Backtesting Expected Shortfall: A simple multinomial test

Release:2016-06-20 Viewed:77


Theme:Backtesting Expected Shortfall: A simple multinomial test

Time & Place: March 18, 201616:00  The forth lecture hall

Lecturer:Marie Kratz, ESSEC Business School, CREAR (risk research center)

Abs:Risk measure has become a fundamental tool to compute the solvency of financial institutions. Therefore, it has caused many controversies regarding the choice of the best measure to use in practice. Although Expected Shortfall (ES) has been widely accepted as a risk measure that is conceptually superior to Value-at-Risk (VaR) because of its mathematical properties, it has also been proved to be non elicitable (Gneiting, 2011), leading to less straightforward backtesting methods than for VaR. Nevertheless, it satisfies the property of conditional elicitability, a concept introduced in Emmer et al. (2015). A popular backtesting procedure for VaR is a binomial test, based on a violation process. Following the idea by Emmer et al. (2015) of considering an empirical approach that consists in replacing ES by a set of a small number of quantiles for the backtesting, comes the natural proposition of a new simple multinomial backtesting test for ES, as handy as for VaR. This talk is based on two studies: one (in Journal of Risk, 2015) with S. Emmer (Crédit Suisse, Zurich) & D. Tasche (PRA - Bank of England), the other with Y. Lok & A. McNeil (Heriot Watt Univ., Edinburgh).