The Strength of Humility

A conversation with Rohini Somanathan

Earlier this year, I sat down for a long, wide-ranging conversation with my friend and former student, the economist Rohini Somanathan for the Stanford’s Center for Advanced Study in the Behavioral Sciences’ Human Centered podcast. You can listen to the entire thing here, but I also wanted to provide a partial transcript below (it picks up at 34:08 in the conversation).

When Rohini asks me how economists can address the political polarization in our society by trying to ameliorate the insularity that can afflict academic economists, I have a simple response: Read widely and deeply. Economists should resist the temptation to bury their heads in statistics and mathematics. We’re social scientists. We ought to concern ourselves not merely with clever models and technical innovation but also, and more importantly, with the broader social landscape those models and techniques seek to represent. That is, with actual people living actual lives and interacting with each other in the real world, with the complex histories that brought about that real world as it presently exists, and with the deep, often counterintuitive psychological structures that motivate people to act in ways that figures and graphs cannot satisfactorily capture. Thinkers in every field have been wrestling with these issues for millennia—we economists should have the humility and the wisdom to listen to what they have to say.

That advice will not solve all of our problems, either as a discipline or a society. But taking it would be a good start for economists, and for everyone else. Having principles of our own should not preclude us from thinking seriously about new ideas, wherever and whenever they originate. In the excerpt below, Rohini’s instigation—and that of CASBS Communications Director Mike Gaetani—prods me to think about the development of my own ideas and thinking. As you may note, they often emerge from places of uncertainty rather than any hardened position. How might we best describe racial inequality? How might we best educate the young people of our nation? What is the proper remit of economics? No one person can answer these questions alone, and whatever answers we find will be, in all likelihood, provisional and subject to change. But then so are we, as humans, provisional and subject to change. We ought to regard that as a strength, not a weakness.

---

This post is free and available to the public. To receive early access to TGS episodes, an ad-free podcast feed, Q&As, and other exclusive content and benefits, click below.

---

ROHINI SOMANATHAN: I want to talk a little bit about your thinking on race over the years. And I was particularly struck by this evolution, when you think about the big papers. So, if you think about your affirmative action paper with Steve Coate in the 90s, and then culminating in the Du Bois Lectures and The Anatomy of Racial Inequality, which is such a powerful monograph.

And what I'm struck by is really a humility, Glenn. I know that you're not thought of as a humble person. This is part of why it's so fun to talk to you, because I think people just see you in so many different ways. I've always felt that I can just say to you whatever I think. I can disagree with you. I find you listening to me. And yet, that doesn't always come across. It brings me to this broader question of economists within the discipline. We're often thought of as arrogant, of taking big ideas in sociology, psychology, and dressing them up with with math, and then presenting them to the world, and then getting accolades.

And so, I was struck, in [The Anatomy of Racial Inequality], by how you start by talking about statistical discrimination and taxi drivers in New York, and how, even if there was no difference within populations, you could generate this difference. To me, in that book, you ultimately come to this notion of stigma, which you then can't put a model on, you can't really characterize it. And the reason I say it was a humble position to take was because here you are—you can do the math, you've done it all these years, you've showed the world that it can actually be useful in thinking about race—you are saying, “Look, ultimately, I don't think this gets at the problem, and I don't really know how to model it.”

I think that if people really read that book, they would not think of you as someone who is not open to the discipline. All the references that you bring in, the wide scholarship that you cite. It's really saying that, “Look, we have a problem here that we have to understand together. And I'm not saying I have the answer. I'm just pointing to some of my observations.” So I want you to talk about that transition a little bit, from discrimination to stigma, if you want to call it that.

GLENN LOURY: Yeah, I do want to call it that. And thank you for such close attention to the argument of that 2002 book, Harvard University Press, The Anatomy of Racial Inequality. Actually, it’s based on the lectures I gave at Harvard, the Du Bois Lectures I gave in 2000. The book was published in 2002. I juxtapose the concept of “stereotype” with “stigma.” And I'm talking about race, I won't try to recapitulate the entire argument. But it's true that there is a different tone, a different level of specificity and formal coherence to the stereotype argument that I make. It's a statistical discrimination argument. It's a self-fulfilling prophecy kind of argument.

