An Explanation for the Halloween Effect: Why "Sell in May and Go Away" Works

Abstract

We provide an empirical explanation for Bouman & Jacobsen's (2002) Halloween effect. Our analysis of over 10 million SEC filings from 1994 to 2023 reveals a significant recurring pattern, with its first half aligning with the May-to-October cycle (summer) and the second half with the November-to-April cycle (winter). As SEC filings contain stock price-impacting information and are used by investors, analysts, and the media, we link the winter overperformance to an information surge during this period. Specifically, we observe a 32% and 38% increase in monthly filing volume and insider trading activity, respectively, 84% of all annual reports with audited financial statements disclosed during this period, along with activist investor activity peaking.

Keywords: Halloween Effect, Sell in May and Go Away, Explanation for Halloween Effect, SEC Filings