Market crash recovery time: two clocks and the maths of losses

A market can spend years falling before its recovery even begins. In the supplied S&P 500 total-return series measured in euros, the August 2000 peak was regained in November 2013, after 159 months. Of those, 102 elapsed before the February 2009 trough; the subsequent rebound took 57 months. Some waits are still open: EUR-hedged global government bonds had not regained their July 2020 peak after 69 months, when the dataset ends in April 2026.

Break-even is a reference point, not a growth objective. Geometric wealth growth depends on the full net return path, including the cost of any protection and returns outside the drawdown.

That split matters when reading claims that a crash “takes five years to recover”. Five years measured from a known bottom describes a different experience from five years measured from the original investment. This companion to the asset-class crash study calculates both clocks and the capital needed to get back.

Key takeaways

1. Peak-to-recovery time includes both the fall to the trough and the rebound. For the S&P 500 in euros, 102 of the 159 months passed before the trough.
2. A loss must be recovered on a smaller base. A 50% loss requires a 100% gain, and the 60.2% loss of the S&P 500 in euros required 151.3%.
3. Unfinished recoveries must be recorded as unfinished, not averaged in. The global government-bond loss of 20.2% from July 2020 was still unrecovered at its April 2026 endpoint.
4. Currency, dividends, costs and withdrawals change what recovery means for a real investor. A historical recovery date is an observation, not a forecast.

Decompose the whole wait

We use the worst-drawdown dataset, rather than resetting peaks at crisis-window boundaries. All figures below are nominal euros and month-end observations. The equity series include reinvested dividends: S&P 500 total return and MSCI Japan IMI net return. Gold is a spot-price series, while high yield is an index total-return series.

The derived two-clock CSV records exact dates and checks that the two components sum to the reported total for every completed recovery.

Japan illustrates why the distinction changes interpretation: its May 2003 trough arrived relatively early, but the old peak was not regained until February 2018. Gold's July 1999 trough, in contrast, arrived after most of its eventual 296-month wait had already elapsed.

These are each series' worst drawdown episodes over different available histories. The table does not rank equally long samples or show the average wait after any crash. The US drawdown spans both the dot-com and financial-crisis episodes without regaining its original euro peak between them.

The rebound is earned on a smaller base

If the initial value is 100 and the loss fraction is L, the trough value is 100 × (1 − L). The gain required from there is:

required gain = 1 / (1 − L) − 1

A 60.2% loss in the US euro series leaves €39.80 of every €100. Getting back requires a 151.3% gain, a 2.513-fold increase from the trough. Japan's 66.2% loss requires 195.9%; gold's 54.0% loss requires 117.4%. Those are calculations from the rounded losses, not independently observed rebound returns with exact crossing values.

The formula does not establish that a low-loss investment is always superior. Returns earned outside the drawdown, income, costs and the risk of an alternative holding still matter. It simply prevents comparing percentages with different denominators.

A recovery hurdle is not a recovery forecast

To translate a hurdle into a planning sensitivity, assume hypothetical constant 5% annual net growth, no further shocks and no withdrawals. Then:

years from trough = log(1 / (1 − L)) / log(1.05)

The 60.2% US loss would need 18.9 years at that assumed rate. Its observed trough-to-recovery interval was instead 57 months, or four years and nine months. The 35.5% whole-history high-yield loss would need 9.0 years at 5%, whereas its observed rebound took 13 months.

The mismatch is useful: the historical rebound was not a smooth 5% path. Using a long-run assumed return to forecast an imminent recovery would confuse a sensitivity exercise with evidence. Neither historical speed nor this hypothetical calculation predicts the next episode.

Keep unfinished recoveries unfinished

The euro government-bond series had not regained its December 2020 peak when its data ended in January 2023, 25 months later. The EUR-hedged global government-bond series had not regained its July 2020 peak by April 2026, 69 months later, after a 20.2% fall to October 2023 that requires a 25.3% gain to recover.

These observations are right-censored: the final duration is unknown in the supplied data. Outside the dataset, the EUR-hedged Xtrackers ETF on the same index closed at €206.15 on Xetra on 30 October 2026, the last trading day of the month, still 16.4% below its July 2020 month-end close of €246.69. These are Yahoo Finance daily closes, the source of the corrected series; justETF's price history shows slightly different levels, with a similar gap.

They must not enter an average as zero-month recoveries, as completed 25- or 69-month recoveries, or as unrecovered through October 2026 on the strength of the supplied data alone. Extending the publication date does not extend the dataset. The bond comparison explains why its fixed-window and whole-history losses also differ.

Even completed records are nominal. The gold analysis shows how ownership costs and purchasing-power requirements can leave a shortfall after a spot-price recovery. An investor who withdraws money also owns fewer units; the withdrawal model quantifies that separate problem.

What you can do now

These calculations use your own statements; none of them predicts when a market will recover.

1. Find your portfolio's highest value and the date it was reached. Your statements or account history give the peak that any later fall is measured from.
2. Calculate the gain needed to get back. Divide the peak by today's value and subtract one: a portfolio worth €80,000 after a €100,000 peak needs a 25% gain.
3. Count the two clocks separately. Note how many months passed from the peak to the lowest value, and how many from that low to today.
4. Add back withdrawals before comparing with an index. Money you took out will not be in the balance when the index regains its peak, so list withdrawals since the peak.
5. Test a fall on your current allocation. The Copio app preview can run crash scenarios on your allocation by asset class across accounts.