The Predictive Value of the Signs of January Returns: Evidence of a New January Effect

Abstract

Many investment practitioners rely on the 'January Barometer', namely, they believe in the theory of 'as goes January, so goes the year', which is fundamentally related to the well-known January effect. Unlike many prior studies that examined the well-known January effect per se, we investigate the issue of whether or not the signs of January returns have superior predictive value for future stock returns vis-a-vis the predictive value of the signs of returns in any other calendar month. Using NYSE equal-weighted stock index, we confirm the existence of the January effect in the sample period of 1941-2002 with both unconditional and conditional evidence, and we document two additional manifestations of the well-known January effect. We also introduce a new type of January effect, namely the signs of January returns have superior predictive value vis-a-vis the signs of any other calendar month's returns for the purpose of predicting the next 12-months' returns.