Dimensionality Reduction and State Space Systems: Forecasting the US Treasury Yields Using Frequentist and Bayesian VARs

Using a state-space system, I forecasted the US Treasury yields by employing frequentist and Bayesian methods after first decomposing of varying maturities into its unobserved term structure factors. Then, exploited model to forecast yields, compared performance each using metric - mean squared error, as loss function. Among methods, applied two-step Diebold-Li, principal components, one-step Kalman filter approaches. Likewise, imposed five different priors in VARs -Diffuse, Minnesota, natural conjugate, independent normal inverse-Wishart, stochastic search variable selection priors. After forecasting for 9 horizons, found that BVAR with Minnesota prior generally minimizes augmented above BVARs including macroeconomic variables constructed impulse response functions recursive ordering identification scheme. Finally, fitted sign-restricted dummy observations.