Testing the Tax-Loss Selling Explanation of the January Effect: Evidence from a 'Confiscatory' Tax Implemented in France in 1921

Abstract

For almost a century, we document a significant January effect on the French equity market. We find strong evidences in favor of the tax-loss selling explanation for this phenomenon. Indeed, the January effect was insignificant before the introduction of a “confiscatory tax” on capital gains in 1921, and became strongly significant afterward. Moreover, the rate of taxation is statistically significant to explain the strength of the January effect over time. Studying individual stock returns, a DiffInDiff investigation shows that past losers, outperform past winners at the turn of the year but only after 1921, which reinforces further the tax-loss selling explanation of the January effect.

Keywords: Market efficiency, Calendar anomalies, historical finance

JEL Classification: G1, G12, N23, N24