply chains had optimized for efficiency at the expense of resilience, eliminating the inventory buffers that had historically provided a cushion during demand shocks. Most critically for market dynamics, the explosion of mechanical and systematic trading strategies — risk parity funds, short-volatility strategies, commodity trading advisors (CTAs), and volatility-targeting pension allocations — had created a market microstructure that was architecturally dependent on low and stable volatility. These strategies collectively managed trillions of dollars and shared a common vulnerability: they would all mechanically sell equities when volatility spiked, regardless of fundamental value.\n\nThe trigger sequence began not with COVID but with a warning signal six months earlier that the financial plumbing was already under stress. In September 2019, overnight repo rates spiked to 10% — a four-fold increase from normal levels — for reasons that were never fully explained. The Federal Reserve was forced to intervene with emergency repo operations that it characterized as routine 'technical adjustments' but that market participants recognized as an ad hoc bailout of the repo market. The Fed then initiated a program of Treasury bill purchases that it emphatically insisted was 'not QE' — a distinction that fooled no one and that became a meme for the pretense that the system was functioning normally when it clearly was not. This September 2019 repo spike revealed that the Treasury market's plumbing was too fragile to handle even modest disruptions to collateral flows — a vulnerability that would prove catastrophic six months later. Meanwhile, the yield curve had briefly inverted in May-August 2019, prompting the Fed to cut rates three times in a 'mid-cycle adjustment' — an unusual step that signaled underlying economic weakness beneath the surface calm of record equity prices.\n\nThe COVID trigger arrived into this fragile environment with devastating speed. On December 31, 2019, the WHO office in China received the first report of pneumonia cases in Wuhan. Markets ignored the news entirely. By mid-January 2020, cases were spreading in Asia. Markets remained indifferent. On February 19, 2020, the S&P 500 reached a new all-time high — the market was pricing COVID risk at zero. On February 24, Italy announced a lockdown of northern provinces, and the reality that COVID would not be contained to Asia hit Western markets simultaneously. What followed was a 34% decline in the S&P 500 over just 23 trading days — the fastest bear market in history. But the crash was not primarily driven by fundamental reassessment of COVID's economic impact. It was driven by a mechanical, self-reinforcing deleveraging cascade that the September 2019 repo spike had foreshadowed.\n\nThe compound interaction that defined the COVID crash was 'Dash for Cash' combined with systematic deleveraging — a mechanism that broke the most fundamental assumption of portfolio construction. Lockdown announcements zeroed service-sector revenue overnight. Corporations drew down credit lines simultaneously, creating a cash drain on the banking system. Investors of all types — institutional, retail, sovereign — began selling assets to raise US dollars. The volatility spike triggered automated selling by Risk Parity funds (which mechanically reduce equity allocation when vol rises), Vol-Targeting strategies (same mechanism), and CTA trend-following systems (which flip to short when price trends reverse). The crucial and terrifying development was that these strategies sold Treasuries alongside equities. Risk parity's core premise is that bonds and stocks move inversely — when stocks fall, bonds rise, providing a hedge. In March 2020, this correlation flipped: stocks and bonds fell simultaneously because the selling was driven by volatility mechanics and liquidity needs, not by fundamental analysis. The Treasury market — the bedrock of global collateral and the ultimate safe haven — broke. Bid-ask spreads on US Treasuries, normally 1-2 basis points, widened to 10-15 basis points. Dealers stopped making markets. The Federal Reserve was forced to become the buyer of last resort, ultimately purchasing over $1.5 trillion in Treasuries and launching emergency lending facilities that dwarfed the 2008 response. The VIX surged from approximately 13 in mid-February to an all-time closing high of 82.69 on March 16 — a six-fold increase in three weeks. Investment-grade and high-yield credit spreads gapped so violently that the primary corporate bond market froze entirely: companies could not issue new bonds at any price. The recovery took just 6 months — by August 2020, the S&P 500 had reclaimed its February high — making it the fastest bear market recovery in history, fueled by approximately $5 trillion in combined fiscal stimulus and Fed balance sheet expansion. The earliest actionable warning was the September 2019 repo market spike, which gave 5 months of lead time that the financial system's plumbing was too fragile to withstand any significant shock.",
      "dataSource": "Federal Reserve repo operations data, VIX historical data, BIS Bulletin on March 2020 market turmoil, US Treasury bid-ask spread data (TRACE), academic literature (Vissing-Jorgensen, He & Krishnamurthy), NBER",
      "leadTime": "September 2019 repo spike gave 5 months lead on plumbing fragility. Yield curve inversion (May-Aug 2019) gave 6-7 months lead on economic vulnerability. COVID itself gave approximately 7 weeks from first WHO report to market crash, but only 3 trading days from Italy lockdown to freefall.",
      "reliability": "Very high for the mechanism analysis. The Dash for Cash / systematic deleveraging dynamic is thoroughly documented by BIS, the Fed, and academic researchers. The specific vulnerability (repo fragility + systematic strategy crowding) is directly relevant to current basis trade risk (cm-069). The recovery speed is NOT generalizable — it was entirely a function of the largest monetary and fiscal intervention in history.",
      "historicalPrecedents": [
        "1987 Black Monday (mechanical/portfolio insurance-driven crash)",
        "2008 GFC repo freeze (similar plumbing failure, slower development)",
        "2019 September repo spike (direct precursor and warning signal)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "credit-debt",
        "asset-bubble",
        "general"
      ],
      "tags": [
        "COVID",
        "dash-for-cash",
        "risk-parity",
        "repo-crisis",
        "systematic-deleveraging",
        "fastest-recovery",
        "VIX-82",
        "Treasury-market-break",
        "BBB-cliff",
        "September-2019-repo"
      ]
    },
    {
      "id": 98,
      "sourceId": "cm-099",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "systemic-banking",
      "title": "2022 UK Gilt Crisis — Trigger Sequence Analysis",
      "content": "The September 2022 UK gilt crisis is the most recent and most instructive example of a hidden leverage doom loop nearly destroying a major sovereign bond market in a matter of hours. It is uniquely valuable for crisis monitoring because it revealed a mechanism — liability-driven investment (LDI) margin procyclicality — that is structurally identical to risks currently embedded in the Treasury basis trade (cm-069) and central counterparty margin models (cm-073). The pre-conditions combined macroeconomic stress with a completely hidden structural vulnerability. Globally, inflation had surged to multi-decade highs in the aftermath of COVID stimulus and the Russia-Ukraine energy shock. The Bank of England, like most central banks, was locked into a tightening cycle, having raised rates six times between December 2021 and August 2022. UK gilt yields had been rising steadily throughout 2022 as rates increased and inflation expectations became entrenched. This was stressful but orderly. The hidden vulnerability was in the UK pension system. Over the preceding decade, UK defined-benefit pension funds had adopted Liability-Driven Investment (LDI) strategies on a massive scale — approximately 1.5 trillion pounds in aggregate. LDI strategies use interest rate derivatives (swaps) and leveraged gilt positions to match the duration of pension liabilities. The leverage allowed pension funds to hedge their long-duration liabilities while freeing up capital for return-seeking investments in equities and alternatives. The leverage ratios were significant: many LDI funds held gilts at 3x to 4x leverage, using the gilts themselves as collateral for the borrowing that funded the leveraged position.\n\nThe trigger was political rather than economic, but the mechanism it exposed was purely financial. On September 23, 2022, new Chancellor Kwasi Kwarteng announced a 'mini-Budget' containing 45 billion pounds in unfunded tax cuts — the largest tax cut package in 50 years — at a time when UK inflation was running at 10% and the Bank of England was actively tightening to fight that inflation. The announcement was a direct contradiction of the BoE's monetary policy: fiscal expansion during monetary contraction. Bond markets reacted with extreme violence. Between September 23 and September 26, the pound crashed from $1.12 to $1.03 against the dollar — approaching parity for the first time in history. But the currency move was a sideshow. The real crisis was in gilt yields: 30-year gilt yields spiked approximately 120 basis points in just three trading days, a move of historically unprecedented speed and magnitude for a G7 sovereign bond. In a normally functioning market, this would have been a severe but manageable repricing. In a market with 1.5 trillion pounds in leveraged LDI positions, it was nearly catastrophic.\n\nThe LDI margin call doom loop operated with terrifying mechanical precision. The spike in gilt yields meant that the value of gilts held as collateral by LDI funds dropped sharply. This triggered margin calls from the counterparties (banks and clearinghouses) that had lent to the LDI funds. To meet the margin calls, LDI funds needed to raise cash. The fastest way to raise cash was to sell gilts — the very asset whose price decline had triggered the margin call. But selling gilts drove gilt prices lower, which triggered more margin calls on other LDI funds, which forced more gilt selling, which drove prices lower still. This is the canonical doom loop: falling asset prices leading to margin calls leading to forced selling leading to lower prices leading to more margin calls. The loop was self-reinforcing and, left unchecked, would have continued until either the pension funds were bankrupt or the gilt market ceased to function. By September 27-28, market participants reported that multiple pension funds were within hours of insolvency. The Bank of England's own assessment, later revealed, was that the gilt market was approximately 48 hours from complete dysfunction — a failure that would have rendered UK pension funds unable to pay retirees, a social and political catastrophe on top of the financial one.\n\nThe Bank of England's response was extraordinary and revealed the fundamental tension at the heart of the crisis. On September 28, the BoE abandoned its planned quantitative tightening (gilt selling) and announced emergency quantitative easing — gilt purchases of up to 65 billion pounds — to stabilize the market. This was a central bank that was simultaneously trying to tighten monetary policy to fight 10% inflation and being forced to ease monetary policy to prevent a pension system collapse. The contradiction was unsustainable and was only resolved by the political removal of the trigger: Kwarteng was fired on October 14, Prime Minister Truss resigned on October 20 (after just 45 days in office), and the unfunded tax cuts were reversed. The gilt market stabilized within approximately two months. The indicator readings were dramatic: the MOVE Index (bond market volatility, the fixed-income equivalent of the VIX) surged to 156, its highest level since the 2008 financial crisis. UK sovereign credit default swap spreads widened significantly relative to other G7 nations, pricing in genuine sovereign credit risk for a country that had been considered virtually risk-free. The MOVE Index had been trending upward through the first half of 2022, giving approximately 6 months of lead time that sovereign bond market liquidity was dangerously thin across the developed world — not just in the UK. The critical lesson for current crisis monitoring is that the LDI doom loop mechanism is not unique to UK pensions. Any market structure where leveraged participants hold an asset as both their investment position and their collateral — and where margin calls force selling of that same asset — contains the same doom loop potential. The Treasury basis trade (cm-069), where hedge funds hold leveraged Treasury positions funded by repo using those same Treasuries as collateral, is structurally identical. Central counterparty margin procyclicality (cm-073), where margin requirements increase during volatility and force selling that increases volatility, is the same mechanism at one remove.",
      "dataSource": "Bank of England Financial Stability Reports, House of Commons Treasury Committee hearings, MOVE Index historical data, UK DMO gilt auction data, Pension Protection Fund data, BIS working papers on LDI and leverage",
      "leadTime": "MOVE Index uptrend through H1 2022 gave 6 months lead that bond market liquidity was fragile. The acute crisis developed in 3 trading days from trigger (Sep 23 mini-Budget) to BoE emergency intervention (Sep 28). The doom loop was approximately 48 hours from pension system insolvency when intervention occurred.",
      "reliability": "Very high — recent event with extensive documentation from the BoE, parliamentary inquiries, and market data. The LDI doom loop mechanism is precisely understood and has been analyzed in detail by the BoE, BIS, and academic researchers. The structural parallel to basis trade and CCP margin risks is direct and well-recognized by regulators.",
      "historicalPrecedents": [
        "1992 UK ERM crisis (political trigger → market crisis → policy reversal)",
        "2020 March Treasury market break (bond market liquidity evaporation)",
        "1998 LTCM (leveraged fixed-income positions → margin calls → forced selling)"
      ],
      "compoundsWith": [
        "systemic-banking",
        "yield-curve",
        "currency-inflation"
      ],
      "tags": [
        "UK-Gilt",
        "LDI",
        "Truss",
        "margin-doom-loop",
        "pension-crisis",
        "MOVE-index",
        "48-hours",
        "BoE-emergency",
        "Kwarteng",
        "mini-Budget",
        "leverage-procyclicality"
      ]
    },
    {
      "id": 99,
      "sourceId": "cm-100",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "geopolitical",
      "title": "EU Infrastructure Decoupling: Digital and Financial Sovereignty",
      "content": "Europe is executing a systematic replacement of US-controlled digital and financial infrastructure with EU-governed sovereign alternatives. This is not policy rhetoric -- it is operational infrastructure deployment with regulatory mandates enforcing adoption.\n\nThe digital layer: Euro Office launched March 27, 2026 (open-source ONLYOFFICE fork, EU governance, coalition: Ionos, Nextcloud, XWiki, OpenProject, Open-Xchange). Full DOCX/XLSX/PPTX compatibility eliminates switching costs. Stable release Summer 2026. Germany, Denmark, and Estonia have active sovereign cloud procurement mandates. Gartner projects EU sovereign cloud spend tripling to EUR 23B by 2027. The structural catalyst is the CLOUD Act (2018): US law enforcement can compel data from US companies regardless of physical server location, making US cloud providers legally non-compliant with EU data sovereignty requirements. This is settled law, not a political dispute -- the conflict is structural and permanent.\n\nThe financial layer: SEPA Instant is mandated across all Eurozone banks (receive by Jan 2025, send by Oct 2025). Wero (European Payments Initiative), backed by 16 major banks, is absorbing national payment systems -- iDEAL (Netherlands) migrated January 2026. Account-to-account (A2A) payments cost 30-70% less than card interchange networks. A2A captured 17% of EU e-commerce in 2024, projected EUR 850B by 2026. EU interchange fees are capped by regulation. Visa and Mastercard cannot compete on speed, coverage, or cost within the EU framework.\n\nCrisis monitoring relevance: this decoupling is a slow-moving structural shift (2-5 year lead time) that compounds with dollar hegemony erosion and multipolar financial system fragmentation. When EU payments settle in EUR via Wero/SEPA and EU data runs on sovereign infrastructure, US visibility into European economic activity (transaction patterns, consumption data, credit signals) degrades. This is not a crisis trigger in isolation but a compound amplifier that worsens data quality for US-based risk models and reduces early-warning signal fidelity for institutions monitoring EU economic conditions.",
      "dataSource": "Euro Office coalition press releases (Mar 2026), European Payments Initiative (EPI) Wero reports, SEPA Instant regulation (EC 2023), Gartner sovereign cloud projections, CLOUD Act (2018) legal analysis, Dutch National Bank iDEAL migration announcements",
      "leadTime": "2-5 years for full structural effect. Digital layer: Summer 2026 Euro Office stable release begins broad EU adoption. Financial layer: SEPA Instant mandate operational, Wero absorbing national systems through 2026-2027. Full US data visibility degradation: 3-5 years.",
      "reliability": "High for mandate enforcement (SEPA Instant is law, not suggestion). Medium-high for adoption trajectory (Wero/Euro Office face inertia but have strong incentive structure). Lower for BRICS/Global South adoption spillover (geopolitical logic is sound but timing uncertain).",
      "historicalPrecedents": [
        "SWIFT exclusion of Iran (2012) and Russia (2022) as catalyst for alternative payment infrastructure development",
        "EU GDPR (2018) as precedent for successful regulatory override of US tech platform behavior",
        "China UnionPay development as domestic card network built in response to Visa/Mastercard dependency concerns"
      ],
      "compoundsWith": [
        "currency-inflation",
        "geopolitical",
        "systemic-banking"
      ],
      "tags": [
        "EU",
        "digital-sovereignty",
        "euro-office",
        "SEPA-Instant",
        "Wero",
        "EPI",
        "CLOUD-Act",
        "A2A",
        "geopolitical",
        "decoupling",
        "Visa",
        "Mastercard",
        "iDEAL",
        "sovereign-cloud",
        "infrastructure-replacement",
        "slow-burn-trigger"
      ]
    },
    {
      "id": 100,
      "sourceId": "cm-101",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "geopolitical",
      "title": "EU Decoupling + Dollar Hegemony Erosion: Compound Data Visibility Crisis",
      "content": "The intersection of EU infrastructure decoupling (cm-100) and the existing dollar hegemony erosion trend creates a compound effect that is more significant than either trend in isolation: the progressive degradation of US financial surveillance and early-warning capabilities.\n\nThe mechanism operates through data flows. Currently, US institutions -- including the Federal Reserve, Treasury, major investment banks, and intelligence agencies -- have deep visibility into global economic activity because that activity flows through US-controlled rails. SWIFT transaction data passes through US jurisdiction. Microsoft Azure and Google Cloud process European enterprise data under US law (CLOUD Act). Visa and Mastercard transaction data gives US institutions near-real-time consumption signal across 450M European consumers.\n\nWhen EU payments shift to Wero/SEPA: EUR-denominated, EU-jurisdiction, EU-governed settlement. Transaction data no longer flows through US financial infrastructure. When EU enterprise data shifts to Euro Office + sovereign cloud: EU-jurisdiction storage and processing. CLOUD Act compulsion is no longer applicable. When BRICS nations adopt similar frameworks (the natural extension of the EU open-source sovereignty blueprint): the geographic coverage of US data blindspots expands.\n\nFor crisis monitoring specifically, this compound creates a second-order risk: US-based risk models that currently incorporate European economic signals (PMI, consumption patterns, credit flows) will face increasing data latency and degraded signal quality as EU decoupling progresses. Models calibrated on comprehensive global data visibility will produce less reliable crisis forecasts as that visibility narrows. The irony is that the crisis monitoring infrastructure most dependent on complete data (Fed, IMF, BIS) is most exposed to the visibility degradation.\n\nThis is not an acute crisis trigger -- it is a slow-moving degradation of the early-warning capability that is supposed to detect acute triggers. Lead time to material impact: 3-7 years. The compound with dollar reserve currency erosion (as more trade settles outside USD) is that both financial transaction data AND dollar-settlement data degrade simultaneously, creating a dual-channel visibility reduction.\n\nFor AI-Rev engine relevance: sectors most dependent on US financial data infrastructure (banks, div-fin, insurance) face increasing model risk as their credit and risk assessment frameworks lose European signal quality. This is a tail risk for AI-powered financial modeling tools that assume data completeness.",
      "dataSource": "EU infrastructure decoupling analysis (cm-100), dollar reserve currency erosion literature (BIS working papers, Gita Gopinath IMF research), SWIFT data flow analysis, CLOUD Act jurisdictional analysis, Federal Reserve international financial surveillance frameworks",
      "leadTime": "3-7 years for material degradation of US data visibility into EU economic activity. Acute crisis interaction: if a EU-originated financial stress event (e.g., sovereign debt crisis, banking stress) develops after decoupling is advanced, US early-warning detection capability may be materially reduced, potentially extending reaction time by days to weeks -- significant in fast-moving market events.",
      "reliability": "Medium. The directional logic is sound and follows from infrastructure facts. The timeline and magnitude of impact on US risk model accuracy is speculative -- US institutions will adapt and develop alternative data sources. The compound effect is real but the severity depends on adaptation speed.",
      "historicalPrecedents": [
        "Asian Financial Crisis (1997): US institutions had limited visibility into Thai, Korean, Indonesian financial flows before the crisis -- geographic data blindspots contributed to delayed recognition",
        "Russian financial isolation post-2022: Western institutions lost real-time data on Russian economic activity; sanctions effectiveness assessment became significantly harder"
      ],
      "compoundsWith": [
        "currency-inflation",
        "geopolitical",
        "systemic-banking",
        "credit-debt"
      ],
      "tags": [
        "EU",
        "dollar-hegemony",
        "data-visibility",
        "compound-risk",
        "early-warning-degradation",
        "SWIFT",
        "CLOUD-Act",
        "geopolitical",
        "decoupling",
        "surveillance-capability",
        "Fed",
        "IMF",
        "BIS",
        "AI-Rev",
        "financial-modeling",
        "multipolar",
        "BRICS",
        "slow-burn-compound"
      ]
    },
    {
      "id": 101,
      "sourceId": "cm-102",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "1970s Gold Bugs: End of the Dollar Predictions",
      "content": "Howard Ruff, Harry Browne, and 1970s 'Gold Bugs' predicted total collapse of fiat currency, hyperinflation, and the end of the US Dollar (1978-1980). Claimed: societal collapse, permanent hyperinflation, return to barter and gold. ACTUAL: Paul Volcker raised rates aggressively, crushing inflation. 1980s-1990s saw massive expansion and historic bull market. Gold crashed from 1980 peak, stagnated 20 years.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [
        "1970s stagflation"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "gold-bugs",
        "hyperinflation",
        "fiat-collapse",
        "1970s",
        "wrong-prediction"
      ]
    },
    {
      "id": 102,
      "sourceId": "cm-103",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "1987 Black Monday: Great Depression 2.0 Predictions",
      "content": "Ravi Batra (author of 'The Great Depression of 1990') and financial media predicted Black Monday signaled a direct repeat of 1929 and imminent global Great Depression. ACTUAL: Fed provided immediate liquidity ('Greenspan put'). Real economy largely unaffected. Market recovered to new all-time highs within two years, leading into massive 1990s boom.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [
        "1929 crash",
        "cm-029"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "black-monday",
        "1987",
        "depression",
        "wrong-prediction"
      ]
    },
    {
      "id": 103,
      "sourceId": "cm-104",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "1990s Japan Mirror: US Lost Decade Predictions",
      "content": "Various macroeconomists predicted US real estate and stock markets would suffer a 'Lost Decade' exactly mirroring Japan's 1989 bubble burst (1990-1992). Claimed: decades of deflation, stagnant growth, collapsing real estate. ACTUAL: US economy fundamentally differed from Japan's leveraged, insular banking system. US entered the 'Roaring 90s' driven by tech innovation and PC/Internet boom.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [
        "Japan 1989 bubble"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "japan-mirror",
        "lost-decade",
        "1990s",
        "wrong-prediction"
      ]
    },
    {
      "id": 104,
      "sourceId": "cm-105",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Y2K: Global Infrastructure Collapse Predictions",
      "content": "Gary North, IT survivalists, and mainstream media predicted global infrastructure collapse from Y2K date bug (1998-1999). Claimed: planes falling from sky, power grids failing permanently, banking system collapse, return to agrarian living. ACTUAL: Massive global remediation ($300B) fixed bugs in advance. Jan 1, 2000 passed with only minor, isolated glitches.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "y2k",
        "infrastructure",
        "2000",
        "wrong-prediction"
      ]
    },
    {
      "id": 105,
      "sourceId": "cm-106",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Dot-Com Crash: Internet Is A Fad Predictions",
      "content": "Clifford Stoll, Paul Krugman, and broad tech-bear sentiment claimed internet was a passing fad with no more impact than fax machine (1998-2000). ACTUAL: Stock bubble absolutely popped (Nasdaq fell 78%), but internet fundamentally rewired the entire global economy. Surviving companies (Amazon, eBay) became titans.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - partially correct",
      "reliability": "PARTIALLY RIGHT",
      "historicalPrecedents": [
        "cm-028",
        "cm-096"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "dot-com",
        "internet-fad",
        "2000",
        "partial-prediction"
      ]
    },
    {
      "id": 106,
      "sourceId": "cm-107",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Peak Oil: End of Industrial Civilization Predictions",
      "content": "Matthew Simmons, James Howard Kunstler predicted global oil production had irreversibly peaked, causing $200+ oil, permanent food shortages, and collapse of industrialized society (2005-2008). ACTUAL: High prices incentivized massive innovation. Fracking and horizontal drilling unlocked massive shale reserves. US became net exporter. Oil prices later crashed.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [
        "cm-094"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "peak-oil",
        "energy",
        "2005",
        "wrong-prediction"
      ]
    },
    {
      "id": 107,
      "sourceId": "cm-108",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Burry/Paulson: Subprime Housing Collapse (CORRECT)",
      "content": "Michael Burry, John Paulson, Steve Eisman predicted US housing market built on fraudulent subprime mortgages would collapse, crashing global financial system (2005-2007). ACTUAL: Exactly as predicted. Subprime collapsed, triggering GFC. Lehman, Bear Stearns failed or were bailed out.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "2-3 years (Burry started shorting in 2005, crisis hit 2007-2008)",
      "reliability": "CORRECT",
      "historicalPrecedents": [
        "cm-029",
        "cm-097"
      ],
      "compoundsWith": [
        "credit-debt",
        "asset-bubble"
      ],
      "tags": [
        "doom-scorecard",
        "subprime",
        "burry",
        "2008",
        "correct-prediction",
        "skin-in-game"
      ]
    },
    {
      "id": 108,
      "sourceId": "cm-109",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Post-2008 QE Hyperinflation Predictions",
      "content": "Peter Schiff, Ron Paul, ShadowStats predicted QE and ZIRP post-2008 would immediately cause Weimar-style hyperinflation and total destruction of USD (2009-2011). Claimed gold at $10,000+. ACTUAL: Inflation remained stubbornly low (under 2%) for a decade. QE money sat on bank balance sheets as excess reserves, did not circulate in real economy.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "qe",
        "hyperinflation",
        "schiff",
        "2009",
        "wrong-prediction",
        "skin-in-game"
      ]
    },
    {
      "id": 109,
      "sourceId": "cm-110",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Grexit/Eurozone Collapse Predictions",
      "content": "Nouriel Roubini ('Dr. Doom') and macro hedge funds predicted Greece's default would force Grexit, causing domino contagion (Italy, Spain, Portugal) shattering the Eurozone (2010-2012). ACTUAL: ECB's Draghi promised 'whatever it takes.' Through bailouts, austerity, and monetary policy, Eurozone remained intact.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [],
      "compoundsWith": [
        "geopolitical"
      ],
      "tags": [
        "doom-scorecard",
        "grexit",
        "eurozone",
        "roubini",
        "2010",
        "wrong-prediction",
        "skin-in-game"
      ]
    },
    {
      "id": 110,
      "sourceId": "cm-111",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "China Hard Landing Predictions",
      "content": "Kyle Bass and George Soros predicted China's massive debt pile and shadow banking would trigger imminent catastrophic 'hard landing' and 40% currency devaluation (2015-2016). ACTUAL: China's command economy absorbed shock via capital controls, state-directed lending, targeted stimulus. Yuan devalued only modestly. Structural issues remained but apocalypse was avoided.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction timing was wrong",
      "reliability": "WRONG TIMING",
      "historicalPrecedents": [
        "cm-030"
      ],
      "compoundsWith": [
        "geopolitical"
      ],
      "tags": [
        "doom-scorecard",
        "china",
        "hard-landing",
        "bass",
        "soros",
        "2015",
        "wrong-timing",
        "skin-in-game"
      ]
    },
    {
      "id": 111,
      "sourceId": "cm-112",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "COVID: Greater Depression Predictions",
      "content": "Nouriel Roubini, Harry Dent, and widespread media consensus predicted COVID lockdowns would cause a 'Greater Depression' worse than 1930s with years of double-digit unemployment (2020). ACTUAL: Unprecedented multi-trillion-dollar fiscal and monetary stimulus. 2020 market crash was shortest in history (33 days), followed by explosive bull market and rapid employment recovery.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [
        "cm-098"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "covid",
        "depression",
        "roubini",
        "dent",
        "2020",
        "wrong-prediction"
      ]
    },
    {
      "id": 112,
      "sourceId": "cm-113",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "Post-COVID Weimar Hyperinflation Predictions",
      "content": "Jack Dorsey, Tucker Carlson, crypto-maximalists predicted post-COVID inflation was permanently out of control, entering Weimar/Zimbabwe-style hyperinflation marking end of fiat era (2021-2022). ACTUAL: Inflation peaked at 9.1% US. Fed's fastest rate-hiking cycle in history pushed rates past 5%. Inflation steadily cooled toward historical norms without hyperinflation.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "hyperinflation",
        "post-covid",
        "crypto",
        "2021",
        "wrong-prediction"
      ]
    },
    {
      "id": 113,
      "sourceId": "cm-114",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "2023 Recession: Soft Landing Impossible Predictions",
      "content": "Bloomberg Economics ('100% chance of recession'), Jamie Dimon ('Economic Hurricane') predicted aggressive Fed rate hikes guaranteed severe recession in 2023 (2022-2023). ACTUAL: US achieved elusive 'soft landing.' GDP growth remained robust (>2.5%), unemployment near historic lows, stock market hit all-time highs.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction was wrong",
      "reliability": "WRONG",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "recession",
        "soft-landing",
        "bloomberg",
        "dimon",
        "2023",
        "wrong-prediction",
        "skin-in-game"
      ]
    },
    {
      "id": 114,
      "sourceId": "cm-115",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "doom-scorecard",
      "title": "AI Job Apocalypse / Extinction Predictions",
      "content": "Eliezer Yudkowsky, sensationalized Goldman Sachs reports predicted generative AI would instantly displace hundreds of millions of workers or result in human extinction (2023-2025). ACTUAL: AI adoption proved powerful productivity tool ('copilot') rather than wholesale replacer. Job displacement was localized, aggregate employment remained strong.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - prediction timing wrong so far",
      "reliability": "WRONG TIMING",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "ai-doom",
        "job-displacement",
        "extinction",
        "yudkowsky",
        "2023",
        "wrong-timing"
      ]
    },
    {
      "id": 115,
      "sourceId": "cm-116",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "doom-scorecard",
      "title": "50-Year Doom Scorecard: Meta-Analysis",
      "content": "Overall accuracy rate for 'extreme severity' doom predictions: <5%. While mild recessions and corrections are common, predictions involving dollar collapse, hyperinflation, or end of civilization are almost universally incorrect due to society's capacity for adaptation and institutional intervention.\n\nPATTERN OF PREDICTORS: Dedicated 'permabears' (Schiff, Dent) are consistently wrong because their worldview is ideological rather than analytical. They mistake secular debt cycles for immediate doom. Individuals who are actually correct (Burry) are pragmatic practitioners performing deep, localized data analysis, not macro-philosophers.\n\nWHY DOOM SELLS: Evolutionary loss aversion — we're wired to pay more attention to threats. Doomsaying is lucrative: drives media engagement and allows pundits to sell 'cures' (gold, survival gear, newsletters, crypto).\n\nTHE BURRY PROBLEM: Survivorship bias. Because Burry was spectacularly right once, he's treated as infallible oracle. Ignores thousands who were wrong in 2008, and that Burry himself subsequently predicted several crashes that never materialized. Being right about one idiosyncratic event doesn't grant permanent clairvoyance.\n\nRISK ANALYSIS vs BUSINESS MODEL: Legitimate risk analysis is probabilistic, specific in mechanisms, adjusts when facts change, comes from individuals with capital on the line. Doomsaying-as-business relies on vague catastrophic language, constantly moves goalposts ('the collapse is delayed, not canceled'), and is usually accompanied by subscription or asset class sales.",
      "dataSource": "Gemini research - 50-Year Doom Scorecard",
      "leadTime": "N/A - meta-analysis",
      "reliability": "high - based on 50 years of documented predictions vs outcomes",
      "historicalPrecedents": [
        "cm-102 through cm-115"
      ],
      "compoundsWith": [],
      "tags": [
        "doom-scorecard",
        "meta-analysis",
        "loss-aversion",
        "permabears",
        "burry-problem",
        "doomsaying-economics"
      ]
    },
    {
      "id": 116,
      "sourceId": "cm-117",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "guru-trap",
      "title": "The Guru Trap: Why Big-Call Oracles Fail Next",
      "content": "The 'Guru Trap' or 'Curse of the Big Call': when an analyst correctly predicts a Black Swan, the market crowns them an oracle. But the exact psychological traits and analytical frameworks that allowed them to spot the original crisis become the exact reasons they fail in the next cycle. This is so common in finance it's practically a law. Four distinct failure modes explain why hitting one grand slam leads to subsequent strikeouts:\n\n1. HAMMER AND NAIL BIAS (Over-extrapolation): When an analyst uses a specific tool to predict a crisis (e.g., finding hidden leverage in obscure derivatives), they look for that exact same mechanism in the next cycle. They fight the last war.\n\n2. UNDERESTIMATING INSTITUTIONAL ADAPTATION ('The Fed Put'): Doomsayers identify real mathematical problems but assume the system will sit there and let the math play out. They repeatedly underestimate Central Banks' and politicians' willingness to change rules, print money, or orchestrate bailouts.\n\n3. CONFUSING 'INEVITABLE' WITH 'IMMINENT' (The Timing Problem): A doomsayer correctly identifies unsustainable trajectory (e.g., sovereign debt). But 'unsustainable' can last 30 years. Because they were rewarded for acting early once, they mistake long-term structural headwinds for immediate trading signals.\n\n4. AUDIENCE CAPTURE (The Guru Trap proper): If you predict standard 8% returns, no one invites you on CNBC. Once you achieve 'Oracle' status by predicting doom, media and investors expect you to continue finding dangers. You become incentivized to ignore bullish data and over-index on bearish data to feed the brand.",
      "dataSource": "Gemini research - Guru Trap Analysis",
      "leadTime": "N/A - meta-framework",
      "reliability": "high - pattern confirmed across multiple decades and predictors",
      "historicalPrecedents": [
        "cm-108",
        "cm-102 through cm-115"
      ],
      "compoundsWith": [],
      "tags": [
        "guru-trap",
        "meta-framework",
        "prediction-failure",
        "cognitive-bias",
        "audience-capture",
        "hammer-nail",
        "timing-problem"
      ]
    },
    {
      "id": 117,
      "sourceId": "cm-118",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "guru-trap",
      "title": "Guru Trap Case Study: Michael Burry Post-2008",
      "content": "WHAT HE GOT RIGHT: 2007-2008 Subprime Mortgage Collapse. Did granular research — read hundreds of pages of obscure MBS prospectuses, realized underlying loans were fraudulent and bound to default when teaser rates reset.\n\nWHAT HE GOT WRONG: Repeated premature crash calls. 'Index Fund Bubble' (2019) — claimed passive investing was a liquidity bubble similar to CDOs. Various imminent crash warnings (2022-2023). In 2023, famously tweeted single word 'Sell' then deleted account as market rallied 20%+.\n\nFAILURE MODE: Hyper-Sensitivity to Structural Flaws (Hammer and Nail). Fundamentally wired to find hidden structural rot. After making fortune finding a fatal flaw everyone missed, began seeing fatal flaws everywhere. Underestimated structural inertia of 401(k) inflows and lack of immediate trigger events (like mortgage resets) in passive investing.",
      "dataSource": "Gemini research - Guru Trap Analysis",
      "leadTime": "N/A - case study",
      "reliability": "high - documented public record",
      "historicalPrecedents": [
        "cm-108"
      ],
      "compoundsWith": [],
      "tags": [
        "guru-trap",
        "burry",
        "index-fund-bubble",
        "hammer-nail-bias",
        "case-study"
      ]
    },
    {
      "id": 118,
      "sourceId": "cm-119",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "guru-trap",
      "title": "Guru Trap Case Study: John Paulson Post-2008",
      "content": "WHAT HE GOT RIGHT: 'The Greatest Trade Ever' (2007-2008). With Paolo Pellegrini, realized housing was massive bubble, bought cheap credit default swaps against toxic mortgage bonds. Made $20 billion for his firm.\n\nWHAT HE GOT WRONG: 2010s Hyperinflation and Gold bet. Believed Fed QE would destroy the Dollar. Launched gold fund, bet heavily on hyperinflation and fiat collapse. Lost billions as gold stagnated and US enjoyed decade-long low-inflation expansion.\n\nFAILURE MODE: Misunderstanding Monetary Plumbing (Institutional Adaptation). Took purely historical view of money printing (Weimar Germany analogy). Failed to understand post-2008 QE was largely trapped as excess bank reserves, not circulating in real economy. Applied housing-market level of confidence to a totally different macroeconomic mechanism.",
      "dataSource": "Gemini research - Guru Trap Analysis",
      "leadTime": "N/A - case study",
      "reliability": "high - documented public record",
      "historicalPrecedents": [
        "cm-109"
      ],
      "compoundsWith": [],
      "tags": [
        "guru-trap",
        "paulson",
        "gold",
        "qe",
        "monetary-plumbing",
        "case-study"
      ]
    },
    {
      "id": 119,
      "sourceId": "cm-120",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "guru-trap",
      "title": "Guru Trap Case Study: Meredith Whitney Post-2008",
      "content": "WHAT SHE GOT RIGHT: 2007-2008 Banking Collapse. As Oppenheimer analyst, accurately predicted Oct 2007 that Citigroup was so undercapitalized it would slash dividend and raise billions. Triggered massive sell-off, effectively fired starting gun on financial crisis. Dead right about bank balance sheets.\n\nWHAT SHE GOT WRONG: 2010 Municipal Bond Collapse. On 60 Minutes (late 2010), predicted '50 to 100 sizable defaults' by US cities/counties, totaling hundreds of billions. Muni bond market panicked. Defaults never happened.\n\nFAILURE MODE: Underestimating Institutional Willpower. Applied corporate bank analysis framework to municipal governments. But states and cities are not banks — they have taxing authority, immense political capital, and federal backing. Assumed math alone would dictate outcome, ignoring the political machinery that prevents local government bankruptcy.",
      "dataSource": "Gemini research - Guru Trap Analysis",
      "leadTime": "N/A - case study",
      "reliability": "high - documented public record",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "guru-trap",
        "whitney",
        "municipal-bonds",
        "institutional-willpower",
        "case-study"
      ]
    },
    {
      "id": 120,
      "sourceId": "cm-121",
      "section": "historical-precedents",
      "category": "historical-precedent",
      "subcategory": "guru-trap",
      "title": "Guru Trap Case Study: Nouriel Roubini 'Dr. Doom' Post-2008",
      "content": "WHAT HE GOT RIGHT: 2008 Housing and Systemic Collapse. Speaking at IMF in 2006, warned of devastating US housing bust, oil shocks, deep recession. Correctly identified macro imbalances in global debt.\n\nWHAT HE GOT WRONG: Almost every recovery phase since. Predicted severe 'double-dip' recession (2010-2011). Repeatedly predicted Eurozone collapse. Missed historic 2010s bull market entirely by maintaining perpetual defensive posture.\n\nFAILURE MODE: Audience Capture and Permabear Psychology. Became global celebrity as 'Dr. Doom.' When entire brand, speaking fee structure, and public identity are wrapped in pessimism, it becomes incredibly difficult to publicly pivot and say 'Actually, things look pretty good.' The incentive structure locks you into bearishness.",
      "dataSource": "Gemini research - Guru Trap Analysis",
      "leadTime": "N/A - case study",
      "reliability": "high - documented public record",
      "historicalPrecedents": [
        "cm-110",
        "cm-112"
      ],
      "compoundsWith": [],
      "tags": [
        "guru-trap",
        "roubini",
        "dr-doom",
        "audience-capture",
        "permabear",
        "case-study"
      ]
    },
    {
      "id": 121,
      "sourceId": "cm-122",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Food-Energy Nexus: The Fossil Fuel Content of Food",
      "content": "Modern agriculture converts fossil fuels into calories. Haber-Bosch consumes ~2% of global energy, 3-5% of global natural gas. 48-50% of global population fed exclusively by synthetic nitrogen (Erisman 2008). Without it, earth supports ~3.5-4B people. Systemic ratio: 10 cal fossil fuel per 1 cal food (Pimentel). Farm-gate grain EROI positive (~1:3), beef 35-40:1 negative. Processing uses 36% of food-system energy. Fertilizer is 33-45% of direct crop operating costs (corn/wheat). Diesel is 64% of farm fuel ($9.9B/yr US). Retail food: 3.5% is transport, 11.8% goes to farmer.",
      "dataSource": "USDA ERS, FAO, Pimentel (Cornell), Erisman (Nature Geoscience 2008), Smil",
      "leadTime": "Structural",
      "reliability": "Very high",
      "historicalPrecedents": [
        "2008 oil spike: FAO crisis",
        "2022 Russia-Ukraine: FAO all-time high 159.7",
        "North Korea 1991: fertilizer loss = famine (240K-3.5M dead)"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "food-energy",
        "haber-bosch",
        "nitrogen",
        "fertilizer",
        "diesel",
        "food-prices"
      ]
    },
    {
      "id": 122,
      "sourceId": "cm-123",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "food-energy-nexus",
      "title": "Hormuz Closure: Food Supply Chain Transmission",
      "content": "Hormuz carries 21M bpd oil (21% of global) AND 20% of global LNG. Full closure = 12-14M bpd unmitigated shortfall after max pipeline bypass (6-8.5M bpd). LNG has ZERO pipeline alternatives. 30% of seaborne ammonia and 50% of urea exports transit Hormuz. Price transmission: energy spike hits grocery shelf in 3-6 months, seasonal storage items in 12 months. Prices are sticky downward — fast to rise, very slow to fall.",
      "dataSource": "EIA, IEA, shipping data, market prices Apr 2026",
      "leadTime": "Immediate",
      "reliability": "High",
      "historicalPrecedents": [
        "1984 Tanker War",
        "2019 Iran tanker seizures",
        "Never fully closed in modern era"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus",
        "currency-inflation",
        "geopolitical"
      ],
      "tags": [
        "hormuz",
        "oil",
        "lng",
        "fertilizer-prices",
        "diesel",
        "energy-shock"
      ]
    },
    {
      "id": 123,
      "sourceId": "cm-124",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "food-energy-nexus",
      "title": "Hormuz 6-Month and 12-Month Closure Scenarios",
      "content": "6-month: Brent avg $115/bbl (EIA). Missing 2026 fertilizer window = 10-20% global cereal yield reduction by autumn. Food inflation severe Q4 2026/Q1 2027. Europe fails winter gas storage without Qatari LNG. 12-month: SPR depletion risk. Grain reserves (31.9% of consumption, ~3.5 months) begin synchronized collapse. Multi-year famine in developing nations. Just-in-time food system has no buffer for 12-month fertilizer deficit.",
      "dataSource": "EIA projections, FAO grain reserves, agricultural modeling",
      "leadTime": "6-12 months",
      "reliability": "Medium — scenario projections",
      "historicalPrecedents": [
        "2008: oil + ethanol = FAO crisis",
        "2022: Russia/Ukraine = FAO 159.7",
        "North Korea: fertilizer loss = 30% grain drop"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus",
        "geopolitical",
        "systemic-banking"
      ],
      "tags": [
        "hormuz",
        "scenario",
        "food-crisis",
        "famine-risk",
        "grain-reserves"
      ]
    },
    {
      "id": 124,
