The formula that killed the banks

Neat article from Wired on the Mathematics behind the secutiries at the heart of the credit crisis here.

  • frustrated democrat

    Well that’s clear then it’s written in Cyrillic script not Latin, so it was Putin trying to undermine the whole Western economy, problem is he shot himself in the foot.

  • LJS

    It is all Greek to me!

  • mnob

    Its quite simple, economists thought they could model the world with numbers – in the same way the government did. It worked for a while so people believed in it. Then it fell apart.

  • Irish Supremacist

    Copulas are straightforward and have been around for ages. Maths is only an arsenal and cannot be blamed for misuse. If you want people to blame:

    1. look in the mirror if you were part of the Celtic Tiger.
    2. all the loafers who travelled to Dublin on their free travel passes to protest.
    3. Politicians who are notorious as moral hazard risk augmenters.
    4. All who believe in consumer cultures and the spiralling expectaitons that go with them.

    The Wired article is for wanna be geeks who want to feel they are on top of things. If they can’t even get on top of a woman, they will hardly mount advanced quant stats, which will continue to be centre stage until the mindsets outlined above end.

  • OC

    I’d say that we all just got copula-ted!

    The mathemetician David Li (a Chinese-born New York banker, once with Barclays plc) responsible for all this now disingenuously claims that we musn’t shoot the messenger.

    But he wasn’t just some hack – he was in charge of the quantitative analysis department at financial institutions. If the model was being misused, Li had a duty to bring this to light. I would really like to see his portfolio for the last several years.

    Li attempted to use an insurance model used to predict the probability that a person dies soon after their spouse does.

    This might work in our Western monogamous society, but would it be appliciable in, say, bonobo groups where everyone is everybody else’s spouse?

    It looks to me that a major problem was the synthetic CDO, similar to naked short selling. These SCDOs, being rather more an insurance product, need to be regulated as such, or banned outright.

    This, plus fraudulent real estate appraisals, which were at the heart of the 1990s Savings & Loan crisis in parts of the US, adds up to criminality in my book.

    Also, consider this – the housing boom in the US would not have been possible without the massive use of illegal aliens by construction companies.

    As someone recently suggested, we need a Nuremburg Trial to punish those responsible.

  • cladycowboy

    So, we have the mathematical geniuses of the like of Li whose models were fatally flawed [as he acknowledged a few million dollars later] and the hedge fund managers who sold this but hadn’t a clue what the maths contained.

    So who exactly is the ‘talent’ that would flee if the failed banks didn’t dish out bonuses this year?

  • NCM


  • OC

    BTW let’s not forget the pricing model for these Collateral Debt Obligations, the recently maligned Mark-To-Market method.

    At the moment, banks, et al, are complaining that this method undervalues the underlying assets (ie real-estate) of the CDOs, the Mortgage Backed Security. “If only we could change this method, the banks would instantly have more asset value on their books, and all our problems would be solved!” according to the geniuses that got us to where we are today.

    They had no problem on the upside – they re-valued their MBSs to market and showed amazing balance sheet fortunes! And paid fortunes in bonuses to self-promoting executives.

    The problem is, Mark-To-Market is only appropriate when a liquid, public, and transparent process is available for the trade in the underlying assets, in this case real-estate.

    I don’t know of any such market for real estate. Transactions for homes are pretty limited in price and parties involved. No NYSE or Chicago Mercantile Exchange action here.

    The financial-brainiacs will claim that the underlying assets are the present-values of the cash-flows generated over time by the mortgages (discounted for any increase of risk of default). And that, therefore, the Mortgage Backed Securities aren’t difficult to price, as they are actively traded.

    I totally disagree.

    IMO the pricing model should be Lower-Of-Cost-Or-Market. There wouldn’t be much help today, but it would ensure that real-estate bubbles are limited in the future.

    If LCM had been used instead of MTM, the holders of these CDOs wouldn’t have been constantly bumping up the “market value”, which as we have seen were illusory. (Hah – Ill & Usury!)

    Hopefully Financial Accounting Standard 157-d, Determining the Fair Value of a Financial Asset in a Market That Is Not Active will address these issues.