Stock Market Seasonality: Every Month Ranked Since 1950, and Whether "Sell in May" Still Works

The stock market does have seasons, and the strongest one starts in two weeks. Since 1950 the S&P 500 has averaged 7.0% from the start of November to the end of April and 2.1% from the start of May to the end of October. November is the best single month of the year, up 70% of the time with an average gain of 1.86%. September is the worst, the only month with a negative average, and the only one that falls more often than it rises.

That is the honest version of "sell in May and go away", and it comes with two caveats that matter more than the headline: the summer half still goes up most years, and the gap between seasons is small next to the swing of any single year. We measured all of it from the index's month-end closes, February 1950 through September 2026, and this guide ranks every month, tests the two halves, checks whether the pattern survived the last 26 years, shows where 2026 sits, and says what seasonality should and should not change about the way you trade. It is education, not advice.

Every month ranked, 1950 to 2026

Average price change in each calendar month, with how often the month rose, and the best and worst example of each since 1950:

• November: +1.86% average, up 70% of the time. Best 2020 (+10.8%), worst 1973 (-11.4%).
• April: +1.54% average, up 70% of the time. Best 2020 (+12.7%), worst 1970 (-9.0%).
• December: +1.41% average, up 72% of the time. Best 1991 (+11.2%), worst 2018 (-9.2%).
• July: +1.29% average, up 60% of the time. Best 2022 (+9.1%), worst 2002 (-7.9%).
• January: +1.08% average, up 60% of the time. Best 1987 (+13.2%), worst 2009 (-8.6%).
• March: +0.96% average, up 64% of the time. Best 2000 (+9.7%), worst 2020 (-12.5%).
• October: +0.90% average, up 59% of the time. Best 1974 (+16.3%), worst 1987 (-21.8%).
• May: +0.44% average, up 62% of the time. Best 1990 (+9.2%), worst 1962 (-8.6%).
• June: +0.19% average, up 54% of the time. Best 1955 (+8.2%), worst 2008 (-8.6%).
• August: +0.07% average, up 56% of the time. Best 1982 (+11.6%), worst 1998 (-14.6%).
• February: -0.04% average, up 53% of the time. Best 1986 (+7.1%), worst 2009 (-11.0%).
• September: -0.62% average, up 44% of the time. Best 2010 (+8.8%), worst 1974 (-11.9%).

Three things stand out. The top three months, November, April and December, are also the three most consistent, each rising at least 70% of the time. The bottom of the table is not a cluster of crashes but of flat, coin-flip months: February and June barely move on average, and August's average is dragged to near zero by 1998 while its median is a respectable +1.1%. And September really is different: it is the only month that is negative on average, negative at the median, and down in a majority of years. Our September piece goes through the reasons people give for that.

The famous worst months are not the worst months on average. October holds the two most violent single months in the table, +16.3% in 1974 and the 21.8% drop of 1987, yet averages +0.90% and rises 59% of the time. That is why the October effect is more a memory of 1929, 1987 and 2008 than a statistical tendency.

Does "sell in May" actually work?

The six-month version of seasonality compares the winter half, November through April, with the summer half, May through October. Since 1950 the winter half has averaged 6.95% (median 6.2%) and finished higher 76% of the time. The summer half has averaged 2.08% (median 2.7%) and finished higher 66% of the time. Compounded, $100 invested only during the winter halves since 1950 would have grown to about $11,664, price only; $100 invested only during the summer halves would have grown to about $344.

So the gap is real and large over 76 years. But look at what "sell in May" asks you to give up. The summer half still rose in two of every three years, and its best run was last year: +22.8% from May to October 2025. Its worst was 2008 (-30.1%), which is where the strategy earned its reputation. A rule that steps out of the market every May skips a positive expected return, pays taxes and spreads to do it, and would have missed most of 2025's advance. Measured as a trading rule rather than a factoid, the honest summary is: the market has tended to do less in summer, not to go down in summer.

