A repeated game of IPO underwriting

Most of the recent research in the IPO literature assumes that a group of repeated investors has an informational advantage over every other player. Here, a model is built on the alternative hypothesis, that a risk averse entrepreneur and a risk neutral underwriter have a more precise estimate of the expected value of the stock than any investor. In a one shot game, there is adverse selection and only lower value firms are able to go public in a partially pooling equilibrium. In a repeated game, a group of investors will be able to face the same underwriter and commit on how to play in future IPOs before the underwriter becomes informed about the IPOs expected value. Under certain conditions and under the assumption that there are no side payments in cash, the repeated investors can offer the underwriter a direct revelation mechanism, which can be implemented in trigger strategies. The mechanism incentives are instead provided in terms of a continuation value of the game for the underwriter. When the underwriter plays under the mechanism, both lower and higher value firms find it in their interest to have the underwriter certifying their value. The range of firm types going public becomes broader and firms separate. When the repeated investors are able to coordinate their actions the model yields the expected IPO discount result. The underwriter needs to give allocation priority to the repeated investors in order to expose herself to their punishment and gain credibility. If side payments in cash are allowed, the position of the repeated investors and underwriter can be improved by allowing the underwriter to partially compensate the investors for the loss in the case of a bad realization of the stock price or forego the continuation value of the game with some probability. The first alternative corresponds to the case of market stabilization. Optimally, the underwriter will choose to restrict the benefits of market stabilization to repeated investors, yielding the discrimination result observed in the data. Unlike recent theoretical research the model provides an explanation for why the underwriter does not disclose her market stabilization efforts.