Interview with Glenn Loury

Interview conducted March 7, 2014

Glenn Loury is an exceptional scholar, with important work in income inequality, public finance, discrimination, game theory, natural resource economics and other areas. He is also African American, a rarity in economics. While race has neither defined nor limited Loury’s scholarship, there is no question that it has influenced his path. His doctoral dissertation examined the dynamics of income inequality and argued that “continued racial economic disparities...reflect the social and economic consequences of historical inequity.”

His model of income distribution included “social capital,” a notion that skills and earning potential are highly dependent on family and community background. “An individual’s social origin,” he wrote in 1976, “has an obvious and important effect on the amount of resources which are ultimately invested in his development.” That background, in turn, is shaped by history, which—in the United States—includes the enduring legacy of slavery and segregation.

Income distribution is thus determined in part by factors with long residual impact well beyond individual effort and innate ability. “The eradication of racial income differences [therefore requires] compensatory efforts,” wrote Loury, “within both the educational sphere and the world of work.”

It was a powerful argument, and his skill in making it led to positions at Northwestern, then at the University of Michigan and, in 1982, at Harvard, its first African American economist with tenure. After a decade, Loury moved to Boston University, and since 2005, he has held a chair at Brown University in economics and social sciences. His research over these years has deepened within economics, earning him honors and wide recognition in the field; it has also broadened far outside economics.

That divergence was predictable, suggests Nobel laureate Robert Solow, his MIT thesis adviser. “It was clear to me [in the mid-’70s] that he would be an outstanding economic theorist. But I think it was equally clear to both of us that there would be enormous pressures on him, as an eminent black in a highly technical, uniformly white field, to spend energy on other roles.”

Indeed, Loury writes and speaks widely on topics as diverse as spirituality, U.S. incarceration, slavery reparations and self-censorship in political discourse. Despite this passionate participation in ongoing social debates, Solow observes, he continues “to produce cool analytical economics.”

In the following interview, Loury covers a mere sliver of his wide-ranging scholarship.

Region: You’re best known, of course, for your work on income distribution, racial inequality and discrimination, so I’ll want to focus much of our conversation on that research. But I hope we can also cover your research on why Pigouvian taxes alone can’t deal efficiently with externalities, on game theory and on exhaustible resources. And I’d be remiss as a Fed employee if I didn’t ask about rotating savings and credit associations.

Loury: OK, this is going to be fun.

SOCIAL CAPITAL AND EMPLOYMENT DISCRIMINATION

Region: I suspect we won’t get to all that material, but let’s begin with your research on labor market discrimination and social capital, if we could.

Eminent economists before you—Gary Becker and Ken Arrow, for example—had studied employment discrimination, of course. Becker, I believe, considered discrimination based on employer tastes; Arrow based his theory on the impact of limited information regarding worker productivity.

In your dissertation, you proposed a new approach that focused on the importance of “social capital”—the term you used for family and community background—for skill acquisition and future earnings potential. Could you describe that approach and what it suggested about economic policy to alleviate racial discrimination in the workplace and improve income distribution? And, specifically, what it implied for policies to ensure equality of opportunity.

Loury: My principal point of departure when writing that dissertation in 1975 and 1976, building on the work of Gary Becker, was to “socialize” the human capital investment decision. That is, I wanted to take explicit recognition of the fact that the acquisition of human capital occurred in a social context.

The insight there was that not all of these external influences on the costs and benefits of acquiring human capital are marketable, traded commodities. Indeed, many of these are not “commodities” at all. Some of these external effects, I argued at that time, come about as a consequence of the preexisting social relationships between people within families, social groups of various kinds, identity groups and racial “communities.” (I put that word in inverted commas because I don’t mean only a geographically extended space. I mean a set of social networks.)

My idea was that these networks mediate the spillovers from the human capital investment one individual makes onto the costs and benefits of similar investments of other individuals within the same network. And that effect was not well-represented in the classical Beckerian framework. It was not only Becker, of course, who wrote about human capital; it was Schultz, Mincer and others as well. This school of thought simply posited—and I mean this not as criticism, but as observation—that, in effect, these human capital investments affecting their productivity were based on inputs that people could buy at a price if the returns justified their acquisition.

What I was after in my dissertation was to explain why it is that the African-Americans might lag behind, in an extended way, even after the equal opportunity regime of the Civil Rights Bill was put into place. I was trying to say, “That’s not enough. Equal opportunity of that sort, while welcome and long overdue, is not enough to remedy the long-term inequality problem.”

You could get stuck with the remnants of history because people are embedded in social networks, the nature of which reflects to some degree the effects of past discrimination. Some communities, because of their historical treatment, are impoverished with respect to the human development resources that people must have access to if they are to succeed in the labor market. I speak here not only of material resources, by the way. Also things like, what do the peer groups hold in esteem? What do people derive social benefits from accomplishing?

In short, I felt that the Beckerian characterization of employment discrimination as merely an impediment in the marketplace because some people have a taste for discrimination was a somewhat limited framework, on both the supply and demand side of the labor market.

On the demand side (employers, say), those tastes don’t just come out of the air. They need to be accounted for in some way, so they’re really a product of their history, which in the United States includes a history of slavery. Likewise, on the supply side, the nature of the social networks in which people are embedded that influence their costs and benefits from human capital acquisition also depend on history, identity, geography and so forth.

In the United States, race has a very particular valence in that history. It’s not the same as gender or sexual orientation. I’m not saying that racial discrimination is better; I’m not saying it’s worse. I’m just saying that race is different from some of these other variables, in the context of American history.

