Bayesian versus Maximum Likelihood Estimation in DSGE Modelling

This paper is intended to render some tips relating to the circumstances and benefits of resorting to Bayesian versus maximum likelihood estimation in solving dynamic stochastic general equilibrium models. By using beta, gamma, inverse gamma or normal priors, the Bayesian approach allows for more grounded likelihood estimation, overcoming the often incorrect uniform priors specific to MLE. Besides, the same priors' setting avoids the cases of weak or non-identification of models and leads to risk minimisation, the latter being one of the major issues to be considered in any economic analysis.