Legacy of privilege: David Deming and Raj Chetty on how elite college admissions policies affect who gains power and prestige

 

Preroll: (Ralph Ranalli): PolicyCast explores evidence-based policy solutions to the big problems we’re facing in our society and our world. This podcast is a production of the Kennedy School of Government at Harvard University. One programming note: Today’s episode concerns recent research on legacy admissions at elite colleges and universities, including Harvard. Harvard University has varied policies on admissions among its 13 schools. For example, Harvard College considers legacy status in admissions; the Kennedy School and other Harvard graduate schools do not. 

Intro: (Raj Chetty): And what we found is that it matters greatly in a very particular way. It matters for your chances of reaching society's leadership positions. So being the CEO of a big business, being a leading politician, being on a career trajectory where your decisions might shape many other people's lives. And so the fact that we have many more kids from very high-income families, rather than middle class families, at Ivy League institutions means that society's leaders are less diverse than they otherwise could be. And that I think has implications for everyone in society in terms of the types of laws that are created, the types of products that companies are developing and so on. 

Intro (David Deming): So why does it matter? Well, the reason it matters is because even though they're less than 1% of undergraduates, they're one in eight Fortune 500 CEOs. They're a quarter of all current US senators. They're 71% of Supreme Court justices, since Thurgood Marshall attended an Ivy Plus college as an undergraduate. They're more than a quarter of MacArthur Genius grant winners. So even though there's not very many students in these schools, they highly, highly, highly disproportionately occupy high earning and high-status positions in society. And since we think those positions matter, there's a lot of ink spilled about them, we think therefore it's important to study these schools. 

Intro (Ralph Ranalli): Welcome to the Harvard Kennedy School PolicyCast. I’m Ralph Ranalli. Students spend years preparing to face judgment by colleges and universities as a worthy potential applicant. They strive for report cards filled with A’s in advanced placement courses. They volunteer for service projects and participate in extracurricular activities. They cram furiously high-stakes standardized tests. They do all that only to find a big question many top colleges have is: “Who’s your daddy? And who's your mother? Did they go to school here?” Legacy admissions, the practice of giving preference to children of alumni, particularly at elite schools, were thrust into the spotlight this summer when the U.S. Supreme Court effectively ended affirmative action in college admissions. Thankfully, at the same time, Harvard professors David Deming and Raj Chetty were there with some important answers—having just wrapped up a 6-year study of the impact of legacy admissions at so-called “Ivy-plus” schools. Using data from more than 400 colleges and universities, the two economists found that legacy and other elite school admissions practices significantly favor students from wealthy families and serve a gate-keeping function to positions of power and prestige in society. Professor Deming is the Academic Dean of the Kennedy School and principal investigator at the CLIMB Initiative, which seeks to study and improve the role of higher education in social mobility. Professor Chetty is the director of Opportunity Insights, which uses big data to understand how children from disadvantaged backgrounds can have a better chance of succeeding. They’re here with me today to talk about their findings and why they matter. 

Ralph Ranalli: Raj, David, welcome to PolicyCast. 

Raj Chetty: Thanks, Ralph. Great to be here. 

David Deming: It's great to be here, Ralph. Thanks for having us. 

Ralph Ranalli: So before we talk about the actual research, I wanted to go back a little bit and talk and have you walk us through the process of gathering all the data. Because I hear it wasn't a simple matter of all this data was just sitting there. It was actually quite a five or six year process of having to go out and get it and to put it into a shape that was actually usable. Can you talk a little bit about the process of getting the data that you needed for this study? 

David Deming: Sure. Yeah, Ralph. Well, it was quite a process as you said. We started this project in 2017, so six years ago, and we started sort of one school or one university system at a time asking people if they were interested in participating. Of course, as you might imagine, a lot of people said no, some people said yes. And then having worked with some early partners like the University of Texas, Austin, and the University of California system and other schools like that, once we got a few people saying yes and showed that we were serious, we had a conference called the Climb Conference also in 2017, which led to a surge of interest and some partners. And then it was kind of like a snowball rolling downhill. I mean, again, we didn't get universal participation, but we spent a lot of time shaking hands, knocking on doors, whatever metaphors you want to use. And then eventually we ended up recruiting, I think, more than 400 colleges and universities encompassing about three and a half million students per year. It's about 15% of the undergraduate population in the US. So we ended up with quite a few partners and we're really grateful to them. Without them, this research wouldn't be possible. 