      "sourceId": "cm-125",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "FAO Food Price Index and Political Instability Threshold",
      "content": "NECSI (2011): FAO crossing 210 = widespread violent riots. 2008: crossed 210, 60+ food riots in 30 countries. 2011: hit 230+, catalyzed Arab Spring (Tunisia, Egypt, Libya fell). Current Mar 2026: 128.5 (rising 2.4% MoM). Wheat +4.3%, veg oil +5.1%, sugar +7.2%. Distance to instability threshold: ~64% more rise needed. At current trajectory: reaches danger zone Q1 2027 if Hormuz stays closed.",
      "dataSource": "FAO, NECSI 2011 study",
      "leadTime": "3-6 months from energy spike to food price peak",
      "reliability": "High — 210 threshold held in 2008 and 2011",
      "historicalPrecedents": [
        "2008 food riots (60+ in 30 countries)",
        "2011 Arab Spring",
        "2022 Sri Lanka collapse"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "fao",
        "food-prices",
        "political-instability",
        "arab-spring",
        "threshold"
      ]
    },
    {
      "id": 125,
      "sourceId": "cm-126",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Global Fertilizer Concentration and Chokepoint Risk",
      "content": "Top 5 producers = ~38% global revenue. Potash: Canada/Russia/Belarus 60%+ of reserves. Phosphate: Morocco 70%+ of rock phosphate. Nitrogen: tied to fossil fuel (China, India, US, Russia). 50% of seaborne urea and 30% of ammonia transit Hormuz. 2022: 70% of European ammonia capacity forced offline when gas exceeded production cost. Green ammonia 2-3x grey ($700-1000 vs $300-400/ton). NEOM (largest green project) targets 2027. Organic yield gap: 20-25%, exceeding 40% for corn/wheat.",
      "dataSource": "Industry reports, NEOM project data, 2022 European data",
      "leadTime": "Structural",
      "reliability": "High",
      "historicalPrecedents": [
        "2022 Russia/China export quotas, European shutdowns",
        "Sri Lanka 2021 fertilizer ban disaster"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus",
        "geopolitical"
      ],
      "tags": [
        "fertilizer",
        "concentration",
        "chokepoint",
        "green-ammonia"
      ]
    },
    {
      "id": 126,
      "sourceId": "cm-127",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Regional Food Security Vulnerability Map",
      "content": "US: Most insulated (domestic gas $2.71-2.79, domestic food/energy). But diesel stays high (global crude pricing). Europe: Highly exposed — lost Russian gas, now losing Qatari LNG. Industrial energy crisis repeat. South/SE Asia: Pakistan/Bangladesh ~100% Hormuz LNG dependent. Cannot afford spot replacement. MENA: Egypt massive wheat importer, food prices = regime stability. Sub-Saharan Africa: Most vulnerable — smallholder farmers cannot absorb $1100/ton ammonia. Immediate yield collapse, famine within same calendar year.",
      "dataSource": "EIA, FAO, country import dependency data",
      "leadTime": "3-12 months by region",
      "reliability": "High — structural import dependencies",
      "historicalPrecedents": [
        "2011 Arab Spring (Egypt, Tunisia, Libya)",
        "2022 Sri Lanka",
        "North Korea 1994-1999 famine"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "food-security",
        "regional",
        "vulnerability",
        "europe",
        "africa",
        "asia"
      ]
    },
    {
      "id": 127,
      "sourceId": "cm-128",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "global-south",
      "title": "The Crisis Is Already Here: Global South Permanent Crisis States",
      "content": "Western crisis forecasting assumes a stable baseline that might be disrupted. For 2+ billion people, there is no baseline stability. Sudan (civil war since 2023, 25M facing acute hunger), Yemen (8 years of war, 21M needing humanitarian aid), Haiti (gang control of 80% of Port-au-Prince, complete state collapse), Myanmar (military coup, civil war, 18M needing aid), the Sahel (cascading coups in Mali, Burkina Faso, Niger), and large parts of Central America live in permanent polycrisis. This is not a forecasting exercise for them. The Western crisis indicators in this brain (CAPE, yield curves, insurance withdrawal) are measuring whether comfortable countries might experience what these populations already endure. The analytical value of including this: Western crises do not exist in isolation. Global South instability feeds migration, supply chain disruption, commodity volatility, and geopolitical realignment that directly impacts the indicators we track.",
      "dataSource": "UN OCHA, World Food Programme, UNHCR, World Bank",
      "leadTime": "N/A (current state)",
      "reliability": "High (institutional data)",
      "historicalPrecedents": [
        "Arab Spring 2011 (food prices)",
        "Venezuelan collapse 2015-present",
        "Sri Lanka default 2022"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "global-south",
        "permanent-crisis",
        "sudan",
        "yemen",
        "haiti",
        "myanmar",
        "sahel",
        "western-lens"
      ]
    },
    {
      "id": 128,
      "sourceId": "cm-129",
      "section": "frameworks",
      "category": "framework",
      "subcategory": "global-south",
      "title": "IMF/World Bank Debt Trap Architecture as Moloch Mechanism",
      "content": "The abstract Moloch/coordination-failure framework has a concrete, documented mechanism: the international debt architecture. Structural Adjustment Programs (SAPs) from IMF/World Bank since the 1980s required developing nations to cut social spending, privatize state assets, and open markets to foreign capital in exchange for loans. Result: 52 low-income countries spend more on debt service than on healthcare (Jubilee Debt Campaign, 2023). Zambia, Ghana, Sri Lanka, Ethiopia all defaulted or restructured 2022-2024. The debt is denominated in dollars they cannot print, owed to institutions controlled by creditor nations, with conditions set by the creditors. This is not conspiracy theory — it is documented institutional design. The relevance to Western crisis forecasting: sovereign debt stress in the Global South is a leading indicator for commodity disruption, migration waves, and geopolitical realignment.",
      "dataSource": "Jubilee Debt Campaign, World Bank IDS, IMF Article IV Reports",
      "leadTime": "6-24 months (debt stress to commodity/migration impact)",
      "reliability": "High (institutional data)",
      "historicalPrecedents": [
        "Latin American debt crisis 1980s",
        "Asian financial crisis 1997",
        "Greek debt crisis 2010-2015",
        "Sri Lanka 2022",
        "Zambia 2020"
      ],
      "compoundsWith": [
        "credit-debt",
        "currency-inflation",
        "geopolitical"
      ],
      "tags": [
        "global-south",
        "imf",
        "world-bank",
        "debt-trap",
        "structural-adjustment",
        "sovereign-debt"
      ]
    },
    {
      "id": 129,
      "sourceId": "cm-130",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "global-south",
      "title": "Food Insecurity as Permanent State vs Theoretical Risk",
      "content": "Western crisis models treat food price spikes as triggers. For 735 million people globally (UN FAO 2023), food insecurity is not a trigger — it is the permanent condition. A Hormuz closure that raises Western food prices 15-20% is an inconvenience for Americans and a death sentence for populations already at caloric minimums. WFP funding cuts ($4B shortfall in 2023) mean the humanitarian buffer that previously absorbed commodity shocks is disappearing. When the next major food price spike hits, there is no safety net for the bottom billion. The cascade: commodity spike hits countries already at breaking point, triggers political instability (Arab Spring pattern), generates migration pressure, feeds into Western political polarization, which further degrades international cooperation. The Western viewer experiences this as 'immigration crisis' without seeing the food-price-to-failed-state-to-migration chain.",
      "dataSource": "UN FAO State of Food Security 2023, WFP, FEWS NET",
      "leadTime": "3-12 months (commodity spike to migration wave)",
      "reliability": "High (institutional data)",
      "historicalPrecedents": [
        "Arab Spring 2011 (Russian wheat export ban)",
        "2007-2008 global food crisis",
        "Horn of Africa famine 2022"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "global-south",
        "food-insecurity",
        "wfp",
        "famine",
        "migration",
        "arab-spring-pattern"
      ]
    },
    {
      "id": 130,
      "sourceId": "cm-131",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "global-south",
      "title": "Western Foreign Policy as Crisis Driver (Not Just Geopolitical Trigger)",
      "content": "The geopolitical trigger category treats foreign policy events as external shocks. For the Global South, Western foreign policy IS the crisis, not a trigger for one. Iraq invasion (2003) destabilized the entire Middle East for two decades. Libya intervention (2011) created a failed state and weapons pipeline to the Sahel. Afghanistan withdrawal (2021) collapsed the economy overnight. US sanctions regimes (Iran, Venezuela, Cuba, Syria) create humanitarian crises that compound with existing vulnerabilities. The analytical point is not moral judgment — it is forecasting accuracy. Models that treat Western foreign policy as neutral miss that we are often the cause. Cascade: Western intervention creates instability, instability creates refugee flows, refugee flows create political backlash in the West, backlash elects governments that pursue more aggressive foreign policy. The feedback loop is the crisis.",
      "dataSource": "Costs of War Project (Brown University), UNHCR, Congressional Research Service",
      "leadTime": "1-10 years (intervention to full cascade)",
      "reliability": "High (academic/institutional)",
      "historicalPrecedents": [
        "Iraq 2003 to ISIS to Syrian refugee crisis to Brexit/Trump",
        "Libya 2011 to Sahel destabilization",
        "Afghanistan withdrawal 2021"
      ],
      "compoundsWith": [
        "geopolitical",
        "energy-shock",
        "currency-inflation"
      ],
      "tags": [
        "global-south",
        "foreign-policy",
        "intervention",
        "feedback-loop",
        "refugees",
        "sanctions"
      ]
    },
    {
      "id": 131,
      "sourceId": "cm-132",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "labor-market",
      "title": "Construction & Manufacturing Jobs: The Only Recession Signal That Matters",
      "content": "Total non-farm payrolls are a lagging indicator that provides virtually zero recession warning. The economy has 158 million payrolls, but only 13% of them matter for recessions. Construction and manufacturing (21 million jobs) are responsible for nearly 100% of private sector job losses in recessions and provide the longest lead times. Manufacturing jobs peaked an average of 29 MONTHS before the last three recessions started. Construction peaked an average of 9 months before (20 months before 2008). The headline number treats all 158 million jobs as equal, but education/healthcare went 687 months (57 years) without year-over-year contraction. Including them in recession analysis dilutes the cyclical signal to nothing.\n\nIn 4 of the last 7 recessions, construction + manufacturing accounted for MORE than 100% of all private sector job losses (meaning the rest of the economy actually added jobs during the recession). In 2 more, they were 90%+. Even in 2008 (the worst case), they were 53% of all losses despite being only 18% of payrolls.\n\nWithin these sectors, the signal narrows further: durable goods manufacturing (7.8M jobs, expansion-recession spread 12.2%) matters more than non-durables (4.7M jobs, spread 5.8%). Residential construction is more interest-rate-sensitive and faster-moving than non-residential, making it the most important sub-signal.\n\nThe framework: watch residential construction and durable goods manufacturing. When they peak and start declining, the recession clock has started — even if the headline jobs number looks fine for another 9-29 months.",
      "dataSource": "EPB Research (Eric Basmajian), BLS data analysis",
      "leadTime": "9-29 months (construction 9mo avg, manufacturing 29mo avg)",
      "reliability": "High (based on 7 recessions of BLS data, 1970-2020)",
      "historicalPrecedents": [
        "Every recession 1970-2020",
        "2008 GFC (20-month construction lead)",
        "2001 recession (manufacturing secular decline overlay)"
      ],
      "compoundsWith": [
        "yield-curve",
        "credit-debt",
        "asset-bubble"
      ],
      "tags": [
        "leading-indicator",
        "construction",
        "manufacturing",
        "recession-signal",
        "durable-goods",
        "residential",
        "lagging-indicator",
        "non-farm-payrolls",
        "epb-research"
      ]
    },
    {
      "id": 132,
      "sourceId": "cm-137",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Petroleum Products Beyond Fuel: Plastics, Pesticides, Pharmaceuticals",
      "content": "99% of plastics and 99% of synthetic pesticides are petroleum-derived. A sustained Hormuz closure doesn't just hit fuel and fertilizer — it threatens medical supply chains, food packaging, crop protection, and pharmaceutical manufacturing at the precursor level.",
      "dataSource": "Gemini Deep Research (Apr 2026), industry data",
      "leadTime": "3-6 months for downstream shortages",
      "reliability": "High — petrochemical dependencies are structural facts",
      "historicalPrecedents": [
        "2022 European resin shortages from gas price spike"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus",
        "geopolitical"
      ],
      "tags": [
        "plastics",
        "pesticides",
        "pharmaceuticals",
        "petrochemicals",
        "medical-supply",
        "food-packaging"
      ]
    },
    {
      "id": 133,
      "sourceId": "cm-138",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "US Energy Insulation Reality: The Independence Myth Tested",
      "content": "US is net petroleum exporter but net crude importer (6.4M bpd). Domestic light sweet crude can't fully feed heavy-sour refinery infrastructure. Oil is globally priced at the margin — a Texas producer sells to the highest global bidder, not at a domestic discount. Natural gas genuinely insulated via pipeline, but LNG exports (14.6 Bcf/d record, targeting 23+ Bcf/d) create pull that keeps Henry Hub elevated.",
      "dataSource": "EIA, refinery data, LNG export capacity data",
      "leadTime": "Immediate — already visible in gas/diesel prices",
      "reliability": "High — trade flows and pricing mechanisms are well-documented",
      "historicalPrecedents": [
        "Every global oil shock since 1973 has affected US prices despite domestic production"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus"
      ],
      "tags": [
        "energy-independence",
        "crude-imports",
        "global-pricing",
        "diesel-paradox",
        "lng-exports"
      ]
    },
    {
      "id": 134,
      "sourceId": "cm-139",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Global South Cascading Impacts: Caribbean, Pacific Islands, SE Asia",
      "content": "Beyond MENA/Africa: Dominican Republic (90% imported energy, 37K bpd petroleum), Haiti (diesel-dependent grid, priced out of spot market), Pacific Islands (up to 25% of GDP on fuel imports, imports dropped to 25% of normal). Fishing fleets confined to port. Outer island communities cut off from food/medical supplies.",
      "dataSource": "Gemini Deep Research, UN/WFP data, Pacific Islands Forum data",
      "leadTime": "Immediate — already manifesting",
      "reliability": "High — import dependencies are structural",
      "historicalPrecedents": [
        "2008 food crisis hit small island developing states hardest",
        "2022 Sri Lanka collapse"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "global-south",
        "pacific-islands",
        "caribbean",
        "haiti",
        "fishing",
        "wfp",
        "imf",
        "famine-risk"
      ]
    },
    {
      "id": 135,
      "sourceId": "cm-140",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "food-energy-nexus",
      "title": "The Barrel Breakdown: What 42 Gallons of Crude Becomes",
      "content": "42 gallons of crude yields ~45 gallons of refined products (processing gain). 45% gasoline, 30% diesel/distillates, 10% jet fuel, plus hydrocarbon gas liquids and naphtha. Naphtha is cracked into ethylene, propylene, butadiene — the 'holy trinity' of modern manufacturing: synthetic rubber (tires), synthetic fibers (polyester clothing), adhesives, paints, detergents. Every branch of the product tree is simultaneously affected by a crude supply disruption.",
      "dataSource": "EIA refinery product slate data",
      "leadTime": "Structural — always present",
      "reliability": "Very high — refinery yields are measured precisely",
      "historicalPrecedents": [],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus"
      ],
      "tags": [
        "refinery",
        "naphtha",
        "ethylene",
        "petrochemicals",
        "product-tree",
        "barrel-breakdown"
      ]
    },
    {
      "id": 136,
      "sourceId": "cm-133",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "energy-shock",
      "title": "Hormuz Triple Farm Input Crisis: Nitrogen + Pesticides + Diesel Simultaneously",
      "content": "Farmers face three simultaneous petroleum-derived input crises, not one. 1) NITROGEN FERTILIZER (natural gas feedstock): Haber-Bosch process requires natural gas; Hormuz carries 20% of global LNG. 2) PESTICIDES: 99% of all synthetic pesticides are petroleum-derived. 50-75% of commercial pesticide volume is petroleum 'inert' ingredients (solvents, surfactants, microplastic coatings). Without affordable pesticides, crop losses to insects and blight compound by 20-40% within a single season. 3) DIESEL: At $5.40+/gal (April 2026), tractor fuel and farm equipment costs near record highs. The compound effect is likely multiplicative, not additive: nitrogen-starved crops are MORE vulnerable to pest damage, while high diesel costs make it uneconomical to even harvest reduced yields. US farmers pre-buy contracts but physical delivery depends on shipping through or around Hormuz. Pre-buying a price is not the same as having fertilizer in the barn.",
      "dataSource": "AI-Stocks Hormuz briefing (April 2026), FAO, petroleum industry data",
      "leadTime": "1-2 growing seasons (immediate price impact, yield impact next harvest)",
      "reliability": "High (institutional + industry data)",
      "historicalPrecedents": [
        "Sri Lanka 2022 organic farming ban (30-50% yield drops)",
        "Cuba Special Period 1990s (lost Soviet fertilizer + oil)",
        "2022 Russia/Ukraine fertilizer export disruption"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "hormuz",
        "nitrogen",
        "pesticides",
        "diesel",
        "compound-effect",
        "farm-inputs",
        "food-security",
        "triple-crisis"
      ]
    },
    {
      "id": 137,
      "sourceId": "cm-134",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "energy-shock",
      "title": "US 'Energy Independence' Is a Pricing Myth: Global Marginal Pricing",
      "content": "Despite political claims, the US energy market is structurally tied to global benchmarks. Key facts: 1) US is net exporter of petroleum products but net IMPORTER of unrefined crude (6.4M bpd in 2024). 2) US shale produces 'light sweet' crude but refineries were built for 'heavy sour' crude. US exports its own light crude and imports heavy crude. 3) MARGINAL PRICING: If Brent hits $106/bbl, a Texas producer will NOT sell domestically for $70. They export to the highest bidder. Domestic refiners must pay global rates, passing cost to consumers. 4) NATURAL GAS is genuinely insulated (domestic pipelines, no marine chokepoint). BUT the US is world's largest LNG exporter (14.6 Bcf/d in 2025, targeting 23+ Bcf/d). Hormuz closure taking Qatari LNG offline causes European/Asian buyers to bid up US LNG, pulling domestic prices higher. 5) DIESEL PARADOX: US produces and exports diesel, yet diesel prices track global Brent crude. When Hormuz takes Middle Eastern refined distillates offline, international buyers bid up US diesel to global levels.",
      "dataSource": "EIA, petroleum industry data, AI-Stocks Hormuz briefing April 2026",
      "leadTime": "Immediate (prices respond within days of disruption)",
      "reliability": "High (EIA institutional data)",
      "historicalPrecedents": [
        "2022 Russia/Ukraine oil price spike",
        "1973 OPEC embargo",
        "2019 Saudi Aramco drone attack"
      ],
      "compoundsWith": [
        "geopolitical",
        "currency-inflation",
        "credit-debt"
      ],
      "tags": [
        "energy-independence",
        "marginal-pricing",
        "diesel",
        "lng",
        "crude-oil",
        "hormuz",
        "myth"
      ]
    },
    {
      "id": 138,
      "sourceId": "cm-135",
      "section": "trigger-analysis",
      "category": "trigger",
      "subcategory": "global-south",
      "title": "Pacific Islands and Small States: 25% of GDP on Diesel, Now Priced Out",
      "content": "Pacific Island nations are the world's most diesel-dependent economies. Tuvalu, Samoa, Vanuatu spent up to 25% of GDP importing diesel for power grids, boats, and vehicles BEFORE the crisis. Samoa: two-thirds of energy from imported diesel. Tonga: 80%. As of April 2026, Pacific fuel imports have dropped to 25% of previous levels due to supply hoarding in Asia. Fishing fleets (crucial for Fiji, Solomon Islands) confined to port - devastating employment, caloric intake, and tuna export revenues. Outer island communities rely entirely on diesel ferries for food and medical supplies, effectively cutting them off. Haiti: lowest electrification in Western hemisphere, diesel generators power fragmented grids and telecom towers. Priced out of spot market = rolling national blackouts, loss of water purification, communications dark zones. Dominican Republic: 90% energy from imported fossil fuels, 37,000 bpd petroleum consumption. IMF emergency financing requests from frontier markets in Sub-Saharan Africa and South Asia as dollar reserves deplete.",
      "dataSource": "AI-Stocks Hormuz briefing April 2026, IMF, World Bank, Pacific Islands Forum",
      "leadTime": "Immediate (already occurring as of April 2026)",
      "reliability": "High (institutional data + scenario modeling)",
      "historicalPrecedents": [
        "2008 oil spike impact on Pacific Islands",
        "Sri Lanka 2022 fuel crisis"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "pacific-islands",
        "haiti",
        "diesel-dependency",
        "small-states",
        "imf",
        "global-south",
        "fuel-crisis"
      ]
    },
    {
      "id": 139,
      "sourceId": "cm-136",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "energy-shock",
      "title": "Petroleum Beyond Fuel: Medical Plastics, Pharmaceuticals, Food Packaging",
      "content": "99% of global plastics manufactured from fossil fuels (44% coal, 40% petroleum, 8% natural gas). 4-8% of annual global oil consumption goes directly into plastic production (projected 20% by 2050). MEDICAL: 300 billion gloves, 129 billion masks, 17 billion IV bags per year — all petroleum-derived (polypropylene, PVC). Syringes, tubing, blister packaging universally petrochemical. Hormuz closure causes resin inflation, driving up hospital overhead and disrupting sterile equipment supply. FOOD PACKAGING: Most modern food packaging (PE, PET) is petroleum-derived. Scarce packaging resin = dramatically shorter shelf life for perishables = spike in food waste, compounding agricultural crisis. PHARMACEUTICALS: 99% of pharmaceutical feedstocks and reagents are petrochemical-derived. Benzene, toluene, xylene are foundational building blocks for APIs including antibiotics and aspirin. Even though US imports most finished APIs from India/China, those countries use precursor chemicals from Gulf-refined oil. NAPHTHA: 1.55 gallons per barrel become hydrocarbon gas liquids, 0.34 gallons become naphtha — cracked into ethylene, propylene, butadiene (synthetic rubber for tires, polyester clothing, adhesives, paints, detergents).",
      "dataSource": "AI-Stocks Hormuz briefing April 2026, petroleum industry data, WHO medical supply data",
      "leadTime": "3-6 months (resin supply chains have buffer inventory)",
      "reliability": "High (industry data)",
      "historicalPrecedents": [
        "COVID-19 PPE shortage 2020",
        "2021 Texas freeze disrupting petrochemical plants"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical"
      ],
      "tags": [
        "plastics",
        "medical-supply",
        "pharmaceuticals",
        "naphtha",
        "food-packaging",
        "petroleum-products",
        "hormuz"
      ]
    },
    {
      "id": 140,
      "sourceId": "cm-nitrogen-yield-curve",
      "section": "frameworks",
      "category": "hormuz-compound-crisis",
      "subcategory": "general",
      "title": "Nitrogen Reduction vs Crop Yield: The Non-Linear Curve",
      "content": "The relationship between nitrogen application and crop yield follows a diminishing returns curve, not a linear relationship. The first units of nitrogen produce the most dramatic growth. Cutting nitrogen 30% from peak only drops yields 10-15%. Cutting 60% drops yields 25-30%. Only at 100% removal do catastrophic 40-60% drops occur for cereals. MAJOR EXCEPTION: Soybeans fix their own nitrogen — zero impact from synthetic nitrogen loss. REAL-WORLD EVIDENCE: Sri Lanka (2021-22) banned synthetic fertilizers — rice plummeted 32-54%, tea fell 18-28.7% within ONE season. China long-term field trials: zero-fertilizer plots yield 40-60% less than fertilized. Drop is immediate (first harvest) but worsens over 2-3 years as soil nitrogen bank from previous seasons exhausts. YIELD TABLE (global averages): Corn at -30% N = -10% yield, at -60% N = -30% yield, at -100% N = -58% yield. Rice at -30% N = -9%, -60% N = -25%, -100% N = -49%. Wheat at -30% N = -11%, -60% N = -28%, -100% N = -48%. KEY INSIGHT FOR INFOGRAPHICS: Spreading reduced nitrogen evenly across all acreage is better than concentrating full-rate on fewer acres, because the first units of N per hectare do the most work (diminishing returns curve).",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "nitrogen",
        "yield-curve",
        "diminishing-returns",
        "sri-lanka",
        "haber-bosch",
        "soybeans"
      ]
    },
    {
      "id": 141,
      "sourceId": "cm-haber-bosch-dependency",
      "section": "frameworks",
      "category": "hormuz-compound-crisis",
      "subcategory": "general",
      "title": "The 48% Fed by Haber-Bosch: Who Specifically Is Dependent",
      "content": "48-50% of humanity is fed by the Haber-Bosch process (Vaclav Smil, Erisman et al.). Nearly half of nitrogen atoms in human dietary protein globally originated in a synthetic fertilizer plant. DEPENDENCY MAP: Most dependent = Asia (China, India, Bangladesh, Indonesia) and Middle East/North Africa (Egypt). Diets reliant on corn, wheat, rice are most Haber-Bosch dependent. SUBSISTENCE FARMERS: Sub-Saharan Africa is generally NOT part of the 48% — they lack capital for synthetic fertilizers and already operate at low baseline yields. A Hormuz shock doesn't directly cut their yields but destroys them INDIRECTLY via skyrocketing global grain prices. HORMUZ CHOKEPOINT: 16-20 million metric tons of nitrogen fertilizer (33% of global seaborne trade) transits Hormuz annually. Key exporters: Qatar, Saudi Arabia, Oman, Iran. COUNTRY EXPOSURE: US/Canada buffered for current season (pre-buy in fall/winter, domestic production) but face devastating costs next season. India massively exposed — over half of fertilizer imports from Persian Gulf. Brazil imports 85%+ of nitrogen, roughly half via Hormuz. DOUBLE-HIT COUNTRIES: Egypt, Bangladesh need Gulf LNG and imported urea to grow domestic crops AND import massive wheat. Disruption crushes domestic harvests while pricing them out of international markets.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "haber-bosch",
        "hormuz",
        "nitrogen-dependency",
        "food-security",
        "chokepoint",
        "global-south"
      ]
    },
    {
      "id": 142,
      "sourceId": "cm-historical-fertilizer-shocks",
      "section": "frameworks",
      "category": "hormuz-compound-crisis",
      "subcategory": "general",
      "title": "Historical Precedents: What Happens When Fertilizer Supply Collapses",
      "content": "NORTH KOREA (1990s): Lost Soviet fossil fuels and synthetic fertilizers. Today: rice yield 3.5 tons/ha vs South Korea's 5.2+ tons/ha — nearly 2:1 productivity gap from the same geography. CUBA SPECIAL PERIOD (1990s): Lost 1.3 million tons of Soviet chemical fertilizers and oil supply overnight. Agricultural production crashed 47%. Average daily calorie intake plunged from 2,600 to 1,000-1,500. WWI AND WWII: Nitrogen plants commandeered for TNT/munitions instead of ammonia. Crop yields plummeted across Europe. Farmers forced back to crop rotation with nitrogen-fixing legumes (clover, alfalfa) — slow process that couldn't maintain pre-war yields. 2022 UKRAINE/RUSSIA: Sanctions sent urea futures soaring. Developing nation farmers reduced application rates, directly contributing to global food price inflation and localized shortages.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "north-korea",
        "cuba",
        "historical-precedent",
        "fertilizer-shock",
        "food-crisis"
      ]
    },
    {
      "id": 143,
      "sourceId": "cm-compound-crisis-multiplicative",
      "section": "frameworks",
      "category": "hormuz-compound-crisis",
      "subcategory": "general",
      "title": "The Compound Effect Is MULTIPLICATIVE, Not Additive: Nitrogen + Pesticides + Diesel",
      "content": "The April 2026 agricultural crisis is not a single input shock — it is a synergistic compound crisis. Three simultaneous disruptions: (1) Nitrogen: natural gas feedstock disrupted by Hormuz. (2) Pesticides: 99% of synthetic crop protection chemicals rely on petroleum derivatives (solvents, surfactants). (3) Diesel: high fuel costs restrict tractor passes for tilling, spraying, harvesting. THE PESTICIDE FACTOR: Without pesticides, crop losses to weeds/insects/pathogens leap from ~20% to 40-70%. Sri Lanka ban: 81% of farmers reported higher weed infestation, 73% reported severe insect damage. WHY MULTIPLICATIVE NOT ADDITIVE: Nitrogen-starved plants are physically weaker, grow slower, lack canopy needed to shade out weeds. If a farmer loses 30% to nitrogen shortage AND loses pesticide access, aggressive weeds outcompete the stunted crop for whatever residual soil nutrients remain. Combined effect frequently results in TOTAL CROP FAILURE, not just mathematical 50% reduction. THIS IS THE KEY INSIGHT: A 60% nitrogen reduction alone = ~30% yield loss. But 60% nitrogen reduction + no pesticides + expensive diesel = potentially 60%+ yield loss or total failure. The compound scenario is where the catastrophic numbers live.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "compound-crisis",
        "multiplicative",
        "pesticides",
        "diesel",
        "total-crop-failure",
        "synergistic"
      ]
    },
    {
      "id": 144,
      "sourceId": "cm-141",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "energy-shock",
      "title": "Seaborne Oil Buffer Exhausted: Last Pre-War Tankers Arrived April 20",
      "content": "The world's safety cushion of oil already at sea when the Hormuz conflict began is now gone. The last tankers that crossed Hormuz before fighting started in late February finally reached destinations (Malaysia, California) by April 20. Physical commodity traders warn this is the inflection point — from here, every day of closure is a net draw on reserves with no replacement in transit.",
      "dataSource": "Patrick Boyle (Apr 26, 2026), physical commodity trader sources",
      "leadTime": "Immediate — buffer exhausted",
      "reliability": "High — tanker tracking data is precise",
      "historicalPrecedents": [
        "1973 embargo had similar 'pipeline draining' effect but smaller scale"
      ],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus"
      ],
      "tags": [
        "seaborne-buffer",
        "tanker-tracking",
        "trafigura",
        "supply-loss",
        "2030-equilibrium"
      ]
    },
    {
      "id": 145,
      "sourceId": "cm-142",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "energy-shock",
      "title": "US Shale Producers Refusing to Increase Production (Dallas Fed Survey)",
      "content": "Despite White House pressure, US shale executives are actively resisting calls to increase production. Dallas Fed survey shows operators taking a 'do nothing' approach to 2026 budgets. Rational capital allocation: they can't plan long-term rig deployments when prices swing wildly on presidential tweets. If they overproduce and a peace deal crashes prices, they're left holding the bag. This is learned behavior from prior overdrilling busts.",
      "dataSource": "Patrick Boyle citing Dallas Fed energy survey (Apr 2026)",
      "leadTime": "Months — drilling decisions take 6-12 months to reach production",
      "reliability": "High — Dallas Fed survey is authoritative on shale sector sentiment",
      "historicalPrecedents": [
        "2014-2016 shale overproduction bust",
        "2020 COVID demand destruction aftermath"
      ],
      "compoundsWith": [
        "energy-shock"
      ],
      "tags": [
        "shale-production",
        "dallas-fed",
        "drill-refusal",
        "capital-allocation",
        "policy-whiplash"
      ]
    },
    {
      "id": 146,
      "sourceId": "cm-143",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "energy-shock",
      "title": "Europe Jet Fuel Vulnerability: 70% Self-Sufficiency, 50 Days Reserves",
      "content": "Europe cannot refine enough jet fuel domestically — capacity covers only 70% of airline demand. Currently sitting on ~50 days of reserves (normal operating level). Kepler modeling shows stocks will fall precipitously if Hormuz flows don't normalize by June. If the US bans refined fuel exports to protect domestic prices, Europe's aviation sector hits a wall.",
      "dataSource": "Patrick Boyle citing Politico, Kepler/Economist modeling (Apr 2026)",
      "leadTime": "June 2026 — reserves start critical decline",
      "reliability": "High — refinery capacity data is structural",
      "historicalPrecedents": [],
      "compoundsWith": [
        "energy-shock",
        "food-energy-nexus"
      ],
      "tags": [
        "jet-fuel",
        "europe-aviation",
        "refinery-capacity",
        "kepler-modeling",
        "export-ban-risk"
      ]
    },
    {
      "id": 147,
      "sourceId": "cm-144",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "energy-shock",
      "title": "Qatar Helium Chokepoint: 1/3 Global Supply, No Substitute, No Air Shipment",
      "content": "Qatar produces ~1/3 of global helium supply as a byproduct of natural gas extraction. Helium cannot be safely shipped by air — it must move by sea. With Hormuz blockaded, global helium supply is effectively choked. Helium has the lowest boiling point of any element and extremely high thermal conductivity — irreplaceable as a coolant. Critical uses: MRI machine superconducting magnets, semiconductor chip fabrication (carrier gas in chemical vapor deposition), clean room purge gas. There is no synthetic substitute.",
      "dataSource": "Patrick Boyle (Apr 26, 2026)",
      "leadTime": "Weeks — hospital and chip fab reserves are limited",
      "reliability": "High — helium supply concentration and physics are facts",
      "historicalPrecedents": [
        "2012-2013 helium shortage caused MRI machine shutdowns"
      ],
      "compoundsWith": [
        "energy-shock"
      ],
      "tags": [
        "helium",
        "qatar",
        "mri-machines",
        "semiconductor-fab",
        "noble-gas",
        "no-substitute"
      ]
    },
    {
      "id": 148,
      "sourceId": "cm-145",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Panama Canal Congestion: Grain Pushed to Back of Queue, 40-Day Waits",
      "content": "With ships rerouting around Cape of Good Hope to avoid Hormuz, Panama Canal congestion has intensified. Oil tankers outbid bulk grain carriers for scarce transit slots (canal already restricted by historic drought). Wait times stretched to ~40 days. Grain shipping rates increased 50-60%. Lower-value cargoes (grain, bulk commodities) pushed to back of queue as oil tankers pay millions to skip ahead. This creates a food-specific bottleneck that compounds the fertilizer/diesel crisis.",
      "dataSource": "Patrick Boyle citing FT (Apr 2026)",
      "leadTime": "Immediate — already causing grain delivery delays",
      "reliability": "High — canal transit data is tracked precisely",
      "historicalPrecedents": [
        "2023-2024 Panama Canal drought restrictions"
      ],
      "compoundsWith": [
        "food-energy-nexus",
        "energy-shock"
      ],
      "tags": [
        "panama-canal",
        "grain-shipping",
        "transit-queue",
        "cape-rerouting",
        "shipping-capacity"
      ]
    },
    {
      "id": 149,
      "sourceId": "cm-146",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "food-energy-nexus",
      "title": "Anhydrous Ammonia Price Spike: $800 to $1,050/ton, 70% of Farmers Can't Afford Full Fertilizer",
      "content": "Anhydrous ammonia (primary nitrogen fertilizer) has spiked from ~$800/ton pre-conflict to $1,050/ton. American Farm Bureau Federation survey: ~70% of farmers report being unable to afford all fertilizer needed for this crop cycle. Additionally, sulfur (4th major plant nutrient after N-P-K) is being diverted to higher-value industrial uses like copper smelting, leaving fertilizer producers waiting. This is a compound input crisis: nitrogen expensive + sulfur diverted + diesel expensive = triple squeeze on farming margins.",
      "dataSource": "Patrick Boyle citing AFBF survey, Louis Dreyfus, FT (Apr 2026)",
      "leadTime": "This growing season — crop decisions being made now",
      "reliability": "High — AFBF survey is authoritative on US farm economics",
      "historicalPrecedents": [
        "2008 fertilizer price spike preceded food riots in 30+ countries",
        "2022 post-Ukraine fertilizer shortage"
      ],
      "compoundsWith": [
        "food-energy-nexus",
        "energy-shock"
      ],
      "tags": [
        "ammonia-price",
        "fertilizer-affordability",
        "sulfur-diversion",
        "afbf-survey",
        "crop-cycle-2026",
        "triple-input-squeeze"
      ]
    },
    {
      "id": 150,
      "sourceId": "cm-147",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "geopolitical",
      "title": "Russia Cutting Kazakh Oil Pipeline to Berlin's PCK Refinery During Hormuz Crisis",
      "content": "Russia announced suspension of Kazakh oil flow through the Soviet-era Druzhba pipeline that supplies the PCK Schwedt refinery — the facility providing 90% of Berlin's petrol, kerosene, and heating fuel. Timed precisely to maximize European pain during the Hormuz closure. By cutting this alternative supply line while seaborne Middle Eastern imports are choked off, Russia ensures Europe feels maximum pressure from the conflict.",
      "dataSource": "Patrick Boyle (Apr 26, 2026)",
      "leadTime": "Immediate — pipeline flow cut",
      "reliability": "High — pipeline operations are verifiable",
      "historicalPrecedents": [
        "2022 Russia gas supply cuts to Europe via Nord Stream"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical"
      ],
      "tags": [
        "russia",
        "druzhba-pipeline",
        "pck-schwedt",
        "berlin-fuel",
        "kazakh-oil",
        "dual-squeeze"
      ]
    },
    {
      "id": 151,
      "sourceId": "cm-148",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "asset-bubble",
      "title": "Physical vs Financial Market Disconnect: Equity Markets Pricing Peace While Physical Markets Price War",
      "content": "S&P 500 hit record highs — equity traders treating Hormuz closure as a buying opportunity. Meanwhile physical commodity traders see 1.5B barrel cumulative loss and 2030 recovery timeline. The disconnect has structural causes: (1) Post-2008 conditioning — every correction was rescued, creating 'buy the dip' muscle memory, (2) ~60-70% of equity volume is passive/index — responds to payroll deposits not geopolitics, (3) Career risk asymmetry — fund managers fired for being bearish and early, never fired for being bullish and wrong ('nobody saw it coming'), (4) TACO thesis — Trump Always Chickens Out, investors assume tariff-style reversal. Fatal flaw: trade wars fought with administrative ink (cancellable), shooting wars fought with drones and missiles (not cancellable). 'It takes two to taco.' Krugman observation: 1978 P/E ratios at historic lows (room to absorb shock), today at historic highs with private credit structures that didn't exist — less oil-dependent economy but more financially fragile.",
      "dataSource": "Patrick Boyle (Apr 26, 2026), Paul Krugman, Daniel Yergin (OddLots podcast)",
      "leadTime": "Unknown — disconnect resolves suddenly when it resolves",
      "reliability": "High — the data points are verifiable; the timing of resolution is unknowable",
      "historicalPrecedents": [
        "2007-2008: equity markets ignored subprime signals until Bear Stearns",
        "1973: markets initially shrugged off embargo",
        "Every bubble: physical reality eventually wins"
      ],
      "compoundsWith": [
        "asset-bubble",
        "energy-shock",
        "credit-debt"
      ],
      "tags": [
        "physical-financial-disconnect",
        "passive-flows",
        "taco-thesis",
        "career-risk",
        "pavlovian-conditioning",
        "krugman",
        "yergin",
        "cape-ratio",
        "private-credit"
      ]
    },
    {
      "id": 152,
      "sourceId": "cm-149",
      "section": "leading-indicators",
      "category": "indicator",
      "subcategory": "geopolitical",
      "title": "Dual Blockade: Iran Restricting Hostile Vessels, US Navy Running Counter-Blockade",
      "content": "The Strait of Hormuz is now under dual blockade. Iran restricts passage to 'hostile vessels' from unfriendly countries. US Navy began counter-blockade April 13, targeting ships bound for/departing Iranian ports. The Navy is no longer a neutral guardian of open sea lanes — it's an active combatant running its own blockade. Merchant ships are on their own: turning off GPS transponders, hiding behind Omani flags, navigating Iranian crypto toll demands. Iran seizing MSC container ships. Trump ordered Navy to 'shoot and kill any boat caught laying mines.' The unwritten assumption that the US Navy guarantees free trade is now obsolete.",
      "dataSource": "Patrick Boyle (Apr 26, 2026)",
      "leadTime": "Immediate — dual blockade active",
      "reliability": "High — naval operations and ship seizures are reported in real time",
      "historicalPrecedents": [
        "1980s Tanker War (Iran-Iraq)",
        "2019 Hormuz tanker seizures (limited)"
      ],
      "compoundsWith": [
        "energy-shock",
        "geopolitical"
      ],
      "tags": [
        "dual-blockade",
        "us-navy",
        "irgc",
        "transponder-dark",
        "crypto-tolls",
        "ship-seizure",
        "free-trade-assumption"
      ]
    },
    {
      "id": 153,
      "sourceId": "cm-150",
      "section": "compound-interactions",
      "category": "compound-interaction",
      "subcategory": "geopolitical",
      "title": "The Interdependence Illusion: Economic Integration Failed to Prevent Conflict",
      "content": "Financial markets spent 30 years operating under 'the great illusion' — that economic interdependence naturally prevents conflict. If Hormuz is essential for global survival, no rational actor would close it. This assumption failed from both ends: (1) Iran discovered holding the global economy hostage with cheap drones is effective leverage — unlike nuclear weapons, it's leverage they can actually USE, (2) major economies spent the last two years dismantling trade interdependence through tariffs, export controls, onshoring — removing the very web of relationships that was supposed to deter conflict. Oil intensity of GDP down 70% since 1970s (less physical vulnerability) but the financial system has much lower margin of safety for prolonged inflation shocks.",
      "dataSource": "Patrick Boyle (Apr 26, 2026), Daniel Yergin",
      "leadTime": "Structural — permanent shift in geopolitical risk framework",
      "reliability": "High — the thesis is supported by current events",
      "historicalPrecedents": [
        "Pre-WWI: European economic integration was supposed to prevent war",
        "2022: Russia gas weapon against Europe despite interdependence"
      ],
      "compoundsWith": [
        "geopolitical",
        "energy-shock",
        "asset-bubble"
      ],
      "tags": [
        "interdependence-illusion",
        "great-illusion",
        "drone-leverage",
        "deglobalization",
        "energy-sovereignty",
        "green-rebranding"
      ]
    },
    {
      "id": 154,
      "sourceId": "cm-151",
      "section": "frameworks",
      "category": "asset-bubble",
      "subcategory": "ai-specific",
      "title": "Lab Capex Recoupment: OpenAI/Anthropic Cannot Cover Compute Commitments From Operating Revenue Within Decade",
      "content": "**VERIFICATION AUDIT (May 6 2026):** Many specific dollar figures in this entry circulated in AI commentary but are not all primary-verified. Framework and directional claims are well-supported; specific numbers should be re-verified before public quoting. VERIFIED elements flagged below.\\n\\nOpenAI and Anthropic face structural recoupment math that cannot work on operating cash flow alone. VERIFIED: OpenAI ARR ~$25B Q1 2026 (Remio.ai citing OpenAI February disclosure). VERIFIED: Anthropic ARR $30B gross / ~$22B net per OpenAI dispute (Remio.ai April 2026). VERIFIED: Claude Code alone $2.5B ARR (The Neuron March 2026). VERIFIED: Anthropic February 2026 round at $380B (The Neuron). UNVERIFIED widely-cited figures: OpenAI compute commitments aggregate $600B-$1.15T, per-cloud breakdowns ($250B Azure / $300B Oracle / $38B AWS / $300B AMD), $14-17B 2026 OpenAI losses, $115-218B cumulative burn projection, profitability timeline 2030, $6B Anthropic 2026 burn, gross-margin trajectory 38%→70%, ~96% MSFT-OpenAI passthrough percentage, $1.30/hr vs $3.70 GPU subsidy rates, OpenAI $730-852B post-money valuation, Anthropic $50B/$900B May 2026 raise. CRITICAL FRAMING DISTINCTION: lab equity is being option-priced (binary outcome on transformative AI) not DCF-priced. Public bear analysts (Zitron 'subprime AI crisis,' Doctorow GPU-depreciation accounting fraud thesis, Bernstein/Rasgon 'crash the global economy for a decade,' Marks Dec 2025 memo question on 30-year debt repayment) are doing correct arithmetic on the wrong question for valuation methodology. Sophisticated capital extracts via cloud-revenue passthrough (directional pattern verified, specific 96% figure unverified), vendor financing (Nvidia/AMD), secondary tenders, and IPO at premium — by the time bears are 'right,' sophisticated holders have exited. Bagholder distribution differs from 1999 dotcom because losses won't concentrate in publicly-held equity until after IPO. THE REAL SYSTEMIC RISK is not lab failure — it's coordination failure where one funding source blinks (sovereign withdrawal, hyperscaler CFO halt of subsidies, IPO market closure due to broader recession) triggering cascade.",
      "dataSource": "Gemini deep research May 2026 + Mark Cuban interview Big Technology Podcast May 2026 + Apollo/Bernstein/Pantheon institutional analysis + Ed Zitron + Cory Doctorow critiques",
      "leadTime": "2-3 years to forced inflection (vs 5-10 years for hyperscalers)",
      "reliability": "High — financial commitments and ARR ranges well-documented; option-pricing reframe is analytical synthesis from sophisticated-capital behavior pattern",
      "historicalPrecedents": [
        "1999 dotcom bubble (telecom fiber overbuild, Lucent/Worldcom collapse) — but bagholder distribution was different",
