The Yale Law Journal

May 2015

The Banality of Racial Inequality

Reproducing Racism: How Everyday Choices Lock in White Advantage


author. Charles Keller Beekman Professor of Law at Columbia Law School and Professor (Adjunct) of Law at Yale Law School. I thank Jamal Greene, Paul McGuire and Carol Rose for their insights and assistance. Thanks also to Elizabeth Chao and the other members of the editorial team at the Yale Law Journal for their careful edits and thoughtful comments.


Consider a counterfactual America where there is race but no racism. An impossibility, even conceptually, cynics and skeptics will quite reasonably argue. Nonetheless, suspend your disbelief. Imagine everyone were given a pill that rendered them and their offspring completely and permanently blind to race and all its amorphous indicia. As a side effect they also experienced amnesia about their prior racial identities and those of others. Logs of racial determination in birth certificates and other official and unofficial records were swiped clean through a massive government program. Race remains as an idea, a concept, in the post-pill world, but no one can observe or remember it in others or themselves.1

Daria Roithmayr’s argument in Reproducing Racism,2 taken in its starkest terms, maintains that, without further state intervention, those people who would have been perceived as black before the pill would, well into the future and perhaps indefinitely, experience the same social and economic disadvantages they have faced for generations.3 Absent additional state action, the fate of black Americans would remain locked in a pattern established long before their birth and based on reasons no one currently recognizes or endorses.4

A stark claim, no doubt, but nobody should be surprised by the suggestion that the children of slaves and their descendants bear inequalities carried forward from a more racist past. Libraries have been written on the various ways in which these inequalities are transferred across generations.5 What distinguishes Roithmayr’s contribution to the inequality library is its indictment of the mundane, ostensibly race-neutral practices we all take for granted. Racial inequality will continue even in the post-pill world described above, but not through overt animus or more subtle forms of implicit or institutional discrimination. Rather, Roithmayr argues, our inequality is perpetuated through a set of seemingly innocuous, if not laudable, choices people take pride in making, such as referring a friend to a job or helping a child pay for college or a down payment on a home.6

Friends and families, left to their own devices, will always stand in the way of any real prospects for equality of opportunity. No one likes to think of loved ones in a negative light, but a moment’s reflection reveals this undeniable truth—both as a practical and a theoretical matter. “[A]s long as some form of the family exists,” John Rawls wrote, the project of “fair opportunity can be only imperfectly carried out.”7 George Bernard Shaw was even more doubtful about the prospects of equal opportunity. In a 1913 address to the Liberal Club of London, he taunted his host, saying, “[y]ou, Mr. Chairman, have spoken of equality of opportunity. The difficulty about that is that it is entirely and completely and eternally impossible.”8 To Shaw, inequality was so inevitable and so profound that even a pill that made everyone forget family and friends could not overcome his pessimism concerning the likelihood of equality of opportunity.9

Roithmayr is no pessimist. She sets her gaze on the banalities of inequality—the everyday mundane determinations that reproduce social and economic differences along racial lines—and seeks a remedy.10 Her remedial preference does not call for busting up families, but rather undoing the effects of other combinations, so-called “racial cartels,” that once dominated the political and economic order of the country. Racial cartels may be past their heyday, but, Roithmayr argues, their lingering effects continue to privilege white Americans while disadvantaging certain racial minorities. Her argument proceeds in parts. The first is historical. Roithmayr establishes that historical decisions based on race gave whites an early advantage.11 Of course, every competition has winners and losers. That’s life. But, Roithmayr presses, whites won their early advantage through morally and legally indefensible conduct, acting as cartels to exclude other racial groups from fair competition for desirable resources like jobs,12 education,13 housing,14 and wealth.15 Moreover, the argument continues, after unfairly acquiring market power, white Americans instituted racist practices that reproduced their ill-gotten advantage.16 Racist behavior, however, is not the central point of Roithmayr’s argument: “This book is about why racial inequality persists,”17 even if there is no ostensible racist behavior.