When European immigrant groups were fighting over the bottom rungs of the ladder with the black American migrant groups coming into U.S. industrial cities in the early part of the 20th century, that was a historically specific kind of contestation. I just thought that an abstract specification of an employer’s disutility from hiring blacks, as Becker had argued in his influential book The Economics of Discrimination, didn’t get to the core of what was going on.

That’s the demand side of the labor market. On the supply side, I also thought that standard theory—human capital theory—didn’t capture the full impact of discrimination because one consequence of discrimination was to deprive individuals in the maltreated group of an opportunity fully to develop their human potential.

Region: What does that imply for policy to alleviate employment discrimination based on race, in particular?

Loury: Well, I’m not sure, in terms of what particular bill should Congress pass. But in terms of how to think about policy, maybe the first thing it would say is, if I do see those deficits on the supply side, which I do, then do a proper accounting.

By that I mean, suppose I have a regression equation with wages on the left-hand side and a number of explanatory variables—like schooling, work experience, mental ability, family structure, region, occupation and so forth—on the right-hand side. These variables might account for variation among individuals in wages, and thus one should control for them if the earnings of different racial or ethnic groups are to be compared. One could put many different variables on the right-hand side of such a wage regression.

Well, many of those right-hand-side variables are determined within the very system of social interactions that one wants to understand if one is to effectively explain large and persistent earnings differences between groups. That is, on the average, schooling, work experience, family structure or ability (as measured by paper and pencil tests) may differ between racial groups, and those differences may help to explain a group disparity in earnings. But those differences may to some extent be a consequence of the same structure of social relations that led to employers having the discriminatory attitudes they may have in the work place toward the members of different groups.

So, the question arises: Should an analyst who is trying to measure the extent of “economic discrimination” hold the group accountable for the fact that they have bad family structure? Is a failure to complete high school, or a history of involvement in a drug-selling gang that led to a criminal record, part of what the analyst should control for when explaining the racial wage gap—so that the uncontrolled gap is no longer taken as an indication of the extent of unfair treatment of the group?

Well, one answer for this question is, “Yes, that was their decision.” They could have invested in human capital and they didn’t. Employer tastes don’t explain that individual decision. So as far as that analyst is concerned, the observed racial disparity would not be a reflection of social exclusion and mistreatment based on race.

Region: They simply chose not to complete high school, or go to college, for example.

Loury: Right. But another way to look at it is that the racially segregated social networks in which they were located reflected a history of deprivation of opportunity and access for people belonging to their racial group. And that history fostered a pattern of behavior, attitudes, values and practices, extending across generations, which are now being reflected in what we see on the supply side of the present day labor market, but which should still be thought of as a legacy of historical racial discrimination, if properly understood.

Or at least in terms of policy, it should be a part of what society understands to be the consequences of unfair treatment, not what society understands to be the result of the fact that these people don’t know how to get themselves ready for the labor market. That’s the spirit of what I was trying to get at in 1976.

“EQUALITY OF OPPORTUNITY” PROJECT

Region: Let me ask you about another piece of research, not your own. I’m sure you’re aware of the Harvard/Berkeley “Equality of Opportunity” project.

Loury: Raj Chetty and his colleagues.

Region: Exactly. They’ve just published an NBER paper on “The Geography of Intergenerational Mobility in the United States.”

They found that greater “social capital” was second highest among the five factors best correlated geographically with higher income mobility. The other top factors were areas with less residential segregation, less income inequality, better primary schools and greater family stability.

What are your general thoughts on these empirical findings?

Loury: I just saw Raj Chetty give a lecture at Brown—literally, two weeks ago—on this very paper. Of course, by now it has been widely discussed in the press. I have not, I must say, gone through their paper carefully, though I think I will with my students in a graduate seminar I’m teaching this semester.

So, of course, I won’t want to comment on whether or not I think they got it right. But these are not neophytes. These are serious people, so I think it’s certainly possible for me to take at face value much of what they’re saying. And I thought Raj gave a very convincing presentation.

Region: Does their concept of social capital align with yours? I think they use Robert Putnam’s measure.

Loury: Right, it is Putnam’s concept, and Putnam, a very distinguished political scientist, has been observing variations across U.S. cities in various measures of trust. You know, “Do I know my neighbor?” and the like. He looked at trust within as well as across racial groups and found it to be negatively associated with ethnic diversity at the city level, so that in places where you had a relatively larger presence of minority groups, you tended to find lower measured levels of trust.

Chetty and colleagues are finding lower social mobility in cities with more ethnic diversity. One imagines that the mechanism probably involves things like the lower provision of local public goods in more diverse cities; the quality of the schools kids attend may vary inversely with ethnic diversity measures across metropolitan areas and things of this kind. It is difficult to interpret some of these findings, since I’m not sure how one is supposed to think about regional variation in something like social mobility, given that people are moving around.

Nevertheless, they make an effort in their study to deal with selection effects. These questions came up in Raj’s seminar presentation. But it’s a big, complex general equilibrium system, right? The regions are all interconnected with each other.

They’ve tried, certainly. For example, they measured the location where the kids were when they were 14 to 17 years old, so then even if they move around later, they’ve still got the local effect identified at the formative period of their lives. I’m not sure this is adequate, though. You’ve got two or three leaps here. This is not the typical sneering economist who wants to just dismiss everything. I’m simply saying that it’s complicated, and I’m not sure that it’s wise to leap too quickly to policy conclusions from the work of Chetty et al., although I know that’s what the press wants to do.

Still, I find this work to be very provocative and interesting. For family structure, for example, they looked at part of the sample where the mothers didn’t have any children out of wedlock, and still, if these mothers were in an area where there was a higher local out-of-wedlock birth rate, then their children were less likely to experience movement from the bottom to the top of the income hierarchy. This was true even among kids who were born and raised in intact families.

