Stock Splits and Information: The Role of Share Price

Financial Management (Financial Management Association), Autumn, 1999 by Robert M. Conroy, Robert S. Harris

Our empirical investigation of announced split factors, split announcement returns, and revisions of analysts' earnings forecasts shows that a firm's past history of stock splits plays a crucial role in both the design and effect of current splits. Managers appear to design splits to return their company's stock price to the price level achieved after the last split. Moreover, when managers announce a split factor to achieve an even lower price than in the last split, both investors and analysts interpret this as a signal of especially positive information.

"The purpose (of splits), apparently, is to bring the market price of the split stock down to the desired range--in the nineteen forties and fifties, $15 to $40 a share." -- Arthur Stone Dewing (1953) in the leading finance text of the day

"Managers report that the main motive for issuing stock splits is to move the stock price into a better trading range...the preferred trading range for these managers is from $20 to $35." -- Baker, Phillips. and Powell (1995) in a review article on splits

Stock splits have long been a fixture in US financial markets. Of the stocks listed on the New York Stock Exchange (NYSE) at the end of 1930, almost 20% split at least once in the prior decade. Over a half century later, in a typical year, 5 to 10% of all NYSE firms announce stock splits. A related, though less noted, feature of US financial markets is that average stock prices remain remarkably constant over time. Over the last fifty years, the average NYSE share price has stayed in the $30 to $40 dollar range despite huge increases in consumer prices and corporate equity values. Stock splits are major contributors to such constancy of prices.

Despite their longevity, splits have long puzzled finance theorists. After all, splits are at one level only cosmetic changes, slicing the same pie into smaller pieces but not changing an investor's fractional ownership of the equity interest and votes in the company.

In this paper, we examine the link between splits and share price at the firm level. First, we show that a large proportion of the cross-sectional variation in split prices (price to which a stock splits) can be explained by readily available public information. A significant contributing factor is the stock price level after a firm's last split. Managers appear to engineer splits to return their company's share price to a particular level that is remarkably stable over time. This role for the lagged split price has not been incorporated in prior studies. Second, we take advantage of these regularities in split prices to construct new tests to discriminate between the information and liquidity effects of splits. Using our findings on patterns in split prices, we use public information to estimate an expected split factor for a company. This anticipated factor should reflect liquidity concerns and the average information effect of splits. Unlike prior studies that investigated share returns around split announcements, our approach develops a firm-specific measure of expectations, rather than looking only at the absolute level of share price or split factor. We find that abnormal returns to shareholders are significantly higher when management announces a larger-than-anticipated split factor.

We also find that analysts increase earnings forecasts significantly more when managers announce a split factor larger than anticipated. Unlike share returns, which may be driven by both information and transactions-cost factors, earnings forecasts are direct predictions of corporate performance. Overall, our findings are consistent with splits signaling information to investors.

The evidence suggests an important role for the level of the share price that is typically overlooked in financial models of corporate and investor behavior. One possible rationale for the importance of historical prices is that they capture firm-specific market microstructure factors that are stable over time. For instance, Angel (1997) argues that a firm's optimal price can be understood in terms of maintaining an optimal relative tick size, which itself depends on firm-specific characteristics. Following this train of thought, our evidence can be viewed in terms of management decisions to split below an optimal "liquidity-based" price in order to convey especially positive news to the market. This logic follows the spirit of signaling interpretations of splits (e.g. Brennan and Copeland, 1988). Our results and prior research (e.g. Angel, 1997; and Schultz, 1997) provide limited support for microstructure explanations but such factors do not appear to explain fully the role of lagged split price. A more behavioral interpretation of an optimal price level is that investors frame their decisions about stocks in terms of dollar prices rather than percentage returns. Such framing behavior is well documented in many areas of human conduct and may help explain the puzzling phenomenon of mutual fund stock splits.

A third interpretation of our evidence is that investors (and analysts) have no preconceived preferred price level but simply learn from a firm's past split behavior. For instance, Pilotte and Manuel (1996) provide evidence that investors use a firm's previous post-split earnings performance to interpret a newly announced split. Following this interpretation, suppose that managers think there is an optimal price and split their stock to keep the price at that level over time. Investors then can infer managers' private information based on the announced split price and react to that information. Management belief in a preferred price level is overwhelmingly supported by survey data (e.g., Baker and Powell, 1993) even if the motivation for that belief is less clear.' [1]