Computer Models Can’t Substitute for Real Life

by Peter C. Earle
The American Institute for Economic Research

On the 31st of July — just a few short days away — the six month anniversary of the WHO Declaration of a Global Health Emergency will arrive. And while there are many dates which could serve as the starting point for the start of the crisis, this one seems particularly appropriate: it is shortly thereafter that epidemiologists began plying the quantitative tools of their trade in force, and just after which the conversion from model outputs to policy schemes started to transpire.

We know now how dreadfully awry much of that planning went, not only in terms of harmless (if shockingingly wrong) predictions, but in real terms: lost lives, wealth, and opportunities. To the extent that epidemiological models attempt to predict or ascertain the infectious rate of a given pathogen, there is reason to believe that they can be helpful. But to the extent that those predictions, in turn, hinge upon decisions made by human beings — often by including a characteristic of an agent that replicates making choices — the effort is doomed to fail and, moreover, is an exercise in uncommendable audacity.

Continue Reading at…