So, they were arguing that there was an environmental effect that operates across a metropolitan area, even for families that were not specifically implicated in this particular behavior (i.e., out-of-wedlock childbearing). That’s all very interesting. It really makes you think.

AFFIRMATIVE ACTION POLICIES

Region: You’ve written numerous papers on affirmative action over the years, and several recently with Roland Fryer. In a 2013 Journal of Political Economy piece, “Valuing Diversity,” the two of you consider policy interventions to improve opportunities for the disadvantaged and you look, specifically, at dimensions of visibility and timing. So, whether affirmative action policies are “sighted” or “blind” and whether they intervene before or after worker productivity is basically established.

Loury: Precisely.

Region: Can you briefly describe your results and, particularly, the major difference in timing of intervention if policy is “blind” versus “sighted”? I was surprised by your finding that to be efficient under sighted affirmative action, policy should focus on job slots and not skills acquisition.

Loury: That is a very well-informed question. Thank you. You laid it out exactly.

So, I could only repeat what you just said, which is that we see two different dimensions along which you might usefully differentiate affirmative action policies. One is the stage in the process of developing a productive worker where the policymaker intervenes—either early or late in the cycle of development. That is, one needs to consider whether you intervene with affirmative action at the point where the person is making an investment in human capital or at the point where people—having made their human capital investments: higher education and the like—are competing for access to opportunities. If you are trying to boost the status of a disadvantaged group, at which stage, early or late, should you intervene?

The other way in which we thought it useful to distinguish between policies, in a very broad-based sense, was, as you mentioned, whether they’re “blind” or “sighted.”

“Sighted” affirmative action policies are those where the policymaker is quite prepared to overtly favor some population group—say, African Americans in the United States. So, if you’re interested in boosting the presence of African Americans in college, for example, are you prepared to have different admission standards for the African American applicant and the non-African American applicant? If so, then your policy would be what we call “sighted.” If not, then your policy would be “blind.”

Under a “blind” policy, a policymaker still wants to boost the status of some target beneficiary group, but does not want to engage in overt differential treatment. Such a policymaker would, therefore, need to look for indirect ways of accomplishing his aims—by, say, subsidizing for everyone those activities that are determined in advance to disproportionately favor the target population.

That’s a very broad way of describing the framework of analysis that we develop in this paper. One of our key insights is that under sightedness (again, overt discrimination in favor of a particular group), the very act of boosting people’s access to slots—that is, putting a thumb on the scale in their favor at the point where they compete for positions—implies a subsidy to their acquisition of skills.

Many people have the intuition that an affirmative action policy is not favoring skill investment when putting a thumb on the scale favoring some group in the competition for positions. But, in fact, it is, at least implicitly. This is the insight, if you will, because it implies that if a later intervention is properly anticipated, then an earlier intervention may not be necessary; it may be redundant. Indeed, that’s what we show in the paper.

Now, this result—that we find quite interesting—requires the assumption I just referred to: that when making their decisions about how to invest in the development of their skills, people be farsighted enough to anticipate the consequences of their being favored at the point of slots allocation. That assumption will not be plausible in every case (youngsters can be unnervingly short-sighted...).

Still, given this necessary assumption—that individuals subject to future affirmative action policies anticipate this accurately—then our result is of a piece with other intuitions that come out of applied microeconomics.

For example, in the area of industrial organization, there has been a classic problem with respect to vertical integration [whether the most efficient way for a monopolist to exercise his monopoly power is to raise the price for his input, or to integrate forward by acquiring downstream manufacturers]. The solution to this classic problem is that one wants any necessary distortion to be as close to the end of the chain-linked process of production as possible. That’s the insight.

Another example comes from public finance. Diamond and Mirrlees, in their classic optimal taxation papers in a 1971 AER,¹ proved that, under certain technical conditions,² if the government can tax final commodities and/or intermediate inputs, and if the government seeks to raise a given amount of revenue with the least distortion to social surplus, then the efficient tax system involves no distortion in production. So, no tax on intermediate inputs; tax instead the final commodities. This is called the “efficient public production” result in the public finance literature spawned by Diamond and Mirrlees.

This is similar to the result in my JPE paper on affirmative action with Fryer: The distortion (in our case, preferences for a disadvantaged group) should take place “downstream,” at the point of competition for final positions, rather than “upstream,” at the point where people are investing in their own productivity.

So, those are two examples of similar economic contexts where similar results have been found. Now, you’d think that for affirmative action it might be different, that, well, it’s always better to go early.

Region: Pre-K—as early as that, perhaps?

Loury: Sure, Pre-K is something people are advocating these days. And, indeed, there may be other reasons, not in our model, having to do with cycles of development and so forth, which would explain why early intervention of a different kind is warranted.

But if it’s purely in the framework of our model, I think our finding is explicable in terms of intuitions that you find in other areas of economics.

So, I have said two things really. First, that our conclusion depends on people having the farsightedness to anticipate the consequences of the fact that they’re going to be favored at the slot competition stage, and that this is an implicit subsidy of their investments. And, second, that writers in the industrial organization and public finance literatures have established results similar to ours. Because of that, I was not at all surprised by our finding.

TRANSITION FROM FORMAL SEGREGATION

Region: Let me ask about another piece of recent work, this with Bowles and Sethi. You consider racial and ethnic discrimination in many societies—from the United States to South Africa to South Korea—and develop a model to analyze factors that affect the evolution of income distribution during transition from overt discrimination to equal opportunity.

