Title: A Bayesian nonparametric approach on model combination for short-term interest rates
Authors: John Maheu - McMaster University (Canada)
Qiao Yang - ShanghaiTech University (China) [presenting]
Abstract: The dynamics of short-term interest rates are important input into pricing models of the term structure of interest rates. Most of the works have focused on adding extra components into the model for improving forecast accuracy. We take another direction which uses the model combination framework to improves the forecast performance. We introduce Bayesian nonparametric approach to extend a previous one from fixed number states into infinite dimension and combine popular discrete time short-rate models. The new approach shows significant improvement in density forecasts to existing approaches and strong evidence of interest rated model dynamics are documented.