Saturday, April 5

Death by Mathematical Model

"The Keynesian school of economics uses mathematical models in attempts to predict the economic future of a country. Unlike the Austrian School, which instead analyzes human action, most Keynesian economic forecasts have been dead wrong."
-- Matt Kibbe, from his 2011 discussion of the Federal Reserve's role in the 2008 financial crash

In a 2005 essay for this blog I asked, "How do you run a government when the voters are smarter than you?"  The American government's answer has been to load up on mathematicians to address every problem under the sun.  One problem with this approach is alluded to in this passage from Andrew Exum's 2012 Policy Journals:
I do think many of the articles that are in political science journals would elude the policy professionals who are actually running the government but whose education probably ended with a master's degree from a public policy school or, more likely, a law degree. I am skeptical of a lot of the statistical work being done in Middle East Studies for substantive reasons ... but in addition the math-heavy work featured in a lot of journals raises the bar of admission for potential readers.
What fool first got the idea that applying math to foreign relations would yield rational policies?  Does that explain why U.S. policy on the Middle East has gotten progressively wackier?

The over-application of math, which I touched on in my "Economic Collectivism - Police State" posts, is extremely dangerous when it's hardened into government policies. It's the old IT story of "garbage in, garbage out" but on government-issue steroids.  If the models are outdated or omit key data, and if government acts on such models, as it did in the 2009 Swine Flu outbreak, the consequences can be deadly to large numbers of people.

Of course math is an extremely useful tool for solving myriad problems, as is scientific research, so government shouldn't be barred from using either. What's the tiebreaker, then?  I'd say it's to keep in mind that science is experimental, a quest for knowledge, which means it's constantly correcting itself.  The corrections can't happen if they get frozen into a government policy. The same with math. If the equations are based on a wrong premise, mathematicians need to be allowed to revise their premise.  They can't do this when a set of equations becomes an article of policy faith.


No comments: