A Simple Fix for the Broken College Scorecard

Michael Morgenstern for The Chronicle

April 01, 2013

In an editorial last month, The New York Times encouraged students and their families to use the Department of Education's new online College Scorecard as a shopping guide. We wish we shared the Times's optimism about this new tool, but the cost information in the scorecard is no more useful to an individual student than the much maligned sticker price published in the college catalog. In addition, the measures the new scorecard uses to provide information to students can be gamed by colleges in ways that disadvantage middle-class families.

The first statistic on the scorecard is the average net cost of attendance. This measure seems easy to understand, until you realize that it is misleading for many students. The scorecard's net-cost measure is "the average yearly price actually charged to first-time, full-time undergraduate students receiving student aid [emphasis ours] at an institution of higher education after deducting such aid." At many institutions, students from relatively high-income families find out that they qualify for little or no aid. For them, this average-net-price score is quite useless. The sticker price offers better information.

Likewise, many students from poorer families will discover that they qualify for enough grant aid that the net cost to them is quite low, even at very expensive institutions. These students will find the scorecard measure a frightening overestimate of what they would be charged. Many may choose seemingly less expensive colleges, forgoing the opportunity to attend "pricey" ones that would actually be excellent matches for them.

While it might not be surprising that a "scorecard" does not do a good job as a shopping guide, one would hope that it does a good job of keeping colleges and universities accountable. Unfortunately, the scorecard also provides disincentives for colleges to behave in truly productive ways. If we look again at the net-price statistic, we see that institutions have a powerful incentive to reduce the number of students from middle-income families.

As mentioned earlier, the net-price statistic each college reports is based only on those students who receive aid. Here is an example to show how this changes a college's incentives:

College A has five students and a list price of $40,000. One student gets a $40,000 scholarship, and the remaining four students pay full tuition. The scorecard would list an average net price of $0 for College A, because only the student who received aid was counted. College B also has five students and a list price of $40,000. Instead of concentrating its financial aid on one student, College B gives each of its students an $8,000 scholarship. Even though both institutions have the same $40,000 budget for financial aid, the scorecard would show us that the average net price for College B was $32,000.

College A would show up as a great bargain, and College B would look quite expensive. This is despite the fact that the average net price, considering all students, is $32,000 at both institutions. College B's "mistake" was spreading its aid evenly.

A college intent on avoiding that error would refocus its recruiting on a small number of students who qualify for federal Pell Grants, which the institution would then top off with substantial amounts of additional grant aid. It would balance this by recruiting lots of students who are capable of paying full price. The students who become radioactive in this situation are the ones in the middle and upper middle of the income distribution. This group needs smaller amounts of aid, and enrolling large numbers of these students is a recipe for generating a high net price on the scorecard.

There are already lots of incentives for colleges to enroll more full-pay students. The federal government shouldn't be piling on.

Another number on the scorecard is the median amount borrowed. This, too, is a number that colleges can manipulate. For starters, any institution that offers some of its students grants that cover 100 percent of tuition and fees could lower its median loan score by reducing the aid package to 95 percent grant and 5 percent loan. Indeed, making more students borrow small amounts is a way to look better on the scorecard because it will skew the median.

Worse still, to lower the size of the average student's loan, many colleges may be sorely tempted to reduce the number of enrolled students who qualify for big loans. Again, they can do this by admitting a larger percentage of the student body from the upper end of the income distribution. That is contrary to the mission of most institutions, but these are the incentives.

With one simple change to the scorecard, the federal government could make it more relevant to prospective students while also eliminating incentives for colleges to game the system.

The change? Do not report a single number for a college's net price. Instead, break it down by family income. Report the average amount of grant aid from all sources that students receive if their families earn less than $40,000, between $40,000 and $60,000, and so on. That approach creates numbers that are more accurate and useful than a collegewide average. And there is no way for the institution to fudge the numbers by changing the proportion of its students who come from wealthy families.

Students and their families are eagerly looking for information to make their decision on which college to choose easier. They need a good shopping guide and a good scorecard. Unfortunately, the new tool from the Education Department is neither.

Robert B. Archibald and David H. Feldman are professors of economics at the College of William and Mary. They are the authors of Why Does College Cost So Much? (Oxford University Press, 2011).