Does the British higher education system have too many students?

In the past 15 years Britain has acquired a full-blown system of mass higher education. There are now twice as many 25-year-olds with degrees as there were 18-year-olds with A Levels in 1965. More than 40 per cent of 18-year-olds are set to enter higher education and the government's target is for 50 per cent to do so by 2010.

This should make many people happy. British education has for years been haunted by stories about our backwardness. But in higher education we have now closed the gap with Japan in enrolment and graduation rates; we have higher proportions enrolling than Germany or France, higher proportions graduating than Italy or Sweden. The US still sends proportionately more to college than we do, but we are about where they stood in the early 1980s in terms of enrolments, and are neck-and-neck on graduation rates.

However, British higher education starts the 21st century in a despondent mood. In May, the warden of New College, Oxford, Alan Ryan, leaving for a year's sabbatical at Stanford University, argued that "no rational person would work in the British higher education system". Salford, one of the country's oldest technical universities, became the latest in a long line to scale down their maths and science provision, in an attempt to close a large financial deficit. And Imperial College's rector, Richard Sykes, government adviser and former chairman of GlaxoSmith Kline, pleaded with ministers in the Financial Times to "do something radical" about a system in trouble.

Over £8 billion a year of taxpayers' money is now channelled into higher education. We must surely be getting something from our move to a mass system, whether it is faster economic growth, a fairer society or more cultured citizens. But we urgently need to clarify what it is that we are getting - and what, if things are going wrong, can be done about them?

The most important fact about university education is easy to spot. University pays - or, to be more specific, it pays the individual. On average, all over the world, university graduates are the ones who succeed, in terms of both income and employment. The average earnings gap between those with some higher education and those who never finished upper secondary school ranges from over half as much again in egalitarian Scandinavia to around double for the OECD as a whole, and to more than double in Britain and the US. Moreover, throughout the world, a growing proportion of desirable jobs are now graduate only. We needed more graduates "to avoid losing competitive advantage", as the CBI says; we now have the graduates, so everything is set fair.

Or perhaps not. It is true that the proportion of professional, technical and managerial jobs has increased greatly in recent years. And while non-graduate managers ran most of the developed world's companies in the first half of the 20th century, and often in the second half too, they no longer do so. Nonetheless, in the past 20 years, study after study has confirmed that many of the jobs - typically a quarter to a third - which were once non-graduate and are now graduate have made this change without any equivalent change in the skills required or used.

What is more, within developed countries there is no clear link between student numbers and growth rates, GDP per head or productivity. For example, Switzerland, at the top of the income tree, has the lowest university participation rates in the OECD; while the US, also near the top, has the highest. Big increases in university numbers are at least as likely to follow periods of rapid growth as they are to precede them: Japan is a prime example.

So when a minister asserts that "We need more young people to go to university because it is an economic necessity," he or she would be hard pressed to back up the claim. Employers sometimes do need graduate skills, but often they use graduate entry as a way of "screening" applicants: that is, targeting people who have shown application, and are assumed to be in the top half of the cohort intellectually. They may miss candidates who have both these qualities, and no degree, but finding them is too much trouble.

This is rational behaviour on employers' part, if not much to do with the "knowledge society". But if the 1.9million students and 172 full-degree institutions in our new system are not about labour market skills, might mass higher education at least be making Britain a more open and equal society?

Selecting people on the basis of objectively measured results, not by connection and family inheritance, has been one of the great achievements of democratic societies. Surely one result of mass university education is to deepen this trend. It may mean that degrees operate as a general entry ticket to many jobs that don't need them. But isn't this better than the alternative: management trainees who know a member of the board, articled clerks whose parents can pay the fees, promotion from the ranks dependent on favouritism?

This argument is rarely made by advocates of university expansion. And, as a victory for fairness, this one doesn't stand much scrutiny. In every developed country, expanding higher education has done less for equal opportunity than one might expect - while steering large subsidies towards the middle classes.

It is true that the absolute chances of a child from a working-class family attending university have increased substantially since the 1950s or indeed the 1970s - from about 1-in-50 to 1-in-20 to something close to 1-in-6. But the chances for a middle-class child have grown far more in that same 50 years - from about 1-in-10 to 1-in-2 for the "social class II" children with teachers or middle managers as parents, and from 1in-5 to pretty near universal for the children of the upper middle classes. The result is a student body in which the proportion of undergraduates from non-manual homes is exactly the same as it was before either the Robbins report expansion of "old" universities in the 1960s or the creation of the polytechnics (which became universities in 1992). This is in part because the manual working class now forms a smaller part of the overall population than in 1965; but it is mostly because of differential access.