I think those who spend their lives in academia will tend to underestimate the return to education.
As Card points out the return to education is higher for those who have limited access. Moreover, academics tend to enjoy the education process and will seek education beyond the point where the marginal pecuniary return is exceeds the marginal cost.
This will lead academics to conclude that education doesn't pay.
This is an interesting point - and one I'll address at length in my book. But I was talking about something else: using personal experience to estimate the practicality of education - the extent to which you take the stuff you learned in class and actually use it in real life. And on this topic, it's hard to deny that professors' experience leads them to overestimate. After all, I use my undergrad Industrial Organization every time I teach IO. How often do my fellow classmates from Theodore Keeler's 1992 IO class do the same?
P.S. Karl, if you're reading this, please send me your email address - you're hard to track down. :-)