Weeds in the Ivy: college admissions under preference constraints

ABSTRACT

In a series of cases spanning more than three decades, the courts have attempted to establish boundaries on the permissible use of racial preferences in college admissions. Proponents of these policies believe that race-based preferences are needed to create a diverse student body that facilitates effective learning and social inclusion. Opponents of such policies contend that racial preferences are inherently discriminatory and eliminating them would yield a more able student body. Whereas race-based preferences have garnered the most attention, elite colleges regularly employ other types of preferences, including those for alumni and talented athletes. To inform this important policy debate, we develop a simple model comprised of a rational college administrator that maximizes a linear combination of student body ability and the college endowment through the choice of race, legacy and merit admission shares. We find that relaxing the racial-preferences constraint can produce a ‘less-able’ student body even when the college administrator places greater weight on student body ability than she does on the college endowment. The change in admissions policy may serve only to increase the number of admissions that can be ‘sold’ to wealthy alumni through legacy preferences and thereby foster the growth of weeds in the Ivy.

KEYWORDS:

• Education
• discrimination
• racial preferences
• legacy preferences

JEL CLASSIFICATION:

• I20
• I28

Acknowledgements

We are grateful to Glen Robinson for many fruitful discussions on the topic of this article and to Soheil Nadimi for expert research assistance. We thank the editor and two anonymous referees for helpful and constructive comments that substantially improved the article. Any remaining errors are our own.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ See Jeffries (2003) and Pacelli (2011) for a thought-provoking historical account of the legal rulings circumscribing the use of racial preferences in college admissions. See also Regents of the University of California v. Bakke, 438 U.S. 265 (1978).

² The terms colleges and universities are used interchangeably throughout the analysis.

³ The Court observed that ‘Enrolling a “critical mass” of minority students simply to assure some specific percentage of a particular group merely because of its race or ethnic origin would be patently unconstitutional’ Grutter v. Bollinger (2003Grutter v. Bollinger (2003, 3).

⁴ See Loury (2002) inclusive of Tables 1–4 and Figures 1–2 in the Appendix.

⁵ It should be further noted than the term ‘most qualified’ in terms of applicants is a contested criterion. In Grutter Justice O’Connor upheld racial preferences on the grounds that it produced a more diverse student body. This is not a criteria defined by the individual’s self-standing merits but by her contribution to the student-body as a whole and to the learning experience derived by all students. Justice O’Connor reasoned that a diverse student body ‘promotes learning outcomes and better prepares students for an increasingly diverse work force, for society and for the legal profession’ Grutter v. Bollinger (2003Grutter v. Bollinger (2003, 3).

⁶ On 24 June 2013, the high court returned the case to a lower court without affecting broad constitutional changes with respect to how affirmative action policies would be administered. Fisher v. University of Texas at Austin, Supreme Court of the United States, 2013 U.S. Lexus 12 October 4701, 2012, Argued, 24 June 2013, Decided. On 15 October 2013, the high court heard oral arguments in Schuette v. Coalition to Defend Affirmative Action, 133 S.Ct. 1633 (2014), another case involving affirmative action in Michigan. This time around, the issue before the court is whether a state violates the Equal Protection Clause by amending its constitution to prohibit race-based and sex-based discrimination or preferential treatment in public-university admissions decisions. On 29 June 2015, the Supreme Court agreed to examine admission’s practices at the University of Texas, which is essentially a rehearing of the Fisher case. http://www.wsj.com/articles/u-s-supreme-court-to-consider-affirmative-action-at-public-universities-1435586162. As this article was going to press, the United States Supreme Court issued a decision upholding the right of the University of Texas to consider race in college admissions. Specifically, the Court held that the race-conscious admissions program in use at the time of petitioner’s application is lawful under the Equal Protection Clause. See Fisher v. University of Texas, 2016 Decision.

⁷ In Parents Involved in Community Schools v. Seattle School District, the Supreme Court rejected the argument that mechanical racial rebalancing to obtain specific numerical targets in public schools is a legitimate governmental interest. In addition, the court ruled that the constitutionality of programs designed to enhance diversity along the lines of Grutter were ‘unique to institutions of higher education’.

⁸ Long (2004a) finds that after eliminating affirmative action in California and Texas, the gap between the numbers of SAT score reports submitted by non-minority and minority students to in-state, public colleges widened significantly. See also Dickson (2006). Card and Krueger (2005) find little or no evidence that the elimination of race-based admissions preferences in California and Texas had any effect on the decisions of highly qualified minorities to submit their SAT scores and presumably apply to the selective institutions in the two states.

