Does "sell in May and go away" really work?

What do people believe about selling in May?

The belief is that summer is the weak half of the year for stocks. So sell in May, and buy back around Halloween.

It started as an old London saying, which in full goes "Sell in May and go away, come back on St Leger's Day". St Leger's Day is the day of a famous horse race in September.

In the US, a long-running trading almanac made it famous from 1987 as a "best six months" plan. A well-known 2002 study then found the same pattern in many countries.

What exactly did we test?

We split every year into two halves and tested two claims, with the rules written down before we looked at any results.

Summer runs from the end of April to the end of October. Winter runs from the end of October to the end of April.

Claim one: summer does worse than winter. Claim two: summer does worse than cash. The "go away" part needs claim two, because going away means sitting in cash.

We used the whole US stock market from October 1926 to August 2026. That gives 99 summers and 100 winters.

Every return is measured above what cash in Treasury bills earned at the same time. When we compare two averages, the gap is in percentage points, or points for short.

Is summer really the weaker half?

Yes, summer really has been the weaker half. Summers averaged +2.3% above cash, against +6.1% for winters, a gap of 3.8 points.

It wasn't a fluke of one era. Summer trailed winter by 3.1 points in the first half of our summers, up to 1977, and by 4.6 points in the second half.

The last 20 years agree. Summers averaged +3.9% above cash and winters +6.4%, so summer minus winter was -2.5 points.

So the old London saying spotted something real. That part is math.

How sure can we be?

Fairly sure, but it's a closer call than the saying suggests. The likely range of the gap runs from -7.3 to -0.3 points, which stays clear of zero, but only just.

With about a hundred summers, one wild year can move the average a lot. And we test many calendar beliefs, so some will pass by luck.

We widened the range because we checked several things at once. At 99.67%, it runs from -9.0 to +1.2 points. That includes zero, so we can't rule out luck completely.

Has it faded since it got famous? No. Summer minus winter was -3.5 points before 1987 and -4.4 points since. The gap was a bit bigger after fame, not smaller. In the last 20 years, summer still trailed winter, by 2.5 points.

Neither stretch is sure on its own. Their likely ranges, -8.5 to +1.5 points and -9.0 to +0.1 points, both include zero, because each has only part of the summers.

Does summer lose to cash?

No, and this is where the saying goes wrong. Summer was weaker than winter, but it still averaged +2.3% above cash.

About 67% of summers beat cash. Summers stayed ahead of cash in both halves of the century, at +2.5% and +2.1%.

In the last 20 years, summers averaged +3.9% above cash, and 75% of them beat cash.

Summer is like the weaker half of a good football team. It scores less than the other half, but it still scores.

Some summers were awful, of course. The worst was 1930, when stocks ended 31.8 points behind cash. Bad summers make great stories, which may be why the saying sticks.

So the claim that summer loses to cash is a myth. Its likely range, -0.2 to +4.7 points, sits mostly above zero, and the average points the other way.

Does it hold everywhere?

Summer was weaker than winter in 42 of the 43 industries with enough history. But summer lost to cash in only 7 of them.

Industries all move with the market, so these aren't independent votes. Still, the picture is the same across the board.

By decade, summer trailed winter most of the time. The 1930s, the 1940s and the 2020s so far went the other way.

Which months make the summer weak?

Not the holiday months: June to August held up, and the one clear dip was September, though not clearly in the last 20 years.

The saying is told three ways, so we tested all three at once; picking the best one would be tuning the dates. You can go away from May or June until October and come back in November. Or, as in the UK original, go away for June to August and come back in September.

The chart shows the average month above cash in each block, from 1927 to 2025. December is left out, because it belongs to the Santa and year-end tests.

June to August was the best of the longer blocks: +1.00% a month above cash, and +0.94% in the last 20 years. No stretch of time away was clearly weaker than the months before it.

This doesn't clash with our main result. Summer's shortfall is small for each month. In the US version, the average month away minus January to April was -0.34 points, or -0.20 points in the last 20 years. Cut into smaller blocks, with several things checked at once, no block is clearly weaker except September. And the main test's winter also holds November and December.

November, the month you come back in the US version, looked strong: +1.54% a month above cash, and +2.32% in the last 20 years. It beat the months away by +1.13 points, but one month swings a lot, so even that could be luck.

September is the one clear result. It averaged -1.06% a month against cash, or -0.51% in the last 20 years. Over the century, its gap below June to August stayed clear even after we widened the range because we checked several things at once. In the last 20 years, it could be luck.

The horses run in mid-September; over the whole month, stocks on average stumbled. Whether September is the worst month of all gets its own test.

Why might it seem to work?

People offer reasons such as summer holidays and thinner trading. We didn't test them; they are reasons people give. June to August held up in our month blocks, though.

The memory trick is easier to see. A bad summer, like 1930, sticks in the mind.

Calm summers that quietly made money are much less memorable.

What does this mean for you?

This is history, not a forecast. Summer has tended to be the weaker half, but that tells you little about any single year.

A myth doesn't mean you should do the opposite. Piling into stocks every May because the cash part failed would be its own untested bet.

Going away cost money. Holding stocks only from November to April, with cash in summer, grew +8.2% a year in total return. Staying invested grew +9.9% a year in total return. In the last 20 years of data, the same rule grew +7.4% a year, against +11.5% for staying invested.

The switching plan's worst fall was smaller, -58.6% against -83.9%. That's mostly because it sat in cash about half the time. In the last 20 years, its worst fall was -35.4%, against -54.7%.

What else did we test?

See all the calendar effects we tested, or how we test a myth.