Figure n°1 :

models. However such algorithms require a tractable form of the criterion functions to be optimized. In these lessons we are interested in econometric models analytical expression. In such problems the difficulty generaly arises from the presence of integrals of large dimensions, and the idea is ta circumvent this numerical difficulty by an approach based on simulations. After a survery of potential applications leading the criterion functions containing integrals, the lectures deal with the method of simulated moments, and similar approaches based on loglikelihood or least squares criteria. A different methodology (indirect inference) uses an auxiliary simple misspecified model which gives by itself inconsistent parameter estimates, but indirect inference uses simulation approaches to correct the asymptotic bias. Two chapters givecletailsof applications ta discrete choice and duration models, and to financial models.

   Retour au numéro