Serving the Quantitative Finance Community

 
User avatar
Herd
Topic Author
Posts: 17
Joined: October 2nd, 2003, 12:48 pm

A hell of a lot to learn from this

March 15th, 2013, 9:37 pm

Senate report on London whale
 
User avatar
purbani
Posts: 10
Joined: July 14th, 2002, 3:00 am

A hell of a lot to learn from this

March 15th, 2013, 10:24 pm

Jamie Dimon Email Directly Ties JPMorgan CEO To $6.2 Billion Fiasco"On January 23, the risk management group emailed the chief executive to ask for the authority to temporarily raise the bank?s overall risk limit. The same email informed Dimon that the CIO "has developed an improved" risk model that would cut the CIO's measurement of risk by nearly half. Dimon wrote back to approve the change, according to the report."To me this is the crux of the matterHas anyone else here ever built a model that halved their apparent risk and then bet the ranch on it without having it very thoroughly checked ?People make mistakes, I have made many, and we know that all models are wrong in an important sense. But for a trade of this size to have been able to be "bottom-drawed" or "mismarked" for several months without any alarms going off or the explicit knowledge and indeed collusion of several people speaks to a very poor system.Either way it seems to be a question of whether or not there was intent or simply widespread incompetence.
 
User avatar
Herd
Topic Author
Posts: 17
Joined: October 2nd, 2003, 12:48 pm

A hell of a lot to learn from this

March 15th, 2013, 10:40 pm

section 5-D is completely amazing.VaR optimisation, model minimisation....The quants' reputation is not gonna improve.What's the solution to this new problem?Give more/all power to Risk?
 
User avatar
Herd
Topic Author
Posts: 17
Joined: October 2nd, 2003, 12:48 pm

A hell of a lot to learn from this

March 15th, 2013, 10:51 pm

well there are many many issues I thin, not just one.1. Have you ever heard of an FO guy writing a VaR model?2. Did you see the name of the quant who did it?3. The new model was rushed. They didn't give a shite about Model Review Group. In fact MRG made many critics but they didn't care. For them MRG is just something they are obliged to go through.4. It is clear that they knew what was going on: they wanted to keep the positions and find a way to be below the limit. That s why they developed this new model very quickly and then rushed it into production.5. Then the CRM thing: when Model risk report a very high number, he gets bullied by * (see top of page 193)6. And then * openly goes on about minimisation....Also * was surprised by the 50% fall, but he didn't say "hang on, that s a bit much. Maybe there is something wrong somewhere". Or maybe he was told to keep quiet.This is a massive issue. Yet one cannot talk about it for obvious reasons.
 
User avatar
purbani
Posts: 10
Joined: July 14th, 2002, 3:00 am

A hell of a lot to learn from this

March 15th, 2013, 11:18 pm

Hi HerdYou raise many valid points to be sureIf there was pressure from management to amend the risk numbers downwards and those numbers were presented to any Govt authority then presumably that is fraud. - I am not a lawyerIf the whole purpose of the CIO Office and its location in London was to circumvent Volker and serve as a stealth prop desk rather than 'hedge' any risk positions, as has been alleged - and also suggested by its large contributions to profits which true hedges don't typically provide, then that is some sort of breach also.I would love to say that JPM's 4:15 report on JD reads 11:55 but I suspect that that is not the case and that the US (In)Justice department will hand out another of its "non-admission-of-guilt", guilty plea bargain's ( the legal equivalent of off-balance sheet accounting ?) and a slap on the wrist fine rather than do what it can to promote a sense of faith in justice and respect for the the rule of law. As for morality any halfway decent leader would have fallen on his sword long ago but that virtue seems to be a optional extra in much of modern society and banking in particular.
Last edited by purbani on March 15th, 2013, 11:00 pm, edited 1 time in total.
 
User avatar
ppauper
Posts: 11729
Joined: November 15th, 2001, 1:29 pm

A hell of a lot to learn from this

March 16th, 2013, 9:42 am

>> A hell of a lot to learn from thisagreed
 
User avatar
Alan
Posts: 2958
Joined: December 19th, 2001, 4:01 am
Location: California
Contact:

A hell of a lot to learn from this

March 16th, 2013, 1:45 pm

I read a fair amount of that report. What a mess!My conclusion: The govt cannot trust these big banks to do their own VAR stuff.They need to take it away from them, set up an independent unit within the bank that is hired by and reports to the OCC, and bills back to the bank the cost of the unit. They are on premises full time anyway.
 
