We need more higher education. That is a statement with which few in education policy circles would disagree. It is why, in the UK, universities¡¯ income was protected in 2010, even as the government slashed so much else.

The price, of course, was the tripling of undergraduate tuition fees in England. But that didn¡¯t have the disastrous consequences for social equality that many feared. The poorer students turned out not to be put off, presumably taking the view that the investment would still be worth it. After all, the lifetime graduate earnings premium is widely quoted to be in excess of ?100,000 ¨C and, if it doesn¡¯t materialise, loans are written off.

The calculation is slightly different in the US, where fees are not capped and loan terms are not so forgiving. There, too, people have largely been prepared to take on colossal debt to pay for college, but many policymakers fret that prices are now rising beyond all reach of the poorest. That point is eloquently made by Martha Kanter, Barack Obama¡¯s former under secretary of education, in this week¡¯s opinion pages.

Another concern is that, in reality, the graduate premium varies enormously according to course and institution. The conviction that university applicants need course- and institution-specific data on future earnings to make informed decisions lies behind the UK government¡¯s Longitudinal Education Outcomes project, which uses tax data to probe graduate earnings. Last week¡¯s on law graduates highlighted the issue: while the median salary of University of Oxford graduates is ?61,500 five years after graduation, those from the University of Bradford earn just ?17,500.

³ÉÈËVRÊÓƵ

Universities have argued that when the data for all subjects are published early next year, they should be benchmarked according to student background and local labour market conditions. That makes sense if their purpose is to inform league tables. But, given that tuition fees are not similarly benchmarked, it is surely the raw figure that will most interest would-be students.

The fact that benchmarking is an issue at all relates to another important point underlined by the LEO data (as well as the Social Mobility Commission¡¯s recent ): that higher education, by itself, is a fairly weak lever in social mobility terms. The students at top universities, who are disproportionately from wealthy backgrounds, do better in life than those from lower-ranked ones even when they study the same subject ¨C probably because the support networks that often helped them get there also give them a leg up in their subsequent careers. No amount of higher education expansion, by itself, will change that.

³ÉÈËVRÊÓƵ

None of this is to say that more higher education is not desirable. Employers need educated workers and, as Kanter sets out, education is a good for all kinds of other reasons, social and economic as well as personal. But it does raise questions about how much individuals should pay.

UK universities understandably resist individual pricing of courses because of the implications for their ability to cross-subsidise more expensive ones. One solution might be to delink the income they receive for running particular courses from the fee charged. Perhaps course-specific tuition fees, based on earnings potential, could be paid to the government rather than to universities (in effect converting loans into a variable graduate tax), with the government then paying the university the actual cost of running the course. For some courses, that would be less than the fee, for others it would be more: a cross-subsidy of sorts.

Such a mechanism might only exacerbate fears, raised by the ³ÉÈËVRÊÓƵ and Research Bill, about government control over which courses universities teach. But something along these lines might be fairer to individuals, and perhaps even move efforts to boost social mobility on to more fruitful territory.

paul.jump@tesglobal.com