It is natural for President George W. Bush and his team to argue that their economic policy helped Republicans achieve their mid-term sweep. After all, in his first year in office, Mr Bush sent tax rates in the right direction - downwards. And that step helped to "return confidence" to voters, as he put it, and made the president seem trustworthy.

Some go further, arguing that the 2001 tax legislation helped to stimulate the economy. Since that worked, the argument runs, a further fiscal stimulus would help to keep the economy on track. But while many voters liked the tax cut, it is not clear that it boosted demand. So the administration would be on firmer ground in proposing further tax cuts if it did not base its argument on the need for a fiscal stimulus.

First, a little history. In July last year, the Internal Revenue Service began posting rebate cheques of up to $600 to US households. The White House said the stimulus could "soften recessionary headwinds". The rebate stimulus argument was based on a standard Keynesian model: tax cuts would lead to consumer spending, which would help growth. The White House was putting its money where its theory was: the rebates cost about $38bn, or about half a percentage point of gross domestic product.

The theory was not borne out in practice, as Joel Slemrod and Matthew Shapiro point out in a National Bureau of Economic Research paper published last week.* The economists looked at households receiving rebates, not once but three times. They found few families reporting that they either had spent, or would spend, rebate cash.

In the first survey, conducted before the arrival of the stimulus cheques, only one in five recipients said he would spend his cheque. Three in ten said they would save the money; the remaining five said they would pay down debt. Nine months later, after the money arrived, rebate recipients were asked what they had done with it. The ratios were more or less the same.

In a different survey, families were asked a hypothetical question: what might they do with an even bigger rebate, of $1,000? Again, rather few - 16 per cent - said they would spend the money. Especially striking overall was that lower-income households were less likely than higher-earning ones to spend a rebate. This stands the old textbook on its head.

It could be, of course, that people lied and said they had saved when, in reality, they had binged on post-September 11 "comfort items". But the federal data on the personal savings rates for the same period back up the findings. When they were feeling strapped, people saved more and spent less, explains Mr Shapiro. Sounds familiar? Think of Japan.

The study's results do not mean that the Republicans should drop tax cuts. It simply means that it is time to push through the sort of tax cuts that work. The first is one the White House has been mooting since August: ending the double taxation of stock dividends. Corporations pay taxes on the dividends they issue on their shares. At the taxpayer level, dividends are taxed again - at rates that can be many percentage points higher than those applied to capital gains.

This lopsidedness resulted in a bias on the part of investors towards shares that provided capital growth rather than a dividend yield, exacerbating the bull market craze for growth stocks. Cutting dividend taxes would help to restore the balance between the two forms of equity investment. Naturally, a tax cut for dividends would not hurt the stock market. But the justification for the measure is less a stimulus than the restoration of rationality to capital markets.

Another valuable step would be to make the currently planned tax cuts permanent. Because of budgetary strictures, these cuts to income and estate taxes have two problematic aspects: they phase in gradually; and they expire in 2010. This is bad politics, saying, in effect: "We will give you a tax cut - but not today! - and then we will take it away." Earlier this year, Gallup reported that Americans liked the idea of making the cuts permanent, which is why there were as few Democrats opposing permanence as there were Democrats opposing a war in the mid-term campaign.

Legislating permanence would allow Republicans to show themselves to be the reliable guys and gals whom voters thought they were supporting. Permanent cuts would also help America's relative competitiveness. That would be especially true compared with Germany, which is contemplating killing its equity culture with initiatives such as a new capital gains tax. Since the mid-term elections there has been much alarmist talk about the deficit. But tax cuts that generate growth can mitigate the deficit as they did in the late 1990s.

As they consider such measures, the Republicans should note that short-term actions to spur the economy are about as effective as pouring water down a well. In other words, tax cuts aimed at long-term economic growth are worth trying. But not another "stimulus".

* NBER working paper 9308. www.nber.org

© Copyright 2002 Financial Times