Speculative feedback
Rising prices attract buyers, whose buying raises prices, which proves the story was right. Until it doesn't.
Economic crises · 33 CE to 2023
Boom.Bubble.Bust.
Every generation believes it has learned the lesson. Then prices rise, debts pile up, and a promise somewhere has to break. This is the story of how prosperity turns into panic, told across two thousand years.
42 crises8 types of collapse6 contagion maps2 hands-on simulations
The Broken Cycle · our symbol
Chapter 1 · Anatomy of a bubble
Seven stages, one shape.
Scroll, and watch the curve that almost every speculative crisis has traced, from Amsterdam taverns in 1637 to subprime mortgages in 2007. Not a law of nature, but a habit of human beings.
Stage 1 · Displacement
Something new appears.
A new trade route, a railway, the internet, a financial invention. It is real, and it is exciting. Early investors are right, which is what makes everyone else pay attention.
Stage 2 · Easy credit
Borrowing makes it bigger.
Banks lend against rising assets; the assets rise because banks lend. In 1929 investors bought shares with 10 percent down. In 2006 some mortgages needed nothing down. Leverage turns a good idea into a dangerous one.
Stage 3 · Euphoria
Prices stop meaning anything.
Valuations detach from income. Rare tulip contracts attract extraordinary bids; a company with no revenue is worth billions. The phrase of the moment is always some version of "this time is different."
Stage 4 · Warning signs
The smart money leaves quietly.
Insiders sell. Interest rates rise. Supply catches up with demand. Nothing dramatic happens yet, because the crowd is still buying what the insiders are selling.
Stage 5 · Trigger
A promise breaks.
A bank fails. A fraud is exposed. A default is announced an ocean away. The trigger is rarely the cause; it is the moment when everyone checks whether the money is still there.
Stage 6 · Panic
Everyone sells at once.
Margin calls force sales; sales push prices down; lower prices trigger more margin calls. Banks stop trusting each other. Liquidity, the ability to turn things into cash, vanishes exactly when everyone needs it.
Stage 7 · Response and recovery
Someone has to catch the fall.
Tiberius in 33 CE, J. P. Morgan in 1907, central banks in 2008: a lender of last resort steps in, or does not. The data recovers first. Jobs, savings and trust take years longer.
Stage 1 of 7DisplacementIllustrative curve, not a price series
Chapter 2 · Why crises happen
Different centuries, familiar weaknesses.
A spark makes the headlines. The fuel is older and quieter: eight mechanisms that show up again and again, alone or combined.
Leverage
Borrowed money multiplies gains on the way up and losses on the way down. A 30 percent fall wipes out anyone who put 20 percent down.
Maturity mismatch
Banks owe money today and own assets that take years to sell. If enough people ask for cash at once, even a sound bank cannot pay.
Currency mismatch
Borrow in dollars, earn in baht. When the baht falls by half, the debt doubles overnight.
Sovereign over-borrowing
Governments borrow against future revenue that never arrives, then refinancing stops. Spain did it four times under one king.
Supply shock
An essential input, oil, grain, or people at work, suddenly becomes scarce. Prices rise while output falls. No bubble required.
Debt deflation
Falling prices make every existing debt heavier in real terms. Borrowers sell assets to pay, which pushes prices down further. Irving Fisher's spiral.
Contagion
Losses travel along loans, trade and fear. A default in Moscow reprices risk in Connecticut; a mortgage in Florida sinks a bank in Iceland.
Chapter 3 · Two thousand years of boom and bust
History moves. Money breaks.
Forty-two episodes from the Roman credit panic of 33 CE to the smartphone bank runs of 2023. Filter by what broke, pick any card for the full story and the source.
Chapter 4 · Same word, different crisis
Compare crises, measure the differences.
Choose three episodes to compare. A market crash, a banking crisis and a recession can have very different effects. Missing historical measures are shown as unavailable.
Choose an episode in each column to update the bars. Figures use different indices where stated. Tulip bid: Museum de Zwarte Tulp catalogue, Rijksmuseum ↗ · 1987 context: Federal Reserve History ↗
1929
The Great Depression
A margin-fuelled crash became four years of bank failures and deflation because nobody caught the fall.
2008
The global financial crisis
Securitised mortgages and overnight funding froze the world's banks; governments and central banks stepped in within weeks.
2020
The pandemic shock
The deepest, fastest fall on record and the shortest recession, because income was replaced almost immediately.
