Application of statistics fails

As someone who has spent his entire academic life in computational economics, I would like to comment on the April 12 Bloomberg article by theoretical physicist Mark Buchanan, titled “Beware economists who peddle cute models.”

An economic model is based on the fundamental assumptions in statistics that what happened in the past will be replicated in the future, more or less, by stochastic means with some probability calculated once again from past data or from simulated data. In reality, that cannot happen, whether in your personal life or in the U.S. housing market cited by Buchanan.

A model can reflect the past quite accurately, but it cannot forecast anything, because the purpose of a model is to plan. And planning is anathema to those Western economists who are always on the lookout for the ghosts of socialist economics.

One does not need a modeler to tell you that very low interest rates and high leverage created the housing price bubble that popped in 2008, creating panic and bankruptcy.

Any retired insurance salesman would tell you that a mortgage rate below 7 percent (with a base interest rate of 5 percent), leveraging of more than 75 percent, and greater borrowing limits as a percentage of annual income would cause personal bankruptcies in the future. That has happened.

My experience suggests that the application of statistics is responsible for the decay of economics. When the queen of England asked why economists failed to predict the disaster of 2008 in the financial market, she should have been told that statistics cannot predict any future upswing or downswing. Statistical models are based only on very strong assumptions that everything is normal and predictable.

Few things in this world are predictable.

Western economic theories that generate predictions based on data, their patterns and algebraic equations look very elegant on paper, but their utility is limited. They promote the writer.

Western economics derives from partial pictures of reality, regardless of whether the discussion is about David Ricardo or Milton Friedman.

Economics is not physics but rather engineering, for which one should try to design a new reality with all available tools, mathematical and computational, rather than try to predict a future reality, which is impossible by definition.

dipak basu

The opinions expressed in this letter to the editor are the writer’s own and do not necessarily reflect the policies of The Japan Times.