Causes and Consequences of the Relation Between Split-Adjusted Share Prices and Subsequent Stock Returns

In this manuscript, we document and explain an empirical artifact - a persistent and substantial negative relation between split-adjusted share prices and subsequent stock returns - that has potentially important ramifications for capital markets research design. This relation pervades all commonly-used commercial databases and is insensitive to the choice of database used for either prices or returns. We investigate four potential causes of the empirical regularity: survivorship bias, asymmetric returns to low-priced stocks, extreme returns, and the effects of stock-split adjustments on portfolio classifications. We find that survivorship bias accounts for approximately half of the returns documented to a share-price-based hedge strategy and that re-classifications caused by stock split adjustments account for substantially all of the remaining returns. We do not find that controlling for either low-priced stocks or extreme returns is effective in purging the data of the empirical price artifact. These findings and our explanations thereof are important, as they show that there are potentially troublesome consequences of using share price as a deflator in markets-based research. In particular, we note and illustrate cause for concern when interpreting associations between share-price-scaled variables and subsequent returns as evidence of market inefficiency. Copyright 2007 The Authors Journal compilation (c) 2007 Blackwell Publishing Ltd.