Raj Chetty: And just to add a bit more color on how we got into this, a lot of this data collection started after an earlier paper we had released together with our collaborator, John Friedman at Brown University, on just measuring rates of economic mobility across colleges. And that was not using internal college admissions data, it was using data from federal tax records in the Department of Education to measure the parental income distributions at every college and how kids did after college. And so we were able to put those statistics out publicly and lots of people were able to see how their college stacked up relative to other colleges. And that led to a lot of interest in understanding how we can create more mobility at our colleges. And so that naturally led to springboard to have the types of conversations that Dave described where we were able to say, "If you really want to answer that question, we need more data," which is what researchers usually say in order to go deeper. 

Ralph Ranalli: Great. So how did you two personally come to study inequality and the relationship of education to inequality, which is central now to both your work? It was interesting because I read in your bios that you were both born right around 1980, which is the jumping-off point when economists say our current period of period of widening economic inequality began. That basically means neither of you has ever lived in an America that wasn't in the process of getting more unequal. Can you talk a little bit about your individual paths to studying this area? 

David Deming: Sure. So in my case, Ralph, I was born in 1979, so you're off by one year, but that's okay, I'll take it. I’ll take every year I can get at this point. I grew up in Nashville, Tennessee. My father was a Methodist minister, and my mom was an editor at a religious book publishing company. And so I grew up in the church and kind of with a strong, I guess moral background, spending a lot of time thinking about trying to help people, trying to make things right. But the budding social scientist in me was always wondering, well, how do we know if the things we're doing are really helping people? And that seems like a pretty important question to answer. If you're going to invest your time and your energy and your resources into social programs, other policies, et cetera, you want to know if they're really working. And so for me, it was always kind of the marrying of the sense that I wanted to make the world a better place with a kind of empirical social scientist’s nose for the truth that really led me to this place.  

And for as far as education in particular, I mean, that's something that was really a great lever for mobility for me. Harvard was not something that was ever on my radar screen when I was growing up. But having done well in school, I went to Ohio State, the Ohio State University, on a full scholarship and academic scholarship, and that really opened a lot of doors for me, and I'm really grateful for that opportunity. And I just want to see more middle-class kids get that opportunity as well. And so that's really what's motivated me to do the work that I'm doing. 

Ralph Ranalli: Raj, you were born in India and I know you have a great story about your parents and access to education. Can you talk a bit about that? 

Raj Chetty: That's right, Ralph. So I was also born in 1979, a very different place from Dave, in New Delhi, India. And for me, I think part of the interest in inequality comes from my own experience growing up till I was age nine in Delhi. And then seeing the tremendous contrast with the United States. And of course one starts to wonder why opportunities seems so different for kids in America relative to India and at a broad level that got me interested in studying these kinds of questions.  

But as you noted, it also goes for me back a generation. So I come from a family of academics. I'm the last person in my family to publish a paper after my sisters and my parents. And my parents got those opportunities. They actually grew up in low-income villages in South India where it was typical at the time to only choose one child to get a higher education because families didn't have the money to educate all of their kids. And it so happened my mom was the person to get that higher education in her family—particularly unusual because women were not getting educated at that point as much. And then my dad happened to be the one in his family, and they ended up coming to the United States to University of Wisconsin, another great public institution like Ohio State University. And that really paved the path for them, my dad getting his PhD, my mom, her medical training. And so I also saw the power of education to really shape things generationally. So I've been interested in intergenerational mobility from many different angles for some time, thinking about the context of neighborhoods and childhood environment and various different programs. But education is clearly a key lever to increase mobility and opportunity. 

Ralph Ranalli: So turning to the actual research, you found that children from families in the top 1% of the income distribution are more than twice as likely to attend a so-called Ivy Plus college. And you define Ivy Plus as the eight Ivy League schools—Harvard, Yale, Princeton, Brown, Penn, Columbia, Cornell, and Dartmouth—plus Stanford, MIT, Duke, and the University of Chicago. Students from wealthy families are more than twice as likely to attend those schools as those from middle class families with comparable SAT, ACT scores. I guess the first question is: Why does that matter? 

