Functional Indirect Inference.

The class of paremetric dynamic latent variable models is becoming more and more popular in economics and finance. Dynamic disequilibrium models, latent factor models, switching regimes models, stochastic volatility models are only few examples of this class of models. Inference in this calss may be difficult since the computation of the likelihood function requires to integrate out the unobservable components and to calculate very high dimensional integrals. We propose an estimation procedure which could be applied to any dynamic latent model.The approach is based on the Indirect Inference principle and considers as binding functions conditional expectations of functions of the endogenous variable, given past values of this variable. These conditional expectations are estimated by a nonparametric approach.