Risk Premia in the Yield Curve

I study the sources of risk premia associated with the level bond portfolio by utilizing an international panel of zero-coupon bond data. I replicate a portion of ‘Yield Curve Premia’ by Brooks et al. who utilize principal component analysis to represent the moments of the yield curve and assess the efficacy of asset pricing factors commonly used in equities in explaining variation in bond returns.. I extend the work done in Yield Curve Premia by employing the partial least squares regression procedure in place of principal component analysis. I find that the level, slope, and curvature result is incredibly robust, not only across countries but also across dimensionality reduction methods. To assess the out-of-sample forecasting power of the partial least squares factors, I construct a trading rule using a predictive regression model and find varying return premia across countries in the panel.