Seasonality in firm-size portfolio returns: A nonparametric analysis

In this study the author uses stochastic dominance, a nonparametric method of portfolio performance analysis, to test for seasonality in firm-size portfolio return behavior. Stochastic dominance confirms the January effect, found in previous parametric studies, only for the smallest firm-size portfolio. It statistically eliminates the size effect for the larger firm-size portfolios in January and for all firm-size portfolios in the other months of the year. It is demonstrated that a market proxy problem and normality assumption violation may bias the parametric results. Nonparametric analysis, therefore, suggests that markets may be more efficient than parametric methods imply when model violations exist.