        "Biotech option pricing pattern — most clinical trials fail, winners pay for everything",
        "Manhattan Project framing — strategic-positioning capital not asking ROI questions"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "credit-debt",
        "geopolitical",
        "currency-inflation"
      ],
      "tags": [
        "lab-capex",
        "openai",
        "anthropic",
        "option-pricing",
        "circular-financing",
        "msft-passthrough",
        "bagholder-distribution",
        "coordination-failure",
        "cuban-they-never-get-it",
        "zitron-56pct",
        "stargate"
      ]
    },
    {
      "id": 155,
      "sourceId": "cm-152",
      "section": "frameworks",
      "category": "asset-bubble",
      "subcategory": "ai-specific",
      "title": "AI Capex As GDP Prop: 50-80% of Q1 2026 US GDP Growth Depends on Hyperscaler Deficit-Spending",
      "content": "Hyperscaler AI capex is no longer a sector story — it's a macroeconomic dependency. VERIFIED: Pantheon Macro / Oliver Allen: AI capex drove ~50% of Q1 2026 US GDP growth (Angel Investors Network May 2026; some commentary cites higher figures up to 80% but the 50% number is the only Pantheon-attributed figure that has been primary-confirmed). VERIFIED: Apollo's Torsten Slok: 2026 hyperscaler capex ~$646B = ~2% of US GDP, offsetting tariff drag while corporate capex outside AI is at essentially zero growth. VERIFIED: Jason Furman (Harvard, former CEA Chair): H1 2025 information processing equipment + software (just 4% of the economy) drove 92% of US GDP growth. VERIFIED: Hyperscaler capital intensity 45-57% of revenue (CreditSights, May 2026) — vs 10-20% historical tech average. VERIFIED: Meta dropped ~6% (some reports up to -8.65%) after raising 2026 capex guidance to $145B in April 2026 — first major sign of investor pushback on AI spending. UNVERIFIED at primary level (drop or qualify before public use): Amazon FCF-negative 2026 projection, Meta FCF -89% projection, '60% of operating cash flow to capex' specific figure, Bernstein/Rasgon 'crash the global economy' verbatim quote. The economy is being held up by a small number of companies' willingness to deficit-spend on infrastructure. CORRELATED RISK: if capex decelerates for ANY reason (obsolescence threat from photonic, FCF discipline, recession, regulatory pressure, lab funding crisis), GDP support evaporates simultaneously with the AI sector pullback. Standard 'AI sells off but broader economy is fine' framing is FALSE at current capex concentration — they are the same trade.",
      "dataSource": "Apollo (Torsten Slok), Pantheon Macroeconomics, Jason Furman (Harvard, former CEA), Bernstein (Stacy Rasgon), Q1 2026 hyperscaler earnings calls",
      "leadTime": "Capex deceleration would lag 2-4 quarters from underlying trigger; macro impact lags another 1-2 quarters",
      "reliability": "Medium-High — Pantheon/Furman attribution methodology is contested but directional finding is consensus across multiple institutional analysts",
      "historicalPrecedents": [
        "2000 dotcom: telecom fiber overbuild was ~1% of GDP; current AI buildout is 2x that proportion",
        "1980s S&L buildout-then-bust pattern — concentrated capital flow that masked underlying weakness",
        "Japan 1989 — when capex flow stopped, broader economy fell with it"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "yield-curve",
        "currency-inflation",
        "credit-debt"
      ],
      "tags": [
        "capex-as-gdp-prop",
        "macro-fragility",
        "concentration-risk",
        "apollo-slok",
        "furman",
        "pantheon",
        "rasgon",
        "meta-pushback",
        "correlated-trade"
      ]
    },
    {
      "id": 156,
      "sourceId": "cm-153",
      "section": "frameworks",
      "category": "asset-bubble",
      "subcategory": "ai-specific",
      "title": "Hyperscaler FCF Deterioration: Capital Intensity 45-57% Of Revenue, 2-3x Historical Tech Average",
      "content": "Hyperscaler financial structure is shifting from cash-generation machines to debt-funded capex vehicles. VERIFIED: Capital intensity (capex / revenue) for the top hyperscalers is currently 45-57% of revenue (CreditSights, May 2026) — vs historical tech average of 10-20%. VERIFIED: Per-hyperscaler 2026 capex estimates — Amazon ~$200B (AWS/Anthropic + Trainium), Microsoft ~$190B, Alphabet $180-190B, Meta $125-145B, Oracle ~$50B (multiple sources May 2026). VERIFIED: Meta dropped ~6% after raising 2026 capex guidance to $145B without revenue proof in April 2026 — first major sign of investor discipline asserting itself. UNVERIFIED at primary-source level (qualify before public use): Amazon projected FCF-negative 2026 (no primary financial filing source found), Meta FCF projected -89% YoY (no primary source found), '60% of operating cash flow to capex' specific aggregate. The directional claim — hyperscaler FCF margin of safety is thinning — is well-supported by capital-intensity data and investor reactions; specific FCF projections need primary financial-filing or analyst-report citations before being quoted publicly. THE MARGIN OF SAFETY IS THINNING: these companies traditionally had massive FCF that absorbed economic shocks. With capital intensity at 45-57% of revenue, ability to weather recession or drawdown without cutting capex is reduced. If recession hits in 2027-2028 with capex still at these levels, hyperscalers may be forced to choose between dividend cuts, debt-funded buybacks, or capex deceleration. The first two trigger equity selloffs; the third triggers broader economic shock per cm-152. RECONCILIATION WITH OPTIMISM: hyperscalers DO have real businesses generating real revenue (Microsoft 365, AWS, Google Search, Meta ads). The capex spike is partially funded by these underlying businesses' cash flows — they're not pure speculation. But the buffer is thinner than it looks.",
      "dataSource": "Q1 2026 hyperscaler earnings, Apollo Global Management analysis (Torsten Slok), Bernstein Research, post-earnings stock reactions",
      "leadTime": "FCF deterioration is current; capex-cut decisions would lag 2-4 quarters from any meaningful trigger",
      "reliability": "High — financial data is from public earnings reports",
      "historicalPrecedents": [
        "1999-2000 telecom: capital intensity reached 50%+ at peak; 60-90% drawdowns followed",
        "2008 banks: capital intensity wasn't the issue, but liquidity was — analogy holds for 'thin margin of safety'",
        "1980s real estate REITs: deteriorating FCF + debt funding preceded major writedowns"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "credit-debt",
        "yield-curve"
      ],
      "tags": [
        "hyperscaler-fcf",
        "capital-intensity",
        "amazon-fcf-negative",
        "meta-fcf-89pct-drop",
        "debt-funded-capex",
        "investor-pushback",
        "margin-of-safety"
      ]
    },
    {
      "id": 157,
      "sourceId": "cm-154",
      "section": "frameworks",
      "category": "asset-bubble",
      "subcategory": "ai-specific",
      "title": "Infrastructure Obsolescence Trigger: Photonic + Next-Gen Compute Creates Capex-Decision Risk Before Deployment Timing Matters",
      "content": "The dominant near-term bear case for hyperscaler stocks is NOT 'AI fails.' It's 'the ~$646B/yr we're spending (VERIFIED: Apollo/Slok 2026 estimate) is buying GPU-vintage infrastructure that won't be globally competitive in 2028.' Optical interconnects (chip-to-chip light) are shipping now. Photonic inference accelerators are 2-4 years from meaningful deployment (Lightmatter, Ayar Labs, Celestial AI funded). Pure photonic compute that fully replaces GPUs is 5-7 years out. THE LOAD-BEARING POINT: the bear case doesn't require photonic to actually displace GPUs. It requires only that the THREAT of obsolescence becomes credible enough at moment of capex decision to make hyperscaler boards pause new deployments. CFO logic in 2027: 'do we approve the next major buildout when the next-vintage architecture will be substantially more efficient?' The capex deceleration doesn't need an AI capability failure to trigger — it needs only the WHIFF of architecture obsolescence at decision time. VERIFIED financial constraint: hyperscaler capital intensity 45-57% of revenue (CreditSights) builds the financial pressure that makes that pause more likely. VERIFIED first investor-discipline signal: Meta -6% on raised 2026 capex guidance (April 2026). Cory Doctorow has alleged that the standard 5-year GPU depreciation schedule constitutes accounting fraud, given that GPUs burn out in 2-3 years and intensive training runs can damage them severely — if Doctorow's accounting framework is adopted by SEC enforcement or institutional investors, it would force a wave of writedowns even without photonic emergence. VERIFIED supporting voice: Howard Marks (Oaktree Dec 2025 'Is It a Bubble?' memo) explicitly raises this question: 'Will the investments funded with debt — in chips and data centers — maintain their level of productivity long enough for these 30-year obligations to be repaid?' THE TRIGGER MECHANISM is financial/timing-based at moment of capital allocation decision, NOT capability-based.",
      "dataSource": "Photonic-computing research note from AI-Stocks meta-opinions (May 2026 update), Lightmatter/Ayar Labs/Celestial AI funding rounds, Cory Doctorow GPU depreciation critique, hyperscaler 10-K depreciation schedules",
      "leadTime": "12-24 months from now to first meaningful CFO pause decisions; 24-48 months to material capex deceleration",
      "reliability": "Medium — photonic deployment timing well-tracked, but CFO behavioral response is inferred from precedent rather than observed",
      "historicalPrecedents": [
        "1999 telecom: WDM optical capacity arrived simultaneously with fiber overbuild collapse; stranded capex pattern",
        "Mainframe era: companies that delayed mainframe upgrades for the next-gen architecture won the next decade",
        "Intel-vs-AMD architecture transitions: pause-then-leap pattern at every major shift"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "credit-debt",
        "currency-inflation"
      ],
      "tags": [
        "infrastructure-obsolescence",
        "photonic-threat",
        "cfo-pause-logic",
        "stranded-capex",
        "doctorow-gpu-depreciation",
        "lightmatter",
        "ayar-labs",
        "celestial-ai",
        "capex-decision-timing"
      ]
    },
    {
      "id": 158,
      "sourceId": "cm-155",
      "section": "frameworks",
      "category": "asset-bubble",
      "subcategory": "demographic-flow",
      "title": "Passive Flow Lock-In + Boomer Decumulation Cliff: 2027-2030 Demographic Flip Removes Structural Bid Under Mega-Caps",
      "content": "Vanguard, BlackRock, and State Street collectively hold 15-25% of every major US company including all hyperscalers, Nvidia, semiconductor equipment makers, and AI-themed ETFs. Their flagship index funds are contractually obligated to hold by index weight — they cannot underweight Nvidia for being expensive, cannot reduce hyperscaler exposure for being bubbly. This is structural, not conspiracy. Effects: (1) the bedrock of passive ownership cannot exit at scale without crashing the holdings, (2) the Fed put is implicit at this concentration — disorderly drawdown becomes systemic risk and policy intervention reflexive, (3) retail and active managers can sell, but absent a real liquidity event, prices float on passive flows. THE CRITICAL CAVEAT: passive flows are net positive ONLY as long as 401(k)/IRA contributions exceed retirement decumulation. The boomer decumulation cliff (estimated 2027-2030) flips passive funds to net seller. When the buying force that has absorbed sellers for 40 years reverses direction, the structural floor under concentrated mega-cap holdings (hyperscalers, Nvidia, the Mag 7) erodes — not because anyone decides to sell, but because the mechanism that has held valuations elevated reverses. This is INDEPENDENT of AI capability or capex; it's a demographic mechanism that layers on top of any other bear case. Combined with the infrastructure-obsolescence (cm-154) and lab-capex-recoupment (cm-151) timing windows (also late-2020s), 2027-2030 is the high-risk convergence window. Mitigants: target-date fund glide paths may smooth this somewhat (auto-rotation to bonds rather than full liquidation); millennial/Gen Z 401(k) growth offsets some boomer outflow; foreign capital flows could partially substitute. Magnitude estimates vary widely — direction is durable, velocity is uncertain.",
      "dataSource": "Federal Reserve Z.1 financial accounts, Vanguard/BlackRock annual flow data, ICI retirement statistics, US Census Bureau demographic projections, AI-Stocks meta-opinions (passive-flow-lock-in opinion + boomer-decumulation-cliff research note)",
      "leadTime": "2-4 years to inflection point; 1-2 years for early signs in flow data",
      "reliability": "High on direction (demographic math is mechanical); Medium on timing and magnitude (depends on TDF glide paths, Gen Z contribution rates, foreign substitution)",
      "historicalPrecedents": [
        "Japan 1990s post-bubble — demographic shifts coincided with multi-decade equity stagnation",
        "1968-1982 stagflation — boomer entry into workforce was insufficient to support previous valuation regime",
        "China demographic peak 2020s — beginning to manifest as equity-market headwind"
      ],
      "compoundsWith": [
        "asset-bubble",
        "ai-specific",
        "currency-inflation",
        "yield-curve",
        "systemic-banking"
      ],
      "tags": [
        "passive-flow-lock-in",
        "boomer-decumulation",
        "demographic-cliff",
        "vanguard-blackrock-statestreet",
        "401k-flip",
        "structural-bid",
        "2027-2030-window",
        "convergence-risk"
      ]
    },
    {
      "id": 159,
      "sourceId": "cm-156",
      "section": "frameworks",
      "category": "other",
      "subcategory": "general",
      "title": "Wikipedia: Tariffs in the Second Trump Administration",
      "content": "Wikipedia reference article cataloging the tariff policy actions of the second Trump administration — chronology of tariff orders, country-by-country impositions and retaliations, legal challenges, court rulings, and economic effects. Useful as a structured reference timeline for trade-policy events feeding into trade-war/recession-risk scenario modeling. Index entry only — not deep extraction. Re-index of file previously marked DONE in INTAKE-PLAN; file remained in /intake/.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "trade-policy",
        "tariffs",
        "trump-administration",
        "wikipedia",
        "reference-source",
        "geopolitical",
        "trade-war"
      ]
    },
    {
      "id": 160,
      "sourceId": "cm-157",
      "section": "frameworks",
      "category": "other",
      "subcategory": "general",
      "title": "Executive Summary: U.S. Trade & Tariff Crisis (Late 2025)",
      "content": "Executive summary of late-2025 tariff escalation: 100% China threat, 50% EU threat, $19B crypto crash on tariff news, Q1 2025 GDP contraction (-0.3%), GM $5B loss, Stellantis 4,500 layoffs, Porsche €1.1B profit hit, LA port -35% imports, China-US container bookings -50%, bond market stress, Moody's US downgrade, accelerating de-dollarization, federal court ruling that Trump exceeded tariff authority. Useful trigger-event timeline for trade-policy + recession scenario backtesting.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "trade-policy",
        "tariffs",
        "recession-risk",
        "supply-chain",
        "de-dollarization",
        "bond-market",
        "moodys-downgrade",
        "auto-sector"
      ]
    },
    {
      "id": 161,
      "sourceId": "cm-158",
      "section": "frameworks",
      "category": "other",
      "subcategory": "general",
      "title": "Money Market Funds at Record Highs — Safe-Haven Flow Analysis",
      "content": "PDF article explaining the surge of money market fund (MMF) assets to record levels — typically interpreted as a safe-haven flow signal indicating investor de-risking, defensive cash positioning, and elevated systemic-stress sentiment. Relevant to crisis-monitor liquidity, capital-flight, and risk-off positioning indicators. Index entry only.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "safe-haven",
        "money-market-funds",
        "liquidity",
        "risk-off",
        "capital-flows",
        "investor-sentiment"
      ]
    },
    {
      "id": 162,
      "sourceId": "cm-159",
      "section": "frameworks",
      "category": "other",
      "subcategory": "general",
      "title": "How Safe Havens (Gold, USD, Non-Correlated Assets) Behave in Crisis",
      "content": "PDF analysis of safe-haven asset behavior — gold, US dollar, Treasuries, and non-correlated assets — and how each performs across different crisis types (liquidity vs. solvency, inflationary vs. deflationary, geopolitical shock vs. credit cycle). Useful for crisis-portfolio framing and cross-correlation modeling. Index entry only.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "safe-haven",
        "gold",
        "us-dollar",
        "non-correlated-assets",
        "crisis-portfolio",
        "asset-behavior",
        "currency-inflation"
      ]
    },
    {
      "id": 163,
      "sourceId": "cm-160",
      "section": "frameworks",
      "category": "other",
      "subcategory": "general",
      "title": "VidBrainz Export: Solvency Crisis (12 Videos)",
      "content": "VidBrainz knowledge-base export of 12 video summaries on solvency crises — Ferguson on European bank solvency, Kirkegaard on Spain/Belgium, Hertz/corporate bankruptcies (2020), Winmill on gold + solvency, Rickards on Fed solvency propped up by gold, JustDario on Japan's $1.9T JGB solvency crisis, Real Vision/G30 corporate solvency report, Chris Casey on existential US solvency threat, Tett on central banks' inability to fix solvency, Steven Van Metre on Fed emergency meeting. Theme: distinction between liquidity vs. solvency crises and the limits of central-bank intervention.",
      "dataSource": "",
      "leadTime": null,
      "reliability": "",
      "historicalPrecedents": [],
      "compoundsWith": [],
      "tags": [
        "solvency-crisis",
        "vidbrainz-export",
        "central-banks",
        "japan-jgb",
        "us-debt",
        "european-banks",
        "corporate-bankruptcy",
        "gold"
      ]
    }
  ],
  "topicIndex": {
    "minsky": [
      0
    ],
    "financial-instability": [
      0
    ],
    "ponzi-finance": [
      0
    ],
    "endogenous-crisis": [
      0
    ],
    "leverage-cycle": [
      0
    ],
    "foundational-framework": [
      0
    ],
    "dalio": [
      1,
      37
    ],
    "debt-cycle": [
      1,
      37
    ],
    "deleveraging": [
      1,
      50
    ],
    "bridgewater": [
      1,
      37
    ],
    "long-term-debt-cycle": [
      1
    ],
    "short-term-debt-cycle": [
      1
    ],
    "monetary-policy": [
      1,
      62
    ],
    "reinhart-rogoff": [
      2
    ],
    "credit-contraction": [
      2,
      28
    ],
    "debt-overhang": [
      2
    ],
    "this-time-is-different": [
      2
    ],
    "empirical-crisis-history": [
      2
    ],
    "recovery-timeline": [
      2
    ],
    "network-theory": [
      3
    ],
    "contagion": [
      3,
      29
    ],
    "robust-yet-fragile": [
      3
    ],
    "ofr-contagion-index": [
      3
    ],
    "systemic-risk": [
      3
    ],
    "fire-sale-externality": [
      3
    ],
    "common-asset-shocks": [
      3
    ],
    "input-output": [
      4,
      35
    ],
    "supply-chain": [
      4,
      20,
      35,
      160
    ],
    "production-networks": [
      4
    ],
    "cascade": [
      4,
      35
    ],
    "chokepoints": [
      4,
      35
    ],
    "leontief": [
      4
    ],
    "granular-origins": [
      4
    ],
    "ofr": [
      5
    ],
    "financial-stress-index": [
      5
    ],
    "daily": [
      5,
      9
    ],
    "33-variables": [
      5
    ],
    "5-categories": [
      5
    ],
    "real-time-monitoring": [
      5,
      30
    ],
    "stress-decomposition": [
      5
    ],
    "cleveland-fed": [
      6,
      9
    ],
    "cfsi": [
      6
    ],
    "monthly": [
      6
    ],
    "5-grade-severity": [
      6
    ],
    "11-components": [
      6
    ],
    "trend-indicator": [
      6
    ],
    "st-louis-fed": [
      7
    ],
    "stlfsi4": [
      7
    ],
    "weekly": [
      7
    ],
    "18-series": [
      7
    ],
    "pca": [
      7
    ],
    "interest-rates": [
      7
    ],
    "yield-spreads": [
      7
    ],
    "ecb": [
      8
    ],
    "ciss": [
      8
    ],
    "systemic-stress": [
      8
    ],
    "cross-correlation": [
      8,
      34
    ],
    "portfolio-theory": [
      8
    ],
    "europe": [
      8,
      126
    ],
    "interaction-effect": [
      8
    ],
    "few-false-positives": [
      8
    ],
    "sri": [
      9
    ],
    "distance-to-default": [
      9
    ],
    "banking-system": [
      9
    ],
    "merton-model": [
      9
    ],
    "bank-health": [
      9
    ],
    "federal-reserve": [
      10
    ],
    "banking-vulnerability": [
      10
    ],
    "capital-adequacy": [
      10
    ],
    "fire-sale": [
      10
    ],
    "liquidity-risk": [
      10
    ],
    "run-vulnerability": [
      10
    ],
    "quarterly": [
      10
    ],
    "supervisory-data": [
      10
    ],
    "yield-curve": [
      11,
      12,
      48,
      49,
      63,
      140
    ],
    "10y-3m": [
      11
    ],
    "recession-predictor": [
      11,
      12
    ],
    "treasury-spread": [
      11
    ],
    "most-reliable": [
      11
    ],
    "14-month-lead": [
      11
    ],
    "new-york-fed": [
      11
    ],
    "10y-2y": [
      12
    ],
    "media-watched": [
      12
    ],
    "self-fulfilling": [
      12
    ],
    "term-premium": [
      12
    ],
    "credit-to-gdp-gap": [
      13
    ],
    "bis": [
      13,
      15,
      16
    ],
    "basel-iii": [
      13
    ],
    "countercyclical-buffer": [
      13
    ],
    "long-range": [
      13
    ],
    "hp-filter": [
      13
    ],
    "macro-prudential": [
      13
    ],
    "noise-to-signal": [
      13
    ],
    "cape": [
      14,
      19
    ],
    "shiller-pe": [
      14
    ],
    "valuation": [
      14
    ],
    "long-term-returns": [
      14,
      51
    ],
    "mean-reversion": [
      14,
      40
    ],
    "market-fragility": [
      14
    ],
    "10-year-earnings": [
      14
    ],
    "mispricing-risk": [
      15
    ],
    "complacency": [
      15
    ],
    "low-volatility": [
      15
    ],
    "outperforms-credit": [
      15
    ],
    "2-3-year-lead": [
      15
    ],
    "behavioral-finance": [
      15
    ],
    "debt-service-ratio": [
      16,
      56
    ],
    "flow-measure": [
      16
    ],
    "rate-sensitivity": [
      16
    ],
    "income-coverage": [
      16
    ],
    "mortgage-repricing": [
      16
    ],
    "2-year-lead": [
      16
    ],
    "vix": [
      17
    ],
    "term-structure": [
      17,
      53
    ],
    "backwardation": [
      17,
      53
    ],
    "contango": [
      17
    ],
    "crash-signal": [
      17
    ],
    "real-time": [
      17,
      42
    ],
    "options-market": [
      17,
      53
    ],
    "short-range": [
      17
    ],
    "credit-trigger": [
      18
    ],
    "corporate-debt": [
      18,
      54
    ],
    "consumer-credit": [
      18
    ],
    "sovereign-debt": [
      18,
      58,
      128
    ],
    "spreads": [
      18
    ],
    "sloos": [
      18
    ],
    "maturity-wall": [
      18,
      69
    ],
    "private-credit": [
      18,
      50,
      73,
      151
    ],
    "asset-bubble": [
      19
    ],
    "concentration-risk": [
      19,
      47,
      57,
      74,
      95,
      155
    ],
    "margin-debt": [
      19
    ],
    "mag-7": [
      19
    ],
    "speculation": [
      19
    ],
    "leverage": [
      19,
      28,
      33,
      45,
      54
    ],
    "bubble-deflation": [
      19
    ],
    "geopolitical": [
      20,
      99,
      100,
      159
    ],
    "conflict": [
      20
    ],
    "sanctions": [
      20,
      130
    ],
    "trade-war": [
      20,
      159
    ],
    "gpr-index": [
      20
    ],
    "safe-haven": [
      20,
      88,
      90,
      161,
      162
    ],
    "rapid-onset": [
      20
    ],
    "energy-shock": [
      21,
      122
    ],
    "oil": [
      21,
      122
    ],
    "opec": [
      21
    ],
    "supply-disruption": [
      21
    ],
    "stagflation": [
      21,
      32,
      63,
      78,
      93
    ],
    "inflation-transmission": [
      21
    ],
    "spare-capacity": [
      21
    ],
    "currency-crisis": [
      22
    ],
    "inflation": [
      22,
      80
    ],
    "dxy": [
      22
    ],
    "dollar-strength": [
      22,
      90
    ],
    "de-anchoring": [
      22
    ],
    "em-vulnerability": [
      22
    ],
    "capital-flight": [
      22,
      29
    ],
    "policy-trilemma": [
      22,
      32
    ],
    "banking-crisis": [
      23
    ],
    "bank-runs": [
      23
    ],
    "cre-exposure": [
      23
    ],
    "deposit-flight": [
      23,
      61
    ],
    "duration-risk": [
      23
    ],
    "unrealized-losses": [
      23,
      61
    ],
    "discount-window": [
      23
    ],
    "too-big-to-fail": [
      23,
      28
    ],
    "ai-bubble": [
      24
    ],
    "capex-revenue-gap": [
      24
    ],
    "gpu-depreciation": [
      24
    ],
    "commoditization": [
      24,
      27
    ],
    "nvidia": [
      24
    ],
    "hyperscaler": [
      24
    ],
    "vendor-financing": [
      24,
      27
    ],
    "dot-com-parallel": [
      24,
      51,
      65
    ],
    "carry-trade": [
      25,
      29,
      94
    ],
    "yen": [
      25
    ],
    "boj": [
      25
    ],
    "hidden-leverage": [
      25
    ],
    "stairs-up-elevator-down": [
      25
    ],
    "fx": [
      25
    ],
    "unwind": [
      25
    ],
    "otc-opacity": [
      25
    ],
    "correlation-breakdown": [
      26
    ],
    "diversification-failure": [
      26,
      34
    ],
    "forced-selling": [
      26,
      33
    ],
    "var-breach": [
      26
    ],
    "risk-parity": [
      26,
      37,
      97
    ],
    "amplification": [
      26
    ],
    "cross-asset": [
      26
    ],
    "safe-haven-failure": [
      26
    ],
    "dot-com": [
      27,
      95,
      105
    ],
    "nasdaq": [
      27
    ],
    "telecom-capex": [
      27
    ],
    "amazon-97-percent-drop": [
      27
    ],
    "ai-parallel": [
      27
    ],
    "infrastructure-overbuilt": [
      27
    ],
    "gfc": [
      28
    ],
    "2008": [
      28,
      107
    ],
    "lehman": [
      28
    ],
    "subprime": [
      28,
      96,
      107
    ],
    "securitization": [
      28
    ],
    "systemic-collapse": [
      28
    ],
    "asian-crisis": [
      29
    ],
    "1997": [
      29,
      94
    ],
    "currency-peg": [
      29,
      94
    ],
    "imf": [
      29,
      128,
      134,
      138
    ],
    "common-creditor": [
      29
    ],
    "iran-2026": [
      30
    ],
    "geopolitical-shock": [
      30
    ],
    "oil-spike": [
      30
    ],
    "differentiated-impact": [
      30
    ],
    "em-vs-us": [
      30
    ],
    "compound-analysis": [
      30
    ],
    "meta": [
      31
    ],
    "75-percent-crash": [
      31
    ],
    "recovery": [
      31
    ],
    "panic-selling": [
      31
    ],
    "year-of-efficiency": [
      31
    ],
    "buying-opportunity-vs-permanent-impairment": [
      31
    ],
    "ai-correction-template": [
      31
    ],
    "compound-crisis": [
      32,
      33,
      34,
      45,
      143
    ],
    "oil-shock": [
      32,
      60,
      63,
      66
    ],
    "volcker": [
      32
    ],
    "impossible-choice": [
      32
    ],
    "services-inflation": [
      32,
      80,
      81
    ],
    "credit-freeze": [
      33
    ],
    "doom-loop": [
      33,
      58,
      76,
      94
    ],
    "margin-calls": [
      33,
      72,
      96
    ],
    "shadow-banking": [
      33,
      77,
      96
    ],
    "collateral-destruction": [
      33
    ],
    "non-linear": [
      34,
      41
    ],
    "multi-trigger": [
      34
    ],
    "ecb-ciss": [
      34
    ],
    "systemic": [
      34
    ],
    "meta-indicator": [
      34
    ],
    "semiconductor": [
      35
    ],
    "tsmc": [
      35
    ],
    "rare-earth": [
      35
    ],
    "all-or-nothing": [
      35
    ],
    "covid-validated": [
      35,
      42
    ],
    "ai-compound": [
      36
    ],
    "mag-7-concentration": [
      36
    ],
    "index-mechanical-risk": [
      36
    ],
    "simultaneous-stress": [
      36
    ],
    "sector-rotation-vs-systemic": [
      36
    ],
    "worst-case-scenario": [
      36
    ],
    "pure-alpha": [
      37
    ],
    "all-weather": [
      37
    ],
    "300-decision-rules": [
      37
    ],
    "alpha-beta-separation": [
      37
    ],
    "universa": [
      38
    ],
    "tail-risk": [
      38
    ],
    "black-swan": [
      38
    ],
    "taleb": [
      38
    ],
    "spitznagel": [
      38
    ],
    "put-options": [
      38
    ],
    "convexity": [
      38
    ],
    "vix-term-structure": [
      38
    ],
    "crash-protection": [
      38
    ],
    "aqr": [
      39
    ],
    "asness": [
      39
    ],
    "risk-decomposition": [
      39
    ],
    "alpha-vs-beta": [
      39
    ],
    "momentum-crash": [
      39
    ],
    "hidden-risk": [
      39
    ],
    "volatility-selling": [
      39
    ],
    "leverage-amplification": [
      39
    ],
    "gmo": [
      40
    ],
    "grantham": [
      40
    ],
    "valuation-reversion": [
      40
    ],
    "7-year-forecast": [
      40
    ],
    "patient-capital": [
      40
    ],
    "quality-stocks": [
      40
    ],
    "bubble-calling": [
      40
    ],
    "machine-learning": [
      41
    ],
    "ai-prediction": [
      41
    ],
    "overfitting-risk": [
      41
    ],
    "small-sample": [
      41
    ],
    "goodharts-law": [
      41
    ],
    "augmented-analysis": [
      41
    ],
    "emerging": [
      41,
      43,
      44
    ],
    "nowcasting": [
      42
    ],
    "high-frequency": [
      42
    ],
    "weekly-data": [
      42
    ],
    "gdpnow": [
      42
    ],
    "wei": [
      42
    ],
    "speed-advantage": [
      42
    ],
    "temporal-graph-learning": [
      43
    ],
    "gnn": [
      43
    ],
    "network-dynamics": [
      43
    ],
    "deep-learning": [
      43
    ],
    "default-prediction": [
      43
    ],
    "cascading-default": [
      43
    ],
    "cutting-edge": [
      43
    ],
    "alternative-data": [
      44
    ],
    "satellite": [
      44
    ],
    "sentiment-analysis": [
      44
    ],
    "nlp": [
      44
    ],
    "transaction-data": [
      44
    ],
    "google-trends": [
      44
    ],
    "ship-tracking": [
      44
    ],
    "signal-to-noise": [
      44
    ],
    "industry-vulnerability": [
      45
    ],
    "ai-rev-engine": [
      45,
      47
    ],
    "cyclical": [
      45
    ],
    "capital-intensive": [
      45
    ],
    "sector-analysis": [
      45
    ],
    "28-industries": [
      45
    ],
    "crisis-resilience": [
      46
    ],
    "defensive-sectors": [
      46
    ],
    "defense": [
      46
    ],
    "healthcare": [
      46
    ],
    "utilities": [
      46
    ],
    "consumer-staples": [
      46
    ],
    "flight-to-safety": [
      46
    ],
    "counter-cyclical": [
      46
    ],
    "cross-industry-cascade": [
      47
    ],
    "regime-dependent": [
      47
    ],
    "engine-reversal": [
      47
    ],
    "positive-sum-vs-negative-sum": [
      47
    ],
    "167-cross-effects": [
      47
    ],
    "recession-signal": [
      48,
      49,
      131
    ],
    "un-inversion": [
      48,
      64
    ],
    "FRED": [
      48
    ],
    "10Y-3M": [
      48
    ],
    "current-reading": [
      48,
      49,
      50,
      51,
      52,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      61,
      62,
      63,
      64,
      65,
      66,
      67
    ],
    "bear-steepening": [
      49
    ],
    "10Y-2Y": [
      49
    ],
    "credit-gap": [
      50
    ],
    "BIS": [
      50,
      100
    ],
    "early-warning": [
      50
    ],
    "CAPE": [
      51,
      65
    ],
    "overvaluation": [
      51
    ],
    "Shiller": [
      51
    ],
    "OFR-FSI": [
      52
    ],
    "volatility": [
      52,
      53
    ],
    "stress-index": [
      52
    ],
    "divergence": [
      52
    ],
    "pre-crisis-positioning": [
      52
    ],
    "VIX": [
      53
    ],
    "panic-signal": [
      53
    ],
    "Z1": [
      54
    ],
    "refinancing-wall": [
      54
    ],
    "interest-rate-vulnerability": [
      54
    ],
    "CRE": [
      55,
      76
    ],
    "office": [
      55
    ],
    "delinquency": [
      55
    ],
    "regional-banks": [
      55,
      76
    ],
    "multifamily": [
      55
    ],
    "vacancy-rates": [
      55
    ],
    "household-debt": [
      56
    ],
    "consumer-health": [
      56
    ],
    "mortgage-rates": [
      56,
      91
    ],
    "rate-lock-in": [
      56
    ],
    "Mag-7": [
      57,
      65,
      74
    ],
    "index-fragility": [
      57
    ],
    "tech-bubble": [
      57
    ],
    "Nifty-Fifty": [
      57
    ],
    "cap-weighted": [
      57
    ],
    "fiscal": [
      58
    ],
    "IMF": [
      58,
      100
    ],
    "debt-sustainability": [
      58
    ],
    "interest-burden": [
      58
    ],
    "AI-capex": [
      59
    ],
    "revenue-gap": [
      59
    ],
    "infrastructure-overbuild": [
      59
    ],
    "telecom-parallel": [
      59
    ],
    "Sequoia": [
      59
    ],
    "Goldman-Sachs": [
      59
    ],
    "Hormuz": [
      60,
      66
    ],
    "Iran": [
      60,
      66
    ],
    "energy-inflation": [
      60
    ],
    "Operation-Epic-Fury": [
      60
    ],
    "stagflation-risk": [
      60
    ],
    "SVB": [
      61
    ],
    "FDIC": [
      61
    ],
    "banking-fragility": [
      61
    ],
    "HTM": [
      61
    ],
    "AFS": [
      61
    ],
    "Fed-funds": [
      62
    ],
    "rate-cuts": [
      62
    ],
    "higher-for-longer": [
      62,
      67
    ],
    "policy-trap": [
      62
    ],
    "inflation-constraint": [
      62
    ],
    "Fed-trap": [
      63
    ],
    "1973-parallel": [
      63
    ],
    "1979-parallel": [
      63
    ],
    "compound-scenario": [
      63,
      64,
      65
    ],
    "VIX-backwardation": [
      64
    ],
    "recession-timing": [
      64
    ],
    "2000-parallel": [
      64
    ],
    "2007-parallel": [
      64,
      77
    ],
    "AI-bubble": [
      65
    ],
    "rotation": [
      65
    ],
    "cap-weighted-risk": [
      65
    ],
    "supply-chain-inflation": [
      66
    ],
    "rapid-change": [
      66
    ],
    "transmission-channels": [
      66
    ],
    "90-day-change": [
      66,
      67
    ],
    "rate-expectations": [
      67
    ],
    "Fed": [
      67,
      100
    ],
    "bond-market": [
      67,
      160
    ],
    "repricing": [
      67
    ],
    "financial-conditions": [
      67
    ],
    "basis-trade": [
      68
    ],
    "repo": [
      68,
      96
    ],
    "hedge-fund-leverage": [
      68
    ],
    "LTCM-parallel": [
      68,
      75
    ],
    "Treasury-market": [
      68
    ],
    "collateral": [
      68
    ],
    "refinancing": [
      69
    ],
    "junk-bonds": [
      69
    ],
    "CLOs": [
      69
    ],
    "zombie-companies": [
      69
    ],
    "downgrade-cascade": [
      69
    ],
    "shadow-insurance": [
      70
    ],
    "private-equity": [
      70
    ],
    "annuities": [
      70
    ],
    "Bermuda-reinsurance": [
      70
    ],
    "illiquid-assets": [
      70
    ],
    "regulatory-arbitrage": [
      70
    ],
    "0DTE": [
      71
    ],
    "gamma-squeeze": [
      71
    ],
    "options": [
      71
    ],
    "flash-crash": [
      71
    ],
    "Volmageddon-parallel": [
      71
    ],
    "market-makers": [
      71
    ],
    "procyclical": [
      71,
      72
    ],
    "CCP": [
      72
    ],
    "clearinghouse": [
      72
    ],
    "LDI-parallel": [
      72
    ],
    "dash-for-cash": [
      72,
      97
    ],
    "derivatives": [
      72
    ],
    "PIK": [
      73
    ],
    "amend-and-pretend": [
      73
    ],
    "opaque-lending": [
      73
    ],
    "BDCs": [
      73
    ],
    "subprime-parallel": [
      73
    ],
    "floating-rate": [
      73
    ],
    "passive-investing": [
      74
    ],
    "ETF-liquidity": [
      74
    ],
    "SPY-RSP-spread": [
      74
    ],
    "1929-parallel": [
      74
    ],
    "yen-carry-trade": [
      75
    ],
    "BOJ": [
      75
    ],
    "JGB": [
      75
    ],
    "repatriation": [
      75
    ],
    "USD-JPY": [
      75
    ],
    "extend-and-pretend": [
      76
    ],
    "S&L-parallel": [
      76
    ],
    "office-collapse": [
      76
    ],
    "Trepp": [
      76
    ],
    "Eurodollar": [
      77
    ],
    "offshore-dollar": [
      77
    ],
    "collateral-squeeze": [
      77
    ],
    "FRA-OIS": [
      77
    ],
    "EM-vulnerability": [
      77
    ],
    "S&P-500": [
      78
    ],
    "YTD": [
      78
    ],
    "tech-rotation": [
      78
    ],
    "unemployment": [
      79
    ],
    "BLS": [
      79
    ],
    "labor-market": [
      79,
      85
    ],
    "CPI": [
      80
    ],
    "Core-PCE": [
      81
    ],
    "Fed-preferred": [
      81
    ],
    "re-acceleration": [
      81
    ],
    "ISM": [
      82
    ],
    "manufacturing": [
      82,
      131
    ],
    "PMI": [
      82,
      83
    ],
    "expansion": [
      82
    ],
    "ISM-Services": [
      83
    ],
    "services-boom": [
      83
    ],
    "late-cycle": [
      83
    ],
    "consumer-confidence": [
      84
    ],
    "sentiment": [
      84
    ],
    "demand-destruction": [
      84
    ],
    "jobless-claims": [
      85
    ],
    "stability": [
      85
    ],
    "BDI": [
      86
    ],
    "shipping": [
      86
    ],
    "global-trade": [
      86
    ],
    "China-PMI": [
      87
    ],
    "contraction": [
      87
    ],
    "global-demand": [
      87
    ],
    "export-weakness": [
      87
    ],
    "gold": [
      88,
      118,
      162,
      163
    ],
    "de-dollarization": [
      88,
      160
    ],
    "super-cycle": [
      88
    ],
    "Bitcoin": [
      89
    ],
    "crypto": [
      89,
      112
    ],
    "risk-off": [
      89,
      161
    ],
    "speculative-assets": [
      89
    ],
    "DXY": [
      90
    ],
    "EM-pressure": [
      90
    ],
    "housing": [
      91
    ],
    "affordability": [
      91
    ],
    "Freddie-Mac": [
      91
    ],
    "copper": [
      92
    ],
    "Dr-Copper": [
      92
    ],
    "structural-deficit": [
      92
    ],
    "AI-demand": [
      92
    ],
    "energy-transition": [
      92
    ],
    "1973": [
      93
    ],
    "oil-crisis": [
      93
    ],
    "Yom-Kippur": [
      93
    ],
    "wage-price-spiral": [
      93
    ],
    "Bretton-Woods": [
      93
    ],
    "OAPEC": [
      93
    ],
    "COLA": [
      93
    ],
    "Arthur-Burns": [
      93
    ],
    "yield-curve-inversion": [
      93
    ],
    "Asian-crisis": [
      94
    ],
    "Thailand": [
      94
    ],
    "twin-crisis": [
      94
    ],
    "EMBI": [
      94
    ],
    "BIBF": [
      94
    ],
    "Finance-One": [
      94
    ],
    "current-account": [
      94
    ],
    "NASDAQ": [
      95
    ],
    "CAPE-44": [
      95
    ],
    "equity-funding-squeeze": [
      95
    ],
    "Y2K-liquidity": [
      95
    ],
    "Barrons": [
      95
    ],
    "15-year-recovery": [
      95
    ],
    "Burning-Up": [
      95
    ],
    "rate-hike-trigger": [
      95
    ],
    "2008-GFC": [
      96
    ],
    "Lehman": [
      96
    ],
    "CDO": [
      96
    ],
    "ABX": [
      96
    ],
    "collateral-contagion": [
      96
    ],
    "LIBOR": [
      96
    ],
    "BNP-Paribas": [
      96
    ],
    "TED-spread": [
      96
    ],
    "COVID": [
      97
    ],
    "repo-crisis": [
      97
    ],
    "systematic-deleveraging": [
      97
    ],
    "fastest-recovery": [
      97
    ],
    "VIX-82": [
      97
    ],
    "Treasury-market-break": [
      97
    ],
    "BBB-cliff": [
      97
    ],
    "September-2019-repo": [
      97
    ],
    "UK-Gilt": [
      98
    ],
    "LDI": [
      98
    ],
    "Truss": [
      98
    ],
    "margin-doom-loop": [
      98
    ],
    "pension-crisis": [
      98
    ],
    "MOVE-index": [
      98
    ],
    "48-hours": [
      98
    ],
    "BoE-emergency": [
      98
    ],
    "Kwarteng": [
      98
    ],
    "mini-Budget": [
      98
    ],
    "leverage-procyclicality": [
      98
    ],
    "EU": [
      99,
      100
    ],
    "digital-sovereignty": [
      99
    ],
    "euro-office": [
      99
    ],
    "SEPA-Instant": [
      99
    ],
    "Wero": [
      99
    ],
    "EPI": [
      99
    ],
    "CLOUD-Act": [
      99,
      100
    ],
    "A2A": [
      99
    ],
    "decoupling": [
      99,
      100
    ],
    "Visa": [
      99
    ],
    "Mastercard": [
      99
    ],
    "iDEAL": [
      99
    ],
    "sovereign-cloud": [
      99
    ],
    "infrastructure-replacement": [
      99
    ],
    "slow-burn-trigger": [
      99
    ],
    "dollar-hegemony": [
      100
    ],
    "data-visibility": [
      100
    ],
    "compound-risk": [
      100
    ],
    "early-warning-degradation": [
      100
    ],
    "SWIFT": [
      100
    ],
    "surveillance-capability": [
      100
    ],
    "AI-Rev": [
      100
    ],
    "financial-modeling": [
      100
    ],
    "multipolar": [
      100
    ],
    "BRICS": [
      100
    ],
    "slow-burn-compound": [
      100
    ],
    "doom-scorecard": [
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115
    ],
    "gold-bugs": [
      101
    ],
    "hyperinflation": [
      101,
      108,
      112
    ],
    "fiat-collapse": [
      101
    ],
    "1970s": [
      101
    ],
    "wrong-prediction": [
      101,
      102,
      103,
      104,
      106,
      108,
      109,
      111,
      112,
      113
    ],
    "black-monday": [
      102
    ],
    "1987": [
      102
    ],
    "depression": [
      102,
      111
    ],
    "japan-mirror": [
      103
    ],
    "lost-decade": [
      103
    ],
    "1990s": [
      103
    ],
    "y2k": [
      104
    ],
    "infrastructure": [
      104
    ],
    "2000": [
      104,
      105
    ],
    "internet-fad": [
      105
    ],
    "partial-prediction": [
      105
    ],
    "peak-oil": [
      106
    ],
    "energy": [
      106
    ],
    "2005": [
      106
    ],
    "burry": [
      107,
      117
    ],
    "correct-prediction": [
      107
    ],
    "skin-in-game": [
      107,
      108,
      109,
      110,
      113
    ],
    "qe": [
      108,
      118
    ],
    "schiff": [
      108
    ],
    "2009": [
      108
    ],
    "grexit": [
      109
    ],
    "eurozone": [
      109
    ],
    "roubini": [
      109,
      111,
      120
    ],
    "2010": [
      109
    ],
    "china": [
      110
    ],
    "hard-landing": [
      110
    ],
    "bass": [
      110
    ],
    "soros": [
      110
    ],
    "2015": [
      110
    ],
    "wrong-timing": [
      110,
      114
    ],
    "covid": [
      111
    ],
    "dent": [
      111
    ],
    "2020": [
      111
    ],
    "post-covid": [
      112
    ],
    "2021": [
      112
    ],
    "recession": [
      113
    ],
    "soft-landing": [
      113
    ],
    "bloomberg": [
      113
    ],
    "dimon": [
      113
    ],
    "2023": [
      113,
      114
    ],
    "ai-doom": [
      114
    ],
    "job-displacement": [
      114
    ],
    "extinction": [
      114
    ],
    "yudkowsky": [
      114
    ],
    "meta-analysis": [
      115
    ],
    "loss-aversion": [
      115
    ],
    "permabears": [
      115
    ],
    "burry-problem": [
      115
    ],
    "doomsaying-economics": [
      115
    ],
    "guru-trap": [
      116,
      117,
      118,
      119,
      120
    ],
    "meta-framework": [
      116
    ],
    "prediction-failure": [
      116
    ],
    "cognitive-bias": [
      116
    ],
    "audience-capture": [
      116,
      120
    ],
    "hammer-nail": [
      116
    ],
    "timing-problem": [
      116
    ],
    "index-fund-bubble": [
      117
    ],
    "hammer-nail-bias": [
      117
    ],
    "case-study": [
      117,
      118,
      119,
      120
    ],
    "paulson": [
      118
    ],
    "monetary-plumbing": [
      118
    ],
    "whitney": [
      119
    ],
    "municipal-bonds": [
      119
    ],
    "institutional-willpower": [
      119
    ],
    "dr-doom": [
      120
    ],
    "permabear": [
      120
    ],
    "food-energy": [
      121
    ],
    "haber-bosch": [
      121,
      140,
      141
    ],
    "nitrogen": [
      121,
      136,
      140
    ],
    "fertilizer": [
      121,
      125
    ],
    "diesel": [
      121,
      122,
      136,
      137,
      143
    ],
    "food-prices": [
      121,
      124
    ],
    "hormuz": [
      122,
      123,
      136,
      137,
      139,
      141
    ],
    "lng": [
      122,
      137
    ],
    "fertilizer-prices": [
      122
    ],
    "scenario": [
      123
    ],
    "food-crisis": [
      123,
      142
    ],
    "famine-risk": [
      123,
      134
    ],
    "grain-reserves": [
      123
    ],
    "fao": [
      124
    ],
    "political-instability": [
      124
    ],
    "arab-spring": [
      124
    ],
    "threshold": [
      124
    ],
    "concentration": [
      125
    ],
    "chokepoint": [
      125,
      141
    ],
    "green-ammonia": [
      125
    ],
    "food-security": [
      126,
      136,
      141
    ],
    "regional": [
      126
    ],
    "vulnerability": [
      126
    ],
    "africa": [
      126
    ],
    "asia": [
      126
    ],
    "global-south": [
      127,
      128,
      129,
      130,
      134,
      138,
      141
    ],
    "permanent-crisis": [
      127
    ],
    "sudan": [
      127
    ],
    "yemen": [
      127
    ],
    "haiti": [
      127,
      134,
      138
    ],
    "myanmar": [
      127
    ],
    "sahel": [
      127
    ],
    "western-lens": [
      127
    ],
    "world-bank": [
      128
    ],
    "debt-trap": [
      128
    ],
    "structural-adjustment": [
      128
    ],
    "food-insecurity": [
      129
    ],
    "wfp": [
      129,
      134
    ],
    "famine": [
      129
    ],
    "migration": [
      129
    ],
    "arab-spring-pattern": [
      129
    ],
    "foreign-policy": [
      130
    ],
    "intervention": [
      130
    ],
    "feedback-loop": [
      130
    ],
    "refugees": [
      130
    ],
    "leading-indicator": [
      131
    ],
    "construction": [
      131
    ],
    "durable-goods": [
      131
    ],
    "residential": [
      131
    ],
    "lagging-indicator": [
      131
    ],
    "non-farm-payrolls": [
      131
    ],
    "epb-research": [
      131
    ],
    "plastics": [
      132,
      139
    ],
    "pesticides": [
      132,
      136,
      143
    ],
    "pharmaceuticals": [
      132,
      139
    ],
    "petrochemicals": [
      132,
      135
    ],
    "medical-supply": [
      132,
      139
    ],
    "food-packaging": [
      132,
      139
    ],
    "energy-independence": [
      133,
      137
    ],
    "crude-imports": [
      133
    ],
    "global-pricing": [
      133
    ],
    "diesel-paradox": [
      133
    ],
    "lng-exports": [
      133
    ],
    "pacific-islands": [
      134,
      138
    ],
    "caribbean": [
      134
    ],
    "fishing": [
      134
    ],
    "refinery": [
      135
    ],
    "naphtha": [
      135,
      139
    ],
    "ethylene": [
      135
    ],
    "product-tree": [
      135
    ],
    "barrel-breakdown": [
      135
    ],
    "compound-effect": [
      136
    ],
    "farm-inputs": [
      136
    ],
    "triple-crisis": [
      136
    ],
    "marginal-pricing": [
      137
    ],
    "crude-oil": [
      137
    ],
    "myth": [
      137
    ],
    "diesel-dependency": [
      138
    ],
    "small-states": [
      138
    ],
    "fuel-crisis": [
      138
    ],
    "petroleum-products": [
      139
    ],
