Market Crashes and Recovery
A 30% fall happens about once a decade. How long past crashes took to recover, and what to do, and not do, while one is happening.
Market Crashes and Recovery
Between October 2007 and March 2009 the S&P 500 lost about 57% of its value. Someone putting $500 a month into a broad index fund watched their balance fall for seventeen months. Then the index climbed about 133% off its low and was back above the old peak by early 2013.
What decided whether that period was a disaster or an opportunity was not anyone's forecast. It was whether they were still buying in March 2009, when staying invested felt indefensible.
The recovery arithmetic nobody mentions
A 50% fall needs a 100% gain to get back to even. A 34% fall needs about 52%. A 57% fall needs 133%. That asymmetry is the most useful piece of crash knowledge there is, because it shows that the fall is the smaller half of the problem.
Small falls are almost symmetrical. Deep ones are not: past a 50% loss, the climb back is more than twice the fall.
Arithmetic. Gain needed = 1 ÷ (1 − fall) − 1.
For people who kept buying through the decline, the asymmetry works in their favor. Their earlier contributions fell 57%, but every contribution near the low bought more than twice as many shares as it would have at the peak. An investor 25 years from retirement who kept contributing did not have a loss in any meaningful sense. They had a discount, and a limited window to use it.
Someone already living off the portfolio had a genuine problem, and no chart makes that go away.
Five crashes, and how long they took to undo
| Crash | Peak-to-trough fall | Time to reach the low | Roughly how long to regain the peak |
|---|---|---|---|
| 1929–1932 | about 86% | three years | about 25 years in nominal terms |
| Dot-com, 2000–2002 | about 49% on the S&P 500, far worse on the Nasdaq | two and a half years | about seven years |
| Financial crisis, 2007–2009 | about 57% | seventeen months | about five and a half years |
| COVID, 2020 | about 34% | about one month | about six months |
| 2022 bear market | about 25% | ten months | about two years |
Depth says little about duration. COVID fell 34% and was over in a month. 2022 was the shallowest and still took two years to undo.
Commonly cited nominal peak-to-trough figures for the S&P 500, and for the Dow in 1929-1932.
The table breaks the most comforting story about crashes, that the deep ones always take longest to fix. The 2022 decline was a repricing of nearly everything at once rather than a panic that reversed, which is why a shallow fall lasted longer than a deep one. There is no template, which is a good reason to distrust anyone who tells you what the next crash will look like.
What continuing to buy actually does
The clearest way to see who a crash hurts is to run one month by month. Start with $7,500 invested. The market falls 4% a month for a year, 39% in all, then climbs back to exactly where it started over the following year. One investor adds $500 a month throughout. The other stops.
- Kept adding $500 a month
- Money paid in by that investor
- Stopped contributing
The market ends where it began. The investor who stopped is back to $7,500 after two stressful years. The one who kept buying paid in $19,000 and holds $22,486.
Illustrative. Prices fall 4% a month for 12 months, then rise about 4.2% a month for 12 months back to the start. $500 invested at the start of months 1 to 23.
Look at month 12: the contributor's balance is $9,436 against $13,500 paid in, a loss of about 30%. Contributions do not make a crash painless. What they do is buy three times as many shares as the stopper owns by the end, at an average price far below the starting one, so that a market that merely recovers produces a gain.
- What falling prices mean
- Better: More shares for each dollar
- More shares sold for each dollar
- What the recovery does
- Better: Lifts shares bought cheaply
- Lifts fewer shares than before
- What protects them
- Keeping the contributions going
- A cash reserve and a smaller stock share
VerdictNothing about the crash differs between these two people. Only the direction of the cash flow does, which is why one of them can ignore it and the other needs a plan in advance.
Why people sell at the bottom
Nobody sells at the bottom because they misread a chart. They sell because something real broke: the job, the business, the relationship, their health. Falling prices are only the trigger.
- 01Something real breaksA layoff, a medical bill, a business that stops paying
- 02There is no cash to cover itThe emergency fund was small or already spent
- 03Stocks are the only thing to sellAnd they are down 30%
- 04The loss becomes permanentShares sold at the bottom miss the recovery
Every step before the last one can be fixed in advance, while markets are calm. None of them can be fixed during the crash.
This is why "stay invested" fails so often in practice. It is correct as arithmetic and useless as an instruction, because it assumes you can afford to wait. If your portfolio is your down payment, your bridge through a layoff or next year's tuition, you cannot wait. Selling at the bottom is then not a discipline failure. It is a liquidity failure, and the fix is holding less stock in the first place.
The cruel part is the timing. Just before a crash is when stocks look safest and cash looks most wasteful, so that is when people take on the most risk. Anyone who tells you to be brave in a downturn should also tell you which money you are allowed to be brave with.
The evidence on behavior is consistent and unflattering. Studies of fund flows keep finding that investors earn less than the funds they own, because money arrives after a good run and leaves after a bad one. The gap is the cost of trading on feeling.
Decide your stock share in advance
Write down the share you could hold through a 50% fall without selling. That is your allocation, whatever a questionnaire suggests.
Fund the emergency reserve first
Three to six months of expenses in cash is what stops a bad year from becoming a forced sale.
Automate the contributions
The decision to keep buying should already be made. On payday there is nothing left to decide.
Rebalance on a calendar, not a feeling
Once a year, trim what ran up and add to what fell. It is the only version of buying the dip most people carry out reliably.
Check the balance quarterly at most
Prices are your noisiest input. Watching them daily adds nothing but the urge to act.
What to do during a crash
The protocol is short and boring, which is the point.
Close the broker app. Keep the automatic contribution running, because that is what you built it for. Rebalance once, if your allocation has drifted far enough to matter. Check that nobody in your household is about to lose their income. Then resist the two temptations that feel like intelligence: rotating into whatever is holding up, and waiting in cash until the bottom is confirmed. Nobody has ever reliably confirmed a bottom in real time.
What the record does not promise
Every US crash in that table was eventually recovered, and that is a strong argument for owning stocks across decades. It is not a law. Japan's Nikkei peaked in December 1989 and did not set a new high until 2024, so a Japanese investor who bought at the peak waited about 34 years. Global diversification is what stops one country's lost decades from becoming your whole portfolio.
Surviving a crash is not a personality trait. It is a structure: enough cash, an allocation chosen while calm, contributions you never have to think about, and a plan for what happens if the income stops. People with that structure mostly get through crashes without doing anything at all. That is the goal.
56 more deep dives are in the Navigator library.
