If low-income students drop out, they end up worse off than before, with dreams broken and huge debts piling up.

Over the past three weeks, The New York Times has been running a series of articles on social class in America. Cynics on the Left will be muttering that it's about time, too, while hard-boiled types on the Right will be making derisive noises about wet liberals who go around making excuses for the idle and feckless whose poverty is their own fault and nobody else's.

In fact, we wet liberals are especially alarmed by one feature of the American scene that might be about to play out here. New Labour has been relying on ÖйúAƬ to play a leading role in increasing social mobility - perhaps "restarting" would be a better term because the postwar rise in social mobility ground to a halt in 1979 and has not picked up again. But the role of ÖйúAƬ may turn out to be a mischievous one.

For reality is at odds with public perception. Only a quarter of Americans think that social mobility has declined in the past two decades; almost half think it has increased. This is partly because many of the old barriers to social mobility have disappeared and partly because we think of the success of highly educated MBAs and their peers as a matter of "merit", without asking too many questions about who gets on to the educational fast track and how their "merit" is acquired. The other reason Americans believe in American social mobility is that they find it impossible to believe that social mobility in the US is no higher than it is in Britain; Britain has a Queen, for heaven's sake! You don't get to be Queen on the basis of merit.

ÖйúAƬ

Merit-based success certainly means that the transmission of advantage from one generation to another nowadays requires more than simply having a child to inherit your privileges. On the other hand, highly educated, well-organised, ambitious and energetic individuals are likely to marry one another, and they are likely to try to bring up children who resemble themselves. In this way, advantage - which is not exactly privilege - is hardly any less hereditary than before. "First, choose your parents" has always been good advice.

Here in the UK, we are trying to open access to ÖйúAƬ by just about every means possible. But the US experience with what happens to low-income students in ÖйúAƬ and - worse yet - the US experience with what happens when low-income students drop out suggests that expanding access against the background of the new top-up fee regime may do the opposite of increasing social mobility.

ÖйúAƬ

The script is simple enough. In the US, many students take a long time to graduate. The standard measure of success isn't whether you get through a four-year course in four years, but whether you do it in no more than five years. On that measure, two thirds of middle and upper-income students get a degree in five years, but only 40 per cent of low-income students. In pursuing a degree, poor students spend more time out of the mainstream labour force than better off students, and their grades tend to be worse because they have to undertake more part-time work. In that sense, their education costs them more for a poorer return.

As that suggests, poor students are also at greater risk of dropping out; and it is dropping out that is the real killer. The payoff from ÖйúAƬ - both the social and the individual return - is overwhelmingly from completed ÖйúAƬ. Students who drop out will incur expenses of one sort or another during their abortive attempt to gain a degree and, if they have met those expenses by borrowing, they will have to pay back their debts without the advantages a degree brings. For most Americans, this is not desperately expensive. Working-class students usually go to community colleges where fees are low, everyone lives at home, and everyone has a job while they study.

The drawback is that the main ladder out of working-class employment has been taken away - in the US, as in the UK, that means missing a hike in salary of something like a 40 per cent in your forties. And few students who drop out get a degree later on.

Here, things will surely be worse. In spite of the revival of grants for the very worst off, most students - those who aren't going to get ?4,000 a year at Oxford or Cambridge universities - will still need to borrow a minimum of ?3,000 a year. Given the speed with which the grant tapers off as family income rises, a lot of quite poor students will be borrowing much more than that.

ÖйúAƬ

The evidence is that most students who drop out do so early on, which is perhaps a mercy; but with the dropout rate already about 18 per cent, a whole lot of people can look forward to starting adult life with broken dreams and debts of ?5,000 or ?10,000 around their necks.

Alan Ryan is warden of New College, Oxford.