Estimation of Extreme Risk Measures from Heavy-tailed distributions

International audienceValue-at-risk, conditional tail expectation, conditional value-at-risk and conditional tail variance are classical risk measures. In statistical terms, the value-at-risk is the upper α-quantile of the loss distribution where α ∈ (0, 1) is the confidence level. Here, we focus on the properties of these risk measures for extreme losses (where α → 0 is no longer fixed). To assign probabilities to extreme losses it is assumed that we are in the case of heavy-tailed distributions. We also consider these risk measures in the presence of a covariate. Let us note that the presence of a covariate has already been investigated in extreme value theory. The main goal of this communication is to propose estimators of the above risk measures in the case of heavy-tailed distributions, for extreme losses, and to include a covariate in the estimation. The asymptotic distribution of our estimators is established and their finite sample behavior is illustrated on simulated data and on a real data set of pluviometrical measurements