Summary
- The S&P 500's long-run average annual return is a useful planning shorthand, but it rarely describes any specific year
- The distribution of annual returns is wide and uneven, with very large gains and sharp losses more common than near-average years
- The average is built from a century of extreme variation, and staying invested through that variation is what generates the cumulative return
- Sequence of returns risk means the order of gains and losses matters as much as the average in retirement
The number every investor hears
One of the most frequently cited facts in personal finance is that the U.S. stock market has historically returned around 10 to 12 percent per year over the long run. The figure is real. Over the century from 1926 to 2025, the S&P 500 and its predecessor index produced an arithmetic average annual total return, including dividends reinvested, of approximately that magnitude.
Figures reflect historical data through 2025. Source: Slickcharts S&P 500 historical returns. Past performance does not indicate future results.
What that number does not convey is how rarely any individual year comes close to it.
What the distribution actually looks like
When all 100 years of annual returns from 1926 through 2025 are sorted into a frequency distribution, the shape is striking. The largest single bar in the chart is the one that surprises most people: returns of more than +20 percent, which accounts for 38 of the 100 years. Negative years, while less common than up years overall, account for 26 of the 100 years: roughly one in four.
The bins nearest the long-run average tell a different story. The +10 to +15 percent range, which is the bucket the arithmetic average of 12.35 percent actually falls into, contains only 9 of the 100 years. The two adjacent bins, 0 to +5 percent and +5 to +10 percent, contain just 6 and 8 years respectively. The over +20 percent bin holds 38 years. The stated average sits in one of the histogram's thinner columns.
The year-by-year chart below plots each calendar year as an individual bar, ordered chronologically. This view makes it easier to trace how outcomes clustered during specific periods: the prolonged strength of the 1980s and 1990s, the back-to-back losses of 2000 through 2002, or the sharp single-year drop in 2008 and the immediate recovery that followed. The distribution view shows how often each return magnitude occurred; the year-by-year view shows when those outcomes arrived.
Use the year filter above to explore any time period; both charts update together.
Why the average may mislead
The arithmetic average of a set of numbers is sensitive to extreme values. A sequence like -37% (2008), +26% (2009), -22% (2002), +29% (2003) produces an average that looks moderate and unremarkable, even though none of those four years was either.
This is not a flaw in how the average is calculated. It reflects a genuine property of the market's return history: extreme variation is what generates the long-run average, not a smooth approximation to it. An investor who stayed fully invested through every year in this dataset would have compounded their wealth at a historically strong rate. But they would have done it by living through the full distribution, including every year that looked nothing like the "average."
The distinction matters because financial projections that assume a steady 10 or 11 percent return each year may look very different from projections that accurately represent the actual sequence of gains and losses.
Arithmetic average vs. compound return. The figure shown in the chart is the arithmetic mean: the sum of all annual returns divided by the number of years. The compound annual growth rate (CAGR), which measures what an investment actually grew at over the full period, is lower. The gap exists because losses hurt more than equal gains help: a 50% drop followed by a 50% gain does not return to the starting point; it leaves the investor 25% below it. This effect is called volatility drag, and it means the arithmetic average overstates what an investor actually earned on a dollar-for-dollar basis. The CAGR is the number that reflects real-world compounding; the arithmetic average is useful for understanding the distribution of outcomes year by year.
The accumulation years: dispersion works in the investor's favor
During the accumulation phase, an investor adding regularly to a portfolio actually benefits from dispersion in a specific way. When prices fall sharply, new contributions purchase more shares at lower prices. When prices recover, those shares may participate in the rebound. This is sometimes called dollar-cost averaging, though it is more accurate to say that dispersion creates the opportunity for it.
The year-by-year chart above shows this pattern directly. Some sharp declines were followed by an immediate positive year; others preceded multiple consecutive down years before a rebound arrived. The timing and depth varied considerably across different eras. Exploring specific decades in the chart illustrates how different these sequences looked in practice.
This does not mean losses are acceptable simply because a rebound may eventually follow. It means the historical record shows that investors who stayed invested through the full sequence, including the painful years, were positioned to participate in the eventual gains.
Past patterns do not guarantee future recoveries occur within any specific timeframe.
The retirement years: when dispersion becomes a risk
The situation changes materially once an investor begins drawing down a portfolio rather than adding to it. Two retirees with the same average return over a 20-year period can end up with dramatically different outcomes depending on when the large negative years occur.
A retiree who experiences a severe bear market in the early years of retirement, while drawing living expenses from a portfolio that has just declined significantly, is forced to sell more shares to meet the same dollar withdrawal. Those shares are permanently gone and cannot participate in the eventual recovery. The portfolio starts in a smaller position and compounds from a lower base.
This is the sequence-of-returns risk. It is one of the primary reasons retirement income planning cannot rely on projecting an average return forward and assuming everything averages out. The order of returns matters as much as their average.
Several factors may help manage this risk, including:
- Maintaining a portion of the portfolio in lower-volatility assets in the early retirement years
- Building a near-term spending reserve so that portfolio withdrawals can be reduced or deferred during periods of market weakness
- Considering a flexible withdrawal approach that adjusts spending in response to portfolio performance
- Coordinating Social Security and other income sources to reduce the size of portfolio withdrawals in early retirement
The appropriate combination depends on the specific situation: portfolio size, income sources, spending needs, health, and other factors that vary significantly from one household to another.
What the histogram does not capture
The frequency distribution of annual returns is a useful picture of the past. But it does not address several questions that are directly relevant to planning:
- Sequence: The histogram shows how often each return occurred, not in what order. Two eras with the same distribution can produce very different compound outcomes depending on when the losses fall relative to when withdrawals begin.
- Inflation: Nominal returns are shown. Real (inflation-adjusted) returns have historically been lower. The purchasing power of returns depends on what inflation was doing during the same period.
- Personal portfolio: An individual portfolio rarely mirrors the index perfectly. Asset allocation, international exposure, bond holdings, tax drag, and fees all affect the actual return experienced.
- Future: A century of historical data is the best record available. It does not mean the next century will look the same.
The takeaway for long-term investors
The long-run average return of the U.S. stock market is not a prediction about any individual year. It is a description of what a century of extreme annual variation, including 26 negative years, has added up to when compounded over time.
Understanding this does not change what the long-run average is. But it does change what a realistic financial plan looks like, particularly one designed to generate reliable income through a retirement that may span 25 to 30 years.
A plan that assumes steady returns smooths over the very dispersion that creates those returns in the first place. A plan that accounts for the full range of outcomes, and that addresses how the portfolio behaves when those outcomes arrive, is more likely to produce reliable results regardless of what any individual year delivers.