S&P 500 100-Year Returns: Annual Returns by Year, CAGR & Worst Years (1928–2026)

From 17 points in 1928 to over 5,000 today — what almost a century of S&P 500 data actually says about long-term equity returns.

10.1%
Annualized total return since 1928
72 / 26
Positive vs negative years
−86%
Worst drawdown (1929–1932)

The headline numbers

Across the 98 finished calendar years from 1928 onward, the S&P 500 has delivered:

The gap between the arithmetic mean (11.9%) and the geometric mean / CAGR (10.1%) is the volatility tax: 1.8 percentage points eaten by the math of compounding through losses. A dollar left alone since the end of 1927, dividends reinvested, is $11,976 today.

How often does the S&P 500 go down?

Out of 98 finished calendar years, 72 were positive and 26 were negative. That's a 73% base rate of an up year. But "up year" hides enormous variance:

Worst years for the S&P 500

YearTotal ReturnContext
1931−43.8%Depth of the Great Depression
2008−37.0%Global Financial Crisis
1937−35.3%Premature Fed tightening
1974−25.9%Oil shock + stagflation
1930−25.1%Smoot-Hawley + bank runs
2002−22.1%Dot-com bottom
2022−18.1%Aggressive Fed hiking cycle
1973−14.3%Bretton Woods collapse

Best years for the S&P 500

YearTotal ReturnContext
1954+52.6%Post-war boom
1933+50.0%Snap-back after Depression bottom
1935+46.7%New Deal recovery
1928+43.8%Late Coolidge boom, one year from the top
1958+43.7%Eisenhower-era bull

Note the pattern: the best years are almost always the year after a bottom. The crowd that fled in the worst year typically missed the snap-back. 1928 is the exception that sets up the rest of the table: it was the last full year before the crash that produced 1931.

Rolling-period probability of a positive return

The single most useful framing for long-term investors:

Windows are calendar-year windows over the finished years charted above, so every figure here can be checked against the raw series.

See the interactive buy-year/sell-year return matrix for the full picture.

What about the 2000s "lost decade"?

The S&P 500's total return from year-end 1999 to year-end 2009 was −9% cumulative — a flat-to-negative decade. This is the most cited counter-example to the "buy and hold always works" narrative.

Two important framings:

  1. The decade started at a CAPE of 44× — the most extreme valuation in U.S. history. A flat decade was almost mathematically required.
  2. An investor making monthly contributions through the period — i.e. dollar-cost averaging — earned a positive return because they bought heavily at the 2002 and 2009 lows.

Should you expect 10.1% going forward?

Probably not. The historical 10.1% includes:

None of those tailwinds is guaranteed to repeat. Most return-forecast models built off CAPE / AIAE / Buffett indicator currently predict 3–5% real returns over the next decade — well below the historical average. See the valuation analysis for details.

Total-return figures come from /api/sp500/annual-tr.json, price-only figures from /api/sp500/annual-returns.json, and the drawdown from /api/sp500/drawdowns.json — all refreshed daily, and the numbers above are regenerated from them rather than typed. Free for research and citation.

Further reading