You found, I believe, that the course of evolution depends crucially upon three factors: the degree of segregation in social interaction post-transition, the society’s demographic composition and, third, timing of integration relative to demographic trends.

Can you describe those findings a bit more deeply?

Loury: I love being invited to describe this problem because it goes back to my dissertation. Most dissertations don’t have 35-to-40-year legs! You know what I’m saying? In fact, two papers grew out of my thesis. One was published in Econometrica in 1981, “Intergenerational Transfers and the Distribution of Earnings.” This other was published in a conference volume in 1977 under the title “A Dynamic Theory of Racial Income Differences.” Both are still being cited and, I must admit, I’m quite proud of that.

What I was unable to do in my original “Dynamic Theory” paper was to provide a satisfactory formalization of the process by which racial segregation limits the economic opportunities of the members of a disadvantaged racial group.

There is an ad hoc character to my argument in that paper. I said then, in effect: OK, let’s contrast two worlds: one where an individual’s cost of getting human capital depends only on his family background and the other where an individual’s cost of getting human capital depended on his family background and his community background.

Family background was proxied by the parents’ earnings, and community background was proxied by the average earnings of the racial group to which the individual belonged. I just slapped that specification down on the page and gave no account of how these effects might be derived. That is, I simply posited cost functions for human capital acquisition that had these things as their arguments. I didn’t explain where they came from.

Now, the internal family thing is kind of plausible, and you can tell a story about that. It could be that higher-income households have access to certain goods that make it easier for their kid to get effective education. Something like that.

I also had a story about the community background effects, which had to do with peer group influences and whatnot. But I didn’t try to model that at all. I just said it was there.

And then I asked, What will evolve if it’s only family background, and you have a group that is behind (on average) because of a history of discrimination against them, but now you have equal opportunity for the individual members of that group going forward? So then, yes, initially those families are going to be disproportionately poor because they were behind due to the discrimination. But now that there’s equal opportunity, you’ve got a difference equation, a dynamic system. There will be some mobility; it may not be perfect mobility, but there will be some. And then, let me ask a question in the most generous way: Suppose we take time to infinity. Does the disadvantaged group ever catch up?

What I was able to show was that, if the only influence of the past on the current and future prospects of group members was that their parents had lower incomes, then eventually they would catch up. (There were some technical conditions involving diminishing returns across generations to the benefit of higher parental income and so forth.)

On the other hand, I exhibited a numerical example in the context of my model where, if it was not just the parents’ income but also the community group income which adversely affected the cost of acquiring human capital for members of the disadvantaged group, then you wouldn’t necessarily catch up, even in the longest of long runs.

The result was interesting and, if I may say so, important. But the technique by which it was demonstrated was not very satisfactory by the standards of a modern economic theorist. Still, it was a creative way to pose what many people considered to be a critical question.

Region: And that work from your dissertation was the starting point for Sethi and Bowles.

Loury: Right. Rajiv Sethi and Sam Bowles went back to my “Dynamic Theory” paper and said: Well, let’s try to make explicit how the community income effects that Loury talked about might actually work. They offered a formal but very simple story about that, which was to posit that all individuals have certain other individuals to whom they are connected or with whom they are affiliated. This, they argued, constitutes an individual’s network, and the average earnings of a person’s network affect the cost to that person of acquiring human capital.

Then they suggested the following model of network formation, through which the impact of racial segregation could be made explicit: Absent any racial segregation, with (say) blacks and whites in the society, the chance that a randomly selected person in anyone’s network is a black or a white equals the proportions of blacks or whites in that society.

Thus, with no racial segregation of networks in this sense, even if a group started out behind because of past discrimination, the young people in that group would not be adversely affected by having less beneficial social networks because every individual’s network would be formed as a result of the same random process.

In contrast, the relative burden of belonging to any historically disadvantaged group will be greater, the greater is the degree of in-group bias in the random process of network construction. Segregation bias is modeled by supposing that the chance of a randomly selected member of any person’s network being from that person’s own group is greater than that group’s share of the society’s population.

Region: Certainly seems a realistic way to model bias in one direction or another.

Loury: In the limit, a perfectly segregated society would have the property that everyone’s network consists only of members of that person’s own group, and vice versa. So, to capture the extent of this own-group bias in network formation, Sethi and Bowles set a “segregation” parameter to zero for perfect integration and to one for perfect segregation. The parameter captures the range between perfect integration and perfect segregation, then, in terms of a probability-weighting on how a person’s network is constructed.

Region: Now, in your joint paper with Bowles and Sethi, you do this analysis post-transition from formal segregation to a society in which segregation is no longer legally permitted, but does persist informally.

Loury: Yes. Post whatever it was that caused the groups to be unequal in the first place.

Region: Apartheid or...

Loury: Yes, apartheid or Jim Crow or whatever. And, given this simple depiction of the social segregation process, we’re just saying, all right, now you’ve got a level playing field going forward, but you’ve got initial conditions and you’ve got intergenerational overhang. Let’s see where the thing goes.

Rajiv and Sam conjectured in their earlier paper, and we show in our joint paper, now published in the Journal of European Economics Association, that you can put down a very plausible model in which this social segregation parameter exhibits a kind of threshold influence on the dynamics of what happens after transition.

For segregation below the threshold, historically inherited group inequality eventually withers away in the face of current and ongoing group equality of opportunity. But for segregation above the threshold, the historical inheritance of group inequality might endure forever, notwithstanding the permanent abandonment of racially discriminatory practice in labor (and other) markets.

Moreover, the threshold above which social segregation implies permanent economic inequality between racial groups depends on the relative size of the groups. That’s how demography and segregation interact with each other.

