Value at Risk (VaR) has achieved a pre-eminent place when quantifyin portfolio market risk. VaR calculation can be carried out using several models, with statistical likelihood being the key feature. In this paper we propose a new model for VaR estimation, namely the ARCH factor methodology. We apply this methodology on a set of foreign exchange risk exposed portfolios and test the previously assumed statistical confidence level. Furthermore, we compare our results with those obtained through J.P. Morgan’s Riskmetrics methodology utilisation.

Key words: ARCH models, value at risk, foreign exchange risk.
JEL classification: F31.