If schools are to become "autonomous", what is to stop a complete free-for-all in admissions?
The new schools white paper's tortured attempt to answer this question tries to reconcile terms like "competition", "fair admissions", "strategic leadership" and "choice". But it does not adequately resolve the central problem, which is that the more you give schools autonomy and urge them to compete, the greater the pressure on them to define their own "markets" by seeking to admit the easiest to educate.
Perhaps the starkest consequences of school autonomy in admissions arose from New Zealand's market reforms of the 1990s. Every oversubscribed school was allowed to formulate its own admission policy provided it did not contravene basic human rights (for example by being overtly racist).
This meant, for example, that there was nothing to stop a school from favouring children from an asymmetrical catchment area drawn to include the most privileged neighbourhoods. It also made it possible for schools to use criteria such as interviews, and to give no priority to local children.
The result appears to have been a rise in the tendency for well-off, white families to access the best schools and in some cases for children to be denied entry to any local school.
A reform of this system by a new government in 2000 obliged schools to give priority to local children.
One thing that we should have learned from this story, and our own messy experience of school choice in the past 15 years, is that publicly-funded schools cannot be left to define their own markets.
However autonomous we wish to make them, it will not work to allow schools to pick and choose whom they educate. If we do this under a system of governance dominated by parents' interests, the result will be exclusive school communities using public money to support their own sectional interests.
Without a co-ordinated system ensuring that places are allocated fairly, there are bound to be huge inequities in access to good schools. Many children will also fall through gaps in provision and be prevented from attending a school near their home.
The Government recognises these considerations and acknowledges the importance of "fair admissions". The role of overseeing such fairness is to be transferred from the independent school organisation committees to local authorities, while individual schools can control their own admissions. The idea is to put the local authority in some form of strategic role, and to reduce its role as a provider.
This remains a messy arrangement as some places will continue to be allocated by the local authority, and, when it makes its own proposals for admissions, decisions will be overseen by the independent schools adjudicator.
But can any regulating body adequately co-ordinate school places offered by a bewildering array of competing semi-autonomous institutions?
Each child is physically able to access only a limited number of schools, which are now defining themselves by various overlapping criteria: their faith, their specialism, their geographic intake. In such circumstances, there seems to be no systematic way of ensuring that every child ends up in a suitable school.
Local authorities will have some power to ensure that schools follow the admissions code of practice, introduced in 2003. This does constrain the extent to which schools can use unfair criteria to allocate places, or even ignore completely the advice of the local admissions forum.
Yet school places cannot be planned or co-ordinated through negative measures. The partnerships among local schools needed to produce decent opportunities for all a community's children rely on voluntary actions and goodwill. Both politics and economics could reduce the chance of this happening.
The political language of competition, which has been emanating from central government from Keith Joseph in the 1980s to Andrew Adonis, continues to emphasise the importance of competition as a central dynamic to improve schools. We should not then be surprised if schools put their own interests before those of the wider community.
And competition will be made all the sharper, at a time of falling pupil numbers, by the knowledge that keeping up one's reputation and client base will become ever more important for the economic survival of each individual school.
Donald Hirsch is an international consultant on education policy