Raj Chetty: So the fact is, as you stated Ralph, is exactly right that kids with comparable academic credentials from high income families are far more likely to attend Ivy Plus colleges than kids from the middle class. And so why does this matter? Well, we might care about diversity on campus just in and of itself, but here in this paper, we actually take an outcome-based perspective and ask the question: Does it matter if you attend an Ivy Plus college or not?  

And what we found is that it matters greatly in a very particular way. It matters for your chances of reaching society's leadership positions. So being the CEO of a big business, being a leading politician, being on a career trajectory where your decisions might shape many other people's lives. And so the fact that we have many more kids from very high-income families rather than middle class families, and Ivy League institutions means that society's leaders are less diverse than they otherwise could be. And that, I think, has implications for everyone in society in terms of the types of laws that are created, the types of products that companies are developing and so on. 

David Deming: And just to put some numbers to it, Ralph, you might say, well, it turns out the Ivy Plus colleges, the 12 you described, enroll eight tenths of a percent of all undergraduates in the US. So you might say, well, that's not very many people. Even if we took all the students out of those schools and filled them with low- and middle-income families, it still wouldn't make a dent in overall inequality. So why does it matter? Well, the reason it matters is because even though they're less than 1% of undergraduates, they're one in eight Fortune 500 CEOs. They're a quarter of all current US senators. They're 71% of Supreme Court justices, since Thurgood Marshall attended an Ivy-plus college as an undergraduate. They're more than a quarter of MacArthur Genius grant winners.  

So even though there's not very many students in these schools, they highly, highly, highly disproportionately occupy high earning and high-status positions in society. And since we think those positions matter, there's a lot of ink spilled about them, we think therefore it's important to study these schools. 

Ralph Ranalli: I think one of the reasons that your study got so much attention—not that it wouldn't have gotten a lot of attention just on its own for its important conclusions—is because the Supreme Court ruling in June that essentially ended race-conscious admissions policies. What’s the relationship between this study and the affirmative action ruling? Are your conclusions going to be more important in the absence of affirmative action? 

Raj Chetty: So, colleges are clearly in a position where they have to revisit admissions, right? Because of the Supreme Court decision, our paper is not fundamentally focused on racial diversity, it's focused on socioeconomic diversity, kids from high- versus low-income families getting in at different rates. 

Ralph Ranalli: You made an affirmative decision not to include race in this study. Can you talk about why? 

Raj Chetty: That's right. That's partly because prior studies, including those that were cited in that Supreme Court decision had extensively examined differences by race and rates of admission. What was new and what hadn't been done is examining differences in rates of admission and subsequent outcomes by parental income, the socioeconomic dimension, those two things are obviously related. So Black people, for example, tend to have lower incomes in the United States at the moment. So there's a correlation between the two, but they're not the same. And so what we do is focus on the socioeconomic dimension in particular, and we show that there's a lot of progress to be made on the socioeconomic dimension, which could also have implications for racial diversity, but in our view is important in its own right. 

David Deming: Yeah, I think we should be able to have these two conversations at the same time, Ralph, racial diversity is an important value of American colleges and rightly so. And we think that socioeconomic diversity should be too, if we have schools that are racially diverse but have only high-income families, they're not diverse in some larger sense. And so even though these two things don't operate in the same way as Raj said, that race and income are correlated but imperfectly, it's another dimension on which we ought to think about diversity. 

Ralph Ranalli: So, you said in the study that high income admissions' advantage at private colleges is driven by three factors. The first is admissions or preference for children of alumni, the second is the weight placed on non-academic credentials, which tend to be stronger for students from wealthier families. And the third is recruitment of athletes, particularly in a group of niche sports that open access to elite institutions but aren't the most popular sports out there in terms of what you see on ESPN. But you also said that none of those factors are necessarily predictors of later success. So why do colleges take those factors into account? 

David Deming: Well, I think that colleges are trying to do a pretty complicated process balancing on multiple dimensions. And again, it's not that, for example, legacy students do worse, they don't do better either. And so given that institutions have values other than just maximizing the future outcomes of the class, I think they have different reasons for doing those things. Our purpose is not so much to say they should be doing one thing or another, but to point out the consequences of the choices that they make. And again, there's nothing wrong with placing a value per se, in placing a value on legacy applicants. If the legacy applicants who were admitted to Harvard had the same income distribution as everyone else, I don't actually think people would be that upset about it. It's the relationship between being legacy and being very high income that makes people feel as though the game is rigged. 

