Politically, making smokers pay for children's health insurance is a great idea: Everybody loves children, and everybody hates smokers. But once you get beyond the popularity contest, it's clear that financing an expansion of the State Children's Health Insurance Program (SCHIP) with a big increase in the federal cigarette tax is neither fair nor wise.

As a group, smokers are less affluent than nonsmokers, and a poor person's spending on cigarettes represents a much bigger chunk of his or her income than a rich person's. These facts combine to make cigarette taxes highly regressive.

According to a Tax Foundation analysis, the Senate proposal to pay for a $35-billion SCHIP expansion by raising the federal cigarette tax from 39 cents to $1 a pack is the "least defensible alternative" because "no other federal tax hurts the poor more than the cigarette tax." The foundation's Gerald Prante calculates that "the burden of the proposed cigarette tax hike on the lowest-earning 20 percent of households is 37 times heavier than it would be if the government raised the money with the federal income tax."

Some supporters of higher cigarette taxes argue that smokers should bear a disproportionate fiscal burden because they account for a disproportionate share of taxpayer-funded medical expenses. But researchers such as Harvard economist W. Kip Viscusi estimate that, if anything, smoking saves taxpayers money.

Because smokers tend to die earlier than nonsmokers, they do not consume as much health care in old age or draw on Social Security as much as nonsmokers do. Leaving aside Social Security savings, a 1997 study in The New England Journal of Medicine concluded that total health care spending would go up, not down, if everyone stopped smoking.

Even if smoking does, on balance, increase government outlays, a 1994 report from the Congressional Research Service concluded that cigarette taxes in all likelihood already covered any external costs that reasonably could be attributed to smoking. Since then, the average cigarette tax (state and federal combined) has tripled, rising from 50 cents to $1.46, an increase of more than 100 percent in real terms. And that's not counting the price hike needed to fund the tobacco companies' settlement payments to the states.

Relying on yet another cigarette tax hike could mean that the people paying for SCHIP's expansion will be poorer than the people benefiting from it. The current Senate bill would raise the family income cutoff for SCHIP, currently 200 percent of the official poverty level, to 300 percent. Some legislators prefer a limit of 400 percent, which comes out to $82,600 for a family of four.

A decade ago, SCHIP's supporters sold the program as a way of providing health coverage to children whose parents could not afford it but were not quite poor enough to qualify for Medicaid. Now they are proposing changes that would make SCHIP resemble a middle-class entitlement.

President Bush is not the most credible opponent of a new federal health care entitlement, given his support for the exorbitant Medicare prescription drug benefit. But he is right to oppose SCHIP expansion and the tax hike that comes with it -- a burden that nonsmokers eventually will find themselves bearing as the percentage of the population that smokes continues to dwindle (an explicit goal of higher cigarette taxes).

SCHIP expansion is especially worrisome in light of research by economists David Cutler and Jonathan Gruber, who found that making publicly funded health care more broadly available tends to crowd out private coverage, encouraging people to decline employer-provided insurance or drop coverage of dependents. According to a 2007 paper co-authored by Gruber, "the number of privately insured falls by about 60 percent as much as the number of publicly insured rises."

This research suggests that much, if not most, of the money spent on SCHIP expansion would pay to cover children who already have insurance. That does not seem like a smart use of taxpayers' money, even if the taxpayers are an unpopular minority.