A SIMULTANEOUS TEST OF COMPETING THEORIES REGARDING THE JANUARY EFFECT

Abstract

Abstract In a multivariate context, the January effect appears most significantly related to excess individual liquidity (i.e., high cash balances and low expected taxes), but multicollinearity may obscure the relation between other variables and the effect. In a univariate context, prior February–December returns and the standard deviation of prior‐year returns are most significantly related to the January effect. In both contexts, higher January volume and lower real interest rates are correlated with higher January returns. I find no evidence that window dressing by professional managers or macroeconomic seasonality (other than real interest rate seasonality) are significantly related to the January effect.