After describing the historical backdrop of unfair play through which whites gained an early advantage, Roithmayr then considers in the second part of her argument how the initial leg-up has reproduced itself over time across several domains (wealth, education, social networks, and housing) through family and other feedback loops.18 The third part of the argument turns to a more theoretical discussion of her lock-in model and develops the claim that the model provides new insight into America’s persistent racial inequality.19 The fourth and final part focuses on the responses and remedies that might arrest and counter the unmeritorious, essentially automated and mundane patterns of distributing advantage.20

I. models and metaphors

Before turning to the details of the book’s argument, a word of clarification about the title is warranted. Roithmayr is largely concerned with everyday choices that reproduce inequality. The inequality that these decisions perpetuate is correlated with race, and importantly so,21 but racist ideologies are not themselves motivating those decisions. She does not assert that racism is a thing of the past.22 Far from it. Roithmayr’s argument, rather, is that the racism of the past is no longer required, even if it is still present, to maintain the differences it initially brought about.23 Moreover, her argument allows that racist ideologies may be maintained or reproduced indirectly through these everyday choices,24 but her primary targets of concern are path-dependent mechanisms that reproduce inequality.25

To make the idea of path-dependent racial inequality broadly accessible, Roithmayr relies on a number of metaphors throughout the book. From Polya urns26 to QWERTY keyboards,27 she draws liberally and frequently on various models, examples, and analogies. Her master metaphor, however, is monopoly, or, more precisely, unfairly acquired, self-perpetuating market power.28 Unfair competition, she urges, is the basic source of persistent racial advantages, the justification for intervention, and the best means for appreciating the remedies that ought to follow.29 “In the context of monopoly,” Roithmayr concludes on the last page, observers “understand quite quickly the need for significant government intervention.”30 Before arriving at that conclusion, however, she devotes a significant portion of the book to advancing a particular anti-competitive narrative—the lock-in account of racial inequality—that she has over a number of years developed in law review articles.31

Monopoly wrongly acquired and then ossified through network externalities is the core of Roithmayr’s racial lock-in argument. Microsoft’s antitrust ligation from the 1990s provides a key illustration.32 “According to the allegations, Microsoft engaged in a range of very bad (and illegal) behavior,” Roithmayr writes with an intentional colloquialism that runs throughout the book.33 “Microsoft’s bad behavior,”34 combined with the basic network structure of the software industry,35 “went on to trigger a ‘positive feedback loop’ in the operating systems market.”36 Thereafter Microsoft’s early unfair “advantage snowballed” and eventually “became locked in.”37 Today, Roithmayr argues, “white economic advantage has become institutionally locked in, in much the same way as Microsoft’s monopoly advantage did.”38

Microsoft’s anti-competitive behavior is not, however, an ideal fit for Roithmayr’s argument. Strictly speaking, Microsoft was a single, hierarchically organized entity, not a horizontal cartel of multiple actors, as Roithmayr characterizes “whites [who] formed racial cartels during slavery and Jim Crow to gain monopoly access to key markets.”39 Yet, cartels do organize to exert market power like a monopolist,40 so the analogy, while not perfect, provides some insight. This close-but-not-quite-exact quality holds for many of the models, metaphors, and analogies that Roithmayr uses. A large part of the reason for this misalignment is that Roithmayr is attempting something novel and distinctive for which there are no canned applications. Still, Roithmayr presents a compelling array of supporting arguments and analogies to make her points effectively for the lay reader.

To preserve scope for broad analogical connections, Roithmayr at times glances past legal details and definitions. For example, although she draws at times upon actual cartel conduct, she spends little time presenting and parsing legal arguments that could have been used to generate liability for this anticompetitive behavior, historically or in the recent past.41 Monopolization is more metaphor than legal fact here. Similarly, while much is made of the state’s role in facilitating residential racial discrimination,42 there is surprisingly little discussion of the state action doctrine.43 No doubt the publisher’s interest in not sacrificing the general audience to satisfy the legal one motivated some of these oversights. Let the law geeks read law reviews. Yet as noted next, even general readers might have appreciated clearer definitions of “racial cartels” and “white cartels,” which are central constructs of the lock-in argument.

