Skip to main content

Presidential / Election Cycle

Updated Jun 24, 2026 at 2:35pm

Research Draft Medium 1,142 words

The Presidential (or Four-Year) Election Cycle Theory holds that U.S. stock returns follow a roughly four-year rhythm tied to the U.S. presidential term: weak in the first two years (post-election and midterm) and stronger in the back half, with the third year (pre-election year) historically the best of the four. The proposed mechanism is political incentive — an incumbent administration front-loads painful or austere policy early in the term and stimulates the economy as re-election approaches, leaving markets to climb into the campaign. Its core tension is that the pattern is one of the most-cited seasonal anomalies in markets, yet the causal claim rests on an extremely small sample (only one cycle every four years) and is regarded by much of academic finance as fragile or spurious.

How it's formed (the mechanics)

The cycle is purely calendar-driven. The four years of a presidential term are labelled:

1. Year 1 — Post-election year (inauguration year) 2. Year 2 — Midterm year 3. Year 3 — Pre-election year 4. Year 4 — Election year

The theory was popularised by Yale Hirsch, founder of the Stock Trader's Almanac (the cycle framing dates to roughly 1967). The behavioural story: in Years 1–2 a new administration tackles unpopular agenda items (tax changes, spending restraint, monetary tightening); by Years 3–4, with re-election in view, policy turns stimulative, boosting growth, confidence, and equities. There is no formula — the "indicator" is simply which year of the term it is, and the proposition is that average returns differ systematically by that label.

How it's used in practice

Practitioners treat the cycle as a low-frequency seasonal tilt, not a trade trigger. Typical uses:

  • A backdrop overlay: being modestly more constructive in pre-election (Year 3) and election (Year 4) years and more cautious in the midterm (Year 2) year, while letting price, valuation, and macro do the heavy lifting.
  • Combining it with other calendar effects (the Best Six Months / "Sell in May," midterm-year-bottom seasonality) — the Almanac famously highlights the stretch from the Year-2 midterm October low into the Year-3 high as one of the strongest seasonal windows on record.
  • Setting expectations and managing client psychology around election noise rather than sizing positions on the cycle alone.

Because the signal updates only annually and the sample is tiny, no serious user runs a standalone "buy Year 3, sell Year 2" system with leverage.

Adoption, debate & evidence

The descriptive pattern is real in the historical record. Multiple independent compilations agree the third year has been the strongest. Charles Schwab's analysis (researcher Lee Bohl) found Year 3 historically produced the largest average gains; commonly cited Stock Trader's Almanac figures show the Dow and S&P 500 up roughly 15% on average in the third year since 1943. Schwab-cited data found the S&P 500 rose in about 90% of third years (1933–2023), well above the unconditional positive-year rate of roughly two-thirds. Academic work by Beyer, Jensen & Johnson (Journal of Portfolio Management) documents a robust term-cycle pattern — returns concentrated in the back half, sharpest in Year 3 — and argues it survives controls for business conditions, attributing much of it to more accommodative Fed policy in Year 3. These numbers vary widely by source and start date (e.g. one Beyer/Jensen/Johnson window cites a ~21% Year-3 S&P average over 1957–2012), so treat them as "commonly cited" rather than precise constants.

But the causal/forward claim is contested, and two distinct phenomena get conflated. The honest "folklore vs measured" verdict:

  • Tiny sample. Since 1928 there have been only ~24 complete four-year cycles. The third-year strength is driven heavily by a handful of observations, and dropping outlier years (e.g. crash-affected periods) materially weakens it.
  • Recent breakdowns. The "weak first year" leg has repeatedly failed in the 21st century — markets rose in the first years under Bush, Obama, and Trump — so the pattern's reliability as a forward predictor is far lower than its historical averages suggest.
  • The four-year term cycle vs. the party premium are two different things — and the famous "spurious" critique targets the party premium, not the term cycle. The "are returns higher under Democrats or Republicans?" debate (Santa-Clara & Valkanov; Pástor & Veronesi's Political Cycles and Stock Returns, JPE 2020) is a distinct phenomenon from the four-year term calendar. It is this party premium that Arnott, Cornell & Kalesnik (Research Affiliates, 2017) and the Cocquemas & Whaley (2016) line argue is largely spurious: the party regressor is highly persistent, and tested across five international markets (Australia, Canada, France, Germany, UK) the party result does not replicate. Pástor & Veronesi instead defend a Democrat-minus-Republican gap via time-varying risk aversion (high risk aversion → elect Democrats → higher subsequent returns). Do not let either phenomenon borrow the other's credibility, and do not cite the party-premium international-replication failure as if it disproved the term cycle.

Bottom line: a robust descriptive regularity for the term cycle with a plausible (largely monetary-policy) story, but a tiny sample, weak recent out-of-sample performance, and averages that mask huge dispersion — useful context, not a dependable forward edge. The separate party premium is the one most strongly flagged as spurious.

Strengths & limitations

When it's useful: as a soft contextual prior over multi-quarter horizons, especially fused with other seasonality and with macro/monetary conditions (which arguably do the real work — the cycle may be a proxy for the Fed and fiscal policy timing rather than an independent force).

When it fails: as a standalone timing rule, in any single cycle, and whenever a dominant macro shock (recession, financial crisis, pandemic, rate shock) overrides the calendar. The pattern's averages mask wide dispersion — knowing it's "Year 3" tells you very little about this Year 3.

The #1 misuse: treating averages from ~24 cycles as a dependable forecast and sizing risk on it, or conflating the term cycle with the party-in-power return gap. Small-sample averages are not probabilities.

Sources

Disputes flagged: the descriptive third-year strength of the four-year term cycle is broadly agreed and has a defensible monetary-policy mechanism (Beyer/Jensen/Johnson), though its small sample and weak recent forward performance limit it. The separate party-in-power return gap is the phenomenon most strongly challenged as spurious (Arnott et al.; Cocquemas & Whaley), and the failed five-country international replication applies to the party effect, not the term cycle — the two must not be merged.