How Recessions Begin
Recessions originate in a contraction of spending and production, but the trigger varies by cycle, and the labels post-applied by economic historians can give the misleading impression that the cause was obvious at the time. It rarely was.
- Monetary tightening. The Federal Reserve raises rates to combat inflation, the cost of credit rises across the economy, business investment slows, household borrowing softens, and aggregate demand contracts. The 1980-1982 double recession is the canonical case: Paul Volcker pushed the federal funds rate above 19 percent to break double-digit inflation, and the economy contracted twice in three years. The point is sometimes treated as if Volcker had a free hand, which he did not; the political cost of the disinflation was considerable, and there were periods when one could not have been confident the policy would survive its own consequences.
- Financial crisis. A breakdown in the banking or shadow-banking system freezes credit and destroys household and corporate wealth more or less simultaneously. The 2008 recession was triggered by the subprime mortgage collapse and accelerated by the failure of Lehman Brothers in September 2008. The mechanism is rather harder to reverse than monetary tightening because the damage is to balance sheets rather than to incentives.
- Asset bubble burst. When a speculative bubble unwinds — dot-com equities in 2000, residential real estate in 2007 — the wealth destruction transmits into reduced consumption and capital expenditure, with a lag that is variable and a magnitude that depends on how widely the bubble’s gains had been distributed.
- External shock. Oil price spikes (1973, 1979), pandemics (2020), and geopolitical disruptions can push an economy into contraction by raising input costs or by suppressing activity outright. These are the easiest recessions to identify after the fact and the hardest to forecast in advance.
Most cycles involve more than one of these factors operating at once, which is part of why the NBER takes its time. By the time a starting date is announced, the underlying causes have generally been compounding for months or years, and the announcement is closer to a piece of historical record-keeping than a piece of news.
Warning Signs
A handful of indicators have a track record of preceding recessions before the contraction is formally dated. None of them is sufficient on its own, and a portfolio decision rests on the cluster rather than any single line.
| Indicator | Warning Signal | Lead Time |
|---|---|---|
| Yield curve inversion | 10-year Treasury yield falls below 2-year yield | 12-18 months |
| Rising initial jobless claims | 4-week average sustained above 300K | 3-6 months |
| ISM Manufacturing below 50 | Sustained readings below the expansion threshold | 1-3 months |
| Leading Economic Index (LEI) | Three or more consecutive monthly declines | 6-12 months |
| Consumer confidence drop | Sharp decline from a recent peak | 3-6 months |
The yield curve inversion is the most reliable single indicator, in the sense that every US recession since 1955 has been preceded by an inversion of the 10-year/2-year spread. The lag from inversion to the start of contraction has typically been twelve to eighteen months, which is a wide enough window for portfolio managers and policy staff to make adjustments and a long enough one for the indicator to feel wrong for most of its waiting period. The 2022 inversion produced a more contested recession dating debate than usual, and the signal is not infallible. It is, however, treated as a serious data point by the FOMC itself, and the historical record is strong enough that one would be bloody-minded to ignore it on the grounds that this time looks different.
US Recessions Since 1980
| Period | Duration | Peak Unemployment | S&P 500 Decline | Cause |
|---|---|---|---|---|
| Jan–Jul 1980 | 6 months | 7.8% | −17% | Fed rate hikes against inflation |
| Jul 1981–Nov 1982 | 16 months | 10.8% | −27% | Volcker disinflation |
| Jul 1990–Mar 1991 | 8 months | 7.8% | −20% | S&L crisis and Gulf War oil shock |
| Mar–Nov 2001 | 8 months | 6.3% | −49%* | Dot-com bust and 9/11 |
| Dec 2007–Jun 2009 | 18 months | 10.0% | −57% | Subprime mortgage crisis |
*The S&P 500 figure includes the broader bear market; the 2001 recession itself was mild in GDP terms.
