December has been the S&P 500’s most dependable month. From 1928 to 2025 the index rose in 71 of 98 Decembers (72%), with an average gain of 1.27% and a median of 1.48%, against 0.65% and 60% for all months. Nearly all of that gain has come in the second half of the month, and the famous Santa Claus window has been much weaker in the last decade than its long record suggests.
All numbers were calculated from daily S&P 500 closes (price index, no dividends). A monthly return runs from the prior month’s last close to this month’s last close. More calendar studies are on the Market Stats hub.
December vs the other eleven months
December ranks first of the twelve months by share of up years and second by average return, behind July.
| Month | n | Average | Median | % positive | Rank by average |
|---|---|---|---|---|---|
| January | 98 | 1.21% | 1.55% | 62% | 4 |
| February | 98 | -0.10% | 0.23% | 52% | 11 |
| March | 98 | 0.48% | 1.05% | 61% | 9 |
| April | 98 | 1.26% | 0.93% | 65% | 3 |
| May | 98 | 0.08% | 1.06% | 60% | 10 |
| June | 98 | 0.80% | 0.39% | 57% | 6 |
| July | 98 | 1.69% | 1.69% | 61% | 1 |
| August | 98 | 0.68% | 0.87% | 58% | 7 |
| September | 98 | -1.12% | -0.42% | 45% | 12 |
| October | 98 | 0.55% | 1.03% | 59% | 8 |
| November | 98 | 1.01% | 1.63% | 62% | 5 |
| December | 98 | 1.27% | 1.48% | 72% | 2 |
December also swings less. Its returns have a standard deviation of 3.71 points against 5.33 for all months combined, so the typical December has been a modest gain rather than a big one. Starting in 1950 instead of 1928 gives almost the same picture: December averaged 1.41% and rose 72% of the time (n = 76), against 0.75% and 61% for all months (n = 912).
December by decade
The edge shows up in most decades but not all, and the samples here are tiny: ten Decembers per decade, six so far in the 2020s.
| Decade | n | Average | Median | % positive | Worst December |
|---|---|---|---|---|---|
| 1920s (1928-1929) | 2 | 1.41% | 1.41% | 100% | 0.29% (1928) |
| 1930s (1930-1939) | 10 | -1.12% | 0.25% | 50% | -14.53% (1931) |
| 1940s (1940-1949) | 10 | 2.49% | 3.28% | 90% | -4.71% (1941) |
| 1950s (1950-1959) | 10 | 2.47% | 3.54% | 80% | -4.15% (1957) |
| 1960s (1960-1969) | 10 | 0.65% | 0.64% | 70% | -4.16% (1968) |
| 1970s (1970-1979) | 10 | 2.27% | 1.57% | 80% | -2.02% (1974) |
| 1980s (1980-1989) | 10 | 0.91% | 1.50% | 60% | -3.39% (1980) |
| 1990s (1990-1999) | 10 | 2.95% | 1.66% | 90% | -2.15% (1996) |
| 2000s (2000-2009) | 10 | 0.63% | 0.77% | 70% | -6.03% (2002) |
| 2010s (2010-2019) | 10 | 0.48% | 0.92% | 70% | -9.18% (2018) |
| 2020s (2020-2025) | 6 | 0.67% | 1.83% | 50% | -5.90% (2022) |
| All, 1928-2025 | 98 | 1.27% | 1.48% | 72% | -14.53% |
Since 2000 the December average has been well under 1%. December 2018 (down 9.18%) is the worst December since 1931, and it fell in the decade after the original version of this page was written.
First half vs second half of December
The back half of the month has done the heavy lifting. Here the first half runs from November’s last close to the last close on or before December 15, and the second half from there to December’s last close.
| Window | n | Average | Median | % positive |
|---|---|---|---|---|
| Dec 1-15, 1928-2025 | 98 | 0.02% | 0.77% | 60% |
| Dec 16-31, 1928-2025 | 98 | 1.25% | 0.97% | 74% |
| Dec 1-15, 1953-2025 | 73 | 0.05% | 0.45% | 58% |
| Dec 16-31, 1953-2025 | 73 | 1.26% | 1.04% | 77% |
| Dec 1-15, 2000-2025 | 26 | -0.18% | 0.06% | 50% |
| Dec 16-31, 2000-2025 | 26 | 0.76% | 0.63% | 69% |
The second half beat the first in 56 of 98 years, so the gap in averages is bigger than the year-by-year win rate. The 1953-2025 rows are there because the NYSE traded on Saturdays until September 29, 1952 (MathWorks’ NYSE calendar), and this data has weekday closes only.
The Santa Claus rally, defined precisely
The Santa Claus rally is a specific seven-session window: the last five trading days of December plus the first two trading days of January, a definition credited to Yale Hirsch of the Stock Trader’s Almanac (Yahoo Finance). The script measures it from the close of the sixth-to-last December session to the close of the second January session. The table compares it with every seven-session window in the same years.
| Window | n | Average | Median | % positive |
|---|---|---|---|---|
| Santa window, 1928-2025 | 98 | 1.60% | 1.38% | 77% |
| Any 7 sessions, 1928-2025 | 24616 | 0.22% | 0.39% | 57% |
| Santa window, 1928-1952 | 25 | 2.74% | 2.75% | 80% |
| Any 7 sessions, 1928-1952 | 6243 | 0.13% | 0.40% | 55% |
| Santa window, 1953-1999 | 47 | 1.43% | 1.34% | 77% |
| Any 7 sessions, 1953-1999 | 11834 | 0.27% | 0.37% | 57% |
| Santa window, 2000-2025 | 26 | 0.80% | 0.90% | 73% |
| Any 7 sessions, 2000-2025 | 6539 | 0.21% | 0.45% | 59% |
| Santa window, 2015-2025 | 11 | 0.23% | 0.43% | 64% |
| Any 7 sessions, 2015-2025 | 2766 | 0.35% | 0.54% | 62% |
Over the long run the window has beaten an ordinary week and a half by a wide margin. The edge has shrunk in each era, and in 2015-2025 it disappeared: an average of 0.23% against 0.35% for any seven sessions. Eleven windows is a small sample, so that could be noise, but the long-run figure no longer describes recent years well. The last three windows were all down:
| December of | From close of | To close of | Return |
|---|---|---|---|
| 2020 | Dec 23, 2020 | Jan 05, 2021 | 1.00% |
| 2021 | Dec 23, 2021 | Jan 04, 2022 | 1.43% |
| 2022 | Dec 22, 2022 | Jan 04, 2023 | 0.80% |
| 2023 | Dec 21, 2023 | Jan 03, 2024 | -0.88% |
| 2024 | Dec 23, 2024 | Jan 03, 2025 | -0.53% |
| 2025 | Dec 23, 2025 | Jan 05, 2026 | -0.11% |
What a missing Santa rally has meant for the next year
Hirsch’s old warning was that a failed Santa rally is a bad sign for the new year. The data gives that a mild yes at best.
| Santa window | n | Next-year average | Next-year median | Next year % positive |
|---|---|---|---|---|
| Down or flat | 22 | 6.07% | 5.03% | 64% |
| Up | 75 | 8.29% | 12.40% | 68% |
After the 22 down or flat windows (December 1928 to December 2024), the next calendar year still rose 64% of the time. The median gap is wider than the average gap, but with 22 cases one or two years move these numbers a lot. December’s record also ties into two other studies here: S&P 500 52-week highs set in December and the third year of the presidential cycle, which lands in 2027.