Similarly with athletes, it's not like people who have a lot of money are just better athletes naturally. It's that people understand how this game is played. And so not everyone can from an early age, from an early age of their kids, invest in private coaches, travel teams, finding sports that are hard to play with specialized equipment and so on. And so it's really reflective of the high stakes of college admission and the fact that if you're someone with a very high income and a high disposable income, there's almost nothing you'd rather spend it on than your child's future. And so this kind of really competitive process leads us to this point. Again, not because there's anything wrong with caring about athletics or caring about legacies or caring about, as David Leonhardt calls it, private school polish. It's that these are all things that people, these are cracks in the system that people with means can manipulate if they choose. And many of them do choose to do so understandably because they're trying to secure a future for their child. And so I think our study points that out, points out the ways in which some of the things that show up on applications as signifying merit may actually signify privilege. 

Raj Chetty: I agree with everything Dave said, and I might just add that if you put yourself in the shoes of an admissions officer, and we talked to many admissions deans and other folks involved in this process along the way, they face a very complicated problem, as Dave said, right? I mean, you get tens of thousands of applications from many highly qualified applicants, and you've got the very difficult task of trying to choose between them. And so I think all of us, if we were in their shoes, would naturally look to things like the student has a really polished essay or a very strong teacher recommendation letter, has done amazing things outside the classroom to really distinguish themselves. Turns out those things correlate also, maybe not surprisingly, exposed with coming from a high-income family, going to a private school that can give you those kinds of resources, give you a teacher who can really write that very personalized recommendation letter. 

And as a result, that signal gets diluted in terms of really picking up that sort of privilege, which is not necessarily predictive of later outcomes. In fact, the child who has gotten to kind of a similar place, but without all that support, may be equally meritorious and may have even better chances of succeeding in the long run in some sense. And so I think that's why that instinct of putting weight on these factors, which is totally understandable, it's actually useful to see these facts in the data that it's not that predictive of success and we're hoping that will reshape what colleges do going forward. 

Ralph Ranalli: Yes. You found that the average income difference is actually quite small compared to the difference between achieving that those very high positions of prestige and status and at a high-level firm or a position of power. How much of an advantage does being wealthy and being a legacy give you in quantitative terms over say students who are just rich or students who are just legacy and maybe middle class and students who are non legacy? How does that break down? 

David Deming: So it depends on, there's a bunch of different comparisons one can make, let me make one. So legacy applicants after controlling for other things that predict admission, things like academic ratings, which encompass SAT scores and GPA and a bunch of things, legacy applicants, all else being equal, are about four times more likely to be admitted with among those who have the same credentials as non-legacy applicants. That legacy boost is actually slightly larger for high income legacies. So that's a pretty large advantage. Now, four times more likely might be five percentage points versus 20 percentage points. So it's not like everybody who's admitted is a legacy. A lot of legacy applicants don't get admitted and a lot of people who aren't legacies do get admitted. So there's a lot of slippage in there, but nonetheless, it's a big boost to be a legacy applicant, all else being equal. 

Similarly, these non-academic gradings tend to make not quite that large of an impact, but it's a more common phenomenon. So it adds up to being similar in magnitude as a contribution to the total increase in the top 1% that we see in these classes. I don't know. Raj, do you want to add anything to that? 

Raj Chetty: Yeah, I mean I think that four times statistic that Dave just gave gives you a sense of the magnitude of these differences. If I were to tell you here's something that you can do that's going to increase your odds of getting into Harvard by factor four, you'd probably be pretty excited about trying to figure out how to do that. So this is an important advantage. 

Now, one thing I would add, which I found particularly striking in this analysis is, you might wonder, well, those legacy applicants, their parents went to a great college, they must be highly qualified. They've been kind of well-trained since childhood, maybe it makes sense that they get in at higher rates. The power of having the data that we have from multiple colleges is we can do sharp tests of whether that kind of logic really makes sense. We can ask if you are, say a legacy applicant, whose parents went to Stanford, are you more likely to get into Princeton? If you were truly more qualified, you might think that other colleges would value those credentials as well. And it turns out the answer is basically no. That fourfold admissions advantage that you have only applies at the college your parents went to and is absolutely irrelevant at other peer institutions. And so that kind of evidence really strongly suggests that this is an advantage that's really about your parents going to that particular college and I think it should make us question, do we want to keep this type of disparity going forward? 