    "diminishing-returns": [
      140
    ],
    "sri-lanka": [
      140
    ],
    "soybeans": [
      140
    ],
    "nitrogen-dependency": [
      141
    ],
    "north-korea": [
      142
    ],
    "cuba": [
      142
    ],
    "historical-precedent": [
      142
    ],
    "fertilizer-shock": [
      142
    ],
    "multiplicative": [
      143
    ],
    "total-crop-failure": [
      143
    ],
    "synergistic": [
      143
    ],
    "seaborne-buffer": [
      144
    ],
    "tanker-tracking": [
      144
    ],
    "trafigura": [
      144
    ],
    "supply-loss": [
      144
    ],
    "2030-equilibrium": [
      144
    ],
    "shale-production": [
      145
    ],
    "dallas-fed": [
      145
    ],
    "drill-refusal": [
      145
    ],
    "capital-allocation": [
      145
    ],
    "policy-whiplash": [
      145
    ],
    "jet-fuel": [
      146
    ],
    "europe-aviation": [
      146
    ],
    "refinery-capacity": [
      146
    ],
    "kepler-modeling": [
      146
    ],
    "export-ban-risk": [
      146
    ],
    "helium": [
      147
    ],
    "qatar": [
      147
    ],
    "mri-machines": [
      147
    ],
    "semiconductor-fab": [
      147
    ],
    "noble-gas": [
      147
    ],
    "no-substitute": [
      147
    ],
    "panama-canal": [
      148
    ],
    "grain-shipping": [
      148
    ],
    "transit-queue": [
      148
    ],
    "cape-rerouting": [
      148
    ],
    "shipping-capacity": [
      148
    ],
    "ammonia-price": [
      149
    ],
    "fertilizer-affordability": [
      149
    ],
    "sulfur-diversion": [
      149
    ],
    "afbf-survey": [
      149
    ],
    "crop-cycle-2026": [
      149
    ],
    "triple-input-squeeze": [
      149
    ],
    "russia": [
      150
    ],
    "druzhba-pipeline": [
      150
    ],
    "pck-schwedt": [
      150
    ],
    "berlin-fuel": [
      150
    ],
    "kazakh-oil": [
      150
    ],
    "dual-squeeze": [
      150
    ],
    "physical-financial-disconnect": [
      151
    ],
    "passive-flows": [
      151
    ],
    "taco-thesis": [
      151
    ],
    "career-risk": [
      151
    ],
    "pavlovian-conditioning": [
      151
    ],
    "krugman": [
      151
    ],
    "yergin": [
      151
    ],
    "cape-ratio": [
      151
    ],
    "dual-blockade": [
      152
    ],
    "us-navy": [
      152
    ],
    "irgc": [
      152
    ],
    "transponder-dark": [
      152
    ],
    "crypto-tolls": [
      152
    ],
    "ship-seizure": [
      152
    ],
    "free-trade-assumption": [
      152
    ],
    "interdependence-illusion": [
      153
    ],
    "great-illusion": [
      153
    ],
    "drone-leverage": [
      153
    ],
    "deglobalization": [
      153
    ],
    "energy-sovereignty": [
      153
    ],
    "green-rebranding": [
      153
    ],
    "lab-capex": [
      154
    ],
    "openai": [
      154
    ],
    "anthropic": [
      154
    ],
    "option-pricing": [
      154
    ],
    "circular-financing": [
      154
    ],
    "msft-passthrough": [
      154
    ],
    "bagholder-distribution": [
      154
    ],
    "coordination-failure": [
      154
    ],
    "cuban-they-never-get-it": [
      154
    ],
    "zitron-56pct": [
      154
    ],
    "stargate": [
      154
    ],
    "capex-as-gdp-prop": [
      155
    ],
    "macro-fragility": [
      155
    ],
    "apollo-slok": [
      155
    ],
    "furman": [
      155
    ],
    "pantheon": [
      155
    ],
    "rasgon": [
      155
    ],
    "meta-pushback": [
      155
    ],
    "correlated-trade": [
      155
    ],
    "hyperscaler-fcf": [
      156
    ],
    "capital-intensity": [
      156
    ],
    "amazon-fcf-negative": [
      156
    ],
    "meta-fcf-89pct-drop": [
      156
    ],
    "debt-funded-capex": [
      156
    ],
    "investor-pushback": [
      156
    ],
    "margin-of-safety": [
      156
    ],
    "infrastructure-obsolescence": [
      157
    ],
    "photonic-threat": [
      157
    ],
    "cfo-pause-logic": [
      157
    ],
    "stranded-capex": [
      157
    ],
    "doctorow-gpu-depreciation": [
      157
    ],
    "lightmatter": [
      157
    ],
    "ayar-labs": [
      157
    ],
    "celestial-ai": [
      157
    ],
    "capex-decision-timing": [
      157
    ],
    "passive-flow-lock-in": [
      158
    ],
    "boomer-decumulation": [
      158
    ],
    "demographic-cliff": [
      158
    ],
    "vanguard-blackrock-statestreet": [
      158
    ],
    "401k-flip": [
      158
    ],
    "structural-bid": [
      158
    ],
    "2027-2030-window": [
      158
    ],
    "convergence-risk": [
      158
    ],
    "trade-policy": [
      159,
      160
    ],
    "tariffs": [
      159,
      160
    ],
    "trump-administration": [
      159
    ],
    "wikipedia": [
      159
    ],
    "reference-source": [
      159
    ],
    "recession-risk": [
      160
    ],
    "moodys-downgrade": [
      160
    ],
    "auto-sector": [
      160
    ],
    "money-market-funds": [
      161
    ],
    "liquidity": [
      161
    ],
    "capital-flows": [
      161
    ],
    "investor-sentiment": [
      161
    ],
    "us-dollar": [
      162
    ],
    "non-correlated-assets": [
      162
    ],
    "crisis-portfolio": [
      162
    ],
    "asset-behavior": [
      162
    ],
    "currency-inflation": [
      162
    ],
    "solvency-crisis": [
      163
    ],
    "vidbrainz-export": [
      163
    ],
    "central-banks": [
      163
    ],
    "japan-jgb": [
      163
    ],
    "us-debt": [
      163
    ],
    "european-banks": [
      163
    ],
    "corporate-bankruptcy": [
      163
    ],
    "instability": [
      0,
      32,
      52,
      68,
      124,
      127,
      129,
      130
    ],
    "hypothesis": [
      0
    ],
    "foundational": [
      0,
      139
    ],
    "framework": [
      0,
      1,
      2,
      3,
      10,
      28,
      30,
      37,
      38,
      39,
      40,
      58,
      72,
      93,
      99,
      119,
      128,
      131,
      154,
      157
    ],
    "understanding": [
      0,
      31,
      46,
      47
    ],
    "crises": [
      0,
      1,
      2,
      4,
      8,
      10,
      11,
      13,
      16,
      18,
      21,
      22,
      23,
      26,
      28,
      37,
      41,
      42,
      45,
      46,
      47,
      50,
      53,
      94,
      96,
      97,
      127,
      130,
      136,
      163
    ],
    "periods": [
      0,
      2,
      7,
      8,
      13,
      15,
      26,
      38,
      40,
      42,
      47,
      51,
      52,
      53,
      79,
      89
    ],
    "core": [
      0,
      31,
      40,
      62,
      63,
      73,
      74,
      78,
      80,
      81,
      84,
      88,
      90,
      91,
      97
    ],
    "insight": [
      0,
      1,
      3,
      4,
      34,
      38,
      40,
      46,
      47,
      64,
      72,
      140,
      143
    ],
    "itself": [
      0,
      1,
      2,
      6,
      12,
      19,
      23,
      26,
      40,
      48,
      50,
      59,
      62,
      64,
      65,
      67,
      88,
      91,
      93,
      96,
      156
    ],
    "prolonged": [
      0,
      6,
      7,
      32,
      49,
      53,
      63,
      64,
      65,
      80,
      82,
      93,
      153
    ],
    "calm": [
      0,
      7,
      8,
      15,
      26,
      38,
      52,
      68,
      75,
      79,
      97
    ],
    "borrowers": [
      0,
      2,
      16,
      18,
      28,
      34,
      36,
      50,
      55,
      56,
      67,
      69,
      73,
      77,
      94,
      96
    ],
    "lenders": [
      0,
      2,
      12,
      33,
      36,
      73,
      76,
      77,
      94
    ],
    "regulators": [
      0,
      3,
      13,
      25,
      43,
      68,
      70,
      72,
      76,
      96
    ],
    "recent": [
      0,
      11,
      39,
      55,
      89,
      98
    ],
    "experience": [
      0,
      62,
      127
    ],
    "tells": [
      0,
      14,
      48,
      54
    ],
    "behavior": [
      0,
      18,
      89,
      92,
      94,
      145,
      162
    ],
    "natural": [
      0,
      4,
      12,
      35,
      40,
      44,
      53,
      86,
      100,
      121,
      133,
      136,
      137,
      139,
      143,
      147
    ],
    "predictable": [
      0,
      46
    ],
    "response": [
      0,
      17,
      28,
      30,
      32,
      48,
      54,
      58,
      59,
      62,
      64,
      71,
      72,
      80,
      97,
      98
    ],
    "stable": [
      0,
      1,
      6,
      13,
      16,
      29,
      40,
      43,
      46,
      61,
      73,
      79,
      85,
      97,
      99,
      127
    ],
    "environment": [
      0,
      18,
      20,
      32,
      36,
      39,
      40,
      49,
      54,
      67,
      69,
      76,
      77,
      88,
      95,
      97
    ],
    "financing": [
      0,
      24,
      27,
      36,
      58,
      62,
      69,
      93,
      138,
      154
    ],
    "positions": [
      0,
      7,
      19,
      25,
      26,
      29,
      33,
      36,
      39,
      68,
      71,
      75,
      79,
      85,
      89,
      95,
      98
    ],
    "three": [
      0,
      2,
      3,
      7,
      28,
      32,
      33,
      34,
      41,
      51,
      64,
      65,
      68,
      69,
      84,
      88,
      93,
      96,
      97,
      98,
      131,
      136,
      143
    ],
    "regimes": [
      0,
      11,
      20,
      29,
      47,
      94,
      130
    ],
    "hedge": [
      0,
      1,
      38,
      46,
      68,
      71,
      89,
      94,
      96,
      97,
      98,
      109
    ],
    "income": [
      0,
      1,
      2,
      16,
      18,
      21,
      25,
      39,
      56,
      66,
      73,
      84,
      91,
      96,
      98,
      128
    ],
    "covers": [
      0,
      146
    ],
    "principal": [
      0,
      7,
      16,
      73
    ],
    "interest": [
      0,
      1,
      7,
      10,
      16,
      22,
      25,
      28,
      29,
      31,
      37,
      39,
      51,
      54,
      56,
      58,
      61,
      63,
      67,
      68,
      69,
      71,
      72,
      73,
      75,
      76,
      77,
      81,
      82,
      83,
      90,
      95,
      98,
      131
    ],
    "payments": [
      0,
      16,
      37,
      56,
      58,
      69,
      70,
      73,
      99,
      100
    ],
    "speculative": [
      0,
      1,
      19,
      86,
      88,
      89,
      93,
      94,
      95
    ],
    "requiring": [
      0,
      13,
      15,
      44,
      62,
      81
    ],
    "rollover": [
      0,
      18,
      22,
      36
    ],
    "ponzi": [
      0,
      1,
      93
    ],
    "neither": [
      0,
      63
    ],
    "borrower": [
      0,
      5,
      73,
      76
    ],
    "depends": [
      0,
      4,
      20,
      46,
      58,
      64,
      65,
      136,
      155
    ],
    "appreciation": [
      0,
      15,
      22,
      75
    ],
    "refinance": [
      0,
      18,
      69,
      76
    ],
    "sell": [
      0,
      10,
      19,
      23,
      26,
      29,
      31,
      33,
      36,
      61,
      68,
      71,
      74,
      89,
      91,
      96,
      97,
      98,
      115,
      117,
      119,
      137,
      158
    ],
    "dynamic": [
      0,
      5,
      12,
      13,
      22,
      24,
      25,
      27,
      42,
      43,
      45,
      47,
      58,
      61,
      64,
      68,
      69,
      70,
      73,
      74,
      75,
      76,
      79,
      83,
      85,
      87,
      90,
      94
    ],
    "economies": [
      0,
      5,
      16,
      22,
      29,
      30,
      32,
      58,
      87,
      90,
      138,
      153
    ],
    "naturally": [
      0,
      153
    ],
    "dominant": [
      0,
      58,
      157
    ],
    "expansions": [
      0,
      18
    ],
    "returns": [
      0,
      6,
      11,
      14,
      19,
      20,
      26,
      38,
      39,
      40,
      43,
      45,
      46,
      47,
      51,
      68,
      70,
      73,
      75,
      93,
      116,
      140
    ],
    "validates": [
      0,
      37,
      38
    ],
    "aggressive": [
      0,
      49,
      55,
      60,
      62,
      63,
      67,
      88,
      113,
      130,
      143
    ],
    "positioning": [
      0,
      15,
      17,
      21,
      25,
      38,
      40,
      52,
      64,
      78,
      161
    ],
    "moment": [
      0,
      5,
      16,
      59,
      94,
      95,
      96,
      157
    ],
    "occurs": [
      0,
      17,
      19,
      24,
      25,
      48,
      49,
      64,
      76
    ],
    "dependent": [
      0,
      66,
      93,
      97,
      100,
      126,
      134,
      138,
      141,
      151
    ],
    "units": [
      0,
      140
    ],
    "longer": [
      0,
      11,
      16,
      17,
      28,
      48,
      49,
      53,
      56,
      67,
      69,
      73,
      76,
      77,
      79,
      81,
      86,
      91,
      93,
      95,
      100,
      152,
      155
    ],
    "sustained": [
      0,
      2,
      6,
      25,
      51,
      52,
      53,
      54,
      58,
      72,
      74,
      81,
      82,
      85,
      86,
      87,
      90,
      132
    ],
    "fueled": [
      0,
      28,
      73,
      97
    ],
    "buying": [
      0,
      4,
      25,
      31,
      33,
      39,
      54,
      71,
      74,
      136,
      151,
      157,
      158
    ],
    "suddenly": [
      0,
      18,
      25,
      26,
      58,
      77,
      85,
      96
    ],
    "face": [
      0,
      21,
      22,
      24,
      32,
      33,
      36,
      40,
      41,
      45,
      52,
      55,
      63,
      67,
      70,
      76,
      79,
      84,
      89,
      90,
      91,
      96,
      100,
      136,
      141,
      154
    ],
    "forced": [
      0,
      3,
      10,
      20,
      23,
      26,
      29,
      33,
      36,
      39,
      48,
      60,
      62,
      68,
      69,
      70,
      72,
      74,
      77,
      94,
      96,
      97,
      98,
      125,
      142,
      156
    ],
    "selling": [
      0,
      10,
      15,
      19,
      20,
      26,
      33,
      39,
      68,
      69,
      70,
      71,
      74,
      75,
      94,
      97,
      98
    ],
    "resulting": [
      0,
      27,
      32,
      63,
      75,
      84,
      91,
      95
    ],
    "self": [
      0,
      6,
      10,
      12,
      15,
      22,
      24,
      25,
      27,
      29,
      33,
      44,
      58,
      61,
      64,
      66,
      67,
      71,
      84,
      93,
      94,
      97,
      98,
      146
    ],
    "reinforcing": [
      0,
      10,
      15,
      25,
      29,
      58,
      61,
      64,
      67,
      71,
      76,
      93,
      94,
      97,
      98
    ],
    "falling": [
      0,
      28,
      33,
      36,
      37,
      71,
      94,
      96,
      98,
      104
    ],
    "trigger": [
      0,
      3,
      4,
      8,
      9,
      18,
      19,
      20,
      21,
      22,
      23,
      24,
      25,
      26,
      33,
      34,
      36,
      46,
      52,
      68,
      69,
      70,
      71,
      72,
      73,
      75,
      77,
      87,
      89,
      93,
      94,
      95,
      96,
      97,
      98,
      99,
      100,
      110,
      117,
      129,
      130,
      156,
      157,
      160
    ],
    "margin": [
      0,
      14,
      19,
      20,
      22,
      26,
      29,
      33,
      36,
      39,
      40,
      51,
      56,
      68,
      69,
      72,
      89,
      96,
      98,
      133,
      153,
      154,
      156
    ],
    "calls": [
      0,
      20,
      22,
      26,
      29,
      31,
      33,
      36,
      39,
      68,
      72,
      96,
      98,
      117,
      145,
      155
    ],
    "force": [
      0,
      10,
      11,
      56,
      71,
      75,
      79,
      85,
      86,
      98,
      109,
      157,
      158
    ],
    "drives": [
      0,
      3,
      33,
      47,
      115
    ],
    "lower": [
      0,
      1,
      2,
      11,
      13,
      14,
      19,
      20,
      21,
      29,
      33,
      37,
      38,
      40,
      62,
      66,
      67,
      76,
      98,
      148,
      153
    ],
    "transition": [
      0,
      6,
      21,
      52,
      85,
      92,
      95
    ],
    "boom": [
      0,
      2,
      13,
      50,
      55,
      59,
      74,
      83,
      102,
      103
    ],
    "faster": [
      0,
      2,
      12,
      13,
      42,
      44,
      49,
      50,
      56,
      62,
      63,
      71,
      131
    ],
    "buildup": [
      0,
      13
    ],
    "amplifies": [
      0,
      15,
      25,
      34,
      45,
      64,
      67,
      74
    ],
    "particularly": [
      0,
      5,
      7,
      8,
      15,
      16,
      18,
      22,
      26,
      32,
      39,
      43,
      44,
      45,
      46,
      47,
      52,
      55,
      58,
      60,
      69,
      75,
      77,
      81,
      87,
      89,
      90,
      91
    ],
    "relevant": [
      0,
      15,
      16,
      27,
      32,
      33,
      39,
      40,
      75,
      81,
      161
    ],
    "monitoring": [
      0,
      1,
      2,
      4,
      5,
      6,
      9,
      12,
      16,
      17,
      18,
      19,
      20,
      21,
      23,
      24,
      26,
      29,
      30,
      31,
      32,
      33,
      34,
      35,
      36,
      37,
      38,
      39,
      40,
      41,
      42,
      44,
      46,
      47,
      50,
      61,
      80,
      81,
      82,
      83,
      84,
      86,
      88,
      89,
      92,
      98,
      99,
      100
    ],
    "capex": [
      0,
      19,
      24,
      31,
      36,
      39,
      45,
      47,
      57,
      59,
      65,
      78,
      92,
      95,
      154,
      155,
      156,
      157,
      158
    ],
    "cycle": [
      0,
      1,
      14,
      15,
      16,
      18,
      19,
      24,
      27,
      37,
      48,
      50,
      53,
      54,
      59,
      61,
      62,
      63,
      64,
      65,
      67,
      70,
      79,
      81,
      82,
      83,
      88,
      91,
      92,
      93,
      95,
      97,
      98,
      112,
      116,
      149,
      162
    ],
    "companies": [
      0,
      9,
      18,
      19,
      21,
      22,
      24,
      27,
      28,
      29,
      31,
      40,
      45,
      47,
      51,
      54,
      57,
      59,
      65,
      69,
      70,
      73,
      94,
      95,
      97,
      99,
      105,
      155,
      156
    ],
    "borrow": [
      0,
      75,
      96
    ],
    "heavily": [
      0,
      42,
      64,
      70,
      80,
      96,
      118
    ],
    "build": [
      0,
      20,
      52
    ],
    "projected": [
      0,
      27,
      58,
      99,
      139,
      156
    ],
    "future": [
      0,
      11,
      28,
      43,
      45,
      67,
      84
    ],
    "revenues": [
      0,
      73,
      87,
      138
    ],
    "materialize": [
      0,
      20,
      24,
      25,
      48
    ],
    "annual": [
      0,
      24,
      139,
      158
    ],
    "actual": [
      0,
      24,
      59,
      65,
      70,
      73,
      81,
      86,
      89,
      96,
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114
    ],
    "revenue": [
      0,
      19,
      24,
      27,
      31,
      35,
      36,
      45,
      46,
      47,
      54,
      59,
      65,
      73,
      78,
      97,
      125,
      154,
      155,
      156,
      157
    ],
    "represents": [
      0,
      5,
      6,
      7,
      18,
      19,
      23,
      24,
      43,
      44,
      53,
      54,
      55,
      61,
      62,
      66,
      74,
      75,
      76,
      77,
      78,
      88
    ],
    "classic": [
      0,
      24,
      25,
      29,
      30,
      33,
      63,
      64,
      68,
      70,
      74,
      83,
      89,
      91
    ],
    "investment": [
      0,
      2,
      12,
      15,
      18,
      19,
      20,
      24,
      27,
      28,
      31,
      33,
      36,
      39,
      40,
      44,
      45,
      54,
      57,
      58,
      59,
      62,
      69,
      70,
      72,
      73,
      74,
      95,
      96,
      97,
      98,
      100
    ],
    "expectations": [
      0,
      12,
      22,
      32,
      47,
      48,
      49,
      53,
      61,
      62,
      65,
      66,
      67,
      68,
      81,
      84,
      90,
      91,
      93,
      98
    ],
    "current": [
      0,
      6,
      7,
      11,
      15,
      18,
      19,
      20,
      22,
      23,
      24,
      27,
      29,
      33,
      36,
      39,
      40,
      41,
      42,
      48,
      49,
      50,
      51,
      52,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      61,
      62,
      63,
      64,
      65,
      66,
      67,
      68,
      71,
      76,
      77,
      78,
      79,
      80,
      81,
      82,
      83,
      84,
      85,
      86,
      87,
      88,
      89,
      90,
      91,
      92,
      94,
      95,
      96,
      98,
      124,
      141,
      155
    ],
    "cash": [
      0,
      24,
      26,
      46,
      54,
      57,
      63,
      65,
      68,
      69,
      72,
      73,
      89,
      95,
      96,
      97,
      98,
      154,
      155,
      156,
      161
    ],
    "flows": [
      0,
      4,
      20,
      22,
      24,
      25,
      30,
      41,
      44,
      46,
      57,
      65,
      69,
      71,
      86,
      90,
      97,
      100,
      130,
      146,
      156,
      158
    ],
    "academic": [
      0,
      3,
      4,
      15,
      26,
      27,
      28,
      29,
      39,
      40,
      41,
      43,
      46,
      93,
      94,
      95,
      97
    ],
    "literature": [
      0,
      3,
      4,
      28,
      29,
      93,
      94,
      95,
      97,
      100
    ],
    "working": [
      0,
      2,
      15,
      23,
      41,
      43,
      47,
      50,
      94,
      98,
      100
    ],
    "papers": [
      0,
      2,
      3,
      15,
      34,
      39,
      40,
      41,
      43,
      50,
      94,
      98,
      100
    ],
    "economics": [
      0,
      41,
      43,
      92,
      113
    ],
    "archives": [
      0,
      93,
      95
    ],
    "developed": [
      1,
      5,
      14,
      16,
      22,
      32,
      37,
      41,
      51,
      98
    ],
    "decades": [
      1,
      11,
      14,
      21,
      61,
      75,
      103,
      130
    ],
    "research": [
      1,
      2,
      3,
      5,
      8,
      11,
      12,
      13,
      15,
      16,
      18,
      20,
      26,
      34,
      37,
      39,
      40,
      41,
      43,
      44,
      46,
      50,
      51,
      52,
      57,
      59,
      68,
      94,
      100,
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115,
      116,
      117,
      118,
      119,
      120,
      130,
      131,
      132,
      134,
      154,
      156,
      157,
      158
    ],
    "associates": [
      1,
      37
    ],
    "identifies": [
      1,
      37,
      39,
      41,
      46,
      116
    ],
    "cycles": [
      1,
      16,
      32,
      37,
      46,
      60,
      115
    ],
    "dynamics": [
      1,
      2,
      6,
      15,
      19,
      27,
      29,
      35,
      36,
      39,
      43,
      45,
      47,
      50,
      54,
      57,
      60,
      71,
      75,
      93,
      97
    ],
    "short": [
      1,
      10,
      11,
      12,
      14,
      15,
      17,
      24,
      29,
      38,
      39,
      46,
      48,
      49,
      53,
      64,
      67,
      68,
      89,
      93,
      94,
      96,
      97
    ],
    "runs": [
      1,
      10,
      23,
      99,
      157
    ],
    "driven": [
      1,
      5,
      16,
      17,
      22,
      26,
      27,
      31,
      36,
      40,
      46,
      49,
      53,
      54,
      60,
      61,
      62,
      63,
      66,
      69,
      70,
      72,
      74,
      76,
      77,
      78,
      81,
      82,
      84,
      88,
      90,
      92,
      93,
      97,
      98,
      103
    ],
    "central": [
      1,
      2,
      8,
      13,
      16,
      21,
      22,
      32,
      34,
      43,
      49,
      60,
      63,
      64,
      72,
      77,
      88,
      90,
      98,
      116,
      127,
      163
    ],
    "adjustments": [
      1,
      97
    ],
    "full": [
      1,
      15,
      16,
      41,
      45,
      63,
      64,
      93,
      95,
      96,
      99,
      122,
      140,
      149,
      158
    ],
    "maximum": [
      1,
      93,
      150
    ],
    "eventual": [
      1,
      76
    ],
    "currently": [
      1,
      47,
      48,
      49,
      50,
      51,
      52,
      53,
      54,
      56,
      57,
      60,
      71,
      74,
      75,
      98,
      100,
      146,
      156
    ],
    "deep": [
      1,
      21,
      38,
      43,
      49,
      53,
      63,
      64,
      74,
      100,
      115,
      120,
      132,
      134,
      154,
      159
    ],
    "arguably": [
      1,
      51,
      68,
      77
    ],
    "began": [
      1,
      28,
      29,
      48,
      49,
      60,
      64,
      93,
      94,
      95,
      96,
      97,
      117,
      144,
      152
    ],
    "distinguishes": [
      1,
      66
    ],
    "austerity": [
      1,
      109
    ],
    "spending": [
      1,
      12,
      16,
      21,
      24,
      28,
      42,
      44,
      45,
      46,
      56,
      59,
      60,
      81,
      84,
      90,
      128,
      155,
      157
    ],
    "cuts": [
      1,
      12,
      48,
      49,
      50,
      60,
      61,
      62,
      64,
      66,
      67,
      81,
      87,
      89,
      93,
      96,
      98,
      129,
      156
    ],
    "restructuring": [
      1,
      73,
      87
    ],
    "defaults": [
      1,
      3,
      18,
      21,
      22,
      43,
      69,
      70,
      76,
      94,
      96,
      119
    ],
    "money": [
      1,
      4,
      8,
      10,
      24,
      27,
      28,
      29,
      33,
      38,
      53,
      59,
      61,
      70,
      72,
      74,
      77,
      93,
      96,
      108,
      116,
      118,
      154,
      161
    ],
    "printing": [
      1,
      118
    ],
    "monetization": [
      1,
      27
    ],
    "wealth": [
      1,
      21,
      58,
      63,
      73,
      84,
      88
    ],
    "calibrated": [
      1,
      47,
      100
    ],
    "correctly": [
      1,
      8,
      116,
      120
    ],
    "typically": [
      1,
      2,
      5,
      8,
      15,
      16,
      22,
      25,
      26,
      39,
      41,
      46,
      64,
      73,
      74,
      84,
      86,
      161
    ],
    "either": [
      1,
      6,
      13,
      24,
      25,
      32,
      33,
      36,
      48,
      67,
      69,
      71,
      76,
      77,
      85,
      86,
      90,
      94,
      98,
      100
    ],
    "deflationary": [
      1,
      162
    ],
    "inflationary": [
      1,
      21,
      37,
      49,
      60,
      62,
      63,
      67,
      93,
      162
    ],
    "weimar": [
      1,
      108,
      112,
      118
    ],
    "germany": [
      1,
      58,
      96,
      99,
      118
    ],
    "zimbabwe": [
      1,
      112
    ],
    "skill": [
      1,
      39
    ],
    "policymakers": [
      1,
      32,
      42
    ],
    "balancing": [
      1,
      37
    ],
    "determines": [
      1,
      3,
      21,
      31
    ],
    "manageable": [
      1,
      16,
      31,
      32,
      33,
      55,
      58,
      60,
      71,
      98
    ],
    "catastrophic": [
      1,
      28,
      39,
      94,
      96,
      97,
      98,
      110,
      115,
      140,
      143
    ],
    "critical": [
      1,
      2,
      4,
      8,
      9,
      20,
      21,
      27,
      29,
      31,
      34,
      35,
      36,
      46,
      47,
      52,
      53,
      54,
      55,
      60,
      61,
      64,
      69,
      73,
      98,
      147,
      154,
      158
    ],
    "peaks": [
      1,
      14,
      55,
      61,
      79,
      86
    ],
    "higher": [
      1,
      10,
      14,
      18,
      21,
      25,
      29,
      30,
      32,
      33,
      34,
      40,
      51,
      53,
      54,
      56,
      58,
      60,
      61,
      66,
      67,
      69,
      70,
      72,
      75,
      81,
      84,
      88,
      90,
      91,
      92,
      93,
      96,
      137,
      143,
      149,
      155
    ],
    "last": [
      1,
      12,
      17,
      40,
      52,
      53,
      60,
      74,
      80,
      97,
      116,
      131,
      144,
      153
    ],
    "borrowing": [
      1,
      18,
      23,
      25,
      29,
      39,
      58,
      62,
      68,
      69,
      75,
      77,
      87,
      94,
      96,
      98
    ],
    "next": [
      1,
      7,
      24,
      58,
      71,
      88,
      116,
      129,
      141,
      157
    ],
    "effect": [
      1,
      8,
      30,
      32,
      47,
      54,
      56,
      59,
      67,
      77,
      91,
      100,
      136,
      143
    ],
    "pushes": [
      1,
      17,
      25,
      32,
      62,
      71
    ],
    "total": [
      1,
      3,
      10,
      34,
      55,
      60,
      75,
      101,
      108,
      131,
      143
    ],
    "further": [
      1,
      15,
      26,
      54,
      57,
      61,
      62,
      66,
      67,
      68,
      71,
      75,
      76,
      79,
      82,
      88,
      91,
      93,
      94,
      95,
      129,
      131
    ],
    "monetary": [
      1,
      11,
      16,
      21,
      22,
      25,
      32,
      37,
      50,
      70,
      75,
      80,
      81,
      83,
      87,
      88,
      90,
      93,
      95,
      98,
      109,
      111,
      118
    ],
    "stimulus": [
      1,
      28,
      48,
      50,
      58,
      63,
      87,
      97,
      98,
      110,
      111
    ],
    "becomes": [
      1,
      2,
      3,
      4,
      22,
      36,
      42,
      45,
      47,
      54,
      57,
      59,
      71,
      73,
      77,
      120,
      135,
      157,
      158
    ],
    "pushing": [
      1,
      48,
      70,
      71,
      87
    ],
    "string": [
      1,
      87
    ],
    "problem": [
      1,
      10,
      23,
      28,
      41,
      68,
      72,
      96,
      115,
      116
    ],
    "zero": [
      1,
      5,
      7,
      11,
      25,
      28,
      33,
      52,
      54,
      56,
      62,
      69,
      71,
      74,
      75,
      94,
      97,
      122,
      131,
      140,
      155
    ],
    "negative": [
      1,
      11,
      14,
      16,
      19,
      22,
      26,
      40,
      44,
      47,
      49,
      50,
      51,
      71,
      75,
      84,
      92,
      95,
      121,
      155,
      156
    ],
    "primary": [
      1,
      13,
      14,
      28,
      29,
      41,
      47,
      55,
      56,
      68,
      73,
      77,
      83,
      87,
      91,
      97,
      149,
      154,
      155,
      156
    ],
    "tool": [
      1,
      14,
      26,
      28,
      44,
      62,
      67,
      114,
      116
    ],
    "managing": [
      1,
      70,
      71,
      72
    ],
    "exhausted": [
      1,
      29,
      37,
      94,
      144
    ],
    "provides": [
      1,
      5,
      13,
      15,
      17,
      28,
      31,
      34,
      37,
      40,
      44,
      46,
      50,
      62,
      73,
      131
    ],
    "specific": [
      1,
      2,
      19,
      23,
      24,
      25,
      27,
      30,
      33,
      35,
      36,
      37,
      38,
      39,
      40,
      41,
      44,
      50,
      53,
      57,
      59,
      93,
      95,
      96,
      115,
      116,
      148,
      154,
      155,
      156
    ],
    "metrics": [
      1,
      15,
      18,
      19,
      21,
      22,
      23,
      24,
      26,
      33,
      40,
      43,
      51
    ],
    "track": [
      1,
      4,
      11,
      18,
      30,
      35,
      37,
      40,
      44,
      71,
      73,
      127,
      137
    ],
    "ratios": [
      1,
      9,
      10,
      18,
      28,
      56,
      68,
      70,
      76,
      91,
      98,
      151
    ],
    "service": [
      1,
      2,
      16,
      27,
      56,
      73,
      77,
      93,
      97,
      128,
      130
    ],
    "costs": [
      1,
      8,
      21,
      24,
      31,
      32,
      35,
      56,
      66,
      69,
      86,
      90,
      91,
      93,
      99,
      121,
      130,
      136,
      141,
      143
    ],
    "capacity": [
      1,
      4,
      16,
      18,
      21,
      35,
      50,
      76,
      86,
      93,
      95,
      115,
      125,
      133,
      146
    ],
    "room": [
      1,
      32,
      62,
      69,
      147,
      151
    ],
    "exists": [
      1,
      36
    ],
    "distribution": [
      1,
      10,
      154
    ],
    "across": [
      1,
      2,
      3,
      4,
      5,
      8,
      10,
      11,
      16,
      21,
      24,
      26,
      28,
      30,
      32,
      34,
      35,
      37,
      44,
      45,
      46,
      50,
      52,
      54,
      55,
      59,
      63,
      67,
      68,
      69,
      75,
      76,
      86,
      87,
      93,
      95,
      96,
      98,
      99,
      100,
      140,
      142,
      162
    ],
    "categories": [
      1,
      5,
      8,
      23,
      34,
      44,
      45,
      52
    ],
    "change": [
      1,
      39,
      42,
      43,
      59,
      66,
      67,
      75,
      76,
      88,
      90,
      115,
      116
    ],
    "rapid": [
      1,
      6,
      15,
      17,
      22,
      24,
      25,
      48,
      53,
      55,
      60,
      61,
      65,
      66,
      75,
      85,
      111
    ],
    "followed": [
      1,
      2,
      6,
      11,
      30,
      51,
      62,
      65,
      84,
      97,
      111
    ],
    "sudden": [
      1,
      13,
      18,
      19,
      21,
      22,
      25,
      39,
      44,
      55,
      68,
      77,
      94
    ],
    "dangerous": [
      1,
      2,
      3,
      6,
      10,
      14,
      20,
      22,
      26,
      32,
      34,
      36,
      43,
      47,
      48,
      49,
      59,
      60,
      63,
      74,
      76
    ],
    "principles": [
      1,
      30,
      38,
      41
    ],
    "navigating": [
      1,
      152
    ],
    "changing": [
      1,
      43
    ],
    "order": [
      1,
      69,
      100
    ],
    "reinhart": [
      2,
      18,
      28,
      58
    ],
    "rogoff": [
      2,
      18,
      28,
      58
    ],
    "great": [
      2,
      33,
      48,
      51,
      63,
      64,
      102,
      153
    ],
    "subsequent": [
      2,
      14,
      27,
      28,
      50,
      51,
      57,
      67,
      116
    ],
    "understand": [
      2,
      118
    ],
    "understood": [
      2,
      61,
      72
    ],
    "contractions": [
      2,
      11,
      29
    ],
    "recessions": [
      2,
      6,
      12,
      16,
      21,
      46,
      64,
      84,
      115,
      131
    ],
    "distinction": [
      2,
      39,
      97,
      154,
      163
    ],
    "matters": [
      2,
      3,
      10,
      66,
      68,
      84,
      131,
      157
    ],
    "decline": [
      2,
      9,
      10,
      24,
      27,
      28,
      33,
      36,
      37,
      42,
      49,
      51,
      61,
      67,
      73,
      74,
      75,
      78,
      86,
      89,
      93,
      96,
      97,
      98
    ],
    "output": [
      2,
      4,
      16,
      21,
      35,
      45,
      47
    ],
    "within": [
      2,
      7,
      20,
      23,
      29,
      30,
      31,
      36,
      42,
      44,
      50,
      58,
      59,
      65,
      66,
      71,
      77,
      78,
      95,
      98,
      99,
      102,
      126,
      131,
      136,
      140,
      154
    ],
    "destruction": [
      2,
      45,
      84,
      108
    ],
    "defaulting": [
      2,
      55,
      72
    ],
    "values": [
      2,
      5,
      9,
      23,
      28,
      33,
      41,
      55,
      74,
      76,
      94
    ],
    "collapsing": [
      2,
      34,
      42,
      60,
      84,
      103
    ],
    "takes": [
      2,
      9,
      109,
      137,
      151
    ],
    "capita": [
      2,
      28
    ],
    "return": [
      2,
      14,
      28,
      31,
      37,
      38,
      39,
      40,
      51,
      59,
      67,
      70,
      74,
      76,
      82,
      98,
      101,
      104
    ],
    "dataset": [
      2,
      14,
      50,
      95
    ],
    "spans": [
      2,
      69,
      76
    ],
    "making": [
      2,
      10,
      12,
      13,
      14,
      18,
      20,
      34,
      35,
      37,
      46,
      96,
      97,
      99,
      117,
      131
    ],
    "comprehensive": [
      2,
      5,
      17,
      35,
      52,
      100
    ],
    "empirical": [
      2,
      11,
      15
    ],
    "ever": [
      2,
      118
    ],
    "include": [
      2,
      4,
      18,
      19,
      21,
      22,
      23,
      35,
      44
    ],
    "banking": [
      2,
      3,
      6,
      9,
      10,
      13,
      16,
      18,
      23,
      25,
      28,
      33,
      36,
      41,
      45,
      50,
      55,
      61,
      76,
      77,
      94,
      95,
      96,
      97,
      103,
      104,
      110,
      119
    ],
    "average": [
      2,
      4,
      5,
      6,
      7,
      11,
      14,
      21,
      51,
      52,
      53,
      55,
      56,
      65,
      68,
      74,
      78,
      79,
      85,
      93,
      131,
      142,
      155,
      156
    ],
    "real": [
      2,
      5,
      6,
      13,
      17,
      18,
      19,
      20,
      22,
      23,
      25,
      27,
      30,
      31,
      33,
      35,
      37,
      40,
      41,
      42,
      44,
      45,
      51,
      55,
      56,
      61,
      67,
      71,
      72,
      74,
      76,
      80,
      84,
      86,
      93,
      94,
      95,
      96,
      98,
      100,
      102,
      103,
      108,
      116,
      118,
      140,
      154,
      156,
      158,
      163
    ],
    "equity": [
      2,
      5,
      6,
      8,
      9,
      10,
      20,
      25,
      26,
      28,
      30,
      33,
      34,
      36,
      40,
      51,
      52,
      55,
      59,
      60,
      64,
      65,
      67,
      68,
      69,
      70,
      72,
      74,
      75,
      76,
      81,
      84,
      89,
      92,
      93,
      95,
      97,
      151,
      154,
      156
    ],
    "fall": [
      2,
      45,
      49,
      68,
      97,
      122,
      141,
      146
    ],
    "rises": [
      2,
      13,
      71,
      79,
      97
    ],
    "percentage": [
      2,
      13,
      19,
      23,
      46,
      56,
      66,
      79,
      154
    ],
    "points": [
      2,
      13,
      20,
      46,
      54,
      62,
      68,
      71,
      79,
      94,
      95,
      96,
      97,
      98
    ],
    "following": [
      2,
      39,
      60,
      62,
      66,
      88,
      95,
      97
    ],
    "onset": [
      2,
      13,
      41,
      48
    ],
    "worst": [
      2,
      3,
      32,
      49,
      54,
      63,
      81,
      93,
      131
    ],
    "case": [
      2,
      3,
      11,
      24,
      28,
      30,
      31,
      32,
      48,
      94,
      96,
      117,
      118,
      119,
      120,
      131,
      157,
      158
    ],
    "scenarios": [
      2,
      30,
      32,
      41,
      47,
      123
    ],
    "averages": [
      2,
      40,
      58,
      140
    ],
    "concept": [
      2,
      9,
      37
    ],
    "exceeds": [
      2,
      13,
      17,
      50,
      51,
      54,
      56,
      57,
      65,
      76
    ],
    "median": [
      2,
      5,
      52,
      91
    ],
    "point": [
      2,
      9,
      18,
      54,
      59,
      73,
      74,
      75,
      76,
      91,
      95,
      129,
      130,
      144,
      147,
      157
    ],
    "falls": [
      2,
      9,
      71
    ],
    "considerably": [
      2,
      53
    ],
    "critique": [
      2,
      157
    ],
    "identified": [
      2,
      15,
      57,
      94,
      120
    ],
    "errors": [
      2,
      22
    ],
    "broader": [
      2,
      6,
      19,
      30,
      31,
      34,
      36,
      40,
      46,
      52,
      57,
      66,
      70,
      74,
      85,
      89,
      91,
      154,
      155,
      156
    ],
    "finding": [
      2,
      3,
      4,
      8,
      15,
      44,
      79,
      85,
      116,
      117
    ],
    "associated": [
      2,
      5,
      6,
      7,
      8,
      14,
      20,
      26,
      32,
      49,
      53,
      69,
      84
    ],
    "held": [
      2,
      23,
      43,
      61,
      69,
      75,
      77,
      94,
      98,
      154,
      155,
      158
    ],
    "multiple": [
      2,
      3,
      4,
      8,
      11,
      16,
      20,
      28,
      29,
      34,
      45,
      50,
      52,
      65,
      67,
      68,
      69,
      74,
      77,
      78,
      85,
      87,
      95,
      98,
      156
    ],
    "studies": [
      2,
      11,
      15,
      27
    ],
    "mechanism": [
      2,
      5,
      11,
      15,
      16,
      21,
      23,
      26,
      29,
      32,
      33,
      35,
      45,
      48,
      49,
      50,
      60,
      61,
      63,
      66,
      68,
      70,
      72,
      73,
      77,
      84,
      93,
      95,
      96,
      97,
      98,
      100,
      116,
      118,
      128,
      157,
      158
    ],
    "straightforward": [
      2,
      11,
      13,
      48,
      70,
      75
    ],
    "requires": [
      2,
      35,
      36,
      37,
      68,
      71,
      90,
      91,
      92,
      136,
      157
    ],
    "leaving": [
      2,
      93,
      149
    ],
    "less": [
      2,
      10,
      12,
      13,
      14,
      15,
      16,
      27,
      30,
      32,
      37,
      55,
      56,
      57,
      70,
      72,
      74,
      77,
      99,
      100,
      140,
      151,
      153
    ],
    "consumption": [
      2,
      42,
      56,
      81,
      99,
      100,
      123,
      138,
      139
    ],
    "provide": [
      2,
      42,
      44,
      46,
      63,
      66,
      71,
      72,
      77,
      81,
      131
    ],
    "insights": [
      2,
      37,
      39,
      40
    ],
    "sentence": [
      2,
      129
    ],
    "finance": [
      2,
      3,
      4,
      15,
      39,
      41,
      73,
      75,
      77,
      94,
      116
    ],
    "features": [
      2,
      18,
      20,
      24,
      27,
      41,
      43
    ],
    "sophisticated": [
      2,
      25,
      44,
      95,
      154
    ],
    "arguments": [
      2,
      51
    ],
    "historical": [
      2,
      6,
      7,
      10,
      19,
      21,
      27,
      31,
      32,
      34,
      41,
      43,
      44,
      45,
      46,
      47,
      48,
      50,
      51,
      52,
      53,
      56,
      57,
      59,
      64,
      65,
      69,
      70,
      71,
      74,
      77,
      78,
      91,
      92,
      93,
      94,
      95,
      96,
      97,
      98,
      112,
      118,
      142,
      155,
      156
    ],
    "patterns": [
      2,
      5,
      41,
      43,
      44,
      46,
      60,
      76,
      81,
      86,
      99,
      100
    ],
    "apply": [
      2,
      51,
      66,
      76
    ],
    "technology": [
      2,
      20,
      24,
      27,
      31,
      51,
      59,
      65,
      154
    ],
    "instruments": [
      2,
      38,
      69,
      70,
      71,
      73
    ],
    "regulatory": [
      2,
      68,
      70,
      72,
      77,
      99,
      155
    ],
    "best": [
      2,
      13,
      70
    ],
    "predictor": [
      2,
      11,
      12,
      14,
      15
    ],
    "relative": [
      2,
      5,
      10,
      16,
      18,
      19,
      22,
      25,
      33,
      37,
      50,
      52,
      53,
      55,
      56,
      57,
      59,
      61,
      72,
      75,
      77,
      90,
      91,
      94,
      98
    ],
    "grows": [
      2,
      50
    ],
    "significantly": [
      2,
      6,
      11,
      13,
      23,
      26,
      30,
      36,
      43,
      45,
      53,
      55,
      57,
      58,
      72,
      74,
      77,
      87,
      92,
      98
    ],
    "question": [
      2,
      17,
      20,
      23,
      31,
      59,
      63,
      78,
      154,
      157
    ],
    "liquidation": [
      2,
      71,
      73,
      89,
      158
    ],
    "nber": [
      2,
      11,
      93,
      95,
      97
    ],
    "global": [
      2,
      5,
      13,
      20,
      21,
      22,
      25,
      27,
      28,
      29,
      30,
      33,
      35,
      37,
      45,
      52,
      57,
      58,
      59,
      60,
      66,
      68,
      69,
      72,
      73,
      75,
      77,
      82,
      86,
      87,
      90,
      92,
      93,
      94,
      96,
      97,
      100,
      102,
      104,
      105,
      106,
      107,
      120,
      121,
      122,
      123,
      125,
      126,
      127,
      128,
      130,
      133,
      134,
      136,
      137,
      139,
      140,
      141,
      142,
      147,
      153,
      154,
      155,
      156
    ],
    "reports": [
      2,
      10,
      21,
      23,
      24,
      29,
      31,
      44,
      70,
      72,
      98,
      99,
      114,
      125,
      128,
      155
    ],
    "network": [
      3,
      27,
      35,
      43,
      77,
      155
    ],
    "represent": [
      3,
      19,
      23,
      35,
      36,
      41,
      54,
      56,
      65,
      69,
      71,
      74,
      88
    ],
    "paradigm": [
      3,
      51
    ],
    "shift": [
      3,
      10,
      25,
      51,
      55,
      68,
      70,
      76,
      85,
      87,
      95,
      99,
      100
    ],
    "analysis": [
      3,
      4,
      7,
      14,
      20,
      27,
      28,
      30,
      32,
      34,
      35,
      39,
      40,
      41,
      43,
      44,
      45,
      46,
      52,
      59,
      71,
      73,
      74,
      93,
      94,
      95,
      96,
      97,
      98,
      99,
      100,
      115,
      116,
      117,
      118,
      119,
      120,
      131,
      154,
      156,
      161,
      162
    ],
    "moving": [
      3,
      5,
      20,
      42,
      72,
      79,
      85,
      86,
      99,
      100,
      131
    ],
    "individual": [
      3,
      8,
      9,
      10,
      19,
      23,
      31,
      34,
      40,
      43,
      73,
      77
    ],
    "institutions": [
      3,
      33,
      43,
      52,
      72,
      75,
      76,
      77,
      99,
      100,
      128
    ],
    "interconnections": [
      3,
      43,
      72
    ],
    "modeled": [
      3,
      35
    ],
    "nodes": [
      3,
      43,
      72
    ],
    "funds": [
      3,
      7,
      28,
      32,
      36,
      54,
      58,
      61,
      62,
      67,
      68,
      69,
      72,
      73,
      74,
      75,
      77,
      88,
      89,
      94,
      96,
      97,
      98,
      109,
      158,
      161
    ],
    "insurers": [
      3,
      70,
      75
    ],
    "clearinghouses": [
      3,
      72,
      98
    ],
    "edges": [
      3,
      43
    ],
    "obligations": [
      3,
      56,
      72,
      94,
      96,
      157
    ],
    "loans": [
      3,
      9,
      16,
      23,
      55,
      67,
      69,
      73,
      76,
      77,
      94,
      117,
      128
    ],
    "contracts": [
      3,
      23,
      46,
      53,
      68,
      71,
      93,
      94,
      136
    ],
    "agreements": [
      3,
      68,
      96
    ],
    "common": [
      3,
      5,
      7,
      10,
      18,
      29,
      33,
      38,
      39,
