For years investors have talked about a "January effect" in equity markets. We have just completed a research study of what happens to markets in January. Our efforts have produced a forecasting tool that might shed some light on whether January 2002 is likely to give investors something more to talk about.
The January effect scenarios described by different analysts vary, but the usual implication is that in January something exceptional (but not unexpected) could happen to stocks. Think of it as an "El Nino" in the markets.
We started our study by asking if there was a January effect and if so, what was it. A search of the literature reveals there are at least three notions in circulation: a performance effect, a capitalization effect and a reversal effect.
The performance version of the January effect might be the most widely cited. It holds that markets tend to perform better in January than in the rest of the year. When we looked at averages of monthly returns over 75 years this hypothesis appeared correct, at least by a few percentage points. But the year-to-year fluctuations in market return swamp this minor difference in statistical terms for all but small-capitalization stocks. Even if strictly true, this form of the January effect seems to be more of a statistical oddity than something that could be exploited by institutional investors who need to be fully invested year-round.
We focused our attention on the other two forms that conceivably could be the subject of tactical or strategic portfolio action.
The capitalization form of the January effect posits a tendency for small-cap stocks to outperform large-capitalization stocks in January. On average, this version also appears to be correct. Although the effect is not present every year, small-cap stocks have outperformed big caps by at least some small amount in more than 80% of the Januarys on record.
In the reversal form of January effect, stocks that have displayed a strong price trend in the previous months tend to reverse in January and re-establish their trends in February. This has been market lore for many decades. The reversal form flies in the face of typical price behavior during the rest of the year. It is well known that in the U.S. markets stocks generally persist in six- to 12-month price trends. This can be easily confirmed by anyone with access to modern financial databases.
The reversal version of the January effect seems to be a tax-related phenomenon. For example, suppose that in mid-December a taxable investor holds a stock that has been up strongly for the year and is considering selling it to take profits. If sold before year-end, taxes on the gains must be paid four months later in April. If the investor delays selling until January, the tax is due in April of the next year. The time value of the tax not paid is obvious. Similarly, stocks with large declines for the year, if sold before year-end, produce losses that are valuable to offset gains and reduce taxes. Investors have a similar tax-driven incentive to recognize those losses in December rather than wait until early January when the benefits won't be seen for another year. A further imperative for sales to offset existing realized gains comes from the uncertainty of gains in the following year.
It's easy to see how this tax-induced effect might make stocks that have been strongly up for the year drop in early January, while stocks down strongly for the year might be forced even lower in December and rebound a bit in January.
Institutions are alleged to engage in year-to-year window dressing. The window dressing effect has institutions eager to display the year's big winners in their holdings, and equally eager to remove losers. If true, this works to strengthen the January reversal effect, but it is, of course, not tax driven.
Some see shift
Some have suggested the January effects have shifted into December or November. It's certainly possible that the delay and acceleration of sales could begin early, but pressure won't be released until the first of January.
It is less obvious, but the capitalization version of the January effect might be tax driven, too. Consider two stocks with identical capitalizations at the beginning of the year. Suppose one doubles in price while the other declines 50%. Of course the gainer will be capitalized larger than the loser, and the reversal effect will make it appear the small-cap stock outperformed the big cap at year end.
If a tax-induced reversal effect occurred every year, it would be exploited quickly by low transaction cost market participants. However, this form of January effect is distributed very strangely.
A few years display an extreme effect, but there seems to be no January with an equally strong effect in the opposite direction. Over the past three decades just a few years - 1971, 1974, 1975, 1992 and 2001 - are the most extreme.
It appears that in two-thirds of the Januaries there is no strong effect; weak and strong stocks exhibit normal fluctuation about a near zero average. In roughly one-third of Januarys there is a substantial effect.
In our efforts to develop a predictor for the January reversal effect, we examined valuation/momentum cycles, a wide range of fundamental factors, economic activity, market direction and volume before settling on our current formulation. Extrapolating from conditions in mid-November, it appears our indicator will forecast a strong January reversal effect for 2002. We anticipate a relative strength reversal measured by worst/best performance deciles of 20% with a standard deviation of 10%.
John S. Brush is president of Columbine Capital Services Inc., Colorado Springs, Colo.