Skip to main content

Horrors! There are only 26 shopping days between Thanksgiving and Christmas. 

Is the retail industry doomed to a poor holiday shopping season because it's so short this year?

That's the argument being made by some industry experts, given that the number of days between Thanksgiving and Christmas is six days shorter this year than last. One typical headline -- this one from CNBC -- reads: "This year's calendar shortchanges retailers this holiday shopping season."

A dismal holiday shopping season would be worrying, of course, since so many retailers are dependent on it in order to be profitable for the entire year.

Wouldn't it be ironic for the economy to be tripped up by the mere timing of Thanksgiving, after having (so far successfully) dodged the twin bullets of an inverted yield curve and a brewing trade war with China?

Yes. But you can relax.

There is no correlation between the length of the holiday shopping season and the performance of the retail sector.

I started my review by analyzing the S&P Retail Select Industry Index, which is the index underlying the SPDR S&P Retail ETF (XRT)  . Unlike the ETF, whose inception was in 2006, the index dates back to 1999. For every year since, I measured its return over the Thanksgiving-to-Christmas period.

I then measured the correlation between the index's returns and the length of those periods. I came up empty.


To appreciate this finding, consider that the worst index return since 1999 during this holiday shopping period came last year, when it lost 13.8%. Yet that was when there were 32 shopping days between Thanksgiving and Christmas -- the theoretical maximum length that this period can have (given that the earliest calendar date Thanksgiving can occur is Nov. 22).

Next, consider the second-worst performance for this index in the Thanksgiving-to-Christmas period. That came in 2002, when it lost 13.1%. Yet in that year this shopping period was just 26 days long, the same as it will be this year. (26 days is the theoretical minimum length that this shopping period can have. Such shortest periods come when Thanksgiving falls on Nov. 28, the latest day on which it can possibly occur).

Similarly for the other years since 1999. My PC's statistical package could detect no correlation between length of shopping season and the performance of the retail sector.

To be sure, it would be ideal, from a statistical perspective, to have more than just two decades of return data. But that's all we have for the S&P Retail Select Industry Index.

As a robustness check, I therefore re-ran the analysis focusing on the Dow Jones Industrial Average back to 1896, when it was created. Now there were more than 120 data points. And, once again, there was no correlation between performance and the length of the Thanksgiving-to-Christmas shopping period.

The bottom line? The upcoming shopping season may be disappointing. But if it is, it won't be because there are only 26 shopping days between Thanksgiving and Christmas.

The broader investment implication: It's important to subject to critical scrutiny the stories we tell in an attempt to make sense of the markets. They remind me of Rudyard Kipling's "Just So" stories, such as the one about how the leopard got his spots. They have superficial plausibility, but break down when they are subjected to any critical scrutiny.