Capitalisn’t: College admission (in)justice—Harvard and beyond

Oct 24, 2019

Sections Public Policy

Getting into the right college is arguably more important than ever, which has put the justice or injustice of admissions processes in the spotlight. On this episode of the Capitalisn’t podcast, hosts Kate Waldock and Luigi Zingales give a fresh perspective on a recent admissions controversy involving Harvard, explain its implications for college admissions in general, and ask whether the way elite universities choose their students is an example of capitalism working or failing.

Luigi: So, Kate, we have to admit that we failed two weeks ago when we tried to predict the Nobel Prize.

Kate: Yeah. Sorry, guys and gals. I don’t think we came close at all, because none of the people who won the economics Nobel Prize were mentioned on our Nobel Prize predictions episode, which is a little embarrassing.

Luigi: But, in our justification, two things. One is they went really young. This is the youngest Nobel Prize in Economics in history. So, I think that’s pretty remarkable. And second, they did go very much for hedgehogs, people who have dedicated all their life to one topic and a very important one, which is how to fight poverty.

Kate: Yes. So many congratulations from the Capitalisn’t team to Abhijit Banerjee, Esther Duflo, and Michael Kremer, who are this year’s winners of the Nobel Prize. 

Getting to the right college these days is arguably more important, or at least more competitive, than ever. There’s increasing demand for college educations, but the number of spots at the top universities, at least in the United States, has been relatively fixed.

Luigi: So, what we want to discuss today is a legal case that involves admissions and biases in admission.

Speaker 3: A lawsuit challenging Harvard University’s applications process may be headed to the Supreme Court.

Speaker 4: So, this case started actually back in 2014, when a group of Asian American students sued the university, saying that they were being held to higher academic standards in admissions than white and other minority students.

Luigi: Since Kate is a Harvard alumna, we should start by asking her about her experience.

Kate: For those of you who have never seen me, I am half Asian and half white. My mom is Korean and my dad is of mixed Irish, English, German heritage. And when I was applying to college, I remember very distinctly that you had to check some combination of boxes to indicate your ethnicity, and I checked white, because my last name indicates that I’m white. I am not sure that they would have otherwise been able to tell. And most importantly, I was afraid that if I put Asian in any way, that there would be some sort of bias that would hurt my chances of admission.

Luigi: And this case is not just about Harvard. The same group has been suing other universities, and this is an issue about admission policy in America in general that can have consequences for all our universities and also important consequences for the mobility of our society. And so, understanding whether there is a racial bias in this process might be very important for understanding whether the opportunity of social mobility is biased one way or another.

Kate: And even though in the first round of this case being heard before the courts, the judge decided in favor of Harvard, this case will be appealed. It’s going to continue to go to appellate court. And some even think that this has the possibility of reaching the Supreme Court.

Luigi: So, from the University of Chicago, this is Luigi Zingales.

Kate: And from Georgetown University, this is Kate Waldock. You're listening to Capitalisn’t, a podcast about what’s working in capitalism today.

Luigi: And, most importantly, what isn’t.

Kate: One of the major flashpoints in the Harvard trial was competing testimony from two economists,  one who claimed that Harvard’s admission process is discriminatory and another who disputed those facts.

Luigi: Those two depositions are actually quite interesting pieces of economics.

Kate: Yeah. This is the economics equivalent of a major sporting event. The playoffs. This is the econ playoffs.

Luigi: What we’re going to do here is we’re going to take the two sides, Peter Arcidiacono and David Card.

Kate: Luigi, why don’t you discuss the plaintiff’s case, Professor Arcidiacono?

Luigi: Yes, Arcidiacono, very good.

Kate: OK. There you go. I did a pretty good job of saying that, but I didn’t want to have to say it over and over again, which is why Luigi is going to take that side.

Luigi: So, first of all, it’s important to understand, if you look at the raw data, Asian Americans and whites tend to be admitted at the same rate, in fact, slightly more Asian Americans than white. The problem is that the pool of applicants is of very different quality. So, the average Asian American has an SAT score which is 25 points higher than the typical white applicant, 154 points higher than the typical Hispanic applicant, and 218 points higher than the typical African American applicant.

