<html><head><meta http-equiv=Content-Type content="text/html; charset=UTF-8"><META name="Author" content="GroupWise WebAccess"><style type="text/css">
body p
{
margin: 0px;
}
</style></head><body style='font-family: Helvetica, Arial, sans-serif; font-size: 13px; '><div id="GroupWiseSection_1592231643000_pit.goetz@wiwi.uni-halle.de_1B992E0002C000009867776461663435_" class="GroupWiseMessageBody"><div><div>Hello everyone,</div><div><br></div><div>I work at a university in
germany and we are also currently working on forecasting ES and (of
course) backtesting of said forecasts.</div><div><br></div><div>Over the last few months some students, who are writing their masters thesis at our chair, had to some litarature research.</div><div>Thats why I wanted to give you a very brief overview of their findings:</div><div><br></div><div>The
most widely applied ES backtests seems to be the backtest by McNeil,
Frey and Embrechts (2000), implemented for example in the rugarch
package.</div><div>(the test was already mentioned here by Alexios)</div><div><br></div><div>In addition to the already mentioned tests and the paper by Acerby and Szekely I wanted to add the following:<br></div><div><br></div><div>A
Hitsequence based backtest was introduced for by Du, Escanciano (2017).
As far as I am concerned, this test has not yet been implemented in a
package, but their code is available online. In a broader view, this
test is a special case of a spectral measure test by Costanzino, Curran
(2014), which was then extended to a Basel-Like traffic light approach
in 2018 (Not sure about the availability of code).</div><div><br></div><div>In
Emmer et al. (2015) it is suggested, that a suitable ES forecast can be
approximated by only 4 different VaR forecasts. This also suggests,
that you can backtest ES, forecasted by a model that forecasts both, ES
and VaR, such as GARCH, by backtesting th 4 different VaR forecasts.</div><div>However this approach seems to need more empirical valuation.</div><div><br></div><div>I
also wanted to mention the paper by Gneiting (2011), showing that the
ES lacks elicitability property. This can lead to complications, when
you try to backtest the ES itself as a point forecast.However, this
property can be used to construct a model comparison like backtest as in
Fissler et al. (2015).</div><div><br></div><div>More reacently, a
quantile regression based approach has been suggested by Coupier,
Leymarie (2020). I have not yet read said paper and therefore I can not
tell you anything about it.</div><div><br></div><div>I hope that this message gives you some new insights and some usefull information.</div><div><br></div><div>Best regards,</div>Pit</div><span> </span><span class="GroupwiseReplyHeader"><br><br><br/><div style='clear: both;'><div><p class="MsoNormal" style="font-size: 11.0pt;font-family: Calibri , sans-serif;"><span style="font-size: 9.0pt;color: gray;">Research Associate</span></p><p class="MsoNormal" style="font-size: 11.0pt;font-family: Calibri , sans-serif;"><b><span style="font-size: 9.0pt;color: gray;">Martin-Luther-Universität Halle-Wittenberg</span></b><span style="font-size: 9.0pt;color: gray;"></span></p><p class="MsoNormal" style="font-size: 11.0pt;font-family: Calibri , sans-serif;"><span style="font-size: 9.0pt;color: gray;">Chair of Finance & Banking</span></p><p class="MsoNormal" style="font-size: 11.0pt;font-family: Calibri , sans-serif;"><span style="font-size: 9.0pt;color: gray;"> </span></p><p class="MsoNormal" style="font-size: 11.0pt;font-family: Calibri , sans-serif;"><span style="font-size: 9.0pt;color: gray;">Große Steinstraße 73 | D-06108 Halle | Germany</span><span style="font-size: 9.0pt;"><br><span style="color: gray;">Tel 0049 345 5523452</span></span></p></div> </div><br/>>>> Daniel Cegiełka <daniel.cegielka@gmail.com> 10.06.20 21.49 Uhr >>><br></span>śr., 10 cze 2020 o 21:14 alexios galanos <alexios@4dscape.com> napisał(a):<br>><br>><br>><br>> On 6/10/20 11:08 AM, Daniel Cegiełka wrote:<br>> > śr., 10 cze 2020 o 19:23 Brian G. Peterson <brian@braverock.com> napisał(a):<br>> >><br>> >> On Wed, 2020-06-10 at 15:08 +0530, Christofer Bogaso wrote:<br>> >>> I was looking for an idea how banks backtest their models for<br>> >>> Expected<br>> >>> Shortfall. Backtesting VaR is well documented but I failed to get any<br>> >>> practical idea about backtesting ES.<br>> >>><br>> >>> Any pointer towards the best practice will be really helpful.<br>> >><br>> >> If you are using Normal VaR, then you know the Expected Shortfall<br>> >> estimate too.<br>> >><br>> >> If you are using a different mechanism, then of course the mean loss<br>> >> when the loss exceeds the VaR may be significantly different than the<br>> >> Normal ES.<br>> >><br>> >> So, to backetesting... the newest Basel standard replaces VaR with ES,<br>> >> and requires that banks justify their use of a particular ES model that<br>> >> they are using to calculate required regulatory capital.<br>> ><br>> > In my opinion, there is one aspect that introduces some confusion. ES<br>> > (CVaR) is now common, but many people, perhaps out of habit, maybe for<br>> > historical reasons, still use the term VaR instead of the correct name<br>> > (ES).<br>><br>> Not sure I follow. VaR and ES are different measures. VaR is a<br>> quantile while ES is the average loss conditional on that quantile<br>> (i.e. the expected loss conditional that the loss is greater than<br>> the quantile of the loss distribution).<br><br>I agree that these names should not be confused. However, I<br>encountered that the _name_ "VaR" is used for ES. In my opinion, this<br>is due to a mental shortcut, or it's a historical habit. Such<br>imprecise use of the names often leads to misunderstanding.<br><br>Daniel<br><br>> Regards,<br>><br>> Alexios<br>><br>> ><br>> > Best regards,<br>> > Daniel<br>> ><br>> ><br>> >> Regards,<br>> >><br>> >> Brian<br>> >><br>> >><br>> >> --<br>> >> Brian G. Peterson<br>> >> ph: +1.773.459.4973<br>> >> im: bgpbraverock<br>> >><br>> >> _______________________________________________<br>> >> R-SIG-Finance@r-project.org mailing list<br>> >> https://stat.ethz.ch/mailman/listinfo/r-sig-finance<br>> >> -- Subscriber-posting only. If you want to post, subscribe first.<br>> >> -- Also note that this is not the r-help list where general R questions should go.<br>> ><br>> > _______________________________________________<br>> > R-SIG-Finance@r-project.org mailing list<br>> > https://stat.ethz.ch/mailman/listinfo/r-sig-finance<br>> > -- Subscriber-posting only. If you want to post, subscribe first.<br>> > -- Also note that this is not the r-help list where general R questions should go.<br>> ><br><br>_______________________________________________<br>R-SIG-Finance@r-project.org mailing list<br>https://stat.ethz.ch/mailman/listinfo/r-sig-finance<br>-- Subscriber-posting only. If you want to post, subscribe first.<br>-- Also note that this is not the r-help list where general R questions should go.<br></div></body></html>