اجرای بهینه برنامه خرید مجدد بازار آزاد سهام

This paper formalizes the following intuition about open-market share repurchases. Firms do open-market share repurchases to return free cash, which would otherwise be wasted. However, when the firm goes to buy its own shares with this cash, it has inside information and hence the actual execution is characterized by adverse selection. The market knows that the firm has inside information, and consequently the ask price is high to compensate for this adverse selection problem. This implies that, all else equal, the greater the adverse selection problem compared to the cash waste problem, the higher the ask price, and, therefore, the wider the bid–ask spread and the lower the share repurchase completion rate. We test this implication on a sample of U.S. firms and report evidence consistent with the model.

Over the last two decades, announcements of open-market stock repurchase programs (henceforth, “open-market programs”) have become common practice (see, for example, Grullon and Michaely, 2002). Yet, empirical evidence suggests there is great variability associated with their execution. First, there is great variability documented about actual completion rates. In the United States, Stephens and Weisbach (1998) and Jagannathan et al. (2000) document that average actual repurchase rates are only 70–80%. Frequently only a small fraction of the quantity announced is actually repurchased, and many announcing firms do not repurchase at all. Chan et al. (2005) report similar results. Actual repurchase rates are even lower outside the U.S. Ikenberry et al. (2000) find average actual repurchase rates to be as low as 28% in Canada, and Rau and Vermaelen (2002) find average actual repurchase rates of only 37% in the United Kingdom.

In this paper we perform a theoretical investigation of the execution of open-market stock repurchase programs. Our model suggests that the execution depends on the agency costs of free cash that are realized, and information asymmetry that develops only after the program is announced. The results highlight important properties of open-market stock repurchase programs that other payout methods (e.g., self-tender-offers, dividends) do not have: they give the firm the option to avoid payout when cash is needed for operations, yet they also disburse free cash as long as the stock is not severely overpriced. Because they are performed at management discretion, however, repurchase programs also redistribute wealth among shareholders. The model generates predictions about the completion rate of repurchase programs and about the bid–ask spread during the repurchase period that might explain inconsistencies in earlier empirical studies. One new and testable prediction of our theory is that program completion rates and bid–ask spreads are negatively correlated. Tests using a sample of U.S. firms yield evidence consistent with this prediction.