Britain needs to expand nursery education. This much is agreed. Increasing education provision for the under fives has become the public policy equivalent of motherhood and apple pie: no one could possibly be against it. So how on earth has such a good idea become badly snared by ideological, administrative and financial difficulties?

The legacy of the Thatcher years still casts a long shadow over education. Right-wing think-tanks still clatter about the battlements. Local authorities remain a “problem”. Everything in the public sector must be done on the cheap. The Department for Education wants more powers though without direct responsibility.

Thus apparently simple policy initiatives become bogged down in efforts to implement them in particular, peculiar ways. Morever, the existing arrangements for funding schools are an unholy muddle, involving partial control by the Government and partial control by local authorities. Under-fives services are even more complex, with local government resources split between education and social services, and some private provision.

The expansion of nursery education has been thrust into this political and administrative tangle. The DFE has been faced with a number of decisions. First, how many additional places are needed? Is the expansion simply designed to supply some kind of educational provision to all four-year-olds who currently have nothing? Or is it supposed to create nursery school education for all four-year-olds? Or to create some extra places while simultaneously changing existing patterns of provision?

It has never been entirely clear what the Government is seeking to achieve with its new under-fives policy. But it has become obvious that expansion is only part of the objective. Adding market rigour and keeping local government away from any new money have become key second-order purposes. Thus were vouchers born.

There are several ways in which the Government could expand - or radically reform - under-fives provision. ln each case additional funding would have to be provided by Whitehall, though the impact of different funding methods could lead to wholly different patterns of services.

The simplest way of pushing more money into nursery provision would be by increasing the Standard Spending Assessment (SSA) for under-fives education (Pounds 963 million in 1995-96) to provide the places needed to fulfil the Government’s commitment.

But there are problems. The under-fives SSA is not ring-fenced for use solely on nursery services. Some authorities receive SSA in line with their population aged 0-4, while spending little or nothing on education for this group. Extra SSA could - and doubtless would - be used by some to fund other services, or to keep local taxes down. Of course, if provision for four-year-olds became a statutory duty, at least part of the under-fives SSA would have to be used for the purpose for which it was intended. However, some councils already have sufficient (or nearly sufficient) provision for all their under-fives. Extra under-fives SSA for these authorities would, in many cases, be transferred to other services. Simply raising the under-fives SSA would not guarantee that all (or even most) of the additional cash was spent on nursery services.

Thus, allocating additional resources for the under-fives through the existing SSA is unappealing both to the Government and to many in education. A second possible way would be by use of specific grants. In education the bulk of such funding is concentrated under the Grants for Education Support and Training (GEST) programme.

GEST grants in 1995-96 will total Pounds 251 million in England. Targeting extra money to schools using GEST would represent a significant addition to the resources allocated in this way. Grant would be paid to local authorities specifically for additional nursery places. The Government could even use its powers under the 1944 Education Act to make grants to private-sector institutions.

Targeted grants also have drawbacks. First, it would be necessary to decide whether or not to give grants to those authorities (or to private providers in those areas) where there was already sufficient under-fives services. Not to do so would be grossly unfair to local taxpayers who would find that the Government was using national tax resources to finance under-fives expansion in areas where councils had chosen not to provide nursery places.

On the other hand, if the DFE paid additional specific grants to authorities that already made provision for most children under five, these authorities would presumably gratefully accept the new money and switch existing nursery funding to other purposes. Some of the new under-fives money would thus be lost.

Another potential difficulty with GEST-type grants is that they are time-limited. Generally they are used for pump-priming purposes, with a life-span of a few years. Unless specific grants for additional under-fives provision were made permanent, authorities might cut back nursery places once the grant had run its course.

GEST-type funding on a large scale would also require the DFE (or a new funding organisation) to develop a financing mechanism in parallel to the existing local authority system. Part of under-fives schooling would be funded by the LEA and another significant part by DFE via GEST. Such a split would create problems of the complexity now arising from having different methods of financing for grant-maintained and LEA schools.

One way round these difficulties with targeted grants would be to transfer the existing under-fives SSA away from local government. Any new resources could then be added to the SSA to create a new, larger, total of money for under-fives education. The DFE or a funding council could then allocate the money - using specific grants - to schools and private providers. The downside would be the massive change in resource distribution it would cause for existing providers, both LEAs and schools.

The third potential way of funding an expansion of nursery education would be through vouchers. In fact, the debate about nursery expansion has been hopelessly tangled in a second, and separate, wrangle about vouchers.

Those who strongly support vouchers doubtless saw the expansion of nursery education by a right-of-centre government as an ideal opportunity to push vouchers back into the political limelight.

Vouchers could indeed be used by the Government to allow parents to choose their preferred under-fives service. My colleagues at the London School of Economics, Howard Glennerster and Julian Le Grand, have explained (Guardian, April 4) that vouchers have left-wing as well as right-wing credentials. They are “basically egalitarian in nature”. Vouchers could discriminate to the advantage of the poor. Vouchers can offer real choice. Indeed, existing age-weighted pupil allocations to school are half way to a voucher system.

But the debate about vouchers must be separated from the argument about under-fives expansion. What the Government is seeking to do - according to its public statements - is to achieve a modest expansion in nursery education. Attempting to introduce a voucher system at the same time would require a massive shake-up of nursery funding. The existing SSA for under fives, plus any new money, would have to be put together and then allocated as an “Pounds x per child” voucher. Whatever the value of the voucher, a massive redistribution of money would be inevitable.

Such a redistribution would be likely to be exaggerated if vouchers could be used to purchase provision in private nursery schools, play groups or other forms of non-LEA institution. If the value of the voucher were set relatively low, it might encourage parents to use lower-cost (and perhaps more flexible) providers in the private sector. A low-value voucher would also fall well short of existing levels of funding per pupil in most LEAs.

It would be impossible to use vouchers to fund only the under-fives expansion. Somehow, the Government would have to give vouchers only to parents whose children are currently without a place.

Of the three possible ways of financing additional under-fives provision, SSAs are the best for local government, though there could be no certainty that extra resources would be used for the under fives. Vouchers would present insuperable problems unless the whole of under-fives provision were to be funded in such a way. The disruption for schools of moving to a voucher system would be politically unacceptable.

Thus specific grants emerge, by elimination, as the least worst way of maintaining existing services while allowing expansion of under-fives education in a way that would allow the Government to appease its right-wing. Crucially, no existing schools or LEAs need be threatened.

Recent press reports in the Sunday Telegraph and Daily Mail suggest there is still a will to proceed with vouchers, though not immediately. Under-fives provision will first be expanded (possibly including capital grants for some providers) and then, some years later, vouchers may be introduced. All the difficulties outlined above of moving to such a system would still apply.

Time is ticking away for Mrs Shephard if she is to fulfil the Government’s commitment to expand under-fives provision in advance of the next election.

There is political advantage to be gained by the party that ensures comprehensive provision for four-year-olds. Will the Conservatives act fast enough to benefit from it?

Tony Travers is director of a research centre at the London School of Economics