A History of Crashes and Recoveries

Every generation of investors gets a crash that feels like the end of the system. Six of them are worth studying closely, because the pattern of what happened, what people believed at the bottom, and how long the repair took is remarkably consistent.

By Karanvir Singh, founder of Quantix Invest

Why crash history is worth your time

Reading about crashes is not morbid curiosity. It is preparation. The single largest determinant of what an ordinary investor earns is not fund selection or asset location or clever tax moves. It is whether they were still holding their portfolio at the bottom of the last bear market. That decision is made under emotional pressure, in real time, with terrible information. The only useful defense is having already decided, calmly, in advance, what you will do. History is how you build that decision.

A few definitions before the tour.

• Correction: a decline of 10% or more from a recent peak. These happen roughly once a year on average in US stocks.
• Bear market: a decline of 20% or more. Historically these arrive every few years, though the spacing is wildly irregular.
• Drawdown: the peak-to-trough percentage loss. This is the number that matters for how a portfolio feels.
• Recovery time: how long from the peak until the portfolio was worth its old peak again. This is where two very different numbers exist, and the difference is enormous.

That last point deserves care. Almost every scary "it took 25 years to recover" statistic uses price indexes only, in nominal terms, ignoring dividends. Real investors received dividends and reinvested them. Real investors also cared about purchasing power, so deflation and inflation change the answer again. Wherever possible, the tables below distinguish price recovery from total-return recovery, because the honest version of history is much less bleak than the version repeated in headlines.

1929: the Great Crash and the Great Depression

What happened. The 1920s produced a genuine technological boom (electrification, automobiles, radio) and a genuine speculative mania on top of it. Stock could be bought on margin with as little as 10% down, which meant a 10% decline could wipe out an investor completely and force liquidation, which pushed prices lower, which forced more liquidation. Prices peaked in September 1929. The famous days came in late October. But the crash itself was not the disaster; the following three years were. The Dow Jones Industrial Average bottomed in July 1932, down roughly 89% from its 1929 peak.

What made it catastrophic. Not the stock market. The policy response was. Roughly a third of US banks failed, taking deposits with them because there was no federal deposit insurance until 1933. The money supply contracted sharply while the Federal Reserve, constrained by gold-standard thinking, did not offset it. The Smoot-Hawley tariff of 1930 helped collapse world trade. Unemployment reached about 25%. Prices fell (deflation of roughly 25% cumulatively), which crushed borrowers and made cash a superb asset.

The recovery numbers, honestly stated. On price alone, in nominal terms, the Dow did not regain its 1929 peak until 1954, which is where the "25 years" figure comes from. That figure is misleading in three ways. Dividend yields in the 1930s were very high, often 5% to 10%, and dividends were a large share of return. Severe deflation meant a dollar in 1932 bought considerably more than a dollar in 1929. And an investor adding money monthly bought their largest share counts at the lowest prices in modern history. Accounting for reinvested dividends and deflation, a lump-sum investor at the exact 1929 peak was back to breakeven in purchasing-power terms in roughly the mid-1930s, and a steady monthly investor was ahead far sooner than that.

The lesson. The permanent damage came from leverage and from bank failure, not from owning diversified equity. Investors who were unlevered, diversified, and still employed enough to keep contributing did fine over the following two decades. The structural fixes that came out of it (deposit insurance, securities regulation, a central bank willing to expand the money supply in a panic) are precisely why later crises did not reproduce it.

1973 to 1974: inflation, oil, and the death of the Nifty Fifty

What happened. The early 1970s combined the end of the Bretton Woods gold link, an OPEC oil embargo that quadrupled crude prices, wage and price controls, and inflation that peaked around 12% in 1974. The S&P 500 fell about 48% from its January 1973 peak to its October 1974 trough. The Dow fell roughly 45%.

