A Semiparametric Estimator for Dynamic Optimization Models

We develop a new estimation methodology for dynamic optimization models with unobserved state variables Our approach is semiparametric in the sense of not requiring explicit parametric assumptions to be made concerning the distribution of these unobserved state variables We propose a two-step pairwise-difference estimator which exploits two common features of dynamic optimization problems: (1) the weak monotonicity of the agent's decision (policy) function in the unobserved state variables conditional on the observed state variables; and (2) the state-contingent nature of optimal decision-making which implies that conditional on the observed state variables the variation in observed choices across agents must be due to randomness in the unobserved state variables across agents We apply our estimator to a model of dynamic competitive equilibrium in the market for milk production quota in Ontario Canada