Complexity

Law professor Michael Greenberger recently appeared on Fresh Air to explain the subprime meltdown and the overall state of the economy. He said that, while the US economy used to be based on investments in tangible assets and enterprises, it is now largely based on financial instruments that are essentially bets on the directions that the tangible assets and enterprises will move in value. He called it a "shadow economy" because a large percentage of its overall value is based on these instruments - derivatives, monetized baskets of assets such as loan portfolios, etc. In other words, much of our economic activity is based on what the tangible assets represent rather than on the assets themselves. I'd call it a second-order economy, perhaps.

It seems to me that this movement away from tangible asset to intangible derivative of the asset owes much of its existence to the Black-Scholes option pricing model and its relatives. This model made pricing an option apparently as reliable as pricing the underlying asset. It inspired so much faith that whole financial industries were built on its foundations. Perhaps the early collapse of Long Term Capital Management, which required the last large Wall Street bailout prior to Bear Stearns, should have been a warning that option pricing wasn't as reliable as it appeared to be.

Nonetheless, heedless of the LTCM fiasco, derivatives became the basis for huge investments in, among other things, baskets of mortgages. Because banks were able to sell their loans to aggregators that collected thousands of loans into securities that they could slice into derivatives and resell, the banks that originated the loans had no further incentive to ensure that the loans would actually be repaid, and every incentive to make and sell as many loans as possible. Consequently, no one knows how many of these loans will actually be repaid.

So here's the problem. Because no one knows how many of these loans will actually be repaid, no one knows what the true values of the derivatives are. When Citi Group or Merrill Lynch or Bear Stearns take writedowns at the end of a quarter, they don't know if those writedowns will be sufficient to cover the actual losses or not. No one knows - without knowing the actual risk underlying the instrument, there's no principled way to value it.

Here's where LTCM comes in. The mathematical models that generate derivative prices are so complex that humans can't understand them - hence, Long Term Capital Management encounters a scenario that they didn't anticipate and their house of cards crumbles around them. The underpinnings of the derivative market suddenly look shaky today, and the economy collapses around the vacuum created by a relatively small number of unpaid mortgages.

All of this makes me wonder - if the economy is built on financial instruments that no one actually understands, how reliable can it be? If software and operating systems are so complex that no one can actually understand how they work, is that why Windows and some applications are so unreliable? Is that why my old Windows Mobile phone kept having problems? Are we condemned to a world where our basic tools are so complex that, in the end, we can't trust them?