The specific bubble. The pre-crash favorites were the "Nifty Fifty," large, high-quality, universally admired growth companies (Xerox, Polaroid, Avon, Coca-Cola, IBM, Kodak) that investors decided were "one-decision" stocks: buy them, never sell, price does not matter. Several traded above 50 times earnings, a handful above 80. The businesses were mostly real. The prices were not. When multiples compressed, many of those stocks fell 70% to 90% even though earnings held up reasonably well. This is the cleanest historical demonstration that a wonderful company can be a terrible investment at the wrong price.

Why it hurt more than the number suggests. This was the worst decade for balanced portfolios in modern US history because bonds offered no shelter. Inflation destroyed the real value of fixed coupons at the same time stocks were falling. In real, inflation-adjusted terms, US stocks did not durably recover their 1973 peak until the early 1980s. Cash lost purchasing power too. There was nowhere good to hide, which is a possibility investors trained only on 2008 tend to underrate.

1987: the one-day crash

What happened. On Monday, October 19, 1987, the Dow fell 22.6% in a single session, still the largest one-day percentage decline in its history. There was no recession, no banking failure, no earnings collapse. The proximate causes were mechanical: "portfolio insurance" strategies that mechanically sold futures as prices fell, computerized program trading, and a market structure that could not handle the resulting order flow. Specialists could not make orderly markets. The tape ran hours behind.

The recovery. This is the most encouraging episode in the set. The S&P 500 finished calendar year 1987 positive, because the year had been up sharply before October. The index regained its August 1987 peak in about two years. An investor who checked their statement annually would barely have noticed that the worst single day in market history had occurred.

The lesson. Not every violent decline signals economic damage. Some are liquidity and plumbing events. The structural response, exchange-wide circuit breakers that halt trading after set percentage declines, is still in force today and is why later panic days paused instead of free-falling.

2000 to 2002: the dot-com bust

What happened. The late 1990s internet boom pushed valuations to levels with no precedent in US data. The Nasdaq Composite roughly quintupled between 1995 and its March 2000 peak. Companies with no earnings and, in some cases, no revenue went public at multibillion-dollar valuations, justified by metrics like "eyeballs" and "page views." The S&P 500's cyclically adjusted price-to-earnings ratio hit about 44, its highest reading ever recorded, well above 1929's.

The damage. The Nasdaq fell about 78% from March 2000 to October 2002. The S&P 500 fell about 49%. Individual names were worse and often permanent: many high-flyers never recovered at all because the businesses did not survive. Even the survivors took a long time. A well-known example is that a large networking equipment company's stock, at its 2000 peak, remained below that peak for over two decades, despite the company continuing to earn substantial profits the entire time.

The crucial distinction. The internet thesis was correct. The internet did change everything. The valuations were still wrong. Being right about the technology and wrong about the price is a complete investment failure. This is worth remembering during every subsequent technology enthusiasm.

What worked. Investors who held anything other than large-cap US growth had a startlingly good stretch. Value stocks, small-cap stocks, real estate investment trusts, and international equities all posted positive returns across 2000 to 2002 while the headline index was cut in half. This is one of the few periods where diversification within equities, not just between stocks and bonds, provided obvious real-time protection.

2007 to 2009: the global financial crisis

What happened. A housing boom financed by loosely underwritten mortgages was packaged into securities, tranched, rated highly by agencies paid by issuers, and then leveraged many times over inside banks and shadow banks. When house prices stopped rising in 2006, the loss cascaded through a system that had no idea where the exposure sat. Bear Stearns was absorbed in March 2008. Lehman Brothers failed in September 2008. Credit markets froze; even ordinary companies briefly could not roll commercial paper.

The damage. The S&P 500 fell about 57% from October 2007 to March 2009, the worst US equity drawdown since the Depression. Global stocks fell similarly. Home prices fell about a third nationally. Unemployment reached 10%.

What held. US Treasuries rallied hard as everything else fell, which is the textbook case for high-quality government bonds as a diversifier. A 60/40 portfolio's drawdown was roughly half the all-equity drawdown. Investment-grade corporate bonds fell modestly; high-yield bonds fell like stocks, confirming that junk bonds are a risk asset wearing a fixed-income costume.

