The crises of 1929, 2008, 2020, 2022 - anatomy of a crash

Every crisis looks different and every crisis has the same structure. If you learn the anatomy of a bubble, you will recognise it in any era, whether it is railway stocks in 1840, dot-coms in 1999, real estate in 2007 or SPACs in 2021. History does not repeat itself, but it rhymes - Mark Twain. A trader who does not know the anatomy of a crash pays into every one of them.

The 5 phases of a bubble - Minsky’s model

The economist Hyman Minsky described a cycle that repeats every 7-10 years:

Phase 1: Displacement A new technology, regulation, a monetary shock. For example the internet in 1995, low Fed rates at 1% in 2003, QE plus zero rates in 2020.

Phase 2: Boom Prices rise, the media write about it, the first participants make money. Credit loosens. Leverage grows.

Phase 3: Euphoria “Everyone has to own it.” A taxi driver recommends stocks. Valuations detach from fundamentals. New metrics to justify them: eyeballs in 1999, price per square metre rising 30% year on year in 2007, “community adjusted EBITDA” in 2019. The “this time is different” phase.

Phase 4: Distribution / Unease (Profit taking) Smart money sells. Volume rises while volatility falls. Cracks appear: Bear Stearns in March 2008, Evergrande in 2021. But the narrative is still bullish.

Phase 5: Panic (Minsky Moment) The moment when everyone wants to leave at once and liquidity disappears. Margin calls, no buyers. A 30-50% decline within weeks.

1929 - the Great Depression - the lesson of leverage

Mechanics: in the 1920s you could buy stocks on 10% margin (10:1 leverage). When the market rose 20% a year, the return on equity was 200%. When it fell 10% in a single day, margin call and bankruptcy. Banks had no FDIC, 9,000 banks failed. US GDP -29%, unemployment 25%.

The policy mistake: the Fed raised rates in 1928-29 to curb speculation, then did not add liquidity after the crash. A 30% contraction in the money supply deepened deflation. Milton Friedman’s lesson: the Great Depression was a failure of the Fed, not of capitalism.

The regime for a trader: buy and hold dies for 10 years. The DJIA returned to its 1929 peak only in 1954 - 25 years.

2008 - the Global Financial Crisis - the lesson of securitisation

Anatomy:

1. 2001-2004 the Fed cuts from 6.5% to 1% after the dot-com bust. Cheap credit.
2. Banks create MBS - packages of subprime loans rated AAA (the mistake of the rating agencies).
3. NINJA loans - No Income, No Job, no Assets - and you still get credit.
4. US house prices rise, because everyone can buy. Leverage at investment banks 30:1.
5. 2006 prices stall. 2007 Bear Stearns funds collapse. 2008 Lehman - nobody wants to buy MBS, nobody knows what they are worth.

The mechanics of contagion - how 2% of the subprime market wrecks the world:

Bank A holds USD 10 billion of MBS, leverage 30:1 → capital USD 0.33 billion
A 5% fall in MBS → a loss of USD 0.5 billion > capital → insolvent
Other banks do not know who holds how much MBS → they lend to nobody → the interbank market freezes (LIBOR-OIS spread from 10bp to 350bp)
The real economy loses credit → recession.

In Poland: CHF/PLN jumped from 2.0 to 3.5, because of CHF mortgage borrowers (frankowicze) plus the carry trade. WIG20 -70%, because Poland as an emerging market loses capital.

The lesson: a liquidity crisis turns into a solvency crisis within 72 hours. When there is no buyer, the price is 0 regardless of “value”.

2020 - the COVID crash - the lesson of speed

The fastest crash in history: -34% in 33 days. The fastest recovery: new highs within 5 months. Why?

• A supply shock and a demand shock at the same time.
• A gigantic monetary and fiscal response: the Fed added USD 5 trillion to its balance sheet, the US government sent USD 3,200 cheques per person, a total stimulus of 25% of GDP. In Poland the anti-crisis shields (tarcze) were about 10% of GDP.
• Algorithms and passive funds accelerate moves.

The lesson for a trader: in 2020 what counted was cash on the day of the crash. Those who had a plan “what I will buy when the S&P 500 is -30%” made money. Those who had margin went bankrupt.

2022 - the inflationary bear market - the lesson of regime

This was not a credit crash, but a change of regime: from disinflation and 0% rates to inflation and 5% rates within 12 months. Valuations:

Price = Earnings / (discount_rate) simplified.