Ralph Ranalli: The second half of your paper focuses on waitlisted students. What were the special characteristics of those waitlisted students that made them particularly interesting and important to study? 

David Deming: Yeah, so I think the reason why we focused on waitlisted students, Ralph, is actually because we think that within the wait list, the students who are admitted and the students who are not admitted are pretty similar. And we do a lot of work in the paper to try to convince the reader that's true. And I think it is true that when you look at academic credentials and other non-academic credentials, almost anything you can look at, it seems like the wait list admits and the wait list rejects are pretty similar, but also quite different from the students who were just straight up admitted and the students who were straight up rejected. And so we do a lot of work to try to convince you in the paper that you should think about the people who just barely get in off the wait list as being kind of similar from the ones who don't, except that they fill some idiosyncratic need in terms of class composition.  

So one example might be if it turns out that you are picking amongst students at the very end and you're trying to fill a class in one school, you might need someone to play the oboe because the oboe player just graduated. So you have one applicant who plays the oboe, another one who plays the saxophone, but the saxophone, that spot is already filled in the orchestra and so you admit the oboe player and not the saxophone player. Now the key implication of that is that it shouldn't predict admission at another school. If the reason you got into Harvard is because you play the oboe rather than the saxophone, it shouldn't matter that you play the oboe whether you get into Yale or Princeton for example, because they may need a saxophone player. And so that's a key test that we perform in the paper. We ask whether being admitted off the wait list to one school predicts being admitted to other schools, and we find that it doesn't. 

And so we argue based on that—and based on a bunch of other things that you can read in the paper if you care to—that these two groups, these wait list admits and these wait list rejects are pretty good comparisons for each other. And then we use that to try to draw out the implications of barely being admitted versus barely being rejected from an Ivy Plus school. 

Raj Chetty: So basically what Dave is saying is we're looking for experiments here, but ideally what we'd like to do is flip a coin and let some kids into an Ivy League college and others not, and compare what their outcomes are in order to figure out the causal effects of attending these places. That's of course very hard to do in this context and in the social sciences more generally. And so the kinds of methods Dave just described are ways to try to approximate that sort of experiment with the data we have. And based on that kind of analysis, we find that if you sort of got lucky and you played the right musical instrument and got in off the wait list, your outcomes look very different. In particular, your chances of reaching the top 1% of the income distribution 10 years after college, your chances of attending a top graduate school working at a very prestigious firm, these odds go up dramatically relative to if you got unlucky and didn't end up getting it. 

Ralph Ranalli: I'm really interested in the feedback or pushback you've gotten. You mentioned the initial pushback that when you were going around trying to get this—when people said no, that we're not going to cooperate with you, what did they say? And was it more unspoken or did they actually come out and say why they didn't want to participate? 

David Deming: Yeah, so the first thing you say, Ralph, is that I think actually most of the folks that we talked to really did want to participate. Sometimes there were institutional constraints that prevented them from doing it, and sometimes there were data constraints. They didn't have the data we needed, or it wasn't recorded anymore. Nobody ever says, no, we don't like what you're doing. They just don't say yes, and you kind of give up. That's the way it works. And I think the reality is these colleges are trying to do lots of things and they're very busy and they don't have any particular incentive to participate. And so often it just relies on some goodwill and us convincing them that we're going to be a trusted partner, which I think we have been. So I don't really attribute anything negative to the people who don't want to participate. It just turns out to be more trouble than it's worth in some cases.  

And we were fortunate that we had enough people who really do care about these issues. I think a lot of people who work in college admissions offices or people who are directors of research or even college presidents actually are in this business because they want to make a difference and they want to participate in a study like this. And we've had a lot of partners who've been pretty enthusiastic about the work we've done. So I wouldn't say there's been a lot of pushback from our... I don't think there's been any pushback from our partners. There's been some pushback on some of the implications of the work in terms of solutions, and I think that reflects the fact this is a genuinely difficult problem. We could talk more about those in particular, but I actually think it's been fairly well received. 

