IPO Information Production, New Issue Quality and Participation Restrictions

We endogenize both 1) investment bankers' screening of new issues and 2) information production of investors, making ours the ̄rst model to characterize double-sided information production. The two sides are nontrivially linked ¡ if investors' information production were low, the banker would sell to an essentially uninformed pool. However, sale outcomes uncorrelated with true value would give no reason to screen out bad ̄rms in the ̄rst place if doing so is costly. This reasoning implies that the banker wishes to structure a sale conducive to investors' research; otherwise, his own moral hazard problem prevents high equilibrium prices. We provide a speci ̄c analytical sense in which this investor learning serves as a complement to bank reputation in ensuring that good quality ̄rms are taken public. As one application of this link between the sides of information production, we demonstrate that bigger pools of investors are not always optimal ¡ in contrast to most IPO models. As the number of investors rises, each investor does less evaluation. Although aggregate informativeness may go up or down as the pool grows, under certain conditions it goes to zero and the moral hazard problem results in a dramatic open sale failure. This may help explain a participation restriction puzzle: investment bankers do not open sales up to all bidders even though doing so would maximize competition and apparently limit underpricing.