The strongest stretch of all is the fourth quarter. October through December has averaged 4.2% and finished higher 80% of the years since 1950, with November and December together averaging 3.3% and rising 78% of the time. The worst fourth quarter was 1987 (-23.2%), the best 1998 (+20.9%). Our Q4 2026 outlook puts this year's version in that frame.

Has it changed since 2000?

Seasonal patterns have a way of fading once everyone knows them, so we re-ran the table on the 26 years since January 2000. November is still the best month, +2.12% on average and up 77% of the time. September is still the worst, -1.28% and up only 48% of the time. April (+1.89%, up 70%) and July (+1.48%) kept their strength.

What faded is January. From 1950 to 1999 the so-called January effect was real at the index level; since 2000 January has averaged -0.02% and risen only 52% of the time, with the 2009 and 2008 declines doing most of the damage. February has been the weakest month of the modern era after September, at -0.51%. And May, the month the saying tells you to sell, has been one of the most reliable since 2000: up 74% of the time. The broad shape, strong winter, flat early summer, weak September, strong finish, has held. The month-by-month detail has moved around, which is the strongest argument against trading any single month on its reputation.

Where 2026 sits

Through September, 2026 has followed the script more than it has broken it: January +1.4%, February -0.9%, March -5.1%, April +10.4%, May +5.2%, June -1.1%, July -0.1%, August +2.6%, September -0.5%. A weak February and March, a very strong April (the best month of the year, as it is on average), a flat early summer, and a slightly negative September. October so far has delivered the kind of swings the month is known for. The seasonally strong half starts on November 1, with the Fed's October 28 decision and the end of the Q3 earnings season sitting right at the handoff.

What seasonality should change about your trading

• Not your direction. The worst month still rises 44% of the time and the best month still falls 30% of the time. A seasonal edge of a percent or two per month is swamped by the normal month-to-month swing of four to five percent. If a calendar tendency is making you buy or sell, the position is being sized by the wrong number.
• Maybe your expectations. Knowing that September is flat-to-down on average and that the winter half carries most of the year's return is useful for patience: a quiet summer is normal, not a signal that something is wrong. How often the market goes up gives the base rates by holding period; seasonality is a small tilt on top of them.
• Your risk budget, a little. September and October hold the largest drawdowns and the largest single-month swings. If you run a fixed daily loss limit or a per-trade stop, those are the months in which they earn their keep, which argues for keeping them tight through the fall rather than loosening them to "let it run". Our drawdown study shows how deep the fall episodes have gone.
• Not your long-term plan. For money with a multi-year horizon, the winter-versus-summer gap is smaller than the cost of being out of the market during a summer like 2025. Lump sum versus dollar-cost averaging covers that decision with the same data.

If you would rather have the stop, the daily loss brake and the position cap enforced for you in every month, that is what JorgAI does inside the brokerage account you already have: set the rules once and they hold in September and November alike.

Quick answers

What is the best month for the stock market? November. Since 1950 the S&P 500 has averaged +1.86% in November and risen 70% of the time, with April (+1.54%) and December (+1.41%) next. Since 2000 November is still first.

What is the worst month for stocks? September. It is the only month with a negative average (-0.62%) and the only one that has fallen in a majority of years (56% down). Since 2000 it averages -1.28%.

Does "sell in May and go away" work? The November-to-April half has averaged 7.0% since 1950 against 2.1% for May to October, so the tendency is real. But the summer half still rose 66% of the time and gained 22.8% in 2025, so selling in May has meant giving up a positive return, not avoiding a negative one.

Is the fourth quarter good for stocks? Historically the strongest quarter: October through December has averaged 4.2% and finished higher 80% of the years since 1950.

Is the January effect still real? Not at the index level. January averaged over 1% from 1950 to 1999 but has averaged roughly zero since 2000, rising in about half of the years.

Data: S&P 500 daily closes from Yahoo Finance (ticker ^GSPC), January 3, 1950 through October 15, 2026; monthly returns use the last close of each month, February 1950 through September 2026 (October 2026 is excluded because it is incomplete). Price only, no dividends. Averages are arithmetic means; the six-month and quarterly figures compound the monthly closes. Past patterns do not predict future returns. This article is education, not investment advice.