Again, I won't try to recapitulate it. But the basic idea is that beliefs on one side of a market and incentives for action on the other side of the market are related to each other. Employers have beliefs about groups of workers. Workers have incentives to acquire skills. And those interactions can have interesting, perhaps anomalous consequences, one of which is that if populations are discernibly distinct one from another—like black and white in a multi-racial setting—even if the populations in their fundamentals—in terms of the distribution of abilities and preferences and so on—are similar, employers can end up believing differently about the populations and as a consequence, creating incentives that induce behavior in those respective populations that confirm the employer's different beliefs. And so we get inequality almost out of the air, almost out of the ether.

But it's about incomplete information and about the consequences of people in games that interact with each other under incomplete information with beliefs and incentives being interrelated, coming to the kind of outcome that I describe, an outcome of inequality, even though the groups are not fundamentally distinct. So that idea is an economics idea. Kenneth Arrow would be among the people, my predecessors—great man, the late Kenneth Arrow—who tried to formally model this. And Steve Coate and I develop, I think, a very useful apparatus for formalizing this kind of thing that I was just describing in words. In the stereotype chapter of that book, I described that way of thinking. 

But I actually got asked a question by William “Sandy” Darity, who was a colleague of mine when I was a graduate student at MIT. He is a left-of-center black man who's an economist at Duke, a prominent economist, whose politics are different from mine. But, with respect, he once asked me a question about statistical discrimination models that really stuck with me, and it leads to the stigma part that you've made reference to, which I acknowledged as less formally developed in my thinking and perhaps less amenable to formalization.

But in any case, Sandy said to me, “Oh, so you got one of these statistical equilibria where there's self-fulfilling prophecies and the Blacks are on the short end. How come nobody experiments and learns about the fact that it's a statistical discrimination equilibrium, which is a kind of artifact of self-fulfilling prophecy and that it really is not indicative of fundamental distinctions between groups? People don't experiment. They don't learn in your world, in your and Kenneth Arrows’s world.” This is Sandy. He's contemptuous, a little bit, of neoclassical economics. So he takes the opportunity to take a shot.

And I never took that question very seriously when I was a graduate student, ‘cause Ken Arrow is Ken Arrow, and Sandy Darity is Sandy Darity, and never the ‘twain shall meet. No disrespect intended, but come on, that's Ken Arrow. I didn't take it as seriously, I think, as I should have. Because actually, the formal framework—you know, we postulate a game, we say people have the information that they have, and they formulate beliefs or rational expectations and whatnot—doesn’t admit to the kind of treatment of learning experimentation. Maybe people have advanced to the point where they can do this now, I don't know, but not to my knowledge.

I was very impressed by this sociologist I read. His name was Barry Barnes. He’s a Brit. And he says, “How does a baby learn language? Is it rational?” And he says, “Well, it can't be rational, because ‘rational’ presupposes that the baby has the capacity to formulate thought.” He says that there's a there's a sense in which, fundamentally, it's conventional. That is, the baby learns language by being embedded within a network of social affiliation and assimilating itself as a consciousness into that network by the acquisition of language capacity, and that to try to make that rational all the way down is a fool's errand.

And I thought, he seems to be putting his finger on something really, really profound about the limitations of our imputing to the agent a kind of rationality that makes their behavior necessarily the consequence of deductive reasoning. And it's exactly that framework that's employed in the stereotype theory, the multiple equilibria, self-fulfilling expectations theory that I developed in the previous chapter and that Sandy was raising his questions about. And that made me think, okay, so how does that apply to race? And I was influenced by some of these sociologists.

Orlando Patterson is another one. His book Slavery and Social Death, which is a monumental achievement, published 1982. Again, Harvard University Press. It's a survey of the institution of slavery from antiquity into the twentieth century, with deep exploration of slavery in Sub-Saharan Africa, as well as slavery in the Americas, of the captive Africans, and slavery in Byzantium and slavery in the Orient, and slavery in Istanbul in 1600, and slavery in Greece and Rome and so on. It's a great achievement, this book.

