How long do you have to hold stocks to not lose money? We ran every window since 1928.

The popular version of this answer is too optimistic, and the reason is that almost nobody adjusts for inflation.

The nominal numbers

Every overlapping window in the 98-year series, total return with dividends reinvested:

Holding period	Windows	Lost money	Worst total return	Worst annualized
1 year	98	26 (26.5%)	−43.8%	−43.8%
3 years	96	15 (15.6%)	−61.6%	−27.3%
5 years	94	11 (11.7%)	−49.3%	−12.7%
10 years	89	5 (5.6%)	−15.5%	−1.7%
15 years	84	1 (1.2%)	−3.0%	−0.2%
20 years	79	0 (0.0%)	+59.8%	+2.4%
30 years	69	0 (0.0%)	+898.2%	+8.0%

This is the table that gets screenshotted, and on its own it supports a clean story: hold for 20 years and you have never lost money.

Two things about that story are worth noticing before you rely on it. The worst 20-year period still only compounded at 2.4% a year — technically positive, and a long way from the 10% average. And the worst 3-year stretch lost 61.6%, which is the kind of number that ends plans regardless of what the 20-year row says.

The real numbers, which are the ones that matter

Prices rose 18.5-fold over this period — 3.02% compounded. A nominal gain smaller than inflation is a loss, so here is the same analysis in purchasing power:

Holding period	Lost purchasing power	Worst real total return
1 year	31 of 98 (31.6%)	−38.9%
5 years	22 of 94 (23.4%)	−39.1%
10 years	10 of 89 (11.2%)	−34.0%
15 years	3 of 84 (3.6%)	−5.5%
20 years	0 of 79 (0.0%)	+14.5%

The 10-year row is the headline. Measured in what the money can buy, a decade in the US stock market failed to break even about one time in nine, and the worst decade destroyed a third of your purchasing power. That's not a rounding adjustment to the nominal table; it doubles the failure rate and turns a −15.5% worst case into −34.0%.

Twenty years remains unbroken, even in real terms, with the worst case gaining 14.5% in purchasing power across two decades — roughly 0.7% a year. Positive, and nothing like what most people mean by "stocks return 7%."

The corollary is the planning rule: money you need inside 10 years has a real chance of being worth less when you need it, and money you need inside 5 years has a large one. That's a statement about the time dimension of risk, and it's more useful than any single average return. (The averages themselves are in what the average stock market return actually is.)

The serious objection: stocks may not get safer with time at all

Here the data above needs a named counterargument, because there is a rigorous academic case that the whole framing is wrong.

In On the Risk of Stocks in the Long Run (Financial Analysts Journal, May/June 1995), Zvi Bodie argued that the probability of a shortfall is the wrong measure, because it ignores how bad the shortfall is. The right measure, he proposed, is what it would cost to insure against one — and that cost behaves in the opposite direction to the conventional wisdom. Restating the argument in a later paper for the Pension Research Council at Wharton, Bodie puts it this way:

If stocks were truly less risky in the long run, then the cost of insuring against earning less than the risk-free rate of interest should decline as the length of the investment horizon increases. But reality is quite the opposite.

The intuition: as the horizon lengthens, a shortfall becomes less likely but much larger when it happens. The option premium reflects both, and the second effect dominates. Bodie showed this holds even if returns are mean-reverting.

This doesn't invalidate the tables above — the frequencies are what they are. It does mean "no 20-year period has lost money" is a weaker claim than it sounds, for three reasons:

1. The windows overlap. Our 79 twenty-year periods are drawn from 98 years, so they share most of their data and are nowhere near 79 independent observations. The real sample is closer to four or five non-overlapping periods. "Never happened" across that few trials is thin evidence.
2. A positive return is not a sufficient outcome. A 20-year period compounding at 2.4% nominal, or 0.7% real, cleared the "didn't lose money" bar and would still wreck a retirement plan built on 7%.
3. The sample is the luckiest one available. This is US data across the century in which the US became the dominant economy. Other markets have 20-year stretches that were far worse.