From the point of view of academic performance, I think that the Asian Americans are much, much better. In fact, Arcidiacono establishes that if Harvard relied exclusively on the academic side, the Asian American share of its domestic admitted freshman class would be over 50 percent. While, in fact, it’s only 22 percent. And the case for discrimination is based exactly on this issue of given the quality that the average Asian American applicant has, he or she should be admitted at a much higher frequency.

Take an average Asian American who is male. On average, you have a 25 percent chance of admission. However, if you were to change the race of the applicant, keeping all the academic scores the same, into white, then your chance of admission will go up to 36 percent. If you were actually Hispanic, that would go up to 77 percent. If you were African American, 95 percent. That's a pretty stunning difference. 

Now, where is the source of the bias coming from? Apparently, the big source of difference is in this personality score. We have some objective measures like your SAT score, and even the extracurricular activity tends to be more objective, in the sense that if you are a champion in chess, I think that stuff is relatively easy to assess in an objective way. Your personality is something much more difficult to explain. 

The only thing I was able to find is there are only six categories. You are either outstanding or very strong, generally positive, bland or somewhat negative or immature, questionable personal qualities, or worrisome personal qualities. Now, the remarkable fact is that, when you look at the interviews with alumni, it doesn’t look like Asian Americans are performing worse than the other groups. But, in the internal rating, Asian Americans performed worse than the other groups, especially conditional on the academic credentials.

So, in general, if you have a better GPA and a better SAT, you tend to have a higher personality score, probably because you know how to express yourself properly in English and you’re not completely weird. This calculation does not seem to hold as much for the Asian Americans. So, one possibility is that Asian Americans have some strange, weird personality not present among whites and Hispanics and so on and so forth. The other possibility is that Harvard tends to discriminate in this particular arena because it’s so subjective. So, you can’t really discriminate on SAT scores because it’s a number, but you can think about either an explicit bias or an implicit bias.

Kate: But what is Arcidiacono’s position on the percentages of each group, of each racial group that are admitted?

Luigi: If you look at the percentage of African Americans that are admitted every year, that percentage looks awfully identical year after year. He’s 99.8 percent sure that level of similarity across years cannot be driven by chance alone, i.e. that Harvard has an explicit strategy of targeting a minimum flow of African Americans, and that the flow is pretty constant over time. 

So, Kate, now that I rested the case of the plaintiff, what is your response? What is your Card response?

Kate: OK. Well, so first I should say that Professor David Card, you can tell that a lot of his testimony is in direct response and not in the same sense, I guess, it doesn't stand as much on its own, it's more of a reaction. Where Professor Diacono presents six models of whether or not—

Luigi: Arcidiacono.

Kate: Dammit!

Luigi: Call him Archie.

Kate: I don’t want to. 

So, where Arcidiacono really focuses on this academic part of the person, Card comes in and says, “Look, that’s just a relatively small part of the overall calculus.” And, in fact, Arcidiacono is trying to spin this story where Harvard should care mostly about these academic elements, and to the extent that you can’t observe anything else, like the personality parts, the softer parts, the leadership skills. Those are all harder to observe. So, Arcidiacono tries to weave the story of, “Well, because the Asians are stronger on these measurable characteristics of academic excellence, they’re probably also better along those other dimensions.”

Card comes in and says, “No, that’s just the opposite of the truth.” So, not only should you not be focusing on these solely academic measures, as Arcidiacono does in four out of six of his main models, but as you start adding more and more nonacademic variables, this perceived bias against Asian Americans shrinks more and more. To extrapolate, if you had all of the information about these unobservable characteristics of these applicants, then presumably, if you kept adding them into the models, then that bias would shrink to nothing. So, I would say that’s the—

Luigi: That’s a huge extrapolation.

Kate: Maybe that’s a huge extrapolation, but he does show that the bias shrinks, and I think that that would be his major point: what are the assumptions on these unobservables? Arcidiacono is coming from a very different place than Card. 

