Stock Market by President: The S&P 500 From Hoover to Trump

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-10-03

The stock market by president means the S&P 500’s price change from the start of one presidential term to the start of the next. This page measures every term from Herbert Hoover’s in 1929 to Donald Trump’s current term, from daily closes, with the dates taken from the official inauguration record. It reports the numbers and what they can and cannot show; it makes no claim that any president caused them.

How it works

A term is a window of calendar dates. Each one starts on the date listed by the Joint Congressional Committee on Inaugural Ceremonies on its past ceremonies page, read on 3 October 2026. That page also lists the three swearing-ins that started a term mid-way: Harry S. Truman on 12 April 1945, after the death of President Franklin D. Roosevelt; Lyndon B. Johnson on 22 November 1963, after the assassination of President John F. Kennedy; and Gerald R. Ford on 9 August 1974, after the resignation of President Richard M. Nixon.

The measurement. Each term runs from the last close before its start date to the last close before the next start date. A new president’s first trading day therefore counts in their own term. Donald Trump’s current term is measured to the close of 2 October 2026 and is unfinished.

The index series. Daily closes come from Yahoo Finance’s S&P 500 series (^GSPC), which starts on 30 December 1927. The S&P 500 itself was launched on 4 March 1957 and has no performance history before that date. Every close before then, which covers the terms from Hoover’s to the start of Eisenhower’s second and the elections from 1928 to 1956, comes from the earlier history carried in the same series, not from the 500-company index. The series has weekday closes only, so the last close before Hoover’s 4 March 1929 start is Friday 1 March. Per-year figures are compound rates over calendar time.

Price only. The index level leaves out dividends, so every figure here understates what a holder who reinvested dividends received. The same rule applies to every term, so the comparison between terms is like for like.

A worked example

One close, two presidents. The S&P 500 closed at 850.12 on Friday 16 January 2009, the last close before the inauguration of 20 January. That close ended George W. Bush’s second term and started Barack Obama’s first.

For the outgoing term it was the end point. Bush’s second term began from a close of 1,184.63, so $10,000 in the index at that start was worth $7,176.25 at the 850.12 close, a fall of 28.24%.

For the incoming term it was the starting point, and it was not the bottom. The index kept falling after the inauguration, to 676.53 on 9 March 2009, 20.42% below the starting close. It then rose, and the term ended at 1,485.98 on 18 January 2013: $10,000 at the start was worth $17,479.65, a gain of 74.80%.

The lesson is about the window, not the person. The same week of prices made one term look bad and the next look good. Every row in the table below carries the same effect, which is why a term’s number says more about the index level on day one than about anything decided in office.

The original data

The sample. 24,806 daily closes from 30 December 1927 to 2 October 2026, and 28 terms from 4 March 1929.

The last row runs to 2 October 2026 and is unfinished. Per-year figures for short terms swing widely: Roosevelt’s fourth term lasted 0.22 years. Every row, with the start and end dates and the deepest fall inside each term, is in the presidential terms file.

Most terms rose, about as often as any four years. The index ended 22 of the 27 finished terms higher. As a check, every four-year window starting on any trading day from March 1929 was measured the same way: 79.65% of 23,511 windows ended higher, with a median change of +36.29%. A term behaved like a four-year window drawn from the same history, which is what it is.

The five that fell were Hoover’s term, Roosevelt’s second, Nixon’s second and both of George W. Bush’s. All five contained a deep fall inside the term: 86.19% from the high under Hoover, 54.47% in 1937 to 1941, 33.84% in 1973 to 1974, and 43.46% and 51.93% in the two Bush terms. Each of those falls is a fact about the dates; this page does not try to assign them to policy.

Grouped by president, consecutive terms joined. Bill Clinton’s eight years came to +15.12% a year, Barack Obama’s +13.01% and Joseph R. Biden’s four +12.11%. Donald J. Trump’s two terms, compounded and still unfinished, come to +14.46% a year over 5.71 years. Herbert Hoover’s came to -31.01%, George W. Bush’s -5.56% and Richard M. Nixon’s -3.95%. Over the whole span, from 25.83 on 1 March 1929 to 7,722.72 on 2 October 2026, the index rose 6.02% a year on price alone.

Election day

Election day is set by statute. Section 21 of title 3 of the US Code, read on govinfo on 3 October 2026, defines it as “the Tuesday next after the first Monday in November”. The move measured here runs from the last close before election day to the first close after it.

Most were small. Across the 25 presidential elections from 1928 to 2024, the index rose after 11, fell after 13 and was unchanged once. The median move was -0.05%. Six moved 2% or more either way: falls of 4.42% in 1932, 3.32% in 1940 and 4.61% in 1948, and rises of 2.53% in 1996, 4.02% in 2020 and 3.79% in 2024. Up to 1980 the series has no close on election day itself, so those moves span a single session. Every election is in the election day file.

What the table cannot tell you

It cannot separate a president from the timing. A term begins wherever the previous one left the index. The 2009 example above shows one close ending a loss and starting a gain within the same week.

It cannot separate a president from everything else. Interest rates, wars, recessions, laws passed by Congress and decisions by the Federal Reserve can all move the index, and a four-year number contains whatever happened in those years at once.

It cannot be used as a forecast. With 27 finished terms, any split of them into groups leaves a small sample, and the spread inside each group is wide. Five terms that fell sit next to 22 that rose.

When it fails

It fails when the window is too short. Per-year figures for terms under two years, such as Roosevelt’s fourth or Johnson’s first, swing on a few months of prices.

It fails when dividends matter. Every figure here leaves dividends out, so each one understates what a holder who reinvested them received. This page does not measure by how much.

It fails when someone reads cause into it. A table of terms and returns invites the claim that one side is better for markets. The data here supports no such claim, and none is made.

And it fails as a trading input. Moving money in or out because of who won an election means betting on a four-year window, and this history says most four-year windows rose regardless of the result.

Related

The all-time high page uses the same index history to ask what followed record closes. A bear market is the kind of fall behind every losing term in the table. And buy and hold is the approach that holds through every term rather than choosing between them.

This page is educational, not financial advice. Test every idea on your own charts before risking money.