Learn the stock market › Studies
Sell in May and go away: is it true? The market month by month since 1928
Partly. From November to April the S&P 500 did better than from May to October in 65 of 98 years since 1929, on average +5.2% against +2.4%. But May to October rose too, in 65 of 98 years, so the whole year gave more than the winter half alone.
S&P 500: average month since 1928
What the saying claims
"Sell in May and go away" is a market saying: stocks do worse from May to October than from November to April. Here we test it on real prices. The winter half is November to April, and the summer half is May to October.
The market's best and worst months
The table shows how much each month has changed on average, counted from the last close of the month before. For the S&P 500, July was best at +1.7% and September worst at −1.1%. The index rose most often in December, 72% of the years, and least often in September, 44%.
| Month | S&P 500 | Rose | OMXS30 | Companies |
|---|---|---|---|---|
| November | +1.0% | 62% | +1.8% | +4.5% |
| December | +1.3% | 72% | +0.9% | +1.0% |
| January | +1.2% | 63% | +1.7% | +2.9% |
| February | −0.1% | 52% | +1.8% | +0.8% |
| March | +0.4% | 61% | −0.9% | +2.0% |
| April | +1.4% | 66% | +2.4% | +3.0% |
| May | +0.1% | 61% | −0.5% | +3.1% |
| June | +0.8% | 57% | −0.3% | +0.9% |
| July | +1.7% | 61% | +2.8% | +1.9% |
| August | +0.7% | 59% | −1.2% | +2.0% |
| September | −1.1% | 44% | +0.6% | −0.1% |
| October | +0.5% | 59% | +1.2% | +3.1% |
When twelve months are compared, one is always best, by chance too. So the two halves below say more than any single month.
The winter half against the summer half
A year here is one winter half and the summer half right after it. The column counts how many years the winter half gave more, and the numbers to the right are the average per half.
| Series | Winter half ahead | Nov–Apr | May–Oct |
|---|---|---|---|
| S&P 500 | 65 of 98 | +5.2% | +2.4% |
| OMX Stockholm 30 | 12 of 17 | +7.2% | +1.2% |
| The 20 US companies | 237 of 449 | +13.7% | +12.4% |
| Stockholm's 30 companies | 435 of 673 | +13.5% | +3.1% |
For the S&P 500 the winter half gave more in 65 years and the summer half in 33.
Toss a coin 98 times and the number of heads lands between 39 and 59 in 95 cases of 100, so 65 is more than chance usually gives.
The average gap is +2.8 percentage points, and the range of chance is −0.7 to +6.2. That includes zero, because the market years around 1930 swung so much. So how often the winter half did better can be told apart from chance, but not how much better it did.
What the money would have become
The summer half was the weaker one, but it rose more often than it fell. That shows when the halves are added up. From October 1928 to October 2026, $100 in the S&P 500 became:
- in the winter halves only: $7,020
- in the summer halves only: $498
- the whole time: $34,963
These sums leave out dividends, and the money earns no interest during the half outside the market. Fees and tax at each switch are not counted either. What the market has returned over longer spans is in the study of the average return.
Has it always been like this?
No. From 1929 to 1949 the summer half usually gave more, in 12 of 21 years. Since then the winter half has led in every period.
| Years | Winter half ahead | Nov–Apr | May–Oct |
|---|---|---|---|
| 1929–1949 | 9 of 21 | −1.4% | +3.5% |
| 1950–1979 | 24 of 30 | +7.0% | +0.3% |
| 1980–2009 | 19 of 30 | +6.6% | +2.4% |
| 2010–2026 | 13 of 17 | +7.8% | +4.9% |
Only 1950–1979 falls outside the range of chance, with a gap of +6.7 points (+3.0 to +10.4). Since 2010 the gap is +2.8 points (−2.7 to +8.4).
Company by company
This is what the indexes don't show. The companies count with dividends reinvested, from the first close we hold, at the earliest from 2001.
At 10 of the 20 largest US companies on our list, the winter half gave more in most years. At 7 it was the summer half, and the rest are even. At 13 of 20 the winter average was higher. Together that is +13.7% against +12.4%, a gap of +1.3 points with a range of −3.9 to +6.6.
That range includes zero, so it cannot be told apart from chance.
Stockholm looks different. At 28 of the 30 companies in OMX Stockholm 30 the winter half gave more in most years, and at 0 the summer half did. On average that is +13.5% against +3.1%, with a range of +4.2 to +16.7, outside chance. The index itself, over 17 years, gave +7.2% against +1.2%.
A company has at most 26 years in the table, so a single row can shift a lot by chance. The table is sorted by the average gap, largest first.
| Company | Winter half ahead | Nov–Apr | May–Oct |
|---|---|---|---|
| Netflix | 14 of 24 | +30.9% | +21.4% |
| AMD | 14 of 26 | +20.9% | +12.2% |
| Exxon Mobil | 16 of 26 | +9.9% | +1.7% |
| JPMorgan Chase | 18 of 26 | +10.6% | +3.3% |
| Mastercard | 11 of 20 | +17.4% | +10.2% |
| Meta | 7 of 14 | +21.0% | +14.7% |
| Eli Lilly | 18 of 26 | +10.7% | +5.7% |
| Broadcom | 12 of 17 | +22.8% | +18.2% |
| Nvidia | 13 of 26 | +23.2% | +19.6% |
| Visa | 11 of 18 | +12.6% | +9.1% |
| Berkshire Hathaway | 17 of 26 | +7.3% | +3.8% |
| Amazon | 12 of 26 | +15.8% | +13.0% |
| Walmart | 14 of 26 | +6.5% | +4.7% |
| Costco | 12 of 26 | +7.5% | +8.8% |
| Palantir | 2 of 6 | +43.1% | +46.4% |
| Microsoft | 8 of 26 | +5.1% | +9.8% |
| Oracle | 12 of 26 | +1.4% | +9.1% |
| Alphabet | 6 of 22 | +6.9% | +16.9% |
| Apple | 12 of 26 | +12.1% | +23.2% |
| Tesla | 8 of 16 | +22.4% | +33.7% |
How we counted
- Each month's last close to Nov 6, 2026: the S&P 500 from Dec 30, 1927 and OMX Stockholm 30 from Nov 20, 2008. The running month does not count.
- The S&P 500 index began in 1957. Numbers before that come from its predecessor with 90 companies.
- A year counts only when both halves are complete. So the S&P 500 has 98 years, from 1929, and OMX Stockholm 30 has 17, from 2010.
- The indexes leave out dividends and the companies include them. A dividend lands in the half where its ex-date falls.
- We checked dividends on 5 November 2026. Of 593 ex-dates at the US companies since 2015, 292 fell in the winter half, so they tilt neither half. In Stockholm 408 of 619 fell in the winter half, most of them in April, and taking dividends out at the 24 companies where we can shrank the gap from 11.0 to 9.2 points. So they explain some of it there, not most.
- The range of chance is 95% around the average gap. Companies count per year, because many move in the same years.
- We split the year where the saying splits it. A different split gives different numbers.
- The halves are compared with each other, not with interest on an account. Inflation is not taken out.
- The S&P 500 is in dollars, OMX Stockholm 30 in kronor and the companies in the currency the stock trades in.
- 20 of the 20 largest US companies on our list. Companies that went bankrupt or left the exchange are missing, so the averages are higher than reality in both halves.
You are welcome to use the numbers
Link here, quote the table or embed the chart on your own page. Name ChartCompass as the source and keep the link.