Now, the second major point that Card makes is that Arcidiacono just misses a lot of stuff in his model. There’s a lot of important data about each candidate that applied to Harvard that Arcidiacono could have thrown into the model, which predicts whether or not that person got into Harvard, but he chose not to. For example, what did the candidate’s parents do? Did the parents attend Ivy League schools? Is the applicant from a rural area? How much did the applicant work at another job while in high school or towards the end of high school? What’s the median income of a high school that he or she went to?

Luigi: Do they eat sushi or not?

Kate: I just had sushi for lunch. How did you know that?

Luigi: No. I’m just saying that some of these characteristics are a way to get at your ethnicity without asking the question. 

Kate: I know.

Luigi: So, it’s a bit unfair to say, “I don’t discriminate on the color of your skin, but just on curly hair.” And it’s not perfect, but it’s pretty close.

Kate: No.

Luigi: I eat sushi, too.

Kate: Yeah, OK.

Luigi: But I’m not Asian.

Kate: Fair enough. But—

Luigi: Actually, you know that sashimi was invented in Italy, but that’s a different story.

Kate: I somehow highly doubt that. But we can hash that out on a different episode.

All right, Luigi. Let’s get into a debate about which economist is right.

Luigi: I prefer Arcidiacono’s position. There is an intrinsic arbitrariness of what is merit. However, my perception is that in a lot of cases, people define merit in order to define who wins. It’s not the other way around. In particular, I grew up in Italy where knowing Latin was very important. And why? It’s because the elite were going to a classical high school, where they’re spending five years of their life learning Latin and ancient Greek, stuff that is super useful in life, and only the rich kids could have the time to waste to attend this school. So, if you are not particularly well off, you would go to a more practical school. So, that was a way to signal how rich your parents were, not how smart you were. Clearly, this could be constructed as merit—knowing Latin is an element of being a cultured human being and being very rounded and not just being a math geek—but in fact it was very highly correlated with parental income.

I feel that you can easily reconcile the two testimonies under one story, and the story is, Harvard introduces very subjective criteria, precisely because they want to choose the composition of the class the way they want. And I think actually Arcidiacono’s complaining about this, and David Card says, “No, no, that’s part of the objective of Harvard, so there’s no discrimination, because people are rated differently in terms of personality, and personality is very important.”

But nobody wants to say why it’s so important and what is the evidence that personality is so important. One thing I’m surprised by is that we have psychological tests, we have potential evidence exposed on success in the labor market or stuff like that. Can’t we see whether people that rate higher on personality did indeed have higher personality in an objective way? I don’t think there is any evidence of that.

Kate: I sort of agree and I sort of disagree with what you’re saying. Some of the elements that go into the personality measure are things like how well-written is your essay and what sort of recommendations did you get from your college counselor? And I think it’s fair that that’s important information. I don't think it’s the same as, did you do well in Latin? I understand that these are also pretty fuzzy things. They’re subjective criteria that can be easily manipulated. 

But I think somewhat important is the fact that on the personal scale, yes, white students do better than Asian American students, but the differential is much smaller than the differential in athletic abilities. When it comes to athletic ability, whites do much better than Asians. And so, it’s not so much the personality measure that I take issue with. It’s this athletic measure, which is quite important when it comes to admissions processes. 

And I want to say that this is an athletic measure that doesn’t apply to the athletes that are recruited to Harvard. Keep in mind that we’re excluding the professional or quasi-professional athletes from the sample. So, we’re just considering regular kids who are probably not going to play on the varsity team. And yet, this athletic measure is still one of the four crucial components. And Harvard’s justification of this criterion is that if you play sports, you acquire team skills and you learn how to be a leader in certain cases, and you learn how to take risks and put yourself out there and be tough or something like that. And so that’s why they’re including it for everybody. But that’s the one in which there’s the biggest favoring, I suppose, of white students over Asian Americans.

