January Barometer Over the Long Run
Posted in Calendar Effects
January 6, 2009
The conventional wisdom is: “As goes January, so goes the rest of the year.” Does long-run data support this belief? To check, we turn to the very long run dataset of Robert Shiller, which offers monthly levels of the S&P Composite Stock Index since 1871. Using monthly closes for the S&P Composite Stock Index from January 1871 through December 2008 (138 years), we find that:
The following scatter plot relates the return for the S&P Composite Stock Index during February-December to the return for the immediately preceding January over the period 1872-2008. The plot is not tightly packed. The Pearson correlation between the two series is 0.24 and the R-squared statistic is 0.06, indicating that January returns are a barometer for 6% of returns for the balance of the year. Other things (or randomness) explain the other 94% of February-December returns.
Is this (weak) relationship reliable across subperiods?
The next chart shows the Pearson correlations between the S&P Composite Stock Index return for January and the return for February-December of the same year during three subperiods: 1872-1921, 1922-1971 and 1972-2008. While the correlation is consistently positive, its magnitude varies by subperiod.
Is the January barometer more accurate for some ranges of January returns than others?
The next chart recasts the data in the above scatter plot by ordering the S&P Composite Stock Index returns for January from lowest (-6.8%) to highest (+9.2%) during 1872-2008. The horizontal axis is therefore not time-sequential. This view of the data suggests that the January Barometer is subject to high variability of February-December returns across all ranges of January returns and that it is not useful for very high values of January returns.
Is the return for January more predictive of the return for the next 11 months than are returns for other individual months?
The final chart shows the correlations between the S&P Composite Stock Index return for each of the 12 calendar months and the return for the immediately following 11-month interval over the entire 1872-2008 sample period. It shows that returns for the months of April, May, August, November and December are about as good as the return for January in predicting returns for the ensuing 11 months. In other words, January is not special.
In summary, evidence from long-run data indicates that the January return for a broad U.S. stock index is weakly predictive of returns for the ensuing February-December. However, the predictive power of January is not appreciably greater in this regard than that of five other months.
