Stock Splits, Price Changes, and Trading Profits: A Synthesis

Abstract

Over the years the relationship between stock splits and stock price behavior has been a subject of continuing interest to economists and practical men of affairs. Very little systematic empirical evidence on the subject was available, however, until the recent appearance of two papers. The first paper presented the results of a study of splits made by Keith B. Johnson.' While interesting and provocative, Johnson's analysis contained, in our opinion, a rather arbitrary assumption concerning the relevant time period over which to study stock splits and price behavior. We suspected that the substitution of different and, possibly, more logical time periods might significantly alter Johnson's conclusions. With the appearance of a second paper by Fama et al.2 (hereafter FFJR), our suspicions received considerable support. After reading their paper, we were sufficiently confident about our tentative hypotheses to want to use a methodology similar to Johnson's to study the relationship between splits and stock price behavior over different time periods. Furthermore, we felt that by doing so we would shed additional light on several issues raised by FFJR, reconcile some surface differences between their results and Johnson's, and, in general, provide a synthesis of the analysis of stock splits. The results of our endeavors are reported in this paper. To place our analysis in the appropriate context, we begin by discussing Johnson's study of stock splits and price behavior and the results reported by FFJR. Following this, we discuss the data on which our analysis was based and the results we obtained. Finally, we attempt to synthesize our results with those of Johnson and FFJR.