Seasonality in style returns

In this paper, I examine seasonality in returns to style portfolios, which serve as important benchmarks for asset allocation, and investigate its implications for investment. In doing so, I consider monthly returns on the style portfolios classified by the six size/book-to-market sorting and the six size/prior-return sorting over the sample period 1927 - 2006. The key findings are: first, as is well known, small-cap oriented portfolios are subject to the January effect, but also to the (negative) September and October effects. Second, cross-style return dispersion exhibits a seasonal pattern of its own (it is largest in January and smallest in August), suggesting possibly profitable trading strategies. Indeed, my seasonal strategies yield significant profits, as high as about 18.7% per annum. Finally, this profit can be mostly explained by the seasonal autocorrelation in style returns.