      45,
      97,
      115,
      116
    ],
    "holdings": [
      3,
      74,
      75,
      158
    ],
    "structure": [
      3,
      17,
      26,
      38,
      43,
      53,
      58,
      63,
      64,
      73,
      94,
      98,
      120,
      156
    ],
    "shocks": [
      3,
      4,
      14,
      16,
      20,
      21,
      30,
      32,
      34,
      52,
      57,
      60,
      84,
      94,
      97,
      120,
      129,
      130,
      153,
      156
    ],
    "propagate": [
      3,
      4,
      20,
      47
    ],
    "shock": [
      3,
      4,
      6,
      8,
      9,
      14,
      16,
      18,
      19,
      20,
      21,
      22,
      26,
      30,
      32,
      33,
      34,
      35,
      36,
      45,
      49,
      53,
      54,
      60,
      62,
      63,
      64,
      66,
      67,
      68,
      69,
      70,
      71,
      76,
      77,
      78,
      81,
      85,
      93,
      97,
      98,
      110,
      141,
      143,
      151,
      156,
      162
    ],
    "node": [
      3,
      43
    ],
    "ways": [
      3,
      66
    ],
    "disproportionate": [
      3,
      35,
      45,
      55,
      61,
      74
    ],
    "initial": [
      3,
      4,
      26,
      33,
      35,
      72,
      79,
      85,
      95
    ],
    "important": [
      3,
      4,
      6,
      12,
      18,
      20,
      21,
      30,
      38,
      49,
      50,
      54,
      56,
      65,
      68,
      72,
      76,
      89,
      131
    ],
    "theory": [
      3,
      128
    ],
    "applied": [
      3,
      37,
      39,
      44,
      95,
      118,
      119
    ],
    "robust": [
      3,
      11,
      83,
      86,
      113
    ],
    "fragile": [
      3,
      34,
      57,
      93,
      94,
      97,
      151
    ],
    "property": [
      3,
      8,
      23,
      35,
      52,
      55,
      76,
      87,
      94
    ],
    "highly": [
      3,
      19,
      28,
      54,
      126
    ],
    "interconnected": [
      3,
      28,
      29,
      72
    ],
    "networks": [
      3,
      4,
      41,
      44,
      59,
      99
    ],
    "random": [
      3,
      41
    ],
    "failure": [
      3,
      10,
      26,
      27,
      43,
      65,
      76,
      95,
      98,
      116,
      117,
      118,
      119,
      120,
      128,
      143,
      154,
      157
    ],
    "arbitrary": [
      3,
      54
    ],
    "institution": [
      3,
      43,
      77
    ],
    "absorbed": [
      3,
      23,
      36,
      74,
      95,
      110,
      129,
      156,
      158
    ],
    "extremely": [
      3,
      14,
      15,
      62,
      85,
      147
    ],
    "targeted": [
      3,
      96,
      110
    ],
    "connected": [
      3,
      28,
      43,
      96
    ],
    "entire": [
      3,
      4,
      9,
      19,
      28,
      29,
      36,
      54,
      57,
      59,
      66,
      67,
      74,
      75,
      77,
      94,
      96,
      105,
      120,
      130
    ],
    "explains": [
      3,
      92
    ],
    "systemically": [
      3,
      4,
      21,
      72,
      96
    ],
    "sifis": [
      3,
      4
    ],
    "size": [
      3,
      25,
      41,
      43,
      75
    ],
    "measures": [
      3,
      5,
      9,
      10,
      13,
      16,
      41,
      50,
      63,
      64,
      80,
      84,
      87,
      91,
      96
    ],
    "fraction": [
      3,
      59,
      65,
      75
    ],
    "direct": [
      3,
      21,
      30,
      31,
      45,
      64,
      68,
      70,
      73,
      77,
      96,
      98,
      102,
      121
    ],
    "exposures": [
      3,
      37,
      43,
      72,
      73
    ],
    "counterparty": [
      3,
      5,
      43,
      45,
      72,
      98
    ],
    "default": [
      3,
      9,
      16,
      18,
      23,
      29,
      33,
      43,
      54,
      69,
      72,
      73,
      75,
      76,
      96,
      98,
      109,
      117,
      118
    ],
    "fire": [
      3,
      10
    ],
    "sale": [
      3,
      10,
      61,
      76
    ],
    "accounts": [
      3,
      4,
      27,
      28,
      54,
      83,
      95,
      158
    ],
    "channels": [
      3,
      18,
      21,
      45,
      58,
      66,
      72,
      77,
      87
    ],
    "shared": [
      3,
      72,
      97
    ],
    "portfolio": [
      3,
      8,
      15,
      17,
      26,
      34,
      37,
      38,
      39,
      45,
      46,
      57,
      61,
      70,
      71,
      73,
      74,
      80,
      89,
      97,
      162
    ],
    "create": [
      3,
      4,
      10,
      13,
      25,
      32,
      35,
      43,
      65,
      70,
      71,
      74,
      75,
      77,
      79,
      84,
      86,
      89,
      130,
      133
    ],
    "correlated": [
      3,
      8,
      26,
      34,
      96,
      155,
      162
    ],
    "pressure": [
      3,
      6,
      21,
      22,
      40,
      56,
      60,
      70,
      74,
      75,
      82,
      83,
      89,
      90,
      91,
      93,
      94,
      129,
      145,
      150,
      155,
      157
    ],
    "funding": [
      3,
      5,
      8,
      9,
      10,
      25,
      45,
      52,
      61,
      68,
      70,
      75,
      77,
      90,
      94,
      95,
      129,
      154,
      155,
      157
    ],
    "distress": [
      3,
      23,
      60,
      69
    ],
    "others": [
      3,
      5,
      8,
      10,
      29,
      33,
      34,
      43,
      52,
      65,
      72
    ],
    "revealed": [
      3,
      23,
      30,
      97,
      98
    ],
    "hold": [
      3,
      10,
      13,
      15,
      19,
      44,
      55,
      57,
      69,
      71,
      74,
      75,
      76,
      81,
      98,
      158
    ],
    "similar": [
      3,
      10,
      24,
      27,
      29,
      46,
      57,
      58,
      68,
      71,
      75,
      84,
      100,
      117
    ],
    "portfolios": [
      3,
      10,
      19,
      23,
      26,
      37,
      39,
      41,
      57,
      61,
      70,
      74
    ],
    "sells": [
      3,
      115,
      133,
      155
    ],
    "potentially": [
      3,
      21,
      58,
      61,
      68,
      69,
      143
    ],
    "triggering": [
      3,
      16,
      18,
      21,
      29,
      33,
      54,
      59,
      71,
      75,
      76,
      85,
      96,
      97,
      107,
      154
    ],
    "sales": [
      3,
      10,
      27,
      61,
      91,
      115
    ],
    "invisible": [
      3,
      25
    ],
    "traditional": [
      3,
      4,
      15,
      26,
      30,
      37,
      38,
      41,
      42,
      43,
      44,
      58,
      70,
      73,
      89,
      92
    ],
    "apparent": [
      3,
      36,
      59,
      97
    ],
    "demonstrated": [
      3,
      20,
      22,
      26,
      28,
      30,
      37,
      71,
      77,
      85
    ],
    "vividly": [
      3,
      77
    ],
    "correlation": [
      3,
      6,
      8,
      26,
      34,
      36,
      47,
      51,
      89,
      92,
      97,
      162
    ],
    "mortgage": [
      3,
      7,
      16,
      28,
      48,
      56,
      60,
      67,
      91,
      96,
      117,
      118
    ],
    "backed": [
      3,
      11,
      28,
      70,
      77,
      91,
      96,
      99
    ],
    "securities": [
      3,
      23,
      28,
      61,
      69,
      74,
      91,
      96
    ],
    "balance": [
      3,
      9,
      23,
      27,
      37,
      40,
      43,
      48,
      50,
      54,
      59,
      63,
      65,
      70,
      76,
      91,
      95,
      96,
      97,
      108,
      119
    ],
    "sheets": [
      3,
      23,
      27,
      40,
      50,
      54,
      59,
      65,
      70,
      76,
      95,
      108,
      119
    ],
    "created": [
      3,
      10,
      28,
      35,
      62,
      67,
      71,
      94,
      95,
      96,
      97,
      130
    ],
    "wide": [
      3,
      9,
      13,
      36,
      56
    ],
    "single": [
      3,
      4,
      8,
      10,
      11,
      15,
      18,
      19,
      34,
      37,
      46,
      63,
      66,
      68,
      69,
      71,
      75,
      76,
      91,
      96,
      117,
      136,
      143
    ],
    "management": [
      3,
      17,
      26,
      37,
      39,
      73,
      80,
      156
    ],
    "detect": [
      3,
      100
    ],
    "federal": [
      3,
      6,
      7,
      9,
      10,
      11,
      12,
      16,
      18,
      20,
      23,
      32,
      42,
      54,
      56,
      58,
      60,
      62,
      66,
      67,
      68,
      69,
      70,
      72,
      76,
      77,
      91,
      93,
      95,
      96,
      97,
      100,
      119,
      158,
      160
    ],
    "reserve": [
      3,
      6,
      7,
      9,
      10,
      11,
      12,
      16,
      18,
      20,
      21,
      22,
      23,
      42,
      54,
      56,
      58,
      60,
      62,
      66,
      67,
      68,
      69,
      72,
      76,
      77,
      87,
      88,
      91,
      93,
      95,
      96,
      97,
      100,
      158
    ],
    "acemoglu": [
      3,
      4
    ],
    "allen": [
      3,
      155
    ],
    "input": [
      4,
      21,
      32,
      35,
      38,
      41,
      45,
      47,
      69,
      82,
      93,
      136,
      143,
      149
    ],
    "rooted": [
      4,
      15
    ],
    "nobel": [
      4,
      14,
      51
    ],
    "work": [
      4,
      40,
      55,
      76,
      94,
      140,
      154
    ],
    "analyze": [
      4,
      10,
      43
    ],
    "production": [
      4,
      21,
      35,
      66,
      80,
      82,
      92,
      93,
      106,
      125,
      139,
      141,
      142,
      145
    ],
    "unlike": [
      4,
      16,
      17,
      20,
      35,
      43,
      58,
      59,
      73,
      86,
      153
    ],
    "physical": [
      4,
      35,
      44,
      86,
      99,
      136,
      144,
      151,
      153
    ],
    "flow": [
      4,
      16,
      23,
      29,
      54,
      69,
      71,
      74,
      100,
      150,
      154,
      155,
      156,
      158,
      161
    ],
    "goods": [
      4,
      35,
      45,
      83,
      86,
      131
    ],
    "services": [
      4,
      24,
      27,
      32,
      35,
      45,
      46,
      47,
      49,
      59,
      67,
      80,
      81,
      83
    ],
    "industries": [
      4,
      21,
      45,
      46,
      47,
      83
    ],
    "industry": [
      4,
      24,
      35,
      45,
      47,
      70,
      125,
      132,
      136,
      137,
      139
    ],
    "supplies": [
      4,
      134,
      138,
      150
    ],
    "disruption": [
      4,
      21,
      30,
      32,
      35,
      45,
      52,
      60,
      63,
      82,
      127,
      128,
      135,
      141
    ],
    "creates": [
      4,
      7,
      10,
      12,
      14,
      19,
      22,
      25,
      26,
      27,
      32,
      34,
      45,
      47,
      52,
      54,
      56,
      60,
      64,
      69,
      71,
      76,
      77,
      79,
      80,
      81,
      83,
      84,
      90,
      100,
      130,
      148,
      157
    ],
    "downstream": [
      4,
      35
    ],
    "constraints": [
      4,
      47,
      80
    ],
    "chain": [
      4,
      21,
      28,
      35,
      45,
      47,
      55,
      59,
      63,
      65,
      96,
      122,
      127,
      129
    ],
    "supply": [
      4,
      20,
      21,
      30,
      32,
      35,
      44,
      45,
      46,
      47,
      55,
      59,
      60,
      63,
      65,
      75,
      77,
      80,
      87,
      91,
      92,
      93,
      97,
      122,
      127,
      132,
      135,
      138,
      139,
      142,
      147,
      150
    ],
    "side": [
      4,
      60,
      62,
      63,
      71,
      79,
      85
    ],
    "transmit": [
      4,
      30,
      53
    ],
    "demand": [
      4,
      15,
      17,
      27,
      45,
      46,
      47,
      55,
      58,
      60,
      66,
      72,
      73,
      75,
      77,
      80,
      82,
      83,
      84,
      86,
      87,
      88,
      90,
      92,
      97,
      146
    ],
    "consumers": [
      4,
      12,
      21,
      45,
      59,
      66,
      84,
      100,
      137
    ],
    "stop": [
      4,
      18,
      68,
      75
    ],
    "moves": [
      4,
      25,
      39,
      44,
      53,
      66,
      115
    ],
    "upstream": [
      4,
      45
    ],
    "unavailable": [
      4,
      35,
      43
    ],
    "disrupted": [
      4,
      35,
      63,
      66,
      127,
      143
    ],
    "simultaneously": [
      4,
      8,
      10,
      12,
      19,
      25,
      26,
      28,
      30,
      32,
      33,
      34,
      36,
      37,
      45,
      49,
      54,
      60,
      64,
      65,
      68,
      69,
      71,
      72,
      77,
      78,
      84,
      97,
      98,
      100,
      135,
      136,
      155
    ],
    "constrained": [
      4,
      58,
      92,
      93
    ],
    "asymmetry": [
      4,
      151
    ],
    "means": [
      4,
      8,
      9,
      11,
      13,
      15,
      24,
      25,
      26,
      34,
      35,
      36,
      40,
      45,
      47,
      48,
      50,
      53,
      54,
      57,
      64,
      65,
      66,
      71,
      73,
      77,
      80,
      85,
      89
    ],
    "severe": [
      4,
      6,
      11,
      17,
      26,
      27,
      31,
      32,
      33,
      34,
      41,
      46,
      55,
      60,
      64,
      68,
      70,
      91,
      94,
      95,
      98,
      113,
      120,
      123,
      143
    ],
    "equivalent": [
      4,
      36,
      59,
      81,
      91,
      98
    ],
    "magnitude": [
      4,
      27,
      40,
      47,
      51,
      60,
      66,
      69,
      71,
      98,
      158
    ],
    "modern": [
      4,
      28,
      29,
      30,
      43,
      48,
      49,
      57,
      63,
      65,
      70,
      88,
      93,
      95,
      96,
      121,
      135,
      139
    ],
    "enhanced": [
      4,
      47,
      61
    ],
    "granular": [
      4,
      44,
      95,
      117
    ],
    "power": [
      4,
      11,
      12,
      13,
      16,
      22,
      24,
      59,
      63,
      80,
      84,
      91,
      104,
      138
    ],
    "researchers": [
      4,
      15,
      43
    ],
    "cascades": [
      4,
      20,
      35,
      66,
      70,
      73
    ],
    "firm": [
      4,
      70,
      118
    ],
    "revealing": [
      4,
      96
    ],
    "aggregate": [
      4,
      8,
      9,
      10,
      23,
      25,
      33,
      44,
      54,
      55,
      56,
      72,
      73,
      95,
      98,
      114,
      154,
      156
    ],
    "fluctuations": [
      4,
      14,
      53,
      86
    ],
    "large": [
      4,
      9,
      10,
      15,
      16,
      19,
      23,
      29,
      57,
      61,
      71,
      74,
      80,
      96,
      127
    ],
    "major": [
      4,
      7,
      9,
      17,
      18,
      20,
      21,
      24,
      25,
      26,
      27,
      32,
      40,
      55,
      64,
      69,
      72,
      74,
      76,
      84,
      86,
      89,
      90,
      91,
      92,
      98,
      99,
      100,
      129,
      140,
      149,
      153,
      155,
      156,
      157,
      158
    ],
    "producer": [
      4,
      30,
      133,
      137
    ],
    "rare": [
      4,
      17,
      35
    ],
    "earth": [
      4,
      35,
      121
    ],
    "mine": [
      4,
      17,
      92
    ],
    "effects": [
      4,
      10,
      12,
      21,
      25,
      35,
      41,
      47,
      67,
      84,
      158,
      159
    ],
    "documented": [
      4,
      12,
      22,
      38,
      39,
      66,
      72,
      128
    ],
    "macroeconomic": [
      4,
      11,
      60,
      98,
      118,
      155
    ],
    "assumption": [
      4,
      30,
      66,
      74,
      96,
      97,
      152,
      153
    ],
    "identify": [
      4,
      41,
      43,
      116
    ],
    "firms": [
      4,
      59,
      70,
      71
    ],
    "perspective": [
      4,
      57
    ],
    "analogous": [
      4,
      73
    ],
    "fabrication": [
      4,
      35,
      147
    ],
    "concentrated": [
      4,
      17,
      19,
      23,
      34,
      35,
      55,
      57,
      65,
      69,
      71,
      72,
      76,
      158
    ],
    "taiwan": [
      4,
      35
    ],
    "south": [
      4,
      29,
      94,
      126,
      127,
      128,
      130,
      134,
      138,
      142
    ],
    "korea": [
      4,
      29,
      94,
      142
    ],
    "processing": [
      4,
      35,
      41,
      44,
      100,
      121,
      135,
      155
    ],
    "dominated": [
      4,
      37,
      57,
      65
    ],
    "increasingly": [
      4,
      15,
      35,
      37,
      44,
      52,
      73,
      75,
      94
    ],
    "cloud": [
      4,
      24,
      27,
      35,
      45,
      47,
      99,
      100,
      154
    ],
    "computing": [
      4,
      35,
      45,
      47,
      157
    ],
    "event": [
      4,
      6,
      15,
      17,
      18,
      20,
      22,
      25,
      28,
      29,
      30,
      34,
      36,
      45,
      52,
      53,
      55,
      56,
      57,
      69,
      71,
      73,
      74,
      76,
      77,
      85,
      89,
      94,
      96,
      115,
      158,
      160
    ],
    "affecting": [
      4,
      19,
      66
    ],
    "cascading": [
      4,
      21,
      35,
      43,
      47,
      54,
      67,
      68,
      127,
      134
    ],
    "failures": [
      4,
      9,
      23,
      55,
      61,
      76
    ],
    "amplifying": [
      4,
      33,
      47,
      71
    ],
    "original": [
      4,
      23,
      24,
      35,
      54,
      116
    ],
    "absorbing": [
      4,
      92,
      99
    ],
    "oecd": [
      4,
      21,
      58
    ],
    "tables": [
      4,
      35,
      45,
      47
    ],
    "database": [
      4,
      6,
      7,
      35
    ],
    "available": [
      5,
      6,
      7,
      8,
      10,
      13,
      14,
      16,
      17,
      21,
      23,
      37,
      40,
      42,
      43,
      44,
      51,
      61,
      70,
      71,
      86,
      94,
      95
    ],
    "updated": [
      5,
      6,
      7,
      8,
      9,
      10,
      11,
      12,
      14
    ],
    "variables": [
      5,
      41,
      42,
      52
    ],
    "corporate": [
      5,
      6,
      7,
      14,
      16,
      18,
      25,
      36,
      43,
      51,
      54,
      56,
      59,
      60,
      62,
      67,
      69,
      72,
      93,
      95,
      96,
      97,
      119,
      155,
      163
    ],
    "bond": [
      5,
      6,
      7,
      8,
      10,
      11,
      12,
      18,
      37,
      48,
      49,
      58,
      60,
      61,
      69,
      72,
      73,
      75,
      77,
      93,
      94,
      95,
      96,
      97,
      98,
      119,
      160
    ],
    "municipal": [
      5,
      119
    ],
    "earnings": [
      5,
      14,
      18,
      24,
      31,
      40,
      47,
      51,
      65,
      77,
      95,
      155,
      156
    ],
    "spread": [
      5,
      6,
      7,
      11,
      12,
      23,
      29,
      35,
      48,
      49,
      64,
      68,
      69,
      70,
      72,
      74,
      75,
      77,
      91,
      94,
      95,
      96,
      97,
      131
    ],
    "commercial": [
      5,
      6,
      21,
      23,
      55,
      56,
      67,
      76,
      136
    ],
    "paper": [
      5,
      6,
      61
    ],
    "treasury": [
      5,
      6,
      7,
      11,
      12,
      48,
      49,
      58,
      67,
      68,
      72,
      75,
      77,
      91,
      93,
      95,
      96,
      97,
      98,
      100
    ],
    "cross": [
      5,
      8,
      22,
      25,
      26,
      34,
      36,
      42,
      47,
      66,
      75,
      77,
      90,
      162
    ],
    "currency": [
      5,
      22,
      25,
      29,
      30,
      58,
      75,
      77,
      87,
      90,
      93,
      94,
      98,
      100,
      101,
      110
    ],
    "basis": [
      5,
      22,
      54,
      62,
      68,
      75,
      77,
      90,
      91,
      94,
      95,
      96,
      97,
      98
    ],
    "safe": [
      5,
      11,
      20,
      26,
      29,
      30,
      41,
      52,
      69,
      74,
      88,
      89,
      90,
      97,
      161,
      162
    ],
    "move": [
      5,
      8,
      10,
      15,
      19,
      35,
      66,
      68,
      71,
      78,
      88,
      90,
      97,
      98,
      147
    ],
    "constructed": [
      5,
      7
    ],
    "factor": [
      5,
      7,
      24,
      42,
      143
    ],
    "extracts": [
      5,
      154
    ],
    "underlying": [
      5,
      24,
      31,
      41,
      43,
      45,
      54,
      57,
      64,
      71,
      74,
      75,
      77,
      83,
      87,
      97,
      117,
      156
    ],
    "variable": [
      5,
      38,
      61
    ],
    "normalized": [
      5,
      7
    ],
    "history": [
      5,
      7,
      24,
      48,
      49,
      52,
      63,
      70,
      74,
      81,
      82,
      95,
      96,
      97,
      98,
      111,
      112
    ],
    "captures": [
      5,
      8,
      9,
      10,
      13,
      15,
      16,
      18,
      34,
      37,
      64,
      81
    ],
    "deviations": [
      5,
      7
    ],
    "normal": [
      5,
      6,
      8,
      17,
      26,
      32,
      33,
      38,
      42,
      45,
      47,
      53,
      71,
      72,
      74,
      86,
      93,
      97,
      134,
      146
    ],
    "conditions": [
      5,
      6,
      9,
      11,
      12,
      13,
      14,
      15,
      17,
      18,
      22,
      26,
      30,
      33,
      36,
      37,
      39,
      41,
      42,
      46,
      52,
      61,
      64,
      67,
      71,
      72,
      80,
      84,
      90,
      93,
      94,
      95,
      96,
      97,
      98,
      99,
      128
    ],
    "absolute": [
      5,
      51,
      61
    ],
    "reading": [
      5,
      6,
      7,
      10,
      34,
      48,
      49,
      50,
      51,
      52,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      61,
      62,
      78,
      79,
      80,
      81,
      82,
      83,
      84,
      85,
      86,
      87,
      88,
      89,
      90,
      91,
      92
    ],
    "positive": [
      5,
      11,
      13,
      47,
      48,
      49,
      50,
      56,
      61,
      63,
      64,
      89,
      121,
      158
    ],
    "indicate": [
      5,
      9,
      19,
      22,
      23,
      34
    ],
    "above": [
      5,
      6,
      7,
      8,
      9,
      14,
      17,
      19,
      21,
      22,
      26,
      31,
      32,
      34,
      49,
      50,
      51,
      52,
      53,
      54,
      55,
      56,
      58,
      61,
      62,
      75,
      78,
      79,
      80,
      81,
      82,
      83,
      85,
      86,
      88,
      90,
      95
    ],
    "readings": [
      5,
      7,
      8,
      9,
      14,
      22,
      30,
      49,
      53,
      54,
      80,
      84,
      86,
      95,
      96,
      98
    ],
    "historically": [
      5,
      6,
      7,
      8,
      9,
      17,
      18,
      21,
      25,
      26,
      32,
      40,
      46,
      50,
      53,
      54,
      56,
      58,
      64,
      74,
      78,
      79,
      82,
      85,
      87,
      89,
      92,
      97,
      98
    ],
    "significant": [
      5,
      6,
      7,
      10,
      11,
      12,
      13,
      14,
      18,
      21,
      25,
      36,
      38,
      41,
      44,
      52,
      53,
      54,
      55,
      56,
      67,
      70,
      71,
      74,
      75,
      76,
      78,
      83,
      84,
      88,
      89,
      90,
      92,
      93,
      97,
      98,
      100
    ],
    "dislocations": [
      5,
      7,
      17
    ],
    "reached": [
      5,
      14,
      34,
      50,
      56,
      58,
      75,
      76,
      79,
      88,
      90,
      92,
      94,
      95,
      97,
      144
    ],
    "approximately": [
      5,
      9,
      11,
      13,
      14,
      27,
      28,
      37,
      41,
      46,
      48,
      50,
      51,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      61,
      62,
      66,
      67,
      68,
      69,
      70,
      74,
      75,
      76,
      78,
      79,
      81,
      84,
      86,
      88,
      89,
      91,
      92,
      93,
      94,
      95,
      96,
      97,
      98
    ],
    "peak": [
      5,
      8,
      15,
      27,
      51,
      54,
      55,
      56,
      57,
      61,
      62,
      63,
      65,
      78,
      88,
      93,
      95,
      101,
      106,
      131,
      140
    ],
    "spiked": [
      5,
      20,
      30,
      42,
      54,
      72,
      93,
      94,
      97,
      98,
      149
    ],
    "march": [
      5,
      8,
      9,
      10,
      17,
      20,
      23,
      26,
      27,
      30,
      38,
      39,
      45,
      48,
      49,
      50,
      51,
      52,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      61,
      62,
      63,
      64,
      65,
      66,
      67,
      68,
      69,
      72,
      78,
      79,
      80,
      81,
      82,
      83,
      84,
      85,
      86,
      87,
      88,
      89,
      90,
      91,
      92,
      94,
      95,
      96,
      97,
      99,
      154
    ],
    "makes": [
      5,
      6,
      7,
      8,
      9,
      14,
      16,
      17,
      18,
      37,
      38,
      46,
      47,
      66,
      67,
      75,
      76,
      81,
      157
    ],
    "valuable": [
      5,
      9,
      14,
      30,
      38,
      42,
      44,
      98
    ],
    "frequency": [
      5,
      6,
      7,
      42
    ],
    "decomposition": [
      5,
      39,
      52
    ],
    "capability": [
      5,
      24,
      100,
      157,
      158
    ],
    "analysts": [
      5,
      41,
      50,
      93,
      154
    ],
    "contributing": [
      5,
      55,
      56,
      75,
      142
    ],
    "overall": [
      5,
      34,
      36,
      52,
      55,
      115
    ],
    "given": [
      5,
      16,
      52,
      63,
      69,
      157
    ],
    "reveals": [
      5,
      30,
      44,
      45,
      47,
      52,
      93,
      95,
      96
    ],
    "nature": [
      5,
      63,
      70,
      121
    ],
    "implications": [
      5,
      31,
      87,
      91
    ],
    "suggests": [
      5,
      6,
      24,
      37,
      55,
      56,
      78,
      79,
      80,
      82,
      86,
      90,
      92
    ],
    "concerns": [
      5,
      6,
      9,
      60,
      70,
      75,
      77
    ],
    "solvency": [
      5,
      45,
      69,
      96,
      162,
      163
    ],
    "slow": [
      5,
      9,
      24,
      47,
      53,
      54,
      55,
      65,
      68,
      72,
      76,
      99,
      100,
      122,
      142
    ],
    "preceded": [
      5,
      7,
      9,
      17,
      18,
      49,
      52,
      54,
      64,
      74,
      93,
      96
    ],
    "deteriorating": [
      5,
      43,
      48,
      55,
      64,
      73,
      79,
      96
    ],
    "fundamentals": [
      5,
      12,
      33,
      79,
      95
    ],
    "fast": [
      5,
      7,
      47,
      70,
      74,
      85,
      94,
      122
    ],
    "quickly": [
      5,
      16,
      20,
      45,
      50,
      60,
      70,
      85
    ],
    "interbank": [
      5,
      6,
      8,
      9,
      23,
      28,
      52,
      77,
      94,
      96
    ],
    "pattern": [
      5,
      6,
      12,
      15,
      17,
      22,
      25,
      30,
      31,
      34,
      41,
      46,
      48,
      49,
      52,
      63,
      64,
      74,
      83,
      85,
      89,
      97,
      115,
      129,
      154
    ],
    "category": [
      5,
      18,
      20,
      22,
      52,
      130
    ],
    "contributions": [
      5,
      158
    ],
    "widely": [
      5,
      11,
      12,
      23,
      41,
      48,
      51,
      59,
      60,
      62,
      71,
      154,
      158
    ],
    "recognized": [
      5,
      94,
      97
    ],
    "breakdowns": [
      5,
      154
    ],
    "advanced": [
      5,
      35,
      58
    ],
    "begins": [
      5,
      12,
      48,
      56
    ],
    "elevated": [
      5,
      6,
      10,
      11,
      14,
      15,
      16,
      18,
      19,
      20,
      21,
      22,
      23,
      26,
      32,
      33,
      34,
      40,
      41,
      42,
      48,
      49,
      50,
      51,
      52,
      53,
      54,
      58,
      60,
      62,
      65,
      67,
      68,
      69,
      71,
      74,
      77,
      78,
      82,
      86,
      88,
      89,
      91,
      92,
      95,
      133,
      158,
      161
    ],
    "region": [
      5,
      29,
      66
    ],
    "follows": [
      5,
      25,
      32,
      33,
      48,
      50,
      76,
      84,
      96,
      140
    ],
    "trajectory": [
      5,
      7,
      11,
      31,
      48,
      49,
      54,
      55,
      58,
      76,
      81,
      84,
      116,
      124,
      154
    ],
    "synchronized": [
      5,
      68,
      123
    ],
    "asian": [
      5,
      22,
      29,
      87,
      94,
      137
    ],
    "showed": [
      5,
      30,
      42,
      44,
      45,
      52,
      67
    ],
    "clear": [
      5,
      48,
      50,
      52,
      72,
      85
    ],
    "iran": [
      5,
      20,
      30,
      32,
      52,
      130,
      141,
      152,
      153
    ],
    "tensions": [
      5,
      20,
      30,
      32,
      77,
      84,
      88,
      90
    ],
    "differentiated": [
      5,
      30
    ],
    "responses": [
      5,
      37,
      60
    ],
    "https": [
      5,
      16
    ],
    "cleveland": [
      6,
      7,
      9,
      23,
      42
    ],
    "composite": [
      6,
      7,
      8,
      27,
      34,
      36,
      95
    ],
    "tracks": [
      6,
      24,
      37,
      44,
      86
    ],
    "components": [
      6,
      28,
      52,
      64,
      74,
      81
    ],
    "segments": [
      6,
      8,
      30,
      34,
      55,
      96
    ],
    "stock": [
      6,
      9,
      15,
      16,
      19,
      31,
      33,
      34,
      36,
      37,
      39,
      43,
      47,
      56,
      74,
      95,
      103,
      105,
      113,
      156
    ],
    "foreign": [
      6,
      8,
      29,
      58,
      75,
      77,
      90,
      94,
      128,
      130,
      158
    ],
    "exchange": [
      6,
      7,
      8,
      22,
      25,
      29,
      75,
      86,
      90,
      93,
      94,
      128
    ],
    "weighted": [
      6,
      10,
      42,
      57,
      65,
      74
    ],
    "dollar": [
      6,
      22,
      29,
      30,
      58,
      68,
      75,
      77,
      90,
      93,
      94,
      98,
      99,
      100,
      101,
      111,
      115,
      118,
      138,
      154,
      162
    ],
    "providing": [
      6,
      7,
      40,
      50,
      51,
      89,
      96,
      97,
      150
    ],
    "slower": [
      6,
      8,
      42,
      56,
      66,
      143
    ],
    "indices": [
      6,
      7,
      8,
      9,
      15,
      18,
      26,
      36,
      38,
      42,
      52,
      57,
      76,
      82,
      87,
      96
    ],
    "uses": [
      6,
      8,
      34,
      35,
      42,
      43,
      44,
      116,
      121,
      147,
      149
    ],
    "severity": [
      6,
      11,
      26,
      84,
      115
    ],
    "percentile": [
      6,
      10
    ],
    "grade": [
      6,
      18,
      54,
      69,
      97
    ],
    "below": [
      6,
      7,
      9,
      14,
      23,
      31,
      50,
      51,
      52,
      55,
      56,
      67,
      69,
      76,
      79,
      83,
      85,
      86,
      87,
      89,
      92,
      154
    ],
    "indicates": [
      6,
      7,
      19,
      40,
      52,
      74
    ],
    "benign": [
      6,
      13,
      49,
      80,
      81,
      85
    ],
    "flowing": [
      6,
      52,
      89
    ],
    "functioning": [
      6,
      15,
      33,
      52,
      66,
      68,
      97,
      98
    ],
    "typical": [
      6,
      53,
      66,
      92
    ],
    "moderate": [
      6,
      9,
      16,
      26,
      34,
      35,
      64,
      79
    ],
    "flags": [
      6,
      15,
      152
    ],
    "warrant": [
      6,
      12,
      41,
      86
    ],
    "attention": [
      6,
      12,
      86,
      115
    ],
    "widening": [
      6,
      34,
      49,
      74,
      90
    ],
    "lending": [
      6,
      18,
      23,
      28,
      33,
      47,
      70,
      73,
      76,
      77,
      89,
      96,
      97,
      110
    ],
    "showing": [
      6,
      58,
      63,
      83
    ],
    "probability": [
      6,
      11,
      13,
      38,
      41,
      88,
      89,
      93
    ],
    "impact": [
      6,
      10,
      15,
      16,
      20,
      21,
      25,
      30,
      34,
      45,
      57,
      65,
      74,
      87,
      97,
      100,
      105,
      140
    ],
    "extreme": [
      6,
      7,
      15,
      19,
      21,
      25,
      30,
      38,
      39,
      51,
      57,
      65,
      68,
      71,
      74,
      94,
      95,
      96,
      98,
      115
    ],
    "signals": [
      6,
      7,
      9,
      11,
      12,
      13,
      17,
      18,
      26,
      28,
      38,
      44,
      50,
      53,
      61,
      62,
      64,
      77,
      91,
      99,
      100,
      116
    ],
    "grades": [
      6,
      92,
      96
    ],
    "jump": [
      6,
      18,
      66,
      68,
      83
    ],
    "resilience": [
      6,
      10,
      30,
      46,
      56,
      63,
      65,
      70,
      97
    ],
    "gradual": [
      6,
      13,
      18,
      19,
      53,
      55,
      60,
      79,
      91,
      94
    ],
    "drift": [
      6,
      19
    ],
    "building": [
      6,
      7,
      17,
      24,
      27,
      53,
      55,
      56,
      60,
      63,
      64,
      66,
      76,
      94,
      139
    ],
    "cannot": [
      6,
      18,
      23,
      35,
      43,
      50,
      60,
      63,
      66,
      67,
      68,
      69,
      70,
      73,
      81,
      83,
      92,
      99,
      126,
      128,
      146,
      147,
      154,
      158
    ],
    "correct": [
      6,
      13,
      27,
      40,
      65,
      66,
      107,
      115,
      154
    ],
    "compressed": [
      6,
      7,
      15,
      78,
      95
    ],
    "premiums": [
      6,
      7,
      30,
      39,
      66,
      70,
      71,
      74
    ],
    "escalation": [
      6,
      20,
      53,
      80,
      82,
      84,
      85,
      86,
      87,
      88,
      160
    ],
    "progression": [
      6,
      96
    ],
    "visible": [
      6,
      23,
      31,
      44,
      49,
      78,
      80,
      86
    ],
    "better": [
      6,
      43,
      140
    ],
    "trend": [
      6,
      13,
      50,
      61,
      81,
      87,
      88,
      95,
      97,
      100
    ],
    "alarm": [
      6,
      59
    ],
    "alongside": [
      6,
      12,
      26,
      85,
      89,
      97
    ],
    "distinguish": [
      6,
      43
    ],
    "temporary": [
      6,
      20,
      31,
      40
    ],
    "spikes": [
      6,
      8,
      21,
      23,
      34,
      39,
      52,
      63,
      66,
      68,
      80,
      85,
      93,
      129
    ],
    "deterioration": [
      6,
      9,
      18,
      44,
      48,
      55,
      70,
      79,
      83,
      85,
      94,
      156
    ],
    "confirms": [
      6,
      63
    ],
    "weeks": [
      6,
      17,
      20,
      25,
      26,
      42,
      44,
      47,
      52,
      66,
      68,
      75,
      85,
      86,
      97
    ],
    "credible": [
      6,
      59,
      75,
      157
    ],
    "alone": [
      6,
      8,
      25,
      36,
      53,
      61,
      75,
      80,
      88,
      92,
      119,
      143,
      154
    ],
    "fred": [
      6,
      7,
      11,
      12,
      13,
      18,
      22,
      32,
      33,
      48,
      49,
      56,
      93
    ],
    "series": [
      6,
      7,
      11,
      12,
      15,
      24,
      48,
      49,
      56,
      85,
      94
    ],
    "louis": [
      7,
      149
    ],
    "stlfsi": [
      7
    ],
    "fourth": [
      7,
      66
    ],
    "effective": [
      7,
      22,
      46,
      56,
      60,
      66,
      153
    ],
    "yield": [
      7,
      11,
      12,
      18,
      25,
      34,
      36,
      40,
      41,
      48,
      49,
      54,
      58,
      60,
      62,
      63,
      64,
      67,
      69,
      73,
      75,
      91,
      93,
      95,
      96,
      97,
      123,
      125,
      126,
      127,
      140,
      142,
      143
    ],
    "fixed": [
      7,
      8,
      16,
      25,
      56,
      91,
      93,
      94,
      98,
      104
    ],
    "month": [
      7,
      11,
      12,
      17,
      21,
      48,
      53,
      60,
      63,
      64,
      66,
      79,
      82,
      84,
      86,
      87,
      93,
      94,
      96,
      123
    ],
    "minus": [
      7,
      11,
      12,
      22,
      48,
      49,
      93,
      95,
      96
    ],
    "chicago": [
      7,
      72
    ],
    "board": [
      7,
      10,
      20,
      33,
      62,
      67,
      84,
      95
    ],
    "futures": [
      7,
      17,
      21,
      25,
      30,
      53,
      64,
      67,
      68,
      71,
      72,
      89,
      142
    ],
    "implied": [
      7,
      15,
      17,
      26
    ],
    "sector": [
      7,
      9,
      13,
      16,
      19,
      23,
      31,
      36,
      45,
      46,
      47,
      50,
      54,
      55,
      56,
      57,
      60,
      61,
      69,
      76,
      78,
      82,
      83,
      87,
      93,
      95,
      97,
      131,
      146,
      155
    ],
    "component": [
      7,
      32,
      33,
      36,
      39,
      52,
      56,
      64,
      65,
      70,
      80,
      84,
      91,
      93
    ],
    "representing": [
      7,
      43,
      52,
      60,
      74,
      78,
      79,
      82,
      89
    ],
    "mean": [
      7,
      14,
      27,
      40,
      129
    ],
    "unit": [
      7,
      27
    ],
    "standard": [
      7,
      14,
      60,
      66,
      69,
      116,
      155,
      157
    ],
    "deviation": [
      7,
      13,
      21,
      50,
      51
    ],
    "tight": [
      7,
      11,
      80,
      93
    ],
    "indicating": [
      7,
      13,
      53,
      85,
      161
    ],
    "terms": [
      7,
      16,
      52,
      65,
      67,
      76,
      90,
      93,
      94
    ],
    "enough": [
      7,
      11,
      15,
      23,
      60,
      63,
      70,
      74,
      96,
      146,
      157
    ],
    "capture": [
      7,
      41,
      42,
      43,
      47,
      50,
      53,
      70,
      86,
      116,
      120
    ],
    "rapidly": [
      7,
      12,
      13,
      20,
      27,
      35,
      42,
      43,
      52,
      60,
      96
    ],
    "evolving": [
      7,
      43,
      89
    ],
    "events": [
      7,
      8,
      17,
      18,
      19,
      20,
      21,
      22,
      23,
      26,
      35,
      38,
      39,
      41,
      43,
      45,
      46,
      85,
      95,
      117,
      130,
      159
    ],
    "week": [
      7,
      42,
      85,
      91,
      95
    ],
    "noise": [
      7,
      12,
      13,
      44
    ],
    "generate": [
      7,
      39,
      70,
      75,
      96
    ],
    "false": [
      7,
      8,
      11,
      13,
      96,
      155
    ],
    "useful": [
      7,
      15,
      17,
      85,
      159,
      160,
      162
    ],
    "tracking": [
      7,
      34,
      35,
      36,
      44,
      57
    ],
    "feature": [
      7,
      35,
      38,
      41,
      73,
      92
    ],
    "calculated": [
      7,
      13,
      15,
      17
    ],
    "early": [
      7,
      13,
      19,
      21,
      24,
      28,
      40,
      42,
      44,
      49,
      50,
      51,
      57,
      60,
      61,
      62,
      65,
      67,
      77,
      79,
      87,
      88,
      93,
      96,
      99,
      100,
      116,
      151
    ],
    "context": [
      7,
      31,
      46,
      53,
      54,
      61,
      69,
      82,
      85,
      86
    ],
    "extended": [
      7,
      52
    ],
    "episodes": [
      7,
      8,
      16,
      17,
      41,
      52,
      63,
      71
    ],
    "interpretation": [
      7,
      80,
      92
    ],
    "consistent": [
      7,
      48,
      84,
      89
    ],
    "taking": [
      7,
      43,
      50,
      71,
      88,
      137,
      145
    ],
    "stage": [
      7,
      59,
      91
    ],
    "european": [
      8,
      58,
      77,
      87,
      90,
      96,
      99,
      100,
      125,
      137,
      150,
      163
    ],
    "distinct": [
      8,
      10,
      23,
      116
    ],
    "phenomenon": [
      8,
      15,
      19,
      26,
      83,
      93
    ],
    "miss": [
      8,
      31,
      41,
      64,
      130
    ],
    "interaction": [
      8,
      10,
      21,
      28,
      30,
      32,
      34,
      41,
      78,
      93,
      95,
      96,
      97
    ],
    "derived": [
      8,
      9,
      132,
      136,
      139
    ],
    "weights": [
      8,
      81
    ],
    "theoretical": [
      8,
      13,
      23,
      34,
      68,
      129
    ],
    "varying": [
      8,
      34
    ],
    "correlations": [
      8,
      9,
      26,
      34,
      44
    ],
    "innovation": [
      8,
      41,
      42,
      43,
      103,
      106
    ],
    "designed": [
      8,
      37,
      38,
      72
    ],
    "stressed": [
      8,
      34,
      75
    ],
    "reflects": [
      8,
      11,
      17,
      39,
      48,
      49,
      50,
      52,
      54,
      61,
      80,
      82,
      86,
      88,
      89,
      91
    ],
    "fundamentally": [
      8,
      9,
      15,
      26,
      31,
      54,
      57,
      85,
      103,
      105,
      117
    ],
    "segment": [
      8,
      18,
      34,
      55,
      61,
      73,
      97
    ],
    "spike": [
      8,
      18,
      21,
      22,
      26,
      45,
      48,
      52,
      54,
      60,
      66,
      68,
      72,
      85,
      91,
      96,
      97,
      98,
      122,
      129,
      139,
      149,
      156
    ],
    "correction": [
      8,
      15,
      24,
      27,
      30,
      31,
      33,
      34,
      36,
      65,
      66,
      78,
      88,
      91,
      151
    ],
    "simultaneous": [
      8,
      27,
      33,
      34,
      45,
      60,
      63,
      69,
      70,
      93,
      94,
      136,
      143
    ],
    "rising": [
      8,
      9,
      13,
      14,
      15,
      16,
      33,
      34,
      36,
      37,
      39,
      49,
      56,
      60,
      62,
      70,
      81,
      84,
      89,
      93,
      95,
      98,
      124
    ],
    "weighting": [
      8,
      34,
      42
    ],
    "increases": [
      8,
      16,
      17,
      26,
      31,
      44,
      46,
      54,
      56,
      58,
      60,
      61,
      67,
      76,
      78,
      82,
      83,
      87,
      90,
      98
    ],
    "rise": [
      8,
      34,
      61,
      66,
      75,
      89,
      97,
      122,
      124
    ],
    "times": [
      8,
      11,
      12,
      22,
      47,
      51,
      96,
      97,
      98,
      131,
      148
    ],
    "offset": [
      8,
      26,
      38,
      67
    ],
    "diversification": [
      8,
      26,
      34,
      37,
      46,
      76,
      88,
      89
    ],
    "together": [
      8,
      53
    ],
    "approaching": [
      8,
      18,
      23,
      60,
      98
    ],
    "dramatically": [
      8,
      24,
      26,
      29,
      38,
      40,
      42,
      47,
      51,
      56,
      58,
      59,
      62,
      69,
      74,
      75,
      92,
      96,
      139
    ],
    "mathematical": [
      8,
      116,
      143
    ],
    "good": [
      8,
      22,
      83,
      86,
      120,
      148
    ],
    "isolated": [
      8,
      104
    ],
    "true": [
      8,
      37,
      68,
      77,
      96
    ],
    "identifiable": [
      8,
      23,
      71
    ],
    "captured": [
      8,
      9,
      99
    ],
    "near": [
      8,
      9,
      17,
      25,
      28,
      33,
      38,
      49,
      53,
      54,
      56,
      62,
      66,
      69,
      73,
      74,
      75,
      81,
      85,
      93,
      100,
      113,
      136,
      157
    ],
    "sovereign": [
      8,
      18,
      22,
      50,
      58,
      73,
      88,
      97,
      98,
      99,
      100,
      116,
      128,
      154
    ],
    "importantly": [
      8,
      67
    ],
    "chinese": [
      8,
      86,
      87
    ],
    "selloff": [
      8,
      17,
      22,
      25,
      26,
      75,
      89
    ],
    "identifying": [
      8,
      40
    ],
    "fewer": [
      8,
      47,
      72,
      140
    ],
    "positives": [
      8,
      31
    ],
    "cost": [
      8,
      21,
      23,
      29,
      38,
      40,
      53,
      63,
      66,
      68,
      69,
      70,
      74,
      75,
      81,
      82,
      86,
      90,
      91,
      93,
      99,
      125,
      137
    ],
    "slightly": [
      8,
      12,
      17,
      50,
      68
    ],
    "detection": [
      8,
      44
    ],
    "confidence": [
      8,
      10,
      12,
      19,
      20,
      22,
      29,
      44,
      45,
      58,
      61,
      73,
      78,
      81,
      83,
      84,
      86,
      87,
      91,
      92,
      96,
      118
    ],
    "statistical": [
      8,
      23
    ],
    "approach": [
      9,
      10,
      24,
      37,
      38,
      39,
      40,
      76,
      145
    ],
    "instead": [
      9,
      38,
      47,
      63,
      65,
      142
    ],
    "measuring": [
      9,
      13,
      127
    ],
    "health": [
      9,
      43,
      76,
      92
    ],
    "directly": [
      9,
      16,
      27,
      29,
      30,
      32,
      39,
      45,
      60,
      63,
      65,
      68,
      69,
      75,
      81,
      127,
      139,
      141,
      142
    ],
    "distance": [
      9,
      23,
      65,
      124
    ],
    "holding": [
      9,
      60,
      67,
      68,
      70,
      79,
      145,
      153
    ],
    "liabilities": [
      9,
      10,
      70,
      77,
      94,
      98
    ],
    "exceed": [
      9,
      11,
      37,
      48,
      55,
      58,
      158
    ],
    "cushion": [
      9,
      49,
      60,
      62,
      63,
      65,
      97,
      144
    ],
    "substantially": [
      9,
      13,
      18,
      19,
      55,
      56,
      62,
      77,
      157
    ],
    "push": [
      9,
      16,
      34,
      54
    ],
    "toward": [
      9,
      21,
      26,
      55,
      58,
      61,
      81,
      87,
      89,
      112
    ],
    "insolvency": [
      9,
      94,
      98
    ],
    "operating": [
      9,
      33,
      45,
      69,
      93,
      95,
      121,
      146,
      153,
      154,
      155,
      156
    ],
    "edge": [
      9,
      34,
      43
    ],
    "dangerously": [
      9,
      98
    ],
    "close": [
      9,
      24,
      79,
      153
    ],
    "collective": [
      9,
      11,
      65,
      67,
      77,
      88
    ],
    "briefly": [
      9,
      51,
      54,
      60,
      97
