The recent flap over the administration's proposal to tax unemployment benefits provides an excellent illustration of why we so rarely get good economic policy in this country.

A sound reform that probably commands the assent of the vast majority of economists was dismissed without a serious airing. Instead of examining the idea on its merits, our political system drowned it in a sea of rhetoric. Its proponents were tried in the media and convicted of being mean-spirited.

Certainly, whoever merchandised the idea deserves part of the blame. Anyone who would suggest taxing unemployment benefits when the unemployment rate is pushing 11 percent probably would try to sell Toyotas at a UAW convention. And leaking the idea to the press on Thanksgiving Day was a charming touch. But is taxing unemployment benefits really such a bad idea?

First, an important distinction must be made. Reducing unemployment benefits is one thing; redistributing a fixed amount of benefits is quite another. Taxing benefits does both if no compensating action is taken. But if we raise the average level of benefits, and then make them taxable, we can redistribute benefits without changing the totals. Prosperous recipients will lose, but needy ones will gain.

The administration's proposal was stillborn before we learned whether raising average benefits was part of the idea. But how, so close to Christmas, could it have been otherwise? So, in the spirit of the season, let's assume that average benefits tax were to be increased.

Advocates of taxing benefits claim that it would increase the incentive to find a job -- an idea that was promptly denounced as crude and insensitive. Let's examine the idea dispassionately.

Suppose Jane Doe, who earns $300 a week, loses her job and qualifies for unemployment benefits of $150 a week. If neither earnings nor benefits are taxed, Jane increases her income by $150 a week by finding a new job and getting off the unemployment rolls.

But suppose earnings are taxed while benefits are not. If Jane is in the 25 percent bracket, then her after-tax earnings while working are only $225 a week, so the gain from finding a job is only $75 a week.

What happens now if benefits are taxed as well? Then the $150 weekly benefit is only worth $112.50 after tax, so the gain from finding a job rises to $225 - $112.50 = $112.50. Taxing benefits thus increases the incentive to find a job.

The example can easily be made more extreme. Suppose John Dough, who also earns $300 per week, has a wealthy wife and lives in a state with a high personal income tax. If they are in the 60 percent (combined state and federal) tax bracket, his $300 weekly paycheck nets him only $120 -- which is $30 less than his unemployment benefit! But if benefits were taxed, he would keep only 40 percent of his benefit, or $60, and could double his after-tax income by taking a job.

Basic economic principles assert that people work more when they have greater incentives to do so. And statistical evidence but-tresses this common-sense idea, suggesting that people who lose their jobs do remain unemployed longer when benefits are higher.

But don't jump to the wrong conclusion. Unemployment insurance is not a bad idea. It puts a more human face on capitalism, helps limit recessions by propping up the purchasing power of unemployed workers, and provides valuable protection that workers cannot easily buy from private insurers. These are considerable achievements.

Please do jump to the right conclusion. Because earnings are taxed while benefits are not, the incentive to find employment is weakest for secondary workers in high-income families. People in high tax brackets gain the least by trading in their (untaxed) benefits for (taxed) earnings.

Taxing benefits like earnings would increase rich John Dough's incentive to work by $90 per week, increase Jane Doe's by $37.50, but would have no effect on the work incentives of Poor Nell, who is too poor to pay income tax. This is a case in which incentives are improved by cutting the benefits of the rich rather than the poor -- a refreshing change from most supply-side prescriptions.

And there is more good news. With the tax revenue collected from rich beneficiaries like John Dough, we can raise unemployment benefits for everyone. If benefits are raised, rich John Dough will pay 60 percent back to the government in taxes; middle-class Jane Doe-will pay back 25 percent; but Poor Nell will keep it all.

For example, a change in weekly benefits from a tax-free $150 to a taxable $200 would increase Nell's after-tax income while unemployed by $50 per week, leave Jane Doe's unchanged, but reduce John Dough's by $70 per week. Merry Christmas!

One further fact illustrates the hysteria that greeted the proposal on Thanksgiving weekend: almost nobody mentioned that partial taxation of benefits has been the law of the land since 1979. Unemployment benefits are now taxable on joint returns with adjusted gross income above $18,000 and on single returns above $12,000.

How have things worked out? Gary Solon, a Princeton graduate student who studied unemployment data from Georgia, found that when benefits became taxable, high-income people, but not low-income, people remained unemployed for a shorter period of time. Furthermore, according to his study, the magnitude of the decline was close to what economic models predicted. A coincidence? I doubt it.

So let's hear it for those great liberals within the administration -- whoever they were -- that wanted to cut unemployment benefits for the rich, raise them for the poor, and improve everyone's incentives in the bargain. Too bad they did not succeed.