A Dynamic Model of Predation

We study the feasibility and profitability of predation in a parsimonious infinite-horizon, complete information setting where an incumbent may face an entrant, in which case it needs to decide whether to accommodate or predate it. If the entrant exits, a new entrant is born with positive probability. We show that there always exists a Markov perfect equilibrium, which can be of three types: accommodation, predation with no future entry, and predation with hit-and-run entry. We use the model to study alternative antitrust policies, derive the best rules for these policies, and compare their welfare effects.