Kevin Carey raises an interesting point:
Bryan is correct that the job market doesn’t provide a lot of partial credit for higher education. But in saying so, he seems to accept Murray’s blanket contention that people don’t graduate from college because they’re not smart enough to handle college-level work. This just isn’t true.
If this really is “Murray’s blanket contention,” then I don’t accept it. There are lots of factors that influence the probability of graduation: IQ, personality, distaste for school, all the facts that Carey mentions—“can’t afford to pay tuition, or they received a terrible high school education, or work and family demands intrude, or the college itself does a poor job of providing an engaging, high-quality education”—and more.
Nevertheless, if all we want to do is calculate the marginal student’s expected return to education, though, it doesn’t really matter why students don’t finish. If a completed year of education pays a 10% return, but the marginal student has a 6% chance of not completing a year of school for any reason, that student’s expected return to education will only be 3.4%. If we’re offering prudential advice to potential students about whether they should pursue a BA, we need to take all these factors into account, not dismiss them.
Think about it this way: If someone is thinking about opening a small business, wouldn’t a good friend warn them about the possibility of failure? Would it matter if the reason for failure was low IQ or “family demands”?
Admittedly, you could say, “We’re talking about social policy, not individual career guidance.” Maybe you think better policies would reduce the dropout rate. Even if that’s correct, though, when you do cost-benefit analysis of your favorite policies, you’ve got to take any remaining dropout rate into account.