Content area
Full text
Anomalous evidence of seasonality in stock market returns presents a serious challenge to the Efficient Market Hypothesis. Previous studies often explain anomalous monthly returns as being caused by various institutional considerations, like taxloss selling or empirical problems tied to making inferences about market efficiency based solely on historical data-the "data snooping" problem. This paper analyzes an anomalous pattern of negative stock-market returns during the month of September. The "September Swoon" cannot be easily dismissed as a reflection of institutional consideration, time period considerations, nor differences in return measurement criteria. As such, it presents a challenge to the EMH.
(ProQuest: ... denotes formulae omitted.)
This paper studies long-term evidence on the seasonal pattern of monthly returns in the stock market. For large cap stocks, we examine 205 years of evidence using 1802-1926 data from Schwert (1990) and Center for Research on Stock Prices (CRSP) value-weighted portfolio returns from 1927 to 2006. The monthly pattern of small-cap stock returns is investigated using 80 years of CRSP equally-weighted portfolio returns from 1927 to 2006. We document an anomalous pattern of negative monthly returns during the month of September. A "September Swoon" is apparent using both value-weighted and equally-weighted portfolio returns, and across a battery of parametric and nonparametric tests. This September Swoon for both large cap and small cap stocks in monthly returns appears persistent and durable like Rozeff and Kinney's (1976) seminal evidence of a January seasonal in monthly returns for small cap stocks. ' As such, this anomalous pattern of September returns constitutes a challenge to the Efficient Market Hypothesis and represents a puzzle for future academic research.
The remainder of the paper is organized as follows. Section I describes the dangers of data-driven inference and related sampling issues. Section II documents consistently negative September returns for both value-weighted and equally-weighted portfolios. In Sections III and IV, a variety of parametric and nonparametric tests are used to establish the anomalous pattern of negative returns during the month of September. Section V offers a behavioral interpretation of the September Swoon, and suggests a prospective test. Section VI discusses the importance of our findings for investment theory and practice, and describes implications for future research.
I. Sampling Issues
A. Dangers of Data-Driven Inference
Both academics and...