Statistics · 2026 · Stocks
Stock Market Seasonality Statistics (2026)
January has been the best month for U.S. stocks, averaging 1.84% since 1872, and October the weakest at 0.02%. Even so, the best month rose in 72.3% of years. Figures use Robert Shiller’s S&P Composite data through June 2026. They are historical tendencies, not forecasts.
Key takeaways
- January has the highest average monthly return since 1872: 1.84%, with gains in 72.3% of 155 years.
- October has the lowest: 0.02%, up in 57.4% of 155 years.
- From 1872 to 2025, November through April beat May through October in 91 of 154 years (59.1%), but the median summer half still gained 5.2%.
- By year of the presidential term, the pre-election year averaged 14.6% and the midterm year 7.4%, with only 38 to 39 years in each group.
- January’s sign matched the rest of the year in 109 of 154 years (70.8%), less often than a rule that always guesses “up” (72.7%).
What is the best month for stocks?
January is the best calendar month for U.S. stocks, averaging 1.84% with a median of 1.74% across 155 years since 1872. The month rose in 72.3% of those years. That is a historical tendency measured on monthly average prices, not a forecast.
Shiller’s price series holds monthly averages of daily prices, not month-end closes. Every monthly return here is the change from one month’s average price to the next, plus one-twelfth of the annual dividend. Averaging smooths prices, so results differ from close-to-close figures. Data: January 1871 to June 2026. Each return needs the prior month, so Februarys through Decembers start in 1871 and Januaries in 1872.
1. January averages 1.84%, the highest of any month
Across 155 Januaries the mean is 1.84%, the median 1.74%, and 72.3% were positive. The next-highest mean is July at 1.27% (WealthyBud data · 1,866 months · June 2026).
2. Since 1950 the leading months are January and November
From 1950 onward (76 to 77 years per month), January and November lead (tied, within 0.1 point): January 1.67%, November 1.61%. By share of years that were positive, the top month is December at 77.6% (WealthyBud data · 1,866 months · June 2026).
3. Every calendar month has a positive average
The 12 monthly means run from 0.02% to 1.84%, a spread of 1.82 points; the standard deviation of all 1,865 monthly returns is 4.05% (WealthyBud data · 1,866 months · June 2026).
| Month | n (all / 1950+) | Mean, all | Median, all | % up, all | Mean, 1950+ | Median, 1950+ | % up, 1950+ |
|---|---|---|---|---|---|---|---|
| January | 155 / 77 | 1.84% | 1.74% | 72.3% | 1.67% | 1.70% | 71.4% |
| February | 156 / 77 | 0.74% | 0.76% | 55.1% | 0.99% | 1.09% | 55.8% |
| March | 156 / 77 | 0.46% | 0.77% | 62.2% | 0.64% | 0.96% | 62.3% |
| April | 156 / 77 | 1.03% | 1.18% | 66.7% | 1.52% | 1.49% | 75.3% |
| May | 156 / 77 | 0.56% | 0.72% | 59.0% | 1.05% | 1.00% | 64.9% |
| June | 156 / 77 | 0.50% | 0.83% | 59.0% | 0.72% | 0.90% | 63.6% |
| July | 155 / 76 | 1.27% | 1.21% | 69.7% | 0.96% | 1.24% | 72.4% |
| August | 155 / 76 | 1.23% | 1.10% | 63.9% | 0.68% | 0.87% | 60.5% |
| September | 155 / 76 | 0.58% | 1.16% | 61.9% | 0.23% | 1.01% | 57.9% |
| October | 155 / 76 | 0.02% | 0.42% | 57.4% | 0.22% | 0.48% | 61.8% |
| November | 155 / 76 | 1.02% | 1.59% | 65.2% | 1.61% | 2.02% | 72.4% |
| December | 155 / 76 | 0.74% | 1.17% | 64.5% | 1.52% | 1.78% | 77.6% |
What is the worst month for stocks?
October is the weakest calendar month, averaging 0.02% across 155 years since 1871, with a median of 0.42%. It was positive in 57.4% of years. The average sits near zero, but the month still rose more often than it fell.
4. October averages 0.02%, the lowest of any month
The standard error of that mean is 0.36 points, so the average is 0.1 standard errors from zero, within normal noise (under 2). Its standard deviation is 4.51% across 155 years (WealthyBud data · 1,866 months · June 2026).
5. Since 1950 the weakest months are October and September (0.22%, 0.23%)
October was positive in 61.8% of years since 1950 (tied, within 0.1 point) (WealthyBud data · 1,866 months · June 2026).
For the long-run picture, see our stock market returns and bear markets guides.
Does “sell in May and go away” work?
November to April beat May to October in 91 of 154 years (59.1%), 1872 to 2025. The winter half averaged 6.1% and the summer half 4.5%. The mean gap is 1.3 standard errors, within normal noise, so the win rate is not a proven edge. Summer still gained on average.
The windows tile the year and use total return (dividends included). Winter runs from the October average price of the prior year to the April average price. Summer runs from the April average to the October average, and each year pairs the winter ending in April with the summer after it. Cash returns are not modeled.
6. Winter averaged 6.1%; summer averaged 4.5%
The mean gap is 1.7 points, with a standard error of 1.3 points. The gap is about 1.3 standard errors, within normal noise (under 2 standard errors), so ordinary year-to-year swings could produce it. Medians: 4.7% and 5.2% (WealthyBud data · 154 years · June 2026).
7. Winter beat summer in 91 of 154 years (59.1%)
Winter was up in 72.1% of years, summer in 65.6% (WealthyBud data · 154 years · June 2026).
8. Since 1950 winter beat summer in 68.4% of years; before 1950, 50.0%
Winter won 39 of 78 years to 1949 and 52 of 76 from 1950 (WealthyBud data · 154 years · June 2026).
