Sell in May and go away: does it work? We tested 30 years of real data

The words JJ read on camera for this test. The final video may differ slightly.

The old stock market saying is right. Over the last thirty years, stocks really did do worse from May through October. About half as well as the other six months. And if you had followed the saying, you would have ended up with about half the money. Today we find out how both of those things can be true.

Hi, I'm JJ, and this is The $10,000 Test. Every week I take a piece of money advice everybody repeats and test it against real market history. Not opinions. What the money actually did. Today's advice: "Sell in May and go away." You hear it every spring. Stocks do badly from May through October, so sell in May, sit in cash for the summer, and come back in November. It's an old saying. In Britain, the long version is "Sell in May, go away and come again on St Leger Day." That's a horse race in September. The version Americans know comes from the Stock Trader's Almanac. The Associated Press says market historian Yale Hirsch first popularized it in the mid-eighties. The Almanac calls it the "Best Six Months," and its website says, quote, "most market gains occur during the months November through April." And people believe it for good reasons. In 2009 the AP reported that since 1950, the Dow averaged a gain of seven point three percent from November through April, and almost nothing, one tenth of one percent, from May through October. The usual explanation: everybody's on vacation, so the summer market is sluggish. And some of the worst stretches in memory hit in those months: September 2001, the summer of 2002, the fall of 2008. So let's test it.

Here's exactly what we did, so you can check it. Ten thousand dollars, starting in January 1996, running right up to October 2026. Plan A: buy the S&P 500 and hold it. We used SPY, an S&P 500 index fund, with every dividend reinvested. You never sell. Plan B: sell in May. You own the same fund from November through April. On the last trading day of April, you sell everything and move it into Treasury bills, about the safest cash there is, earning whatever T-bills paid at the time. On the last trading day of October, you buy back in. Every year, for thirty years. To keep it simple, I'll call May through October "summer," and November through April "winter." Two things we left out: trading costs and taxes. Plan B makes sixty-one trades. Plan A makes none. So everything you're about to see is the best case for selling in May.

First question: is the saying even right about the calendar? Mostly, yes. Over thirty summers, from 1996 through 2025, stocks gained about three point seven percent per summer, on average. Over thirty winters, about seven point three. Summer earned about half as much. But here's what surprised me. Look at how often summer actually lost money: seven times out of thirty. Winter lost money six times. Almost the same. What makes summer look so bad is three terrible ones. Summer 2001, down about sixteen percent. Summer 2002, down about seventeen. And summer 2008, down twenty-nine percent. Ten thousand dollars in stocks on the last day of April 2008 was worth about seven thousand eighty by Halloween. In T-bills, it was ten thousand seventy-four. Outside of those, summer was usually fine. Twenty-three of the thirty summers made money, and twenty-two of them beat T-bills.

Now the part that matters. The money. Plan A, buy and hold: ten thousand dollars became about two hundred eleven thousand dollars. Plan B, sell in May: about one hundred twelve thousand. Selling in May cost you almost ninety-nine thousand dollars for every ten thousand you started with. That's about ten point four percent a year, against about eight point two. Here's the twist. For a long time, it looked brilliant. Sell in May was ahead at the end of every single year from 2001 through 2016. Sixteen years in a row. In June 2012, it was ahead by about eleven thousand dollars: thirty-eight thousand against twenty-seven. Then, in May 2017, it fell behind. And it never caught up. What changed was the summers. From 1996 through 2008, the average summer actually lost a little money, and five of those thirteen summers were losers. From 2009 through 2025, the average summer gained about six point seven percent, and only two of seventeen lost money. And in ten of those summers, T-bills paid less than a fifth of one percent for the whole six months. You'd have sold your stocks to earn next to nothing while the market went up.

"OK, JJ," you might say, "I'm not investing for thirty years. What about ten?" We checked every ten-year stretch in the data, starting every month from January 1996 to October 2016. Two hundred fifty of them. Sell in May won ninety-three. More than a third. But look where they are. Every one of those ninety-three wins started before November 2005. For the one hundred thirty-two ten-year stretches that started after that, buy and hold won every single one. And here's the year that shows why this is so hard to follow. May 2009. The market had just crashed. The Associated Press wrote that selling in May "might sound awfully tempting this year." If you sold that May, you missed a gain of almost twenty percent. Ten thousand dollars in stocks became about eleven thousand nine hundred seventy by the end of October. In T-bills, you earned seven dollars. To be fair, over 2008 and 2009 together, selling in May still came out ahead. From the end of April 2008 to the end of October 2009, ten thousand dollars fell to about seven thousand seven hundred seventy if you stayed in, and only to about nine thousand two hundred forty if you sold every May. And the most recent full summer, 2025, was the best of all thirty: up almost twenty-four percent. This year, from the end of April to early October, stocks are up about eight percent so far.

So how can a pattern that's real lose this much money? First, weaker isn't the same as bad. The average summer earned about half what the average winter did. But half is still a gain, and usually more than cash paid. Every summer you sit out, you give that up. Add up all those summers, and that's your ninety-nine thousand dollars. Second, crashes don't check the calendar. The 2020 virus crash, down about thirty-four percent, ran from February 19 to March 23. That's the winter half. In the last seven winters, three lost money. In the last seven summers, only one did. Third, the worst fall for buy and hold, about fifty-five percent, ran from October 2007 to March 2009, right across both seasons. Selling in May cut that worst fall to about thirty-five and a half percent. That's real, and it matters. But it didn't come close to making up for the summers you missed.

So what does work? Three things from the same data. One: stay in. Ten thousand dollars that just sat in the S&P 500 all year became about two hundred eleven thousand. No trades, no timing, summer included. Two: if what you really want is a smaller fall in a crash, compare sell in May with a plain mix. We also tested keeping half in stocks and half in T-bills, all year. Its worst fall was about thirty percent, and ten thousand dollars became about seventy-two thousand. Sell in May beat that half-and-half mix in two hundred twenty-one of the two hundred fifty ten-year stretches. So if you were only going to own stocks half the time anyway, winter was the better half to own. But both ended up far behind just staying in. Three: watch the taxes. Our test is the best case, like trading inside an IRA. In a regular account, every May you'd be selling shares you've held about six months. The IRS says gains on anything held one year or less are short-term, and short-term gains are taxed as ordinary income. That would make selling in May look even worse.

So, "Sell in May and go away." Verdict: PARTLY TRUE. The saying is right about the calendar. Summer really was the weaker half. But as advice, it cost you almost half your money. It usually worked if your ten years started between 1996 and 2005, because sitting out dodged the summer crashes of 2001, 2002 and 2008. For every ten-year stretch that started after October 2005, it lost. So next May, when you hear it again, remember: summer was the weaker half. But it still paid.

Every week I send a free email with the full numbers behind each test, the tables, the charts, and what we're testing next. Free weekly email at 10Ktest.com. And tell me in the comments what money advice you want tested next. This video is for education and entertainment, not personal financial advice. Your situation is different; talk to a professional before changing how or when you invest. Disclosure: I don't own an S&P 500 index fund. I do own some individual stocks, including Apple, which is part of the S&P 500. Nobody paid for this video.