    ],
    "approached": [
      9,
      24
    ],
    "healthy": [
      9,
      23,
      27,
      36,
      49,
      56,
      78,
      88
    ],
    "substantial": [
      9,
      55,
      65,
      95
    ],
    "buffers": [
      9,
      13,
      70,
      72,
      97
    ],
    "uniquely": [
      9,
      94,
      95,
      96,
      98
    ],
    "sheet": [
      9,
      43,
      48,
      50,
      59,
      63,
      91,
      96,
      97
    ],
    "performing": [
      9,
      39,
      55,
      72,
      115
    ],
    "declining": [
      9,
      16,
      27,
      64,
      89,
      92,
      131
    ],
    "eroding": [
      9,
      22,
      91
    ],
    "capital": [
      9,
      10,
      13,
      15,
      18,
      20,
      22,
      23,
      24,
      29,
      30,
      35,
      39,
      40,
      41,
      45,
      46,
      47,
      51,
      55,
      59,
      61,
      65,
      70,
      72,
      73,
      75,
      76,
      77,
      88,
      89,
      90,
      94,
      98,
      110,
      115,
      119,
      128,
      141,
      145,
      154,
      155,
      156,
      157,
      158,
      161
    ],
    "wider": [
      9,
      34,
      41
    ],
    "provided": [
      9,
      35,
      63,
      72,
      95,
      96,
      97,
      102
    ],
    "meaningful": [
      9,
      44,
      54,
      61,
      62,
      68,
      157
    ],
    "warning": [
      9,
      13,
      15,
      20,
      33,
      50,
      64,
      77,
      88,
      93,
      94,
      95,
      96,
      97,
      99,
      100,
      131
    ],
    "late": [
      9,
      14,
      15,
      18,
      19,
      27,
      30,
      37,
      40,
      48,
      49,
      51,
      56,
      60,
      61,
      64,
      66,
      67,
      78,
      81,
      83,
      84,
      87,
      88,
      89,
      90,
      92,
      94,
      95,
      96,
      119,
      144,
      158,
      160
    ],
    "several": [
      9,
      10,
      12,
      15,
      17,
      18,
      20,
      21,
      22,
      30,
      37,
      38,
      39,
      45,
      50,
      57,
      58,
      67,
      79,
      85,
      86,
      115
    ],
    "months": [
      9,
      11,
      17,
      25,
      29,
      42,
      47,
      48,
      49,
      50,
      58,
      64,
      66,
      75,
      79,
      82,
      85,
      86,
      91,
      93,
      94,
      95,
      96,
      97,
      98,
      122,
      123,
      131
    ],
    "acute": [
      9,
      17,
      53,
      55,
      60,
      85,
      94,
      96,
      100,
      127
    ],
    "phase": [
      9,
      28,
      48,
      49,
      75,
      78,
      95,
      96,
      120
    ],
    "signature": [
      9,
      16,
      23,
      61
    ],
    "republic": [
      9,
      23,
      61,
      134,
      138
    ],
    "combined": [
      9,
      14,
      16,
      23,
      32,
      41,
      61,
      90,
      94,
      95,
      97,
      98,
      143,
      158
    ],
    "monitor": [
      10,
      20,
      33,
      37,
      66,
      74,
      87,
      161
    ],
    "channel": [
      10,
      12,
      18,
      29,
      100
    ],
    "fail": [
      10,
      19,
      26,
      41,
      116
    ],
    "multi": [
      10,
      34,
      37,
      68,
      72,
      91,
      98,
      111,
      123
    ],
    "dimensional": [
      10,
      34,
      41,
      44
    ],
    "sources": [
      10,
      39,
      42,
      44,
      144,
      156
    ],
    "mechanisms": [
      10,
      33,
      36,
      72,
      115
    ],
    "metric": [
      10,
      92,
      141
    ],
    "adequacy": [
      10,
      22,
      70
    ],
    "ratio": [
      10,
      13,
      14,
      16,
      19,
      44,
      50,
      51,
      54,
      56,
      59,
      61,
      65,
      69,
      78,
      95,
      121
    ],
    "buffer": [
      10,
      13,
      21,
      62,
      90,
      123,
      129,
      144,
      156
    ],
    "absorb": [
      10,
      14,
      16,
      20,
      21,
      34,
      79,
      93,
      126,
      151
    ],
    "losses": [
      10,
      16,
      23,
      25,
      26,
      27,
      28,
      29,
      31,
      34,
      38,
      39,
      45,
      59,
      61,
      73,
      75,
      76,
      79,
      94,
      95,
      96,
      131,
      136,
      143,
      154
    ],
    "dimension": [
      10,
      53,
      64,
      66,
      70
    ],
    "exposure": [
      10,
      15,
      23,
      30,
      36,
      37,
      38,
      40,
      43,
      45,
      46,
      55,
      68,
      71,
      96,
      141,
      158
    ],
    "mark": [
      10,
      38,
      73,
      154
    ],
    "mismatch": [
      10,
      29,
      68,
      70,
      94,
      96
    ],
    "liquid": [
      10,
      26,
      70,
      74
    ],
    "fund": [
      10,
      22,
      23,
      29,
      37,
      39,
      45,
      57,
      58,
      65,
      68,
      73,
      77,
      95,
      98,
      117,
      118,
      151,
      158,
      161
    ],
    "wholesale": [
      10,
      114
    ],
    "vulnerable": [
      10,
      14,
      26,
      39,
      45,
      54,
      55,
      61,
      69,
      95,
      126,
      136
    ],
    "proportion": [
      10,
      61,
      74
    ],
    "comes": [
      10,
      13,
      16,
      50,
      56,
      65,
      68,
      82,
      115
    ],
    "uninsured": [
      10,
      61
    ],
    "deposits": [
      10,
      23,
      61,
      77,
      151
    ],
    "runnable": [
      10,
      61
    ],
    "speed": [
      10,
      17,
      25,
      27,
      30,
      41,
      42,
      47,
      52,
      60,
      65,
      66,
      71,
      85,
      97,
      98,
      99
    ],
    "loss": [
      10,
      16,
      22,
      55,
      61,
      72,
      73,
      76,
      96,
      115,
      138,
      140,
      143,
      151,
      160
    ],
    "drain": [
      10,
      97
    ],
    "reported": [
      10,
      12,
      25,
      38,
      44,
      73,
      83,
      98,
      143
    ],
    "gained": [
      10,
      37,
      38
    ],
    "credibility": [
      10,
      22,
      62
    ],
    "quarters": [
      10,
      19,
      37,
      47
    ],
    "unrealized": [
      10,
      23,
      61
    ],
    "increasing": [
      10,
      16,
      43,
      55,
      66,
      74,
      77,
      92,
      95,
      100
    ],
    "incentive": [
      10,
      25,
      73,
      75,
      82,
      91,
      94,
      120
    ],
    "depositors": [
      10,
      61,
      72
    ],
    "yielding": [
      10,
      25,
      70,
      75
    ],
    "alternatives": [
      10,
      98,
      99,
      122
    ],
    "combination": [
      10,
      15,
      16,
      24,
      48,
      53,
      54,
      55,
      61,
      64,
      66,
      69,
      78,
      82,
      84
    ],
    "loop": [
      10,
      27,
      33,
      58,
      61,
      76,
      91,
      93,
      94,
      95,
      98,
      130
    ],
    "accelerate": [
      10,
      39,
      71,
      81,
      93
    ],
    "featured": [
      10,
      24,
      29
    ],
    "deadly": [
      10,
      61
    ],
    "explicitly": [
      10,
      37,
      59,
      80,
      81,
      95,
      157
    ],
    "theoretically": [
      10,
      63,
      93
    ],
    "institutional": [
      10,
      17,
      28,
      37,
      39,
      44,
      50,
      51,
      53,
      57,
      63,
      64,
      69,
      71,
      73,
      75,
      78,
      88,
      97,
      115,
      116,
      118,
      119,
      128,
      154,
      157
    ],
    "assessments": [
      10,
      20,
      37,
      94
    ],
    "published": [
      10,
      16,
      37,
      38,
      39,
      40,
      43,
      50,
      79,
      80,
      95
    ],
    "twice": [
      10,
      34
    ],
    "confidential": [
      10,
      43
    ],
    "supervisory": [
      10,
      43
    ],
    "call": [
      10,
      19,
      23,
      62,
      72,
      94,
      98,
      116
    ],
    "filings": [
      10,
      27,
      31,
      37,
      39
    ],
    "curve": [
      11,
      12,
      34,
      36,
      41,
      48,
      49,
      58,
      60,
      62,
      63,
      64,
      67,
      75,
      91,
      93,
      95,
      96,
      97,
      140
    ],
    "inversion": [
      11,
      12,
      17,
      41,
      48,
      49,
      63,
      64,
      93,
      96
    ],
    "reliable": [
      11,
      12,
      50,
      64,
      68,
      93,
      100
    ],
    "toolkit": [
      11,
      37,
      50
    ],
    "evidence": [
      11,
      58,
      140
    ],
    "independent": [
      11,
      73,
      158
    ],
    "turns": [
      11,
      13,
      33,
      70
    ],
    "inverts": [
      11,
      17,
      38,
      53,
      64
    ],
    "meaning": [
      11,
      13,
      28,
      33,
      48,
      50,
      54,
      57,
      59,
      68,
      71,
      76,
      79,
      86,
      131
    ],
    "expects": [
      11,
      17,
      40,
      67
    ],
    "weakness": [
      11,
      21,
      48,
      49,
      63,
      80,
      84,
      87,
      90,
      91,
      92,
      94,
      97
    ],
    "normally": [
      11,
      52,
      76,
      81,
      97,
      98
    ],
    "bonds": [
      11,
      23,
      25,
      26,
      37,
      54,
      58,
      61,
      63,
      68,
      69,
      70,
      72,
      73,
      97,
      118,
      158
    ],
    "shorter": [
      11,
      15,
      16,
      17,
      40,
      49,
      53,
      139
    ],
    "compensate": [
      11,
      47
    ],
    "investors": [
      11,
      17,
      19,
      26,
      27,
      29,
      31,
      33,
      36,
      37,
      38,
      40,
      42,
      44,
      46,
      51,
      53,
      63,
      64,
      65,
      69,
      71,
      73,
      74,
      75,
      78,
      88,
      89,
      90,
      94,
      95,
      97,
      116,
      151,
      155,
      157
    ],
    "duration": [
      11,
      20,
      23,
      61,
      70,
      77,
      89,
      98
    ],
    "premium": [
      11,
      12,
      17,
      22,
      53,
      58,
      64,
      77,
      154
    ],
    "willing": [
      11,
      77
    ],
    "yields": [
      11,
      48,
      49,
      58,
      61,
      72,
      75,
      98,
      135,
      136,
      140,
      141,
      142
    ],
    "lock": [
      11,
      56,
      91,
      158
    ],
    "remain": [
      11,
      14,
      48,
      49,
      51,
      52,
      61,
      76,
      79,
      85,
      86,
      92,
      143
    ],
    "assessment": [
      11,
      18,
      20,
      40,
      48,
      51,
      98,
      100
    ],
    "record": [
      11,
      37,
      40,
      48,
      51,
      73,
      88,
      97,
      133,
      136,
      151,
      161
    ],
    "remarkable": [
      11,
      50,
      92
    ],
    "inverted": [
      11,
      12,
      17,
      34,
      36,
      48,
      49,
      53,
      63,
      64,
      93,
      95,
      96,
      97
    ],
    "lead": [
      11,
      12,
      13,
      15,
      16,
      17,
      58,
      93,
      94,
      95,
      96,
      97,
      98,
      99,
      100,
      131
    ],
    "critically": [
      11,
      27,
      49,
      59,
      65,
      66,
      84,
      94,
      95,
      97
    ],
    "produced": [
      11,
      20,
      33,
      39,
      40,
      61,
      93
    ],
    "slowdown": [
      11,
      54,
      57,
      84,
      87,
      92
    ],
    "official": [
      11,
      42,
      44,
      87
    ],
    "gives": [
      11,
      56,
      62,
      100
    ],
    "depending": [
      11,
      13,
      70
    ],
    "episode": [
      11,
      69,
      70,
      75
    ],
    "shown": [
      11,
      43,
      44
    ],
    "environments": [
      11,
      37,
      39
    ],
    "bound": [
      11,
      62,
      117,
      152
    ],
    "york": [
      11,
      42
    ],
    "maintains": [
      11,
      56
    ],
    "says": [
      11,
      41
    ],
    "coming": [
      11,
      17,
      40,
      54,
      151
    ],
    "mild": [
      11,
      115
    ],
    "precede": [
      11,
      41,
      43,
      44
    ],
    "vice": [
      11,
      73
    ],
    "account": [
      11,
      22,
      29,
      57,
      58,
      74,
      94,
      99,
      117
    ],
    "recessionary": [
      11,
      79
    ],
    "crashes": [
      11,
      31,
      38,
      39,
      40,
      46,
      115,
      145
    ],
    "coincide": [
      11,
      32
    ],
    "remained": [
      11,
      23,
      31,
      56,
      82,
      83,
      85,
      89,
      94,
      95,
      97,
      108,
      109,
      110,
      113,
      114,
      159
    ],
    "longest": [
      11,
      13,
      48,
      82,
      96,
      131
    ],
    "creating": [
      11,
      21,
      24,
      27,
      32,
      36,
      41,
      54,
      68,
      71,
      73,
      74,
      75,
      77,
      79,
      85,
      91,
      95,
      97,
      100,
      151
    ],
    "ongoing": [
      11,
      20,
      32,
      87,
      91
    ],
    "test": [
      11,
      41,
      54
    ],
    "predictive": [
      11,
      12,
      16,
      41
    ],
    "media": [
      12,
      44,
      45,
      102,
      104,
      111,
      115,
      116
    ],
    "measure": [
      12,
      14,
      22,
      25,
      26,
      44,
      49,
      50,
      52,
      78,
      81,
      95
    ],
    "fact": [
      12,
      30,
      78,
      92
    ],
    "considered": [
      12,
      51,
      56,
      58,
      60,
      69,
      85,
      98
    ],
    "proxy": [
      12,
      86
    ],
    "expected": [
      12,
      15,
      24,
      32,
      40,
      45,
      49,
      51,
      56,
      58,
      62,
      63,
      67,
      91
    ],
    "outperforms": [
      12,
      15
    ],
    "prediction": [
      12,
      37,
      38,
      40,
      41,
      43
    ],
    "difference": [
      12,
      17,
      34,
      42,
      47,
      59,
      69,
      96
    ],
    "modest": [
      12,
      25,
      33,
      42,
      61,
      75,
      78,
      97
    ],
    "advantage": [
      12,
      41,
      42
    ],
    "tends": [
      12,
      83,
      87,
      89
    ],
    "earlier": [
      12,
      30,
      89,
      97
    ],
    "responds": [
      12,
      151
    ],
    "changes": [
      12,
      16,
      20,
      21,
      25,
      31,
      39,
      51
    ],
    "generates": [
      12,
      25,
      39,
      68,
      129
    ],
    "behavioral": [
      12,
      15,
      19,
      76
    ],
    "tighten": [
      12,
      21,
      22,
      45,
      67,
      98
    ],
    "declare": [
      12,
      73
    ],
    "businesses": [
      12,
      18,
      23,
      57,
      59,
      60,
      66,
      82,
      85,
      87,
      156
    ],
    "cautious": [
      12,
      62
    ],
    "hiring": [
      12,
      79,
      82,
      85
    ],
    "reduce": [
      12,
      15,
      21,
      26,
      33,
      37,
      40,
      46,
      47,
      58,
      71,
      72,
      87,
      89,
      97,
      158
    ],
    "standards": [
      12,
      18,
      19,
      28,
      32,
      53,
      78,
      92
    ],
    "partial": [
      12,
      41,
      60
    ],
    "fulfilling": [
      12,
      22,
      84
    ],
    "beyond": [
      12,
      70,
      94,
      95,
      132,
      134,
      139
    ],
    "pure": [
      12,
      37,
      38,
      41,
      156,
      157
    ],
    "tracked": [
      12,
      28,
      41,
      45,
      66,
      96
    ],
    "reflect": [
      12,
      81
    ],
    "distortions": [
      12,
      51
    ],
    "genuine": [
      12,
      39,
      86,
      89,
      98
    ],
    "equally": [
      12,
      63
    ],
    "deeply": [
      12,
      22,
      40,
      96
    ],
    "imminent": [
      12,
      17,
      53,
      61,
      63,
      64,
      102,
      110,
      116,
      117
    ],
    "pricing": [
      12,
      24,
      31,
      49,
      53,
      67,
      73,
      88,
      93,
      95,
      96,
      97,
      98,
      126,
      137,
      141,
      151
    ],
    "emergency": [
      12,
      28,
      54,
      72,
      77,
      95,
      96,
      97,
      98,
      138,
      163
    ],
    "basel": [
      13,
      18
    ],
    "reference": [
      13,
      18,
      28,
      159
    ],
    "committee": [
      13,
      98
    ],
    "embedded": [
      13,
      21,
      64,
      98
    ],
    "regulation": [
      13,
      99
    ],
    "private": [
      13,
      16,
      18,
      23,
      28,
      50,
      58,
      70,
      73,
      77,
      96,
      131,
      151
    ],
    "growing": [
      13,
      16,
      44,
      50,
      52,
      65,
      93
    ],
    "becoming": [
      13,
      37,
      77
    ],
    "signaling": [
      13,
      56,
      60,
      94,
      96
    ],
    "must": [
      13,
      16,
      20,
      22,
      24,
      26,
      30,
      32,
      33,
      42,
      58,
      61,
      67,
      68,
      69,
      70,
      71,
      73,
      74,
      76,
      85,
      137,
      147
    ],
    "eventually": [
      13,
      17,
      18,
      27,
      29,
      32,
      94
    ],
    "reverse": [
      13,
      19,
      22,
      39,
      47,
      61,
      97
    ],
    "gradually": [
      13,
      58,
      62,
      66,
      67,
      85
    ],
    "escalating": [
      13,
      88,
      90
    ],
    "reversal": [
      13,
      47,
      82,
      151
    ],
    "range": [
      13,
      15,
      25,
      40,
      41,
      59,
      62,
      85,
      86,
      88
    ],
    "potential": [
      13,
      20,
      22,
      36,
      41,
      51,
      52,
      73,
      89,
      98
    ],
    "leading": [
      13,
      18,
      74,
      98,
      102,
      128
    ],
    "allows": [
      13,
      66,
      70,
      115
    ],
    "activate": [
      13,
      78
    ],
    "actually": [
      13,
      15,
      16,
      17,
      19,
      31,
      39,
      49,
      94,
      115,
      120,
      131,
      153,
      157
    ],
    "hits": [
      13,
      21,
      26,
      31,
      59,
      66,
      73,
      122,
      129,
      137,
      146,
      156
    ],
    "practical": [
      13,
      34,
      42,
      58
    ],
    "limitations": [
      13,
      14,
      44
    ],
    "estimated": [
      13,
      24,
      33,
      90,
      158
    ],
    "methodology": [
      13,
      34,
      37,
      40,
      42,
      43,
      154
    ],
    "sample": [
      13,
      41,
      44
    ],
    "produces": [
      13,
      32,
      35,
      60,
      137,
      147
    ],
    "estimates": [
      13,
      18,
      24,
      25,
      26,
      42,
      59,
      68,
      69,
      75,
      156,
      158
    ],
    "differ": [
      13,
      47
    ],
    "additionally": [
      13,
      18,
      149
    ],
    "performs": [
      13,
      162
    ],
    "poorly": [
      13,
      72
    ],
    "experienced": [
      13,
      15,
      16,
      29,
      30,
      75
    ],
    "structural": [
      13,
      24,
      30,
      40,
      51,
      55,
      57,
      63,
      71,
      72,
      74,
      75,
      76,
      77,
      80,
      86,
      87,
      88,
      90,
      92,
      94,
      96,
      97,
      98,
      99,
      110,
      116,
      117,
      128,
      151,
      154,
      158
    ],
    "shifts": [
      13,
      22,
      44,
      67,
      76,
      100
    ],
    "permanently": [
      13,
      55,
      76,
      104,
      112
    ],
    "raises": [
      13,
      16,
      32,
      34,
      129,
      157
    ],
    "despite": [
      13,
      21,
      44,
      49,
      50,
      57,
      78,
      86,
      89,
      90,
      92,
      131,
      137,
      145
    ],
    "alternative": [
      13,
      26,
      44,
      47,
      66,
      73,
      150
    ],
    "comparable": [
      13,
      60
    ],
    "acceptable": [
      13,
      95
    ],
    "accuracy": [
      13,
      41,
      43,
      115,
      130
    ],
    "statistics": [
      13,
      16,
      19,
      25,
      42,
      44,
      77,
      93,
      96,
      158
    ],
    "gaps": [
      13,
      50
    ],
    "shiller": [
      14,
      19,
      51,
      65,
      78,
      95,
      96
    ],
    "cyclically": [
      14,
      51
    ],
    "adjusted": [
      14,
      16,
      40,
      51,
      69,
      93
    ],
    "laureate": [
      14,
      51
    ],
    "robert": [
      14,
      51
    ],
    "smooths": [
      14,
      51,
      85
    ],
    "profits": [
      14,
      25,
      95
    ],
    "misleading": [
      14,
      80
    ],
    "extremes": [
      14,
      57,
      97
    ],
    "cheap": [
      14,
      15,
      33,
      38,
      69,
      118,
      153
    ],
    "expensive": [
      14,
      38,
      44,
      58,
      77,
      90,
      94,
      143,
      149,
      158
    ],
    "crash": [
      14,
      17,
      27,
      31,
      38,
      39,
      40,
      45,
      51,
      53,
      65,
      69,
      71,
      95,
      96,
      97,
      105,
      111,
      117,
      154,
      155,
      160
    ],
    "reaching": [
      14,
      18,
      69,
      95
    ],
    "japan": [
      14,
      25,
      58,
      75,
      103,
      163
    ],
    "throughout": [
      14,
      98
    ],
    "until": [
      14,
      16,
      23,
      28,
      66,
      85,
      93,
      98,
      154
    ],
    "poor": [
      14,
      58
    ],
    "timing": [
      14,
      15,
      36,
      37,
      40,
      51,
      116,
      157,
      158
    ],
    "occur": [
      14,
      38,
      40,
      47,
      93,
      140
    ],
    "compound": [
      14,
      20,
      30,
      32,
      33,
      35,
      36,
      45,
      63,
      64,
      65,
      78,
      87,
      91,
      93,
      95,
      96,
      97,
      99,
      100,
      130,
      136,
      143,
      149
    ],
    "tightening": [
      14,
      16,
      18,
      22,
      32,
      33,
      50,
      53,
      60,
      64,
      67,
      77,
      90,
      91,
      93,
      95,
      98
    ],
    "cocktail": [
      14,
      33
    ],
    "valuations": [
      14,
      19,
      27,
      33,
      36,
      41,
      51,
      55,
      62,
      65,
      67,
      69,
      73,
      76,
      95,
      158
    ],
    "error": [
      14,
      62
    ],
    "surprise": [
      14,
      25,
      62,
      71,
      93
    ],
    "triggers": [
      14,
      19,
      20,
      21,
      22,
      23,
      34,
      36,
      43,
      51,
      68,
      74,
      78,
      79,
      100,
      129,
      156
    ],
    "larger": [
      14,
      15,
      25,
      30,
      36,
      37,
      69,
      71,
      72,
      76,
      77,
      78
    ],
    "corrections": [
      14,
      115
    ],
    "interpreted": [
      14,
      52,
      85,
      161
    ],
    "fragility": [
      14,
      56,
      57,
      74,
      83,
      94
    ],
    "cause": [
      14,
      35,
      36,
      38,
      48,
      95,
      108,
      111,
      130
    ],
    "outsized": [
      14,
      39
    ],
    "damage": [
      14,
      64,
      66,
      136,
      143,
      157
    ],
    "yale": [
      14,
      51
    ],
    "mispricing": [
      15
    ],
    "measured": [
      15,
      30,
      54,
      57,
      68,
      71,
      79,
      96
    ],
    "dominate": [
      15,
      70
    ],
    "excess": [
      15,
      69,
      108,
      118
    ],
    "condition": [
      15,
      25,
      26,
      77,
      94,
      129
    ],
    "participants": [
      15,
      25,
      28,
      41,
      64,
      74,
      96,
      97,
      98
    ],
    "systematically": [
      15,
      47,
      67,
      95
    ],
    "underestimate": [
      15,
      47,
      116
    ],
    "microstructure": [
      15,
      17,
      97
    ],
    "steadily": [
      15,
      56,
      98,
      112
    ],
    "insurance": [
      15,
      46,
      53,
      66,
      69,
      70,
      71,
      100,
      127
    ],
    "required": [
      15,
      32,
      63,
      68,
      71,
      72,
      76,
      95,
      128
    ],
    "hedging": [
      15,
      38,
      52,
      53,
      71,
      88
    ],
    "needed": [
      15,
      26,
      33,
      49,
      70,
      79,
      98,
      124,
      143,
      149
    ],
    "sellers": [
      15,
      72,
      158
    ],
    "offer": [
      15,
      58
    ],
    "cheaper": [
      15,
      68,
      94
    ],
    "protection": [
      15,
      17,
      37,
      38,
      40,
      46,
      53,
      64,
      73,
      98,
      132,
      143
    ],
    "puts": [
      15,
      38
    ],
    "leveraged": [
      15,
      18,
      19,
      21,
      26,
      28,
      33,
      39,
      65,
      68,
      69,
      71,
      73,
      75,
      89,
      93,
      98,
      103
    ],
    "strategies": [
      15,
      26,
      36,
      39,
      71,
      94,
      97,
      98
    ],
    "profit": [
      15,
      38,
      40,
      51,
      71,
      88,
      93,
      160
    ],
    "carry": [
      15,
      25,
      29,
      36,
      39,
      45,
      75,
      86
    ],
    "trades": [
      15,
      25,
      60,
      71,
      75
    ],
    "parity": [
      15,
      26,
      36,
      37,
      97,
      98
    ],
    "attract": [
      15,
      58
    ],
    "result": [
      15,
      39,
      49,
      59,
      68,
      70,
      74,
      93,
      114,
      128
    ],
    "appears": [
      15,
      39,
      55
    ],
    "positioned": [
      15,
      53,
      64
    ],
    "unlikely": [
      15,
      76,
      80
    ],
    "placing": [
      15,
      48
    ],
    "strategic": [
      15,
      17,
      21,
      88
    ],
    "precise": [
      15,
      59,
      71,
      74,
      95,
      96
    ],
    "flagged": [
      15,
      23,
      59,
      68,
      154
    ],
    "ahead": [
      15,
      48,
      93,
      148
    ],
    "related": [
      15,
      27,
      31
    ],
    "stocks": [
      15,
      19,
      26,
      31,
      34,
      36,
      37,
      39,
      40,
      57,
      63,
      65,
      74,
      84,
      97,
      136,
      137,
      138,
      139,
      146,
      157,
      158
    ],
    "unusually": [
      15,
      16,
      39,
      57,
      80
    ],
    "exactly": [
      15,
      48,
      51,
      70,
      103,
      107
    ],
    "gains": [
      15,
      26,
      40,
      72,
      78
    ],
    "small": [
      15,
      19,
      25,
      38,
      41,
      42,
      47,
      57,
      73,
      138,
      155
    ],
    "mega": [
      15,
      57,
      65,
      74,
      78,
      158
    ],
    "tech": [
      15,
      24,
      31,
      33,
      36,
      47,
      51,
      57,
      64,
      78,
      95,
      103,
      105,
      155,
      156
    ],
    "magnificent": [
      15,
      19,
      36,
      57,
      65,
      74,
      78
    ],
    "predicted": [
      15,
      26,
      101,
      102,
      103,
      104,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115,
      119,
      120
    ],
    "realized": [
      15,
      26,
      39,
      71,
      93,
      117,
      118
    ],
    "share": [
      16,
      19,
      27,
      55,
      61,
      74,
      95
    ],
    "devoted": [
      16,
      56
    ],
    "servicing": [
      16,
      54,
      56,
      90
    ],
    "stretched": [
      16,
      91,
      96,
      148
    ],
    "burden": [
      16,
      22,
      58,
      94
    ],
    "sensitive": [
      16,
      21,
      34,
      72,
      83,
      131
    ],
    "increase": [
      16,
      42,
      54,
      66,
      68,
      69,
      75,
      84,
      86,
      91,
      93,
      96,
      97,
      98,
      145
    ],
    "simple": [
      16,
      68,
      74
    ],
    "declines": [
      16,
      24,
      26,
      30,
      36,
      44,
      53,
      96
    ],
    "marginal": [
      16,
      34,
      58,
      137
    ],
    "sensitivity": [
      16,
      45,
      117
    ],
    "hiking": [
      16,
      22,
      53,
      61,
      112
    ],
    "mechanically": [
      16,
      36,
      71,
      97
    ],
    "increased": [
      16,
      17,
      20,
      29,
      30,
      73,
      80,
      94,
      98,
      148
    ],
    "operates": [
      16,
      17,
      21,
      33,
      46,
      66,
      76,
      77,
      84,
      86,
      100
    ],
    "existing": [
      16,
      20,
      24,
      30,
      33,
      56,
      85,
      91,
      92,
      97,
      100,
      130
    ],
    "unaffected": [
      16,
      35,
      102
    ],
    "mature": [
      16,
      69,
      76
    ],
    "refinanced": [
      16,
      54,
      56
    ],
    "example": [
      16,
      19,
      29,
      32,
      41,
      43,
      93,
      98
    ],
    "rolling": [
      16,
      26,
      69,
      138
    ],
    "mortgages": [
      16,
      56,
      70,
      91,
      96,
      107
    ],
    "dominance": [
      16,
      51
    ],
    "transmission": [
      16,
      21,
      23,
      28,
      30,
      45,
      66,
      84,
      91,
      122
    ],
    "household": [
      16,
      46,
      56,
      63,
      84,
      96
    ],
    "relies": [
      16,
      68,
      115
    ],
    "floating": [
      16,
      69,
      73,
      96
    ],
    "powerful": [
      16,
      65,
      114
    ],
    "projections": [
      16,
      27,
      58,
      59,
      99,
      123,
      156,
      158
    ],
    "stagnant": [
      16,
      103
    ],
    "precursor": [
      16,
      132,
      139
    ],
    "continued": [
      16,
      22,
      57,
      59,
      75,
      88,
      91,
      93,
      94,
      95,
      98
    ],
    "canonical": [
      16,
      29,
      32,
      93,
      95,
      98
    ],
    "relationship": [
      17,
      26,
      34,
      37,
      43,
      58,
      73,
      89,
      140
    ],
    "days": [
      17,
      25,
      42,
      44,
      52,
      66,
      67,
      71,
      72,
      74,
      77,
      97,
      98,
      111,
      146,
      148
    ],
    "dated": [
      17,
      53,
      68
    ],
    "trade": [
      17,
      20,
      25,
      29,
      30,
      36,
      39,
      53,
      68,
      71,
      73,
      75,
      77,
      82,
      86,
      90,
      96,
      98,
      100,
      118,
      141,
      151,
      152,
      153,
      155,
      159,
      160
    ],
    "reflecting": [
      17,
      30,
      39,
      53,
      62,
      67,
      77,
      80,
      82,
      86,
      87,
      91,
      93
    ],
    "uncertainty": [
      17,
      20,
      30,
      44,
      52,
      53,
      69,
      77,
      78,
      82,
      86,
      90,
      91
    ],
    "distant": [
      17,
      64
    ],
    "horizons": [
      17,
      53,
      94
    ],
    "baseline": [
      17,
      52,
      127,
      141
    ],
    "expectation": [
      17,
      67
    ],
    "generated": [
      17,
      24,
      27,
      59
    ],
    "need": [
      17,
      38,
      46,
      54,
      77,
      141,
      156,
      157
    ],
    "intense": [
      17,
      93
    ],
    "inverting": [
      17,
      60
    ],
    "fear": [
      17,
      44,
      95
    ],
    "urgency": [
      17,
      42,
      56
    ],
    "happen": [
      17,
      20,
      31,
      64
    ],
    "relatively": [
      17,
      25,
      38,
      46,
      57,
      72,
      78,
      85
    ],
    "spends": [
      17,
      53
    ],
    "vast": [
      17,
      69,
      72,
      77,
      90
    ],
    "majority": [
      17,
      18,
      19,
      56,
      69,
      72,
      95
    ],
    "persists": [
      17,
      25
    ],
    "coincided": [
      17,
      60
    ],
    "october": [
      17,
      27,
      28,
      61,
      75,
      85,
      93,
      94,
      96,
      98
    ],
    "august": [
      17,
      75,
      77,
      83,
      93,
      96,
      97,
      98
    ],
    "february": [
      17,
      60,
      66,
      71,
      79,
      80,
      82,
      83,
      84,
      86,
      87,
      94,
      95,
      96,
      97,
      144,
      154
    ],
    "volmageddon": [
      17,
      71
    ],
    "operate": [
      17,
      72,
      92,
      141
    ],
    "information": [
      17,
      21,
      30,
      44,
      71,
      88,
      155
    ],
    "happening": [
      17,
      30
    ],
    "right": [
      17,
      59,
      115,
      117,
      118,
      119,
      120,
      154
    ],
    "serves": [
      17,
      74,
      86
    ],
    "coal": [
      17,
      86,
      139
    ],
    "cboe": [
      17,
      26,
      30,
      38,
      53,
      71
    ],
    "curves": [
      17,
      41,
      64,
      127
    ],
    "providers": [
      17,
      27,
      44,
      57,
      71,
      99
    ],
    "front": [
      17,
      53,
      71,
      82
    ],
    "encompasses": [
      18,
      59
    ],
    "broad": [
      18,
      36,
      46,
      83,
      105
    ],
    "originating": [
      18,
      19,
      20,
      21,
      22,
      23,
      75
    ],
    "excessive": [
      18,
      28
    ],
    "domain": [
      18,
      41
    ],
    "consumer": [
      18,
      21,
      44,
      45,
      46,
      48,
      56,
      60,
      62,
      78,
      81,
      82,
      83,
      84,
      86,
      87,
      91,
      92
    ],
    "accelerating": [
      18,
      67,
      78,
      79,
      81,
      83,
      91,
      160
    ],
    "surveys": [
      18,
      42,
      62
    ],
    "senior": [
      18,
      69,
      96
    ],
    "loan": [
      18,
      27,
      28,
      56,
      68,
      69,
      73,
      76,
      96
    ],
    "opinion": [
      18,
      95,
      158
    ],
    "survey": [
      18,
      22,
      25,
      44,
      79,
      91,
      145,
      149
    ],
    "quality": [
      18,
      26,
      40,
      43,
      57,
      70,
      75,
      77,
      99,
      100
    ],
    "slowly": [
      18,
      19,
      20,
      25,
      47,
      50
    ],
    "worsens": [
      18,
      22,
      99,
      140
    ],
    "tipping": [
      18,
      60
    ],
    "external": [
      18,
      19,
      76,
      130
    ],
    "accounting": [
      18,
      73,
      154,
      157
    ],
    "causes": [
      18,
      20,
      35,
      77,
      85,
      137,
      139,
      151
    ],
    "reassess": [
      18,
      36
    ],
    "freeze": [
      18,
      33,
      44,
      45,
      46,
      47,
      52,
      59,
      65,
      77,
      96
    ],
    "issuance": [
      18,
      19,
      33,
      54,
      58,
      95
    ],
    "depend": [
      18,
      47
    ],
    "maturing": [
      18,
      61,
      67,
      76
    ],
    "find": [
      18,
      24,
      36,
      117
    ],
    "affordable": [
      18,
      136
    ],
    "whose": [
      18,
      19,
      43,
      69,
      98
    ],
    "viable": [
      18,
      45,
      67
    ],
    "entirely": [
      18,
      58,
      81,
      95,
      96,
      97,
      120,
      138
    ],
    "cutting": [
      18,
      43,
      48,
      62,
      64,
      67,
      79,
      81,
      93,
      95,
      138,
      140,
      150,
      156
    ],
    "causing": [
      18,
      25,
      26,
      32,
      106,
      109
    ],
    "sharp": [
      18,
      25,
      39,
      44,
      91
    ],
    "rated": [
      18,
      69,
      96,
      97
    ],
    "lowest": [
      18,
      69,
      97,
      138,
      147
    ],
    "portion": [
      18,
      71
    ],
    "issued": [
      18,
      54,
      69,
      72
    ],
    "maturity": [
      18,
      23,
      43,
      54,
      61,
      67,
      68,
      69,
      70,
      76,
      94,
      96
    ],
    "wall": [
      18,
      23,
      54,
      69,
      70,
      71,
      76,
      146
    ],
    "concrete": [
      18,
      19,
      23,
      128
    ],
    "grown": [
      18,
      73,
      77
    ],
    "enormously": [
      18,
      31,
      38,
      42,
      59,
      75,
      84,
      94
    ],
    "opacity": [
      18,
      25,
      73
    ],
    "difficult": [
      18,
      20,
      41,
      75,
      120
    ],
    "publicly": [
      18,
      23,
      37,
      40,
      42,
      43,
      67,
      69,
      120,
      154,
      156
    ],
    "traded": [
      18,
      69,
      71,
      74
    ],
    "moody": [
      18,
      69,
      96,
      160
    ],
    "bubble": [
      19,
      20,
      24,
      27,
      28,
      33,
      36,
      40,
      44,
      45,
      51,
      65,
      94,
      95,
      103,
      105,
      117,
      118,
      157
    ],
    "encompass": [
      19,
      20,
      21,
      22,
      23
    ],
    "deflation": [
      19,
      33,
      36,
      45,
      58,
      65,
      103
    ],
    "overvalued": [
      19,
      33,
      51,
      95
    ],
    "equities": [
      19,
      25,
      26,
      30,
      37,
      72,
      75,
      89,
      95,
      97,
      98
    ],
    "estate": [
      19,
      23,
      33,
      45,
      55,
      56,
      67,
      76,
      94,
      103
    ],
    "classes": [
      19,
      26,
      40,
      73
    ],
    "fundamental": [
      19,
      26,
      36,
      40,
      41,
      65,
      72,
      89,
      92,
      94,
      97,
      98
    ],
    "danger": [
      19,
      50,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      63,
      64,
      65,
      68,
      94,
      124
    ],
    "unwinding": [
      19,
      25,
      36,
      75
    ],
    "collapse": [
      19,
      27,
      28,
      29,
      42,
      44,
      45,
      70,
      71,
      74,
      77,
      84,
      86,
      92,
      94,
      96,
      98,
      101,
      104,
      106,
      107,
      109,
      115,
      117,
      118,
      119,
      120,
      123,
      126,
      127
    ],
    "norms": [
      19,
      34,
      77,
      91,
      112
    ],
    "weight": [
      19,
      36,
      65,
      74,
      80,
      94,
      158
    ],
    "exceeded": [
      19,
      24,
      31,
      55,
      94,
      95,
      125,
      160
    ],
    "seen": [
      19,
      79,
      84,
      86
    ],
    "activity": [
      19,
      20,
      23,
      32,
      44,
      76,
      77,
      82,
      83,
      86,
      92,
      99,
      100
    ],
    "form": [
      19,
      40,
      43,
      73,
      93
    ],
    "drive": [
      19,
      57
    ],
    "appear": [
      19,
      29,
      39,
      71
    ],
    "diversified": [
      19,
      26,
      39,
      57,
      74
    ],
    "bets": [
      19,
      37,
      71
    ],
    "handful": [
      19,
      35,
      69,
      72
    ],
    "news": [
      19,
      20,
      30,
      44,
      83,
      88,
      97,
      160
    ],
    "company": [
      19,
      24,
      31,
      35,
      41,
      43,
      57,
      70,
      73,
      74,
      94,
      158
    ],
    "disappointment": [
      19,
      36,
      65,
      78
    ],
    "deliver": [
      19,
      62
    ],
    "justified": [
      19,
      24,
      45,
      74
    ],
    "catalyst": [
      19,
      51,
      95,
      99
    ],
    "fraud": [
      19,
      154,
      157
    ],
    "panic": [
      19,
      28,
      31,
      52
    ],
    "disappoints": [
      19,
      65
    ],
    "forces": [
      19,
      57,
      61,
      73,
      76,
      88,
      90,
      92
    ],
    "multiples": [
      19,
      65,
      74
    ],
    "international": [
      20,
      21,
      22,
      29,
      50,
      75,
      77,
      90,
      92,
      94,
      100,
      128,
      129,
      137,
      141
    ],
    "conflicts": [
      20,
      46
    ],
    "wars": [
      20,
      151
    ],
    "regime": [
      20,
      39,
      47,
      54,
      75,
      88,
      91,
      93,
      126
    ],
    "disrupt": [
      20,
      21
    ],
    "little": [
      20,
      72,
      73
    ],
    "forecast": [
      20,
      40
    ],
    "factors": [
      20,
      28,
      31,
      39,
      46,
      62,
      80,
      82,
      86
    ],
    "geographic": [
      20,
      76,
      100
    ],
    "lanes": [
      20,
      152
    ],
    "fabs": [
      20,
      35
    ],
    "commodity": [
      20,
      21,
      46,
      72,
      87,
      92,
      93,
      97,
      127,
      128,
      129,
      144,
      151
    ],
    "especially": [
      20,
      34
    ],
    "caldara": [
      20,
      44
    ],
    "iacoviello": [
      20,
      44
    ],
    "threats": [
      20,
      115
    ],
    "text": [
      20,
      44
    ],
    "reduced": [
      20,
      31,
      55,
      63,
      76,
      87,
      136,
      140,
      142,
      156
    ],
    "havens": [
      20,
      26,
      89,
      162
    ],
    "landscape": [
      20,
      28,
      30,
      50,
      56,
      63,
      66,
      67,
      77
    ],
    "active": [
      20,
      37,
      53,
      63,
      64,
      77,
      99,
      152,
      158
    ],
    "competition": [
      20,
      27,
      61,
      94
    ],
    "export": [
      20,
      87,
      94,
      133,
      137,
      138,
      153,
      163
    ],
    "controls": [
      20,
      35,
      110,
      153
    ],
    "chains": [
      20,
      87,
      97,
      132
    ],
    "friend": [
      20,
      95
    ],
    "tariff": [
      20,
      53,
      78,
      82,
      84,
      85,
      87,
      151,
      155,
      159,
      160
    ],
    "currencies": [
      20,
      25,
      29,
      30,
      72,
      75,
      88,
      90
    ],
    "sold": [
      20,
      26,
      30,
      31,
      70,
      97
    ],
    "haven": [
      20,
      26,
      30,
      88,
      89,
      90,
      97,
      161,
      162
    ],
    "rallied": [
      20,
      30,
      117
    ],
    "hours": [
      20,
      25,
      44,
      98
    ],
    "vulnerabilities": [
      20,
      28,
      29,
      30,
      96,
      130
    ],
    "reverses": [
      20,
      24,
      45,
      47,
      158
    ],
    "integrated": [
      20,
      44
    ],
    "geopolitically": [
      20,
      46
    ],
    "announcements": [
      20,
      97,
      99
    ],
    "military": [
      20,
      60,
      63,
      66,
      127
    ],
    "intelligence": [
      20,
      24,
      100
    ],
    "open": [
      20,
      23,
      24,
      31,
      68,
      71,
      99,
      100,
      128,
      152
    ],
    "source": [
      20,
      24,
      25,
      31,
      70,
      99,
      100,
      154,
      156
    ],
    "disruptions": [
      21,
      35,
      46,
      49,
      95,
      97,
      143
    ],
    "movements": [
      21,
      34,
      44
    ],
    "remains": [
      21,
      22,
      28,
      29,
      55,
      56,
      59,
      61,
      66,
      78,
      80,
      83,
      89,
      90,
      92
    ],
    "virtually": [
      21,
      63,
      65,
      67,
      78,
      93,
      98,
      131
    ],
    "transportation": [
      21,
      32,
      66,
      93
    ],
    "agriculture": [
      21,
      32,
      93,
      121
    ],
    "importing": [
      21,
      138
    ],
    "nations": [
      21,
      93,
      98,
      100,
      123,
      128,
      138
    ],
    "producers": [
      21,
      92,
      125,
      145,
      149
    ],
    "reducing": [
      21,
      30,
      60,
      75,
      86
    ],
    "disposable": [
      21,
      56
    ],
    "raising": [
      21,
      22,
      29,
      75,
      90,
      93,
      95,
      155,
      156
    ],
    "brent": [
      21,
      60,
      66,
      123,
      137
    ],
    "crude": [
      21,
      60,
      66,
      93,
      126,
      133,
      135,
      137
    ],
    "petroleum": [
      21,
      93,
      132,
      133,
      134,
      136,
      137,
      138,
      139,
      143
    ],
    "inventory": [
      21,
      24,
      82,
      95,
      97
    ],
    "intensity": [
      21,
      44,
      153,
      155,
      156,
      157
    ],
    "declined": [
      21,
      28,
      30,
      54,
      55,
      58,
      76,
      79,
      87,
      89,
      92
    ],
    "raise": [
      21,
      72,
      75,
      89,
      97,
      98,
      119,
      154
    ],
    "intensive": [
      21,
      45,
      51,
      80,
      92,
      157
    ],
    "transport": [
      21,
      45,
      121
    ],
    "chemicals": [
      21,
      93,
      139,
      143
    ],
    "compressing": [
      21,
      60
    ],
    "margins": [
      21,
      24,
      40,
      45,
      46,
      51,
      57,
      60,
      69,
      93,
      149
    ],
    "feed": [
      21,
      116,
      133
    ],
    "forcing": [
      21,
      25,
      26,
      67,
      72,
      89
    ],
    "trap": [
      21,
      22,
      24,
      49,
      60,
      62,
      64,
      67,
      77,
      81,
      94,
      116,
      117,
      118,
      119,
      120,
      128
    ],
    "compression": [
      21,
      65,
      69
    ],
    "heating": [
      21,
      32,
      150
    ],
    "discretionary": [
      21,
      44,
      45,
      46,
      60,
      84
    ],
    "hitting": [
      21,
      62,
      65,
      116
    ],
    "retail": [
      21,
      44,
      45,
      71,
      74,
      76,
      88,
      95,
      97,
      121,
      158
    ],
    "entertainment": [
      21,
      44,
      45
    ],
    "shortages": [
      21,
      35,
      106,
      142
    ],
    "physically": [
      21,
      143
    ],
    "responsible": [
      21,
      131
    ],
    "contributed": [
      21,
      68,
      78
    ],
    "embargo": [
      21,
      63,
      93
    ],
    "triggered": [
      21,
      25,
      28,
      60,
      70,
      72,
      75,
      94,
      95,
      97,
      98,
      119
    ],
    "iranian": [
      21,
      60,
      63,
      66,
      152
    ],
    "revolution": [
      21,
      63
    ],
    "ukraine": [
      21,
      32,
      53,
      60,
      88,
      98,
      142
    ],
    "pushed": [
      21,
      52,
      72,
      90,
      93,
      112,
      148
    ],
    "impossible": [
      21,
      22,
      25,
      32,
      63,
      76,
      79,
      93,
      113
    ],
    "choice": [
      21,
      55,
      63,
      73,
      79
    ],
    "fighting": [
      21,
      144
    ],
    "worsen": [
      21,
      34,
      77
    ],
    "supporting": [
      21,
      80,
      93,
      157
    ],
    "spiral": [
      21,
      22,
      32,
      72,
      93
    ],
    "administration": [
      21,
      159
    ],
    "agency": [
      21,
      91
    ],
    "cftc": [
      21,
      25,
      68
    ],
    "commitments": [
      21,
      25,
      154
    ],
    "traders": [
      21,
      25,
      48,
      53,
      62,
      64,
      68,
      71,
      94,
      144,
      151
    ],
    "depreciation": [
      22,
      24,
      154,
      157
    ],
    "persistent": [
      22,
      32,
      38,
      39,
      48,
      49,
      58,
      77,
      80,
      84,
      88
    ],
    "target": [
      22,
      62,
      80,
      81,
      83,
      158
    ],
    "affect": [
      22,
      47,
      95
    ],
    "entrenched": [
      22,
      62,
      63,
      98
    ],
    "strengthening": [
      22,
      25,
      30
    ],
    "tightens": [
      22,
      77
    ],
    "denominated": [
      22,
      29,
      75,
      77,
      90,
      94,
      100,
      128
    ],
    "dollars": [