The recovery. The bottom, March 9, 2009, arrived when the news was still uniformly terrible. Unemployment kept rising for months afterward. The S&P 500 returned to its 2007 peak on a total-return basis in roughly three years and on a price basis in about five and a half. Then it produced one of the longest bull markets on record. Anyone who sold in early 2009 to "wait for clarity" missed a first-year rebound of more than 60% off the low, and clarity did not arrive until prices were far higher.

2020: the pandemic crash

What happened. Between February 19 and March 23, 2020, the S&P 500 fell about 34%: the fastest 30%-plus decline in US market history, roughly 23 trading days. Entire industries stopped operating. Oil futures briefly traded below zero. Nobody had a credible model for what was coming.

The recovery. Also the fastest on record. Massive fiscal transfers and immediate central bank intervention (including a commitment to backstop corporate credit markets) stabilized prices. The index regained its February peak by August 2020, about five months from top to new high, and finished the calendar year up about 18% despite the deepest quarterly output collapse since the Depression.

The lesson. The market is a discounting machine looking a year or more ahead, not a scoreboard for current conditions. Stock prices bottomed in March while infections, deaths, and unemployment were all still climbing steeply. Waiting for the news to improve was, once again, a losing strategy. It also demonstrated that recovery timelines are unpredictable: an investor who had internalized "recoveries take five years, so I have time to get back in" would have missed almost the entire move.

The drawdown and recovery table

Approximate figures for US large-cap stocks (Dow for 1929, S&P 500 thereafter). Recovery times are stated on a total-return basis where noted, since that is what an actual investor experienced.

Two things stand out. First, the deepest declines came with either extreme leverage in the system or extreme valuations at the start, and usually both. Second, the length of the decline and the length of the recovery are only loosely related. The 1987 crash was violent and brief. The 1973 bear was slower and far more damaging in real terms.

The arithmetic of staying invested

Two pieces of math govern everything in this guide, and they push in opposite directions.

The first: losses require larger gains to undo. This is why avoiding catastrophic drawdowns matters, and it is the honest argument for owning some bonds.

The formula is simply 1 / (1 minus the loss) minus 1. A 50% loss needs a 100% gain because you are computing the gain on a much smaller base.

The second: the recovery gains are concentrated in a few days, and those days sit next to the worst days. Missing them is far more costly than sitting through them.

Why market timing fails: two worked examples

Worked example 1: the cost of missing the best days. Take a hypothetical $100,000 invested in a broad US index for a 30-year stretch that compounds at 10% per year if you stay fully invested the whole time.

• Fully invested: $100,000 x 1.10^30 = $1,744,940.
• Historical studies across long US periods consistently find that missing just the 10 single best days cuts the annualized return by roughly 2 percentage points. At 8% instead of 10%: $100,000 x 1.08^30 = $1,006,266.
• Missing the 20 best days typically costs about 3.5 points. At 6.5%: $100,000 x 1.065^30 = $661,437.

So sitting out roughly 20 trading days out of about 7,500 cut the final balance by more than 60%. The reason this trap is so effective: the best days cluster inside the worst periods. Several of the largest single-day gains in S&P 500 history occurred in October 2008 and March 2020, in the middle of the panics. To capture them you had to be holding while it felt worst. An investor who sold on a terrible day and planned to return "when things calm down" was structurally guaranteed to miss the rebound, because the rebound is what calming down looks like.

Worked example 2: the world's unluckiest investor. Suppose someone had a genius for terrible timing and invested a lump sum only at the exact market peak before each of the last several crashes. Say they put $10,000 into a total US stock index at the top in 1973, again at the top in 1987, again in March 2000, again in October 2007, and again in February 2020. Every single purchase was immediately followed by a bear market, several of them severe. They never sold, and they reinvested dividends.

