A Computable Dynamic Oligopoly Model of Capacity Investment

We analyze a class of dynamic models that has several recent applications, where each period, each firm receives a private shock to the marginal cost of investment and chooses among many ordered capacity levels. Simulation methods to compute these models can result in non-existence of pure strategy equilibrium, while including multinomial shocks leads to counterintuitive predictions. We provide a computationally fast method to calculate optimal investment probabilities given value functions that iteratively finds investment levels chosen with positive probability and cutoffs of private information shocks across choices. Our method is useful to practitioners seeking to estimate models with continuous firm choices.