Raj Chetty: One of the things we're able to do for our partners is provide detailed information on their own colleges' admission situations. What is it at your college? Which kinds of students are you missing? Where might you be able to look for more students if you want to increase diversity on a certain dimension? We've learned that while there are broad patterns that we've been discussing, those patterns differ across different colleges. To give you one example, we see in the federal data that this pattern of higher admissions rates for kids from high income families is true at most Ivy League colleges and Ivy Plus colleges, but it's not true, it turns out, at MIT, which tends to have a much flatter distribution of admissions across the income distribution. So when we're able to work with college's internal data directly, we're able to give much sharper advice in some sense on how to improve things going forward. And so I think that leads to actually a good partnership and is a good model going forward. 

Ralph Ranalli: Did that help you? That you were giving them information that they didn't have and that was useful to them? 

David Deming: I think so. I mean, I think that, as I said before, people don't... And this is true for any data partnership you engage in as a researcher, people don't often have much direct... They're not necessarily benefiting from it. And so our being able to offer something back to them I think makes a big difference not only for their own purposes, but also is a good faith gesture that we truly are partners, not just people who are going to take the data and disappear and that's not the spirit. We really try to be partners to our institutions and value their contributions. So I think that did make a big difference and we intend to keep doing that as we move forward with the work. 

Ralph Ranalli: I think people tend to have generally sort of a very jaded view of college admissions offices and the people who work there. Was it eye opening having a lot of contact with and working with those offices and a lot of those folks? 

Raj Chetty: Yeah, I mean I would say it was eye opening, but I certainly don't share the view after interacting with a lot of these folks that there's any negative intent or that they don't share a lot of the values that many of us would have. I think people are genuinely trying to do the best job they can in a complicated environment facing lots of constraints, and I think there's a lack of information. You have an instinct that this seems like a pretty good predictor of the student doing well. The student would be a good match at this college. But honestly, we just have not had the data to see systematically how will kids, say from middle class families, do if we admit them at higher rates at Harvard? Maybe they will do very well as we're finding in this study, but it's equally plausible to have thought, ex abundante, that they've suffered from various disparities from birth until the point they're applying to college, going to different schools and so on. Maybe some of those students are actually going to struggle if we put them in an environment like this. And so that's why I think it's so important to actually look at these data and we've found that people are interested in listening and possibly making changes in light of that. 

David Deming: Yeah, absolutely. And just on the information gap, it's important to understand that in some cases we have more information than the college admissions offices do. So they know whether students qualify for financial aid, if they apply for it, but beyond that, they don't actually have direct income data. So they may have a general sense from where the student grew up or what kind of school they went to, but they really can't tell the difference between somebody who comes from a family that earns $300,000 a year or $3 million a year or $30 million a year. They don't actually know, even if they could guess. And so I do think that we're providing some genuinely new information to schools, and I do think that they will be more conscious of income going forward. 

Ralph Ranalli: So some schools had dropped legacy even before your study came out, some schools have dropped it after your study came out. Johns Hopkins, Carnegie Mellon, Amherst College, Pomona College, I mean those may not be Ivy-plus schools, but maybe Ivy-plus-plus? Those are good schools. Have you received any feedback, any information anecdotal or otherwise sort of in the before and after sense of reporting back from colleges that have dropped legacy admissions and how it's affected them? 

Raj Chetty: So the first thing to say here is a number of the colleges you pointed out don't formally put them in the Ivy Plus group, but the patterns we document in this paper apply more broadly to elite private colleges. Public colleges look very different to be clear, places like Berkeley and University of Michigan, but private colleges even outside the Ivy Plus exhibit very similar patterns. And so it's very interesting to think about the places that have ended legacy admissions or made other changes in their admissions practices to see what happens over time. 

We have not done that type of analysis yet, partly because a lot of these changes have been extremely recent, but I think that's exactly how we can learn going forward. Do things in some sense really fall apart in some way when we end legacy admissions? One thing people often understandably worry about is what are the implications for fundraising? Our sense, just at least anecdotally from looking at some of the cases you just described, is those institutions are not struggling to raise funds after they've made those decisions. Now does that apply universally or not? Is that really a systematic pattern? I think more work needs to be done to figure that out. 

David Deming: Let me say one other thing, Ralph, which is that, while I certainly wouldn't defend the practice of legacy admissions. And I do think, I mean it's clear from our study that it contributes to disproportionate representation among the rich. If all you did was end legacy admissions, but you changed nothing else about the admissions process, it would have at best a very modest effect on income diversity at US colleges. And that's because most of the high-income students at colleges are not legacies and not all legacies are high income. And so there's a lot of slippage in that measure. 