Region: That interaction would account for the potential post-transition differences in, say, Bangor, Maine, and Baltimore, Maryland—as mentioned in your paper—since their black-white ratios are so different.

Loury: Exactly, though anything I’d say along those lines would be speculative because I have not done any careful empirical investigation of those cases.

I’m happy about this paper with Sam Bowles and Rajiv Sethi because, in a way, it completes in a rather satisfying fashion a project that I began almost 40 years ago. And it does it by formalizing this idea that too much social segregation can get in the way of a natural recovery from a history of discrimination against groups. The overriding idea here is that social discrimination and economic discrimination are complements to each other in terms of holding a group back.

In a world where there is some economic discrimination, the existence of social discrimination can amplify and perpetuate the effects of that economic discrimination. That’s how all of this is relevant to this idea of “transition.”

TRANSITION FROM SEGREGATION POLICY

Region: You discuss tentative implications in that paper about what this might mean for different U.S. cities, for example, or within South Africa.

Loury: Again, the policy implications come at a fairly high level of conceptualization, not at the level of programmatic implementation. Essentially, how does one want to think about policy?

I’m not a South African, and I don’t pretend to have deep knowledge of the history or the social structure of the society. But I think I can say with confidence that equalizing labor market opportunities and capital market access for all South African people will be insufficient to overcome the enduring consequences of generations of apartheid, which entailed the exclusion and marginalization of black South Africans.

Mind you now, this is not simply a sophisticated way of advocating for affirmative action in South Africa, the United States or anywhere else. To me, the implications of this analysis are far more profound than that. People who would think affirmative action might be a sufficient solution to the problem of making a just transition after a history of racial discrimination are not thinking deeply enough about the interplay between social and economic structures.

Rather, the policy responses that might be effective in overcoming the problems that we identify are more likely to involve increasing public expenditure on infrastructure in the rural communities where large numbers of black people in South Africa live. Or it might entail a concerted effort to close the gap in primary and secondary education between racial groups in the United States. That would be much more demanding on the society than merely setting aside a few seats in a college class for the beneficiaries of affirmative action.

Or to give another example of what might be required to overcome the legacies of generations of social exclusion, it might require noticing that a regime of social control and public safety, law enforcement, prisons has been created with a vastly disproportionate incidence across racial groups, which requires being reformed before true racial equality of opportunity can be achieved.

Of course, I’m not saying here that there should be an exemption from criminal punishment for those whose life chances have been diminished by past discrimination. I’m not making excuses for criminals. I’m just saying, how one deals with lawbreakers—the severity and sweep of punishment policies—can be one of the ways that group inequality gets locked in.

SOURCES OF RISING INEQUALITY

Region: I’d like to ask you about rising U.S. income inequality—a very prominent issue these days, for obvious reasons. Many theories are put forth as to the causes of rising inequality over recent decades—since you wrote your dissertation, actually—from broad structural changes like technological change and a transformed international economy to those perhaps more amenable to policy intervention.

What are your general thoughts on the factors behind rising U.S. inequality?

Loury: Well, you know, I’m not in as good a position as are some of my colleagues to address this. Larry Katz and Claudia Goldin at Harvard have looked a lot at skill-biased technology change. David Autor at MIT as well could give you a detailed accounting of what all these various studies are showing about the decomposition of inequality trends across different explanatory factors.

I have the impression, or even the belief, that an increase on the premium on skills is a big part of it and that the college-high school wage gap has increased as a result. I do believe that skill-biased technical change is a real thing.

I also can’t help but think that globalization and the ability to market services across borders and create these very high returns for people with specialized skills plays a part.

I don’t know what share of increased inequality is being driven by the financial sector, but I know it’s going to be a quantitatively measurable effect. I don’t know how much is due to excessive executive compensation. I tend to think that that gets overstated.

On the other hand, I also tend to think that Sherwin Rosen’s 1981 AER paper, “The Economics of Superstars,” offered a profound insight. If you can manage, and manage better than this guy over here, so now you’re managing over $500 billion worth of stuff instead of $500 million, you know, that’s going to imply a convex function measuring the linkage between financial rewards and managerial aptitude.

Region: And Rosen’s theory doesn’t just apply to managers, of course—it applies to “stars” of all sorts: athletes, singers, actors...

Loury: Right, it does! And it illustrates the beauty of the economics of the Chicago school. I was trained at MIT but I have always found there to be an elegance in Rosen’s and Becker’s pioneering contributions to labor economics. One reads those papers and one thinks—you know—this is Economics.

What’s the difference between economics and mathematics? I mean, it’s not all about, you know, “proving hard stuff.” Some of it is just about getting the right curves, seeing the right trade-offs, modeling the right margins and seeing implicit markets where no explicit markets exist—stuff like that. That I think is very characteristic of the Chicago style, and I’ve always admired it.

RAISING THE MINIMUM WAGE

Region: I’d like to ask for your thoughts on one option put forth to address rising inequality. President Obama and others are floating proposals to increase the minimum wage. Several states have passed, or are considering, such legislation. Back in 1981, you analyzed the impact of raising the minimum wage, in research with your future wife, Linda Datcher.

Loury: Right, we were married two years after that, in ’83.

Region: Are the findings from that study still relevant? Would minimum wage legislation now help reduce inequality?

Loury: I don’t think so.

Region: Your 1981 paper suggested that implementing a minimum wage hike would discriminate against young black people.

Loury: Yes. It wasn’t absolutely definitive, but that’s what the data we were looking at suggested at that time, and I guess that’s what I still believe, although it’s become ideologically unacceptable to say so. I think there’s a kind of anti-intellectualism that’s afoot, especially in the Democratic Party—and I’m a Democrat—which economists should be alarmed by.

