When companies announce that they may buy back their own shares in the market, stock prices increase by about 3% on average. The most commonly accepted explanation for this abnormal return is that by announcing a buyback program, firms signal that their shares are undervalued. However, there are two problems with this ''signaling'' hypothesis. First, share buyback programs are not firm commitments, i.e, companies can decide not to buy back any shares at all, so the costs of ''false signaling'' are not clear. Second, during the last three years (1994 to 1996) more than 3,000 companies have announced open market share repurchase programs for more than $300 billion. It is difficult to imagine that so many managers would find their stock undervalued at the same point in time. In this paper we provide an alternative explanation. We argue that 1) companies only buy back stock when their shares are undervalued and 2) by authorizing an open market share repurchase program, the board provides the management with an option to engage in ''insider'' trading with company funds. As a result, we prove that share buyback program announcements should increase stock prices to reflect the value of these expected ''insider'' trading benefits. Even if managers have no superior information at the time of the announcement of the buyback authorization, stock prices should increase. Although the repurchase decision has no impact on the total operating cash flows of the firm, our model does not assume a free lunch. The crucial insight is that, eventually, after the repurchase option expires, the stock price is determined by the cash flows per share. The ability of the firm to reduce its shares outstanding when they trade below true value is the source of the option value. In contrast to the predictions of the signaling literature, our paper argues that companies want to avoid signaling in order to be able to buy back shares below the true value. This can be achieved by systematically authorizing open market programs in advance of any perceived mispricing. The conclusion that the value of the firm is larger than the value of its expected operating cash flow is similar to the conclusions of the ''real'' options literature. Specifically, if a company has the option to defer investing in a project until more accurate information about the project's cash flow becomes available, the value of the project is larger than its net present value (i.e., the value of the project if the investment decision is made immediately). Similarly, ifa company, acting in the interest of its long-term shareholders, can wait for better information before choosing to buy shares from liquidity traders, the value of such a repurchase program to longterm investors becomes larger than its (zero) net present value. We model the buyback option as an option to exchange the market price of the stock for the true value. The value of this option increases with the volatility of the stock, the fraction of shares the company is authorized to buy back, and the maturity of the option. The option value is negatively related to the correlation between the ''true'' return and the return on the stock observed in the market (which can be interpreted as a measure of market efficiency). Depending on the assumptions one is willing to make about the volatility of the stock, the fraction of shares the company is authorized to buy back, the maturity of the option, and the degree of market efficiency, the option effect has the potential to explain a significant fraction of the announcement returns observed after announcements of buyback programs. For example, if a company announces that it may buy back 10% of its own shares during the next two years and 1) the volatility of its own stock is 40% per year and 2) the correlation between the true return and the (observed) market return is 60%, then our model predicts that the company's stock price should increase by 3% when this buyback authorization is announced. Employing a sample of 892 open market share repurchase announcements, we find results consistent with the theory: announcement returns are positively correlated with the fraction of shares the company may buy back, the volatility of the stock, and a proxy for the extent of mispricing in the future. Although these results are also consistent with the predictions of the signaling literature, we believe that our theory better explains why thousands of firms are announcing buyback programs and why many of them are not completed. Hence, our model suggests that, as repurchase programs contain valuable options and are relatively simple to establish, all companies should adopt them. Recently, this has indeed been the case. In 1996, Securities Data Corporation reports more than 100 share repurchase program announcements per month. As repurchase programs often have maturities of several years or more, we estimate that currently thousands of US firms have in place the option to buy back their own shares.
