Friday, December 04, 2009

The price of pretty good college 

A report shows a troubling trend continuing: Students are leaving Minnesota colleges with record debt.

Minnesota's 2008 graduate had an average of $25,558 in debt -- the sixth-highest in the nation, according to the Project on Student Debt. (The state ranked fifth last year.) The proportion of graduates with debt -- 72 percent took out loans -- puts the state fourth in the nation (up one spot from last year).

...Students in the District of Columbia, Iowa and Connecticut had the three highest average debt loads in 2008, while those in Utah, Hawaii and Kentucky had the lowest.

Here's the project that studied it. SCSU students leave here with a little less than $25,000 of debt, with about 2/3 of it federally guaranteed.

Is this a good return on investment? It turns out to depend on where you go, at least according to Caroline Hoxby's latest NBER study. Because students are now willing and able to select higher education in a national rather than regional market, colleges and universities compete more intensely for the best students. At the very best schools, that competition actually drives down the cost of an education that is borne by a student (or parents.) The remainder is paid by an endowment; each graduate in turn pays for good students to come after them. Even though the donations to the alumni fund are higher at the best schools, the total investment made on the best student, and the rate of return on the investment, generate both larger alumni funds and more income for those students.

So what happens to the remainder? It turns out the remaining schools are less selective, insofar as the accomplishments of an incoming freshman (based on NAEP scores) are now lower. Students have re-sorted themselves, she says, so that the top schools have a more uniform level of student achievement, and the best students who used to choose the nearby schools now travel further away. From the paper (ungated copy):
Average tuition paid as a share of student-oriented resources falls for every selectivity group, but the patterns differ. The least selective colleges start out with average tuition paid being about 60 percent of resources, and this statistic vacillates, ending up at about 44 percent. Most of these colleges are public colleges whose students have modest incomes. Thus, tuition paid is not a large share of resources because tax dollars make up the difference. Colleges at 51st through 60th percentile of selectivity have tuition paid fall from 88 percent of resources to about 65 percent of resources. This is a substantial decrease but students at such colleges (and other middling selectivity colleges) continue to finance most of their own investments in human capital through the tuition they pay.

In contrast, students at the most selective colleges paid tuition equal to only 46 percent of their human capital investment even in 1967. By 2007, they were financing only 21 percent of their investment through tuition!
Thus, the share you pay to go to the less-selective college is higher, and the difference between the amount of spending done by a very selective college and a not selective college (per student) widened from a ratio of about 4 to 1 to nearly 8 to 1 (even though all schools spend more per student nowadays.) Getting into a 99th percentile school just gets you access to so much more, and it seems to get you access to mostly high-quality classmates, where there may be complementarities.

This means that for our public four-year schools, there are fewer high-quality students around; our students are borrowing more to come here, and relatively speaking getting less of an advantage than they would have thirty years ago. Increasing specialization should lead to better delivery of high-quality education to those who can best use it. The paper is not as clear on whether the person borrowing $25,000 to come to a public four-year school would have been better off putting that money in the stock market (it appears not to be so for the most selective schools.)

Labels: ,