Stock Market Statistics by Month: Best and Worst Times for Returns

Many investors have heard that September is a weak month for stocks. Others follow the old rule to “sell in May and go away.” Some of these sayings hold up in the data, and some do not.

This article looks at stock market statistics by month and by time period. It draws on six published sources, and each figure keeps its own time frame. Those time frames differ, so the numbers should not be mixed together.

One point appears in every source. The US stock market has gone up more often than it has gone down over almost any long stretch.

How Often the Stock Market Goes Up

Stocks do not rise every day, but they rise more often than they fall. Investment writer Ben Carlson shared win rates for the US stock market going back to 1926.

On a daily basis, stocks have risen 56% of the time. On a monthly basis, the figure is 63%. On a yearly basis, it is 75%.

The share grows as the time frame gets longer. Over five-year periods, stocks have gained 88% of the time. Over 10-year periods, the figure is 95%.

Over every 20-year period since 1926, returns have been positive. That is a win rate of 100%. Carlson is clear that no one can promise these rates will hold in the future.

These numbers show why short-term moves can feel random. A single day is close to a coin flip, with a small lean toward gains. A single year is more likely to be up than down, but losing years still happen about one time in four.

 

Carlson also describes betting on a crash as a low-probability bet. A crash story sounds smart, but it is hard to time. The market has risen most of the time.

Bear Markets and Crashes Since World War II

A bear market is usually defined as a drop of 20% or more from a recent high. Carlson counted 13 of them since World War II, including the one under way when he wrote in May 2023.

That works out to about one bear market every six years. Falls of 30% or worse have happened four times, or about once every 13 years.

Falls of 50% or worse have happened three times. That is about once every 26 years. These events are rare, even though they get a lot of attention.

The story of investor Carl Icahn shows the cost of betting on a crash. The Financial Times reported that Icahn began positioning part of his portfolio for a large fall in 2017. Those bets cost him nearly $9 billion over six years.

At times, the value of the securities he bet against passed $15 billion, according to regulatory filings. Icahn told the paper that no one can pick the market over the short or medium term. He added that he made a mistake by not following his own advice.

Icahn blamed part of the loss on the Federal Reserve, which put trillions of dollars into markets during the Covid pandemic. Carlson’s point is simple. Crashes are real, but nobody can say when the next one will start.

Best and Worst Months From 1926 to 2025

Morningstar’s Dan Lefkovitz and Michael Wang studied more than 100 years of monthly US stock returns. They used data from the Center for Research in Security Pricing (CRSP), with all stocks weighted by company size.

September is the weakest month. It is the only month with a negative average return over the full period. Its median return is positive, but it is still the lowest of the 12 months.

A few very bad Septembers pull the average down. There were double-digit losses in the 1930s and in 1974. Other hard Septembers came in 2001, 2002, 2008, 2011 and 2022.

The causes differ by year. They include the 9/11 attacks, accounting scandals, the Lehman Brothers collapse, a US credit rating downgrade, and inflation with Fed rate hikes. The Morningstar US Total Market Index also fell 0.76% in September 2026.

October is the second-weakest month. Big crashes took place in October 1929, 1987 and 2008. The month has also marked market bottoms.

November is the best month on both average and median returns. December and July are also strong. Theories for the year-end strength include holiday spending, buying by professional investors, post-election rallies and tax-loss selling.

Batting average by month

A weak month is not always a losing month. Of the past 101 Septembers, 52 were positive, or about 51%. September 2024 and September 2025 were both up.

December has the best record, with 21 negative Decembers in 100 years, or about 79% positive. Even so, December lost money in 2022 and 2024. The market has risen more often than it has fallen in every calendar month.

Across all months, the market has produced an average annual return of roughly 10%. That helps explain why even weak months tend to finish up more often than down.

Sell in May and go away

The saying is said to date to 17th-century London. It was tied to summer holidays more than to returns. Wang tested it over 100 years and found that November through April beat May through October on both average and median returns.

May through October was not negative, though. An investor who skipped it would have given up an average six-month return of nearly 4%. July has made up a large part of that return.

In the 2020s so far, May through October has beaten November through April. Morningstar links that to rough first quarters, including the pandemic panic of 2020, the inflation shock of 2022 and tariff turmoil in 2025.