And I think that this athletic measure is just, for the most part, ridiculous. I think that that’s the closest equivalent to what you just talked about with the inclusion of Latin. I think that that’s one where kids who grew up in the suburbs and played lacrosse or field hockey or something, it’s easy to game this measure. I think that certain kids are much more likely to have the opportunity to play sports based on their socioeconomic background. And I think it’s a measure that for the most part should be excluded from the analysis.

Luigi: It’s funny you mentioned lacrosse, because until recently I did not even know it was a sport. I think it exists only to give an advantage to insiders in college applications, because only if you are a white-bread American do you actually know what lacrosse is, let alone can you play it.

Kate: Yeah, well, guess who was captain of Harvard’s club lacrosse team for two years.

Luigi: You?

Kate: Yeah.

Luigi: Wow. Congratulations.

Kate: Thank you. But yeah, so, do you agree with me about this athletic measure being pretty absurd?

Luigi: My concern is that these qualities are not validated against any objective metric. If you can show me that captains of the lacrosse team are more successful in life afterwards, then that’s a great criterion and is an objective criterion. And, actually, Harvard has the data to do that. They follow all the alumni very rigorously, and so they can determine whether the kid that was admitted with a pretty low academic score but with a fantastic personality score because he was captain of a team, on average, does very well or very poorly both in class and in life.

I think all of this is feasible. Now, you have to be careful, because of course, discrimination is also present in life. So, maybe this undermines it. But at least it is a bit less arbitrary than to just decide that, oh, I love people who play polo and I think they are great. Not to mention, if you sail on a yacht or you go and explore exotic places on vacation.

Kate: All of this conversation loses a little bit of its meaning if we don’t exactly know what Harvard’s objective is as an institution. And that, to me, is a little bit unclear. What is Harvard trying to achieve as a university with its undergraduate class? What do you think that the answer to that is, Luigi?

Luigi: I agree. It’s a very difficult question. Let me say what they probably should achieve and probably is also what they try to achieve, but I think that they should provide the most valuable and formative education to their students. And I agree that being exposed to diversity is important. I disagree that diversity is mostly a racial issue. I know a lot of people with different racial backgrounds, but they all come from the same socioeconomic status and they’re pretty similar. On the other hand, when you leave the two coasts and Chicago and you go into the real America, you meet a lot of people who may all be white or mostly white, but they’re very, very different in terms of background, way of thinking, et cetera, than most of the people you will meet on campus. So, diversity is an important element of your experience, but diversity is multidimensional. It’s not just ethnic-based or racially based.

Kate: I think that part of the challenge with measuring outcomes is that Harvard is trying to maximize a pretty complicated objective function. I think it wants to generally maximize the power of its students. When I went to Harvard, the academic education for a huge segment of the student body was a relatively minor, insignificant part of your college experience. Really, the important parts of college were in the extracurriculars, in the bonds that you formed with other people, and in the sort of relationships that you would form to leverage later on in life so that you could all help each other. You could all lift each other up as Harvard men and women. And there were some particularly crass experiences that I went through where people would be trading favors about who’s going to get to be what in the cabinet, if somebody gets to be president, for example, based on how many shots you can take that evening or how grotesque you could be in some sort of public . . . I don’t know, I don’t want to expose too much, but people would just get really drunk and get really outlandish and then say, “OK, yeah, I’m going to make you secretary of the treasury because you did this ridiculous thing.”

And obviously, this isn’t true, but I think that there are elements of truth there to this idea that important bonds are formed when you’re in college and also in other professional schools, but particularly when you’re in college. And I think Harvard for the most part turned a blind eye to all of that. They wanted that stuff to go on. They want people to form these deep networks amongst powerful people so that they would then go on to leverage those connections, and then it would all sort of benefit Harvard through this complicated system that was not simply donations but a measure of the influence that the university has in the world.

Luigi: So, you’re saying that they maintain this subjective score on purpose to hide what their objective criteria are, because if they were to make it more objective, that would not sell well to the public at large.

Kate: Yes.

Luigi: That’s very clever. What you are saying explains certainly the enormous fraction of legacy students that they admit, and the fact that they have a special list of donors and friends that have a much higher probability of being admitted. So, I am reading from the report. While the average nonlegacy, nonathlete guy has a 6 percent probability of being admitted, if you are on the deans and directors list, you have a 42 percent chance of being admitted. If you are a child of a faculty member, you have a 46 percent chance of being admitted. If you are an athlete, you have an 86 percent chance of being admitted. And if you’re a legacy, a 33.6 percent chance of being admitted.

So, they do have disproportionate favored treatment for people that belong to, if you want, powerful groups. So, if you are the son of the king of Saudi Arabia, chances are that you’re going to be the next king. Now the interesting thing is, and Saudi is a good example, things do change over time. So, in 30 years, maybe there will not be a king in Saudi Arabia. And if you look back, it’s not clear that Harvard did the best on this dimension because they should have admitted many more women early on. They should have admitted many more African Americans early on and they didn’t.

Now, we are fortunately away from the time where there was explicit discrimination. There was a dean at Yale that had this rule of admission, say, never admit more than five Jews, take only two Italian Catholics, and take no blacks at all.

Kate: That’s very bad.

Luigi: That was the rule back in the ’30s. Hopefully we’ve improved on that. But all of this legacy is really favoring mostly white Americans. In fact, mostly Anglo-Saxon Americans. And so, if your goal is diversity, in the other groups, you should compensate on that dimension.

Kate: Yeah, I think that that’s a good point that you’re making. So, to the extent that this case continues to go to higher courts, I think that this is a point that Arcidiacono should probably raise, which is that those special kids account for such a big fraction of the admitted pool, but for the most part they are white. And so, if you want to make things evened out on an overall basis, then you should give the nonwhite groups a little bit of an advantage in the other applicant pool, the nonspecial pool, and it seems like this is somewhat true for black and Hispanic students, but it’s not true for Asians.

Now, I will say that Card says, he does this year-by-year analysis and he shows that in a couple of years, I think it was actually four years, there is a slight advantage to Asian students, and so there’s a little bit of this . . . I think that Card could say that there’s a tiny bit of evidence that there’s a favorability bias actually for Asian Americans, but I don’t think it’s statistically significant. Certainly not to the extent that that’s true for black and Hispanic students.

Luigi: Well, one point is pretty clear, if you’re white and nonlegacy, you’re screwed.

Kate: Sure. In really technical terms, you’re screwed.

Luigi: No, no, I think that is something that we should seriously consider as a negative effect, because the moment you start to have all these groups and you try to compensate on diversity, et cetera, you end up inevitably penalizing some other group. So, the Asian Americans are a recognized group and they complain, and now they’re probably going to get better treatment whether they win the case or not. The publicity is such that Harvard will be much more careful in turning down Asian Americans.

However, who is the likely loser of this? The rural white Americans that are not necessarily a separate category. They might invent one pretty soon in order to be protected, but it gets worse and worse and worse, because every group has an incentive to differentiate itself and finding some quota protection, and that becomes a constant battle in that dimension.

Kate: Yeah, for the most part, I agree. I actually think that it’s not the rural white kids, though, that would be the most disadvantaged from an application perspective. I think it would be regular, upper-middle-class white kids who grew up in cities who were not otherwise rich or legacies or whatever. They were just pretty smart and pretty well-rounded. I think that they would be the ones against whom there would be the most bias.

Luigi: The last question we need to discuss, and we touched upon it in, I think, our first or second podcast, is whether a private institution like Harvard should have the right to pick whatever policy they want, or do they have some responsibility to our society to be fair and unbiased?

Kate: I think that they have a responsibility toward society to be fair and unbiased. Now, everything, I think, goes back to the question of what sort of objective should the institution have. What sort of objective should society have? What sort of role should colleges play in society overall? Is it OK if, let’s say, a handful of elite Ivy League universities-plus are trying to maximize the objective of maintaining power? Whereas the rest of the schools are trying to just give students the best education possible? Is this the sort of system that we have today? And if so, is that sort of system ideal?

