Relationships Among Financial and Commodity Markets on Economic Growth: New Evidence from Bayesian Estimation of Seemingly Unrelated Time Series Equations

This paper uses Bayesian estimation of the Seemingly Unrelated Time Series Equations (BSUTSE) model to present the relationships of explanatory variables on economic growth. From the Bayesian model average (BMA) analysis, the total considered independent variables is eleven. It was found that there are three variables that can explain the best model of yield curves and real GDP that consists of gold price, and stock from industrial, resource, financials, property and construction, and technology. The BSUTSE presents time-varying coefficients estimation on each economic growth variable. For the yield curve, most variables can explain economic growth more than ever. For real GDP, the return from property construction, industry, and resources significantly explains economic growth and changes in the opposite direction with real GDP. Gold price and the financial industry can explain changes in the same direction with real GDP. Only coefficients from technology parameters have more explain real GDP in the same direction. The BSUTSE is the model that allows coefficients to vary in time in an equation system. The advantage of The BSUTSE model, it can estimate time-varying coefficients.