Bayesian Forecasting using Stochastic Search Variable Selection in a VAR Subject to Breaks

This paper builds a model which has two extensions over a standard VAR. The first of these is stochastic search variable selection, which is an automatic model selection device which allows for coefficients in a possibly over-parameterized VAR to be set to zero. The second allows for an unknown number of structual breaks in the VAR parameters. We investigate the in-sample and forecasting performance of our model in an application involving a commonly-used US macro-economic data set. We find that, in-sample, these extensions clearly are warranted. In a recursive forecasting exercise, we find moderate improvements over a standard VAR, although most of these improvements are due to the use of stochastic search variable selection rather than the inclusion of breaks. Classification-JEL: