Common factors, principal components analysis, and the term structure of interest rates

Abstract

This paper studies common factor structure of bond returns from the US, UK and Germany. We estimate factors using both principal components analysis and common principal components analysis (CPCA), and construct factor mimicking portfolios to provide interpretations for some of these factors. A regression analysis of these portfolios shows that the common factors relate mostly to the level and steepness of the term structure in the US, with the first common factor explains approximately 90% of the variation. We use simulations to show that the power of the CPC test statistic to detect similarities in the factor structure which comprises our sample is limited.