He has a theory of slavery which I thought relevant to the discussion of racial inequality in the United States because of our history. And his theory postulates that slavery is fundamentally about dishonor. It's about non-personhood. It's about the capacity to categorize and see people and not accord them standing, in some way. He thinks what he calls “natal alienation” is a fundamental part of what is at the essence of slavery, when natal alienation means, across generations, the progeny are alienated from their forebears by the intercession of the master’s property claim, so that the baby is a property of the master rather than being issue that comes through the familial line. 

And he has other very, very interesting things to say. He says an athlete who is owned by a sports team can't ply their trade without the forbearance of the guy that's got the contract is not a slave—although they do not own their labor in some literal sense—because they don't suffer the same degree of social alienation and dishonor. I'm sorry, I'm taking a long time to say this, Rohini, but we're in the longform podcast format, so I can I can do that. Thank you.

The thing about the importance of this perception of honor being critical in thinking about slavery and slavery being critical in thinking about racial inequality in America and thinking about the stigma of that being critical for understanding the persistence of racial inequality and seeing that it's something that lies outside of our capacity to formally model, is that if slavery is not simply property and people but instead involves this kind of hierarchic relationship of social status and of meanings, then emancipation is not simply a legal act where the authorities—Abraham Lincoln in the case at hand in the United States—pronounce these people are no longer property, they have been freed.

Well, yes, you can abrogate the contract. But you don't erase the dishonor for having abrogated the contract. The contract rested in part on a social convention of dishonoring people based upon their African descent. And that did not get reversed. When you abrogated the contract, that still was lingering there. That was the stigma that I was making reference to. That was the thing that keeps people from experimenting and asking questions about their interactions with others when these seemingly irrational and certainly inefficient statistical discrimination equilibria emerge. And Sandy was asking a good question when he asked, “Why don't we interrogate this framework?” That was part of what it was that I was trying to point toward. And I'll stop. But I'll just say: read sociology, whoever is out there.

MIKE GAETANI: You just tweeted within the past week about your January 2016 public symposium at CASBS, which was sort of the launch of your memoir project. And there was a mantra throughout that talk—and I'm just gonna paraphrase it—you were saying several times, “What's a good upstanding black intellectual to do?” And you asked that several times throughout. What’s the black intellectual to do about the persistence of racial inequality? That's been a through line. And then you, a few years later in 2020, appeared on a CASBS webcast, and that was the title of the episode, “The Persistence of Racial Inequality.” And the conversation revolved more or less around an essay you wrote for the Manhattan Institute. And it was a great panel involving Josh Cohen and Frank Fukuyama and Alondra Nelson. I just want to know: now, in 2023, on the persistence of racial inequality, how do you feel? Do you feel that the needle has been moving?

It's a deep question for me. And one that's a little disquieting, actually, because it gets into the psyche a bit. You know, we touched on this at the beginning of the conversation with Rohini, that the conflict between the science part of it and the role that you play in public politics and public policy. Wanting to be a good black economist, helping the situation of my people and also wanting to be a respected technical social scientist who practices economic theory at a high level. 

But when you get down to a question like incarceration—which is, by the way, the subject that led me into this self-accountability reflection that issued in my asking the question, “What is a self-respecting black economist to do?” You take incarceration. So, it's crime, punishment, policing, law and order, deterrence, incentives. Should you outlaw drugs? What's the role of poverty and neighborhoods? What are the root causes of crime?

You can't get away from the racial disparity and the application of the punishment in the society. I mean, blacks are 10%, 12% of the population; we’re 40%, 50%— depending on where you look and how you count—of the people who are being locked up. When I gave the Tanner Lectures on Human Values at Stanford in 2007, the great French sociologist Loïc Wacquant, who's a professor at Berkeley and in Paris, and a protege of Pierre Bourdieu—you see, I do read sociology—flew overnight from Paris to be able to be present to comment on my [lectures], because he was lecturing the day before. He was half asleep as he stumbled through his presentation, but fortunately he had written his remarks down and they are included in the comments in the volume on those lectures—that was published by MIT Press—called Race, Incarceration, and American Values. 