The defensible conclusion is narrower than the usual one: a long horizon reliably reduces the frequency of loss, and does not cap its severity.

What to do with this

1. Match the horizon to the money, not to your confidence. Under 3 years: it shouldn't be in equities. 3–10 years: partially, knowing an 11% real failure rate over a decade is live. 10+ years: equities are appropriate.
2. Plan in real terms. Nominal targets quietly assume inflation away, and inflation did a third of the damage in the worst decade here.
3. Don't let "20 years has never lost" become a reason to stop thinking. It rests on about four independent observations from the most favorable market in history.
4. Decide the near-term money's allocation while markets are calm. Fixing it during a decline is the expensive version — that's the whole argument in what to do when the market drops.

The horizon only helps if you're still there

Every number above assumes you held. That assumption is doing more work than the statistics are.

A 10-year holding period with an 11% real failure rate is survivable. The thing that actually converts a temporary decline into a permanent loss is selling partway through — and the periods containing the worst windows above are precisely the ones that made people sell. The historical record of the asset is not the same as the historical record of the investors in it.

That's the gap Sydnical measures. You run real positions at live prices with nothing at risk, and the grading is on patience, sizing, concentration, and discipline rather than a balance dominated by luck. The crash scenarios put you inside the worst windows in these tables — 2000–02, 2008, 2022 — day by day, with no knowledge of where the bottom is. Finding out that you sell is the single most valuable thing a simulator can tell you, and it costs an afternoon instead of a decade.

FAQ

How long do you need to hold stocks to not lose money?

Historically, about 20 years for near-certainty. Across 1928–2025, none of the 79 overlapping 20-year periods in the S&P 500 lost money nominally or after inflation, while 11.2% of 10-year periods lost purchasing power and 23.4% of 5-year periods did. Ten years is commonly quoted but has a real failure rate of roughly one in nine.

Has the stock market ever lost money over 10 years?

Yes. Five of 89 overlapping 10-year periods since 1928 lost money in nominal terms, and 10 of 89 lost purchasing power after inflation — the worst of those down 34.0%. The 2000s are the best-known example, but they aren't unique.

Is it safe to invest money I need in 5 years?

The data argues against it. Nearly a quarter of 5-year periods since 1928 lost purchasing power, and the worst lost 39.1%. Money needed within about three years generally shouldn't be in equities at all, and a 5-year horizon carries a real and substantial chance of having less than you started with.

Do stocks get less risky the longer you hold them?

The frequency of loss falls sharply with horizon, but the severity does not. Zvi Bodie's 1995 Financial Analysts Journal paper showed that the cost of insuring against a shortfall actually rises as the horizon lengthens, because shortfalls become rarer but much larger. Both facts are true, and the second one is usually left out.

How reliable is the "no 20-year period has lost money" claim?

Weaker than it sounds. The 79 twenty-year windows in a 98-year series overlap heavily, so they amount to roughly four or five independent observations from the single most favorable market in modern history. It's suggestive, not a law — and the worst of those periods still only compounded at 2.4% nominal.

Sources

• Aswath Damodaran — Historical Returns on Stocks, Bonds and Bills, NYU Stern, updated January 2026
• Federal Reserve Bank of Minneapolis — Consumer Price Index, 1913–
• Zvi Bodie, "On the Risk of Stocks in the Long Run," Financial Analysts Journal 51, no. 3 (May/June 1995)
• Zvi Bodie — "An Analysis of Investment Advice to Retirement Plan Participants", Pension Research Council, Wharton (source of the quoted passage)

All holding-period frequencies are our own computation over overlapping windows in the Damodaran series. Overlapping windows are not independent observations — a caveat we treat explicitly above rather than in a footnote.

---

Twenty years of data says the asset recovers. It says nothing about whether you'll still be holding it. That part is testable →