The Groundhog Day stock market anomaly

Abstract

This paper discovers a distinct calendar anomaly on the US stock market associated with the Groundhog Day prognostication tradition across 1928–2021. There are significant positive abnormal returns around the “prediction” of an early spring, while buy-and-hold returns around the “prediction” of a long winter are 2.78% lower. The results are robust in subsamples, to a set of placebo tests for international stock indices, and cannot be explained by January effect, the “halloween Indicator”, turn-of-the-month effect, or other seasonalities. The findings imply major and persistent irrational optimism of US investors revolving around Groundhog Day early spring prognostications.