Monday, April 19, 2010
Posted by Jeff Lipshaw
Well, gosh, I haven't had this much fun reading the Wall Street Journal and the New York Times on a Monday morning in a long time. First of all, I want to note that while I use the first rate Choi & Pritchard, Securities Regulation 2d (great teacher's manual!), as the casebook for my class, I disagree with the idea of teaching Rule 10b-5 litigation before teaching the 1933 Act stuff on public offerings. So I teach the book out of order. The benefit is that we are moving into the elements of Rule 10b-5 in the next two weeks, and I'm thinking about scrapping my prepared materials in favor of this case. (A more devious me would use it as a question on the final exam.) I just want it to be noted that my syllabus is evidence I predicted this crisis, and now am in a position to benefit from it.
Second, I can't help being amused by some of the public commentary. The Wall Street Journal carries a headline referring to a statement by Gordon Brown (the British Prime Minister) that Goldman was "morally bankrupt." Oh, come on. That puts Goldman in the class of all other occupations that make money on the churn, like real estate brokers, executive recruiters, Las Vegas casinos, every state that conducts a lottery or allows parimutuel betting on horses, car dealers, and advertising agencies.
Third, in the past I advocated a standard of conduct in which one ought not to engage in conduct that one would be embarrassed to see highlighted on the front page of the Journal or the Times. I think I need to amend that to include that one ought not to be in a business in which one cannot explain the products being sold if they were to appear on the front page of the Journal or the Times. Let me give a breaking news example. My friend Erik Gerding has run a series of very helpful commentaries over at Conglomerate. I am pretty sure I disagree with a number of his conclusions (perhaps because having done deals in the real world for a long time I am more jaundiced about the number of times anybody actually gets led down the primrose path, and particularly in the never-never land of financial products). But that's what makes betting on horse races or synthetic CDOs. I've commented on one of his more interesting insights, but I decided to bring it out of the hinterlands of the way comments work over there. (I am grateful to Erik for having found the article and opened the debate on this aspect!)
Erik highlights this morning a paper by Arora et al., entitled "Computational Complexity and Information Asymmetry in Financial Products" to the effect that it really was material to the synthetic CDO investors that Mr. Paulson selected the Reference Portfolio. The gist of the paper is that it's very easy to create a computationally complex system from simple factors, but almost impossible to work the other way and select the factors that gave rise to the resulting system. Hence, conclude the authors, if an arranger of CDOs wants to hide "lemon" assets among the good ones, it's an intractable computational problem to find the lemons. See this blog post to which Erik links for another good explanation.) Thus, if Paulson had a hand in selecting the Reference Portfolio he really did have an advantage over those poor victims at IDK Deutsche Industriebank and ABN Amro.
Granted that blogging often is to research what journalism is to literature; nevertheless we don't always believe what we read in the newspaper and we need to be careful in assessing real-time commentary. The gist of my comment over at Conglomerate is that I think there's a flaw in Erik's move from applying the "Intractability Theory" in cash CDOs to a justification for a conclusion that who selected the Reference Portfolio in a synthetic CDO is material. In a cash CDO, there is only a long position - that is, the arranger has no interest other than in having investors believe that the underlying assets and the collateral are all good. The lemons in that case are the underlying mortgage assets. (Indeed, much of the math in the Arora paper is beyond me, but I believe that the authors argue the computational complexity increases the farther you get from the actual lemons, say by creating CDOs out of the CDOs.) Part of the problem may be terminology: the cash CDOs themselves are "derivatives" because their value derives from the value of the underlying assets. The point is that if the arranger-seller did slip in some lemons, the buyer would never be able to discover it.
How does that carry over to somebody who isn't actually compiling a portfolio of mortgages to package in a CDO to sell to investors, but instead is selecting the securities on which to bet in a synthetic CDO? Let's assume Paulson did select the synthetic portfolio. He doesn't want to slip a lemon asset in among the good assets. He wants ALL the assets to be lemons, not because he's trying to hide a pig in a poke (as if he were the actual arranger of a CDO), but because he wants to bet against the whole portfolio. He doesn't accomplish his goal at all if he gets IKB and ABN Amro to bet on really good assets with a lemon hidden among them. He ought to be worried that there are GOOD assets baked in there that he can't find!
Even assuming that Arona et al. have a point, I suspect Paulson doesn't qualify as the arranger who had the information advantage. The Reference Portfolio consisted of fewer than 100 already assembled CDO securities, each with a notional value of $22,222,222, and each being made up of many, many underlying mortgages (indeed, the flip book refers to the CDO security as "midprime" if the average weighted FICO score was above 625, and as "subprime" if the equivalent number was below 625). To put it more simply, if you are the bettor looking from the outside at a synthetic CDO portfolio, looking either to be long like IKB or ABN Amro or short like Paulson, and not the actual arranger of the cherries, peaches, plums, and lemons that went into the portfolio, you are no better off, informationally speaking, in trying to gauge whether you want to bet for it or against it.