The post-1980 average recession runs about eleven months, and the average peak-to-trough decline in the S&P 500 across recessionary periods is roughly thirty percent. Those averages conceal a fairly wide dispersion: 2001 was mild for the real economy but severe for technology-heavy portfolios, while 2008-2009 was severe on both fronts. Duration and equity drawdown do not track each other linearly, and the popular intuition that a long recession produces a deep market decline is supported by the data only loosely.
Recessions and the Stock Market
The relationship between recessions and equities is structured by the market’s forward-looking character, which is why the calendar of price moves and the calendar of GDP releases never quite line up.
- Equities decline before the recession is dated. The S&P 500 typically begins its drawdown six to nine months before the NBER announces a starting date. By the time the word recession enters the headlines, a substantial share of the equity decline has already happened, and the investor reading the announcement is being told what the market knew rather earlier.
- Equities bottom before the recession ends. The market low generally precedes the trough in GDP by three to six months. Prices recover while unemployment is still rising, earnings are still depressed, and the news flow is uniformly bleak. This is the phase of the cycle that feels most counter-intuitive while it is happening, and most obvious in retrospect.
- The recovery move is steep. The twelve months following a cyclical bear-market low have produced average gains of forty to fifty percent across post-1945 recessions. Investors who queue for an unambiguous all-clear have, on the historical record, missed most of that recovery.
The timing structure is one of the more useful pieces of information a long-horizon investor can absorb. Favourable entry points coincide with the moments the economy feels worst, and the discomfort is not a defect of the strategy but the reason the prices are available. There is something of the Test match in this: the period when the score looks worst is rather often the one in which the position is being quietly rebuilt, and the spectator who walks out at tea misses the relevant session.
Investing Through a Recession
- Shift toward defensive sectors. Sector rotation into healthcare (XLV), consumer staples (XLP), and utilities (XLU) tends to provide relative protection during contractions, since demand for medicine, food, and electricity is comparatively inelastic to the cycle. Relative protection is the operative phrase; defensive sectors decline in absolute terms during severe drawdowns.
- Raise cash gradually. When the warning indicators activate — inversion, rising claims, weak ISM — trim positions in stages. Going fully to cash at the first signal is a mistake the historical record punishes, since markets often continue to advance for many months after the initial warnings appear.
- Build a watchlist. Use the contraction to assemble a list of quality businesses trading at compressed valuations. These tend to be the stronger performers in the subsequent expansion, and the time to do the work is before the prices arrive, not after.
- Avoid leverage. Margin positions during recessions are unforgiving. Drawdowns are deeper, volatility is higher, and the margin call arrives at the worst possible price. Cash-financed positions only.
- Keep contributing. For investors with horizons of ten years or more, continuing to add capital through a recession via dollar-cost averaging into broad index funds or ETFs has produced strong long-term outcomes. The shares purchased during the contraction are bought at compressed prices that compound across the next expansion, and this is the structurally simplest way an ordinary investor benefits from the cycle without trying to time it.
Recession versus Depression
A depression is a severe and prolonged recession, conventionally defined by a peak-to-trough GDP decline greater than ten percent or a contraction lasting more than three years. The United States has had only one event that meets the threshold: the Great Depression of 1929-1933, in which real GDP fell roughly thirty percent and the unemployment rate reached twenty-five. The development of modern monetary and fiscal policy tools — an active Federal Reserve able to cut rates and expand its balance sheet, and a federal government able to run substantial deficits without immediate funding crises — has made a depression-scale event materially less likely, though not impossible. The 2008-2009 contraction came closer to that threshold than any other postwar episode, and the policy response in the autumn of 2008 was, in the round, an attempt to prevent precisely that escalation (one imagines the next serious test of these tools will arrive sooner than the policy community would prefer). The base case is that the lesson holds; the tail risk lies in discovering, at an inconvenient moment, that it does not.
See also: GDP · The Federal Reserve and Interest Rates · What Is a Bear Market? · The 2008 Financial Crisis · Sector Rotation
