The Presidential Cycle: Does the Election Calendar Move Markets?

Among the more debated seasonal patterns is the idea that the 4-year US presidential election cycle correlates with stock market performance — with some years in the cycle historically tending to perform better or worse than others.

The four years of the cycle

The cycle is typically broken into four roles, based on the US presidential term:

• Election year (years divisible by 4, e.g. 2024, 2028) — the year a president is elected.
• Post-election year (the year after) — the first full year of a new or continuing administration.
• Midterm year — the year of Congressional midterm elections, roughly the cycle's midpoint.
• Pre-election year — the year before the next presidential election.

Why might this happen?

The theories here lean more political-economic than calendar-driven:

• Policy timing. Administrations may have incentives to front-load difficult or unpopular economic policies (e.g. spending cuts, rate-sensitive decisions) earlier in a term, and shift toward growth-friendly policies as the next election approaches.
• Reduced uncertainty after elections resolve. Markets sometimes dislike uncertainty more than any specific outcome — once an election is decided, some of that uncertainty discount can unwind.
• Pre-election stimulus incentives. Incumbent administrations may have political incentives to support economic conditions heading into an election year.

As with all such explanations, these are plausible narratives fitted to historical patterns after the fact — not proven causal mechanisms. The actual political and economic environment differs enormously from cycle to cycle.

The honest statistical problem: small samples

This is the most important thing to understand about presidential-cycle analysis: a 4-year cycle means you only get one "pre-election year" every 4 years. Over a 24-year lookback — already a long window — that's only 6 data points for any single phase of the cycle.

Six observations is a very small sample. A couple of unusual years (a financial crisis, a pandemic, an unprecedented policy response) can swing the average for an entire phase of the cycle significantly. This doesn't mean the pattern is meaningless — but it does mean the statistical confidence behind any presidential-cycle claim is inherently much lower than, say, a monthly seasonal pattern with 15 years of data (15 observations per month).

How to use cycle filtering responsibly

Presidential-cycle analysis is best treated as a curiosity to be aware of — "is this an election year, and does that historically correlate with anything for this asset?" — rather than a basis for a standalone strategy. With only a handful of observations per phase, it's one of the seasonal techniques most prone to being overfit to a small number of historical events.

It can be more interesting when combined with monthly seasonality — for example, checking whether a specific month's pattern looks different in election years versus other years — though again, the sample sizes involved mean any difference should be treated as a loose observation, not a rule.