A Model for Measuring Financial Risks

In this paper, after reviewing the regulatory conditions for the use of internal models, such as the specification of market risk factors, quantitative standards (including backtesting) and stress testing, we put forward a general continuous-time model. This is used to determine measures of volatility and «turbulence» that serve as synthetic instruments for monitoring the riskiness of a portfolio. We then look at the significance of value-at-risk (VAR) and propose an analytical approximation. A discussion follows of the reasons (including the empirical evidence contrary to the assumption of normality) which led the Basle Committee to prescribe a multiplicative factor for VAR in determining the capital requirement for market risks. VAR is then compared with the value of a put option guaranteeing a minimum level of shareholder's equity. This leads to a call for the present regulations to be developed in the direction of an integrated approach that will determine the value of a portfolio and its volatility as a function of the various financial and credit factors (and of the concentration of exposures towards given sectors/ borrowers). The paper concludes with the presentation of some applications showing how the proposed approach can be implemented in practice.