Scenario Based Principal Component Value-at-Risk: An Application to Italian Banks' Interest Rate Risk Exposure

Abstract

The paper develops a Value-at-Risk methodology to assess Italian banks' interest rate risk exposure. By using 5 years of daily data, the exposure is evaluated through a Principal Component VaR based on Monte Carlo simulation according to two different approaches (parametric and non-parametric). The main contribution of the paper is a methodology for modelling interest rate changes when underlying risk factors are skewed and heavy-tailed. The methodology is then implemented on a one year holding period in order to compare the results from those resulting from the Basel II standardized approach. We find that the risk measure proposed by Basel II gives an adequate description of risk, provided that duration parameters are changed to reflect market conditions. Finally, the methodology is used to perform a stress testing analysis.

Keywords: Interest rate risk, VAR, PCA, Non-normality, Non parametric methods

JEL Classification: C14, C19, G21