Economist Michael Kremer lectured at my department yesterday, reviewing the literature on evaluation of health and education aid in the developing world. It's an exciting literature, and one I may try to get in on. Among the points he made, not originally his own, one caught my notice.
Is it possible, wondered an economist, that people systematically underestimate the value of education? And how could that happen?
Well, we know people tend to live in neighborhoods and municipalities sorted by wealth. Thus, a poor child (in a poor neighborhood) is likely to know poor people. Some of them will have college degrees - but are more likely to be low-end outliers in income, and perhaps other measures. That is, the child's estimation of the return to college is biased by his implicit sampling method (those whom he lives near).
Does the same hold in a wealthy neighborhood? Yes! A child growing up in a wealthy neighborhood will know some non-college-educated people, but they are the ones who made good. Thus, he perceives less difference between college- and non-college-educated workers than he would if he sampled correctly.
If this story is true, than there is a positive externality to mixed-income neighborhoods and take-your-kid-to-work days. We already recognize that it's good to take poor children to educated workplaces - but maybe it's also valuable to take middle-class children to unskilled workplaces.