Monday, May 25, 2009

Mathematics of Credit Crisis


I was disquieted when Wall Street types began talking about advanced mathematics as the core tool for producing economic models that priced product. I simply know it can not work that way. Our mathematica is based on continuity, while our economic world is based on discrete events that can only be estimated by mathematica.

Large numbers tend to make the estimating mathematica fit better but it does not repeal the role of human decision making over discrete events or that they communicate and can catastrophically develop schooling behavior.

Over the long term, even extreme schooling events can be compensated for, except that no credit system is actually able to withstand such an extreme event. In this case, it was the classic push to increase leveraged assets and the loosing of regulatory ratios that allowed the capital base to become far too small to handle what is after all a run on the bank. We just do not do line ups any more. So in the short term the global banking system lost cash holder’s confidence and paid the price.

Having been involved with portfolios and the like for years, I have always found portfolio riskiness analysis to be rubbish. They fail for one very sound reason. If the risk of failure is 5%, then the sheer weight of time will reward you with failure. A portfolio must have dynamic management to prosper and the winners must be able to provide a tenfold return over the long term to offset real losers.

From this it should be clear that if your bond portfolio has no upside save the interest earned, then the lower the rate, the less risk that you can actually accept. Sustained low cost money should cause a strong improvement in corporate capital positions, but it is also not the time to load up on debt that will need to be refinanced with expensive money later.

Cheap rates drove Wall Street to manufacture attractive yield bearing securities with fraudulent credit worthiness because the demand was there and there was no supply.

It is telling that the Chinese put their US dollar accumulation into US treasuries. That surely meant that the alternative investments offered to them did not pass muster.

Today, I am informed that they have commenced buying raw material displacing demand that used to come from Europe. I would suggest that they are now using physicals to burn off some of the US dollar hoard. This will become clearer over the next few months and it is a luxury that a state has. Repatriation of this cash will shore up corporate balance sheets everywhere.

It is just that it is hard for Americans to imagine that other nations are actively doing things to restore the economic system regardless of US floundering.

I am scheduled to interview Riccardo Rebonato tomorrow for EconTalk. His book,
Plight of the Fortune Tellers is one of the best things I've read about the crisis. Maybe the best. Written before the crisis, Rebonato warns of the dangers of the techniques being used at the time by both firms and regulators to assess the riskiness of their portfolios. He has a lot to say about probability, risk, and the seductive romance of the ill-suited applications of advanced mathematics. Best of all, it's very well-written and though at times, very ambitious, it is always accessible to the non-practitioner.

He has a fascinating discussion of "economic capital," the amount a firm would hold on its own to avoid the risk of bankruptcy. He argues that firms will hold too little capital because they will rationally ignore the spillover effects a collapse of their firm would have on other insititutions, so-called systemic risk. But a firm doesn't want to go bankrupt. It may take too much risk and end up bankrupt anyway.

Rebonato also points out that bondholders and stockholders have different goals for the firm.
Bondholders want enough profitability to get their money back but do not share in any upside risk. Stockholders generally wnat more risk even though there is the risk of bankruptcy. For a naif like myself, this raises the question of why these institutions have both stakeholders with such conflicting goals. And the managers in these publicly traded companies would seem to have different goals as well.

A few semi-random questions, a few of which I hope to ask Rebonato tomorrow.

If it's hard to assess risk, and therefore hard for regulators to specify what is "enough" capital, how would an unregulated firm choose economic capital to reassure bondholders and stockholders that their firm is a good risk?

Why did firms choose such radically different levels of riskiness when they faced similar constraints?

Some people argue that the reason firms took on so much risk was because they were publicly traded. Didn't investors know about the moral hazard problem?

Was "too big to fail" an important, crucial, or irrelevant factor in the risk profiles these firms ended up with?

What role does mark-to-market play in risk assessment?

How did the ratings agencies distort choice if at all? (He seems to think it did.)

How much did Basel II distort choice, if at all? (He seems to think it did.)

Rebonato also observes that the 1990s reduced profit margins because of increased competition, encouraging innovation as a way to achieve yield. That's generally good. But Rebonato implies that firms continued to take bigger and bigger risks as a way to sustain yield. Why didn't self-regulation slow or stop this?