⁹ Banning race-based admissions does not imply that universities will cease their efforts to achieve diversity targets through other means. For thought-provoking analyses of the efficiency consequences of achieving diversity through non-race-based preferences, see Chan and Eyster (2003), Fryer, Loury, and Yuret (2007) and Epple, Romano, and Sieg (2008).

¹⁰ As Schuck (2002, 26) observes, ‘A preference for legacies and athletes, for example, may maximize the alumni contributions and loyalty that in turn support the institution’s academic mission’.

¹¹ Espenshade, Chung, and Walling (2004, 1431) find that ‘The bonus for African-American applicants is roughly equivalent to an extra 230 SAT points (on a 1600-point scale), to 185 points for Hispanics, 200 points for athletes, and 160 points for children of alumni’. See also Golden (2006) for an account of the pervasive use of legacy preferences in admissions at prestigious colleges.

¹² There is another interesting perspective on the issue raised by an anonymous referee. As highlighted in the review by Kniffin (2007), parental education plays a very important role in college admissions, doctoral program admissions, and becoming faculty members at national research universities. While we do not consider the parental education dimension (other than the legacy admissions) in this article, ‘more weeds’ necessarily means ‘less non-weeds’ in a static framework. An outstanding question concerns whether over the long run, in a dynamic framework, the aggregate societal gains associated with departing from pure merit-based admissions passes the cost-benefit test. This important and thought-provoking question is beyond the scope of the current article.

¹³ We use the terminology ‘de facto relaxation of the racial-preferences constraint’ because the Supreme Court has been abundantly clear that both explicit quotas for minorities and the automatic awarding of extra points in the admission process solely due to minority status are patently unconstitutional. Gratz v. Bollinger, 539 U.S. 244 (2003). It is noteworthy, however, that Chief Justice Rehnquist observed in his dissent in Grutter that the Michigan Law School admissions process essentially made use of quotas. Grutter v. Bollinger, 539 U.S. 383. In addition, some scholars have suggested that alternative rules, such as the Texas Top 10% Rule (the top 10% of graduates of Texas public high schools are automatically admitted to the University of Texas) may simply be quotas in disguise (Banks 2007). Long (2004b) finds that race-neutral programs like the one instituted in Texas are not likely to produce minority representation in top-tier public institutions of the same magnitude as race-based admissions policies. These findings are consistent with those of Cancian (1998). The top 10% rule in Texas was passed by the Texas legislature in the aftermath of Hopwood v. Texas, a case in which the 5^th Circuit ruled that diversity in law school admissions was not a compelling governmental interest.

¹⁴ Fu (2006) shows that handicapping the stronger applicants for admission may increase the competition for admissions. In this manner, racial preferences may help top-ranked colleges increase the academic quality of their students.

¹⁵ This article differs from Li and Weisman (2011) and Weisman and Robinson (2012) in formally demonstrating that this result can occur as an equilibrium response by college administrators to a relaxation of the racial-preferences constraint.

¹⁶ A precise definition of student ability/quality is not essential for our purposes. It is sufficient that the ability metric, which is a proxy for student quality, be a continuous numerical variable that can be observed. Examples include IQ, SAT and GPA.

¹⁷ These include Blacks, Hispanics and Native Americans.

¹⁸ This ordering follows from the assumption that merit-based admissions are, on an average, likely to have greater earning power than race-based admissions.

¹⁹ We thank an anonymous referee for suggesting this equivalent formula that is easier to follow.

²⁰ As Loury (2002, 132) observes, ‘Selective institutions will naturally try to reject the least qualified of the otherwise admissible nonblack applicants while admitting the most qualified of those black applicants who would otherwise have been rejected’.

²¹ The asterisks indicate equilibrium values.

²² The fact that $w_L(\alpha )+w_M(\alpha )=1$ follows directly from Proposition 4(iii). In addition, it follows directly from Lemma 2(i) and 2(ii) that $\hspace{0.17em}\hspace{0.17em}{w_L}^{\prime }(\alpha )<0$ and ${w_M}^{\prime }(\alpha )>0\hspace{0.17em}.$ μ.

²³ All simulations were performed using MATLAB.

²⁴ Recognize that the simulation results are biased against finding that a relaxation of the racial-preferences constraint results in a less-able student body. This is the case because it is assumed that the administrator moves sequentially from more-able to less-able legacy admissions. It is perhaps more plausible to believe that a larger ‘bribe’ would induce the administrator to admit a markedly less-able (as opposed to a marginally less-able), legacy applicant consistent with Proposition 1.