II. abstract and actual cartels

What, exactly, is a racial cartel? The question is raised,44 but never fully answered. “Economists,” Roithmayr tells us, “typically define a cartel as a group of actors who work together to extract monopoly profits by manipulating price and limiting competition.”45 But what makes a cartel racial? A shared racial identity among cartel members would certainly be too inclusive a criterion for the kind of argument she is making. Otherwise, OPEC, Standard Oil, the De Beers diamond mongers and most other known cartels might be considered racial cartels. Perhaps the key is not the membership, but the identity of those excluded. But then every private club excluding minorities would be a racial cartel, which may be what Roithmayr has in mind but nevertheless seems too broad. All cartels are exclusive clubs, but surely not all exclusive clubs are cartels.

Moreover, when Roithmayr writes of “white cartels,” it is often unclear how far this white economic conspiracy extends.46 Sometimes “white cartels” seem to implicate white Americans generally,47 and at other times just those of a certain generation, or of a particular, though still abstract, place (such as the Jim Crow South48 or northern communities with racial covenants49). When she writes of “the white organizations of Jim Crow[,] white unions, political parties, school districts, and associations among them,”50 one has a clear sense of what made them white organizations (their “whites only” signs were superfluous), but it is still not obvious what made them racial cartels.

The puzzle of “racial cartels,” as used in the book, is resolved by viewing it as a framing device rather than in terms of strict legal or economic criteria. “To be sure,” Roithmayr has written elsewhere, “the analogy between racial cartels and ordinary market cartels is far from perfect. . . . More specifically, the [racial] cartel account does not conform to the precise technical definitions of cartel conduct.” Yet viewed metaphorically, argues Roithmayr, “the analogy captures far better than standard accounts the competitive advantage and collective dynamic of racial exclusion.” Of course she is right. As metaphor, “racial cartels” reveal important features of our racial history too often obscured by conventional accounts. But in this history there are also actual cartels, more than metaphors, engaged in racialized restraint of trade. Following the Civil War, for instance, plantation owners explicitly conspired to restrict labor opportunities of recently freed slaves—“You won’t hire my niggers, and I won’t hire yours”51—and these agreements, often supported by the state, continued well into the twentieth century.52 Roithmayr herself provides other compelling examples of actual cartel conduct aimed at excluding competition from racial minorities.53 The book’s argument is sharpest and most convincing in these specific cases. But in the back and forth between analogy and actual conduct, it is easy to lose sight of the level of argument suggested at any particular point in the book’s use of racial cartels.

III. racial covenants and white primaries: the central accounts

Roithmayr argues that throughout the twentieth century, a number of “private business organizations functioned together in ways that resembled racial cartel activity.”54 Among these private actors, Roithmayr describes the tactics of railway workers, who organized to exclude blacks;55 “agricultural growers” who “collaborated to segregate Mexican laborers”; crafts unions, citizens’ councils, and even parent-teacher associations, all engaging in “collective action to exclude nonwhite groups from key education, labor, and political markets.”56 These cases are briefly mentioned, more as general sketches, rather than as concrete examples of racial cartels. For a more sustained discussion, Roithmayr looks at voting in Texas57 and housing in Chicago.58

Finding racially discriminatory market manipulation in housing is, sadly, not a difficult task, even today.59 Locating discrimination in historical records is even more easily accomplished, thanks to a great corpus of research developed by scholars of race and housing over the past century.60 Local and national real estate boards surely operated as cartels61—setting policies, coordinating members, and disciplining defectors—and Roithmayr effectively describes their market-rigging practices.62

