Stock market seasonality is the idea that returns differ by time of year. The best-known patterns are the "best six months" (November through April beating May through October, the flip side of "sell in May and go away"), the Santa Claus rally at the turn of the year and the January effect. In the Nasdaq Composite, November-April rose 8.8% on average from 1971 to 2026 and May-October 3.6% [2]. In the S&P 500 since November 2016, the order reversed: 6.8% for November-April and 7.4% for May-October [1]. The patterns show up in long histories, but they are too weak and irregular to rely on in any single year.
What is the "best six months" or "sell in May" pattern?
The Stock Trader's Almanac's "Best Six Months" strategy holds the Dow Jones Industrial Average from November 1 to April 30 each year and switches to fixed income for the other six months. The Almanac says most market gains since 1950 came from November through April, and that it first identified the pattern in 1986 [3]. Academic research reached a similar conclusion: Sven Bouman and Ben Jacobsen, in the American Economic Review in December 2002, found the "sell in May" effect in 36 of the 37 stock markets they studied, using data from 1970 to 1998 [5][6]. They called it the Halloween indicator, because the strong half starts at the end of October.
DataPorium tested the pattern on daily closing prices. Each November-April return runs from the last close of October to the last close of April, and each May-October return from the last close of April to the last close of October. Returns are price changes, without dividends.
| Index and half-year | Periods | Average | Median | Share positive |
|---|---|---|---|---|
| Nasdaq Composite, November-April (1971-72 to 2025-26) | 55 | 8.8% | 9.5% | 73% (40 of 55) |
| Nasdaq Composite, May-October (1971 to 2025) | 55 | 3.6% | 2.4% | 65% (36 of 55) |
| S&P 500, November-April (2016-17 to 2025-26) | 10 | 6.8% | 6.5% | 70% (7 of 10) |
| S&P 500, May-October (2017 to 2025) | 9 | 7.4% | 8.0% | 89% (8 of 9) |
Source: DataPorium calculation from FRED daily closes [1][2].
The Nasdaq Composite, which covers all stocks listed on the Nasdaq market, is the longest daily index history in the public FRED database; the free S&P 500 series there starts in October 2016 [1][2]. Even in the Nasdaq data, the November-April half beat the May-October half before it in only 34 of 55 years, or 62% of the time [2].
How did the pattern do in recent years?
In the S&P 500, the "worse" half was often the better one. May-October gained 22.8% in 2025 and 13.3% in 2024, while November 2024 to April 2025 lost 2.4% [1]. The winter half lost 10.3% from October 2021 to April 2022, at the start of that year's bear market, and 4.1% from October 2019 to April 2020, which included the pandemic crash [1]. Across the nine full cycles from 2017 to 2025, November-April beat the May-October before it only four times [1].
The latest winter half, from October 31, 2025 to April 30, 2026, gained 5.4% (6,840.20 to 7,209.01). From April 30 to October 12, 2026, the summer half was up 8.3%, with the last weeks of October still to come [1].
What is the Santa Claus rally?
The Santa Claus rally, devised by Yale Hirsch in 1972 and first published in the 1973 Stock Trader's Almanac, is the last five trading days of the year plus the first two trading days of the new year [4]. It is not the same as a general year-end rally. The Almanac reports an average S&P 500 gain of 1.3% over these seven trading days since 1950 [4]. For the window that ran from December 24, 2025 to January 5, 2026, DataPorium measured a 0.1% decline in the S&P 500 (6,909.79 at the December 23 close to 6,902.05 on January 5) [1].
| Santa Claus rally window | S&P 500 change |
|---|---|
| 2016-17 | +0.4% |
| 2017-18 | +1.1% |
| 2018-19 | +1.3% |
| 2019-20 | +0.3% |
| 2020-21 | +1.0% |
| 2021-22 | +1.4% |
| 2022-23 | +0.8% |
| 2023-24 | -0.9% |
| 2024-25 | -0.5% |
| 2025-26 | -0.1% |
Source: DataPorium calculation from FRED daily closes, from the close before the last five December trading days to the second January trading day [1].
Over these 10 windows, the S&P 500 rose 7 times, with an average gain of 0.5%; the last three were negative [1]. In the Nasdaq Composite from 1971-72 to 2025-26, the window rose in 40 of 55 years, 73%, with an average gain of 1.6% [2]. For comparison, all seven-trading-day stretches in the same Nasdaq history rose 59% of the time, with an average of 0.3% [2]. The Santa Claus window has been better than a typical week and a half, but it has also failed in about one year in four.
What is the January effect?
