Persistence and Reversal in Herd Behavior: Theory and Application to the Decision to Go Public

We model rational herd behavior when the underlying value changes over time, with payoffs that are either dependent or independent of the underlying value. We show that herding does not last forever and is not monotone in signal quality. High correlation among agents' actions does not necessarily imply herding. This suggests alternative empirical methods are needed to detect herding. The model has many applications, including the IPO decision in which payoffs are state dependent. The model implies that the decision to go public is more likely associated with herding than the decision to delay an IPO. Copyright 2002, Oxford University Press.