Using Input-Output Information for Bayesian Forecasting of Industry Employment in a Regional Econometric Model

Bayesian estimation is used to incorporate regional input-output information into the employment block of a regional econometric model. The Bayesian approach used borrows both from previous work on embedding input-output information within econometric models and from Bayesian vector autoregression forecasting. The accuracy of out-of-sample forecasts produced by the alternative Bayesian models and some traditional forecast approaches are compared.