User avatar
Herd
Topic Author
Posts: 17
Joined: October 2nd, 2003, 12:48 pm

A hell of a lot to learn from this

March 17th, 2013, 10:02 am

Alan,I think main problem here is FO shouldn't do VaR. And Model Validation should not get pressure to approve. And their recommandations shld be followed.In fact Risk should have more power, not be something that needs to exist in order to comply.If we follow your reasoning, then why not nationalisation? or have govt officials behind every desk?we also know that nationalisation and govt control is no way a guarantee against fuck ups.
 
User avatar
DevonFangs
Posts: 0
Joined: November 9th, 2009, 1:49 pm

A hell of a lot to learn from this

March 17th, 2013, 10:49 am

QuoteOriginally posted by: ppauper>> A hell of a lot to learn from thisagreedbut still, they stole my thread so I'll call the senate
 
User avatar
farmer
Posts: 61
Joined: December 16th, 2002, 7:09 am

A hell of a lot to learn from this

March 17th, 2013, 10:59 am

QuoteOriginally posted by: AlanThe govt cannot trust these big banks to do their own VAR stuff. They need to take it away from themOr, the government could just sell their stock in JPMorgan.
 
User avatar
farmer
Posts: 61
Joined: December 16th, 2002, 7:09 am

A hell of a lot to learn from this

March 17th, 2013, 11:27 am

If people are so worried about the banking system, shouldn't they just put their money in a safety deposit box? I envision a business called "farmer's extreme-security storage." For $5000 a year, I can give them a space big enough to hold $1 million in cash. That is a lot less than the amount I see rich people's taxes have gone up, to keep these regulators in clam chowder while they make their 2%.
 
User avatar
DevonFangs
Posts: 0
Joined: November 9th, 2009, 1:49 pm

A hell of a lot to learn from this

March 17th, 2013, 12:12 pm

QuoteOriginally posted by: HerdAlan,I think main problem here is FO shouldn't do VaR. And Model Validation should not get pressure to approve. And their recommandations shld be followed.In fact Risk should have more power, not be something that needs to exist in order to comply.If we follow your reasoning, then why not nationalisation? or have govt officials behind every desk?we also know that nationalisation and govt control is no way a guarantee against fuck ups.Like for the gaussian copula, I believe this is again mostly a business issue, rather than a model one. Even good model val wouldn't change much, probably. VaR, IRC and CRM are just indicators built on the ground of some assumptions. Sky-rocketing VaR, IRC or CRM numbers can be unrealistic, but the business shouldn't dismiss them as unreasonable without looking into why the models are behaving wildly.
 
User avatar
DevonFangs
Posts: 0
Joined: November 9th, 2009, 1:49 pm

A hell of a lot to learn from this

March 17th, 2013, 12:13 pm

QuoteOriginally posted by: AlanI read a fair amount of that report. What a mess!My conclusion: The govt cannot trust these big banks to do their own VAR stuff.They need to take it away from them, set up an independent unit within the bank that is hired by and reports to the OCC, and bills back to the bank the cost of the unit. They are on premises full time anyway.OR acknowledge that if they let the banks free to develop their own models, they are going to do regulatory arbitrage.
 
User avatar
ppauper
Posts: 11729
Joined: November 15th, 2001, 1:29 pm

A hell of a lot to learn from this

March 17th, 2013, 12:18 pm

QuoteOriginally posted by: DevonFangsQuoteOriginally posted by: AlanI read a fair amount of that report. What a mess!My conclusion: The govt cannot trust these big banks to do their own VAR stuff.They need to take it away from them, set up an independent unit within the bank that is hired by and reports to the OCC, and bills back to the bank the cost of the unit. They are on premises full time anyway.OR acknowledge that if they let the banks free to develop their own models, they are going to do regulatory arbitrage.let the banks do their own VaR, just don't bail them out when the bank failes
 
User avatar
farmer
Posts: 61
Joined: December 16th, 2002, 7:09 am

A hell of a lot to learn from this

March 17th, 2013, 12:53 pm

If I took my dog to the vet, and two weeks later the dog was living at the vet hospital and sicker than he had ever been, I would just realize the doctor doesn't know what the fuck he is doing and save money on medical bills. When did the banking system develop HIV or terminal cancer or whatever, that it spends every damn day of its life in the hospital with some quack trying to cure it?