Landsburg A-loose

In a lot of science fiction movies there is a robot or mad scientist who has logically deduced the necessity of some crime against humanity. Impeccable logic proves his case and it/he must act on it. Genocidal tyrants and their minions make similar claims at their war crimes trials, and so do terrorists.

Steven Landsburg is a mathematically educated economist who has written well on a variety of topics, including this nice (but hardly error free) note on the major twentieth century mathematician Andrei Weil. Weil, among other activities, was a founding member of Bourbaki, a coalition of French mathematicians dedicated to sucking the life out of mathematics.

On the web, Landsburg is probably most noted for his Slate articles, each of which seems to take a deeply counterintuitive notion and then defend it with rigid economic logic. His conclusions are often not only deeply counterintuitive but distinctly inhuman - remniscent of Bourbaki in their aridity and of the mad scientist/robot/Nazis in their detachment from any normal human feeling. This column, Do the Poor Deserve Life Support? illustrates style and method in starkest form.

Landsburg has written a short article in the Atlantic explaining how he signed on as a McCain supporter. I won't dissect it, but I will note one bit of his logic.

As that famously mad robot, HAL, used to say: "It's usually human error, Dave." And so it is with Prof Dave. An excerpt from his Atlantic suffices to demonstrate the point:

The reason poor Americans get too little health care is that rich Americans get too much. The reason rich Americans get too much is that they're overinsured, and therefore run to the doctor for minor problems. The reason they're overinsured is that employer-provided health benefits aren't taxed, so employers overprovide them.

It has been clear for decades that the single most effective way to control health care costs is to eliminate the tax break for employer-provided health care. According to one careful study by my colleague Charles Phelps (admittedly several years old, but I'm not sure anything relevant has changed), this single reform could reduce health care costs by 40% with essentially no effect on health care outcomes.

By the "It has been clear" remark he means that one economist made a model from which his improbable result emerged. Of course only a tiny minority of economists, mostly in the glue-sniffing libertarian branch of the profession, find that result persuasive, and there is zero empirical evidence to back it up, and a large amount of evidence to the contrary.

The notion that poor Americans get too little health care because the insured (not the rich, they don't need insurance) use too much is based on the very peculiar idea that health care is fixed quantity resource. Are Somalis starving because Americans are too fat? Probably so in Landsburg's world.

Economists need to believe in their models at least a little, of course, but those with sense don't trust them, especially when the evidence isn't there. Libertarian economics, though, is heavily into faith based reasoning. Their predictions, however, are wrong at least as often as anyone else's.