Tuesday, October 6, 2009

Robert Rubin Understands Dancing Variables

The most dangerous step that econometricians take is to assume that a variable is a constant. This will result in equations that "work" until the assumed constants start to dance and act like variables.

In other words, they work under "normal" circumstances, but the eqations do not to take into account what might happen under extraordinary circumstances. Long Term Capital Management and the subprime crisis were to significant degrees problems in equations that were designed assuming certain variables were constants (Although the subprime crisis was accompanied by a full blown business cycle downturn.) .

Former Clinton Treasury Secretary Robert Rubin is one of the very few who understands the problems of these faulty econometric models with assumed constants, at least Rubin understands it on a gut level.

Ryan Lizza in his profile in New Yorker on Larry Summers pulls this qute from Rubin's memoir, In an Uncertain World: Tough Choices from Wall Street to Washington, written in 2002:
Larry [Summers] thought I was overly concerned with the risks of derivatives,” His argument was characteristic of many students of markets, who argue that derivatives serve an important purpose in allocating risk by letting each person take as much of whatever kind of risk he wants. Larry’s position held together under normal circumstances but seemed to me not to take into account what might happen under extraordinary circumstances.


1 comment:

  1. This is just CYA by Rubin. The only problem with derivatives during the financial crisis is that they were based on bundles of mortgage obligations that included sub-prime credits foisted on the financial community by Rubin's short-sighted politicos.

    GIGO - garbage in garbage out, an acronym most "students of markets" fully understand. Apparently not Rubin!