I think it makes sense from the school’s perspective to try and maximize something other than just pure education. Whether or not this is good for society overall, I think is less clear. My sense is that the answer definitely tends in the direction of no. Should Harvard and Yale and Stanford and Chicago, et cetera, should they be tasked not with providing the best educations, but with picking the winners of society, the people that are going to go on to be the titans of industry and president and Supreme Court justices and the most notable activists? I don’t think that that’s something that should fall in the hands of universities. I think it’s something that should be decided later on, once everyone has the same sort of access to education and once everyone’s been given the best education possible. 

So, having said all of that, I think that there should be some light stepping-in of the government or regulators and policymakers in offsetting the system a little bit. Now, what policies should be taken to offset the system? That’s a whole different question that I think is equally as complicated. What do you think, Luigi?

Luigi: Actually, I think it’s easy. My view is, to the extent these are private institutions that do not benefit at all from any tax subsidy, they can do whatever they want. The moment they are subsidized by the government, they need to provide a minimum level of fairness in admission. And because all major universities are hugely subsidized, because donations are subsidized, the return on endowment is not taxed and so on and so forth, because of these enormous tax subsidies that they receive, they need to choose. 

If I were king for a day, what I would say is, either you adopt an equal admission policy based on objective criteria that might include merit, might also include disadvantaged backgrounds, or you lose your tax subsidies. And if you do not have any tax-favored status, you can do whatever you want. 

So, Kate, do you think that this case is a capital-is or capitalisn’t?

Kate: I think it’s a pretty complex case. I think there are several different levels on which it should be analyzed. One is just looking at the regular applicant pool. Is there systematic discrimination against Asians? The Arcidiacono testimony would say yes. That focuses mainly on the characteristics on which Asians are better at, particularly academic achievement. And the Card testimony would say, no, there’s no systematic discrimination or bias against Asians, because Harvard considers all of these other criteria that are important for generating a holistic leader, and Asians are not as good on some of those other dimensions.

So, just based on that analysis, I think that this whole system is a capitalisn’t, because some of the other measures of whether you’re a well-rounded leader, such as your athletic ability and to some extent this personality measure, which is subjective and could potentially be biased, even though it’s hard to prove, I think that those elements shouldn’t be given as much weight, or, if anything, they should be wrapped up into the other measures like extracurriculars. So, I would tend to side with Arcidiacono on that one, even though I acknowledge the point that there’s more to a person than just GPA and grades.

Luigi: So, Kate, is Harvard discriminating against Asian Americans?

Kate: I think we still don’t know.

Luigi: We do know its revealed preferences. You did not put your ethnicity down when you applied, and I think because the probability of being discriminated against was large enough.

Kate: Yeah. Is that going to hold up in court as evidence that Harvard discriminates? I don’t think so.

Luigi: No, but it proves that at least you think so.

Kate: That’s fair. I think that Card is right in some of his points about missing data from the Arcidiacono model. I would like to see the Arcidiacono model, but I would also like to see different measures of or different weights on these other criteria. 

Now, there’s the larger point that we talked about, which is, what about the other, over a third of admitted students who come from rich, privileged legacy families? Should they be given the advantages that they’re given, and should they make up as large of an accepted pool as they do? I think the answer is definitely not.

Luigi: There’s at least a one-third chance if you’re coming out of Harvard that you didn’t belong there. So, the quality of a Harvard student is not as high, sorry, Kate, as people think.

Kate: Hey, I’m not going to argue with that.

Luigi: If you come from a state school, a very competitive state school, and you did very well there, chances are, you are a much better candidate than if you came from Harvard and you don’t deserve it. I think that the great merit of the American litigation system is to bring up these topics and disclose the data and expose the problem. We who are in the profession had a sense of how widespread this legacy business was, but now we have the data, and it’s massive, and I think that something should be done. If rich people want to preserve their status, they have the right to do so, but not with the subsidy of the federal government.