But anyway, I'm saying all that to say, why did Loïc get on a plane in Paris and fly to San Francisco just so he could comment on my talk before he had to get on another plane and fly back? Because a black intellectual of real prominence and distinction was addressing himself full-throatedly to the incarceration problem in a way that Loïc, in 2007, lamented hadn't been done. I'm not going to name names, but if Loïc were here, he’d be saying, “Why didn't so-and-so give this lecture? Why didn't so-and-so? Finally, somebody's giving [the lectures].”

And when I gave the lectures, I asked the audience in effect, “What am I to do up here?” I had a photo of an African American with dreadlocks with jail bars behind him at a window, and a bird representing freedom visible through the bars, but him longingly looking through the bars out to whatever. He's a criminal and he's being confined and he's a “thug” and he's one of these people that we are going to protect ourselves from. He symbolizes the marginal status of the hundreds of thousands—indeed, over the years, millions—in ghettos in cities around the country who end up on the wrong side of the law and then behind bars. And who speaks for them? And shouldn't I speak for them? What is a respectable black intellectual to do? 

And if you combine that with the fact that the things that you have to say may not line up with whatever the latter day “fad and fancy” tells you that a thinking person is supposed to say, you might end up saying it's not structural racism, and that that's a silly pipe dream dead end, talking about something as abstract as structural racism. What are his peers doing that he comes to value that leads him to make choices that end up with him being in a fix?

I'm putting it very, very starkly. I'm saying, on the one hand, attribute the morally objectionable and socially pathological outcome—over-representation of blacks in prison—to the system that denies opportunity to the person. That is, of course, a very appealing progressive posture. Or consider that each and every one of us exercises discretion over the things that we control. And moreover, maintenance of order in the society is not going to be possible unless you hold people responsible for the actions that they take. The incentives your response to this condition create will either help to ameliorate it or will give people reason to double down on what they're doing. And so, as an economist who bears bad news, I'm here to tell you: Black Lives Matter have got it wrong! That was the alternative argument. 

This is all hypothetical, people. I'm not actually taking a position. I'm describing the dilemma of the black self-respecting economist, and what is he to do. Is he to decide with his people in the fight for justice in a form that his intellect tells him is woefully incomplete for capturing the subtle complexities of the situation? And I have had, throughout my life, to make that choice.

---

This post is free and available to the public. To receive early access to TGS episodes, an ad-free podcast feed, Q&As, and other exclusive content and benefits, click below.

---

ROHINI SOMANATHAN: One of the striking conceptual distinctions that you made was on problems of contact versus contract. It was just so beautifully put. And I think you came to the conclusion that the problems of racial inequality are problems of contact and not primarily of contract. We've made progress on the contract front, but we really need to make progress on the contact front.

I want to move that discussion a bit forward and end with us [economists] and the other social sciences again. Because where does that put us? What we're good at is the contract stuff. If we have to reinvent ourselves, if we have to speak to the others and have to really bring in their insights and yet contribute in the way that I think economists really can contribute … We think about strategic behavior. We think about equilibrium. We think about aggregation, very importantly—and Tom Schelling is the one I would really thank for bringing this into the discipline. He moves us from micro-motives to macro-behavior, and shows us that that link isn't at all obvious. So I think there is a lot we can contribute.

But I'd like to go back to the title of the essay you wrote for the co-edited book that I did with Danielle Allen, Difference Without Domination. Your chapter was called “Relations before Transactions.” And I want you to say a little bit about how you think we, as economists, should start that project. I think it's really a project we need to begin. I don't think it's a project that's underway. Or if it is, it's still in adolescence or childhood. We still haven't really developed that idea. How should we go about it? And how should we talk to the other social science disciplines? We're living in a very polarized world, Glenn. And I think we can't do much about a lot of the political polarization. But I think if we had less polarization in academia, that would at least allow us to fulfill our purpose as public intellectuals better. How should we go about it with the wisdom of all these years? Help us get started.

