Many portfolio strategies are “event driven” (i.e., designed to
benefit from price movements caused by corporate events, such as a merger).
These strategies involve payoffs with discontinuous and skewed distributions
that conventional risk methods do not measure well. This article develops
methods to measure the forward-looking risk, based on current positions, of
portfolios exposed to such discrete events. The method is applied to independent
events and to the more realistic case of events that are not independent. For
mergers and acquisitions, empirical estimates of deal-break correlations are
positive but low, which implies that most of this event risk is idiosyncratic
and diversifiable. The methodology allows assessment of the risk and return of
various portfolio structures for event-driven funds.