Devlin's basic idea, that the cost of a college education will continue to rise, is unsupported, and his analogy with the auto-industrial era leaves much to be desired. Much about the cost of college will depend on how we redesign our educational provisions - including how MOOCs enter into the equation - and what will count as a "college education" or even a "first class college education".
That said, the rising cost of this "first class college education" raises questions that Keith does not address here. The first is, how will "society" pay for the rising cost of college, when the current cost is already ringing alarm bells? The second is, how many people will have access to college on Devlin's assumptions?
One current problem is the eradication of many previously "college level" jobs by outsourcing and technology. Fewer jobs + rising cost will both contribute to a declining rate of return on private investment on a college education.
Tuition dollars, however, now fund an increasing share of college education. The declining rate of private return implies that fewer people (at least those with an iota of economic rationality) will make the investment, driving down the flow of tuition dollars. Without these, how will colleges get funded?
So this leaves the question: what fragment of the young adult population should go to college?
With the declining private economic benefit, we have to ask about the public benefit of a college educated population. If there is a social benefit in an educated population beyond the economic growth provided by an educated workforce, then this cost should be borne by public investment, not private tuition fees. I don't see anyone arguing for a free or highly subsidized higher education anymore, because the emerging workforce no longer needs many of what came to be regarded as "college level" skills. The social benefit that most college professors assure us of appears invisible to policy makers and the masses of tax-conscious citizens.
If our economic policy and occupational arrangements emerge in such a way as to richly reward that declining number of people possessing college-level skills the economy really needs, then those few who attain those skill levels will be paid back with interest for their investment in education, and should pay the requisite private tuitions. But then what happens to those who will be excluded on that basis?
The most obvious answer is that we should (1) imagine new educational provisions for those priced out of college and unable to gain from college level skills in the emerging economy; and (2) ask those who eventually share directly in the wealth created by economic growth to contribute, through higher taxes, for expanded social insurance - and perhaps a social minimum of support - for those facing the risks of work in the contingent labor force: low wages, lack of job security and benefits and possible lifelong unemployment. .