You would probably be surprised how many banks still operate with “Default Only” as opposed to “Mark-to-Market” models which is the technical term for the distinction you are making. On a similar note you would also be surprised how that some companies still claim that Binomial Expansion Technique (BET) is an adequate method for assessing option-style-risk, but this is a different story.

— you have to have a theory that (1) the probability of default of the underlying bonds is separate from (2) changes in prices of the credit default swaps on those bonds – but (2) is nothing more than the market’s assessment of (1)
—
not quite; (2) is the current compensation paid by the market for taking on the risk of (1), but this usually does include a risk premium and that risk premium can be volatile (and is arguably so) meaning that (2) can vary independently of (1). What you are positing would imply the following statement for the equity and bond markets “whether I invest $100 in equities, or in long term bonds, or in cash, I will always expect to make the same amount” which very few people believe (to be precise: it is not exactly the same – what you are saying is that there might be a risk premium, but that has been the same from the beginning of time to the end of the universe)

More importantly, though, we need better mechanisms to protect ourselves from companies that make bad decisions, because there will always be such companies. One way to do that is transparency. If everyone knew what AIG’s CDS portfolio looked like, maybe fewer companies would have been willing to do business with them. Maybe bond investors would have driven up the cost of AIG’s debt long before, forcing AIG to take more conservative positions. Maybe equity investors would have dumped their shares in fear, forcing AIG management (which, in all likelihood, only cared about their share price) to take more conservative positions.

There is a downside to transparency in a trading business, which is that it hurts traders if other people know what they are holding. But I think we need to move toward more transparency, because we should never be dependent on hoping that these banks and hedge funds know what they are doing.

I think your article was wonderful, especially your analysis of why AIG Financial Products’ models failed. You basically took the information contained in the story and turned it into useful knowledge, which we, your readers, can use in the future. (i.e. lessons learned from AIG’s mistakes)

Your analysis and conclusions lead me to the next question: What should “the brilliant minds” in these companies do in the future? – build better models which factor in the unforseeable or should the people using their models put less faith in these arcane models and use more of their own judgment in their decision making. (e.g. Jamie Dimon at JP Morgan Chase)

What would you recommend to the top-level risk managers of these financial companies? Is the mixture of complicated derivatives with arcane computer models toxic? And how will the regular investor be protected from these kinds of mistakes?

Add me to the list of those who are outraged about the risk manager from Bear Stearns landing a job with the Fed Reserve Bank. Have you no shame, sir??

The underlying assumption, that the process is gaussian, is where the flaw arises. In the kinds of probability distributions which properly model market behavior, the term “sigma” is meaningless.

See Benoit Mandelbrot, _The Misbehavior of Markets: A Fractal View of Risk, Ruin & Reward_ and
Nassim Taleb, _Fooled by Randomness_.

My opinion is that I think it’s fine if a company or companies has a model that is unable to take into account the remote probability of an extreme, highly correlated credit event like the one we are going through now. The impact of having an insufficiently conservative model is that you will take extraordinary risks in achieving ordinary returns. And that appears to be what most companies during the real estate cum credit boom.

But the problem that causes so much upheaval in the markets especially around the time of the LEH/AIG bankruptcy/near-bankruptcy was that it became apparent that nobody knew what their counterparty risk was. CDS are completely bi-lateral agreements and not traded on any kind of exchange. Therefore, when a company enters into a CDS it is counting on the other party to pay in the case where a covenant is triggered. But what if that party is a hedge fund that has recently gone out of business? Or, more likely, what if that party is a hedge fund that “invested” its CDS payment in other high-risk derivatives and is nearly out of business due to the combination of investor withdrawals and reduced asset valuation? In the case of a CDS payment that could result in a payment of 90c on the dollar, that event itself could push some hedge funds under (which it likely did).

I am not trying to pick on hedge funds but rather making the point that the CDS “market” is, and continues to be, a house of cards just waiting for the next bankruptcy to come crumbling down. Investors and institutions of all shapes and sizes have used CDS to hedge themselves against all kinds of risk. But what happens if these hedges aren’t really hedges at all and, in the event of a CDS covenant being triggers, they find that their counterparties are not able to pay or, worse yet, non-existent.

This was the problem facing Treasury when it bailed out AIG. Although I have no first hand knowledge, my impression is that AIG had written CDS on debt of all kinds including Lehman bonds. When Lehman went under, AIG was forced to post more collateral (which we all know). Given that AIG didn’t have that liquidity available, it was close to seeking bankruptcy protection as well. And, if Treasury hadn’t stepped in, the house of cards would have come crumbling down as number of companies that had written CDS insurance on AIG bonds was apparently very large.

I am saying that I agree with Treasury’s action to bail out AIG as Lehman and AIG are two very big dominoes and that would have cascaded into something none of us would have wanted to see. At the same time, however, I think we need to look at the CDS market in general and implement some important fixes. Specifically, we need to get these bi-lateral contracts onto an exchange where prices, payments, and positions can be clearly identified and guaranteed. This would be similar to the options market (and I believe the CBOE is angling to be the clearinghouse for CDS.)

You see, there is nothing wrong with a company that has a faulty risk model. In the end, they will pay the price if it is sufficiently insufficient. But there is no reason why the folly of a few companies should bring down the entire financial system. And when a $60B market is being traded in back rooms and not on an exchange, that is exactly the danger we are facing.

The actions of the Fed and the Treasury have probably saved us from the Great Depression or worse. But unless CDS are moved to some kind of a clearinghouse or exchange, we have only postponed the crisis to another day. It is inconceivable to live in a capitalist society where no large institution is allowed to go bankrupt due to its impact on the financial system as a whole. That is not “survival of the fittest” – it is “survival of the largest” which is truly horrible foundation upon which to build an economy.