Lefkovitz cautions against reading too much into monthly patterns. The best month in 100 years was April 1933, when President Franklin D. Roosevelt suspended the gold standard. Within a few months, the market was losing again.

Seasonal Patterns in Three Major Indexes

Cory Mitchell, CMT, looked at monthly patterns in three indexes. The NYSE Composite covers all stocks listed on the New York Stock Exchange. The S&P 500 holds the largest US companies, and the Nasdaq 100 holds large companies, with about 65% in technology or telecommunications.

Seasonality is an average of past behavior. Mitchell calls it a guide, not a forecast. His 25-year summary, covering 2001 to 2025, lists September as the first of the worst months for all three indexes.

November is the first of the best three months for all three indexes. April and July also rank high for the NYSE Composite and the S&P 500. The Nasdaq 100 list includes October and July.

The last 10 years tell a slightly different story. For the S&P 500, April through August and November have been even stronger. January improved, and December weakened.

The Nasdaq 100 shows a similar shift. May to August and November are stronger, and a few weak months are weaker. Mitchell lists February, September and December as its weak months over 10 years.

Long-term returns put these months in context. The S&P 500 has returned about 10.5% per year over 100 years. The Russell 2000 has returned close to 9% per year.

For the Nasdaq 100, Mitchell gives two figures. One is 15% per year for the QQQ fund over 20 years, and the other is 16.6% per year over the same span. Both are listed in the article.

Mitchell also warns about averages. If April rose 80% of the time, it went up in 16 of the last 20 years, but it may not rise this year. A few large moves can also skew an average.

The January Barometer and October

The January Barometer says that how January goes, so goes the rest of the year. Mitchell reports S&P 500 data going back to 1950.

When January was an up month, the average return for that year was about 17%. When January was down, the average yearly return was -1.7%. Mitchell adds, “Take from it what you will.”

January 2026 ended with the S&P 500 up about 1.5% and the Nasdaq 100 slightly positive. January 2025 was also up, and the SPY fund gained 16.96% that year.

A rule built on one month has limits. A gap in averages does not mean each year follows the pattern. Mitchell says most seasonal patterns lack enough data to be proven, and he calls them ideas with some evidence.

October has a reputation for crashes, but Mitchell’s recent data is milder. Over the last 20 Octobers, the S&P 500 averaged a gain of 1.1%. The Nasdaq 100 averaged 1.7%.

The Morningstar data lists October as the second-weakest month over 100 years. The two views differ because the time frames differ. It helps to check the period before drawing a conclusion.

Holiday and Weekday Patterns

Some patterns last days rather than months. Mitchell describes a pre-holiday pattern. A trader buys one or two days before a long weekend and sells one or two days after.

Trading volume tends to be lower before long weekends, which may help prices drift up. Testing by QuantifiedStrategies found that most holidays do not produce a big jump. July 4th, Thanksgiving and Christmas were more reliable.

The same firm found that buying at the close the day after a holiday and selling at the next close showed a steady rise. Mitchell describes the Santa Claus rally as the best documented pattern. It means holding for the last four or five days of the year and selling two or three days into the new year.

The average gain has been about 1.1% per trade in an index such as the S&P 500. The exact number of days can change because of weekends and market closures.

The weekend effect

The weekend effect is the tendency for stock returns to be lower on Mondays than on the Friday before. Frank Cross wrote about it in the Financial Analysts Journal in 1973. He found that average Friday returns beat average Monday returns.

A Federal Reserve study found that weekend returns were often negative before 1987. The negative return disappeared from 1987 to 1998. Weekend volatility has risen again since 1998, and the cause is still debated.

Several reasons have been offered. Companies may release bad news after the market closes on Friday. Individual investors may sell more on Mondays, and short selling may play a part.

Other research points to a reverse effect, where Monday returns are higher than those on other days. Some studies find that small companies do worse on Mondays while large companies do better. One view is that the reverse effect happens only in US markets.

The Four Stages of the Market Cycle

Months and weekdays are one way to view the market. Another is the cycle that stocks move through over years. One source describes four stages: accumulation, markup, distribution and markdown.