      22,
      29,
      68,
      71,
      72,
      75,
      76,
      77,
      94,
      97,
      128
    ],
    "targets": [
      22,
      80,
      81,
      125
    ],
    "university": [
      22,
      84,
      130
    ],
    "michigan": [
      22,
      84
    ],
    "nominal": [
      22,
      37,
      54,
      65,
      93
    ],
    "purchasing": [
      22,
      63,
      80,
      84,
      91,
      97
    ],
    "swaps": [
      22,
      72,
      75,
      96,
      98,
      118
    ],
    "additional": [
      22,
      25,
      26,
      33,
      34,
      54,
      62,
      67,
      68,
      76,
      84,
      95
    ],
    "trilemma": [
      22,
      32
    ],
    "choose": [
      22,
      72,
      156
    ],
    "continuing": [
      22,
      85
    ],
    "fight": [
      22,
      79,
      93,
      98,
      116
    ],
    "risking": [
      22,
      79,
      161
    ],
    "deeper": [
      22,
      73
    ],
    "downturn": [
      22,
      45,
      64,
      94
    ],
    "support": [
      22,
      50,
      63,
      75,
      76,
      79,
      81,
      87,
      89,
      92,
      155
    ],
    "acceleration": [
      22,
      55,
      81,
      90
    ],
    "tension": [
      22,
      72,
      98
    ],
    "came": [
      22,
      87,
      96
    ],
    "follow": [
      22,
      46,
      63
    ],
    "inflows": [
      22,
      29,
      74,
      88,
      94,
      117
    ],
    "deficits": [
      22,
      29,
      37,
      49
    ],
    "hikes": [
      22,
      75,
      95,
      113
    ],
    "local": [
      22,
      29,
      90,
      94,
      119
    ],
    "governments": [
      22,
      28,
      119,
      130
    ],
    "borrowed": [
      22,
      29,
      33
    ],
    "difficulties": [
      22,
      36
    ],
    "defending": [
      22,
      29,
      90
    ],
    "crushes": [
      22,
      32,
      141
    ],
    "domestic": [
      22,
      29,
      58,
      75,
      87,
      90,
      94,
      126,
      133,
      137,
      141,
      146
    ],
    "live": [
      22,
      127,
      143
    ],
    "border": [
      22,
      25,
      77,
      90
    ],
    "claims": [
      22,
      42,
      77,
      79,
      85,
      137,
      154
    ],
    "communications": [
      22,
      67,
      138
    ],
    "freezes": [
      23,
      33,
      36,
      77
    ],
    "dysfunction": [
      23,
      45,
      68,
      98
    ],
    "lend": [
      23,
      96
    ],
    "lose": [
      23,
      24,
      79,
      100
    ],
    "access": [
      23,
      77,
      143
    ],
    "sharply": [
      23,
      56,
      84,
      89,
      94,
      98
    ],
    "free": [
      23,
      27,
      29,
      98,
      152
    ],
    "libor": [
      23,
      96
    ],
    "deposit": [
      23,
      61
    ],
    "outflows": [
      23,
      29,
      61,
      74
    ],
    "discount": [
      23,
      67,
      74,
      133
    ],
    "window": [
      23,
      24,
      93,
      123,
      158
    ],
    "usage": [
      23,
      27,
      76,
      77
    ],
    "themselves": [
      23,
      24,
      74,
      96,
      98
    ],
    "hiding": [
      23,
      152
    ],
    "sitting": [
      23,
      146
    ],
    "massive": [
      23,
      24,
      27,
      31,
      42,
      45,
      49,
      50,
      54,
      55,
      59,
      60,
      62,
      63,
      65,
      71,
      81,
      88,
      94,
      95,
      97,
      98,
      101,
      102,
      104,
      106,
      110,
      118,
      119,
      126,
      141,
      156
    ],
    "destroying": [
      23,
      33,
      98
    ],
    "community": [
      23,
      55,
      61,
      76,
      92
    ],
    "vacancy": [
      23,
      55
    ],
    "post": [
      23,
      28,
      33,
      48,
      50,
      51,
      54,
      55,
      72,
      76,
      79,
      82,
      94,
      108,
      112,
      117,
      118,
      119,
      120,
      151,
      154,
      156
    ],
    "pandemic": [
      23,
      35,
      50,
      54,
      55,
      56,
      58,
      62,
      76,
      79
    ],
    "originated": [
      23,
      76,
      141
    ],
    "underwriting": [
      23,
      96
    ],
    "assumptions": [
      23,
      24,
      36,
      41
    ],
    "disproportionately": [
      23,
      57,
      65,
      66,
      74
    ],
    "exposed": [
      23,
      29,
      45,
      47,
      89,
      98,
      100,
      126,
      141
    ],
    "cases": [
      23,
      48,
      49,
      64,
      97
    ],
    "distributed": [
      23,
      68
    ],
    "broadly": [
      23,
      34
    ],
    "fdic": [
      23,
      55,
      61,
      76
    ],
    "profile": [
      23,
      43,
      55,
      61,
      76
    ],
    "release": [
      23,
      71,
      79,
      80,
      81,
      99
    ],
    "costar": [
      23,
      76
    ],
    "overinvestment": [
      24
    ],
    "artificial": [
      24,
      82
    ],
    "cumulative": [
      24,
      62,
      92,
      151,
      154
    ],
    "expenditure": [
      24,
      59,
      65
    ],
    "annually": [
      24,
      31,
      58,
      141
    ],
    "bull": [
      24,
      31,
      49,
      78,
      96,
      101,
      111,
      120
    ],
    "correcting": [
      24,
      34
    ],
    "match": [
      24,
      98
    ],
    "reality": [
      24,
      67,
      73,
      74,
      76,
      80,
      97,
      133
    ],
    "bear": [
      24,
      28,
      31,
      49,
      57,
      78,
      92,
      96,
      97,
      105,
      107,
      154,
      157,
      158
    ],
    "parallels": [
      24,
      27,
      63,
      64,
      71,
      96
    ],
    "previous": [
      24,
      31,
      51,
      138,
      140
    ],
    "bubbles": [
      24,
      31,
      40
    ],
    "striking": [
      24,
      27,
      65
    ],
    "vendor": [
      24,
      27,
      154
    ],
    "telecom": [
      24,
      27,
      47,
      59,
      65,
      138
    ],
    "equipment": [
      24,
      27,
      35,
      47,
      59,
      95,
      136,
      139,
      155,
      158
    ],
    "makers": [
      24,
      27,
      45,
      71,
      158
    ],
    "cisco": [
      24,
      27,
      31,
      57,
      65,
      95
    ],
    "lucent": [
      24,
      27,
      95
    ],
    "nortel": [
      24,
      27,
      95
    ],
    "lent": [
      24,
      27,
      29,
      98
    ],
    "customers": [
      24,
      27,
      45,
      82
    ],
    "products": [
      24,
      31,
      46,
      70,
      75,
      90,
      132,
      135,
      137
    ],
    "circular": [
      24,
      27
    ],
    "collapsed": [
      24,
      28,
      29,
      61,
      70,
      74,
      84,
      93,
      94,
      107,
      130
    ],
    "dried": [
      24,
      59
    ],
    "hyperscalers": [
      24,
      59,
      156,
      158
    ],
    "microsoft": [
      24,
      27,
      31,
      57,
      59,
      65,
      74,
      100,
      156
    ],
    "google": [
      24,
      27,
      31,
      42,
      44,
      59,
      100,
      156
    ],
    "amazon": [
      24,
      27,
      31,
      57,
      59,
      105,
      155,
      156
    ],
    "massively": [
      24,
      141
    ],
    "gpus": [
      24,
      27,
      157
    ],
    "partially": [
      24,
      26,
      32,
      40,
      75,
      156,
      158
    ],
    "financed": [
      24,
      27,
      46,
      56,
      58,
      94
    ],
    "benefit": [
      24,
      46,
      98
    ],
    "slows": [
      24,
      67
    ],
    "purchases": [
      24,
      56,
      67,
      84,
      88,
      97,
      98
    ],
    "narrative": [
      24,
      31,
      65,
      67,
      89,
      92
    ],
    "referential": [
      24,
      27
    ],
    "appreciated": [
      24,
      36
    ],
    "generation": [
      24,
      58,
      59,
      156
    ],
    "chips": [
      24,
      35,
      45,
      157
    ],
    "arrive": [
      24,
      59
    ],
    "invested": [
      24,
      59,
      70,
      96
    ],
    "billions": [
      24,
      27,
      65,
      70,
      72,
      118,
      119
    ],
    "clusters": [
      24,
      59
    ],
    "writedowns": [
      24,
      95,
      157
    ],
    "justify": [
      24,
      36,
      51,
      70,
      88
    ],
    "efficient": [
      24,
      157
    ],
    "obsolete": [
      24,
      152
    ],
    "fully": [
      24,
      49,
      64,
      66,
      97,
      133,
      157
    ],
    "pace": [
      24,
      55,
      62,
      79
    ],
    "chip": [
      24,
      35,
      59,
      147,
      157
    ],
    "development": [
      24,
      27,
      55,
      66,
      73,
      82,
      97
    ],
    "roughly": [
      24,
      48,
      53,
      60,
      68,
      75,
      77,
      78,
      83,
      88,
      91,
      141
    ],
    "improvements": [
      24,
      82
    ],
    "obsolescence": [
      24,
      155,
      157,
      158
    ],
    "adds": [
      24,
      53,
      60,
      66,
      69,
      70
    ],
    "layer": [
      24,
      36,
      66,
      69,
      99
    ],
    "performance": [
      24,
      26,
      45,
      47,
      57,
      65,
      74
    ],
    "proprietary": [
      24,
      44
    ],
    "inference": [
      24,
      157
    ],
    "erode": [
      24,
      76
    ],
    "narrows": [
      24,
      100,
      131
    ],
    "spent": [
      24,
      27,
      138,
      153
    ],
    "hundreds": [
      24,
      27,
      55,
      65,
      70,
      72,
      88,
      114,
      117,
      119
    ],
    "compresses": [
      24,
      75
    ],
    "destroys": [
      24,
      33,
      141
    ],
    "utilization": [
      24,
      36,
      93
    ],
    "deceleration": [
      24,
      59,
      82,
      156,
      157
    ],
    "signs": [
      24,
      63,
      67,
      76
    ],
    "project": [
      24,
      125,
      130
    ],
    "enterprise": [
      24,
      45,
      58,
      100
    ],
    "disclosures": [
      24,
      35,
      70,
      72,
      73
    ],
    "semi": [
      24,
      35,
      94
    ],
    "gartner": [
      24,
      99
    ],
    "limited": [
      24,
      38,
      43,
      44,
      68,
      70
    ],
    "availability": [
      24,
      38,
      43,
      52
    ],
    "invest": [
      25,
      70,
      75
    ],
    "hidden": [
      25,
      39,
      73,
      74,
      77,
      95,
      98,
      116,
      117
    ],
    "japanese": [
      25,
      29,
      75,
      94
    ],
    "swiss": [
      25,
      30
    ],
    "franc": [
      25,
      30
    ],
    "profitable": [
      25,
      31,
      59,
      65,
      75,
      94,
      95
    ],
    "differential": [
      25,
      90
    ],
    "devastating": [
      25,
      29,
      39,
      62,
      63,
      64,
      94,
      95,
      96,
      97,
      120,
      138,
      141
    ],
    "characteristic": [
      25,
      28,
      52,
      53
    ],
    "described": [
      25,
      36,
      67
    ],
    "steady": [
      25,
      38,
      39,
      60,
      70,
      81
    ],
    "cover": [
      25,
      67,
      73,
      95,
      154
    ],
    "violence": [
      25,
      98
    ],
    "unwinds": [
      25,
      68
    ],
    "nearly": [
      25,
      27,
      51,
      56,
      81,
      86,
      91,
      93,
      94,
      98,
      131,
      141,
      142
    ],
    "precisely": [
      25,
      26,
      36,
      38,
      49,
      56,
      60,
      75,
      83,
      85,
      94,
      150
    ],
    "forwards": [
      25,
      72
    ],
    "trillion": [
      25,
      27,
      54,
      58,
      68,
      69,
      73,
      74,
      75,
      76,
      77,
      90,
      97,
      98,
      111
    ],
    "minor": [
      25,
      104
    ],
    "reaction": [
      25,
      30
    ],
    "normalization": [
      25,
      75,
      91
    ],
    "away": [
      25,
      57,
      74,
      87,
      97
    ],
    "loose": [
      25,
      28,
      32,
      93
    ],
    "maintain": [
      25,
      38,
      46,
      73,
      142,
      157
    ],
    "funded": [
      25,
      45,
      59,
      68,
      75,
      95,
      96,
      98,
      156,
      157
    ],
    "perhaps": [
      25,
      67,
      74
    ],
    "hike": [
      25,
      32,
      75
    ],
    "july": [
      25,
      29,
      93,
      94
    ],
    "mini": [
      25,
      98
    ],
    "rally": [
      25,
      26,
      34,
      78,
      90
    ],
    "scale": [
      25,
      55,
      59,
      68,
      70,
      71,
      96,
      98,
      158
    ],
    "spreading": [
      25,
      97,
      140
    ],
    "breakdown": [
      26,
      36,
      135
    ],
    "suffer": [
      26,
      34,
      39,
      45,
      103
    ],
    "exceeding": [
      26,
      75,
      94,
      125
    ],
    "commodities": [
      26,
      86,
      148
    ],
    "class": [
      26,
      40,
      63,
      67,
      68,
      73,
      115
    ],
    "converge": [
      26,
      87,
      95
    ],
    "fails": [
      26,
      77,
      123,
      157
    ],
    "reduction": [
      26,
      35,
      100,
      123,
      140,
      143
    ],
    "whatever": [
      26,
      109,
      143
    ],
    "impaired": [
      26,
      76,
      94
    ],
    "unrelated": [
      26,
      95
    ],
    "impacted": [
      26,
      55
    ],
    "treasuries": [
      26,
      30,
      58,
      68,
      75,
      77,
      89,
      97,
      98,
      162
    ],
    "ultimate": [
      26,
      97
    ],
    "everything": [
      26,
      59,
      66
    ],
    "estimate": [
      26,
      42,
      56,
      75,
      92,
      157
    ],
    "meet": [
      26,
      29,
      39,
      61,
      68,
      70,
      72,
      74,
      89,
      98
    ],
    "limits": [
      26,
      163
    ],
    "contribution": [
      26,
      37,
      39
    ],
    "benefits": [
      26,
      38,
      46
    ],
    "similarly": [
      26,
      59
    ],
    "caught": [
      26,
      62,
      152
    ],
    "strategy": [
      26,
      38,
      39,
      68,
      70,
      71,
      94
    ],
    "sectors": [
      26,
      30,
      34,
      35,
      36,
      46,
      80,
      83,
      100,
      131
    ],
    "selloffs": [
      26,
      156
    ],
    "breaks": [
      26,
      74,
      77,
      92
    ],
    "suggest": [
      26,
      62,
      69
    ],
    "engaged": [
      26,
      27
    ],
    "analytics": [
      26,
      44,
      46,
      47,
      71,
      74
    ],
    "platforms": [
      26,
      31,
      44,
      73,
      78
    ],
    "precedent": [
      27,
      31,
      48,
      65,
      71,
      88
    ],
    "rose": [
      27,
      28,
      31,
      67
    ],
    "internet": [
      27,
      59,
      65,
      95,
      103,
      105
    ],
    "destroyed": [
      27,
      61,
      63
    ],
    "trough": [
      27,
      63,
      65
    ],
    "prominent": [
      27,
      95
    ],
    "lost": [
      27,
      31,
      63,
      77,
      103,
      118,
      126,
      142
    ],
    "went": [
      27,
      85,
      95,
      131
    ],
    "bankrupt": [
      27,
      94,
      98
    ],
    "laying": [
      27,
      79,
      152
    ],
    "fiber": [
      27,
      59
    ],
    "optic": [
      27,
      59
    ],
    "centers": [
      27,
      35,
      37,
      38,
      40,
      157
    ],
    "traffic": [
      27,
      59,
      60,
      66
    ],
    "largely": [
      27,
      43,
      68,
      70,
      77,
      102,
      118
    ],
    "grow": [
      27,
      59,
      68,
      95,
      141,
      143
    ],
    "exponentially": [
      27,
      59
    ],
    "wildly": [
      27,
      145
    ],
    "wrong": [
      27,
      31,
      40,
      59,
      65,
      96,
      115,
      117,
      118,
      119,
      120,
      151,
      154
    ],
    "meant": [
      27,
      95,
      96,
      98
    ],
    "worldcom": [
      27,
      59
    ],
    "crossing": [
      27,
      59,
      60,
      124
    ],
    "tens": [
      27,
      72
    ],
    "compute": [
      27,
      154,
      157
    ],
    "accelerant": [
      27,
      60
    ],
    "booked": [
      27,
      86
    ],
    "faced": [
      27,
      70,
      93,
      96
    ],
    "identical": [
      27,
      65,
      98
    ],
    "buyers": [
      27,
      58,
      71,
      72,
      88,
      91,
      96,
      137
    ],
    "startups": [
      27,
      36
    ],
    "destroy": [
      27,
      31,
      66,
      96,
      118
    ],
    "essential": [
      27,
      31,
      46,
      47,
      153
    ],
    "ultimately": [
      27,
      32,
      62,
      76,
      81,
      93,
      94,
      96,
      97
    ],
    "ofek": [
      27,
      95
    ],
    "richardson": [
      27,
      95
    ],
    "brunnermeier": [
      27,
      28,
      95
    ],
    "defining": [
      28,
      92,
      94
    ],
    "none": [
      28,
      34,
      96
    ],
    "individually": [
      28,
      32,
      34,
      71,
      96
    ],
    "produce": [
      28,
      35,
      42,
      57,
      71,
      94,
      97,
      100,
      140
    ],
    "meltdown": [
      28,
      96
    ],
    "occurred": [
      28,
      60,
      71,
      75
    ],
    "home": [
      28,
      33,
      91,
      96
    ],
    "risen": [
      28,
      67
    ],
    "widespread": [
      28,
      35,
      73,
      96,
      111,
      124
    ],
    "shadow": [
      28,
      33,
      70,
      96,
      110
    ],
    "entities": [
      28,
      68,
      70,
      72,
      77,
      88
    ],
    "sivs": [
      28,
      96
    ],
    "cdos": [
      28,
      96,
      117
    ],
    "operated": [
      28,
      96,
      98
    ],
    "complex": [
      28,
      37,
      41,
      46,
      90,
      96
    ],
    "opaque": [
      28,
      73
    ],
    "unfolded": [
      28,
      93
    ],
    "sequence": [
      28,
      72,
      93,
      94,
      95,
      96,
      97,
      98
    ],
    "template": [
      28,
      31,
      58,
      63
    ],
    "warnings": [
      28,
      117
    ],
    "delinquencies": [
      28,
      55,
      56
    ],
    "starts": [
      28,
      48,
      91
    ],
    "contained": [
      28,
      97
    ],
    "paribas": [
      28,
      77,
      96
    ],
    "froze": [
      28,
      72,
      86,
      97
    ],
    "stearns": [
      28,
      96,
      107
    ],
    "seized": [
      28,
      70
    ],
    "clearly": [
      28,
      87,
      96,
      97
    ],
    "september": [
      28,
      37,
      48,
      62,
      72,
      90,
      97,
      98
    ],
    "bankruptcy": [
      28,
      119
    ],
    "broke": [
      28,
      83,
      94,
      97
    ],
    "completely": [
      28,
      31,
      98
    ],
    "bailouts": [
      28,
      109,
      116
    ],
    "guarantees": [
      28,
      152
    ],
    "timeline": [
      28,
      59,
      66,
      78,
      151,
      154,
      159,
      160
    ],
    "predictions": [
      28,
      101,
      102,
      103,
      104,
      105,
      106,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115
    ],
    "surged": [
      28,
      39,
      50,
      63,
      83,
      90,
      91,
      93,
      95,
      97,
      98
    ],
    "quantitative": [
      28,
      39,
      72,
      91,
      98
    ],
    "easing": [
      28,
      62,
      67,
      81,
      98
    ],
    "successfully": [
      28,
      66
    ],
    "prevented": [
      28,
      54
    ],
    "moral": [
      28,
      41,
      94,
      130
    ],
    "hazard": [
      28,
      41,
      94
    ],
    "zombie": [
      28,
      69,
      76
    ],
    "shaped": [
      28,
      97
    ],
    "decade": [
      28,
      31,
      32,
      37,
      70,
      73,
      84,
      91,
      94,
      98,
      103,
      108,
      118,
      154
    ],
    "influenced": [
      28,
      80
    ],
    "fcic": [
      28,
      96
    ],
    "bernanke": [
      28,
      96
    ],
    "geithner": [
      28,
      96
    ],
    "transform": [
      29,
      65
    ],
    "localized": [
      29,
      76,
      94,
      96,
      114,
      115,
      142
    ],
    "thailand": [
      29,
      94
    ],
    "baht": [
      29,
      94
    ],
    "reserves": [
      29,
      55,
      77,
      88,
      90,
      94,
      106,
      108,
      118,
      123,
      125,
      138,
      144,
      146
    ],
    "indonesia": [
      29,
      94,
      141
    ],
    "malaysia": [
      29,
      94,
      144
    ],
    "philippines": [
      29,
      94
    ],
    "russian": [
      29,
      126
    ],
    "ltcm": [
      29,
      68,
      75
    ],
    "setup": [
      29,
      63,
      64,
      65
    ],
    "offered": [
      29,
      40
    ],
    "strong": [
      29,
      31,
      40,
      46,
      51,
      54,
      57,
      65,
      70,
      83,
      90,
      114
    ],
    "primarily": [
      29,
      44,
      47,
      49,
      57,
      72,
      77,
      78,
      89,
      90,
      91,
      97
    ],
    "earning": [
      29,
      39
    ],
    "corporations": [
      29,
      77,
      97
    ],
    "made": [
      29,
      42,
      62,
      94,
      95,
      96,
      118
    ],
    "cracked": [
      29,
      135,
      139
    ],
    "devaluations": [
      29,
      87
    ],
    "appeared": [
      29,
      64,
      81,
      94
    ],
    "solvent": [
      29,
      72,
      94
    ],
    "attempt": [
      29,
      71
    ],
    "crushed": [
      29,
      93
    ],
    "flight": [
      29,
      46,
      61,
      75,
      88,
      94,
      161
    ],
    "contracted": [
      29,
      54
    ],
    "problems": [
      29,
      31,
      60,
      116
    ],
    "caused": [
      29,
      61,
      75,
      90,
      94,
      96
    ],
    "pegs": [
      29,
      94
    ],
    "redemptions": [
      29,
      70,
      73,
      74
    ],
    "creditor": [
      29,
      128
    ],
    "applicable": [
      29,
      100
    ],
    "today": [
      29,
      57,
      68,
      69,
      70,
      71,
      74,
      75,
      76,
      77,
      92,
      142,
      151
    ],
    "radelet": [
      29,
      94
    ],
    "sachs": [
      29,
      57,
      59,
      94,
      114
    ],
    "corsetti": [
      29,
      94
    ],
    "pesenti": [
      29,
      94
    ],
    "serve": [
      30,
      77
    ],
    "systems": [
      30,
      44,
      92,
      97,
      99
    ],
    "immediate": [
      30,
      53,
      60,
      64,
      102,
      115,
      116,
      117,
      126,
      140
    ],
    "minutes": [
      30,
      71,
      119
    ],
    "algorithmic": [
      30,
      71
    ],
    "trading": [
      30,
      37,
      53,
      74,
      88,
      89,
      90,
      92,
      95,
      97,
      98,
      116
    ],
    "priced": [
      30,
      38,
      62,
      65,
      91,
      96,
      133,
      134,
      138,
      154
    ],
    "almost": [
      30,
      48,
      51,
      53,
      58,
      65,
      81,
      115,
      120
    ],
    "instantly": [
      30,
      94,
      114
    ],
    "middle": [
      30,
      32,
      58,
      73,
      80,
      84,
      87,
      88,
      90,
      130,
      137,
      141,
      150
    ],
    "eastern": [
      30,
      88,
      137,
      150
    ],
    "imports": [
      30,
      86,
      87,
      134,
      137,
      138,
      139,
      141,
      150,
      160
    ],
    "partly": [
      30,
      39,
      54,
      56,
      74
    ],
    "status": [
      30,
      34,
      58,
      97,
      116
    ],
    "lesson": [
      30,
      31,
      40,
      98
    ],
    "crashing": [
      31,
      74,
      107,
      158
    ],
    "recovering": [
      31,
      84
    ],
    "highs": [
      31,
      98,
      102,
      113,
      136,
      151,
      161
    ],
    "play": [
      31,
      116
    ],
    "behave": [
      31,
      162
    ],
    "convergence": [
      31,
      64,
      65,
      68,
      88,
      158
    ],
    "slowing": [
      31,
      32,
      49,
      78,
      84,
      90
    ],
    "apple": [
      31,
      57,
      74
    ],
    "targeting": [
      31,
      60,
      71,
      97,
      133,
      137,
      152
    ],
    "pivot": [
      31,
      120
    ],
    "bottom": [
      31,
      39,
      69,
      129
    ],
    "thesis": [
      31,
      151,
      154
    ],
    "proved": [
      31,
      114
    ],
    "business": [
      31,
      45,
      51,
      62,
      73,
      83,
      92,
      93,
      95,
      96,
      115
    ],
    "aggressively": [
      31,
      56,
      62,
      67,
      79,
      93,
      94,
      101
    ],
    "efficiency": [
      31,
      47,
      97
    ],
    "recovered": [
      31,
      95,
      102
    ],
    "locked": [
      31,
      48,
      56,
      69,
      73,
      76,
      98
    ],
    "turned": [
      31,
      74
    ],
    "thing": [
      31,
      45
    ],
    "enormous": [
      31,
      68,
      69,
      71,
      72,
      77,
      95,
      96
    ],
    "profound": [
      31,
      87,
      97
    ],
    "analytical": [
      31,
      115,
      116,
      127,
      130
    ],
    "permanent": [
      31,
      99,
      101,
      106,
      115,
      127,
      129
    ],
    "impairment": [
      31,
      76
    ],
    "opportunity": [
      31,
      40,
      47,
      151
    ],
    "beginning": [
      31,
      51,
      77,
      93,
      96
    ],
    "audience": [
      31,
      116,
      120
    ],
    "possible": [
      31,
      66,
      94
    ],
    "sound": [
      31,
      57
    ],
    "analyst": [
      31,
      44,
      116,
      119,
      156
    ],
    "deepest": [
      32,
      48,
      63,
      88
    ],
    "downturns": [
      32,
      49,
      64,
      82
    ],
    "eliminates": [
      32,
      60,
      67,
      81,
      99
    ],
    "tools": [
      32,
      42,
      47,
      59,
      100
    ],
    "works": [
      32,
      38,
      43,
      77
    ],
    "gasoline": [
      32,
      66,
      135
    ],
    "indirectly": [
      32,
      141
    ],
    "zone": [
      32,
      50,
      54,
      56,
      57,
      64,
      124
    ],
    "committed": [
      32,
      45,
      59,
      65,
      94
    ],
    "continue": [
      32,
      46,
      75,
      76,
      86,
      116
    ],
    "pause": [
      32,
      85,
      157
    ],
    "ease": [
      32,
      98
    ],
    "risks": [
      32,
      39,
      52,
      55,
      62,
      66,
      67,
      70,
      76,
      81,
      96,
      98
    ],
    "anchoring": [
      32,
      62,
      69
    ],
    "wage": [
      32,
      80,
      83,
      93
    ],
    "option": [
      32,
      39,
      69,
      71,
      96,
      154
    ],
    "raised": [
      32,
      35,
      70,
      73,
      93,
      94,
      98,
      101,
      157
    ],
    "break": [
      32,
      43,
      62,
      63,
      71,
      81,
      89,
      90
    ],
    "burns": [
      32,
      62,
      63,
      81,
      93
    ],
    "tried": [
      32,
      94
    ],
    "eroded": [
      32,
      63,
      84,
      94
    ],
    "living": [
      32,
      70,
      93,
      104
    ],
    "resolve": [
      32,
      80
    ],
    "remaining": [
      32,
      38,
      74,
      85,
      95
    ],
    "east": [
      32,
      58,
      80,
      84,
      87,
      88,
      90,
      130,
      141
    ],
    "space": [
      32,
      44,
      55,
      58
    ],
    "unable": [
      32,
      49,
      62,
      63,
      98,
      149
    ],
    "outlook": [
      32,
      34,
      58,
      83,
      87
    ],
    "occurrence": [
      33,
      94
    ],
    "reinforces": [
      33,
      65
    ],
    "feedback": [
      33,
      91,
      93,
      130
    ],
    "overwhelm": [
      33,
      47
    ],
    "circuit": [
      33,
      52
    ],
    "breakers": [
      33,
      52
    ],
    "backing": [
      33,
      119
    ],
    "doom": [
      33,
      58,
      61,
      76,
      94,
      95,
      98,
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115,
      116,
      120
    ],
    "pull": [
      33,
      52,
      84,
      133
    ],
    "drops": [
      33,
      140
    ],
    "lines": [
      33,
      52,
      54,
      77,
      97
    ],
    "renewed": [
      33,
      61,
      78,
      84,
      90,
      91
    ],
    "prevents": [
      33,
      91,
      119,
      153
    ],
    "amplifier": [
      33,
      39,
      69,
      73,
      99
    ],
    "smaller": [
      33,
      55
    ],
    "essentially": [
      33,
      39,
      44,
      54,
      65,
      80,
      96,
      155
    ],
    "loaded": [
      33,
      77,
      82,
      96
    ],
    "spring": [
      33,
      87,
      124,
      129
    ],
    "globally": [
      33,
      77,
      98,
      129,
      133,
      141,
      157
    ],
    "volumes": [
      33,
      42
    ],
    "flash": [
      33,
      71
    ],
    "regardless": [
      33,
      36,
      38,
      46,
      59,
      65,
      66,
      81,
      95,
      97,
      99
    ],
    "broker": [
      33,
      96
    ],
    "dealer": [
      33,
      71
    ],
    "activation": [
      34,
      79
    ],
    "greatest": [
      34,
      118
    ],
    "principle": [
      34,
      40
    ],
    "applies": [
      34,
      39
    ],
    "consider": [
      34,
      41
    ],
    "reduces": [
      34,
      47,
      75,
      76,
      77,
      91,
      99
    ],
    "ability": [
      34,
      37,
      41,
      43,
      47,
      58,
      66,
      80,
      156
    ],
    "linear": [
      34,
      41,
      52,
      140
    ],
    "independently": [
      34,
      92
    ],
    "informative": [
      34,
      89
    ],
    "propagates": [
      35,
      59
    ],
    "interventions": [
      35,
      54
    ],
    "demonstration": [
      35,
      72
    ],
    "lockdowns": [
      35,
      111
    ],
    "auto": [
      35,
      56,
      158
    ],
    "container": [
      35,
      152,
      160
    ],
    "manufacturers": [
      35,
      59,
      82
    ],
    "cars": [
      35,
      84
    ],
    "medical": [
      35,
      46,
      132,
      134,
      138,
      139
    ],
    "appliances": [
      35,
      84
    ],
    "industrial": [
      35,
      46,
      86,
      92,
      93,
      106,
      126,
      149
    ],
    "telecommunications": [
      35,
      45,
      95
    ],
    "imported": [
      35,
      134,
      138,
      141
    ],
    "amplified": [
      35,
      47,
      71,
      74,
      90
    ],
    "concentrate": [
      35,
      154
    ],
    "mining": [
      35,
      44,
      92
    ],
    "pipeline": [
      35,
      64,
      122,
      130,
      133,
      150
    ],
    "routes": [
      35,
      66,
      86
    ],
    "training": [
      35,
      41,
      157
    ],
    "logistics": [
      35,
      44,
      45
    ],
    "baltic": [
      35,
      86
    ],
    "ship": [
      35,
      44
    ],
    "occurring": [
      36,
      58,
      60
    ],
    "drag": [
      36,
      47,
      60,
      67,
      87,
      155
    ],
    "etfs": [
      36,
      74,
      158
    ],
    "passive": [
      36,
      37,
      65,
      74,
      117,
      151,
      158
    ],
    "mechanical": [
      36,
      71,
      97,
      98
    ],
    "automatically": [
      36,
      42,
      72,
      73,
      93
    ],
    "weakened": [
      36,
      75,
      76,
      92
    ],
    "assumed": [
      36,
      72,
      96,
      119
    ],
    "breaking": [
      36,
      48,
      92,
      129
    ],
    "ecosystem": [
      36,
      71,
      74
    ],
    "center": [
      36,
      45,
      47,
      59,
      92
    ],
    "intersection": [
      36,
      45,
      100
    ],
    "otherwise": [
      36,
      56,
      57,
      58,
      76,
      92
    ],
    "likely": [
      36,
      65,
      81,
      94,
      136,
      157
    ],
    "painful": [
      36,
      57,
      65
    ],
    "coincides": [
      36,
      69
    ],
    "relying": [
      36,
      42
    ],
    "founded": [
      37,
      38,
      39,
      40
    ],
    "rules": [
      37,
      116
    ],
    "weather": [
      37,
      44,
      85,
      156
    ],
    "perform": [
      37,
      39
    ],
    "alpha": [
      37,
      39,
      71
    ],
    "systematic": [
      37,
      39,
      97,
      99
    ],
    "machine": [
      37,
      41,
      105,
      147
    ],
    "decision": [
      37,
      57,
      67,
      157
    ],
    "rule": [
      37,
      79
    ],
    "reportedly": [
      37,
      38
    ],
    "returned": [
      37,
      67
    ],
    "date": [
      37,
      78,
      95,
      104,
      158
    ],
    "macro": [
      37,
      59,
      71,
      75,
      83,
      94,
      109,
      115,
      120,
      155
    ],
    "beta": [
      37,
      39
    ],
    "argues": [
      37,
      40
    ],
    "disguised": [
      37,
      39
    ],
    "actionable": [
      37,
      93,
      94,
      95,
      96,
      97
    ],
    "stages": [
      37,
      40,
      42,
      60,
      87
    ],
    "extraordinary": [
      37,
      38,
      98
    ],
    "implies": [
      37,
      53
    ],
    "believe": [
      37,
      53
    ],
    "books": [
      37,
      76
    ],
    "investor": [
      37,
      38,
      73,
      74,
      155,
      156,
      157,
      161
    ],
    "letters": [
      37,
      38,
      40
    ],
    "investments": [
      38,
      98,
      157
    ],
    "tail": [
      38,
      39,
      52,
      57,
      100
    ],
    "continuous": [
      38,
      95
    ],
    "black": [
      38,
      102,
      116
    ],
    "maintaining": [
      38,
      63,
      120
    ],
    "easy": [
      38,
      74
    ],
    "program": [
      38,
      97
    ],
    "allocation": [
      38,
      97,
      145,
      157
    ],
    "predict": [
      38,
      43,
      116
    ],
    "simply": [
      38,
      39,
      72,
      73,
      76
    ],
    "ensures": [
      38,
      150
    ],
    "offers": [
      38,
      89
    ],
    "highest": [
      38,
      76,
      83,
      91,
      95,
      98,
      133,
      137
    ],
    "feel": [
      38,
      84
    ],
    "view": [
      38,
      118
    ],
    "gaussian": [
      38,
      96
    ],
    "cliff": [
      39,
      59,
      158
    ],
    "parts": [
      39,
      46,
      127
    ],
    "seemingly": [
      39,
      95
    ],
    "momentum": [
      39,
      78,
      88
    ],
    "achieve": [
      39,
      43,
      116
    ],
    "position": [
      39,
      56,
      68,
      76,
      98
    ],
    "differentials": [
      39,
      90
    ],
    "bearing": [
      39,
      157
    ],
    "plummet": [
      39,
      61
    ],
    "surge": [
      39,
      60,
      61,
      62,
      66,
      69,
      70,
      93,
      161
    ],
    "accompanied": [
      39,
      115
    ],
    "defensive": [
      39,
      46,
      64,
      120,
      161
    ],
    "lagged": [
      39,
      44,
      82
    ],
    "discipline": [
      39,
      73,
      155,
      156,
      157
    ],
    "creation": [
      39,
      47,
      50,
      74
    ],
    "violently": [
      39,
      52,
      72,
      97
    ],
    "journal": [
      39,
      41,
      43
    ],
    "reversion": [
      40
    ],
    "built": [
      40,
      68,
      95,
      96,
      107,
      137
    ],
    "forecasting": [
      40,
      127,
      128,
      130
    ],
    "called": [
      40,
      68,
      92
    ],
    "termed": [
      40,
      96
    ],
    "remarkably": [
      40,
      85,
      94,
      95
    ],
    "direction": [
      40,
      47,
      70,
      71,
      75,
      79,
      81,
      158
    ],
    "fair": [
      40,
      51
    ],
    "dividend": [
      40,
      119,
      156
    ],
    "projects": [
      40,
      55,
      92,
      94,
      99
    ],
    "annualized": [
      40,
      51
    ],
    "profitability": [
      40,
      154
    ],
    "needs": [
      40,
      63,
      66,
      81,
      97,
      157
    ],
    "sustainability": [
      40,
      94
    ],
    "globalization": [
      40,
      51
    ],
    "sometimes": [
      40,
      96
    ],
    "outperform": [
      40,
      46
    ],
    "missing": [
      40,
      123
    ],
    "forecasts": [
      40,
      100
    ],
    "learning": [
      41,
      42,
      43
    ],
    "methodological": [
      41,
      80
    ],
    "frontier": [
      41,
      44,
      138
    ],
    "offering": [
      41,
      61,
      72
    ],
    "relationships": [
      41,
      43,
      153
    ],
    "econometric": [
      41,
      42
    ],
    "span": [
      41,
      66
    ],
    "knowledge": [
      41,
      163
    ],
    "base": [
      41,
      74,
      77,
      163
    ],
    "learn": [
      41,
      43
    ],
    "assume": [
      41,
      100,
      116,
      151
    ],
    "relate": [
      41,
      42
    ],
    "neural": [
      41,
      43
    ],
    "discover": [
      41,
      73
    ],
    "challenges": [
      41,
      159
    ],
    "past": [
      41,
      48,
      55,
      58,
      63,
      66,
      67,
      70,
      75,
      76,
      81,
      86,
      112
    ],
    "types": [
      41,
      46,
      55,
      97,
      162
    ],
    "adopted": [
      41,
      98,
      157
    ],
    "exist": [
      41,
      71,
      127,
      151
    ],
    "promising": [
      41,
      43
    ],
    "applications": [
      41,
      59
    ],
    "human": [
      41,
      71,
      114,
      141
    ],
    "recognition": [
      41,
      73,
      76
    ],
    "inability": [
      41,
      45,
      163
    ],
    "anticipate": [
      41,
      48
    ],
    "fintech": [
      41,
      43
    ],
    "publications": [
      41,
      43,
      59,
      70,
      73,
      74
    ],
    "agencies": [
      42,
      96,
      100
    ],
    "decisions": [
      42,
      62,
      66
    ],
    "released": [
      42,
      82,
      83,
      85,
      87,
      91
    ],
    "quarter": [
      42,
      81
    ],
    "card": [
      42,
      44,
      56,
      99
    ],
    "transaction": [
      42,
      44,
      99,
      100
    ],
    "electricity": [
      42,
      46
    ],
    "search": [
      42,
      156
    ],
    "trends": [
      42,
      44,
      97
    ],
    "mixed": [
      42,
      46,
      76
    ],
    "confirmed": [
      42,
      155
    ],
    "travel": [
      42,
      44,
      79,
      81
    ],
    "employment": [
      42,
      47,
      79,
      82,
      85,
      87,
      111,
      114,
      138
    ],
    "allowed": [
      42,
      62,
      73,
      76,
      81,
      98,
      116
    ],
    "respond": [
      42,
      58,
      62
    ],
    "atlanta": [
      42,
      72
    ],
    "informational": [
      42,
      44
    ],
    "versus": [
      42,
      80
    ],
    "later": [
      42,
      77,
      97,
      98,
      106
    ],
    "territory": [
      42,
      51,
      54,
      64,
      82,
      87
    ],
    "anonymized": [
      42,
      44
    ],
    "temporal": [
      43
    ],
    "graph": [
      43
    ],
    "carries": [
      43,
      58,
      88,
      122,
      136
    ],
    "ratings": [
      43,
      69,
      70,
      96
    ],
    "evolve": [
      43,
      75
    ],
    "mutual": [
      43,
      69
    ],
    "results": [
      43,
      143
    ],
    "predicting": [
      43,
      116
    ],
    "detailed": [
      43,
      96
    ],
    "outside": [
      44,
      60,
      68,
      77,
      100,
      155
    ],
    "content": [
      44,
      96,
      121
    ],
    "language": [
      44,
      51,
      115
    ],
    "commentary": [
      44,
      154,
      155
    ],
    "around": [
      44,
      52,
      71,
      86,
      96,
      136,
      148
    ],
    "harder": [
      44,
      45
    ],
    "tanker": [
      44,
      66
    ],
    "light": [
      44,
      51,
      133,
      137,
      157
    ],
    "correlates": [
      44,
      45
    ],
    "occupancy": [
      44,
      76
    ],
    "delayed": [
      44,
      115
    ],
    "payment": [
      44,
      73,
      91,
      99
    ],
    "afford": [
      44,
      126,
      149
    ],
    "feeds": [
      44,
      67,
      127,
      129
    ],
    "backtesting": [
      44,
      160
    ],
    "labs": [
      44,
      157
    ],
    "engine": [
      45,
      46,
      47,
      100
    ],
    "differs": [
      45,
      65,
      76,
      154
    ],
    "attacks": [
      45,
      86,
      94
    ],
    "withdrawal": [
      45,
      61,
      95,
      127,
      130,
      154
    ],
    "once": [
      45,
      115,
      116
    ],
    "immediately": [
      45,
      66,
      68,
      93,
      95,
      108
    ],
    "thin": [
      45,
      69,
      71,
      72,
      98
    ],
    "heavy": [
      45,
      61,
      133,
      137
    ],
    "double": [
      45,
      52,
      54,
      65,
      83,
      91,
      94,
      95,
      111,
      120,
      141
    ],
    "operations": [
      45,
      46,
      70,
      94,
      95,
      96,
      97
    ],
    "adjacent": [
      45,
      47,
      57,
      65
    ],
    "unique": [
      45,
      97,
      98
    ],
    "operators": [
      45,
      145
    ],
    "programs": [
      45,
      82,
      128
    ],
    "decelerates": [
      45,
      155
    ],
    "step": [
      45,
      54,
      97
    ],
    "memory": [
      45,
      151
    ],
    "items": [
      45,
      122
    ],
    "faces": [
      45,
      57,
      69,
      79,
      90
    ],
    "sits": [
      45,
      49,
      78
    ],
    "operational": [
      45,
      62,
      99
    ],
    "resilient": [
      46,
      56
    ],
    "counter": [
      46,
      72,
      152
    ],
    "certain": [
      46,
      58,
      77,
      96
    ],
    "safety": [
      46,
      129,
      144,
      153,
      156
    ],
    "function": [
      46,
      60,
      97,
      98
    ],
    "advance": [
      46,
      50,
      104
    ],
    "insulated": [
      46,
      56,
      126,
      133,
      137
    ],
    "guaranteed": [
      46,
      113
    ],
    "effectively": [
      46,
      56,
      57,
      66,
      68,
      75,
      81,
      86,
      90,
      95,
      119,
      138,
      147
    ],
    "impair": [
      46,
      76
    ],
    "hospital": [
      46,
      139
    ],
    "coverage": [
      46,
      69,
      99,
      100
    ],
    "mass": [
      46,
      56
    ],
    "water": [
      46,
      83,
      138
    ],
    "seek": [
      46,
      90
    ],
    "food": [
      46,
      80,
      81,
      93,
      106,
      121,
      122,
      123,
      124,
      126,
      127,
      129,
      132,
      134,
      138,
      139,
      142,
      148
    ],
    "basic": [
      46,
      95
    ],
    "personal": [
      46,
      56,
      81
    ],
    "holds": [
      46,
      61,
      83
    ],
    "type": [
      46,
      76
    ],
    "transformation": [
      47,
      70,
      71,
      92
    ],
    "adoption": [
      47,
      99,
      114
    ],
    "generally": [
      47,
      56,
      141
    ],
    "mapped": [
      47,
      96
    ],
    "pullback": [
      47,
      78,
      84,
      86,
      155
    ],
    "material": [
      47,
      100
    ],
    "anchor": [
      47,
      66,
      93
    ],
    "happened": [
      47,
      58,
      119
    ],
    "rest": [
      47,
      131
    ],
    "drivers": [
      47,
      82,
      88
    ],
    "mode": [
      47,
      52,
      117,
      118,
      119,
      120
    ],
    "largest": [
      47,
      55,
      56,
      57,
      61,
      65,
      66,
      75,
      82,
      87,
      94,
      98,
      125,
      137
    ],
    "stands": [
      48,
      51,
      53,
      54,
      55,
      56,
      62,
      65,
      78,
      79,
      80,
      81,
      86
    ],
    "inversions": [
      48,
      96
    ],
    "treated": [
      48,
      88,
      115
    ],
    "story": [
      48,
      54,
      93,
      95,
      155
    ],
    "expect": [
      48,
      84,
      116
    ],
    "forward": [
      48,
      94
    ],
    "looking": [
      48,
      72
    ],
    "anticipated": [
      48,
      71,
      82
    ],
    "pulling": [
      48,
      49,
      137
    ],
    "producing": [
      48,
      83
    ],
    "sticky": [
      48,
      49,
      78,
      79,
      80,
      83,
      84,
      122
    ],
    "stark": [
      48,
      95
    ],
    "december": [
      48,
      62,
      64,
      85,
      93,
      95,
      97,
      98
    ],
    "unprecedented": [
      48,
      54,
      57,
      97,
      98,
      111
    ],
    "strength": [
      48,
      59,
      70,
      82,
      83,
      86,
      90,
      92
    ],
    "actively": [
      48,
      52,
      53,
      63,
      98,
      145
    ],
    "steepening": [
      48,
      49
    ],
    "things": [
      49,
      84,
      120
    ],
    "looks": [
      49,
      73,
      131,
      156
    ],
    "surface": [
      49,
      52,
      55,
      79,
      80,
      97
    ],
    "anxiety": [
      49,
      52
    ],
    "contrast": [
      49,
      82
    ],
    "configuration": [
      49,
      63,
      64
    ],
    "summer": [
      49,
      64,
      94,
      99
    ],
    "officially": [
      49,
      64
    ],
    "judgment": [
      49,
      88,
      130
    ],
    "soft": [
      49,
      113
    ],
    "landing": [
      49,
      110,
      113
    ],
    "weakening": [
      49,
      62,
      63,
      64,
      66,
      87,
      94
    ],
    "pressures": [
      49,
      69,
      87,
      93
    ],
    "prevent": [
      49,
      98,
      153
    ],
    "blow": [
      49,
      60,
      66
    ],
    "softening": [
      49,
      55,
      62,
      67,
      87
    ],
    "labor": [
      49,
      62,
      67,
      79,
      80,
      81,
      85,
      93
    ],
    "argue": [
      49,
      50,
      51,
      60,
      62
    ],
    "reads": [
      50,
      52
    ],
    "places": [
      50,
      54,
      70
    ],
    "expanded": [
      50,
      83,
      95,
      97
    ],
    "demonstrates": [
      50,
      51
    ],
    "genuinely": [
      50,
      51,
      56,
      57,
      59,
      61,
      133,
      137
    ],
    "reassuring": [
      50,
      52,
      85
    ],
    "caveats": [
      50,
      56
    ],
    "shifted": [
      50,
      67,
      96
    ],
    "turn": [
      50,
      56
    ],
    "swing": [
      50,
      145
    ],
    "understates": [
      50,
      68
    ],
    "dating": [
      51,
      93,
      95
    ],
    "eliminate": [
      51,
      59
    ],
    "supports": [
      51,
      121
    ],
    "brutal": [
      51,
      93
    ],
    "clarity": [
      51,
      93,
      96
    ],
    "peaked": [
      51,
      53,
      93,
      95,
      96,
      106,
      112,
      131
    ],
    "compared": [
      51,
      78
    ],
    "mirror": [
      51,
      103
    ],
    "downside": [
      51,
      53,
      64
    ],
    "ycharts": [
      51,
      78
    ],
    "modestly": [
      52,
      110
    ],
    "concerning": [
      52,
      55,
      58,
      79,
      80,
      82,
      91
    ],
    "begun": [
      52,
      64,
      75,
      85
    ],
    "possibility": [
      52,
      81
    ],
    "masks": [
      52,
      55,
      56,
      58,
      80,
      139
    ],
    "matter": [
      52,
      72,
      75,
      98,
      131
    ],
    "fastest": [
      52,
      61,
      97,
      98,
      112
    ],
    "headline": [
      52,
      55,
      62,
      71,
      80,
      86,