So, I do worry a little bit that colleges will end legacy missions but not really systematically change the affirmative action for the rich essentially that we see throughout the admissions process. So I would actually prefer to see a more focused effort on making colleges more income diverse along whatever dimensions each school feels works for them. There's a bunch of different things they could do. The three factors we mentioned, there are others for some schools that may make more sense than others. Again, I don't think that people would really be that bothered by legacy preferences if it didn't result in the income disparities that we see. So again, I am not defending the practice, but I don't think it's a panacea for the larger patterns that we document in our study. 

Raj Chetty: Just to underscore the point Dave just made about legacy, I find there's a lot of focus on legacy at the moment, maybe rightly so, but it's critical to understand that that's only a piece of the puzzle and not just in an accounting sense, in the sense that it explains a small part of the high-income advantage, but also that people are likely to respond in terms of their choices when they apply to college if you don't do anything else. So to give you an example, suppose you eliminate legacy admissions, so you no longer have that sort of leg up in the admissions process, as an affluent parent and as a child applying to college, you now have even more incentive to go do the extra extracurriculars and start a nonprofit and invest in embellishing your application because that other pathway now becomes less likely. So unless we address the system, the incentives it creates more broadly, I don't think we're actually going to address the issues of inequality. 

Ralph Ranalli: Yeah, that was actually my next question was in terms of helping to create a society where we are making progress on the problem of inequality and we're trying to get more diversity in those upper echelons of decision making and income earning and controlling a lot of things that affect us in society—what characteristics does a more optimal admissions process have? 

David Deming: Yeah, so I'll give you one answer, Ralph, and I think basically sunlight is the best disinfectant. I think that if we had a system of reporting the income distribution of the admitted classes of US colleges and holding them publicly accountable for it in the same way that we do for race and gender and other demographic characteristics, I think that would make a big difference because all of these schools care immensely about their reputation. In some sense, the reason Harvard is Harvard is because everyone wants to go there. And so things that affect the reputations of schools, if they're seen as places where they practice affirmative action for the rich rather than places that reward merit, that's bad for their long-term reputation.  

And so I do think if we had that system, if colleges reported those statistics, I think you would see them start to think about balance on income the way they think about it on race and gender and other characteristics. So that would be one, I'm not pretending that would solve the problem by itself, but I actually think it would make a huge difference and I would strongly advocate for finding some way to report those data at the college level by year and we could actually help with that. 

Raj Chetty: Yeah, I think that sort of transparency is incredibly valuable and there are steps we can take basically building on the architecture we've developed to do this study, to put out those kinds of data publicly, and that's something we and others hopefully will be able to work toward.  

What kinds of steps might colleges then take concretely when faced with that sort of public pressure? I think it's important to look at the data and see what are the types of factors that are predictive of student success that are less manipulable. So to give you an example, we find in this study that contrary to maybe popular perception at the moment, SAT scores, standardized test scores like on the SAT and ACT are highly predictive of students' outcomes in college and beyond. And so they're actually, in some sense, a better measure if you had to pick between SAT or ACT scores and various non-academic credentials or even high school grades. Those things are much more manipulable and much more correlated with income and are also poorer signals of later success. And so I think starting to understand, I'm not saying we should just make admissions decisions on the basis of SAT and ACT, but starting to understand what are the factors that have those kinds of features and putting more weight on those. That's the kind of approach I think that can create more equity going forward. 

David Deming: Raj raises a really important point, but I think is a bit of a subtle one, so I just want to dig into it a little bit. So it could very well be, and in fact there's some evidence that this is true, that the SAT is biased in favor of higher income applicants in the sense that the distribution of high scoring students is strongly high income. However, colleges know this and colleges admissions offices know that a 1450 on the SAT doesn't mean the same thing when it comes from a high income applicant as it does from a lower or middle income applicant. And so they have a way of adjusting for that. And we can see that in the data that they do actually put a thumb on the scale for high-scoring, low-income students. 

