A mystery of the January effect : single or multiple driving forces?

This study examines the joint effect of potential explanatory factors of the January effect. Four major hypotheses are tested in the study, including the tax-loss selling/parking-the-proceeds, window dressing, overreaction, and liquidity hypotheses. The results of this study make the following contributions to the literature: [1] A new, potentially more powerful proxy for potential tax-loss selling (PTS) is derived. This new proxy allows us to examine directly the tax-loss selling effect on cross-sectional stock returns and trading volumes. [2] We simultaneously test the four major hypotheses in a regression. Previous studies fail to simultaneously control for the most potentially important factors in their analyses. Consequently, their empirical results and conclusions are often contradicting to one another, and make these studies on the January effect less conclusive. By inspecting the sign of the estimated coefficients, based on the prediction of each major hypothesis, we reexamine the effects of different possible explanatory factors and identify the underlying explanations of the January effect. Our evidence strongly supports the liquidity and the overreaction hypotheses. [3] We add an institutional-ownership interaction term for each tested factor in the simultaneous test. Since all the four tested hypotheses resort to trading behavior of either individual or institutional investors as the explanation of the January effect, the evidence from institutional ownership plays an indispensable role in testing these hypotheses. It also provides an opportunity to distinguish between the tax-loss selling and the window dressing hypotheses, which is unavailable in prior studies. [4] We examine the trading behavior of management, institutions, and insiders around year-end. If the January effect is exploitable, then its continuous existence is obviously evidence against the efficient market hypothesis, which is the foundation of many modem finance theories. Since the January effect is a well-known phenomenon, these stock market participants may exploit this opportunity with their private information or lower transaction costs in stock trading. Because of the advantage in stock trading, they are the most likely candidates of exploiting the January effect. However, our results provide no evidence that these participants exploit the January effect.