If you say that demand curves slope downward and supply curves slope upward—in other words, if you insist that there’s no free lunch and that people respond to incentives—you get labeled by many Democrats these days as some kind of neo-liberal ideologue who thinks the poor are lazy.

Now, I’m no fan of Rand Paul, but I think he had a point when he said that if you extend long-term unemployment benefits, you make it easier for people to be able to subsist without having to find jobs. And he suggested that one consequence may be to extend the duration of unemployment for some people to such an extent that their subsequent employability is undermined.

That may or may not be true. It’s an empirical question, a difficult empirical question. But one thing is for sure: Such a claim is certainly not a moral statement. And yet, the way many op-ed writers, politicians and talking heads on MSNBC and other places responded to that was, “Oh, my God! He just said that the poor people who don’t have jobs don’t want to work; he just called them lazy!”

Well, as someone who has read a principles of economics textbook or two, I wish to insist that he did nothing of the kind.

Furthermore, the 1994 study that David Card and Alan Krueger did comparing restaurateurs in New Jersey and Pennsylvania does not—I repeat, does not—speak to the long-term implications for the fast food and other low-wage and unskilled laborers’ employment of a significant and indexed increase in the minimum wage.

For example, it doesn’t address whether people have the time to conceive and invent new technologies that would economize on low-skilled labor between the time that Card and Krueger’s data started and the time that it ended. It doesn’t allow for the possibility that the higher costs of the services from that industry due to higher wages might alter the consumption habits of consumers over the longer term. Their study doesn’t look at those questions. It’s a small-scale, time-limited quasi experiment.

I’m not saying there’s nothing in the argument that raising the minimum wage won’t result in job loss. The evidence is mixed. David Neumark, labor economist now at University of California, Irvine, has done interesting research in this area with Salas and Wascher that suggests real trade-offs between higher wages and job losses.

Anyway, I’m getting exercised because what I’m trying to say is that economists have an insight here. This is not blind ideology. Some people, like Paul Krugman in the pages of the New York Times, are quick to dismiss the concern I am expressing here as some kind of Milton Friedman-like conservatism. But it’s not. And I think Krugman does the profession a disservice to the extent he is just repeating a party line.

Again, economists have insight on this. By raising the minimum wage, you’re pricing people with few or no skills out of the labor market. I’m talking about people whose labor isn’t worth $10.10 an hour. Such people exist. If nobody wants to pay them $10.10 an hour, but somebody wants to pay them $8.30 an hour, then by raising the minimum wage, you have just priced them out of the labor market.

Now, maybe the net effect of having done so raises social welfare. That’s a separate question. But to observe that this policy is going to be awfully hard on the person whose labor is only worth $8 an hour is not to make an ideological argument. Rather, it is to simply make a statement of fact.

CRIME, INCARCERATION AND INEQUALITY

Region: I’m curious about your work on crime, prisons and racial inequality. You’re a public intellectual. You’ve done a lot of scholarship that’s not strictly economics, and much of this has focused on incarceration and inequality.

Loury: Recently, that’s true.

Region: Perhaps I’m ignorant about this, but you don’t seem to approach it as an economist. I’ve wondered if there’s a specific reason for that. Obviously, Becker and many others have researched the economics of crime. Why not you?

Loury: The stuff I’ve been writing about crime—which you are quite right to say is not economics—in fact, I’ve been mildly chastised from time to time by some colleagues in economics when I send around this or that paper, and they’ll say, “Why don’t we get some data and try to look at this?”

But this work comes out of a very specific origin. I would have never been writing about crime if I hadn’t been invited to give the Tanner Lectures on Human Values at Stanford, which I delivered in 2007.

It was a great honor to have been invited. I had somewhat of a reputation among some political theorists and philosophers because I had been doing this public intellectual work, and I had been writing the occasional essay about reparations for slavery, for example. Different things like that.

And when I got to Boston University in 1991 and became a university professor there a few years later, I could teach outside the economics department. I was invited to do a lot of different collaborative, interdisciplinary teaching.

I started teaching with a philosopher and a political theorist—courses on Adam Smith, Karl Marx, John Maynard Keynes, Joseph Schumpeter. You know, reading classics in political economy. So I was known to some of these people. And when the committee out of Stanford was deciding who they were going to have to be the Tanner lecturer, they asked me.

The Tanner Lectures, created by a foundation in Utah, are very distinguished lectures in political philosophy. They have them at Princeton, the University of Michigan and other major universities. So it was a real honor to have been invited. When I got the invitation, I said “Oh, well, that’s cool” because there were only a few other economists I could find at that time who had done the Tanner Lectures. My dissertation adviser, Robert Solow, was one. Jeffrey Sachs, Amartya Sen and Thomas Schelling were others. I felt this was pretty good company to be in, so I took the assignment very seriously and wanted to challenge myself by choosing my topic judiciously.

My first instinct was, OK, I’ll go to my corpus of work on affirmative action, and I’ll find a way of saying something that philosophers could appreciate out of that. But the more I thought about it, the less I liked that idea. I mean, it was too easy for me [laughs].

And I had been disturbed in a general way about the rising numbers of Americans in jail, particularly African Americans. I’d been teaching undergraduates about race and inequality. I would teach some ethnography and urban sociology, and the crime issue and rising prison numbers would always come up.

So, I decided I was going to give lectures on “Race, Incarceration and American Values.” That’s the title that we came up with. I had two lectures to give and it was a wonderful experience, just a wonderful, wonderful experience. A great triumph, and well—I got fired up while preparing for it.