United States figures. Recession length from NBER peak-to-trough dates (Aug 1929–Mar 1933, Dec 2007–Jun 2009, Feb–Apr 2020). Unemployment peaks from BLS and Federal Reserve History estimates; the 1933 figure is an estimate from reconstructed series. Output fall is real GDP peak to trough, rounded; the 2020 figure is the Q4 2019–Q2 2020 drop. Stock drawdowns: Dow Jones 1929–32, S&P 500 for 2007–09 and 2020.
How deep, and how long until the market got back to its peak
Index levels, not adjusted for inflation or dividends. 1929: Dow Jones fell 89% and regained its 1929 peak in November 1954. 1973: S&P 500 fell about 48%, regained July 1980. 1987: S&P 500 fell about 34% from August to December, regained July 1989. 2000: Nasdaq fell 78%, regained April 2015. 2007: S&P 500 fell 57%, regained March 2013. 2020: S&P 500 fell 34% in 33 days, regained August 2020.
Chapter 5 · The crisis lab
Break a bank yourself.
Two teaching models with the arithmetic laid bare. They ignore interest, fees, deposit insurance and bankruptcy law; what they keep is the part that ruins people.
Model A · The leverage trap
Buy an asset worth 100. How much is borrowed, and how far does the price fall?
This is every margin account in 1929 and every zero-down mortgage in 2006, reduced to two numbers.
−10
Your equityDebt still owedValue lost
Model B · The bank run
A bank holds 100 in deposits. How much is cash, and how many depositors want it today?
The bank lent the rest out as loans that will mostly be repaid, but not this afternoon. Each figure below is a depositor holding 5.
Chapter 6 · How crises travel
Fear has no borders.
Pick a crisis and watch it spread. Each path follows a documented channel, a loan book, a currency peg, a trade route, but the arcs are schematic, not measured flows.
Main channel
Chapter 7 · People, not markets
What a crisis actually looks like.
Charts recover before people do. Twelve pictures from the ground, from a Flemish satire of bulb traders to an empty expressway in 2020, then six numbers.
0%
of American workers unemployed, 1933
Roughly one in four. There was no federal unemployment insurance until 1935.
Federal Reserve History · estimate0
US banks failed, 1930 to 1933
Depositors lost their savings outright. Deposit insurance was created in 1933 so it would not happen again.
Federal Reserve History0
marks to buy one US dollar, November 1923
Workers were paid twice a day and spent the money at lunch before it lost value by evening.
Weimar hyperinflation0
US jobs lost in two months, March to April 2020
More than the previous decade of job growth, erased in eight weeks. Most returned within two years; many did not.
Bureau of Labor Statistics0
of Argentines below the poverty line, 2002
Savings were frozen, then converted to devalued pesos. The middle class took to the streets banging pots.
INDEC · 20020
monthly inflation in Zimbabwe, November 2008
Prices doubled roughly every day. The country gave up its own currency the following year.
Cato Institute estimateChapter 8 · Myth or reality
What everybody knows, checked.
Financial folklore is vivid and often wrong. Tap a card to turn it over.
Chapter 9 · The economic cycle
Expansion, peak, contraction, trough. Repeat.
Economies breathe. The question each theory tries to answer is why a normal exhale sometimes turns into a collapse.
Minsky
Long calm periods encourage riskier borrowing, until finance depends on asset prices rising forever. Stability breeds instability.
Fisher
When prices fall, debts grow in real terms. Paying them down pushes prices lower still. The 1930s in one sentence.
Keynes
If everyone saves at once, nobody sells anything. Spending falls, jobs go, and the economy settles below its potential unless someone spends.
Energy, disease, war
Output falls because inputs vanish, not because confidence does. Prices rise at the same time, which makes the usual cures work badly.
Kondratiev
Fifty-year technological waves. Elegant, contested, and useless for prediction. Treat it as a lens, not a clock.
Sources and method
Evidence before folklore.
Every crisis card links to its Wikipedia article. The figures on this page come from the institutions below; the research file lists every episode with its status.
Image credits
Observed, estimated, illustrative
Numbers are labelled by what they are. The anatomy curve and the wave are illustrations; the comparison and recovery bars use index and NBER data; 1933 unemployment is a reconstructed estimate.
Verified vs. to check
Cards marked "verified" rest on the institutional sources listed here. Cards marked "to check" are summarised from encyclopedic sources and need case-specific review before a numeric chart is built on them.
Not to scale
The era bar compresses two millennia so the early episodes are visible. Map arcs show documented channels, not measured capital flows. A trigger is not a cause.