Back off, Paul O'Neill, treasury secretary, told the business lobbyists seeking to work new breaks for individual sectors into the Bush tax cut plan. This plan is for families; the shark fight over business breaks has to come later.

The current plan would mainly be confined to pulling down marginal rates and "flattening" the progressive rate schedule for families. It would replace a steeply progressive five-rate schedule with a less progressive four-rate schedule.

But though it isn't labeled "for business", the proposals do help emerging companies. In particular, they can assist that crucial subset that generates so much economic growth, new entrepreneurs.

This at least has been the finding of President Bush's new chairman of the Council of Economic Advisers, R Glenn Hubbard, a professor at Columbia University, who has done significant research on the complex interactions of tax rates, tax schedules, and entrepreneurship.

In "Tax Policy and Entrepreneurial Entry" Professor Hubbard and Columbia colleague, William M. Gentry, looked at data on taxes and self-employment for the head of household for the period 1979-1993. They found that the Clinton administration's 1993 tax increase lowered the level of entry into self-employment for higher income households by 20 per cent.

In contrast to many of the classic free marketeers and supply-siders, the authors did not find that shifts to higher rates alone necessarily stop the creation of new businesses. Rather they concluded that changes in the shape of the rate schedule matter most - the more progressive, or convex, the rate schedule, the greater the deterrent to new business formation.

The argument makes sense for a number of reasons. The first is that business startups in America are sensitive to the income tax, since many new entrepreneurs still receive and generate earnings that are subject to the income tax regime (rather than to corporate taxes). The second reason the Hubbard/Gentry hypothesis rings true is that startups do tend to experience a progressive rate schedule as a series of stairsteps which they must climb to success. The steeper, or more progressive, the staircase, the more likely they are to give up.

This work is an important addition to a field which generally tends to concentrate on simpler areas, such as the relationship between labour supply and tax rates. Classic research for example has shown that male labour is relatively inelastic: that men want to work regardless of the tax rate. Female labour, by contrast, is relatively elastic. When taxes on the second earner in a household, usually a woman, mount, that earner tends to alter her economic behaviour.

Thus the 1990s tax hikes, both of which increased the top marginal rate, deterred many women from working or continuing to work. Or, to put the problem in layperson's language, they did "The Babysitter Math". If paying the babysitter cost more than the gains from the job, women stayed home.

All this speaks well of the Bush plan, which reduces both progressivity for all earners and progressivity for two-income households (reducing the "Babysitter" problem). It means that if the plan passes there will likely be more startups in the future - not that this trend will satisfy the ravenous lobbyists. But then, no one ever accused them of fighting for the general good.

© Copyright 2001 Financial Times