The president should resist any impulse to please Democrats that puts at risk the best part of his educational reforms.

Democrats these days are feeling far friendlier towards President George Bush than most could have imagined back in bitter November. The most visible signs of a budding special relationship have shown up in the matter of public schools, where America's record borders on the tragic.

Democrats abhor the voucher component of the proposed Bush reform, which would give parents of pupils attending failing schools some form of tax break or cash so they might pull their children from the public systems. But the party is thrilled with Mr Bush's overall plan. His general idea is to increase Washington's contribution to town or state budgets while putting pressure on schools to perform better. Even Senator Ted Kennedy of Massachusetts, normally critical of Republican programmes, has said he finds "overwhelming areas of agreement" on education.

And no wonder. Mr Bush is stressing the accountability aspect of reform. Still, his is also a centralising programme. The accountability he seeks is to Washington, which will have more power over schools because it is increasing its share of their budgets. And centralising, or federalising, America's schools is an ancient part of the Democratic agenda.

But if winning on centralisation would be a victory for the Democrats, it would be a defeat for the pupils. In America, federal involvement in the schools has been small; less than 10 per cent of school budgets comes from Washington. Still, what federal spending there has been has caused outsized damage.

The trouble is that the goal of national projects, as opposed to local ones, nearly always concentrates on inputs rather than outcomes: this translates to mean that the environment where poor children study ought to be every bit as hospitable as the environment of schools in suburbs.

This goal is praiseworthy in the abstract but so grand and costly in practice that it obscures specific aims such as better reading scores. What is more, aid from distant Washington to the nation's scattered schools is rarely efficient, since Washington tends to see the schools as its clients, and ignores the true customers: parents.

Consider the record. America first woke up to its education troubles in the late 1950s with the Soviet launch of Sputnik. The Eisenhower administration passed the National Defense Education Act to improve competitiveness in science and maths, but most of the budget went to university-level projects. Still, the effort is notable because it was the last time in modern memory Washington targeted achievement.

The next national education project, far more important, came in the 1960s as part of the civil rights movement. Lawmakers injected millions into schools serving poor kids nationwide, with the aim of supplying better textbooks, teachers, and so on. This was seen as a revolutionary decision that would narrow the black-white score gap and by doing so, democratise the culture and enrich the education of all. At the time there was talk - most memorably from Sen Kennedy himself - that schools would be made accountable to Washington for student results.

But the accountability component fell casualty to the Zeitgeist, which held that demanding improved results of poor people, particularly minorities, was too much to ask. The result of the open-ended spending was that the country got little for its outlays. In 1980, 17-year-old blacks scored about 10 per cent lower than whites on reading tests, the same gap that obtained a decade prior.

And the poor were not the only ones hurt by the emphasis on equipment and teachers; the overall education culture turned away from performance. Average scores on the SAT, the entrance exam favoured by higher-end universities, dropped year after year.

Unconditional spending nonetheless remained the federal rule. When it didn't want to spend the cash itself, Washington, or federal judges, simply required that localities subsidise federal laws and regulations. Meanwhile, curriculums and some important national tests - most notably, the SAT - were dumbed down to conform to the culture of equality and to engineer more socially acceptable outcomes. Parents, including many black and poor parents, protested vehemently but had no leverage to force change.

Again, the results were disturbing. By 1995, America spent more per child under 13 than other Group of Seven leading industrial countries (60 per cent more than Britain). Yet a broad study from the same period found that 77 per cent of nine-year-olds from poor cities could not read at a basic level. More than four decades after the Sputnik challenge, US high school graduates performed worse in maths than pupils from Lithuania, Slovenia, or the Russian Federation - not to mention Singapore.

What then would make Mr Bush's programme any different? Only two things. The first is that Mr Bush has a record of goading public schools into producing better students. He achieved this in Texas, by shining a humiliating spotlight on schools that failed to produce improving test results. But to repeat this success on a national scale, the carrot of more spending alone will not do. The amiable Mr Bush will probably have to turn tough and use his stick, withdrawing funding from schools. He must also use significantly challenging tests, and prove his accountability is genuine and meaningful.

The second hope is the opt-out component of his programme. This would introduce a dramatic reality check into the decades-old dream of social equity pursued by educators and their Washington subsidisers. Once parents can walk out, they have a weapon to fight for better schools.

Mr Bush may shortly have to decide between romancing Democrats and preserving the controversial aspect of his education reform. It's a hard choice, but given the case history, an obvious one.

© Copyright 2001 Financial Times