Why I Worry about an Affirmative Action Ban

Two cases challenging the existence of Affirmative Action — one case against Harvard and the other against The University of North Carolina — may be heading to the Supreme Court, possibly together. Affirmative Action seeks to lessen racial and ethnic inequality in education by giving preference to certain underrepresented students — typically Black, Hispanic, and Native American — over others with similar admissions profiles. You know, sort of like Legacy Admissions for rich kids. Hey, where’s that Supreme Court case? Weird.

Anyway, Affirmative Action is obviously a controversial topic. And, I think rightfully so. In the preceding paragraph, I was pretty clear that it gives some groups preference “over others.” Members of those “other” groups — in these legal cases Asian and White individuals — claim that the policy discriminates against them. Given this controversy, it’s worth thinking through what Affirmative Action does.

The post proceeds as follows. First, I want to discuss how Affirmative Action works. Then, I want to consider how it influences college admissions and outcomes for underrepresented students. Finally, I want to discuss the issue of the policy’s fairness and unfairness. Some background reading that could help on that last bit can be found here and here.

What does Affirmative Action Do?

Two broad kinds of Affirmative Action exist: race-based and colorblind. The race-based approach gives an admission advantage to certain groups of students based on their race/ethnicity. A “colorblind” approach benefits students from lower-income backgrounds regardless of race. The Supreme Court cases target race-based Affirmative Action.

The simplest way to think about race-based Affirmative Action is by considering two students with the same SAT scores. An underrepresented student under an Affirmative Action policy will have a higher probability of admission given this score than a White or Asian student. This difference can be large. For example, one study found that a student from an underrepresented group was 15 percentage points more likely to gain admission than a similarly scoring Asian or White student.

The figure below provides an illustration. It comes from a book on Affirmative Action, “The Shape of the River“, and compares admission rates for White students versus Black students under two admission regimes. The actual regime, and then a “Hypothetical” regime without any Affirmative Action. It clearly shows that Black applicants had a higher admission rate, but would have had a much lower one without Affirmative Action.

Figure 1. Admissions Rates for White and Black Students in Five Selective Colleges

The Effect on Outcomes

In this section of the post, I want to walk through the impact of Affirmative Action on outcomes ranging from undergraduate attendance and graduation, to graduate school attendance, and ultimately to wages.

Undergraduate Attendance and Graduation

The purpose of Affirmative Action is to improve the educational outcomes of underrepresented students. Critics of the policy often argue that the reverse could happen. That by admitting students who might be less prepared into more selective schools, those students are set up to fail. Is this criticism of “mismatch” true?

Probably not. While it is true that Affirmative Action shifts underrepresented students from less selective schools to more selective ones, it does not seem to set up failure. A very recent study by Zachary Bleemer examined the issue in the context of the University of California system. California voters banned Affirmative Action in the late 1990s, and in doing so set up a “natural experiment” that Bleemer exploits. What happened to underrepresented students who applied to the UC system before and after the ban?

Bleemer’s answer is that the ban caused these students “to cascade into lower-quality colleges.” If a “mismatch” had existed, this cascade would be good for underrepresented students. But, it wasn’t. If anything, students were slightly less likely to earn undergraduate degrees after the ban.

This figure, again from “The Shape of the River”, illustrates one reason for potentially improved performance. If anything, more selective schools are better at graduating students. The figure divides Black students up by SAT bins, and shows graduation rates by level of selectivity. More selective “Sel-1” schools have higher graduation rates, even for students with low SAT scores.

Figure 2. Graduation Rates for Black Students Based on School Selectivity (Sel-1 is Most Selective, Sel-3 Least Selective).

Admission to Graduate School

So, banning Affirmative Action results in admission to lower quality schools for underrepresented groups. Anyone who has ever applied to graduate school knows the implication of going to a less selective undergraduate program — it’s much harder to get into graduate school.

Thus, it shouldn’t be surprising that the California Affirmative Action ban had an impact on graduate school completion. The Bleemer study found that attainment of graduate degrees by underrepresented students dropped by a statistically significant 1.3 percentage points. If that doesn’t seem like a lot, note that roughly 10 percent of underrepresented college applicants complete graduate school. So, a 1.3 percentage points decline is a large drop.

