If you thought that Bush economics was an oxymoron before, wait until you read this morning’s lead story in the Washington Post.

President Bush will propose that Americans be allowed to take tax deductions on more of their out-of pocket medical expenses, as part of an initiative the White House believes will rein in soaring health costs

Evidently, the Administration believes that health care is a Giffen good, so that if you subsidize it, total spending on it will decline.The thinking is explained by John F. Cogan, R. Glenn Hubbard, and Daniel P. Kessler in this book.

Allowing out-of-pocket health-care spending to be tax deductible has two opposing effects on health care spending. First, expanding deductibility lowers the overall price of health care relative to other goods and services and, thereby, increases spending…Second, [it] induces people to shift to health plans with higher deductibles and coinsurance rates which, in turn, lowers spending.

In appendix D, the authors claim to prove, using three pages of equations, that the second effect is larger than the first. With hundreds of billions of dollars at stake, I sure hope they didn’t, you know, forget a term or accidentally flip a sign or something.

Forgetting the equations, here is the intuition. If there were no tax distortions, I would pay for $500 of health care. With the current tax code, my employer might buy $1000 of health insurance for me, at an after-tax cost of $700. I have no incentive at the margin to conserve on health care spending, so I spend $800.

With the new tax code, I take $700 in take-home pay instead of health insurance, and I pay for health care out of pocket. Because the marginal cost of health services is no longer zero, I don’t spend $800 on health care. However, because medical expenses are tax deductible, I buy $650 of health care services instead of the $500 I would buy if they were not deductible and I had no insurance. So health care spending does not drop to the no-distortion level of $500, but it still declines from the current level of $800.

I think that there are good reasons to doubt that the results would shake out this way. My guess is that people would sort themselves into tax clienteles in such a way as to increase total health care spending, not reduce it. People with high marginal tax rates would pay out of pocket, while people with low marginal tax rates would stick with employer-provided health insurance.

When your hope for reducing health care spending rests on a sort of Giffen’s Paradox, you really are grasping at straws.