The S&P 500 over 100 years: what the track record shows

CostMe Research Desk · June 30, 2026

The long-run average annual return of the U.S. stock market is cited everywhere in personal finance. Ten percent. That number gets used in retirement calculators, compound interest explainers, and financial planning projections. What it contains, and what it does not contain, is less often explained.

The 10% figure comes from historical S&P 500 data going back to the early 20th century. It is an average, and like most averages, it conceals enormous variation. Understanding what the record actually shows is useful context before relying on that number for decisions.

What the 100-year record contains

The S&P 500 in its current form dates to 1957, but Standard and Poor's had predecessor indices going back to 1926, and broader U.S. equity market data extends further. Across that record, the approximate 10% nominal annual return (before inflation) holds up reasonably well as a long-run average.

But the record also contains: the Great Depression, during which the market fell about 89% from peak to trough between 1929 and 1932; two world wars; more than 25 recessions; a decade of near-zero real returns in the 1970s due to high inflation; a 57% decline from 2007 to 2009; and the 2020 pandemic decline of about 34% in roughly five weeks. The 10% is the average of a sequence that included all of those events.

What a long flat period looks like inside the average

The most important thing the average obscures is multi-year flat periods. From January 2000 to January 2013, the S&P 500's nominal return was approximately zero. Thirteen years. An investor who started in January 2000 and checked their balance in January 2013 saw roughly the same number they started with, before accounting for inflation. The long-run average was intact if you included the full century, but it was invisible for a long stretch.

Similarly, from 1966 to 1982, U.S. stocks returned close to zero in real (inflation-adjusted) terms over 16 years. An investor who retired at the start of that period and relied on historical averages for their planning faced a different reality.

This matters for understanding the 10% figure. The average holds over very long periods, but the individual decades inside that average vary enormously. A 30-year investment period that ends in 1999 looks different from one that ends in 2009.

Why the index approach has held up

The academic literature on active management is fairly consistent: most actively managed funds underperform their index benchmark over long periods, net of fees. The reasons are structural. Active managers must beat a benchmark by enough to cover higher costs, and because markets incorporate information quickly, the edge needed to do that consistently is very hard to maintain.

Low-cost index funds that track the S&P 500 capture the historical return (minus a small fee) without requiring correct predictions about individual stocks. This is the core argument behind passive investing, and it is supported by the data. You can read more about it in the article on why active investors tend to underperform.

What the record does not guarantee

Historical returns are context, not forecasts. The 10% figure reflects the specific political, economic, and demographic conditions of 20th and early 21st century American capitalism. It was produced by a particular set of circumstances that may or may not persist.

Many researchers who study long-run equity returns believe the forward return expectation for the next several decades may be lower than the historical average, owing to lower expected productivity growth, higher starting valuations, and demographic shifts. Others argue the long-run structural forces that drove the historical return remain in place. The honest answer is that no one knows the future return with precision.

What the historical record does show is that long holding periods reduced, though did not eliminate, the probability of nominal loss. In most 20-year periods for which there is data, an investor who held U.S. equities for the full period ended with more than they started with. The longer the horizon, the more that statement has historically held. That is meaningful context. It is not a guarantee.

How CostMe uses this data

The 30-year projection in CostMe uses long-run average market returns as the rate assumption. This article provides the historical context behind that rate: where it comes from, what it contained, and what its limitations are. The projection is not a prediction. It is a tool for making the opportunity cost of a purchase concrete and comparable. See what the full app includes at the pricing page, and read more on how compounding works in the compound interest explained article.

The science behind it

Dimson, E., Marsh, P., and Staunton, M., 2002, "Triumph of the Optimists: 101 Years of Global Investment Returns," Princeton University Press. The most comprehensive long-run study of equity returns across multiple countries. Found that the U.S. equity premium was real but somewhat exceptional compared to other developed markets over the same period.

French, K.R. and Fama, E.F., 2010, "Luck versus Skill in the Cross-Section of Mutual Fund Returns," Journal of Finance. Showed that after fees, most actively managed funds underperformed low-cost index alternatives, and that the small number that outperformed were not reliably distinguishable from luck.

Goetzmann, W.N. and Ibbotson, R.G., 1994, "Do Winners Repeat?", Journal of Portfolio Management. Early analysis of return persistence showing that past performance in actively managed funds did not reliably predict future outperformance, supporting the case for passive index exposure.

This is general education, not financial advice.