Market timing · Report
Election Juice
Does the incumbent party juice the markets and economy in year 4 of their term, the election year itself? Across 23 terms since Coolidge, the honest answer: the average bump is small, the variance is large, and the data cannot distinguish the sample from noise.
Last updated November 06, 2026
Is there a market bump in the election year?
Folklore says the incumbent party manages policy to juice the economy and markets in the election year itself. The data weighs in across 23 presidential terms from Coolidge (1928) to Biden (2024).
Statistical significance (paired t-test, two-sided):
- SPX juice: p=1.000 — noise (t = +0.47)
- UE juice: p=0.328 — noise (t = +1.39)
- FF juice: p=1.000 — noise (t = +0.19)
The small positive SPX juice (+1.69pp average) sounds supportive — but p = 1.000 means it is indistinguishable from random noise. In 23 elections, year-4 beats the year-1-through-3 baseline just 57% of the time — barely better than a coin flip. The folklore is mostly folklore.
The 23 terms — juice per series
Each row is one presidential term (end-year labeled). The “juice” column shows the year-4 outcome minus the year-1-through-3 average, so positive = better in the election year than the baseline. Color convention: green = pro-incumbent direction, red = anti-incumbent. UE: negative = unemployment fell more in year 4 than the baseline. FF: direction is context-dependent (loose policy helps asset prices, tight policy fights inflation).
The six biggest juice outliers
Where the averages come from: a handful of extreme years dominate the distribution. Two of the top three (2008 GFC, 2000 dot-com) are negative — election-year markets collapsed under external shocks, dragging down the aggregate.
Does the juice depend on the incumbent’s party?
A quick split by party of the sitting president. Across the 23-term sample, neither party shows a reliable election-year SPX edge over its own 1-3 baseline. The variance within each party group is larger than the gap between them.
Methodology
The juice calculation. For each presidential term, compute:
- y1-3 baseline: average of years 1, 2, and 3 outcomes for the series (SPX return, unemployment delta, Fed Funds rate delta).
- y4 outcome: year-4 (election year) value for the same series.
- Juice: y4 minus y1-3 baseline. Positive = election year outperformed the baseline.
Statistical test. Paired two-sided t-test with the null hypothesis that mean juice equals zero. A p-value below 0.05 would reject the null (the juice is real); above 0.10 means the data cannot distinguish the sample mean from random noise.
Sign conventions.
- SPX juice: positive = election year had higher return. Clearly pro-incumbent (markets like incumbent survival).
- UE juice: positive = unemployment rose more than baseline. Anti-incumbent. Negative = unemployment fell = pro-incumbent.
- FF juice: positive = Fed hiked more in election year. Direction is contextual — tight policy helps dollar + fights inflation but hurts equities.
Data sources.
- SPX annual returns:
symbol_data.dbdaily closes, computed start-of-year to end-of-year. - Unemployment rate: FRED
UNRATE, Dec year-1 to Dec year-end delta. - Fed Funds target: FRED
DFEDTARU, same window.
Caveats.
- Sample size (23) is small for a hypothesis test. A real effect of +5pp or more would likely still fail to clear p=0.05 at this n.
- The baseline is averaged across just 3 years, which is noisy. A rolling 5-year or 10-year baseline would tighten the test.
- External shocks (2008 GFC, 2000 dot-com, 1929 crash) land in election years randomly. They inflate variance without signaling policy intent.
Related reports
- October Premium — the companion piece: is election-year Aug-Oct SPX vol elevated? (No.)
- Midterm Seat Loss Ranking — voter-side verdict at the mid-point of each term.
- 2026 Midterm Forecast — three-model ensemble for House seat swing.
- Back to Presidential Performance library