“…Moreover, several analyses quantify estimation risk through determining confidence intervals for Value-at-Risk estimates under various model assumptions, i.e., the unconditional normal distribution, historical estimates, conditional GARCH models with different innovation assumptions such as normal or Student t errors, the Hill estimator, filtered historical simulation, Gram-Charlier and Cornish-Fisher expansion (see Pritisker, 1997;Christoffersen and Goncalves, 2005;Chan et al, 2007). Besides this type of model risk quantification, Gourieroux and Zakoyan (2013) and Boucher et al (2014) examine model risk through adjustment factors which need to be added to the original risk measure estimate in order to derive the desired behavior.…”