Roithmayr takes homeowners’ associations to be “the poster children for racial cartels.”63 Chicago’s homeowners’ associations were “a model in efficient racial exclusion” she writes, “[t]ightly coordinated, well run, and legally armed with the restrictive covenant.”64 There are, however, reasons to question whether these organizations were as efficient and effective as Roithmayr suggests. First, while some Chicago neighborhood associations were better organized than others, collective action problems plagued them all. The more successful ones had to rely on the energy of a few motivated or entrepreneurial individuals, along with significant institutional support, such as the sort given by the University of Chicago to the Woodlawn Property Owners’ Association,65 which Roithmayr discusses at length. Second, neighborhood association-driven racially restrictive covenants in Chicago and other cities were often sloppy affairs replete with legal defects.66 The historian Arnold Hirsch argued that it was for this reason that covenants had little impact in maintaining traditional racial barriers in Chicago.67 Hirsch no doubt understated the impact of covenants with this logic, since much of their effect was social, rather than legal, and didn’t depend on court enforcement. Still, the covenants imposed by neighborhood associations were often less effective than the members sought.

Efficiency and efficacy in the use of covenants were better realized by developers than neighborhood associations. Roithmayr overlooks this aspect of the covenant account, failing to address the role of racially restrictive deeds installed by developers in planned communities as distinct from the neighborhood-initiated restrictive agreements organized by existing communities.68 Developer deed restrictions, which existed in both middle class tracts, like the Levittowns developed by William Levitt, and more upscale developments, such as those developed by the W.C. and A.N. Miller Company in the Washington, D.C. area, were able to cost-effectively subject whole communities to racial restrictions in one fell swoop, and these racially restrictive devices tended to avoid the legal defects common to neighborhood agreements.69

Class was also a significant factor in neighborhood agreements. Given the costs of establishing neighbor-driven racially restrictive covenants, working class neighborhoods, like Chicago’s Back-of-the-Yards district, generally did not bother with them.70 Racial covenants operated in wealthier neighborhoods,71 where white homeowners could afford the expense of covenant drives and the costs of their enforcement.72 By the same token, minority purchasers drawn to covenanted neighborhoods were generally themselves middle and upper-class buyers—buyers with enough cash or access to loans to purchase homes in covenanted neighborhoods notwithstanding the barriers erected by the Fair Housing Administration’s (FHA) discriminatory policies.73

Racial covenants in Chicago, and in the rest of the country, had a more complicated existence than is suggested in Roithmayr’s discussion. They were not simply established legal tools used by racist cartels to achieve their limited and foul objectives. Racial covenants were both more corrosive (generating, promoting, disseminating, and legitimating racist ideologies through housing policy and practice) and more constitutive of communities, including integrated ones, than is suggested by Roithmayr’s analysis. For instance, racially restrictive covenants may have played a stabilizing role in some communities that were already somewhat racially integrated by preventing the neighborhood from “tipping” and thereby slowing or preventing white flight.74

Texas’s all-white primary may have been a more straightforwardly exclusionary cartel tool. Roithmayr presents an abbreviated, but still illuminating, account of the collective efforts of Texas Democrats and Fort Bend County Jaybirds to disenfranchise black voters.75 These accounts of housing and voting cartels are designed to get her readers to see racial inequality in a new light. “Describing the homeowners’ association and the Fort Bend County Jaybirds as racial cartels,” she writes, “serves to highlight several aspects of racial exclusion that conventional theory obscures.”76 She argues that key insights are gained when we adopt this new perspective: “First, a cartel story emphasizes that whites benefited from racism.”77 Second, the “cartel story also focuses on the way that whites worked together collectively to exclude, and the techniques they used to keep each other in line.”78 Third, “a cartel story focuses on the effect of all this exclusion on fair competition.”79

