Abstract:
In this paper we adopt a principal components analysis (PCA) to reduce the dimensionality of the term structure and employ autoregressive models (AR) to forecast principal components which, in turn, are used to forecast swap rates. Arguing in favor of structural variation, we propose data driven, adaptive model selection strategies based on the PCA/AR model. To evaluate ex-ante forecasting performance for particular rates, different forecast features such as mean squared errors, directional accuracy and big hit ability are considered. It turns out that relative to benchmark models, the adaptive approach offers additional forecast accuracy in terms of directional accuracy and big hit ability.
