To call something a postcode lottery is a choice insult. It usually refers to a public service allocated by geography rather than need.
When Tim Yeo addressed the annual conference of the National Union of Teachers, the union's general secretary, Doug McAvoy, laughed when it was suggested that the Tories might want to abolish catchment areas. "What will you do instead?" he asked. "Draw them out of a hat?" Yeo backtracked at once - but he should have said that was exactly what he would do.
The alternative to the postcode lottery, it is said, is to give parents more choice. And both Labour and the Tories are now devoted to expanding choice. But does it work? Since the Education Reform Act 1988, the school admissions system for England and Wales has been based on the principle that parents have the right to express a preference in the choice of their child's school. Its rationale was to weaken the middle-class hold on school admissions. It has clearly failed.
We should be clear that the flaw is not parental choice, but catchment areas. When a popular school is oversubscribed, the admissions authority (this is often the local education authority; sometimes it is the school itself) has to decide how to allocate its places. After giving priority to siblings or to children with special needs, it will usually do so geographically.
Formal catchment areas for favoured schools have dwindled to a tiny radius as well-informed parents have crowded into the houses immediately surrounding the school. In some parts of British cities we have the absurd spectacle of admissions officers walking around with a tape measure, as if the school admissions procedure were like working out who has won a game of bowls. At the moment, if you cannot afford to live inside the enchanted area, you have no chance of getting your child into the school.
The market is working very efficiently: either you pay school fees or you pay through a high mortgage. A 10 per cent improvement in Key Stage 2 results in primary schools is associated with a 7 per cent increase in house prices.
The usual left-of-centre solution is to force children to go to their nearest school. This would have two consequences. First, any allocation system that is based on a zone of right of access to your local school inflates the housing market. Second, many middle-class parents who fail to get their children into the desired school will leave state education.
There is a well of untapped demand for private education, and we do not yet know how deep it is. The number of children attending private schools has gone up in every one of the past nine years. If the private sector decides to expand capacity rather than increase its prices, don't be surprised if the places fill up.
Instead of turning away from parental choice we need a new device for settling conflicts. There are many possible ways of allocating pupils when a school is oversubscribed. For instance, the education select committee of the House of Commons recently suggested that schools should recruit in ability bands, so that none of them has more than its "fair" share of either above-average or below-average children.
There is no need for such tortuous bureaucracy. There is a time-honoured way of selecting people who all want the same thing. It will produce a better social mix than the current system, it allocates an equal weight to everybody's choice, and it is indisputably fair between parents: it is a ballot.
This is the conclusion that a commission at the Social Market Foundation has come to. The system would work as follows. There would be just one admissions authority for the whole country, rather like Ucas, which handles university admissions. The 154 LEAs would lose their admissions function, as would all those schools that currently act as their own authority. Parents would be asked to list six preferences, in order. They would be able to choose any school they wanted. The only constraint on this choice would be that the transport subsidy, redeemable only on public transport, would be limited.
Once the applications were in, all schools with more first preferences than places would operate a ballot. If parents do not get their first choice, their child would be allocated to their second-choice school. Where schools get more second choices than they have remaining places, they would hold another round of ballots.
The same procedure would operate for third, fourth and fifth choices, if necessary. It is highly unlikely that it would come down to fifth choices. The majority of preferences would be settled by the second choice. Parents would have two weeks to confirm their acceptance of an offer. There would be a right to appeal only on the grounds that the ballot had been administered wrongly.
More than 92,000 appeals against allocation of school places were lodged in 2002-2003, a number which has risen 50 per cent since 1997. Professional appeals consultancies have sprung up to encourage and advise parents in appealing. This costly and time-consuming process could be avoided.
There is one potentially decisive objection to the balloting scheme: that it removes an advantage enjoyed by the voluble middle class. There is no doubt that it does. It would erode the premium that has grown on the value of their houses and it would mean that access to a good school could no longer be purchased with any certainty through the housing market.
But this is the perennial problem of social-democratic politics, not a specific deficiency of a ballot for school admissions. A ballot is palpably fair. And it would test us in the most important context - when our children are at stake.
Philip Collins is director of the Social Market Foundation, which will publish a pamphlet on school ballots this month