Estimación de modelos de equilibrio general en economías dinámicas por métodos de Monte Carlo y Cadenas de Markov

this paper describes a general procedure to do bayesian inference based on the likelihood evaluation of the stochastic general equilibrium models (mege) through markov chain monte carlo methods (mcmc). the proposed methodology involves log linearizing the model, transformed into state space form, then use the kalman filter to evaluate the likelihood function and finally apply the metropolis hastings algorithm to estimate the posterior distribution parameters. technique is illustrated using the stochastic growth of basic model, considering quarterly data on the venezuelan economy between the first quarter of (1984) through the third quarter of (2004). the empirical analysis made allows us to conclude that the algorithms used to estimate the model parameters work efficiently and low computational cost, the estimates obtained are consistent, that is, estimates of the predictions adequately reflect the behavior of the product, employment, consumption and investment per capita in the country. the graphs of the estimated histograms show bimodal and skewed distributions