But do we really need a “cartel story” to know that some whites benefited by excluding blacks and other racial minorities, that there was collusion among vested interests to achieve this exclusion, and that the conduct and outcomes were unfair? Scarcity and economic competition—familiar frames in the literature on American race relations—suggest all these points. Nonetheless, the cartel account may be the best way to tie these factors together in an accessible manner, and that’s a valuable contribution to the literature. Moreover, and perhaps most importantly, as Roithmayr suggests, the cartel story “can justify government moves to dismantle the effects of such cartel behavior as a kind of antitrust intervention.”80 This too is a valuable contribution, but to make it more than just a suggestion, the book would have had to engage the legal arguments and details that were perhaps sacrificed to make the book accessible to the lay reader.81

IV. wealth, networks, and feedback loops

According to Roithmayr’s lock-in account, unfair practices led to racial inequalities, and then networked feedback loops kept the differences going.82 Social networks are the engine behind the distribution of resources, Roithmayr writes. “[M]ost of us derive our well-being from our networks . . . . [W]e rely on four basic types of social ‘network’ arrangements—(1) families, (2) friends and colleagues, (3) neighborhoods, and (4) workplaces and other kinds of market-based institutions.”83 These networks create feedback loops, which are the key to the lock-in model of racial inequality. “Through positive feedback loops, inequality reproduces itself from decade to decade, automatically and in the absence of intentional discrimination.”84 Roithmayr describes the operation of networks and feedback loops in generating and maintaining inequalities in wealth, institutional access, professional outlooks, and housing opportunities.85 Let’s begin with wealth.

A. The Racial Wealth Gap

What explains the persistent racial wealth gap? Researchers have sought to identify the sources of the wealth gap by focusing on earnings and education, cultural practices, and, finally, family background and networks. Roithmayr favors this last account. Empiricists have had widely varying degrees of success in explaining wealth differences by race, identifying as little as five percent to more than one hundred and twenty percent of the gap.86 The variation is largely due to the multiplicity of identification strategies and data employed.87 Nonetheless, some consistent patterns are observed across studies.88 Most studies find income or earnings to be the single best predictor of the wealth gapexplaining, for example, between twelve and seventy-two percent of the black-white wealth difference.89 Neither these differences in earnings, nor differences in education, however, can fully account for the significant racial differences in wealth and wealth accumulation.90

Some commentators have claimed that cultural differences account for some of the racial wealth gap. Implicit in this claim are often assertions of dysfunctional behaviors among blacks and some other racial minorities that undermine their ability to save and invest.91 Yet the evidence of racially or culturally contingent savings and investment behaviors is inconsistent and in any event too insignificant to account for the generally observed racial wealth gap.92 While a number of early studies suggested that blacks actually saved at higher rates than whites, current analyses tend to find little or no differences in racial savings rates. Maury Gittleman of the Bureau of Labor Statistics and his collaborator, Edward Wolff, however, did report a higher savings rate among whites.93Yet with income taken into account, this difference in the savings rate explained only one percent of the wealth gap.94 Differences in investments, if not in savings, might better explain the wealth gap. There is some evidence that whites earn a higher return on their investments than blacks and that blacks are more likely than whites to hold liquid (and other lower-return) assets.95 Here again, however, the evidence is mixed and seems largely mediated by income. As blacks get richer, their asset portfolios and rates of return tend to look a lot like those of whites in the same income brackets.96

But income doesn’t smooth away all observed racial differences in asset portfolio and returns. To fill in the gap, researchers have turned their attention to the effect of family transfers, including gifts and inheritances, as the likely source of wealth differences between blacks and whites.97 Some regression estimates and simulations suggest that family financial transfers may account for up to a quarter of the wealth gap. One might predict that gifts and inheritances would have a large effect on the wealth of advantaged households. Less well-off families, black or white, simply cannot pass on such benefits to their children.98 Family financial transfers, if they play a meaningful role in generating the wealth gap, might be expected to have a larger impact among the advantaged.

