Sequential Learning

Two players sequentially and privately examine a project of unknown quality. Launching the project requires mutual consent and the first player values the project more than the second player does. The combination of the conflict of interest and private learning leads to moral hazard. We show that an efficient equilibrium must take one of two forms as a function of the prior: either one player relinquishes control of the project, thereby rendering the collaboration moot, or the first player occasionally makes false claims about achieving positive findings. In the latter case, the players' relevant beliefs diverge as time progresses. In addition, we show that projects for which an initial examination failed to generate positive findings may be launched, and that projects known to be good by the first player may be delayed or even aborted.