Novo Working Paper - Bayesian Factor Selection in Dynamic Term Structure Models

Bayesian Factor Selection in Dynamic Term Structure Models

Márcio Laurini (IBMEC Business School)

This paper discusses Bayesian procedures for factor selection in dynamic term structure models through simulation methods based on Markov Chain Monte Carlo. The number of factors, besides influencing the fitting and prediction of observed yields, is also relevant to features such as the imposition of no-arbitrage conditions. We present a methodology for selecting the best specification in the Nelson-Siegel class of models using Reversible Jump MCMC.