A Practical Guide to Market Risk Model Validations (Part II - VaR Estimation)

Abstract

The VaR (Value at Risk) concept has emerged back in 1994 when JP Morgan started routinely using it in its daily reporting. Simply said, it represents a lower bound of large rare losses. The VaR metric became an industry standard for measuring market risk because it is intuitive and easy to interpret. This led to the adaptation of VaR for market risk capital calculations in the 1996 market risk amendment (also know as Basel II). Following the failure of this capitalization approach during the 2008 financial crisis, Basel Committee strengthened capital requirements by introducing stressed VaR (in Basel 2.5) and tail VaR (in Fundamental Review of the Trading Book) metrics. The VaR process involves a number of steps that include input processing, curve building, pricing, hedging, risk factor identification, simulation, and VaR estimation. In this paper, we will focus on the VaR estimation only. All VaR estimation models can be categorized by their simulation technique, simulation object and revaluation methodology. We will define a broad validation framework that includes assessment of the conceptual soundness and performance of various VaR models. Common modeling issues and practical solutions will be discussed as well.