Topic:
No Exit*:
Failure to Exit Under Uncertainty

Delayed exit is a substantial economic problem.  Studies indicate if VCs exited ventures optimally, returns would triple, and if corporations divested underperforming business units, GDP would increase 13.6%.  A prevalent explanation for delayed exit is behavioral bias associated with escalated commitment.  In general however exit will be "delayed" even absent bias.  This arises from decision maker efforts to avoid Type I error while discovering the long run prospects of an endeavor.  Solutions differ depending upon which source of delay predominates.
            While exit delays are problematic in numerous contexts, we examine them in the simplest real-world context—entrepreneurial exit of banks.  Entrepreneurial exit comprises a well defined objective function (maximizing over choice over entrepreneurial profits versus wages) evaluated by a solitary actor with perfect incentives.  Thus the context avoids problems of a larger (and generally unknowable to the researcher) opportunity set, as well as goal conflict and/or incentive misalignment among a group of non-owner managers.  We decompose exit delay into a rational component (Bayesian updating under uncertainty) and behavioral bias.  We find most of the delay in this context is rational.  While we do find evidence of behavioral bias, results run counter to predictions for escalated commitment.