Comments on: Kerviel at last http://www.prospect-magazine.co.uk/blog/franceprofonde/kerviel-at-last/ Tim King on French politics Mon, 13 Oct 2008 22:49:39 +0000 http://wordpress.org/?v=2.0.5 by: Tim http://www.prospect-magazine.co.uk/blog/franceprofonde/kerviel-at-last/#comment-33839 Thu, 07 Feb 2008 16:46:31 +0000 http://www.prospect-magazine.co.uk/blog/franceprofonde/kerviel-at-last/#comment-33839 Quickly on your second paragraph (I find the first one very helpful and good to think about), I agree absolutely - but that is the classic difference between the market and the state. I find it extraordinary that the statists cannot see that the best way to improve things is by making mistakes and taking the consequences. Being molly-coddled never got anyone anywhere. But to my great sadness I can see that France simply is not ready for that. More on that soon.... Quickly on your second paragraph (I find the first one very helpful and good to think about), I agree absolutely - but that is the classic difference between the market and the state. I find it extraordinary that the statists cannot see that the best way to improve things is by making mistakes and taking the consequences. Being molly-coddled never got anyone anywhere. But to my great sadness I can see that France simply is not ready for that. More on that soon….

]]> by: Sean http://www.prospect-magazine.co.uk/blog/franceprofonde/kerviel-at-last/#comment-33793 Thu, 07 Feb 2008 00:02:49 +0000 http://www.prospect-magazine.co.uk/blog/franceprofonde/kerviel-at-last/#comment-33793 I've also held off commenting for the same reasons you described. I do know a bit about banking and trading (having run trading businesses at a large investment bank) and perhaps can add some insight. I won't go so far as to guess at the sequence of events that resulted in SocGen's loss - I simply don't have enough information to make an intelligent addition - however I can speculate as to why a typical large bank's risk management department might fail more generally. In my experience, the risk management functions of large banks have a significant bias towards analysing, opining upon (second guessing) and generally 'backseat' trading the 'declared' risk of the bank. That is to say that the focus is on monitoring and interpreting the risk spit out by the system. Insofar as the "system" is concerned, the focus is on designing, implementing and auditing the financial and mathematical models used to produce the risk numbers - in particular the VaR (Value at risk.) Notwithstanding the enormous limitations of VaR as a risk measurement (see Taleb's Black Swans for an eloquent and accurate summary of the intrinsic flaw of VaR as a risk tool,) once these measurements are 'hard coded' into the systems, they numbers are seen as robust and risk management focuses on compliance vs established limits and policies and too often sees their role as being a devil's advocate to the traders and management as to the appropriateness and 'rightness' of the risk. So when the problem arises because the inputs are wrong (fraudulent) there is a much higher chance of it being 'missed.' I suspect further that when the underlying risk is simple (as seems to be the case with SG - ie long listed futures) ironically the chances of this being overlooked is higher: it's not sexy from a risk management point of view. If he had been mucking about with long dated barrier options, even with fraudulent inputs, falsification, etc. I believe there would have been a higher chance of risk management paying attention and digging deeper, earlier. Alot has been written about how SG - this super smart equity derivatives house - has been bitten by its own cleverness but actually, their losses were not due to derivatives magic gone wrong - Kerviel was just long shares - there wasn't anything wrong with models, volatility assumptions, etc. I also think the government's response is misguided. The best way to protect against a repeat (one can't insure against this ever happenning again no matter how many volumes of regulations are written imo) is to let the market work - ie if SocGen is taken over as a result of its weakened position, that would be the strongest possible signal for other banks to do their best adapt their procedures as needed to ensure this can't happen. Ironically by seeming to protect SG from these market forces, Sarkozy's government is implicitly underwriting other (French) banks risk management. Ironically, the best policy is aptly described by a great French expression..."pour encourager les autres..." I’ve also held off commenting for the same reasons you described. I do know a bit about banking and trading (having run trading businesses at a large investment bank) and perhaps can add some insight. I won’t go so far as to guess at the sequence of events that resulted in SocGen’s loss - I simply don’t have enough information to make an intelligent addition - however I can speculate as to why a typical large bank’s risk management department might fail more generally. In my experience, the risk management functions of large banks have a significant bias towards analysing, opining upon (second guessing) and generally ‘backseat’ trading the ‘declared’ risk of the bank. That is to say that the focus is on monitoring and interpreting the risk spit out by the system. Insofar as the “system” is concerned, the focus is on designing, implementing and auditing the financial and mathematical models used to produce the risk numbers - in particular the VaR (Value at risk.) Notwithstanding the enormous limitations of VaR as a risk measurement (see Taleb’s Black Swans for an eloquent and accurate summary of the intrinsic flaw of VaR as a risk tool,) once these measurements are ‘hard coded’ into the systems, they numbers are seen as robust and risk management focuses on compliance vs established limits and policies and too often sees their role as being a devil’s advocate to the traders and management as to the appropriateness and ‘rightness’ of the risk. So when the problem arises because the inputs are wrong (fraudulent) there is a much higher chance of it being ‘missed.’ I suspect further that when the underlying risk is simple (as seems to be the case with SG - ie long listed futures) ironically the chances of this being overlooked is higher: it’s not sexy from a risk management point of view. If he had been mucking about with long dated barrier options, even with fraudulent inputs, falsification, etc. I believe there would have been a higher chance of risk management paying attention and digging deeper, earlier. Alot has been written about how SG - this super smart equity derivatives house - has been bitten by its own cleverness but actually, their losses were not due to derivatives magic gone wrong - Kerviel was just long shares - there wasn’t anything wrong with models, volatility assumptions, etc.

I also think the government’s response is misguided. The best way to protect against a repeat (one can’t insure against this ever happenning again no matter how many volumes of regulations are written imo) is to let the market work - ie if SocGen is taken over as a result of its weakened position, that would be the strongest possible signal for other banks to do their best adapt their procedures as needed to ensure this can’t happen. Ironically by seeming to protect SG from these market forces, Sarkozy’s government is implicitly underwriting other (French) banks risk management. Ironically, the best policy is aptly described by a great French expression…”pour encourager les autres…”

]]>