9. $1 held only in winters grew to $3,921; only in summers, $243
Both compound the 154 windows from October 1871; holding all the time grew $1 to $953,237. Cash interest is ignored (WealthyBud data · 154 years · June 2026).
10. A later working paper reports that Bouman and Jacobsen found a winter edge in 36 of 37 countries
Bouman and Jacobsen (American Economic Review 92(5), 2002). We read only this summary in a later working paper by Jacobsen and Zhang, which says the 2002 study found winter returns “significantly higher than during summer (May-October) in 36 out of the 37 countries in their study.”
| Period | Years | Winter mean | Summer mean | Winter median | Summer median | Winter beat summer |
|---|---|---|---|---|---|---|
| 1872–2025 | 154 | 6.1% | 4.5% | 4.7% | 5.2% | 91 (59.1%) |
| 1872–1949 | 78 | 3.9% | 4.8% | 2.5% | 6.4% | 39 (50.0%) |
| 1950–2025 | 76 | 8.4% | 4.1% | 9.1% | 4.9% | 52 (68.4%) |
Compare low-cost index ETFs
How do stocks perform in election years?
Election years averaged 11.8% total return across 39 years from 1872 to 2025, and 79.5% were positive. The pre-election year had the highest average, 14.6%. With under 40 years per group and swings of 17 points or more, the gaps between groups are small.
Election years are calendar years divisible by 4 (1872, 1876 and so on); inauguration dates are ignored. Returns run December average to December average, 1872 to 2025, dividends reinvested.
11. The pre-election year averaged 14.6%; the midterm year, 7.4%
By median the order is pre-election year highest (19.2%) and midterm year lowest (7.5%). All years average 10.9%, median 13.2% (WealthyBud data · 154 years · June 2026).
12. Election years: 11.8% mean, 79.5% positive, n = 39
Since 1950 (19 election years) the mean is 12.5% and 84.2% were positive (WealthyBud data · 154 years · June 2026).
13. The best-to-worst term-year gap is 7.2 points; standard errors reach 3.1 points
Within-group standard deviations are 17 to 20 points. The gap is 1.7 standard errors of the difference, within normal noise (under 2) (WealthyBud data · 154 years · June 2026).
- The post-election year (year % 4 = 1) averaged 9.9% with a median of 12.5% across 39 years, and 64.1% were positive.
- The midterm year (year % 4 = 2) averaged 7.4% with a median of 7.5% across 38 years, and 65.8% were positive.
- The pre-election year (year % 4 = 3) averaged 14.6% with a median of 19.2% across 38 years, and 84.2% were positive.
- The election year (year % 4 = 0) averaged 11.8% with a median of 15.2% across 39 years, and 79.5% were positive.
Is January a good predictor of the year?
January’s direction matched the direction of the following eleven months in 109 of 154 years (70.8%), 1872 to 2025. That is above the 60.0% expected by chance, but below the 72.7% you get by always guessing that the rest of the year rises.
The rule: a year is a “hit” when January’s total return (December average to January average) has the same sign as the rest of the year (January average to December average). Classic versions use closing prices. The chance rate assumes the two are independent.
14. January’s sign matched the rest of the year in 109 of 154 years (70.8%)
Since 1950 the hit count is 54 of 76 (71.1%), against 76.3% for always guessing “up” (WealthyBud data · 154 years · June 2026).
15. After an up January, the rest of the year rose 80.2% of the time (n = 111)
After a down January it rose 53.5% of the time (n = 43). Average rest-of-year returns were 11.4% and 2.1% (WealthyBud data · 154 years · June 2026).
How reliable are seasonal patterns?
Seasonal patterns are weak guides. Monthly returns have a standard deviation of 4.05%, while the gap between the best and worst month’s average is 1.82 points. January beat October in 99 of 154 years (64.3%), so even the strongest contrast fails often.
16. Each month’s average carries a standard error of 0.23 to 0.43 points
With 155 to 156 observations per month, the best-to-worst gap of 1.82 points is 4.2 times the standard error of the difference, beyond the 2-standard-error threshold. But the best and worst months were picked from 12, which flatters the gap (WealthyBud data · 1,866 months · June 2026).
17. Month rankings before and after 1950 correlate at only 0.18
The Spearman rank correlation of the 12 monthly means before 1950 and from 1950 on is 0.18 (1 is identical order). Before 1950 the top month was January; from 1950 on, January and November (WealthyBud data · 1,866 months · June 2026).
What this means for investors
Treat calendar effects as context, not a signal. The best-to-worst monthly gap is 1.82 points against a 4.0% typical swing. The summer half still gained 4.5% on average and was positive in 65.6% of years. Selling in May gave up that return unless cash interest made it up.
Don’t build a plan on 39 election years. Groups are small and results vary widely. A broad fund such as SPY lets you stay in the market through every season.
Time in the market beats timing. Lori Schock, a former director of the SEC’s Office of Investor Education and Assistance, wrote on Investor.gov (a page marked no longer updated): “It’s time in the market that counts, not timing the market.” Browse the stocks hub for company-level data.
More Stocks statistics
Bear Market Statistics (2026)
Every U.S. bear market on the monthly record since 1871 — how often they happen, how deep they go, how long they last, and the bull markets between.
View ResearchStock Market Returns Statistics (2026)
Average U.S. stock market returns since 1871 — by decade, over rolling 10-, 20- and 30-year periods, before and after inflation.
View ResearchStock Market Statistics (2026)
Combined market cap, revenue and margins across 103 large-cap U.S. stocks, computed from SEC EDGAR filings.
View HubStock Fundamentals & Investor Scores
103 large-cap companies ranked by SEC-filed fundamentals.
View