
The Sunday Times magazine had an article on understanding risk in the capital markets. It was an interesting article. It discussed a system referred to as VaR or value at risk. In a nutshell, the system is set up so that a portfolio manager can assess the potential P&L of a given position as its market value fluctuates. VaR is a good system - variations of it have been in use at banks for many years. It has been considered so reliable, that bank regulators have endorsed it and the Basel Committee has used it to determine compliance with regulatory capital requirements. Remember though, that the model was developed for banks ... banks, that historically bought only treasury securities and took all of their risk in lending to corporate clients.
What does this have to do with the mortgage crisis and why should we care? Simple ... if one has read what I have written in the past, one knows that the limitation of any trading system, is the people that are involved with it. The Times article seemed to miss this point. To actually believe that any valuation system can be relied upon all of the time illustrates, not only the folly of relying on computer models, but the limitations of the small minds in Wall Street.
The following is the lead-in to one of my posts from a while back: avarice |ˈavəris| noun: extreme greed for wealth or material gain - Harry's definition; an uncontrollable urge to look the other way when, in the face of making huge amounts of easy money, otherwise astute managers allow their staff to do deals they do not understand - even remotely.
The way that the deals get done, is to remind everyone just how reliable the models are and how careful everyone has been as it relates to VaR and, therefore, the soundness of the firms capital.
In summary: remember, that Wall Street firms are like all corporations, e.g., their organizational chart looks like a pyramid. Further, owing to time constraints, the higher you go on the chart, the further away from reality the managers become. Now, let's pretend that I'm a senior trader (who has always been profitable) and that I see an opportunity to make a big score. Generally, these opportunities occur due to disruptions in the market value of a security that is relative to another security or sector of the market - the security is said to be cheap. So, I can buy the cheap collateral, restructure it into collateralized debt obligation, like a CMO, get it rated and place it in the market and make a two or three percent profit. To the extent that the deal is $2 billion in size, two percent equals $40,000,000 - nice days work, non? Have you all read Catch 22? The catch in this deal is, that in order to get it rated, I have to come up with a buyer for the bottom most tranches ("the equity") of it. Let's say that a major institutional buyer - AIG say, tells me that they'll take 50% of the equity if we take the other 50%. Let's further assume, the equity is two percent of the deal, or $40 million. My pitch to management now becomes, "hey, this deal is so sound, that AIG will take 50% of the equity if we take 50%" management, "what's in it for us"? me, "$40 million" management, " how much do we have to buy and what's the risk"? me, "half, $20 million and the models tell us the risk is manageable" management, "so we put up $20 million, that we can borrow, and we can make $40 million? ... done! it's a no-brainer"
Now, the preceding scenario was carried out dozens of times during the last three or four years at every firm that was involved in the mortgage market. The astute among you will see the flaw, that is, "that we can borrow". What happened, is that each of the firms that are getting bailed out, and those that did not, borrowed the cash that was necessary to finance the portion of the deals that they, and their idiotic models, decided had risk characteristics that were manageable.
The models can't measure liquidity. When liquidity dries up, it does so from the bottom up. This means, that the lowest rated securities loose their ability to finance first. As soon as lenders realized that there was no market for the MBS, they pulled their lines and the firms that owned the securities had to use their own capital or sell the securities. The loses exceeded the firms capital and down we go.
I'm making this all seem pretty simple. In fact, it was. If any of the firm's management took the time to think about the nature of the collateral they were using to collateralize these securities, we hope, that the deals never would have gotten done. But, look back at my definition of avarice - could everyone have been blinded by greed? Can the models measure that variable?
I'll state this one more time; any reasonable person knows that real estate is the most vulnerable market of all to economic cycles - it's always been boom and bust. How, therefore, could a reasonably prudent manager look at a collateral pool filled with loans having LTV's of 100% - at the obvious top of a cycle - and think that the deal was sound?
Here is the best example that I know of: in 1989 or 90, on behalf of the trust of the same name, Goldman Sachs sold Rockefeller Center to a large Japanese investor. Rockefeller Center IS THE trophy piece of real estate in the US. The sale price was $1.3 billion and the date coincided with a peak in the cycle for real estate. In 1993 or 4, post Japanese bubble, a group led by Goldman and the Rockefeller Trust bought it back for about $600 million. So, the best piece of real estate in the US lost more than half of its value in about five years. Everybody in the US real estate market knows this. How then, could they let the single family market get so far away from reality?
Could it be avarice?