Midterm years are weak. The year after is the trade.
Since 1950 the midterm year is the weakest of the four and the year after is the strongest. In 2026 the swing factor is the 10-year yield, not the polls.
- $SPX
- $SPY
Presidential cycle · 1950–2025 average
- Year 1 +8.4%
- Midterm +4.5% 2026 · now
- Year after +16.1% the window
- Election +9.2%
Is a midterm year a great year to own US stocks? The pattern people remember is real, but it is attached to the wrong year. The midterm year itself is the weakest of the four-year presidential cycle. The strong stretch is the 3 to 12 months after the vote.
I’m about 85% confident in the pattern itself. I’m only 55–60% confident 2026 repeats the historical average, because this year the stock market already ran ahead and the 10-year yield just closed at its highest level since 2002.
A. What the history actually says
I rebuilt the cycle from raw S&P 500 prices rather than quoting a table: Shiller’s monthly series up to 2016, FRED daily closes after that, 76 calendar years from 1950 to 2025. Price return only.
The midterm year is the weakest of the four; the year right after it is the strongest, up in 17 of 19.
The midterm year averages +4.5% with a median of just +2.4%, and it finished higher in only 10 of 19 cases. The year after it averages +16.1% and was up in 17 of 19. Published studies land in the same place; Fidelity, for example, counts the S&P 500 higher in the 12 months after a midterm about 95% of the time since 1938.
The shape inside the midterm year is the more useful picture:
A typical midterm year goes nowhere for three quarters and does its work after October. 2026 did not wait: it was already up about 12% before the vote.
A typical midterm year goes sideways for three quarters and does most of its work from October on. In my data the fourth quarter of a midterm year averages +4.6%, against +2.7% for all years. 2026 broke that template: the index was already up about 12% before October.
B. After the vote
Measured from the election month:
Average +14.9% after 12 months, up 18 of 19 times. The 6-month window has never been negative.
The one 12-month miss is 1986, and it is flat rather than down: the window ends right after the October 1987 crash. The 6-month window has never been negative in 19 tries. That is the part of the pattern worth trading.
C. Why it tends to work
The usual explanation is not that markets prefer one party. It is that uncertainty ends.
Political uncertainty rises → companies and investors defer decisions → the risk premium widens → the vote settles it → the policy path gets clearer → the risk premium compresses → money goes back into equities.
There is also a second-order effect. Midterms often produce a divided government, which lowers the odds of a sudden tax, regulation or fiscal shift. Markets tend to like a government that can’t do anything big, because earnings forecasts get steadier.
D. What is different in 2026
Three facts change the setup:
- The rally came early. The S&P 500 closed at 7,722.72 on October 2, up 12.8% for the year and about 1% below its August 13 record close.
- Rates are moving the wrong way. The Fed raised rates on September 16 to 3.75–4.00%, its first hike since 2023. September payrolls came in at just 29,000, well under forecasts, so the economy is cooling while policy tightens.
- The 10-year is at a 24-year high.
From 3.97% in February to 5.29% on September 30, the highest close since May 2002. September alone added 54 basis points.
So the variable that matters this year is not “midterm, therefore stocks go up”. It is “the election uncertainty clears, and does the 10-year top out?” If the 10-year sits at 5.2–5.5% or above, a high risk-free rate can easily cancel the seasonal tailwind. That is the part I think the market is underweighting.
Two reminders from recent history that the vote is not the low by itself: in 2018 stocks kept falling into December after the midterm, and in 2022 the year’s closing low came in October, before the election. The pattern supports building a 12-month position. It does not tell you which day to buy.
E. When to buy this year
This is a timing sleeve, not a core holding, and the shares below are shares of that sleeve.
- Oct 5 – 13 25%A small first clip Only if you are flat and worried about missing it. Already long? No need to add for the calendar.
- Oct 14 – 31 · primary 50%The main build, after CPI After CPI the long yield stops making new highs; the index pulls back and holds, or breaks out and the retest holds; guidance does not deteriorate.
- Early Nov · confirm 25%Top up after the Fed and the vote Once the FOMC and the election are digested, the rally broadens beyond a few megacaps and bonds are not pushing back.
- Oct 7FOMC minutes (Sep)
- Oct 14 · 8:30September CPI
- Oct 27–28FOMC meeting
- Nov 3Midterm elections
I would rather buy while some uncertainty is still on the table but bad news has stopped pushing the index lower, than wait for every question to be answered. The cost is buying early, which is why it is staged.
The ideal setup is a 3–5% pullback from the recent high while the long yield stabilises and earnings estimates hold. That is a discount I’m willing to wait for, not a forecast or a backtested threshold. If conditions keep improving without a dip, there is no need to wait for a price that may never come.
One correction to a common shortcut: stocks falling while yields fall is not automatically a buy. If both are pricing a recession and weaker earnings, that calls for more caution, not less.
F. Three paths from here
| Scenario | Odds | What happens | What it means for the entry |
|---|---|---|---|
| Base: chop, then stabilise | 55% | 10Y steadies or eases, the vote clears; S&P about +6% to +10% over 6 months | Build through late October |
| Bull: no real dip | 25% | Inflation cools, 10Y below ~4.8%, earnings hold; +12% to +18% | Small first clip, add on a breakout that holds its retest |
| Bear: rates bite | 20% | 10Y above 5.5%, oil and inflation re-accelerate, the Fed keeps hiking; −10% to −15% first | Delay the build; November alone is not a reason |
These odds are my judgement for the next few weeks, not historical frequencies.
Risk and invalidation
The call is wrong if three things happen together: the 10-year holds above 5.5%, core inflation re-accelerates, and earnings estimates get cut month after month. If all three show up, I drop the seasonal trade and don’t average down.
Sizing for the sleeve: at most 20% of the account, with a risk exit 5% below the sleeve’s average cost, about a 1% planned loss on the account. Gaps and slippage can make it worse. Count any existing long exposure toward the same risk, and no leverage.
Calendar for the next weeks
- Oct 7: minutes from the September FOMC, for how the committee talks about further hikes.
- Oct 14, 8:30 ET: September CPI. This is the trigger for the main build; watch how bonds and stocks hold the move over the session, not the first minute.
- Oct 27–28: FOMC meeting.
- Nov 3: midterm elections.
The one-line version: the midterm year isn’t the good year. The run-up to the vote is hard to trade, and the 6 to 12 months after it have an unusually high hit rate. In 2026, let the 10-year tell you when.
Opinion, not investment advice. Data: Robert Shiller’s S&P 500 dataset, FRED (SP500, DGS10), BLS, Federal Reserve.
— RELATED READING