Kevin Carey, today, in the Chronicle:
Finally, and most important, the Obama administration should expand its vision of what publicly supported higher learning can mean. The MOOC provider Coursera recently announced that it would charge students relatively small sums, on the order of $100, for verified certificates of learning. The marginal cost of making a MOOC available on the Internet is $0, which is why those courses are free to students. By contrast, the marginal cost of assessing student learning, at a MOOC or elsewhere, is more than $0. Authentic, reliable evaluation costs money, and so someone must pay for it. Evaluation is also crucial to granting students college credits that lead to credentials.
Why can’t students use their Pell Grants to pay for MOOCs, even those offered by the world’s most prestigious universities? Because MOOCs are not colleges, traditionally defined. But that definition is an artifact of federal policy. Students can give their federal-aid dollars only to accredited colleges, and existing colleges control the accreditation process. So new colleges, eager for aid dollars, dress themselves up as traditional colleges, with commensurate expense to students, even if that makes no sense in the digital age.
I’ve been talking about this for a while, I know, but the big political impact of MOOCs is likely to be something along these lines. What is and isn’t an institution of higher education in this country right now is pretty bright-line. There’s not all that much room for disagreement.
Government funding for MOOCs would change that. The funding could come at the state level or the federal level.
Is this a good thing? It’s good and bad, in ways we can talk about later. But this is likely to be the primary political legacy of MOOCs, left for us to deal with long after MOOCs are gone: the number and variety of experiences and certifications that can be funded by government dollars will expand. And once that is true, capitalism will do the rest. As we stated above, it’s a mixed bag for students (and society), but it’s likely a substantial downward pressure on enrollment and tuition for existing institutions. I wonder how many institutions are ready for that?
Hapgood 
I’ve not got much to say on policy except my concern these become shovel ware courses.
But I can’t help but wonder about the assertion that it costs almost $0 to make a Coursera course available online- Audrey Watters relayed a quote from Koller that their average cost was $50k per course not counting for faculty time. Maybe that is extrapolating over the idea that 100k students register (yet maybe 1k finish)? Is there an illusion that assessment at scale is going to be as cheap as broadcasting?
Yeah — marginal cost of zero is not the same as zero (see, for example, graph here: http://en.wikipedia.org/wiki/Marginal_cost ). But here’s the thing — that only comes into play as a big factor at a certain level of scale. A digital textbook I write may have a marginal cost of zero after I spend a year writing it, and that’s pure profit at a certain point. But at an audience of dozens, marginal cost is pretty irrelevant, because it doesn’t recoup my loss.
And I think what we are looking at here, if we look at *completers* is that the power of marginal cost is not currently kicking in all that much, which is why the total cost per student is still quite high. It would be interesting to compute, given noncompletion, how many students would have to take this class to get down to a cost that is truly disruptive.