      131
    ],
    "prior": [
      52,
      74,
      78,
      82,
      84,
      145
    ],
    "spot": [
      53,
      116,
      126,
      134,
      138
    ],
    "plus": [
      53,
      91,
      95,
      135
    ],
    "unambiguous": [
      53,
      54
    ],
    "paying": [
      53,
      64
    ],
    "happens": [
      53,
      81,
      142
    ],
    "threat": [
      53,
      64,
      155,
      157,
      160,
      163
    ],
    "invasion": [
      53,
      60,
      130
    ],
    "ones": [
      53,
      79,
      95
    ],
    "fears": [
      53,
      78,
      84,
      89,
      91
    ],
    "slowdowns": [
      54,
      92
    ],
    "drew": [
      54,
      97
    ],
    "rolled": [
      54,
      68
    ],
    "expense": [
      54,
      97
    ],
    "squeeze": [
      54,
      71,
      77,
      95,
      149
    ],
    "motion": [
      54,
      76
    ],
    "accrual": [
      55,
      73
    ],
    "figure": [
      55,
      56,
      59,
      61,
      69,
      80,
      154,
      155
    ],
    "worse": [
      55,
      68,
      84,
      111
    ],
    "remote": [
      55,
      76
    ],
    "hybrid": [
      55,
      76
    ],
    "nationally": [
      55,
      96
    ],
    "affected": [
      55,
      135
    ],
    "properties": [
      55,
      76
    ],
    "injecting": [
      55,
      63
    ],
    "severely": [
      55,
      157
    ],
    "experiencing": [
      55,
      56,
      92
    ],
    "deliveries": [
      55,
      83
    ],
    "started": [
      55,
      131,
      144
    ],
    "rent": [
      55,
      81,
      91
    ],
    "tier": [
      55,
      61,
      71,
      97
    ],
    "accelerates": [
      55,
      61,
      78,
      88,
      89
    ],
    "meaningfully": [
      56,
      67
    ],
    "observed": [
      56,
      85
    ],
    "households": [
      56,
      60,
      66,
      87
    ],
    "homeowners": [
      56,
      91
    ],
    "unusual": [
      56,
      87,
      97
    ],
    "situation": [
      56,
      66,
      73,
      74,
      76,
      79
    ],
    "economists": [
      56,
      63,
      79,
      93
    ],
    "inadvertently": [
      56,
      72
    ],
    "masking": [
      56,
      83
    ],
    "pain": [
      56,
      81,
      150
    ],
    "alphabet": [
      57,
      156
    ],
    "capitalization": [
      57,
      65,
      74
    ],
    "seven": [
      57,
      65,
      74
    ],
    "control": [
      57,
      58,
      112,
      127
    ],
    "collectively": [
      57,
      61,
      71,
      97,
      158
    ],
    "thirds": [
      57,
      138
    ],
    "exceptional": [
      57,
      88
    ],
    "competitive": [
      57,
      82,
      87,
      90,
      157
    ],
    "action": [
      57,
      68
    ],
    "restructure": [
      57,
      73
    ],
    "significance": [
      57,
      81
    ],
    "caps": [
      57,
      65,
      158
    ],
    "precedents": [
      57,
      142
    ],
    "group": [
      57,
      60,
      66,
      92
    ],
    "dramatic": [
      57,
      58,
      60,
      70,
      94,
      98,
      140
    ],
    "losing": [
      57,
      62,
      94,
      126
    ],
    "intel": [
      57,
      59,
      65
    ],
    "walmart": [
      57,
      65
    ],
    "represented": [
      57,
      96
    ],
    "suggesting": [
      57,
      74,
      80,
      82,
      83,
      84
    ],
    "equal": [
      57,
      65,
      74,
      131
    ],
    "jones": [
      57,
      93
    ],
    "goldman": [
      57,
      59,
      114
    ],
    "congressional": [
      58,
      130
    ],
    "budget": [
      58,
      98
    ],
    "continues": [
      58,
      68,
      73,
      76,
      87
    ],
    "buyer": [
      58,
      91,
      97
    ],
    "italy": [
      58,
      97,
      109
    ],
    "savings": [
      58,
      76,
      84
    ],
    "appetite": [
      58,
      89
    ],
    "deficit": [
      58,
      77,
      92,
      94,
      123,
      155
    ],
    "surpassing": [
      58,
      93,
      95
    ],
    "destabilizing": [
      58,
      66
    ],
    "reflexive": [
      58,
      74,
      158
    ],
    "psychological": [
      58,
      116
    ],
    "greece": [
      58,
      109
    ],
    "repeatedly": [
      58,
      116,
      120
    ],
    "avoided": [
      58,
      110
    ],
    "sustain": [
      58,
      59
    ],
    "frequently": [
      58,
      143
    ],
    "cited": [
      58,
      154
    ],
    "oracle": [
      59,
      115,
      116,
      154,
      156
    ],
    "announced": [
      59,
      93,
      96,
      97,
      98,
      150
    ],
    "progress": [
      59,
      61,
      81,
      95
    ],
    "networking": [
      59,
      92,
      95
    ],
    "meanwhile": [
      59,
      68,
      97,
      151
    ],
    "user": [
      59,
      65
    ],
    "software": [
      59,
      155
    ],
    "paid": [
      59,
      82
    ],
    "vary": [
      59,
      158
    ],
    "place": [
      59,
      64,
      68,
      75
    ],
    "parallel": [
      59,
      63,
      65,
      68,
      69,
      70,
      74,
      75,
      81,
      93,
      95
    ],
    "bankruptcies": [
      59,
      73,
      163
    ],
    "downs": [
      59,
      68
    ],
    "transformative": [
      59,
      154
    ],
    "claim": [
      59,
      156
    ],
    "financially": [
      59,
      151
    ],
    "overbuild": [
      59,
      65
    ],
    "begin": [
      59,
      60,
      64,
      66,
      85,
      123
    ],
    "backward": [
      59,
      72
    ],
    "circulated": [
      59,
      154
    ],
    "justifies": [
      59,
      65
    ],
    "generating": [
      59,
      68,
      93,
      156
    ],
    "barrel": [
      60,
      66,
      69,
      93,
      135,
      139,
      151
    ],
    "west": [
      60,
      130
    ],
    "texas": [
      60,
      76,
      133,
      137
    ],
    "intermediate": [
      60,
      72,
      96
    ],
    "attributable": [
      60,
      87
    ],
    "closure": [
      60,
      122,
      123,
      129,
      132,
      137,
      139,
      144,
      150,
      151,
      154
    ],
    "strait": [
      60,
      66,
      152
    ],
    "campaign": [
      60,
      128
    ],
    "transit": [
      60,
      66,
      86,
      122,
      125,
      144,
      148
    ],
    "wave": [
      60,
      69,
      70,
      76,
      157
    ],
    "acting": [
      60,
      116
    ],
    "trapped": [
      60,
      63,
      73,
      118
    ],
    "conflicting": [
      60,
      62
    ],
    "mandates": [
      60,
      62,
      97,
      99
    ],
    "ignore": [
      60,
      116
    ],
    "outcome": [
      60,
      63,
      119,
      154
    ],
    "economically": [
      60,
      67
    ],
    "adjust": [
      60,
      66,
      71
    ],
    "demands": [
      61,
      152
    ],
    "erodes": [
      61,
      80,
      158
    ],
    "reliance": [
      61,
      74
    ],
    "table": [
      61,
      90,
      140
    ],
    "requirements": [
      61,
      70,
      72,
      89,
      98,
      99
    ],
    "tested": [
      61,
      96,
      133
    ],
    "coinciding": [
      61,
      70
    ],
    "fomc": [
      62,
      67
    ],
    "meeting": [
      62,
      72,
      82,
      163
    ],
    "moderating": [
      62,
      67
    ],
    "slowed": [
      62,
      94
    ],
    "reductions": [
      62,
      87
    ],
    "bringing": [
      62,
      67
    ],
    "clean": [
      62,
      147
    ],
    "escape": [
      62,
      67,
      69
    ],
    "injected": [
      62,
      67,
      95
    ],
    "impulse": [
      62,
      67,
      75,
      95
    ],
    "repeat": [
      62,
      95,
      102,
      126
    ],
    "arthur": [
      62,
      63,
      81,
      93
    ],
    "premature": [
      62,
      81,
      117
    ],
    "paul": [
      62,
      101,
      105,
      108,
      151
    ],
    "constraint": [
      62,
      157
    ],
    "violent": [
      62,
      67,
      71,
      75,
      124
    ],
    "recalibration": [
      62,
      67
    ],
    "premised": [
      62,
      67,
      95
    ],
    "fedwatch": [
      62,
      67
    ],
    "governors": [
      62,
      67
    ],
    "doing": [
      63,
      96,
      154
    ],
    "fuel": [
      63,
      66,
      70,
      86,
      92,
      121,
      125,
      132,
      134,
      135,
      136,
      138,
      139,
      143,
      146,
      150
    ],
    "instructive": [
      63,
      68,
      95,
      98
    ],
    "arab": [
      63,
      93,
      124,
      129
    ],
    "overnight": [
      63,
      68,
      94,
      95,
      96,
      97,
      130,
      142
    ],
    "avoid": [
      63,
      73,
      76,
      148
    ],
    "normalize": [
      63,
      146
    ],
    "entered": [
      63,
      64,
      79,
      103
    ],
    "word": [
      63,
      117
    ],
    "previously": [
      63,
      69,
      91,
      129,
      159
    ],
    "believed": [
      63,
      93,
      118
    ],
    "fell": [
      63,
      65,
      75,
      87,
      95,
      97,
      105,
      124,
      140
    ],
    "accelerated": [
      63,
      94
    ],
    "resolution": [
      63,
      87
    ],
    "differences": [
      63,
      80
    ],
    "import": [
      63,
      126,
      141
    ],
    "synthesis": [
      63,
      64,
      65
    ],
    "firmly": [
      64,
      87
    ],
    "flipped": [
      64,
      97
    ],
    "execution": [
      64,
      71
    ],
    "flip": [
      64,
      97,
      158
    ],
    "bust": [
      64,
      120
    ],
    "steepened": [
      64,
      96
    ],
    "marks": [
      64,
      154,
      157
    ],
    "leads": [
      64,
      116
    ],
    "style": [
      64,
      108,
      112,
      151
    ],
    "members": [
      65,
      67,
      72
    ],
    "triple": [
      65,
      136,
      149
    ],
    "justification": [
      65,
      75
    ],
    "materialized": [
      65,
      115
    ],
    "structurally": [
      65,
      73,
      98,
      137
    ],
    "buildout": [
      65,
      92,
      157
    ],
    "took": [
      65,
      93,
      94,
      95,
      96,
      97,
      118
    ],
    "recover": [
      65,
      93,
      94
    ],
    "respect": [
      65,
      76
    ],
    "includes": [
      65,
      72
    ],
    "retirement": [
      65,
      158
    ],
    "breakout": [
      66
    ],
    "exploded": [
      66,
      71,
      94,
      96
    ],
    "surging": [
      66,
      94
    ],
    "kind": [
      66,
      73
    ],
    "exogenous": [
      66,
      97
    ],
    "adaptation": [
      66,
      115,
      116,
      118
    ],
    "impacts": [
      66,
      127,
      134
    ],
    "petrochemical": [
      66,
      139
    ],
    "feedstock": [
      66,
      136,
      143
    ],
    "fertilizers": [
      66,
      140,
      141,
      142
    ],
    "constrains": [
      66,
      80,
      91
    ],
    "resolved": [
      66,
      98
    ],
    "adding": [
      66,
      86,
      88
    ],
    "cargoes": [
      66,
      148
    ],
    "january": [
      67,
      78,
      79,
      80,
      82,
      83,
      84,
      87,
      88,
      92,
      93,
      95,
      97,
      99
    ],
    "catalysts": [
      67,
      93
    ],
    "politically": [
      67,
      77
    ],
    "deteriorated": [
      67,
      84
    ],
    "stayed": [
      67,
      96
    ],
    "facing": [
      67,
      127
    ],
    "extending": [
      67,
      76
    ],
    "waiting": [
      67,
      149
    ],
    "persist": [
      67,
      82
    ],
    "counting": [
      67,
      69
    ],
    "mechanics": [
      68,
      71,
      74,
      75,
      97
    ],
    "deceptively": [
      68,
      79
    ],
    "notional": [
      68,
      71,
      72,
      75,
      96
    ],
    "domiciled": [
      68,
      70
    ],
    "cayman": [
      68,
      70
    ],
    "islands": [
      68,
      70,
      134,
      138
    ],
    "oversight": [
      68,
      70
    ],
    "gross": [
      68,
      154
    ],
    "repurchase": [
      68,
      96
    ],
    "staggering": [
      68,
      88
    ],
    "failed": [
      68,
      94,
      95,
      107,
      118,
      129,
      130,
      153
    ],
    "auction": [
      68,
      98
    ],
    "counterparties": [
      68,
      72,
      96,
      98
    ],
    "haircuts": [
      68,
      77
    ],
    "require": [
      68,
      71,
      157
    ],
    "liquidate": [
      68,
      70,
      72,
      74,
      89,
      96
    ],
    "coordination": [
      68,
      128,
      154
    ],
    "entity": [
      68,
      73
    ],
    "bailed": [
      68,
      107
    ],
    "visibility": [
      68,
      77,
      99,
      100
    ],
    "authority": [
      68,
      70,
      119,
      160
    ],
    "orderly": [
      68,
      78,
      98
    ],
    "wind": [
      68,
      92
    ],
    "council": [
      68,
      73
    ],
    "automatic": [
      68,
      93
    ],
    "prime": [
      68,
      98
    ],
    "directional": [
      68,
      71,
      90,
      154,
      156
    ],
    "freezing": [
      68,
      96
    ],
    "scheduled": [
      69,
      76
    ],
    "line": [
      69,
      77,
      115,
      150
    ],
    "eliminating": [
      69,
      97
    ],
    "downgrades": [
      69,
      96
    ],
    "junk": [
      69,
      70,
      96,
      97
    ],
    "pool": [
      69,
      96,
      97
    ],
    "pension": [
      69,
      72,
      73,
      75,
      97,
      98
    ],
    "mandated": [
      69,
      72,
      99
    ],
    "fallen": [
      69,
      85
    ],
    "angel": [
      69,
      155
    ],
    "insufficient": [
      69,
      73
    ],
    "thousands": [
      69,
      115
    ],
    "collateralized": [
      69,
      96
    ],
    "clos": [
      69,
      70
    ],
    "struggling": [
      69,
      82
    ],
    "tranches": [
      69,
      96
    ],
    "process": [
      69,
      100,
      136,
      141,
      142
    ],
    "surviving": [
      69,
      105
    ],
    "qualify": [
      69,
      155,
      156
    ],
    "definition": [
      69,
      81
    ],
    "consecutive": [
      69,
      82,
      84,
      87,
      88
    ],
    "annuity": [
      70
    ],
    "consequential": [
      70,
      75
    ],
    "life": [
      70,
      75,
      139
    ],
    "apollo": [
      70,
      154,
      155,
      156,
      157
    ],
    "affiliated": [
      70,
      73
    ],
    "structured": [
      70,
      75,
      96,
      159
    ],
    "latest": [
      70,
      81
    ],
    "illiquid": [
      70,
      94,
      96
    ],
    "arbitrage": [
      70,
      74
    ],
    "achieved": [
      70,
      113
    ],
    "offshore": [
      70,
      77,
      90
    ],
    "weaker": [
      70,
      87,
      143
    ],
    "transparency": [
      70,
      72,
      73
    ],
    "reach": [
      70,
      84,
      96
    ],
    "retirees": [
      70,
      98
    ],
    "structures": [
      70,
      73,
      96,
      151
    ],
    "dual": [
      70,
      100,
      152
    ],
    "executive": [
      70,
      160
    ],
    "california": [
      70,
      76,
      144
    ],
    "optimized": [
      70,
      97
    ],
    "liability": [
      70,
      72,
      98
    ],
    "millions": [
      70,
      114,
      148
    ],
    "distressed": [
      70,
      76
    ],
    "gamma": [
      71
    ],
    "volume": [
      71,
      74,
      86,
      136,
      151
    ],
    "barely": [
      71,
      73,
      78,
      83
    ],
    "trillions": [
      71,
      72,
      75,
      97
    ],
    "seeking": [
      71,
      98
    ],
    "dealers": [
      71,
      96,
      97
    ],
    "orders": [
      71,
      82,
      83,
      86,
      87,
      95,
      159
    ],
    "round": [
      71,
      93,
      154
    ],
    "running": [
      71,
      75,
      80,
      82,
      88,
      91,
      93,
      98,
      152
    ],
    "trying": [
      71,
      98
    ],
    "procyclicality": [
      72,
      98
    ],
    "derivative": [
      72,
      96
    ],
    "ensuring": [
      72,
      95
    ],
    "aftermath": [
      72,
      76,
      98
    ],
    "clearing": [
      72
    ],
    "totaling": [
      72,
      119
    ],
    "dash": [
      72,
      97
    ],
    "gilt": [
      72,
      98
    ],
    "gilts": [
      72,
      98
    ],
    "england": [
      72,
      98
    ],
    "halt": [
      72,
      154
    ],
    "amend": [
      73
    ],
    "pretend": [
      73,
      76
    ],
    "product": [
      73,
      135
    ],
    "marketed": [
      73,
      74
    ],
    "lender": [
      73,
      76
    ],
    "technical": [
      73,
      97
    ],
    "added": [
      73,
      131
    ],
    "marking": [
      73,
      81,
      87,
      112
    ],
    "exchanges": [
      73,
      74,
      89
    ],
    "controlled": [
      73,
      99,
      100,
      128
    ],
    "secondary": [
      73,
      154
    ],
    "outright": [
      73,
      84
    ],
    "illusion": [
      73,
      74,
      94,
      153
    ],
    "investing": [
      74,
      117
    ],
    "contains": [
      74,
      98
    ],
    "analogy": [
      74,
      118
    ],
    "uncomfortably": [
      74,
      79
    ],
    "obscure": [
      75,
      116,
      117
    ],
    "threatens": [
      75,
      132
    ],
    "responded": [
      75,
      76
    ],
    "signaled": [
      75,
      96,
      97,
      102
    ],
    "willingness": [
      75,
      116,
      155
    ],
    "normalizing": [
      75,
      80
    ],
    "giving": [
      75,
      78,
      98
    ],
    "catalyzed": [
      75,
      95,
      124
    ],
    "statements": [
      75,
      95
    ],
    "secular": [
      76,
      78,
      93,
      115
    ],
    "headwinds": [
      76,
      90,
      92,
      116
    ],
    "commerce": [
      76,
      86,
      99
    ],
    "compounds": [
      76,
      79,
      89,
      90,
      99,
      148
    ],
    "reflected": [
      76,
      93
    ],
    "math": [
      76,
      116,
      119,
      154
    ],
    "rational": [
      76,
      77,
      145,
      153
    ],
    "maturities": [
      76,
      94
    ],
    "finally": [
      76,
      83,
      93,
      95,
      96,
      144
    ],
    "issue": [
      76,
      95,
      97
    ],
    "unsustainable": [
      76,
      88,
      93,
      94,
      98,
      116
    ],
    "restrict": [
      76,
      94,
      143
    ],
    "restricted": [
      76,
      148
    ],
    "drove": [
      76,
      98,
      155
    ],
    "bailout": [
      76,
      97
    ],
    "green": [
      76,
      125
    ],
    "street": [
      76,
      158
    ],
    "eurodollar": [
      77,
      90
    ],
    "plumbing": [
      77,
      90,
      97,
      118
    ],
    "jurisdiction": [
      77,
      100
    ],
    "sovereigns": [
      77,
      88,
      90
    ],
    "debts": [
      77,
      93
    ],
    "settle": [
      77,
      99
    ],
    "catastrophically": [
      77,
      96
    ],
    "paradoxically": [
      77,
      78
    ],
    "frozen": [
      77,
      91
    ],
    "volatile": [
      77,
      80
    ],
    "hoard": [
      77,
      96
    ],
    "hoarding": [
      77,
      138
    ],
    "shortage": [
      77,
      143
    ],
    "marked": [
      77,
      96,
      159
    ],
    "swap": [
      77,
      90,
      98
    ],
    "greater": [
      77,
      111
    ],
    "dollarization": [
      77,
      88,
      160
    ],
    "accumulated": [
      77,
      93
    ],
    "dependency": [
      77,
      126,
      141,
      155
    ],
    "death": [
      77,
      129
    ],
    "having": [
      77,
      88,
      89,
      98,
      136
    ],
    "stagflationary": [
      78,
      79
    ],
    "defined": [
      78,
      97,
      98
    ],
    "notable": [
      78,
      82,
      83,
      88
    ],
    "drawdown": [
      78,
      93,
      156,
      158
    ],
    "consolidation": [
      78,
      88
    ],
    "panicked": [
      78,
      95,
      119
    ],
    "upward": [
      79,
      81,
      83,
      91,
      98
    ],
    "firing": [
      79,
      85
    ],
    "employers": [
      79,
      85
    ],
    "workers": [
      79,
      85,
      114
    ],
    "population": [
      79,
      80,
      121
    ],
    "jobless": [
      79,
      85
    ],
    "equation": [
      79,
      85
    ],
    "jobs": [
      79,
      131
    ],
    "excluding": [
      80,
      81
    ],
    "refusing": [
      80,
      96,
      145
    ],
    "complete": [
      80,
      96,
      98,
      100,
      127
    ],
    "education": [
      80,
      83,
      131
    ],
    "present": [
      80,
      84
    ],
    "fees": [
      80,
      81,
      99
    ],
    "driving": [
      80,
      90,
      139
    ],
    "owners": [
      81,
      91
    ],
    "accurately": [
      81,
      119
    ],
    "gauge": [
      81,
      89
    ],
    "weakens": [
      81,
      85
    ],
    "plunging": [
      81,
      84
    ],
    "contracting": [
      81,
      82,
      87
    ],
    "textbook": [
      81,
      93
    ],
    "preceding": [
      82,
      98
    ],
    "productivity": [
      82,
      114,
      142,
      157
    ],
    "warrants": [
      82,
      86
    ],
    "responding": [
      82,
      93
    ],
    "organic": [
      82,
      97,
      125
    ],
    "pass": [
      82,
      93
    ],
    "amid": [
      82,
      88,
      89,
      90
    ],
    "durable": [
      82,
      131,
      158
    ],
    "expanding": [
      82,
      87
    ],
    "collapses": [
      83,
      84,
      142
    ],
    "specifically": [
      83,
      100,
      141
    ],
    "reason": [
      83,
      155
    ],
    "psychology": [
      84,
      120
    ],
    "constitutes": [
      84,
      157
    ],
    "delay": [
      84,
      95
    ],
    "compounding": [
      84,
      139
    ],
    "security": [
      84,
      126,
      129
    ],
    "ending": [
      85,
      93
    ],
    "lows": [
      85,
      86,
      91,
      113,
      151
    ],
    "seasonal": [
      85,
      86,
      122
    ],
    "adjustment": [
      85,
      93,
      97,
      128
    ],
    "isolation": [
      85,
      99,
      100,
      127
    ],
    "layoffs": [
      85,
      93,
      95,
      160
    ],
    "bulk": [
      86,
      148
    ],
    "iron": [
      86,
      87
    ],
    "grain": [
      86,
      121,
      123,
      141,
      148
    ],
    "ships": [
      86,
      148,
      152
    ],
    "supported": [
      86,
      154,
      156
    ],
    "hope": [
      86,
      148
    ],
    "feeding": [
      86,
      159
    ],
    "disrupting": [
      87,
      139
    ],
    "schedules": [
      87,
      157
    ],
    "downward": [
      87,
      94,
      122
    ],
    "india": [
      87,
      88,
      125,
      139,
      141
    ],
    "driver": [
      87,
      130
    ],
    "exporting": [
      87,
      93
    ],
    "brazil": [
      87,
      141
    ],
    "devaluation": [
      87,
      93,
      94,
      110
    ],
    "yuan": [
      87,
      110
    ],
    "various": [
      87,
      103,
      117
    ],
    "deployed": [
      87,
      89
    ],
    "corrected": [
      88,
      92
    ],
    "doubled": [
      88,
      94
    ],
    "tonnes": [
      88,
      92
    ],
    "parabolic": [
      88,
      95
    ],
    "fiat": [
      88,
      93,
      101,
      112,
      118
    ],
    "hard": [
      88,
      110
    ],
    "bitcoin": [
      89
    ],
    "digital": [
      89,
      99
    ],
    "stubbornly": [
      89,
      108
    ],
    "evaporates": [
      89,
      155
    ],
    "perpetual": [
      89,
      120
    ],
    "psychologically": [
      89,
      90
    ],
    "technically": [
      89,
      94
    ],
    "euro": [
      90,
      99,
      100
    ],
    "pound": [
      90,
      98
    ],
    "exporters": [
      90,
      141
    ],
    "determined": [
      91,
      93
    ],
    "evaporation": [
      91,
      96
    ],
    "removes": [
      91,
      158
    ],
    "putting": [
      91,
      94
    ],
    "determine": [
      91,
      96
    ],
    "supposed": [
      92,
      100,
      153
    ],
    "grid": [
      92,
      134
    ],
    "fossil": [
      92,
      121,
      125,
      138,
      139,
      142
    ],
    "vehicles": [
      92,
      96,
      138,
      156
    ],
    "electrification": [
      92,
      138
    ],
    "mines": [
      92,
      152
    ],
    "shortfall": [
      92,
      122,
      129
    ],
    "closed": [
      92,
      93,
      124
    ],
    "removing": [
      93,
      153
    ],
    "white": [
      93,
      145
    ],
    "house": [
      93,
      98,
      145
    ],
    "pursued": [
      93,
      94
    ],
    "industrialized": [
      93,
      106
    ],
    "overheated": [
      93,
      95,
      96
    ],
    "precision": [
      93,
      98
    ],
    "june": [
      93,
      94,
      95,
      96,
      146
    ],
    "bill": [
      93,
      97
    ],
    "egypt": [
      93,
      124,
      126,
      141
    ],
    "syria": [
      93,
      130
    ],
    "launched": [
      93,
      99,
      118
    ],
    "coordinated": [
      93,
      95
    ],
    "tied": [
      93,
      96,
      125,
      137
    ],
    "deepened": [
      93,
      94
    ],
    "widened": [
      93,
      94,
      95,
      96,
      97,
      98
    ],
    "surpass": [
      93,
      96
    ],
    "earliest": [
      93,
      94,
      95,
      96,
      97
    ],
    "gave": [
      93,
      94,
      95,
      96,
      97
    ],
    "bureau": [
      93,
      149,
      158
    ],
    "imbalances": [
      94,
      120
    ],
    "prove": [
      94,
      97
    ],
    "facility": [
      94,
      96,
      150
    ],
    "exports": [
      94,
      122,
      133,
      137,
      146
    ],
    "burst": [
      94,
      103
    ],
    "entering": [
      94,
      112
    ],
    "burning": [
      94,
      95
    ],
    "dropped": [
      94,
      98,
      134,
      138,
      155,
      156
    ],
    "announcement": [
      94,
      96,
      98
    ],
    "devalued": [
      94,
      110
    ],
    "instant": [
      94,
      99
    ],
    "defaulted": [
      94,
      128
    ],
    "tripling": [
      95,
      99
    ],
    "burn": [
      95,
      154,
      157
    ],
    "stopped": [
      95,
      96,
      97
    ],
    "half": [
      95,
      98,
      141
    ],
    "passed": [
      95,
      104
    ],
    "article": [
      95,
      128,
      159,
      161
    ],
    "arithmetic": [
      95,
      154
    ],
    "april": [
      95,
      136,
      137,
      138,
      139,
      143,
      144,
      152,
      154,
      155,
      156,
      157
    ],
    "filing": [
      95,
      156
    ],
    "degradation": [
      96,
      100
    ],
    "synthetic": [
      96,
      121,
      132,
      135,
      136,
      139,
      140,
      141,
      142,
      143,
      147
    ],
    "crashed": [
      96,
      98,
      101,
      106,
      142
    ],
    "dedicated": [
      96,
      115
    ],
    "citing": [
      96,
      145,
      146,
      148,
      149,
      154
    ],
    "start": [
      96,
      131
    ],
    "relevance": [
      96,
      99,
      100,
      128
    ],
    "substitute": [
      96,
      147,
      158
    ],
    "closing": [
      96,
      97
    ],
    "seizing": [
      96,
      152
    ],
    "historic": [
      97,
      101,
      113,
      120,
      148,
      151
    ],
    "downgrade": [
      97,
      160
    ],
    "reasons": [
      97,
      116
    ],
    "arrived": [
      97,
      144
    ],
    "western": [
      97,
      127,
      128,
      129,
      130,
      138
    ],
    "crucial": [
      97,
      138
    ],
    "terrifying": [
      97,
      98
    ],
    "bedrock": [
      97,
      158
    ],
    "purely": [
      98,
      118
    ],
    "left": [
      98,
      145
    ],
    "removal": [
      98,
      140
    ],
    "fired": [
      98,
      119,
      151
    ],
    "sovereignty": [
      99,
      100
    ],
    "replacement": [
      99,
      126,
      144
    ],
    "governed": [
      99,
      100
    ],
    "deployment": [
      99,
      157
    ],
    "spend": [
      99,
      128,
      155
    ],
    "enforcement": [
      99,
      157
    ],
    "dispute": [
      99,
      154
    ],
    "sepa": [
      99,
      100
    ],
    "wero": [
      99,
      100
    ],
    "visa": [
      99,
      100
    ],
    "mastercard": [
      99,
      100
    ],
    "hegemony": [
      99,
      100
    ],
    "erosion": [
      99,
      100
    ],
    "degrades": [
      99,
      129
    ],
    "legal": [
      99,
      159
    ],
    "passes": [
      100,
      143
    ],
    "azure": [
      100,
      154
    ],
    "storage": [
      100,
      122,
      123
    ],
    "frameworks": [
      100,
      116
    ],
    "modeling": [
      100,
      123,
      146,
      159,
      162
    ],
    "bugs": [
      101,
      104
    ],
    "howard": [
      101,
      106,
      157
    ],
    "harry": [
      101,
      111
    ],
    "claimed": [
      101,
      103,
      104,
      105,
      108,
      117
    ],
    "stagnated": [
      101,
      118
    ],
    "gemini": [
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115,
      116,
      117,
      118,
      119,
      120,
      132,
      134,
      154
    ],
    "scorecard": [
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115
    ],
    "monday": [
      102
    ],
    "north": [
      104,
      141,
      142
    ],
    "grids": [
      104,
      138
    ],
    "passing": [
      105,
      137
    ],
    "civilization": [
      106,
      115
    ],
    "society": [
      106,
      115
    ],
    "incentivized": [
      106,
      116
    ],
    "shale": [
      106,
      137,
      145
    ],
    "exporter": [
      106,
      133,
      137
    ],
    "michael": [
      107,
      117
    ],
    "john": [
      107,
      118
    ],
    "fraudulent": [
      107,
      117
    ],
    "nouriel": [
      109,
      111,
      120
    ],
    "spain": [
      109,
      163
    ],
    "apocalypse": [
      110,
      114
    ],
    "displace": [
      114,
      157
    ],
    "universally": [
      115,
      139
    ],
    "mistake": [
      115,
      116
    ],
    "wired": [
      115,
      117
    ],
    "survival": [
      115,
      153
    ],
    "bias": [
      115,
      116
    ],
    "doesn": [
      115,
      132,
      141,
      157
    ],
    "facts": [
      115,
      137
    ],
    "guru": [
      116,
      117,
      118,
      119,
      120
    ],
    "oracles": [
      116
    ],
    "hammer": [
      116,
      117
    ],
    "nail": [
      116,
      117
    ],
    "underestimating": [
      116,
      119
    ],
    "print": [
      116,
      128
    ],
    "bullish": [
      116,
      151
    ],
    "bearish": [
      116,
      151
    ],
    "brand": [
      116,
      120
    ],
    "fatal": [
      117,
      151
    ],
    "flaw": [
      117,
      151
    ],
    "missed": [
      117,
      120
    ],
    "seeing": [
      117,
      129
    ],
    "lack": [
      117,
      141,
      143
    ],
    "meredith": [
      119
    ],
    "nexus": [
      121
    ],
    "fuels": [
      121,
      138,
      139,
      142
    ],
    "haber": [
      121,
      136,
      141
    ],
    "bosch": [
      121,
      136,
      141
    ],
    "erisman": [
      121,
      141
    ],
    "farm": [
      121,
      131,
      136,
      149
    ],
    "crop": [
      121,
      132,
      136,
      140,
      142,
      143,
      149
    ],
    "corn": [
      121,
      125,
      140,
      141
    ],
    "wheat": [
      121,
      124,
      125,
      126,
      140,
      141
    ],
    "goes": [
      121,
      139
    ],
    "farmer": [
      121,
      143
    ],
    "smil": [
      121,
      141
    ],
    "seaborne": [
      122,
      125,
      141,
      144,
      150
    ],
    "ammonia": [
      122,
      125,
      126,
      142,
      149
    ],
    "urea": [
      122,
      125,
      141,
      142
    ],
    "shelf": [
      122,
      139
    ],
    "winter": [
      123,
      141
    ],
    "qatari": [
      123,
      126,
      137
    ],
    "developing": [
      123,
      128,
      142
    ],
    "agricultural": [
      123,
      139,
      142,
      143
    ],
    "crossed": [
      124,
      144
    ],
    "libya": [
      124,
      130
    ],
    "stays": [
      124,
      126
    ],
    "canada": [
      125,
      141
    ],
    "offline": [
      125,
      137
    ],
    "bangladesh": [
      126,
      141
    ],
    "mena": [
      126,
      134
    ],
    "importer": [
      126,
      133,
      137
    ],
    "saharan": [
      126,
      138,
      141
    ],
    "farmers": [
      126,
      136,
      141,
      142,
      143,
      149
    ],
    "humanitarian": [
      127,
      129,
      130
    ],
    "port": [
      127,
      134,
      138,
      160
    ],
    "populations": [
      127,
      129
    ],
    "realignment": [
      127,
      128
    ],
    "unhcr": [
      127,
      130
    ],
    "architecture": [
      128,
      157
    ],
    "moloch": [
      128
    ],
    "lanka": [
      128,
      140,
      143
    ],
    "conspiracy": [
      128,
      158
    ],
    "insecurity": [
      129
    ],
    "treat": [
      129,
      130
    ],
    "caloric": [
      129,
      138
    ],
    "treats": [
      130,
      131
    ],
    "weapons": [
      130,
      153
    ],
    "neutral": [
      130,
      152
    ],
    "fine": [
      131,
      155
    ],
    "packaging": [
      132,
      139
    ],
    "pharmaceutical": [
      132,
      139
    ],
    "insulation": [
      133
    ],
    "independence": [
      133,
      137
    ],
    "sweet": [
      133,
      137
    ],
    "sour": [
      133,
      137
    ],
    "bidder": [
      133,
      137
    ],
    "pacific": [
      134,
      138
    ],
    "dominican": [
      134,
      138
    ],
    "fleets": [
      134,
      138
    ],
    "confined": [
      134,
      138
    ],
    "outer": [
      134,
      138
    ],
    "island": [
      134,
      138
    ],
    "communities": [
      134,
      138
    ],
    "forum": [
      134,
      138
    ],
    "gallons": [
      135,
      139
    ],
    "refined": [
      135,
      137,
      139,
      146
    ],
    "distillates": [
      135,
      137
    ],
    "hydrocarbon": [
      135,
      139
    ],
    "liquids": [
      135,
      139
    ],
    "propylene": [
      135,
      139
    ],
    "butadiene": [
      135,
      139
    ],
    "rubber": [
      135,
      139
    ],
    "tires": [
      135,
      139
    ],
    "polyester": [
      135,
      139
    ],
    "clothing": [
      135,
      139
    ],
    "adhesives": [
      135,
      139
    ],
    "paints": [
      135,
      139
    ],
    "detergents": [
      135,
      139
    ],
    "pesticide": [
      136,
      143
    ],
    "solvents": [
      136,
      143
    ],
    "surfactants": [
      136,
      143
    ],
    "insects": [
      136,
      143
    ],
    "season": [
      136,
      140,
      141
    ],
    "tractor": [
      136,
      143
    ],
    "additive": [
      136,
      143
    ],
    "starved": [
      136,
      143
    ],
    "crops": [
      136,
      141
    ],
    "pest": [
      136
    ],
    "harvest": [
      136,
      140
    ],
    "briefing": [
      136,
      137,
      138,
      139
    ],
    "domestically": [
      137,
      146
    ],
    "intake": [
      138,
      142,
      159
    ],
    "rely": [
      138,
      143
    ],
    "plastic": [
      139
    ],
    "scarce": [
      139,
      148
    ],
    "gulf": [
      139,
      141
    ],
    "application": [
      140,
      142
    ],
    "rice": [
      140,
      141,
      142
    ],
    "plummeted": [
      140,
      142
    ],
    "drop": [
      140,
      155
    ],
    "soil": [
      140,
      143
    ],
    "plant": [
      141,
      149
    ],
    "tons": [
      141,
      142
    ],
    "soviet": [
      142,
      150
    ],
    "chemical": [
      142,
      147
    ],
    "plants": [
      142,
      143
    ],
    "numbers": [
      143,
      154
    ],
    "tankers": [
      144,
      148
    ],
    "patrick": [
      144,
      145,
      146,
      147,
      148,
      149,
      150,
      151,
      152,
      153
    ],
    "boyle": [
      144,
      145,
      146,
      147,
      148,
      149,
      150,
      151,
      152,
      153
    ],
    "dallas": [
      145
    ],
    "plan": [
      145,
      159
    ],
    "deployments": [
      145,
      157
    ],
    "peace": [
      145,
      151
    ],
    "sufficiency": [
      146
    ],
    "shipment": [
      147
    ],
    "extraction": [
      147,
      159
    ],
    "choked": [
      147,
      150
    ],
    "panama": [
      148
    ],
    "canal": [
      148
    ],
    "congestion": [
      148
    ],
    "queue": [
      148
    ],
    "waits": [
      148
    ],
    "wait": [
      148
    ],
    "anhydrous": [
      149
    ],
    "kazakh": [
      150
    ],
    "berlin": [
      150
    ],
    "disconnect": [
      151
    ],
    "managers": [
      151,
      158
    ],
    "trump": [
      151,
      152,
      159,
      160
    ],
    "drones": [
      151,
      153
    ],
    "daniel": [
      151,
      153
    ],
    "podcast": [
      151,
      154
    ],
    "blockade": [
      152
    ],
    "restricting": [
      152
    ],
    "hostile": [
      152
    ],
    "vessels": [
      152
    ],
    "navy": [
      152
    ],
    "interdependence": [
      153
    ],
    "integration": [
      153
    ],
    "recoupment": [
      154,
      158
    ],
    "figures": [
      154,
      155
    ],
    "entry": [
      154,
      159,
      161,
      162
    ],
    "verified": [
      154,
      155,
      156,
      157
    ],
    "unverified": [
      154,
      155,
      156
    ],
    "projection": [
      154,
      155
    ],
    "framing": [
      154,
      155,
      162
    ],
    "doctorow": [
      154,
      157
    ],
    "bernstein": [
      154,
      155,
      156
    ],
    "memo": [
      154,
      157
    ],
    "cory": [
      154,
      157
    ],
    "prop": [
      155
    ],
    "torsten": [
      155,
      156
    ],
    "slok": [
      155,
      156,
      157
    ],
    "creditsights": [
      155,
      156,
      157
    ],
    "guidance": [
      155,
      156,
      157
    ],
    "sign": [
      155,
      156
    ],
    "photonic": [
      155,
      157
    ],
    "note": [
      157,
      158
    ],
    "opinions": [
      157,
      158
    ],
    "boomer": [
      158
    ],
    "decumulation": [
      158
    ],
    "demographic": [
      158
    ],
    "court": [
      159,
      160
    ],
    "summary": [
      160
    ],
    "vidbrainz": [
      163
    ],
    "videos": [
      163
    ],
    "video": [
      163
    ]
  },
  "categoryIndex": {
    "framework": [
      0,
      1,
      2,
      3,
      4,
      115,
      116,
      127,
      128
    ],
    "composite-index": [
      5,
      6,
      7,
      8,
      9,
      10,
      52
    ],
    "indicator": [
      11,
      12,
      13,
      14,
      15,
      16,
      17,
      48,
      49,
      50,
      51,
      53,
      54,
      55,
      56,
      57,
      58,
      59,
      60,
      61,
      62,
      66,
      67,
      78,
      79,
      80,
      81,
      82,
      83,
      84,
      85,
      86,
      87,
      88,
      89,
      90,
      91,
      92,
      121,
      124,
      125,
      126,
      131,
      132,
      133,
      134,
      144,
      145,
      146,
      147,
      148,
      149,
      150,
      152
    ],
    "trigger": [
      18,
      19,
      20,
      21,
      22,
      23,
      24,
      25,
      26,
      68,
      69,
      70,
      71,
      72,
      73,
      74,
      75,
      76,
      77,
      99,
      122,
      129,
      137,
      138
    ],
    "historical-precedent": [
      27,
      28,
      29,
      30,
      31,
      93,
      94,
      95,
      96,
      97,
      98,
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      117,
      118,
      119,
      120
    ],
    "compound-interaction": [
      32,
      33,
      34,
      35,
      36,
      45,
      46,
      47,
      63,
      64,
      65,
      100,
      123,
      130,
      135,
      136,
      139,
      151,
      153
    ],
    "institutional-investor": [
      37,
      38,
      39,
      40
    ],
    "emerging-method": [
      41,
      42,
      43,
      44
    ],
    "hormuz-compound-crisis": [
      140,
      141,
      142,
      143
    ],
    "asset-bubble": [
      154,
      155,
      156,
      157,
      158
    ],
    "other": [
      159,
      160,
      161,
      162,
      163
    ]
  },
  "subcategoryIndex": {
    "credit-debt": [
      0,
      1,
      2,
      13,
      16,
      18,
      28,
      33,
      50,
      54,
      56,
      58,
      69,
      73,
      91
    ],
    "systemic-banking": [
      3,
      9,
      10,
      23,
      43,
      55,
      61,
      68,
      70,
      72,
      76,
      96,
      98
    ],
    "general": [
      4,
      5,
      6,
      7,
      8,
      26,
      34,
      35,
      37,
      38,
      39,
      41,
      42,
      44,
      45,
      46,
      47,
      52,
      79,
      82,
      83,
      84,
      85,
      86,
      88,
      89,
      92,
      97,
      140,
      141,
      142,
      143,
      159,
      160,
      161,
      162,
      163
    ],
    "yield-curve": [
      11,
      12,
      48,
      49,
      64
    ],
    "asset-bubble": [
      14,
      15,
      17,
      19,
      27,
      31,
      40,
      51,
      53,
      57,
      65,
      71,
      74,
      78,
      95,
      151
    ],
    "geopolitical": [
      20,
      30,
      87,
      99,
      100,
      150,
      152,
      153
    ],
    "energy-shock": [
      21,
      32,
      60,
      63,
      66,
      93,
      136,
      137,
      139,
      144,
      145,
      146,
      147
    ],
    "currency-inflation": [
      22,
      62,
      67,
      77,
      80,
      81,
      90
    ],
    "ai-specific": [
      24,
      36,
      59,
      154,
      155,
      156,
      157
    ],
    "carry-trade": [
      25,
      29,
      75,
      94
    ],
    "doom-scorecard": [
      101,
      102,
      103,
      104,
      105,
      106,
      107,
      108,
      109,
      110,
      111,
      112,
      113,
      114,
      115
    ],
    "guru-trap": [
      116,
      117,
      118,
      119,
      120
    ],
    "food-energy-nexus": [
      121,
      122,
      123,
      124,
      125,
      126,
      132,
      133,
      134,
      135,
      148,
      149
    ],
    "global-south": [
      127,
      128,
      129,
      130,
      138
    ],
    "labor-market": [
      131
    ],
    "demographic-flow": [
      158
    ]
  },
  "compoundsWithIndex": {
    "asset-bubble": [
      0,
      3,
      5,
      7,
      9,
      13,
      14,
      15,
      17,
      18,
      20,
      23,
      24,
      25,
      26,
      27,
      28,
      30,
      31,
      33,
      34,
      36,
      38,
      39,
      40,
      45,
      47,
      48,
      49,
      50,
      52,
      53,
      56,
      59,
      62,
      64,
      67,
      68,
      71,
      72,
      74,
      75,
      78,
      89,
      95,
      96,
      97,
      107,
      131,
      151,
      153,
      154,
      155,
      156,
      157,
      158
    ],
    "credit-debt": [
      0,
      1,
      2,
      3,
      5,
      6,
      7,
      8,
      9,
      10,
      11,
      12,
      13,
      14,
      15,
      16,
      19,
      21,
      22,
      23,
      24,
      26,
      27,
      28,
      32,
      33,
      34,
      36,
      37,
      39,
      40,
      43,
      45,
      47,
      51,
      52,
      55,
      57,
      59,
      61,
      62,
      63,
      64,
      65,
      66,
      67,
      69,
      70,
      73,
      76,
      79,
      91,
      94,
      95,
      96,
      97,
      100,
      107,
      128,
      131,
      137,
      151,
      154,
      155,
      156,
      157
    ],
    "systemic-banking": [
      0,
      1,
      2,
      3,
      5,
      6,
      7,
      8,
      9,
      10,
      13,
      16,
      17,
      18,
      19,
      25,
      26,
      28,
      29,
      33,
      34,
      36,
      43,
      45,
      47,
      48,
      49,
      50,
      52,
      54,
      56,
      58,
      60,
      62,
      63,
      64,
      66,
      67,
      68,
      70,
      71,
      72,
      73,
      74,
      76,
      77,
      94,
      96,
      97,
      98,
      99,
      100,
      123,
      158
    ],
    "ai-specific": [
      0,
      4,
      14,
      15,
      19,
      27,
      31,
      34,
      35,
      36,
      40,
      45,
      47,
      51,
      54,
      57,
      65,
      78,
      92,
      95,
      154,
      155,
      156,
      157,
      158
    ],
    "currency-inflation": [
      1,
      2,
      5,
      6,
      8,
      11,
      12,
      16,
      18,
      20,
      21,
      23,
      25,
      29,
      30,
      32,
      34,
      37,
      48,
      49,
      50,
      52,
      53,
      54,
      55,
      56,
      58,
      60,
      61,
      63,
      64,
      65,
      66,
      69,
      77,
      79,
      80,
      81,
      83,
      84,
      88,
      90,
      91,
      93,
      94,
      98,
      99,
      100,
      121,
      122,
      124,
      126,
      127,
      128,
      129,
      130,
      134,
      136,
      137,
      138,
      154,
      155,
      157,
      158
    ],
    "yield-curve": [
      1,
      7,
      11,
      12,
      14,
      16,
      18,
      22,
      32,
      37,
      51,
      53,
      54,
      55,
      57,
      58,
      59,
      60,
      61,
      62,
      63,
      65,
      66,
      67,
      68,
      72,
      75,
      81,
      91,
      93,
      95,
      96,
      98,
      131,
      155,
      156,
      158
    ],
    "carry-trade": [
      3,
      5,
      10,
      15,
      17,
      18,
      19,
      20,
      22,
      23,
      24,
      26,
      28,
      29,
      33,
      34,
      36,
      38,
      39,
      75,
      90,
      94
    ],
    "geopolitical": [
      4,
      17,
      21,
      22,
      25,
      29,
      30,
      32,
      34,
      35,
      44,
      46,
      53,
      57,
      58,
      60,
      63,
      66,
      77,
      84,
      86,
      87,
      88,
      90,
      93,
      99,
      100,
      109,
      110,
      121,
      122,
      123,
      124,
      125,
      126,
      127,
      128,
      129,
      130,
      132,
      134,
      136,
      137,
      138,
      139,
      150,
      152,
      153,
      154
    ],
    "energy-shock": [
      4,
      20,
      22,
      30,
      32,
      34,
      35,
      45,
      46,
      48,
      49,
      51,
      52,
      53,
      62,
      64,
      67,
      69,
      72,
      76,
      80,
      84,
      86,
      92,
      93,
      121,
      122,
      123,
      124,
      125,
      126,
      127,
      129,
      130,
      132,
      133,
      134,
      135,
      136,
      138,
      139,
      144,
      145,
      146,
      147,
      148,
      149,
      150,
      151,
      152,
      153
    ],
    "general": [
      4,
      8,
      20,
      21,
      24,
      26,
      34,
      35,
      36,
      37,
      38,
      39,
      41,
      42,
      44,
      46,
      47,
      82,
      85,
      87,
      97
    ],
    "food-energy-nexus": [
      122,
      123,
      125,
      126,
      132,
      133,
      134,
      135,
      144,
      146,
      148,
      149
    ]
  }
}