Mean-CVaR vs Mean-Variance portfolio comparison

Discussion relating to general questions on Market Risk Analysis
Forum rules
DISCLAIMER: We do not warrant or represent that this forum or its content is free of viruses, worms or other code that might be contaminating or destructive. We cannot guarantee that documents or files downloaded from the Site will be free from viruses and we do not accept any responsibility for any damage or loss caused by any virus. Accordingly, for your own protection, you must use virus-checking software when using the forum. You must not post or provide to us via the forum, any document or file which you believe may contain a virus. You must virus check any document or file which you intend to post or provide to us via the forum. You must ensure that any document or file you intend to post to the forum does not contravene any applicable laws or contravene any person's legal rights. We do not accept any responsibility for any damage or loss you may suffer.
kelyos
Posts: 5
Joined: Sun Jul 29, 2012 9:11 pm

Mean-CVaR vs Mean-Variance portfolio comparison

Postby kelyos » Mon Jul 30, 2012 2:57 pm

"Thesis will concern the Conditional Value at Risk of a portfolio.It will evaluate the benefits of considering minimising CVaR compared to minimising portfolio variance"

I'm struggling a little bit since a week with this subject and I don't really know yet how to start with that. I have looked some paper from uryasev etc, so I understand basis of CVaR (equation..) and also Mean Variance but I'm thinking on how to compare each other.

Also I've already implement the return over 1 year in the past and put both method in a mathematical programming software called AMPL.

Is someone could help me with some idea on how to compare after implementation could be very helpful for me.
Thank you

coalexander
Posts: 815
Joined: Sun Sep 28, 2008 10:30 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby coalexander » Tue Jul 31, 2012 6:46 pm

Are you looking for out-of-sample portfolio evaluation methods, eg Sharpe Ratio and other performance measures?

Carol

kelyos
Posts: 5
Joined: Sun Jul 29, 2012 9:11 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby kelyos » Tue Jul 31, 2012 10:03 pm

What do you mean by out-of-sample ? The past asset return I have over 1 year is it not enough ?

Thank you

coalexander
Posts: 815
Joined: Sun Sep 28, 2008 10:30 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby coalexander » Wed Aug 01, 2012 7:27 am

No, you need about 10 years of daily returns. Otherwise, the out-performance of model A over model B is just for a specific market ;period, and the performance ranking could change over a different sample period.

Even with 10 years, the 'best' model could have won by chance. Hence, DATA SNOOPING checks need to be made like White's REALITY CHECK

An ex-post analysis (which model did best using historical data) is no good. You need an ex-ante analysis ---use some data to decide on allocations, then pretend you don't know the future and see how model performed on next bit of data. That is out-of-sample.

Carol

kelyos
Posts: 5
Joined: Sun Jul 29, 2012 9:11 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby kelyos » Wed Aug 01, 2012 9:08 am

Thank you Carole

Oh really but how can I decide which asset has to be chosen in portfolio ?

Also I don't know how to sample without data or this year (july 2011 to july 2012) could be fine ?

That's strange because I read a lot of paper as to regard to that and most of them took their data from the past, so ex-post.

coalexander
Posts: 815
Joined: Sun Sep 28, 2008 10:30 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby coalexander » Wed Aug 01, 2012 10:25 am

Take a look at the introduction to portfolio theory in Volume I - I think it is chapter 5....

kelyos
Posts: 5
Joined: Sun Jul 29, 2012 9:11 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby kelyos » Wed Aug 01, 2012 2:58 pm

It is chapter 6

Thank you I will take a look.
Anthony

kelyos
Posts: 5
Joined: Sun Jul 29, 2012 9:11 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby kelyos » Sun Aug 26, 2012 1:41 pm

Hi Carole according my subject (Mean-CVaR vs Mean-Variance),

Do you think it is enough to do the backtesting in my in-sample part, that means number of exceedances is roughly the same as required by Basel to show that the model could be validated for the out-of-sample period ??

coalexander
Posts: 815
Joined: Sun Sep 28, 2008 10:30 pm

Re: Mean-CVaR vs Mean-Variance portfolio comparison

Postby coalexander » Sun Aug 26, 2012 2:16 pm

No, you should at least perform the conditional coverage tests of Christoffersen (these are explained with excel workbook in Vol IV). Furthermore, the mean-CVaR criterion should perform better in ETL tests because of its design, so also do ETL tests (as explained in Volume IV)

Hope this helps, Carol


Return to “General”

Who is online

Users browsing this forum: No registered users and 1 guest