Financial Crises

graph TD
  %% Phase 1: The Boom
  CheapMoney[Cheap money from<br/>central banks] -->|easy loans| MoreBorrowing[Everyone borrows more]
  MoreBorrowing -->|drives up prices| PricesRise[Prices go up<br/>houses, stocks]
  MoreBorrowing -->|lax lending| TooMuchDebt[Banks and investors<br/>take on too much debt]
  PricesRise -->|collateral looks valuable| MoreBorrowing

  %% Phase 2: The Trigger
  SomethingBreaks[Something breaks<br/>Fed raises rates / big default / shock] -->|confidence cracks| Panic[People panic]
  Panic -->|everyone sells at once| FireSale[Prices crash<br/>forced selling]
  TooMuchDebt -->|debts exceed value| Insolvency[Banks or countries<br/>go bust]

  FireSale -->|prices fall more| FireSale

  %% Phase 3: The Contagion
  Insolvency -->|depositors demand cash| BankRun[Everyone wants<br/>their money back NOW]
  BankRun -->|banks hoard cash| LendingFreezes[Banks stop<br/>lending to each other]
  LendingFreezes -->|mistrust spreads| PanicSpreads[Healthy institutions<br/>also get dragged down]
  PanicSpreads -->|more withdrawals| BankRun

  %% Phase 4: The Fallout
  LendingFreezes -->|can't borrow| CreditCrunch[No one can borrow<br/>businesses starve]
  CreditCrunch -->|layoffs, closures| Recession[Recession<br/>companies close, jobs lost]

  %% Protection
  CentralBank[Central bank<br/>prints money to lend] -->|emergency loans| LendingFreezes
  Gov[Government<br/>guarantees deposits] -->|calms fear| BankRun
  Rules[Rules forcing banks<br/>to keep cash buffers] -->|absorbs losses| Insolvency

How banks work

Banks transform short-term deposits into long-term loans — maturity transformation.

Simplified balance sheet (10% reserve ratio):

Normally 1-2% of depositors show up daily. Reserves cover it.

Vulnerability: assets (loans) are illiquid, liabilities (deposits) are demandable. If a depositor believes others will run, running is rational — even if the bank is solvent.

Diamond-Dybvig model (Nobel 2022)

Two equilibria:

1. Good: everyone waits → bank functions, depositors get full value
2. Bad: everyone runs → fire sales → bank insolvent → depositors lose

Deposit insurance eliminates equilibrium 2. Before FDIC (1933), runs were rational.

2008 Global Financial Crisis

Shadow banking

By 2007, ~50% of lending was through shadow banks — no deposits, no deposit insurance, no Fed access:

Repo works like a deposit: “I lend $100M overnight, you give me Treasuries as collateral.” Rolled daily. If collateral value is doubted, lender raises haircut or refuses to roll.

The trigger

• 2004-2006: 5M subprime mortgages originated → packaged into MBS → CDOs → CDO-squared. Rating agencies gave AAA to toxic pools.
• 2006: US house prices peak.
• 2007: Delinquencies rise. BNP Paribas freezes redemptions (Aug 9 — first signal). Bear Stearns hedge funds fail.
• Mar 2008: Bear Stearns acquired by JPMorgan (Fed guarantees losses).

September 2008 — everything breaks

Why Lehman was catastrophic

Lehman held $600B assets and was counterparty to everyone — derivatives, repo, swaps.

1. Repo market froze: lenders demanded Treasuries-only. Non-Treasury collateral unusable. Hedge funds couldn’t fund → forced selling → crash.
2. CDS chain: AIG wrote $440B in CDS on MBS. Lehman triggered margin calls AIG couldn’t meet.
3. CP market froze: GE, Caterpillar couldn’t roll commercial paper to fund payroll. Wall Street crisis → Main Street recession.

The plumbing that saved the world

Key insight: 2008 was a wholesale funding run — not retail depositors, but institutional money (MMFs, repo, CP). Same logic as a bank run, no deposit insurance.

Post-2008 regulation

Great Depression (1929-1933)

• Oct 1929 crash: stocks -40% from peak.
• 1930-1933: Four banking panics. 9,000 banks failed (40% of total).
• No deposit insurance. Runs were rational.
• Fed did nothing — even raised rates in 1931 to defend gold standard.
• Money supply fell 33%. Deflation ~10%/year.
• Friedman & Schwartz (1963): failure of the Fed to provide liquidity.

Lesson learned: 2008 Fed acted aggressively because the Great Depression showed the cost of inaction.

How sovereign debt crises differ

A country with its own currency can’t have a classic bank run (it can print money). But:

Minsky’s Financial Instability Hypothesis

Hyman Minsky argued that financial systems are inherently unstable — stability breeds instability. Over a prolonged expansion, three types of borrowing units emerge:

1. Hedge units: cash flows cover both principal and interest. Safe, conservative.
2. Speculative units: cash flows cover interest but not principal. The borrower must roll over debt — relies on continued access to credit.
3. Ponzi units: cash flows cover neither principal nor interest. The borrower relies entirely on asset price appreciation to repay. The slightest disruption triggers default.

As the expansion matures, the composition shifts from hedge → speculative → Ponzi. The system becomes fragile. When the first Ponzi units fail, asset prices fall, which forces more units into distress, and the system collapses into a debt-deflation spiral.

Minsky’s key insight: stability is destabilizing. The longer the good times last, the more risk is accumulated in the tail of the distribution. This contrasts with the prevailing pre-2008 view that markets are self-stabilizing.

The 2008 crisis is a textbook Minsky moment: the housing boom shifted speculative and Ponzi finance to unprecedented scale (subprime mortgages, CDOs, SIVs), and the first price declines triggered a chain reaction. Post-crisis regulation (Dodd-Frank, Basel III) attempts to contain this cycle, but regulatory arbitrage continually pushes risk into less-regulated parts of the system.

Crisis propagation pattern

Credit boom → bubble → trigger → fire sales → insolvency → run → freeze → credit crunch → recession

Each crisis follows this pattern. The only difference is the trigger and the type of run:

• 1929: stock crash → retail bank run
• 2008: housing crash → wholesale funding run
• 2020 (COVID): health shock → corporate bond/Treasury dash-for-cash run (Fed stepped in immediately)
• 2023 (SVB): rate hike → unrealized bond losses → social-media-fueled depositor run on uninsured deposits

2022 UK Gilt Crisis: A Modern Shadow-Banking Run

In September 2022, the UK government announced large unfunded tax cuts (the “mini-budget”). Bond markets panicked — gilt (UK government bond) yields surged by 100+ basis points in days. This triggered margin calls on LDI funds (liability-driven investment), which are pension fund hedging vehicles:

1. UK pension funds use LDI funds to hedge interest rate risk — they hold gilts and interest rate swaps as collateral
2. When gilt yields spike, the value of LDI collateral falls → LDI funds face margin calls on their swap positions
3. To meet margin calls, LDI funds must sell gilts → selling gilts pushes yields even higher → more margin calls
4. The Bank of England was forced into emergency gilt purchases — quantitative easing to stop a crisis, not to stimulate the economy

The 2022 gilt crisis is a classic shadow-banking run: maturity and liquidity transformation in non-bank entities (LDI funds), with no central bank backstop, requiring emergency intervention. It closely mirrors the 2008 repo run (see shadow-banking.md), confirming that the crisis propagation pattern is the same regardless of the trigger.

Related

• Capital Flows & Crises — hot money, Asia 1997, EM crises
• Shadow Banking — money market funds, repo, SIVs
• Sovereign Debt & IMF — default mechanics, IMF bailouts
• The Dollar System — Fed swap lines, reserve currency