Oh, you put me on the spot now. Let's see what I can do, though. I think, humility, understanding the limits of our discipline and being open-minded and listening and all of that. I think, humility. I think, a breadth of interest. I think, read widely, and this kind of thing. And I don't know that the culture of our profession—economics, academic economics profession—admits enough of that: of encouraging a breadth of intellectual interest in our charges so that they grow through the course of their professional careers and become not just adept practitioners of the particular craft of model-building and statistical inference, but real thinkers who engage with society, with the problems of society broadly. Remembering that we're a social science. We’re not statistics, although we use statistics; we're not mathematics, although we use mathematics. We’re a social science.

Read the history of economic thought. Read some of these classic [texts]. Read Ricardo or read Smith. Or read Marx! You know, argue with him if you want to. Read Keynes. Read Marshall. I don't know if anybody even knows who Marshall was, nowadays. Read Joan Robinson. I don't mind engaging the Cambridge capital theory argument, or whatever. There's a lot of stuff to read. And if you take yourself out of the small world of technique—it's very important, but it's a relatively narrow berth—and you put yourself in the shoes of somebody sitting in 1850 trying to figure out what the terms of trade are supposed to be and why. Or trying to figure out whether a bridge is worth building and how you’d even pose the question in a meaningful way and then answer it. “The areas under these curves actually have welfare significance,” or something like that. You put yourself in the position of these people thinking about well, what could determine the terms?

I'm thinking of Irving Fisher right now, or whatever or Paul Samuelson, some of the great, epic engagement with the economics of public goods and how to frame [the questions]. I'm sorry, I'm rambling. Read widely. The questions are not, at the end of the day, about technique. The questions are about social life and about what happens when we put people together in a system of interaction, and so on. So I know that I have not answered, '“How can we facilitate the development of economics in a way that's congenial to and knowledgeable about what it is that other disciplines have to offer?” I think the answer has to lie in pedagogy and what we teach our students and how we encourage them to develop their intellectual gifts.

ROHINI SOMANATHAN: Bringing us back to the world of policy, and how do we offer education to everyone rather than just to a select few. Here I want to go back to your conceptual distinction between reward bias and development bias. And the question is, really, how do we encourage development? How do we remove development bias? And one thing that's talked about a lot in politics these days is the unaffordability of a college education. And you've been through many institutions, you haven't just been to the very elite institutions. You've been to community college. You went to community college, and then you went to Northwestern. At a time when you were a young father, you had a full-time job, and you've seen a lot of different people in a lot of different places. If we are really to make opportunities widely available, do you have thoughts on how we might do that, in terms of changes in educational policy, the educational infrastructure? I know, that's a big question. So just pick any part of it you want. And then we’ll end there.

When I think about my own development, it's true that I was a college dropout. I was in college and my girlfriend became pregnant and we married. I dropped out of school and went to work. I was working at a factory. And then I started clawing my way back into academia. The first step in that process was a community college, and it was absolutely transformative for me. It was basically free. I mean, I think a couple hundred dollars a year or something. In the late 1960s, it was pretty high quality.

I had a calculus teacher who was a former engineer, Mr. Andres, who had retired from his job and was taking up teaching calculus—and he was really quite good as a pedagogue and knowledgeable about the applied mathematics of it—who took a liking to me and encouraged me and then ultimately referred me to Northwestern, where I ended up getting a full scholarship. But community college was important for me.

I think, based on what I hear from people who study these matters more closely than I do that, the proliferation of community college as a avenue for personal development—with the idea that people would go for a couple of years as I did, and then transfer to a four year institution, as I did—it’s not for everybody. A lot of the young people coming along don't really have the skills of quantitative and verbal reasoning that makes a college path nearly as plausibly sensible for them to pursue in terms of resource allocation, likelihood of success. Because the completion rates for four year for these kids are not as high as you would like them to be as might be the case if they were instead diverted into vocational education to acquire some kind of skill that will allow them to work at some craft, or something like that.

