Is September Really the Worst Month of the Year for the Stock Market?

Every year, as August ends, headlines repeat that September is the worst month for stocks. The claim has a basis in the numbers. The question is how strong it is, and whether you can do anything useful with it.

What people on X are asking

In September 2026 a wealth manager asked why September is historically the worst month for the stock market, noting that there is no single explanation for the "September effect". In August 2025 one investor joked back at the headline: "So you're saying September is the best month of the year to buy stocks?" And in September 2025 a trader pointed out that September is historically the worst month, but since 1995 there have been 14 green Septembers in the S&P. Here is what a long data series shows.

The fact-checked answer

The data. We used FRED's daily Nasdaq Composite closes, which start in February 1971, and computed each month's return from the last close of the previous month. That gives 55–56 years for each calendar month, through October 2026. FRED's S&P 500 series only covers the last ten years, so we use it as a check, not the main test. We could not check the 14-green-Septembers-since-1995 figure, because our S&P data start in 2016.

Nasdaq Composite return by calendar month, 1971–2026

UXS calculation from FRED NASDAQCOM month-end closes, price only (no dividends). November 2026 is not complete and is excluded. Tap a column header to sort.

Month	Years	Average	Median	Up years
January	55	+2.5%	+2.9%	37
February	55	+0.4%	+0.1%	28
March	56	+0.6%	+1.3%	35
April	56	+1.6%	+1.7%	37
May	56	+1.4%	+2.9%	35
June	56	+1.0%	+1.1%	32
July	56	+0.9%	+1.1%	32
August	56	+0.4%	+1.6%	32
September	56	−0.7%	+0.4%	30
October	56	+0.7%	+1.9%	30
November	55	+2.1%	+2.4%	39
December	55	+1.5%	+0.6%	33

What it shows.

• September is the only month with a negative average, and the worst ranked by average. That part of the claim holds.
• Its median is positive. The typical September is roughly flat to slightly up; the average is negative because the bad ones were very bad.
• It has one of the lowest counts of up years (30 of 56, the same as October; only February, at 28 of 55, had fewer), so it still rose more often than it fell.
• The worst Septembers are clustered: 2001 (−17.0%, including the market closure after September 11), 2000 (−12.7%), 2008 (−11.6%), 2002 (−10.9%) and 1974 (−10.7%). Remove the 2000–2002 bear market and 2008 and much of the effect disappears.
• It depends on the period. From 1971 to 1999, September averaged +0.1% and rose 17 times in 29 years. From 2000 to 2025 it averaged −1.7% and rose 12 times in 26.

UXS chart

Nasdaq Composite since 1971 (log scale)

Nasdaq Composite: latest

25,253

November 9, 2026

Highest in chart period

27,600

October 6, 2026

Lowest in chart period

54.87

October 3, 1974

Change over 1 year

+9.8%

vs November 7, 2025

Source: FRED, Federal Reserve Bank of St. Louis (FRED NASDAQCOM). Retrieved November 10, 2026. Chart computed and drawn by UXS from the raw series. Daily closes from FRED. Log scale, so the 2000–2002 and 2008 falls that dominate the September average are visible in proportion. Download the data (CSV)

Year by year: Nasdaq Composite

Computed by UXS from the same data as the chart. Tap a column header to sort. * = year to date (latest data November 9, 2026).

Year	Nasdaq Composite: last value	Change vs prior year	Year high	Year low
2026*	25,253	+8.7%	27,600	20,795
2025	23,242	+20.4%	23,958	15,268
2024	19,311	+28.6%	20,174	14,510
2023	15,011	+43.4%	15,099	10,305
2022	10,466	-33.1%	15,833	10,213
2021	15,645	+21.4%	16,057	12,609

Recent Septembers. On the Nasdaq Composite: 2020 −5.2%, 2021 −5.3%, 2022 −10.5%, 2023 −5.8%, then 2024 +2.7%, 2025 +5.6% and 2026 +1.9%. On the S&P 500 over 2017–2026, September averaged −1.4% and rose in 5 of 10 years, while November averaged +4.2%. So the run of four bad Septembers in a row helped keep the claim popular, but the last three were positive.

Why it might happen. Explanations offered include fund managers selling losers before fiscal year-ends, investors returning from summer and repositioning, tax-loss selling and less buying from retirement and bonus flows. None is proven, and if the effect were large and dependable, traders would buy in late September and sell in August until it vanished.

Is it worth acting on? The September spread is small against normal monthly swings: the standard deviation of monthly Nasdaq returns is about 6 percentage points, the same in September as in other months. Selling every August and buying back in October would have meant paying trading costs and taxes, and missing Septembers like 1998 (+13.0%) and 2010 (+12.0%). For long-term investors, a steady schedule and an asset mix you can hold through bad months matter far more; see lump sum vs dollar-cost averaging and how often the Nasdaq falls 10%. Seasonal stories also exist for December (the Santa Claus rally) and for bitcoin (best and worst months for bitcoin).

Monthly effects are small next to the S&P 500's average annual return.

Another calendar claim is the election cycle; see whether election years are better for stocks.

Decision helper

Common mistakes

• Treating an average as the typical outcome.
• Forgetting that a few crashes drive the result.
• Quoting one index or period as if it were universal.
• Ignoring trading costs and taxes when timing the calendar.

FAQ

Is September the worst month for stocks?

On the Nasdaq Composite since 1971 it has the lowest average return (−0.7%), but the median September is slightly positive.

How often does the stock market go up in September?

The Nasdaq Composite rose in 30 of 56 Septembers.

What was the worst September?

2001, when the Nasdaq Composite fell 17.0%.

Was September bad in 2026?

No; the Nasdaq Composite rose 1.9% and the S&P 500 fell 0.5%.

What is the best month for stocks?

On average, January (+2.5%) and November (+2.1%) on the Nasdaq Composite.

Should I sell before September?

The data don't support it once costs, taxes and missed gains are counted.