Preparing took nine months to do. I instituted a course at Brown to help me get ready, which I taught in the fall of 2006; I gave these lectures in the spring of 2007. So, my course on punishment had a ton of books from a variety of scholars. Like Michel Foucault, you know? “Discipline and Punish”—man, it makes your head hurt trying to read that stuff.

It was mostly sociology and criminology because, what else would it be? Not that there isn’t any economics in it, but I wasn’t going to limit myself to the economics literature to be able to engage in this question for what were basically lectures in ethics.

I read and thought broadly, and I composed an argument that was not an economics argument. It was basically a kind of Rawlesian justice argument, at the end of the day—applied, very applied.

So, that’s why I approached crime from that particular point of view.

But I also want to say that others are doing very good economics in this area. There is a program at the NBER on the economics of crime. I know these people. These are very good economists: Phil Cook, Justin McCrary, Jens Ludwig. There are interesting lines of investigation, studies that people are doing with applied micro approaches. Steven Levitt has made his career, in part, writing articles about this area.

So, I would not disparage this line of work at all, not at all. However, sometimes I think these economic issues and effects are of second-order magnitude, relative to the first-order issues, which are basically value questions.

Who are we as a people? And what are we going to do with this conundrum, that we’ve got these undeveloped individuals in our midst, in our cities? I mean, these people who are bad actors. Some of the discussion about this issue is just very discouraging because people are not facing up to the facts. I’m talking about discussion on the left, OK? I’m talking about people who are against prisons, who think—as I do—that we are over-incarcerated.

But crime is a real issue, and there are empirical questions. Does the death penalty deter people from committing murder? That’s a classical question of inference from whatever the data might be. It’s a microeconomic and sort of analytic, quantitative sociology, criminology-type question. It’s a question that experts need to answer, and it’s subtle. I want to hear what Dan McFadden or Charles Manski has to say, because they’re among the people who are going to know how to best judge what these robust statistical models will tell us. So, those are technical questions.

But, there are just other things that are going on. I mean, how long should these sentences be? Should we allow felons to vote? Does it make sense to disqualify them from housing subsidy and educational subsidy programs, Pell grants and things like that after they get out? What do you do with a juvenile? Life without the possibility of parole as a sentence to a juvenile offender, is that something that you actually want to do? What about solitary confinement? “Administrative segregation,” I think that’s the sanitized term.

I didn’t think economics by itself reached broadly enough or deeply enough to allow me to cover the terrain that I wanted to cover, which is, What has happened to my country here? How did we get to be a nation of jailers?

“ROSCAS”

Region: With Timothy Besley and Stephen Coate, you’ve written a couple of papers on “roscas”—rotating savings and credit associations—an informal financial intermediation institution that seems ubiquitous.

They’ve developed in areas as varied as Indonesia, Sri Lanka, Korea, Ethiopia, west Africa and Bolivia, and among immigrant groups in the United States. As you note in the paper, even cooperative barn-raising can be viewed as a type of rosca.

You three were probably the first economists to formally analyze their role and efficiency relative to other financial intermediaries.

Could you briefly explain what they are, what function they serve in these many countries and why—though they’re not efficient in general, and their financial role could be better performed by credit markets—roscas (or at least one variant of them) can still provide some surplus value relative to credit markets?

Loury: Roscas vary in the details of their operation, but basically, they involve a group of people agreeing to make periodic monetary contributions into a fund, the proceeds from which are then made available, in turn, to each member. To illustrate, a dozen people might meet monthly over the course of a year. At each meeting, everyone would contribute $100, thereby creating a fund of $1,200 that one of them would be able to use for whatever purpose. By year’s end, every member would have been in the position of receiving the joint contributions for one of the months. Of course, early recipients are obliged to continue attending meetings and making contributions, though they could expect to receive no further benefit from the association in that year.

Roscas serve the function of aggregating the collective savings of a group of people and putting that money to immediate use. In the absence of formal banking institutions and without access to roscas, people would be saving in isolation from one another. As a result, they would not be able to enjoy the benefits of financial intermediation that those of us with access to formal credit markets take for granted.

Thus, in the example I just gave, were those individuals unable to cooperate through some kind of rosca, it would take 12 months for anyone to have access to a $1,200 fund, which is worse for all but one of them. So, as a practical matter, these institutions are a principal means by which people raise the money they need to finance personal and business expenses, when they lack access to conventional bank loans.

In our original paper on this subject, Besley, Coate and I study two cases—what we call random roscas and bidding roscas. In a random rosca, those who have yet to receive the monthly contribution pool draw lots to determine who will be the next recipient. In a bidding rosca, priority of access to the collective savings is auctioned off at the outset—with early recipients agreeing to make larger monthly contributions than later recipients.

As far as the efficiency issue goes, a formal credit market would perform the intermediation function more effectively than an informal rosca if one could overcome the moral hazard and adverse selection problems that arise because of the inherent asymmetry of information between borrowers (who know their credit worthiness and intentions to repay) and lenders (who can only guess about such things).

But, these informational problems are often insurmountable in the developing country context. So, in the absence of access to a formal credit market, roscas are a “second-best” response that constitutes an improvement over autarky (where individuals save on their own in isolation from one another).

Random roscas are simpler but less flexible than bidding roscas, because they are less able to accommodate differences in the preferences and unanticipated needs of individual participants.

PIGOUVIAN TAXES AND EXTERNALITIES

Region: In two papers with Dennis Carlton, you analyze the classic solution to negative externalities, proposed by Pigou in 1920: a constant per unit tax on output.

Loury: Man, those papers were written more than 30 years ago. It’s nice to know that somebody still cares!

