What January Effect?

The first month used to be the best month to buy stock. Not anymore.

In the waning weeks of 2002, channel-surfers who happen upon CNBC on their way to Frosty Saves Christmas will hear lots of hastily considered remarks about the “January effect.” Like EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) the January effect is one of those investing terms that is cheerfully batted about but rarely explained or examined in detail. And like EBITDA, the January effect may be increasingly useless for making investment decisions.

First of all, there is not one January effect. There are two. They concern different historical stock market phenomena and matter to different groups of investors.

The first holds that the action in the market in the month of January predicts the general market direction for rest of the year. For 45 of the previous 51 years, when the Standard & Poor’s 500 rose in January, it was up for the year; and when it was down in January, it was down for year. Barring a miraculous bull run next week, 2002 will make it 46 of 52. (What’s the logic of this January effect? Click here for an explanation.)

The second January effect interests short-term traders. It turns out that January has been the most rewarding month for investors. In the ‘70s, professors Michael Rozeff and William R. Kinney Jr. examined average monthly returns of New York Stock Exchange stocks between 1904 and 1974. They found that the average monthly returns for January were more than five times greater than the average monthly returns of the other months.

Yes, the same dynamics of optimism and new cash could have something to do with the comparatively positive returns in January: If everybody is out shopping for stocks at the same time, they’ll get more expensive. But there’s more to the story. For it turns out that small stocks tend to perform far better than bigger stocks in late December and early January. Since 1979, according to research by Salomon Smith Barney’s U.S. Quantitative Strategy unit, the average return spread between the Russell 2000 Index and the Russell 1000 Index (the former is a proxy for small stocks and the latter a proxy for large stocks) for the 24 Januaries since 1979 was .82 percent—meaning if the Russell 1000 went up 2 percent, the Russell 2000 went up 2.82 percent. For all the other months, however, the spread was a negative .1 percent—meaning if the Russell 1000 went up 1 percent, the Russell 2000 was up only .9 percent. The returns are further concentrated into a few trading days. If you look at 1979 through January 1995 and examine the last two trading days of December and the first five trading days of January, small stocks did even better.

The theory? As the year ends, investors dump beaten-down stocks. They want to recognize tax losses to offset capital gains or simply purge portfolios of stinkers. At the same time, they tend to hold on to stocks that have performed well, to avoid taking capital gains. (Mutual funds have been known to add solid-performing stocks at the last minute so that when they report their holdings as of Dec. 31, they will include winners.)

Add it all up, and the stocks that have been poor performers are likely to experience comparatively more selling in December, and the good performers are likely to see less selling. This action may create temporary inefficiencies—the small, cheap stocks get inordinately cheaper, and the big, expensive stocks get more expensive—that are rectified as new money floods into the market in early January. As the market regains its balance the small, cheap stocks do better and the big ones do (comparatively) worse.

But the small-stock outperformance in early January has decreased in recent years. Between 1995 and 2002, the Russell 2000 beat the Russell 1000 in that seven-day December-January trading period just three times, and the average spread was a slim .34 percent, according to Salomon Smith Barney.

This is no surprise. Once an investing theme trickles down from the professionals to the masses, it’s probably no longer a good idea. Markets have a way of pricing in expectations. In the late ‘90s, investors came to understand that the stocks of Initial Public Offerings would rise when the 30-day quiet period ended and the stock’s underwriters would come out with a buy recommendation. So, savvy traders would start to build a position a few days before the event. The end result was that the impact of the event itself ultimately became a nonevent.

With so many people hip to the January effect, the same may be happening to it. An  investor, wanting to ride the updraft, would buy beaten-down small stocks in late December and then sell them in early January. But if everyone is buying beaten-down small stocks in late December, they won’t be beaten down anymore. So, a shrewd investor might start buying those stocks in mid-December. Pretty soon you have a December effect, then a November effect …