Value at risk computation by means of principal component analysis: the european case

VaR analysis for fixed income securities requires a preliminary mapping phase which allows to decompose actual trading positions into virtual exposures to several chosen key rates. The question arises whether a more parsimonious method can be found. The goal of this paper is to make an attempt in this direction. We initially test a method that exploits the possibility to depict the historical dynamics of interest rates with just few latent factors and which is also able, by means of a simulation approach, to provide plausible scenario analysis for the term structure. We also propose a new solution to express bonds' sensitivity to the extracted principal components which makes simulating bond portfolio value changes in response to many generated interest rate scenarios extremely simple and quick.