Do stocks rise before holidays?

What do people believe about the day before a holiday?

The belief is simple: stocks tend to rise on the last trading day before a market holiday.

The usual stories are holiday cheer, a short trading day, and traders closing their bets before a long weekend. Some say the gain on these few days is many times an ordinary day's.

Researchers noticed it as far back as 1934. It became famous after academic studies in 1988 and 1990, and a long-running trading almanac has repeated it for years.

What exactly did we test?

We compared the market's return on each pre-holiday day with every other trading day.

A pre-holiday day is the last trading day before a market holiday. We let the data find the holidays: a day counts if the market was shut on a named holiday, like Thanksgiving, Christmas or Labor Day.

One-off closures, such as storms or the 1933 bank holiday, don't count, because they weren't regular named holidays. That gave us 939 pre-holiday days from July 1926 to August 2026, against 25,378 other trading days.

We measure every return above what cash in Treasury bills earned on the same day. When we compare two averages, the gap is in percentage points, or points for short.

What did we find?

The day before a holiday really has been special over the century, and if anything it stayed ahead in the last 20 years. Since 1926 it averaged +0.26% above cash, against +0.02% on other days. That's 11.7 times an ordinary day's average.

Pre-holiday days also beat cash more often: 68% of the time, against 53% of other days.

The gap of 0.24 points has a likely range of +0.18 to +0.30 points, well clear of zero. It showed up in the early half of our history (+0.33 points) and the later half (+0.16 points).

In the last 20 years the gap was smaller and less clear. Pre-holiday days averaged +0.17% above cash, against +0.04% on other days, a gap of +0.13 points. That's 4.3 times an ordinary day's. They still beat cash more often: 64% of the time, against 54%.

How did each year do?

The day before a holiday beat an ordinary day in 86 of 100 years since 1926, and in 17 of the last 20.

Each column is one year: that year's average day before a holiday, minus its average ordinary day. A column above the line is a winning year, when the day before a holiday beat an ordinary day.

From 1926 to 1980, every year was a winning year: 55 in a row. Before the belief got famous in 1988, it won 60 of 62 years.

Then came a surprise. Soon after the studies made it famous, it went 6 years in a row without a winning year, from 1992 to 1997. Each of those years, the day before a holiday trailed an ordinary day. Then the winning years came back. Since 1988 it has won 26 of 38 years, and 17 of the last 20.

A year holds only a handful of pre-holiday days, so one bad day can sink a year.

Did it fade once it became famous?

It shrank after 1988 but didn't fade by our verdict rule, and if anything it stayed ahead in the last 20 years. Since 1988 and in the last 20 years, that could be luck. Before 1988 the gap was +0.32 points; since then it has been +0.10 points. In the last 20 years it was +0.13 points.

The likely range since 1988 is +0.01 to +0.20 points. That stays above zero, but only just. We widened the range because we checked several things at once: -0.04 to +0.24 points. That includes zero, so the edge since 1988 alone could be luck. The verdict rests on the full period, and its range stays clear of zero even widened.

The decade chart tells the same story. The edge was big for decades, then small from the 1990s to the 2010s. It was bigger again in the 2020s so far, if anything, and it never turned negative in any decade.

One common story is that once a pattern gets famous, people trade on it, and some of the edge goes. Here, the average stayed above zero, if anything.

Did every holiday join in?

Since 1926 the day before all 8 big holidays averaged more than an ordinary day, Presidents' Day only just; 6 did in the last 20 years. The big holidays are the 8 the market has closed for from the 1920s to today.

Over the century, the days before Labor Day and Independence Day were among the strongest, at +0.36% and +0.37%. The day before Labor Day beat cash 75% of the time. The day before Thanksgiving averaged +0.20%, and the last trading day before Christmas +0.28%.

The last 20 years shuffled the order. The day before Labor Day averaged -0.05%. New Year's Day's averaged -0.11%, against +0.22% over the century. Good Friday's was the strongest, at +0.59%, against +0.33%. Independence Day's averaged +0.21%, Thanksgiving's +0.31% and Christmas's +0.14%.

Each holiday comes only once a year, so one bad day moves a recent average a lot.

Presidents' Day was the party pooper. The day before it averaged only +0.06% and beat cash 53% of the time, about like an ordinary day (53%). In the last 20 years it averaged +0.13%.

Does it hold everywhere?

It held across the market over the century. Pre-holiday days beat other days in all 43 industries with enough history, out of the 49 US industry portfolios we checked.

Industries all move with the market, so these aren't independent votes. Still, it's a sign the effect isn't one odd corner.

Why might it work?

Nobody knows for sure, and we didn't test the reasons. People point to good moods before a break, and to traders closing bets against stocks before a long weekend.

Some of it may be mechanical. Pre-holiday sessions are often shortened, and a short, quiet day may simply behave differently.

What's the story in numbers?

939 days before a market holiday since 1926. They averaged +0.26% above cash, against +0.02% on an ordinary day; in the last 20 years, +0.17% against +0.04%. Winning years: 86 of 100, with 55 in a row from 1926 to 1980, and 17 of the last 20.

Three things to take away. The day before a holiday really has beaten an ordinary day over the century. The edge shrank after it became famous: +0.32 points before 1988, +0.10 points since. But the day before a holiday still won 17 of the last 20 years.

And a real edge can still be worth little. Owning stocks only on those days, with cash otherwise, grew +4.6% a year in total return, against +9.9% staying invested. In the last 20 years: +2.2% against +11.5%. It grew slowly and fell far less for the same reason: it was in stocks on only 4% of days. Its worst fall was -6.2%, against -83.9%; in the last 20 years, -6.2% against -54.7%.

What does this mean for you?

MATH doesn't mean you should trade it. The edge is a fraction of a percent on a handful of days a year, and that's tiny in money terms.

Costs bite hard. At a quarter of a percent a switch, the pre-holiday rule's total return fell to +0.9% a year. In the last 20 years, its total return fell to -1.5%.

And the opposite bet, selling before holidays, would be its own untested idea. This is history, not a forecast for the next holiday.

What else did we test?

See all the calendar effects we tested, or how we test a myth.