Global financial crisis (2007-2009)

Definition

The 2007-2009 crisis — housing-credit collapse propagating through leveraged, interconnected balance sheets into the deepest US equity drawdown since the Depression (S&P 500 −57% peak to trough) — is the KB’s master case study in SYSTEMIC deleveraging: how funding liquidity, market liquidity, and solvency doubts chain together, and what asset behavior actually looks like when the whole system degrosses at once.

How it works / structure

• The propagation chain (Brunnermeier’s map): housing credit losses → uncertainty about who held them (securitization opacity) → funding runs on shadow-bank balance sheets (repo, commercial paper) → forced asset sales → mark-to-market losses at other institutions → repeat. Liquidity spirals, not the initial loss size, produced the systemic outcome (the FCIC report documents the institutional sequence: Bear, GSEs, Lehman, AIG).
• Market behavior facts (engine-relevant): equity correlation toward one across sectors and geographies (risk-correlation-exposure at maximum); VIX above 80 (regime-volatility crisis-state calibration); Treasuries the one reliably negative-correlated asset (the flight-to-quality regime — ext-bonds-rates); credit spreads as the leading stress gauge; short-selling bans (Sept 2008) changing strategy mechanics mid-crisis (strategy-short-selling regulatory risk).
• The drawdown arithmetic: −57% requires +130% to recover — risk-max-drawdown-budget’s convexity table in lived form; recovery took until 2013.
• Bank-analysis legacy: capital ratios, stress tests (DFAST/CCAR — risk-scenario-analysis canon), and the entire fa-sector-banks metric set are post-2008 artifacts.

When it applies

Cited for crisis-state calibration (correlation, volatility, liquidity parameters at their documented extremes); for the funding-vs-market-liquidity distinction; for financial-sector analysis (the crisis defined its metrics); for drawdown-budget severity floors.

Risk profile & failure modes

• The central lesson: leverage plus opacity converts asset losses into system runs; the equity holder’s risk in a leveraged financial is the LIABILITY side’s confidence, not the asset side’s yield.
• Correlation-of-everything: diversification across risk assets provided little; only duration and cash diversified (port-correlation-budgets stress matrices are calibrated on this).
• Intervention discontinuities: bans, backstops, and rescues rewrote strategy mechanics repeatedly — policy response is a risk factor in both directions.
• Misuse: treating 2008’s Treasury behavior as structural (the 2022 inflation regime inverted it — regime-rate-environments).

Evidence & limits

The FCIC report is the official record; Brunnermeier (2009) is the standard academic map; market data facts are public record. Causal weightings (regulation, monetary policy, fraud, structure) remain politically contested — the KB cites the mechanics, which are documented, and not the blame allocation, which is not settled.

Falsifiable-thesis examples

Illustrations only, not signals:

• “In the next systemic-stress quarter (credit spreads +200bp), this book’s realized correlation to equities stays below 0.5 (diversification-under-stress audit)” — falsified by the episode measurement.
• “Bank X maintains a CET1 ratio above its stress-test minimum through the next DFAST cycle (solvency-buffer thesis)” — falsified by the published results.

Cross-references

• The state calibrations: regime-volatility, risk-correlation-exposure, risk-max-drawdown-budget
• The sector legacy: fa-sector-banks, risk-scenario-analysis (DFAST canon)
• The hedge-asset regime: ext-bonds-rates, regime-rate-environments (and its 2022 inversion)
• The successor stress: episode-banking-stress-2023