Accumulation is the first stage. Prices move sideways in a range, sometimes for years. Large investors buy in small steps so they do not push the price up, and few individual investors take part.

Markup follows when the price breaks above the top of the range. Volume spikes as new buyers arrive. Higher highs and higher lows appear, and the trend can turn parabolic, meaning prices rise fast and at a growing pace.

Distribution is the topping stage. Volume rises but price does not. Early buyers leave, and chart patterns such as a head and shoulders top or a double top can appear.

Markdown is the final stage. Late buyers are underwater and start to sell, and there are few new buyers to take the other side. Selling can feed on itself and send prices down fast.

The stage often ends when a key support level breaks and volume jumps to many times the daily average. At that point, most of the selling is done and the stock can return to accumulation.

The source says the cycle is a pattern, not a rule. Product news and world events also move stocks, so movements can be hard to predict. It suggests studying weekly charts to see how the stages looked in real cases.

Timing the Market: Best Day vs Worst Day

Capital Group compared two hypothetical investors in the S&P 500. The period was 20 years ending December 31, 2025. Each investor put in $10,000 per year, for $200,000 in total.

The first investor picked the best day each year, which was the market low. The average annual return was 14.65%, and the account ended at $1,074,826.

The second investor picked the worst day each year, which was the market high. The average annual return was 12.64%, and the account ended at $801,554.

The gap is about $273,000, so timing did matter. Even so, the investor with the worst timing ended with about four times the money put in.

Both accounts dipped in 2008, 2018 and 2022. In 2006 and 2008, both worst-day totals sat below the amount invested. By 2025, both were far ahead.

Capital Group notes that regular investing does not ensure a profit or protect against a loss. The figures are hypothetical, and no one can pick the exact low each year. The lesson it draws is to focus on time invested, not timing.

How Holding Periods Change the Odds

Capital Group also counted positive and negative periods for the S&P 500 over the past 98 years, ending December 31, 2025. For one-year periods, 72 were positive (74%) and 26 were negative (26%).

For three-year periods, 81 were positive (84%) and 15 were negative (16%). For five-year periods, 83 were positive (88%) and 11 were negative (12%).

The 10-year count covers 89 years. Of those periods, 84 were positive (94%) and five were negative (6%).

These results are close to Carlson’s figures of 75%, 88% and 95% for one, five and 10 years. The small gaps come from different data sets and time spans. Both sources show the same shape, with the odds improving as the holding period grows.

Morningstar reaches a similar view. Colleague Christine Benz told investors, “Get out of the timing business.” Morningstar’s Jeff Ptak has also measured the cost of missing the 10 best trading days of a year.

Lefkovitz closed with a line from Charlie Munger: “The first rule of compounding: Never interrupt it unnecessarily.” Taxes and trading costs also add to the price of guessing wrong.

What the Numbers Say

Stock market statistics by month show real patterns, but they are weaker than the sayings suggest. September has been the weakest month since 1926, yet about half of all Septembers were up. November and December have the best records, yet December still lost money in 2022 and 2024.

Seasonal results change with the time frame and the index. The 10-year and 20-year lists in Mitchell’s data do not match, and the 2020s have favored May through October so far. Averages can also hide a few large moves.

The longer-term numbers are steadier. Stocks rose in 63% of months, 75% of years and every 20-year period since 1926 in Carlson’s data. Even the worst-timed yearly investor in Capital Group’s example ended with about four times the money invested.

None of these figures predict what comes next. They show what has happened, and they point to time in the market as the factor that has mattered most.

---

Sources

1. The Stock Market Usually Goes Up (But Sometimes it Goes Down), Ben Carlson, May 21, 2023
2. The 10 Best and 10 Worst Months of All Time for the US Stock Market, Dan Lefkovitz, Morningstar, October 13, 2026
3. Time, not timing, is what matters, Capital Group
4. Best and Worst Months for the Stock Market – Seasonal Patterns, Cory Mitchell, CMT, September 29, 2026
5. Understanding the Weekend Effect in Stock Markets, Will Kenton, Investopedia, updated July 21, 2026
6. The Four Stages of the Stock Market Cycle, August 29, 2025