Modified Marginal Expected Shortfall under Asymptotic Dependence

J. -J. Cai,V. Chavez-Demoulin,A. Guillou
DOI: https://doi.org/10.1093/biomet/asx005
IF: 3.0279
2017-01-01
Biometrika
Abstract:SUMMARY We propose an estimator of the marginal expected shortfall by considering a log transformation of a variable which has an infinite expectation. We establish the asymptotic normality of our estimator under general assumptions. A simulation study suggests that the estimation procedure is robust with respect to the choice of tuning parameters. Our estimator has lower bias and mean squared error than the empirical estimator when the latter is applicable. We illustrate our method on a tsunami dataset.
What problem does this paper attempt to address?