The authors examine the returns of stocks in the Cowles Industrial Index before and after the introduction of personal income taxes in 1917. This is distinct from earlier studies because they cross-sectionally analyze the relationship between the returns of the individual stocks and measures of tax-loss selling potential and size. The authors find that excess returns at the turn-of-the-year and for the month of January were not significant until after 1917. These results provide strong support for the tax-loss selling hypothesis as an explanation for the January seasonal in the returns of small firms.

We examine the returns of stocks in Cowles Industrial Index before and after the introduction of personal income taxes in 1917. This is distinct from earlier studies because we cross‐sectionally analyze the relationship between the returns of the individual stocks and measures of tax‐loss selling potential and size. We find that excess returns at the turn‐of‐the‐year and for the month of January were not significant until after 1917. These results provide strong support for the tax‐loss selling hypothesis as an explanation for the January seasonal in the returns of small firms.