Region: And why not? It’s fascinating work and still relevant. You showed that Pigouvian taxes alone can’t achieve long-run efficiency—that they have to be supplemented with a lump-sum subsidy or tax.

Could you explain why the two in conjunction—a per unit tax and a lump sum tax (or subsidy)—can achieve efficiency while the per unit tax is insufficient?

Loury: Gladly! Carlton and I began from the observation that the external diseconomies (like pollution) associated with the manufacture of some product might depend in a nonadditive way on the aggregate level an industry’s output and on the distribution of that production among the several firms that constitute the industry. In other words, the external costs of 1,000 units of production might be lower, or higher, if all that output is produced by one establishment rather than if 10 separate establishments produce 100 units each.

We modeled this by supposing that if n firms each produce q units of output, so that industry production was Q = n*q, then the external costs for that industry would be some function of both n and q [i.e., damages = D(n,q)] rather than being a function of Q alone [i.e., damages = D(Q)], as had typically been assumed in the literature on Pigouvian taxation).

We went on to notice that, if one adopted our somewhat more general specification, then a characterization of efficiency would need to attend to both the marginal cost and the average cost of production at the affected firms. That is, both marginal and average cost would need separate modifications to achieve efficiency, and a single per unit tax on a firm’s output would generally not be sufficient to achieve the necessary adjustments on both the intensive (per firm output) and extensive (entry or exit of individual firms) margins.

Thus, the generalized external damage cost function that we were proposing implied the need for both a per unit tax on output and a lump sum entry tax at the firm level in order to achieve efficiency.

Region: What are your general thoughts about current carbon tax proposals? Rep. Henry Waxman and three other congressmen, for example, have drafted a related bill.

Loury: I’m no expert on the economics of environmental regulation. But like most economists, I generally believe that the most effective way to reduce greenhouse gas emissions is by raising the price of carbon across the board and then allowing the resulting change in incentives to induce adjustments by all kinds of firms in many different industries, as well as by the customers of those firms.

Our insight in those old papers—emphasizing the distinction between tax effects at the intensive and extensive margins of an industry—is not a first-order consideration for the formulation of national policy on greenhouse emissions control, in my humble opinion.

PROGRESS IN ECONOMIC THEORY?

Region: My next question is about progress in economics. In your 1977 article that we discussed earlier, “Dynamic Theory of Racial Income Differences,” you wrote that it might “be useful to employ a concept of ‘social capital’ [because it forces] the analyst to consider the extent to which individual earnings are accounted for by social forces. ... However, for precisely this reason such analysis is unlikely to develop within the confines of traditional neoclassical theory.”

You wrote that nearly 40 years ago. Time has passed. Was your pessimism warranted? Has economic theory made any progress in this direction?

Loury: Oh, sure. You have to understand, I wrote that line in 1975 or early ’76. The ’60s were over, but not long over—the ’60s effectively extended into the ’70s in many ways. The Vietnam War, for instance, was winding down, with all that entailed.

But I had grown up in Chicago. I was then a black kid at MIT in the graduate program, and the question, “How are you relevant to being a part of the solution to the struggle for our people?” was being asked. This kind of thing. “Can you be a black and an economist at the same time?” “Have you sold out?”

So there was a lot of faux radicalism in the air, a lot of posing, a lot of, “Yeah, I’m going to get an education and I’m going to become a scientist, but I’m going to be a critical scientist. I’m going to stand a little bit at a remove from the system. I’m not going to just buy it all, not going to drink the Kool-Aid.” That kind of thing.

And in economics, there was neoclassical versus radical economics. Something called the Union of Radical Political Economists really existed. I guess it probably still exists.
Region: Sam Bowles, your co-author now, was prominent in URPE back then.

Loury: Yes, Sam Bowles and Herb Gintis—their book had a big impact on me, Schooling in Capitalist America. I read that book cover to cover. And then there was the controversy of “Bowles and Gintis, Marxists at Harvard, were they going to get tenure?” All that good stuff. And, you know, I wasn’t a radical economist. I was a mainstream economist, but...

So those lines in the article were kind of a pose to say, “I’m in the neoclassical camp, but I’m not of it.” I didn’t really have any critique of neoclassical economics, as such, and, really, that’s an ad hominem comment I’m making there.

But more substantively, you asked me, Have economists done better in this area? Sure. I mean, theory is now a major topic—Stanford’s Matt Jackson, people like that. These guys will get Nobel Prizes one of these days for working out the implications of people’s social connectivities.

On the empirical side, there’s Northwestern’s Charles Manski’s “reflection problem.” This is the idea that if I’m trying to measure the impact of a peer group on an individual, I can’t look at variance across data on individuals’ observations on their peers, because individuals are choosing their peers. That’s a kind of intrinsic endogeneity that creates a very difficult identification problem for inferring the causal effect of association on outcome. People have worked on that. So I’m just saying, there have been advances and advances. There’s much better data.

And a lot of the social economics work is being done not just in the United States, but around the world. The Poverty Action Lab, for example, does its random clinical trials in international settings and around the questions that have social-capital-like themes embedded in them.

So, I don’t think that I would be at all dismissive of the profession in terms of taking seriously the kind of social effects that I was interested in, in those days. But, yes, that line was a little gratuitous.
But, then, the other thing is about markets: “Will markets solve all the problems? Is laissez-faire sufficient?”

Region: And you’re not in that camp.

Loury: No, I’m certainly not saying laissez-faire is OK. Laissez-faire has its issues.

Region: Well, clearly then, that leaves much to discuss. But perhaps we should end on that note. Thank you so much—it’s been a great pleasure.

Endnotes