Will Seasonal Slump Drive Derisking?

The so-called January-Effect is almost at an end and if the market closes near these levels, the S&P 500 will have managed a 4.4% gain or its 20th best January since 1928 (84 years) and best since 1997. The outperformance of banks and sovereigns (LTRO) and the worst-of-the-worst quality names (most-shorted Russell 3000 stocks +9% YTD vs Russell 3000 +5.2%), as Morgan Stanley noted recently, is not entirely surprising since the January effect is considerably larger in mid-cap and junk quality names than any other size or quality cohorts. We have pointed to the seasonal positives in high-yield credit and volatility and along with the obvious short squeeze in S&P futures (which has seen net spec shorts come back to balance recently), we, like MS, are concerned that the tailwinds of exuberance that virtuously reflect from seemingly pivotal securities (such as short-dated BTPs now or Greek Cash-CDS basis previously) very quickly revert to a sense of reality (earnings and outlook changes) and perhaps the slowing rally and rising volatility of the last few days is the start of that turbulence.

The most-shorted stocks (tracked by the red lines on the above chart) have dramatically outperformed the broad markets they are part of with the Russell 3000 most-shorted (thick red) massively outperforming (almost 400bps in the month!).

Morgan Stanley: January Effect

January is often a month for risk taking since optimistic investors believe that any underperformance during the month can be reversed by year-end.

In light of the sharp rally in the equity market thus far this year, we took some time to study the concept of a “January Effect.” Since 1901, the S&P 500 has averaged a 1.2% return during January with a standard variation of 4.3%. In the remaining eleven months of the year, the index has averaged a 0.5% monthly return with a 5.2% standard variation (Exhibit 2).

After accounting for the standard deviations, the return spread between January and the remaining eleven months is marginally statistically significant: With a T-stat of 1.73, it is significant at the 10%-level but insignificant at the 5%-level. In fact, 2012’s rally to date is only a 0.8 standard deviation event, and studying history, we would expect such a move to occur in slightly over 20% of January’s. We studied the “January Effect” by market cap cohort, quality-junk status, and value-growth status. Since 1970, the spread between the January return and the return for February through December has been highest in mid-cap stocks (Exhibit 3). None of the three cap cohort’s return spread is statistically significant—the mid-cap spread has the highest T-stat at 1.46. Year-to-date performance so far this year by cap cohort is consistent with smaller-cap outperformance.

We analyzed returns by quality cohort since 1981 and found that both quality and moderate quality, on average, perform worse in January than during the remainder of the year. Low quality slightly outperforms in January, while junk is by far the largest outperformer on average (Exhibit 4). The more positive performance of junk relative to the other quality quintiles is not surprising given that junk stocks are generally smaller than quality stocks, and the January effect is stronger in these small stocks. Still, none of the quality cohorts’ return spreads are statistically significant after accounting for volatility and the number of observations.