The January effect is the tendency for stocks, especially small companies, to do better in January than in other months. A common explanation is that investors sell losing positions in December to realize tax losses and buy back in January.
In the Nasdaq Composite, January has been the best month on average, though not the most reliable one:
| Month | Nasdaq average, 1971-2026 | Nasdaq share positive | S&P 500 average, 2016-2026 | S&P 500 share positive |
|---|---|---|---|---|
| January | 2.5% | 67% | 2.1% | 70% |
| April | 1.6% | 66% | 2.1% | 70% |
| July | 0.9% | 57% | 3.0% | 90% |
| September | -0.7% | 54% | -1.4% | 50% |
| November | 2.1% | 71% | 4.1% | 90% |
| December | 1.5% | 60% | 0.1% | 60% |
Source: DataPorium calculation from FRED month-end closes; Nasdaq has 55 or 56 observations per month and the S&P 500 9 or 10 [1][2].
September stands out as the only month with an average loss in both series [1][2]. November has been strong in both, which helps explain why the November-April half looks good in long histories.
Why is seasonality weak evidence?
- Small samples. There is one Santa Claus window and one "sell in May" switch per year. The S&P 500 series used here has 10 of each, and even the 55-year Nasdaq history is a modest sample for a pattern that fails one year in three or four [1][2].
- Data mining. When thousands of calendar windows are tested on the same history, some will look strong by chance. A pattern found after the fact needs to keep working on new data.
- Big events dominate. Single episodes such as the 2020 pandemic crash or the 2022 bear market move the half-year averages more than the calendar does [1].
- Costs and taxes. A strategy that sells every spring and buys every fall pays trading costs and, in taxable accounts, may realize gains each year. Back-tested averages ignore both.
- Being out of the market has a price. In the S&P 500, May-October was positive in 8 of the last 9 years, averaging 7.4% [1].
Index levels, sector moves and the market calendar are on DataPorium's stock market page [7].
Seasonal patterns such as "sell in May" and the Santa Claus rally show up in long market histories, but they fail often enough that the calendar is a weak guide for any single year.
Key takeaways
- In the Nasdaq Composite since 1971, November-April averaged 8.8% and May-October 3.6%, but the winter half beat the summer half in only 34 of 55 years.
- In the S&P 500 since 2016, the pattern reversed: May-October averaged 7.4% and November-April 6.8%.
- The Santa Claus rally (last 5 trading days of December plus the first 2 of January) rose in 40 of 55 Nasdaq windows, averaging 1.6%; in the S&P 500 it fell in each of the last 3 windows.
- January has been the best Nasdaq month on average since 1971 (2.5%), and September the only month with an average loss in both indexes.
- With 10 to 55 observations per pattern, seasonality describes the past; it is weak evidence for any one year.
Frequently asked questions
Does "sell in May and go away" work?
Over long histories, November-April has beaten May-October on average: 8.8% versus 3.6% in the Nasdaq Composite since 1971 [2]. But it fails often; in the S&P 500 since 2016, May-October did better on average, at 7.4% [1].
When is the Santa Claus rally in 2026?
By the Stock Trader's Almanac definition, it covers the last five trading days of December 2026 and the first two trading days of January 2027 [4]. The exact dates depend on the exchange holiday calendar.
What is the best month for the stock market?
It depends on the index and period. In the Nasdaq Composite since 1971, January had the highest average return (2.5%) and November the highest share of gains (71%); in the S&P 500 since 2016, November led with a 4.1% average [1][2].
What is the worst month for stocks?
September. It was the only month with an average decline in both the Nasdaq Composite since 1971 (-0.7%) and the S&P 500 since 2016 (-1.4%) [1][2].
Sources & References
- [1] Federal Reserve Bank of St. Louis, FRED, S&P 500 (SP500), daily close, October 13, 2016 to October 12, 2026 (accessed October 13, 2026)
- [2] Federal Reserve Bank of St. Louis, FRED, NASDAQ Composite Index (NASDAQCOM), daily close since February 5, 1971 (accessed October 13, 2026)
- [3] Stock Trader's Almanac, Our Strategy (Best Six Months) (accessed October 13, 2026)
- [4] Stock Trader's Almanac (Jeff Hirsch, editor), Real Santa Claus Rally: It's An Indicator! (December 1, 2025)
- [5] Sven Bouman and Ben Jacobsen, The Halloween Indicator, "Sell in May and Go Away": Another Puzzle, American Economic Review 92(5), December 2002
- [6] The Journalist's Resource (Harvard Kennedy School Shorenstein Center), The Halloween strategy: stock market lore or scary real? (October 16, 2019)
- [7] DataPorium Stock Market (US market overview, sectors and index data)