Again, earlier data available in the Shape of the River can illustrate why this happens. Figure 3 shows that all students — whether Black or White — are more likely to obtain graduate degrees if they graduate from the most selective schools (Sel-1 on the graph below). So, that cascade to lower quality schools is the likely culprit in the decline observed by Bleemer in California.

Figure 3. Percent of Graduates Attaining Doctoral or Professional Degrees (Sel-1 is Most Selective, Sel-3 Least Selective)

Wages

Well, you can guess what happens to wages once Affirmative Action is banned. A shift to less selective schools + less graduate school = lower wages. In California, underrepresented workers’ wages who had applied to the University of California system before the ban earned 4 percent more than those that applied after the ban.

So, the literature suggests that banning Affirmative Action shifts underrepresented groups to less selective schools. A ban makes these students no more likely to graduate with an undergraduate degree, and less likely to obtain a graduate degree. The net effect of this fact combined with undergraduate degrees from less selective schools is lower wages. If the Supreme Court bans the practice, it seems likely that the racial wage gap — a constant for 40 years — would likely be negatively affected.

At the same time, the policy is often criticized as itself being discriminatory. How should we think about this critique?

Affirmative Action: Fair or Unfair?

Here’s the thing about college admission, it’s a zero-sum game. Someone extra gets in…someone who would have gotten in doesn’t. A 2012 study by Peter Hinrichs gets to the heart of the matter. His study finds that eliminating Affirmative Action would reduce Black enrollment in Top 50 Public Universities by 1.74 percentage points and Hispanic enrollment by 2.03 percentage points. But, it increases enrollment in those same schools by 2.93 percentage points for Asian students and 1.43 percentage points for White students. To put it a little differently, some students worthy of admission when no Affirmative Action exists are hurt by the policy. This result is pretty clearly unfair to those students, who had the academic achievements they needed for admission.

Yet, I can’t help but think that this characterization is a bit simplistic. One goal of college admissions is to identify students with the ability and work ethic to succeed at college. On the one hand, it may seem obvious that a Black student and a White student with the same SAT score have the same potential in this respect. But, then again, such a similar outcome may mask inequality in actual underlying effort or talent.

Think about it this way. If the average underrepresented student goes to a lower performing High School compared to the typical White student, then isn’t that same SAT score more impressive for her? It reflects either a higher underlying natural ability or a greater work ethic to overcome a lack of resources. Either way, isn’t it appropriate to incorporate this information into the admissions process? Perhaps better metrics for this sort of ability to overcome adversity exist than race, but this thought exercise can make it clear that Affirmative Action may not be as inherently unfair as it seems.

Affirmative Action Going Forward

My best guess is that race-based Affirmative Action will be gone from public-college admissions, and perhaps private-college admissions in the coming years. What is my opinion on this? As the title of this post suggests, I worry. It’s not that I am in love with Affirmative Action. It strikes me as an attempt relatively late in someone’s life to right some wrong. And, Affirmative Action inspires animosity from one racial group to another, causing White people to feel they are discriminated against. And, in the case of Affirmative Action, some grievance on the part of White or Asian students passed over for admission makes perfect sense.

The reason I worry is that I highly doubt that the United States is willing to do the work to make policies like Affirmative Action unnecessary. As I’ve written above, banning Affirmative Action is likely to damage underrepresented students, who already lag considerably in terms of wages and attainment of graduate degrees. That damage could be undone. For example, the U.S. could ban Affirmative Action, but commit itself to policies like high-quality, Universal Pre-K that might undo some of the damage. Or, we could seek to restrict exclusionary zoning policies or punish real estate companies for steering, practices that keep underrepresented groups out of districts with higher-quality high schools.

John Roberts will ultimately play a big role in any decision. He has written: “[t]he way to stop discrimination on the basis of race is to stop discriminating on the basis of race.” I suppose that’s literally true. It also strikes me as hopelessly premature. Stopping Affirmative Action may end one policy that favors one group over others. But it would do so in a world that has thousands of policies that also favor one group over another, just in a direction that is much more tolerable to the majority of United States citizens.