Optimal capital allocation: VaR, CVaR, spectral measures and beyond in Russian Market

1
Optimal capital allocation VaR, CVaR, spectral
measures and beyond in Russian Market

• A discussion by Branko Uroševic
• Faculty of Economics, University of Belgrade and
  NBS

2
What is the paper about

• This paper aims to compare and contrasts capital
  allocation based on the standard Markowitz
  mean-variance efficient frontier with optimal
  portfolios based on other risk measures
  mean-VaR,mean-CVaR and mean-generalized SRM
• It aims to do so for a portfolio of 25 most
  liquid stocks on the Russian stock exchange.
• In the process, the paper address a plethora of
  estimation issues including various volatility
  forecasting techniques etc, calibrated to the
  Russian market

3
Claimed contribution

• First large portfolio analysis with data from the
  Russian stock market (whatever that means)
• First large-scale comparison of the above methods
  of optimization that include data from the
  current financial crisis
• They plan to check whether Russian ban on short
  selling had a desired effect during the crisis
  time (2008-2009)

4
General remarks

• Power point presentation I have received contains
  carefully outlined summary of the idea and the
  basic concepts.
• It is, however, much more a lecture note than a
  presentation of the original work (at this
  iteration).
• Given that there is no paper yet, it is hard to
  make a definitive judgment on the project.
  However

5
General comments

• The way it is presented, it seems that the
  authors are simply implementing well known
  capital allocation methods in the context of the
  Russian market.
• To make the paper (once it is actually written)
  possible to publish, it is important to focus it
  on the most salient new results.
• These would be, as I understood it, two
• Which capital allocation method performs the best
  in times of financial crisis
• Whether ban on short selling is indeed reducing
  risk to which investors are exposed

6
Other comments

• Use of too many methods for volatility estimation
  distracts from these two main points.
• When writing the paper, the authors should
  probably go over the exposition of the models
  relatively quickly (since these are not the
  original contributions) and focus on the
  empirical testing instead.
• One possibility to broaden the appeal of the
  paper is to consider a broader class of markets
  (say, several emerging markets or perhaps also
  other, developed, markets) and check, overall,
  which of the methods provides a superior
  investment performance in times of crisis. That,
  in itself, would, possibly, make for an
  interesting paper
• If the large scale portfolio is important
  somehow, this should then be discussed clearly.
  I.e. perhaps some of the methods may work better
  in small and other in large scale

7
In conclusion

• I enjoyed reading the slides and I learned some
  interesting things about SRMs.
• I think that there is a potential to make this
  into an interesting paper, but needs a clear
  focus.