Discussion from Whitney Wilson ’90 of “Watch What We Do (and Not What We Say): How Student Aid Awards Vary with Financial Need and Academic Merit,” by Michael S. McPherson and Morton Owen Schapiro.


There are arguments which can made on both sides for either increasing or decreasing the amount of “merit” based aid relative to need based aid. This paper by Michael McPherson and Morty does not seek to address those arguments, pointing out that its not always clear whether a particular aid award is merit or need-based. The authors engage in a fairly detailed and – to the lay-person – complicated analysis of whether, and how, scholarship money (regardless of whether it is labeled as “merit” aid or “need-based aid”) correlates with ability to pay and “merit.” “Merit” is measured by means of SAT, a methodology acknowledged as imperfect but defended as the best available. Of course, as has been pointed out quite frequently on EphBlog, SAT scores are correlated with income. However, I am in agreement with the authors on this point – SAT is likely the best (or least most easily available and broadly comparable) measure of “merit” which can be used for a study like this.

“Ability to pay” is measured using one of two measures: family income or “expected family contribution” (“EFC”) (determined using a formula by the federal need-analysis system.) While EFC undoubtedly has its flaws (the formula used to create it is simply one group’s ideas of what a family’s financial priorities should be), it is probably a better tool for these purposes than income, as it seeks to take into account non-income based wealth, and other family specific factors.

One odd quirk – to my mind – in the analysis is the authors distinction between institutionally awarded grants and government awarded grants (such as Pell grants and state grants awarded directly to students). The result of this breakdown is that it makes it appear as though aid awards are not as strongly correlated with income (or EFC) as they are. That is, institutionally awarded aid is less correlated with income/EFC than government awarded grants.

On page two, the authors exclude athletic grants-in-aid from student aid grant calculations. Its not clear whether this was done because that was how the data were available, or because of a policy choice made by the authors. I don’t know how significant athletic scholarships are relative to all financial aid, so it may be an immaterial point, but from a theoretical perspective, I’m do not think it is appropriate to exclude it. I don’t know how including athletic scholarships would change the data (intuitively, I would think it would increase the correlation between aid and need, but I have no data to support that theory).

One item which will be of interest to some in the EphBlog community is a short section in which the authors argue “there is good reason for government policy to expand opportunities for private colleges to reach agreements on targeting their financial aid resources on needy students without risking antitrust prosecution.”

On the whole, the data presented in this paper is equivocal, suggesting a decrease in the amount of aid targeted at poorer students, but also indicating that net costs for poorer students have decreased at public institutions. However, the trends are quite complicated, and are based on data which is now almost seven years old. The conclusions do not appear to provide much assistance – or even suggestions – in shaping policy.

The problem may simply lie in trying to incorporate too many institutions in a single study. Since so much of the aid available is made available on an institution-by-institution basis, it may make more sense for each college or university to do the same analysis on its own financial aid practices. This would have the benefit of using more up to date information, and would generate a clearer picture of the institution’s practices (rather than the muddle generated by clumping thousands of them together). This in turn would allow each institution to see how its actions match up with its rhetoric, and to make adjustments as necessary. It certainly would be interesting to see how a place like Williams (with its no-merit aid and we-meet-all-demonstrated-need policies) compares with other colleges.

Print  •  Email