Wednesday, March 2, 2011

Voodoo Statistics Part 2 - The Jan-Feb Effect

A few days ago all anyone could talk about was how stocks were virtually guaranteed to go up on the first day of the month. After all, the first trading day for the S&P 500 had been positive for 7 straight months, and 14 out of last 16 months (88% of the time!).

Well we all know what happened on February 1. And hopefully those that read my blog First Trading Day Of The Month - Voodoo Statistics had the sense to protect themselves in case things didn't go according to plan.

Now I'm hearing a lot of folks talk about the Jan-Feb effect, that is when the stock market rises in January and February the market almost always closes higher for the entire year. One study states that since 1938, when the S&P 500 rose in January and February (as it has this year), the index went on to close higher for the entire year 25 out of 26 times. This stat was referenced in Eddy Elfenbein's blog Crossing Wallstreet earlier today. (Note: I am not picking on Mr. Elfenbein in any way, just referencing where I got the data. This stat has also been discussed endlessly on CNBC and elsewhere)

I don't actually have the raw data going back to 1938, but I do have the data going back to 1970. So let's take a look and see what the data actually says:

Since 1970 (41 years), the S&P 500 has been positive in both Jan and Feb 22 times. Of those 22 years, the market went on to close positive 21 times. That seems pretty consistent with the stats referenced above. 21 out of 22, or 95% of the time, the market closed the year higher.

Sweet! Let's mortgage the house and buy high-beta stocks on margin then!

Ok seriously, 95% does sound impressive. And in some ways it is. But let's put the number in proper perspective before we all go out and buy stocks like drunken sailors.

First off, for those 22 years in question, the market averaged gains of 6.3% in the first two months. This year the S&P 500 "only" gained 5.5% in the first 2 months. What if instead we asked the question, "what happens in years where the market gains less than 6% during the first two months?". Certainly that is not any more arbitrary than the original question posted above?

As it turns out the S&P 500 has averaged gains of less than 6% in January and February in 30 out of the last 41 years. And in those years the market went on to post positive gains just 21 out of 30 times. That's only 70% of the time compared to 95% mentioned above. Furthermore in those 30 years the market ended up averaging gains of just 3.5% for the entire year.

Another thing to consider is that since 1970 the market has been up 31 out of 41 years (76% of the time) no matter how January or February fared. Historically stock markets go up. That will put a positive bias on any study.

The point of the above is that data can be sliced and diced in an infinite number of ways and can be made to say almost anything you want it to. And like anything in life, past returns are no guarantee of future results. We certainly saw that on February 1 this year.

No comments:

Post a Comment