The Pre-Holiday Effect: What the S&P 500 Did the Day Before Every Market Holiday Since 1950

The market closes for Thanksgiving in five weeks, and the session before it has one of the most consistent records in the whole calendar. That is not folklore; it is one of the oldest documented anomalies in finance, the pre-holiday effect, and it is rarely measured on current data. We took every S&P 500 close since January 1950, found every weekday the market was shut for a holiday, and measured the session before and the session after. The effect is real, it is the strongest of the calendar patterns we have tested, it survived into the 2020s, and it is still not a reason to trade on a date. Education, not advice.

What is the pre-holiday effect?

The tendency for stocks to rise on the last trading day before a market holiday, by more and more often than on an ordinary day. Robert Ariel documented it in the Journal of Finance in 1990 on data back to 1963, finding that the day before holidays produced a large share of the market's total return. The usual explanations are thin trading, short sellers closing positions before a closure they cannot trade through, and a better mood heading into a long weekend.

How we measured it

We used the S&P 500's daily closes from January 3, 1950 through October 16, 2026, 19,319 sessions. A holiday is any weekday with no session that matches a known market holiday by date: New Year's Day, Martin Luther King Day, Presidents' Day, Good Friday, Memorial Day, Juneteenth, Independence Day, Labor Day, Thanksgiving, Christmas, and the older Election Day, Washington's Birthday and similar closures from the 1950s and 1960s. Unexplained closures, such as the week after September 11, 2001, are excluded. That gives 650 holidays. For each we took the close-to-close return of the session before (the pre-holiday session) and the session after.

What the numbers say

• The pre-holiday session averages +0.21% and rises 65% of the time. An ordinary session averages +0.03% and rises 53% of the time. The day before a holiday has been seven times the average day, with a hit rate twelve points higher.
• The session after the holiday is ordinary. It averages +0.06% and rises 54% of the time, with slightly larger swings than normal (0.74% average absolute move against 0.67%), because two days of news land on one open.
• Those 650 sessions are 3.4% of trading days. Compounded on their own since 1950 they would have turned $100 into about $384, price only. An illustration of how lopsided the pattern is, not a strategy; the spread and the tax on 650 round trips would eat it.

Which holidays are strongest?

• Thanksgiving: the day before averages +0.33% and has risen 76% of the time since 1950, the best pre-holiday session of the year. The half-day after it is also unusually good: +0.28% on average, up 71% of the time.
• Good Friday: the Thursday before averages +0.33% (up 69%), but the Monday after is the weakest post-holiday session, -0.18% on average, because the March jobs report often lands on the closed Friday and is priced on Monday.
• Memorial Day, Labor Day and July 4: +0.29%, +0.27% and +0.24% before, each rising about two times in three.
• Christmas: +0.23% before (up 67%) and +0.28% after (up 64%), the only holiday where both sides are strong, which is the core of the Santa Claus rally.
• Presidents' Day is the exception: the Friday before averages -0.06% and has risen only 43% of the time since the holiday moved to a Monday in 1971, and the Tuesday after averages -0.18%. Mid-February is also the weakest stretch of the month in our seasonality data.

Does the pre-holiday effect still work?

Ariel's paper was published in 1990, and the 1990s were the decade the effect nearly vanished: the pre-holiday session averaged +0.05% and rose only 49% of the time. It then came back. Since 2000 the pre-holiday session averages +0.11% (up 60%) against +0.03% for other days; since 2010, +0.14% (up 63%) against +0.05%; and in the 2020s so far, +0.21% with 67% up. The edge is about half what it was in the 1950s and 1970s, and it is still the largest calendar effect in our data, bigger than the first trading day of the month and far bigger than any weekday pattern.

The session after the holiday is where the change shows. Since 2000 it averages +0.01% and rises exactly half the time; since 2010, +0.01%. Whatever good mood carries into a holiday does not carry out of it.

The worst and best pre-holiday sessions

A 65% hit rate still means one in three falls. The worst since 1950: 2000-02-18 before Presidents' Day (-3.04%), 2009-07-02 before Independence Day (-2.91%), 2018-12-24 before Christmas (-2.71%), 2011-09-02 before Labor Day (-2.53%), 1950-11-06 before Election Day (-2.47%). Christmas Eve 2018 was the bottom of that quarter's 19% drawdown. The best: 1962-05-29 before Memorial Day (+4.65%), 2009-04-09 before Good Friday (+3.81%), 2008-11-26 before Thanksgiving (+3.53%), 1974-08-30 before Labor Day (+3.09%), 1957-11-27 before Thanksgiving (+2.89%). Both of the 2008 and 2009 entries came inside a bear market, which is the usual reminder that a calendar average is made of crisis outliers in both directions.

2026 so far: MLK Day -0.06%, Presidents' Day +0.05%, Good Friday +0.11%, Memorial Day +0.37%, Juneteenth +1.08%, Independence Day +0.00%, Labor Day -0.38%. Four up, two down, one flat, in line with the long-run hit rate. The remaining pre-holiday sessions this year are Wednesday, November 25 before Thanksgiving and Thursday, December 24, a 1:00 PM close, before Christmas, per the NYSE calendar. The half-day after Thanksgiving, Friday, November 27, also closes at 1:00 PM.

What this should change about your trading

Not your entries. Two tenths of a percent on a 65% hit rate is smaller than the spread on a normal day's range and much smaller than a stop. Buying the Wednesday before Thanksgiving because it is the Wednesday before Thanksgiving is a calendar rule, and the record above, like every calendar study we have run, is a record of averages that individual days ignore.

Your expectations, a little. A quiet, slightly positive pre-holiday session is the normal case, and a sharper session after the holiday is also normal. If a position is near its stop going into a long weekend, the day after is when it gets tested.

Your weekend exposure, which is a risk decision. The post-holiday session carries two days of news and swings about 10% more than an average day. Position sizing handles that; a date rule does not. For crypto, which never closes, the holiday is simply a thin-liquidity day with the U.S. desk gone, which is its own reason for tighter sizing.

JorgAI does not time holidays. It enforces the stop, the position cap and the daily loss brake inside your own brokerage account on the Wednesday before Thanksgiving and the Monday after it alike. Set your rules once; the calendar takes care of itself.

Quick answers

Does the stock market go up before holidays? Historically yes. Since 1950 the S&P 500 has averaged +0.21% on the session before a market holiday and risen 65% of the time, against +0.03% and 53% on an ordinary day. Since 2010 the pre-holiday average is +0.14% with 63% up.

Which holiday is best for stocks? Thanksgiving. The session before has risen 76% of the time since 1950 with an average of +0.33%, and the half-day after is up 71% of the time. Presidents' Day is the only holiday with a negative pre-holiday average.

What happens the day after a market holiday? On average nothing special: +0.06% since 1950, up 54% of the time, with slightly larger swings than normal. Since 2000 the day after is a coin flip.

When is the market closed for the rest of 2026? Thanksgiving, Thursday, November 26, and Christmas, Friday, December 25. The sessions on November 27 and December 24 close early at 1:00 PM Eastern.

Data: S&P 500 daily closes from Yahoo Finance, January 3, 1950 through October 16, 2026, our own count; holidays identified as weekday closures matching a market holiday by date rule, with unexplained closures excluded; returns are close-to-close and price-only. Ariel (1990), Journal of Finance, is the original study. Past patterns do not predict future returns. This article is education, not investment advice.