AIG, The Space Shuttle, and Creeping Risk

Listening to all the fully justified outrage about bonuses getting paid to the employees of AIG that took all the risky bets reminds me of the recriminations and second-guessing after the two Space Shuttle disasters.

No doubt if Edward Tufte put his mind to it he could come up with a fascinating graphic about the data leading up to the collapse of AIG, as he did with the launch decision for the ill-fated Space Shuttle Challenger. But in the meantime here are some thoughts based on the extremely thorough book, The Challenger Launch Decision, which I read years ago (a very interesting book, but very long — 566 pages).

Normalization of deviance leading to incrementally increased comfort with greater and greater risk. The Shuttle organization became more comfortable with doing things that lay outside the limits of specifications, or at least edged closer to the limits of specs. Each time this would reset their tolerance for deviation from normal. Ultimately this led to decisions that pushed risk over the edge. Like AIG (and much of the decision-making that led up to the economic collapse), in hindsight these decisions seem outrageous, but within the bubble they seem reasonable and with precedent. Which is not to excuse them by any means, just to note that we need to be especially on-guard, and reward those who are the lone voices questioning the decisions when the stakes are so high.

Assumption that others know the parameters and risks: In the case of the Shuttle, NASA engineers believed that engineers at Thiokol, the maker of the external rockets, knew the limits, when in fact there was a disconnect between the specifications that NASA used for launch decisions, and the specifications the rockets were designed to. Obviously plenty of people at AIG, and the banks who did much of the investing with them, were quite familiar with the risks, but many others were not. Thanks to how the credit rating agencies evaluated Credit Default Swaps, risks seemed lower than they were, hiding the true extreme risks from those less familiar with the ins-and-outs of CDS’s.

A realization that decisions are heavily influenced by emotions. Managers are not “amoral calculators” to use Vaughn’s phrase. Too much in the world of economics also assumes that people act rationally — consumers, bankers, investors alike. (This despite the daily evidence to the contrary, namely the emotion-driven swings of the stock market.) People will do things that are most definitely not in their self-interest (or in the interests of others they are supposedly responsible for) if they get caught up in the emotion enough.