And I think they're much more able to do that both because it's a quantitative metric that's on a common scale across all applicants, and because they have more experience with it, they're much more able to do that for the SAT than they are for guidance council recommendations or sports or some of these other factors. And so it's not that the SAT is some perfect measure of merit, it's that it's better than the alternatives. And so the concern with going test optional across the country is that if you eliminate SAT as a way of judging applicant's ability to achieve academically, what you're left with is all the things that money can buy, which is perfect guidance council recommendations and polished applicant profiles and all these other things. And so again, everything has bias in it, everything has this competitive nature to it, but academics are something that's a bit harder to fake and something that admissions offices have more experience with accounting for. And so that's the sense in which it's better. Not that it's perfect, it's better than some of the alternatives. 

Ralph Ranalli: Plus all the college essays pretty much from now on are going to be written by ChatGPT anyway. 

David Deming: That too. Anyway, that too. Of course it probably could take a test too. So let's not, let's not get too excited. 

Raj Chetty: ChatGPT may be more equitable, I guess, across the income distribution than hiring someone. 

Ralph Ranalli: So, this is the part of our podcast where we put the policy in PolicyCast, so I'm going to ask you both for a couple of concrete policy recommendations to improve this area of our education system that our PolicyCast listeners might want to support. David, do you want to go first? 

David Deming: Sure. I'm going to repeat and expound on something I said earlier because I think it's so important, Ralph. I think that the single biggest thing that could come out of our study is a system of collecting and measuring statistics of income, the distribution of income of the admitted classes of colleges across the country. So how many of the students who attend a school are from the top 1%? How many are from the middle class, how many are low income? We do that now with things like financial aid. So how many people are Pell eligible? But there's no reporting for income levels above low income students.  

And I think that could change. It could change in a couple of ways. One is our team could build a consortium of colleges who agree to report those statistics yearly and we could help facilitate those computations. And then eventually, I think there could be some active legislation or some other policy that enshrines this nationwide. It is definitely possible. It requires some cooperation across federal agencies, but I think that one thing would make a bigger difference than any one lever I can think of to pull. So again, I would strongly support the collection and reporting of standardized data on the income distribution of accepted classes across colleges nationwide. 

Raj Chetty: What I would say is in the current context, we mentioned the Supreme Court decision earlier where colleges are trying to think about how to increase diversity on the racial dimension and on various other dimensions at the same time. One theme I often hear coming out in the public discourse is maybe we should put a thumb on the scale for kids from lower-income families, socioeconomic disadvantage as a way to kind of substitute for the loss of racial diversity. The lesson I feel we've learned from this study is before we think about putting a thumb on the scale for anyone, we should just simply think about how to take the thumb off the scale that we currently have on for kids from high income families, right? The affirmative action for the rich essentially that we currently have in place.  

And so how might we do that? I think it's about revisiting our admissions practices in the ways that we've discussed, trying to think carefully about what is leading to this thumb on the scale for kids from high income families? That's going to, in and of itself create more equity by class, also to some extent by race and ethnicity. And I think that will make a big difference in terms of diversity at elite colleges. 

I also want to add one other point. Our conversation here in this particular study has focused specifically on a very small set of elite private colleges, which we think play an important role in educating society's leaders. But in the grand scheme of things, they educate a very small fraction of students. As Dave mentioned earlier, less than 1% of college students go to these 12 colleges that we focused on here. And so at a bigger picture level, I think in order to increase the contribution of our higher education system to economic mobility more broadly, we need to think about how to improve access and outcomes at a much larger set of institutions, community colleges, public institutions that educate much larger groups of students.  

There's going to be a different set of solutions required there, which might involve state funding, might involve improving retention rates and outcomes at those places. Actually, the key focus of some of our ongoing research, but I want to make sure we don't lose sight of that important goal while we focus on improving access to elite colleges. 

David Deming: That's the next project. Stay tuned. Hopefully it won't take six more years. 

Ralph Ranalli: You heard it on PolicyCast first, folks. Well, thank you both very much for being here. This has been a really interesting and enjoyable conversation and I appreciate it. 

David Deming: Thanks for having us, Ralph. It was a pleasure. 

Raj Chetty: Thanks so much, Ralph. 

Outro: Thanks for listening. Please subscribe to PolicyCast on your favorite podcasting app so you don’t miss any our great upcoming episodes. If you have a comment or a suggestion for the team here at PolicyCast, please drop us an email at PolicyCast@hks.harvard.edu—we’d love to hear from you. And until next time, remember to speak bravely, and listen generously.