One of the oldest cartoon jokes is the hapless cat, dog or coyote lured into chasing its prey straight over the edge of a cliff. For a few moments, it runs in empty space, legs spinning ever more furiously; then it plummets, catastrophically but non-fatally, several thousand feet to earth. Is university recruitment about to do the same?
For years, the "graduate premium" has defied gravity. The same is true in the United States, where confident predictions of its elimination as the output of graduates increased have been falsified for 30-odd years.
While the pay of British academics has responded to the usual economic pressures - the larger the profession, and the higher the proportion of women working in it, the further behind the salaries of their former peers the pay of academics has fallen - the graduate premium apparently hasn't.
There are lots of plausible reasons why the graduate premium has not fallen, none of them conclusive. The great shift away from an economy in which 65-70 per cent of the workforce were manual workers took place a long time ago - between 1945 and 1975. But on an optimistic reading, we increasingly live in a "knowledge economy" and the jobs available for graduates have expanded to meet the supply of graduates to do them.
In precisely what sense jobs outside law, medicine, accountancy and applied research need graduate-level skills is debatable. Indeed, given the extent to which lawyers, doctors and managers of all kinds do most of their learning on the job, it is debatable across the board. But compared with, say, Egypt or India, graduate underemployment has not been a problem.
Another explanation for the persistence of the premium is less a matter of the positive benefits of a degree than of the damage done by not having one.
The old issue of "deskilling" has not been much discussed recently, but trends in management suggest that the pressure to simplify jobs (even white-collar ones), to remove discretion from lower-level staff, to give supervisors more absolute authority over their inferiors, and so on, has not slackened.
The great cleavage between order-givers and order-takers on which analysts such as Ralf Dahrendorf used to focus is as important today as it was in traditional industrial society. The premium reflects that cleavage; students may be pretty uninterested in higher education, but they want to be the right side of that gap.
If you ask the nasty question of how the graduate premium is paid for, you will not get an answer. Whatever else pays for the premium, it is not through an uplift in productivity caused by an expanding university population. The correlation between increases in graduate numbers and increases in productivity almost certainly runs in the opposite direction from that which ministers Bill Rammell and John Denham affect to believe.
A prosperous economy can afford to have 2 million young people unemployed for three years. This is not the same as in some developing countries where the economy cannot absorb all the available labour. If we closed all the higher education institutions, the young could be productively mopped up, although their elders might find the transition pretty difficult.
But what a degree does do is assure employers that a new employee is teachable. The CBI constantly complains that graduates can neither read nor write nor behave with even a modicum of social grace. Yet the CBI's members happily employ these illiterate and graceless creatures, and within a few years they say equally disobliging things about their successors.
If they were ineducable, and their vices incurable, this wouldn't happen. The fact that they have lasted the course attests to their ability to provide what their instructors ask for, to their ability to endure boredom, to persist at tasks whose point is either non-existent or at any rate not immediately obvious. The economy wants employees for whom deferred gratification is second nature.
But some of this might be about to come to an end. The latest numbers suggest that the gap between maths and sciences on the one side and humanities on the other is wider than ever.
And the gap between the premium from going to a high-end university and attending one at the bottom of the reputational heap has widened to the point where for arts graduates the premium at the bottom is either non-existent or too low to pay off the debt incurred for three years out of the workforce, let alone to compensate for three years' lost promotion.
This all makes good, if depressing, sense. The evidence of educability provided by survival at some HEIs is not impressive, and certainly less impressive than survival at some others. Science and maths degrees provide better evidence of an ability to stick to something difficult than do most arts degrees.
The Government's 50 per cent target has always been pretty goofy, given that it implies that students who haven't got five A*-C grades at GCSE, including maths and English, should nonetheless go on to higher education.
Once employers internalise that fact, they are bound to become sceptical about exactly what the possession of an arts degree demonstrates.
All the same, waiting for the Government to make good on its commitment to sensible apprenticeships, properly structured vocational education, and all the other things we have known that we need for a century or so is likely to be a dreary business.
So, if gravity does kick in at last, it will be less dramatic than it is at the movies.