The Santa Claus Rally: what the data actually shows

Most people use "Santa Claus Rally" to mean stocks go up around Christmas. The term is narrower than that, and the difference matters, because the loose version is not a tradeable claim and the strict one is.

What the window actually is

The Santa Claus Rally, as Yale Hirsch defined it in the Stock Trader's Almanac in 1972, is a specific seven-session window: the last five trading days of the year plus the first two of January. Not December. Not "the holidays". Seven sessions.

That definition is doing real work. December as a whole is an unremarkable month. The last week of it is not, and folding the two together produces an average that describes neither.

The three questions worth asking

Any seasonal claim should survive the same three questions, and most do not survive the second:

1. How large is the average move? A pattern that averages a few tenths of a percent is not worth the transaction costs, however reliable it looks.
2. How often does it happen? An average of +1.3% built from nine flat years and one +13% year is not a pattern, it is one event with a mean drawn through it.
3. How bad is the worst case? A window that has been positive 80% of the time can still have a -12% year in it, and that year is the one that decides whether you can hold the position.

The second question is where most seasonal folklore falls over. It is also the easiest to check and the one almost nobody publishes, because a win rate is a much less flattering number than an average.

Why the average misleads here

Here the mean (+0.72%) and the median (+0.62%) sit close together, which is unusual for a seasonal statistic and is the strongest thing you can say about this one: the typical year really does look like the average year. The best window on record is +7.4% and the worst -4.0%, so the distribution is tight by market standards.

Holiday-period returns are still skewed. Volume thins out, so a small number of orders moves price further than usual, and the distribution ends up with a few large outliers in both directions. The mean gets dragged toward whichever tail was heavier over your sample.

The median is the more honest summary. When the mean sits well above the median, the "typical" year looks nothing like the average year, and an average is exactly what gets quoted every December.

The five best and five worst

What to do with it

Treat it as a base rate, not a signal. Knowing that a particular seven-day window has historically been positive more often than not tells you something about the distribution you are trading into. It does not tell you what this year will do, and the sample is small: a few decades of history is a few dozen observations, which is not many to draw a conclusion from.

The useful version of this exercise is not reading someone else's number. It is checking the window on the instruments you actually hold — an index behaves differently from a single name, and a small cap behaves differently from a mega cap. The pattern that shows up in the headline is the index's.

Run this on your own tickers

1. 1Open TradeSeasons and pick your ticker in the symbol selector at the top.
2. 2Set the date range to the last 20 or 30 years. Fewer than ten and the sample is too small to mean anything.
3. 3Go to the Trading Analysis tab and set the period to Dec 24 – Jan 3, which covers the window on most calendars.
4. 4Read the win rate and the worst year, not the average. Those two decide whether the position is holdable.
5. 5The pattern returns chart gives you the bar-per-year view above, for your ticker instead of the index.

A single stock will look nothing like the index here. That is the point: the headline number everyone quotes is the S&P's, and you are not holding the S&P.

Open TradeSeasons

Related calendar effects

The Santa Claus window is one of several dated effects with a long paper trail. The January Effect overlaps its tail end. Sell in May is its seasonal opposite. All three deserve the same three questions, and all three answer them differently.