US Midterms and the S&P 500: the Full Table, and Why We Do Not Trade It
Published 11 September 2026 · Last updated 11 September 2026
The claim circulates every two years: the S&P 500 has risen in the twelve months after every US midterm since 1950. It computes out true. Nineteen midterms, nineteen higher closes a year later, median +14.49 percent. This page prints the whole table, including the two thirds of the count that sit under +20 percent, and then explains why a 19 of 19 record is weaker evidence than it sounds.
The next US midterm is Tuesday 3 November 2026. What follows is the computed record, measured from the S&P 500 close on or before each election day to the close on or after the same date one year later. No party, no winner and no policy enters this page. The only variable is the calendar.
The computed record, 1950 to 2022
| Midterm year | S&P 500 at election | 12 months later | 12 month change | Calendar year change |
|---|---|---|---|---|
| 1950 | 19.36 | 22.49 | +16.17% | +21.68% |
| 1954 | 31.79 | 42.35 | +33.22% | +45.02% |
| 1958 | 51.56 | 57.26 | +11.06% | +38.06% |
| 1962 | 58.35 | 72.81 | +24.78% | -11.81% |
| 1966 | 80.73 | 91.14 | +12.89% | -13.09% |
| 1970 | 84.22 | 94.91 | +12.69% | +0.10% |
| 1974 | 75.11 | 89.15 | +18.69% | -29.72% |
| 1978 | 93.85 | 99.87 | +6.41% | +1.06% |
| 1982 | 137.49 | 164.84 | +19.89% | +14.76% |
| 1986 | 246.20 | 248.96 | +1.12% | +14.62% |
| 1990 | 311.62 | 389.97 | +25.14% | -6.56% |
| 1994 | 465.65 | 591.71 | +27.07% | -1.54% |
| 1998 | 1,110.84 | 1,354.93 | +21.97% | +26.67% |
| 2002 | 915.39 | 1,051.81 | +14.90% | -23.37% |
| 2006 | 1,382.84 | 1,475.62 | +6.71% | +13.62% |
| 2010 | 1,193.57 | 1,237.90 | +3.71% | +12.78% |
| 2014 | 2,012.10 | 2,102.31 | +4.48% | +11.39% |
| 2018 | 2,755.45 | 3,076.78 | +11.66% | -6.24% |
| 2022 | 3,828.11 | 4,382.78 | +14.49% | -19.44% |
Positive in 19 of 19. Median +14.49 percent, mean +15.11 percent. The weakest was 1986 at +1.12 percent and the strongest 1954 at +33.22 percent. Computed from daily closing prices of the S&P 500 index.
The part that usually gets left out
The midterm calendar year itself is unremarkable. Across the same nineteen years the index finished the election year higher only 11 times, with a median of +1.06 percent, against a long run average that is considerably better. Four of those years were double digit declines: 1974 at minus 29.72 percent, 2002 at minus 23.37, 2022 at minus 19.44 and 1966 at minus 13.09.
Put the two facts together and the seasonality reads differently. The pattern is less a statement about elections and more a statement about what tends to follow a weak year, which is a claim with a much larger sample and a much more ordinary explanation.
Why Tempora publishes no forward call on this
Three reasons, stated so the omission is on the record alongside the calls we do make.
The sample is nineteen. A run of nineteen consecutive successes on a coin weighted toward heads is unremarkable. US equities rise in roughly three of every four rolling twelve month windows since 1950, so nineteen in a row drawn from that distribution carries far less information than the unbroken run implies.
The windows overlap the base rate they are measured against. Every twelve month window here is drawn from the same upward drifting series. A rule that says buy any date and hold twelve months would also have produced a strong majority of positive outcomes across this period. The midterm filter needs to beat that baseline to mean anything, and on this sample it does not clearly do so.
There is no mechanism we can state. Tempora publishes dated calls with the reasoning attached, and the reasoning here would be the pattern itself. A pattern with no stated mechanism and a sample of nineteen is a description of the past. We keep those on findings pages and off the tracker.
What we do have dated across this window
The eclipse drawdown window running to 10 November 2026 is on the public board with its condition stated in advance, and the December 2026 FOMC call sits just past it. Those carry mechanisms and falsifiers. This page carries a table.
Method
S&P 500 daily closing prices. For each midterm, the reference is the last close on or before election day and the first close on or after the same calendar date one year later. Calendar year change is the last close of the election year against the last close of the previous year. Nineteen midterms from 1950 to 2022 inclusive.
Questions
Has the S&P 500 really risen after every midterm since 1950?
Yes. Measured from the close on or before election day to the close on or after the same date a year later, all 19 midterms from 1950 to 2022 produced a higher index. The median gain was 14.49 percent and the weakest was 1.12 percent in 1986.
What happens in the midterm year itself?
It is much weaker. Only 11 of the 19 election years finished higher, with a median of 1.06 percent. Four were double digit declines, including 1974 at minus 29.72 percent and 2022 at minus 19.44 percent.
Does Tempora have a forward call on the 2026 midterm?
No, and the reasons are on this page. The sample is 19, the twelve month windows are drawn from the same upward drifting series they are measured against, and we can state no mechanism. Calls on our tracker carry a mechanism and a falsifier.
When is the 2026 US midterm?
Tuesday 3 November 2026.
Is this investment advice?
No. It is a computed historical table published with its limitations stated. Not financial advice.
Read next
This page was first published on 11 September 2026 and last updated on 11 September 2026. It reports computed S&P 500 index returns around US midterm elections. No political outcome is discussed or forecast. A muhurat is personal and no birth chart is computed here. Not financial, medical or personal advice.