La Value-at-Risk: Modèles de la VaR, simulations en Visual Basic (Excel) et autres mesures récentes du risque de marché

Since the end of the nineties, Basle Committee has required that banks compute periodically their VaR and maintain sufficient capital to pay the eventual losses projected by VaR. Unfortunately, there is not only one measure of VaR because volatility, which is a fundamental component of VaR, is latent. Therefore, banks must use many VaR models to compute the range of their prospective losses. These computations might be complex because the distribution of high frequency returns is not normal. This article analyses many VaR models and produces their programs in Visual Basic. It considers also other new measures of market risk and the use of copulas and Fourier Transform for the computation of VaR.