In the last in our Pushing 50 series, Alison Wolf suggests that attempts to widen participation in ÖйúAƬ need to focus on schoolchildren long before they embark on A-level study.

In 1963, the Robbins Committee advocated a "massive" expansion in ÖйúAƬ, from 4 per cent in universities and 8 per cent overall to 10 per cent participation in universities and 17 per cent respectively. It did so in the belief that "courses of ÖйúAƬ should be available for all those who are qualified by ability and attainment to pursue them and who wish to do so". By the early 1980s, we were already past Robbins's targets, with 670,000 students in our universities and polytechnics. In the past ten years, despite a one-third decline in per-student funding, participation rates have risen by almost twice that figure. Today, ÖйúAƬ enrolments, including ÖйúAƬ programmes at further education colleges, top 2 million.

"The good society desires equality of opportunity for its citizens," Robbins argued. Dearing, too, proposed that ÖйúAƬ should "play a major role in shaping a democratic, civilised, inclusive society". New Labour, single-mindedly pursuing higher participation, presumably sees 50 per cent as the point where all who are "qualified" to benefit may do so. So why, as this point approaches, does no one seem happy?

Universities matter now to more people than ever before, if only because degrees so straddle the job market. The world's desirable jobs are overwhelmingly "graduate only", even though study after study suggests that a quarter to a third of graduate jobs have moved from being non-graduate without any change in the skills they demand. In this country, graduates are, by middle age, earning twice as much as those of the unqualified and half as much again as those of people with 5 "good" GCSEs: and the returns on a degree were much higher in the 1990s than in late-1970s Britain. So university has become the necessary ante-chamber to the good life.

ÖйúAƬ

This is not, on the face of it, unfair. Granted, some people may lose because of their lack of formal diplomas: but selecting people on the basis of objectively measured attainments has been one of the foundations of modern democratic societies. A "career open to all the talents" replaces old patterns of jobs for the boys.

But the reality has turned out rather differently. The detailed "cohort studies" that track British children born in 1958, 1970 and 2000, enable us to look in detail at the experiences of successive generations. Between the first two groups - born 12 years apart - there is a significant and depressing difference. Jo Blanden and her colleagues at the London School of Economics and Institute for Fiscal Studies have found that the incomes of the 1970 cohort are, on average, twice as strongly determined by their parents' income (measured at much the same age) as they were for the 1958-born.

ÖйúAƬ

In other words, family circumstances had an increasing effect on life chances as the postwar decades advanced. The reason is the powerful effect of education, and especially ÖйúAƬ, on earnings - and the fact that access to ÖйúAƬ is closely and increasingly linked to family background.

Here and in every other developed country, expanding ÖйúAƬ has been a wonderful thing for the middle classes. True, the absolute chances of a British child from a working-class family attending university have increased substantially since the 1950s or indeed the 1970s - from about one in 50 to one in 20 to something close to one in six. But the chances for a middle-class child have grown far more in that same 50 years - from about one in ten to one in two for "social class II" children with teachers or middle managers as parents, and from one in five to pretty near universal for the children of the upper-middle classes.

In our universities, the overall proportion of undergraduates from non-manual homes is exactly the same as it was in the pre-Robbins era. This is partly 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. The UK is far from unique. In France, entrance into the most selective ÖйúAƬ institutions is dominated by just a dozen or so lycees - all of them state schools, but ones that cater to an upper-middle-class, metropolitan clientele. In Japan's best universities, the percentage of students from high-income homes has increased markedly since the 1960s. British observers often believe that America's top universities are far less socially exclusive than the UK's, but nothing could be further from the truth. Certainly, Harvard and others offer generous packages of aid and loans. But in a country where most of the best universities are also private, fewer than one in five American students from poor backgrounds attends a private institution, compared with almost half of those from families in the top income decile.

The Conservative-Labour consensus on fees at the time of the Dearing report was based partly on an awareness that a minority of the population was doing very nicely, at an individual level, from an education that cost them very little (?7 billion a year pours into the sector from the Treasury, via the ÖйúAƬ funding and research councils; another ?1.5 billion or so from other government sources). In the face of Liberal Democrat attacks, this government appears to be losing its nerve over funding. But the figures quoted here might seem to justify the new preoccupation: widening access. Not, however, as currently pursued.

