Regional Economic Forecasting Model Based on the Bayesian Vector Autoregrssion:Evidence from Qinghai Province

Because of insufficient data observations and other reasons,in China,the traditional regional economic forecasting models can not achieve the expected forecasting results.The Bayesian Vector Autoregrssion(BVAR) model introduces variables' statistical properties as parameter's prior distribution in the traditional VAR model,solving the problem of lacking degree of freedom.Based on Qinghai province's empirical data,this paper illustrated the application of BVAR model,in which national GDP and transfer payment from central authority were introduced as exogenous variables to reflect the economic linkage between national economy and regional economy.In-sample and out-of-sample forecasting error comparison and accurate forecast of the turn point of Qinghai economic growth show that the BVAR forecasting model outperforms other alternative models.