As appealing as it may be to politicians and journalists to think so, the economy is not a motor vehicle. It does not have an "engine," there are no "spark plugs," and you cannot attach "jumper cables" to it. Above all, it is not susceptible to "stimulus," that discarded relic of Depression economics that has somehow remained lodged in popular thinking, a virus passed from half-remembered undergraduate courses through the pronouncements of public figures and into the general run of "what everybody knows." An economy is not a thing. It is a set of relationships -- between buyers and sellers, investors and entrepreneurs, employers and employees -- who exchange goods, services, capital and labour with each other at a price each finds agreeable. How well they collaborate will determine how productive they are and how rich they become. The notion that all of this productive activity on the part of millions of people could be speeded up or slowed down, more or less instantaneously, merely by pulling little levers marked "spending" or "taxes" in some government office somewhere, must count as one of the more risible urban myths.

Yet everyone soberly discusses the need for economic "stimulus," as if they had the slightest idea of what this meant, or how it was supposed to work.

Conservatives, I should say, are at least as guilty of this as liberals.

There was a time when conservatives sneered at talk of "stimulus," when this meant big new spending programs. But ever since Reagan conservatives have been peddling the same miracle cure, only in the form of tax cuts, which are said to "put more money in consumers' pockets" and thus, inevitably, "jump-start" the economy. How I wish even one politician, whether of the left or right, invited to give his views on how he planned to "stimulate" the economy, would say "the question is nonsensical: You might as well ask how I plan to make it rain." That isn't to say that tax cuts aren't good for economic growth. But the effects are mostly in the long-run, and work not by stimulating demand but by raising productivity. Mr. Bush, to his credit, seems to understand this. "The role of government," he said in his speech to the Chicago Economic Club yesterday, "is not to manage or to control the economy from Washington, D.C., but to remove obstacles standing in the way of faster economic growth." Taxes pose an obstacle to growth in two ways. One, as applied to productive resources like labour and capital, they drive a wedge between buyer and seller: After-tax, the buyer pays more and yet the seller receives less. This reduces both the demand and the supply: Purchases (say, of capital) are more costly, while the supply (say, of labour) is less remunerative. And two, so far as tax rates vary across different goods or services or different types of investment, they distort economic decisions: Instead of weighing the real costs and benefits of their choices, people are too often swayed by the tax implications.

Mr. Bush's proposal to abolish the tax on dividends scores well on both counts. Because it reduces the cost of capital, it encourages (more precisely, it does not discourage) investment. And because it equalizes the tax treatment of dividends, capital gains and interest on corporate debt -- the three means by which investors are compensated for the capital they provide to firms -- it removes a significant distortion in the American tax code.

This last point requires some explanation. Investors will still pay tax on capital gains and interest income. How does abolishing one tax make it equal to the other two? Because in whatever form it is paid, the income investors receive comes out of a company's after-tax earnings. That's not a problem when it comes to interest payments: Companies can deduct these from their taxes. Capital gains, likewise, are taxed at a reduced rate, to take account of the tax already paid at the corporate level. But until now dividends in American companies have effectively been taxed twice: once as corporate income, a second time as personal income. Abolishing the tax at the personal level doesn't mean they won't be taxed. It just means they'll only be taxed once.

That won't have much impact in the short-run. But it will greatly improve the efficiency of U.S. capital markets in the long-run. And if investors have more confidence about their longer-term prospects, they may be less skittish about investing today. That's the best sort of "stimulus." A footnote: There's been much clamour for Canada to "match" Mr.

Bush's bold scheme. But in fact the United States is playing catchup here. Canada's tax system already compensates investors in dividend- bearing stocks for the tax paid by the corporation, via the dividend tax credit. Or at least, it does in part: Short of abolishing the tax on dividends, we could simply bump up the credit.