09 November 2011

Economic Models and Forecasting


Recently, I’ve been thinking a lot about how pointless economic models are.  A good portion of this is probably due to being in a college microeconomics class (which uses Mankiw’s book, which Vox picks apart here.).  What passes for mainstream economics is nothing more than gussied-up tautologies and pretty models that don’t actually prove anything.  What irks me most is how the proponents of the mainstream view—in this case my professor and Mankiw’s textbook—act so certain about everything they assert.

For example, Mankiw devotes a chapter to the actions of a monopoly in the free market, complete with very precise mathematical models.  Yet, this entire chapter is utter B.S. because all the models in the world don’t change the fact that said models are based entirely on assumptions and tautologies, and stupid tautologies at that.  The models, then, only say what the programmer thereof tells them to say.  This isn’t science; it’s an appeal to personal authority, writ large.

In the same vein, an economic prediction is only going to be as accurate as the assumptions upon which it is based.  The reason why so many mainstream pundits missed the housing bubble is not because they lacked computational power with which to build their macroeconomic models; it’s because their assumptions about the housing market and banking industry were wrong.  They thought housing growth could be sustained forever, or that the current housing growth was organic instead of government-subsidized, or that the government would be able to subsidize housing forever, etc.

Trying to build on assumptions of arbitrarily-defined variables being correlated to one another is a recipe for failure, as is evidenced by decades of repeatedly doing so.  The time has come to put away the pretentious belief in the need for mathematically precise models; this isn’t physics, after all.  Ultimately, the foundation of economics must be built on understanding Man as an economic actor.  This paradigm shift cannot come soon enough.

No comments:

Post a Comment