So if we're talking about for the everyman, for the median person, I'm not sure that a four year colleges is the path for everybody. But, again, I speak only based on my familiarity with the thinking that some people who are more knowledgeable about the question. 

I mean, I could opine about education. I could talk about the importance of early childhood and neurological development before young people even get to the formal institutions, and how interventions of a variety of kinds that have been tried around the world, direct themselves to that set of issues. I could talk about the provision of educational services by public institutions in the United States where there are real big policy debates going on about charter schools and the choice matrix for parents in terms of where their kids get educated, and they influence a public school teacher unions. [Those questions] are arguable, and I'm not trying to pick a fight with anybody. I do have views about those matters, but they're less important than just calling attention to the fact that the vast bulk of the population, K-12, quality is a key issue and that there are places where the outcomes are not so great. I could talk about parenting and culture and all those kinds of things, which are unspeakable topics, if you address yourself to inequality, because they seem to be blaming the victim.

ROHINI SOMANATHAN: Just to interrupt you, Glenn. One particular question on this is there have been a lot of programs which actually tried to improve opportunities for children living in poor neighborhoods. The Moving to Opportunity program is an example of that. And there's been a fair amount of empirical work now, that shows that has good effects in the long run. Raj Chetty has done a lot of work, and others. There's also been a lot of talk about charter schools and voucher programs. And how do you feel generally about giving a few promising children opportunities and then tracking them rather than raising the overall quality of public education? Do you think there's a tension there? Do you think the rhetoric of providing opportunity to a few has distracted us from trying to really build quality public education, K-12? What's your view on that?

I think it's a hard problem. So it comes up, for example, with the selective tracking of gifted and talented, where people are saying no. It comes up in the charter school debate, because there's a cream-skimming allegation of charter school. Providers deny it. They say we're just peeling people off of a waiting list. But you can't help but notice that the parents who were most proactive in their children's lives, most resourceful, and so forth are going to be the ones to our first in the line to get take advantage. So even if a charter is providing better services to the kids who are lucky enough to get access to it, it might end up diverting resources away from the system, which is a system of last resort for other kids. I think there's a unavoidable conflict in that.

There's clearly the conflict in terms of race diversity, of do I maintain, let's say, in a selective university, a high standard of admission that relies on standardized tests to assess students’ suitability, when I know that because of the history of developmental disadvantage, the populations of color may be underrepresented amongst the kids who have the best chance of doing well at these instruments of assessment and therefore they will be underrepresented amongst those who are selected? Hence, the pressure is to, in one way or another, relax reliance upon instruments of assessment that can be presumed to disadvantage the kids who are in the population that you most want to help. And that, I think, inevitably pits a kind of maintenance of standards and a kind of commitment to elitism and meritocracy. Not inexorably, but I think that tension will be there against the kind of egalitarian sentiment and a desire to open up institutions and access to others. That's a very hard and a very important problem.

And you can get it wrong in both directions. You can get it wrong by being completely indifferent to the fact that large, identifiable segments of your population are going to be underrepresented because you're using, let's say, a standardized test for admission, and they're not doing well on it. “They're not doing well on it” well may be a signal of the fact that, at this point in time, they haven't acquired the mastery over the material to the same extent. But it may also be a reflection of the of the deeper structural conditions in the society, which you are now ratifying by relying upon this entrance.

That's on the one side. But on the other side is a desire to be egalitarian and inclusive that causes you to issue any reliance upon differentiating information, and that waters down your program. It's a kind of road to mediocrity and a loss of the edge of innovation and competitiveness. And this is, of course, one of our conundrums here. Because we're decent people, we economists, but we also know that incentives matter, and everybody can't win the race. I mean, if I have a race, I'm gonna get kids to run faster, because everybody wants to be in front. But then, by definition, somebody's going to be in front, and I can't help but induce a certain amount of disparity in outcome if I create an institutional framework that encourages the effort in the first place. It seems to me this is what makes it interesting.

---