When the rate goes from 0% to 5%, the P/E multiple falls from 30 to 15 for the same earnings.
NASDAQ -38%, because it is long duration (earnings far away).
Long-term bonds TLT -33% - the worst year for bonds in history.

Polish 10Y bonds went from 1.2% to 8.5% in a year - the price fell 35%. Many “safe” investors lost more than they would have on stocks.

Warning indicators - a crisis dashboard

Build yourself a table and update it every month:

None of them gives you day-level timing, but the combination gives you the regime.

Crisis_probability ~ f(system_leverage, valuations, liquidity, concentration)

When 3 out of 4 are red - cut leverage, raise cash.

When the plan says: do not trade

1. When the VIX is >40 and you have no plan for a gap. On such days the market can open 5% lower, the broker widens the spread 10x, and stop-losses are jumped over. This is not an environment for a beginner day trader.
2. When correlations go to 1. In a panic everything falls together - stocks, gold, crypto, even bonds in 2022. Diversification does not work for 2 weeks. If you have no cash, you have no options.
3. When you try to catch the bottom without a ladder. Buying all-in after a -30% fall because “it is cheap already” ends with the market falling another -30%. 1929-32 had 5 false bounces of 20%+ before the final -86% bottom.

Mistakes that cost the most

Mistake 1: Averaging down without a limit (martingale). In 2008 banks fell 90%, then another 90% of what was left (for example Citigroup from USD 55 to USD 1). A ladder without a stop and without a capital limit is bankruptcy. A professional has it defined up front: how many tranches, at what levels, what maximum percentage of the portfolio.

Mistake 2: Believing in mean reversion during a credit crash. Mean reversion works in a sideways market. In a crash momentum rules - declines accelerate. You need a regime rule: if the S&P 500 is <200DMA and the curve is inverted, switch mean reversion off.

Mistake 3: Ignoring liquidity risk. In 2020 the spread on WIG20 futures jumped from 2 points to 30 points. In small caps there were no buyers at all. If your position is 20% of a stock’s daily volume, you will not get out without -20% slippage. Measure ADTV (average daily trading volume) and do not exceed 5%.

A measurable exercise

Goal: Survive a crash on paper.

Steps:

1. Take your current portfolio (real or demo).
2. Simulate the 2008 scenario: -50% on stocks, corporate bonds -20%, PLN weakening 30% against USD, volatility up 3x.
3. Calculate:

Loss = Stock_positions*0.5 + Bonds*0.2 + Margin_call_cost (if leveraged)
Will you survive without a margin call?
How much cash is left?

4. Do the same for 2022: stocks -25%, long bonds -30%, cash -10% through inflation.
5. Write down the conclusion: what maximum leverage and what minimum cash cushion (for example 20-30%) lets you survive both scenarios without liquidation?

If you do not survive on paper, you will not survive in the market.

Test yourself

1. List the 5 phases of a Minsky bubble and match them to 2007-2008 in the US housing market.
2. Why did the collapse of one bank (Lehman) freeze the entire interbank market? Explain the role of uncertainty and 30:1 leverage.
3. Compare the speed of the 1929, 2008 and 2020 crashes. What accelerated the moves in 2020 compared with 1929?
4. Calculate the effect of a rise in the discount rate from 1% to 5% on the valuation of a growth company with a P/E of 40. Use the simplified formula Price = Earnings / rate.
5. What is a yield curve inversion and why is it a recession indicator? Give the mechanism.

Frequently asked questions

Can a crash be predicted? A regime of elevated risk can be predicted; the day and the trigger cannot. Minsky’s model gives you the framework, the indicators (curve, spreads) give you 6-18 months of timing, but they never give you the day.

Why did the market rebound so quickly in 2020 and not in 1929? Because in 1929 the Fed was tightening, while in 2020 it eased without limit plus direct fiscal support. Policy. The market in 2020 was pricing not today’s economy, but tomorrow’s liquidity. That is the lesson: do not fight the Fed when it is printing USD 120 billion a month.

Can Poland have a crisis like the one in 2008? Poland has a different system: less securitisation, more loans held on bank balance sheets, and a fairly conservative supervisor, the Polish Financial Supervision Authority (KNF). But it has currency risk (CHF before 2015), interest-rate risk (variable-rate loans) and growth risk (exports to Germany). The crisis will come by a different route, but it will come - because the credit cycle exists.

How do you tell a -10% correction from a -30% bear market? Correction: no recession, company earnings are growing, the Fed is not hiking aggressively. Bear: a recession in the forecasts, earnings falling for 2 quarters, an inverted curve, rising unemployment. The difference is in the duration and in how defensive sectors behave.

---