A Sieve‐SMM Estimator for Dynamic Models

This paper proposes a Sieve Simulated Method of Moments (Sieve‐SMM) estimator for the parameters and distribution shocks in nonlinear dynamic models where likelihood moments are not tractable. An important concern with SMM, which matches sample simulated moments, is that parametric required. However, economic quantities depend on this distribution, such as welfare asset prices, can be sensitive to misspecification. The Sieve‐SMM addresses issue by flexibly approximating Gaussian tails mixture sieve. asymptotic framework provides consistency, rate convergence, normality results, extending existing results new more general dynamics latent variables. application pricing production economy shows large decline estimates relative risk aversion, highlighting empirical relevance misspecification bias.