Legislating for fairness through central targets, incentive funding, special initiatives and financial penalties is impossible and wasteful: not just in ÖйúAƬ but certainly, and harmfully, there. In this case it is also a smokescreen for the serious barriers to opportunity. Until recently, the main focus of government criticism has been the proportion of independent school pupils at top universities. But in terms of class profiles, this misses a major point: independent school students are overwhelmingly middle-class, but state-school entrants are not much less so. Moreover, school-sector targets - like any targets - quickly generate problems and distortions. Should universities get the same good conduct points for yet another student from Hills Road Sixth Form College in Cambridge (one of country's most successful institutions for A-levels and Oxbridge entry) as for one from the borough of Lambeth (where in 1999, only 31 students received even a single A-level A grade)? Is it fair to put Christ's Hospital - founded in the 16th century for poor pupils and still following this tradition - in the same category as Eton? Should the targets be on the basis of school populations or sixth-form cohorts? Should they be the same for a physics department - where independent schools provide one-quarter of entries and one-third of the A grades - as for business studies, where independent entries run at just 13 per cent?

ÖйúAƬ

One alternative policy with a "widening participation" tag is the postcode premium. Through it, institutions receive a bonus for every student they recruit from a "low-participation" area. In a rough and ready way, this directs money towards universities with above-average numbers of poor students. But as an incentive to change recruitment patterns it has huge limitations, since most admissions tutors do not even know which the relevant postcodes are. The ÖйúAƬ Funding Council for England refused to release the data, but then if it had, the premium probably would not have lasted the year. As anyone who looks around them knows, many low-participation neighbourhoods contain some very select dwellings indeed, with teenage offspring all set for a Russell Group education. Given the data, finding exceptions to the presumed rule and exposing its lack of "fairness" in a newspaper would be the work of an afternoon. If you want to help particular students, then it is those students you need to identify: you don't do it through institutional targets.

Hence, no doubt, the latest idea - giving more money per student for those whose entry grades are low. But this is guaranteed tabloid heaven. A-level grades cannot, unlike postcodes, be kept secret. So very quickly, we will have, on the left-hand side of the page, the rich, public-school educated layabout, whose two Es bring in the premium; and on the right, the dedicated, immigrant child from a single-parent family, whose four As do not. The lesson is the same, if consistently unpalatable. You cannot create fairness through rigid administrative rules or formulae set by central agencies - but in our distrustful, centralised system, that is what universities get.

Rhetoric over participation is, in any case, only partly a recognition that mass universities create their own inequalities and barriers. It is also a diversionary tactic that conceals other public failures. Whatever targets and incentives are introduced through Hefce, there cannot be an instant upsurge of students from poor homes, because the gap between middle-class children and the rest yawns wide well before university application. Children who make it into the sixth form with the same GCSE grades have the same chances of passing their A levels, regardless of parental class. But their previous chances of a given set of GCSE results depend enormously on their family, not least through the quality of the local school.

ÖйúAƬ

The result: only one in seven young people from poor and unskilled homes gets two or more A levels. That small minority is already overwhelmingly likely to go on to university, while the educational gulf at 16 means that there is no large reservoir of qualified, disadvantaged students for universities to draw on. With current population figures and school-performance trends, we could expand participation to 50 per cent tomorrow with no effect on the relative chances of children from manual and non-manual homes.

Most of the world's universities are public-sector institutions where entry depends entirely on certificates and marks. For example, in Germany, Italy, Belgium and the Netherlands, anyone with the correct certificate has the right to enter university; without it, they do not. In other cases, entry is strictly on marks: someone with a 4.58 average is automatically preferred over someone with 4.55. Australia, Sweden and the French grandes ecoles are all examples. These systems have the huge advantage of treating everyone in a formally equal and "fair" way: none of them, for example, can create the endless feelings of being badly treated that characterise Oxbridge entrance procedures. But they also cannot take into account individual circumstances or make allowance for the disadvantaged.

Before long, our increasingly politicised and semi-nationalised universities will surely follow this route. It will be simpler, less arbitrary and far less stressful than their current disputed autonomy. Politicians, tinkering with yet another formula in the face of media attack and disgruntled parent-voters, will see it as defensibly fair (and under control). But admissions tied strictly to academic achievement will do nothing on their own to change the social balance. Part of the problem is a deepening confusion over what our mass system is actually for. Junior ministers obediently trot out the line that, "we need more young people to go to university because it is an economic necessity". But we passed that point long ago. Obviously all developed countries need university sectors: but within the Organisation for Economic Cooperation and Development there is no obvious relationship between levels of university attendance and wealth, or growth rates, or productivity.

Margaret Hodge's view that A-level grades should not matter much implies a belief in a quite different theory: namely that degrees do not attest to substantive skills or standards, but are just about labelling people as intrinsically able and reasonably organised. The Japanese system is heavily biased towards such "screening". Students work like crazy in their teens, enter university (on their marks), do little work as students, get their degrees, and then are hired on that basis. But billions of pounds a year seems an exorbitantly expensive way to hand out gold stars for previous effort. So as we continue headlong towards the highest participation rates in Europe, should we not start to think seriously and fast about where we think we are going?

ÖйúAƬ

Alison Wolf is professor of education at the Institute of Education, London. Her book Does Education Matter? is published by Penguin on May 30, ?8.99.