Stocks Rose in the Year After Every Midterm Since 1950. What the Streak Leaves Out

Research note: S&P 500 daily closes through October 2, 2026. Frozen October 3, 2026. How we checked it · How we use AI.

• The streak is real. 19 of 19 since 1950, and the average is more than twice that of other years.
• The start date flatters it. Counting 1930 to 1946, it is 22 of 24, with a −38.4% loss after 1930.
• Where the market starts matters. Years beginning within 5% of a high rose by a median of 12.4%; years beginning after a deeper fall, 17.5%.
• The path was rarely smooth. 11 of the 19 windows since 1950 included a decline of 10% or more.

The midterm pattern is one of the most quoted statistics in markets, and for good reason. Few calendar effects hold up this well, and this one has a plausible story behind it. Since 1950 the midterm calendar year has been the weakest of a presidential term, up 4.6% on average, and the year before a presidential election the strongest, up 17.2% and higher in 17 of 19. The October-to-October window runs from the end of the first into most of the second.

I checked it on the S&P 500’s daily record back to 1927, and the familiar figure holds: from October 31 to October 31, the average gain after the 19 midterms since 1950 is 16.6%, with no losing year. The question for 2026 is how much of that applies to a market that has not had the kind of midterm-year decline most of those years started from.

Here is where 2026 stands. The S&P 500 is up 12.8% this year and 1.0% below its high. Its worst decline of the year, −9.1% into March 30, 2026, is shallower than the typical midterm year’s −16.0%. The window opens with the close on October 30, 2026, the last trading day before October 31 and four days before the November 3, 2026 election.

Every midterm since 1930

The chart below shows the October-to-October return after each midterm, with the worst decline from a high inside that year beside it. The years before 1950 are faded, because the usual version of the statistic leaves them out; the dots mark years that started within 5% of a high, as 2026 will.

The two losing years were 1930 (−38.4%) and 1938 (−2.6%). Neither is an obscure case: 1930 was the second year of the Depression decline. A rule that starts in 1950 drops them by construction, which is why I would describe the record as 22 of 24, not 19 of 19.

The average path

Laid over each other, the 19 years since 1950 show when the gain arrived. On average the S&P 500 was +7.5% three months after October 31 and +15.0% after six, against +4.4% at six months in other years. The second half added only +1.8% more. 4 of the 19 years were below their starting level at three months; none was at six or twelve.

What the average depends on

The same streak looks different depending on how you slice it. Two things move it most: whether the pre-1950 years are included, and how far the market had fallen by October 31.

Two of the three best years since 1950, 1962 (+30.9%, from 22% below its high) and 1990 (+29.1%, from 18% below its high), started after a deep fall; 1954 (+33.6%) started near a high. The three weakest, 2014 (+3.0%), 1986 (+3.2%), 2010 (+5.9%), all began within 5% of a high. The near-high years still rose every time, so the difference is the size of the gain, not its direction.

The start date matters for the same reason. Measured from election day, the first Tuesday in November, the average falls to 15.1%, still positive in 19 of 19, because the window then starts a few days later, and the market rose between October 31 and election day in 15 of the 19 years. Measured from each midterm year’s lowest close, the average one-year gain is 30.1%. Much of the effect is a rebound from the midterm-year low.

Where I come out

I would lean positive on the year from October 30, 2026: 22 of 24 since 1930 is a strong base rate, and none of the near-high years since 1950 fell. But I would not plan around the headline 16.6%. 2026 has not had the deep midterm-year decline that the biggest years rebounded from, and the near-high group’s median of 12.4% is the better guide.

I would also expect a meaningful decline along the way. 11 of the 19 windows since 1950 had one of 10% or more, and the median worst drop was −10.3%. The pattern has said where the market ended the year, not how calmly it got there.

What would change my read

Toward the headline: a sharp decline before October 31. A market that falls 10% or more into the window would start from the kind of low the strongest years began from, which would make a larger gain more likely, not less.

Toward a weaker year: valuation and narrow participation. The market enters the window near the top of its valuation history and with a minority of stocks above their 50-day average; if both persist into November, I would lean toward the low end of the near-high years’ range.

How we checked it

• The window. S&P 500 price change from the last close on or before October 31 of each midterm year to the last close on or before October 31 of the next. It is the usual way this statistic is measured.
• Price only. Dividends would raise every figure, by more in the decades when yields were higher; they do not change the comparisons between groups.
• Near a high. The October 31 close within 5% of the highest close in the prior 52 weeks. 9 of the 19 post-1950 years qualify.
• Small sample. 19 midterms since 1950 and 24 since 1930: enough to show a pattern, too few to treat any average as a forecast.
• Frozen evidence. Every number on this page comes from the snapshot of October 3, 2026.