Abstract:
|
In this paper, an improved Bayesian smoothing spline (BSS) model is developed to estimate the term structure of Chinese Treasury yield curves. The developed BSS model has a flexible function form which can model various yield curve shapes. As a nonparametric method different from the penalized splines, the BSS model does not need to choose the number and locations for the knots since all distinct design points are treated as knots. Furthermore, the BSS model obtains the smoothing parameter as a by-product that does not need to be estimated. A dimension reduction procedure is developed to guarantee the feasibility of the inverse matrix computation when implementing the BSS model. The BSS model are compared with two versions of the Nelson-Siegel model, two versions of the Svensson extension model, and the penalized spline model. Both the in- and out-of-sample results strongly support the assertion that the BSS model fits the Chinese Treasury yield curves better than all the other methods. More specifically, The BSS model has the smallest yield error and the largest hit rate for both in-sample and out-sa
|