Because US stocks have compounded at roughly 10% nominal over long horizons and every one of those declines was eventually recovered, each of those purchases is worth far more today than what was paid. The 1973 purchase, held through the entire inflation bear, the 1987 crash, the dot-com bust, 2008, and 2020, would have grown many times over. The 2007 purchase, made at the top of the worst crash since the Depression, roughly quadrupled or better over the following seventeen years. The point is not the precise multiples, which depend on your index and end date. The point is that perfectly bad timing plus never selling still produced strong long-run results, because the compounding period was long and the holding was never interrupted.

Compare that to the alternative failure mode: good timing on the way in, panic on the way out. The investor who bought sensibly but sold in March 2009 and returned in 2013 locked in the loss and missed the recovery. Selling is the mistake, not buying at the wrong time.

What actually repeats

The specific trigger is always novel. The structure is not. Six patterns recur across all six episodes.

1. Leverage turns a decline into a collapse. 1929 had 10% margin. 2008 had banks running 30-to-1 balance sheets. When borrowed money is forced to unwind, selling begets selling regardless of fundamentals. Crashes without leverage (1987, 2020) repaired quickly. Crashes with leverage (1929, 2008) did lasting economic damage.

2. The story is always that this time the old rules do not apply. "Permanently high plateau" in 1929. One-decision stocks in 1972. Eyeballs instead of earnings in 1999. Housing never falls nationally in 2006. The phrase changes; the function is identical, which is to excuse paying prices that historic norms cannot justify.

3. Starting valuation drives the depth. The deepest declines followed the most extreme starting valuations. This says nothing useful about when a decline will occur, which is why valuation is a poor timing tool, but it says a great deal about how bad it will be when it comes.

4. The bottom occurs while the news is still awful. March 2009 and March 2020 both bottomed with the data still deteriorating. Prices lead the economy. Waiting for confirmation means buying higher.

5. Something usually works. Treasuries in 2008. Value and small-cap and international in 2000 to 2002. Commodities and TIPS in the inflation shocks. Cash in the 1930s deflation. No single asset covers every scenario, which is the entire argument for holding more than one.

6. The recovery is invisible until it is obvious. By the time a recovery is confirmed on the front page, most of the gain has happened.

Common mistakes

Treating "stocks always recover" as a law. They have, in the diversified US index, so far. Individual stocks frequently do not: many dot-com companies never came back and the businesses no longer exist. Whole national markets have delivered decades of disappointment. Japan's Nikkei peaked in 1989 and took about 34 years to reach that level again in nominal terms. Diversify globally and own indexes rather than single names; the recovery pattern belongs to broad markets, not to whatever you happen to hold.

Confusing volatility with risk. A 34% drawdown you sit through is a temporary quote. A 34% drawdown you sell into is a permanent loss. The market's price movement is the same either way; the risk was created by your behavior and your time horizon, not by the index.

Holding money in stocks that you need within a few years. The single legitimate reason to fear a drawdown is being forced to sell during one. Retirement in eighteen months, a house down payment next spring, and tuition due next year do not belong in equities. This is what the bond and cash allocation is for.

Learning only the most recent crash. Investors trained by 2008 expect bonds to save them, and 2022 punished them. Investors trained by 2020 expect a five-month recovery and get impatient. Study all of them; the range of outcomes is wider than any single memory.

Deciding your risk tolerance during a bull market. Everyone is comfortable with 100% stocks when stocks go up. Set the allocation by asking what you would do if the balance fell by half, then assume your real answer is somewhat worse than your imagined one.

Trying to "sit this one out" and come back. See the worked examples above. The cost of being out for the twenty best days is larger than the benefit of dodging an average bear market, and those days are hidden inside the panic.

Bottom line. Six crashes, six different causes, six different recovery timelines, and one consistent conclusion: the diversified investors who kept contributing and never sold ended up fine, and the ones who used leverage, concentrated in the story stock of the era, or sold at the bottom did not. You cannot control when the next one arrives. You can control how much of your money is in stocks, how much is in bonds, whether you owe anyone money, and whether you have already decided to do nothing.

This is educational material, not individualized financial advice. Your own allocation depends on facts about you (horizon, income stability, obligations, temperament) that no general guide can know.