"Sell in May and go away" is one of the oldest sayings in investing. The idea is that stocks do most of their rising between November and April, so you can sell at the start of May, sit in cash over the summer, and buy back in at the start of November without missing much.

Is there anything to it? We tested the idea against more than 150 years of S&P 500 data, first by looking at how each calendar month has performed, and then by comparing an investor who sells every May with one who simply stays invested.

How we tested it

We used Robert Shiller's monthly S&P 500 dataset with dividends reinvested, from 1871 to mid-2023. For each year we calculated two six-month returns:

  • "Summer" (May–October): from the April price to the October price.
  • "Winter" (November–April): from the October price to the following April price.

Shiller's prices are monthly averages of daily closes, so these are close to — but not exactly the same as — returns from the last day of April to the last day of October.

Average return by month

Month Average return Median return Share of years the month rose
January 1.86% 1.78% 72.4%
February 0.72% 0.70% 54.9%
March 0.51% 0.79% 62.7%
April 1.06% 1.26% 67.3%
May 0.46% 0.66% 58.2%
June 0.45% 0.82% 58.2%
July 1.22% 1.14% 69.1%
August 1.26% 1.12% 64.5%
September 0.56% 1.14% 61.8%
October 0.01% 0.40% 57.2%
November 1.00% 1.58% 64.5%
December 0.70% 1.05% 63.8%
All months 0.82% 0.99% 62.9%

Every month rose more often than it fell. October had the weakest average — pulled down by a handful of severe months such as October 2008 (−20%), 1937 (−14%) and 1987 (−12%) — yet its median return was still positive. July and August, right in the middle of the supposed weak season, were among the strongest months over the full period. On this evidence alone, no month stands out as one to avoid.

Summer vs winter

Period Years Average May–Oct return Average Nov–Apr return May–Oct rose Nov–Apr beat May–Oct
1871–2022 152 4.2% 6.1% 64.5% of years 50.7% of years
1871–1949 79 4.8% 4.1% 62.0% 46.8%
1950–2022 73 3.6% 8.4% 67.1% 54.8%
1990–2022 33 3.0% 8.1% 63.6% 60.6%

Here is where the saying gets some support. Since 1950, the November–April half-year has averaged more than double the May–October half — 8.4% against 3.6%. Before 1950 there was no such pattern; summers were, if anything, slightly better.

But look at the fifth column. Even since 1950, the summer half still made money in about two years out of three. A weaker season is not the same as a losing one.

Does selling every May actually pay?

We compared two investors from 1871 to 2022:

  • Buy and hold: invested all year.
  • Sell in May: invested November to April, out of the market May to October.

The key question is what the sell-in-May investor earns while out of the market. We ran it two ways: with cash earning nothing, and with cash earning the 10-year Treasury yield. The second is generous, because the long-term yield is usually higher than what a savings account or short-term bill pays.

Period Buy and hold Sell in May, cash earns 0% Sell in May, cash earns the 10-year yield
1871–2022 9.1% a year 5.5% 7.9%
1871–1949 7.2% 3.4% 5.2%
1950–2022 11.2% 7.9% 10.8%
1990–2022 10.1% 7.6% 9.9%

Buy and hold came out ahead in every period. With uninvested cash earning nothing, selling in May cost between 2.5 and 3.8 percentage points a year, depending on the period. Even with the generous interest assumption, it still lagged — and that is before trading costs and taxes on selling every spring, which would widen the gap further.

Year by year, the picture is similar. Since 1950, selling in May beat staying invested in only 33% of years with cash earning nothing, and in 45% of years with cash earning the 10-year yield.

Why the saying survives

Seasonal patterns are easy to remember and occasionally look brilliant. The worst May–October periods in the data — 1931 (−33%), 2008 (−29%), 1930 (−28%), 1937 (−26%) and 1974 (−23%) — were all bear-market years when getting out would have helped enormously.

But the strategy has to be followed every year to capture those, and the misses are costly. Recent years show it clearly. On price alone, the S&P 500 rose about 13% from April to October 2024, and about 25% from April to October 2025. An investor who sold in May would have missed both — and the November 2024 to April 2025 half-year, when they would have been invested, fell about 7%.

This mirrors what we found for other timing rules: being out of the market during its strongest stretches is expensive. See what happened after all-time highs and lump sum vs dollar-cost averaging.

Practical takeaways

  1. Don't treat the calendar as a signal. Since 1950 winters have been stronger on average, but summers were still positive most of the time, and selling every May lagged staying invested in every period we tested.
  2. Expect dips in any season. A 10% fall at some point has happened in about one year in three, regardless of the month — see how often the S&P 500 falls 10%, 20% or 30% in a year.
  3. If you want less risk, change your mix, not your timing. Holding more bonds or cash all year is a more reliable way to reduce swings than trying to dodge particular months. Compare risk levels in our calculator.

Notes

Data: Robert J. Shiller, U.S. Stock Markets 1871–Present, via datasets/s-and-p-500. Total returns with dividends reinvested, using monthly average prices, through June 2023; the 2024 and 2025 figures are price-only and use FRED data. "May–October" runs from the April monthly average to the October monthly average, and "November–April" from October to the following April. The 10-year Treasury yield is used as a generous upper-bound proxy for interest on cash. Trading costs and taxes are ignored. Calculations are our own.

This guide is educational and not investment advice. Past performance does not guarantee future results.