There is strong consensus that the family resource advantage of whites plays a large role in creating and maintaining the wealth gap, but there is no agreement about the exact mechanism.99 No doubt this sounds entirely obvious. The children of wealthy parents eventually grow up to become wealthy adults. But the claim is not as straightforward as it might appear. It’s not clear how parental wealth affects the wealth of children and whether it operates differently for blacks and other racial or ethnic minorities as compared to whites. Do wealthy parents contribute to their adult children’s wealth by investing more in their early and continuing education? And might racial differences in the expected returns to education lead to different educational investments by comparable black and white parents? Or do wealthy parents simply enrich their children through direct financial contributions, where for historical reasons, whites have a lot more money to give?

B. Family Feedback Loops

As an empirical matter, the questions above are far from resolved. To the list of possible mechanisms explaining the racial wealth gap, Roithmayr offers a novel rhetorical account: “Early anticompetitive conduct garnered for whites additional wealth, acquired on the backs of slaves or from victims of white cartel conduct in unions, homeowners’ associations and political parties.”100 Thereafter, family feedback loops simply perpetuate a state of affairs wherein “white families continue to pass down the ill-gotten wealth they acquired during Jim Crow and slavery. Families of color have no such racism dividend to pass down.”101 These dividends are distributed not only at death, but also, and perhaps primarily, inter vivos. Roithmayr calls particular attention to two kinds of inter vivos transfers: “First, white parents help their children to make a down payment on a house far more often than do nonwhite parents. Second, white parents help with college tuition, which boosts a child’s future earnings and wealth as well.”102 These are the essential family feedback loops, she argues, that preserve racial differences over time.

There is empirical support for Roithmayr’s argument. Industrial and labor economist Francine Blau and financial economist John Graham were among the first to suggest that inter vivos parental transfers and inheritances from older generation to younger generation whites might explain the black-white gap in subsequent generations.103 Sociologists Melvin Oliver and Thomas Shapiro brought popular attention to this suggestion in Black Wealth/White Wealth, showing that the wealth of blacks tends to be less sensitive to income and other demographic characteristics than the wealth of whites.104 Oliver and Shapiro offer a convincing descriptive account of how the legacy of racial discrimination in the United States might perpetuate such wealth differences through families across generations.105 Their account has been supported, as well as complicated, by a number of more formal statistical analyses conducted by economists and sociologists (including Oliver and Shapiro themselves).106

In addition, Dalton Conley has argued that parental wealth is more important than one’s own income for predicting one’s wealth.107 Taking into account parental wealth, along with several other demographic characteristics, Conley is able to explain fully the average differences in wealth between blacks and whites.108 But the work of some economists points in another direction. Employing a clever approach to control for family background characteristics, Joseph Altonji and Ulrich Doraszelski developed an analysis of wealth accumulation focusing on a sample of siblings.109 If parental wealth is important, Altonji and Doraszelski hypothesize, then their analysis should look very different when it includes a variable indicating that various individuals in their sample come from the same family (a family fixed effects approach) compared to when no such account is made (the standard approach).110 Finding little difference between the fixed effects and standard approach, they reject the claim that parental wealth explains the significant observed differences in the rates at which blacks and whites accumulate wealth.111 Yet they do not deny that family background characteristics might affect other aspects of the wealth function.112 In other words, even if parental wealth doesn’t help whites get richer faster than blacks, it might still start them off with a lot more than blacks; in terms of a (wealth regression) line, whites may have a higher intercept, if not a steeper slope.

C. Education and Mental Models

Roithmayr’s analysis of capital does not focus on only cash, stocks, bonds, and real property. She recognizes that social capital may be as important, if not more important, than financial capital in generating and maintaining wealth.113 Consider, for instance, parental spending on elite education for their children. These expenditures certainly promote their children’s educational and occupational attainment, which is likely to translate into wealth down the road. But there is more to it than that, as David Wilkins observes: “students attending elite schools are also socialized into habits and possibilities of eliteness and granted . . . membership in the elite networks.”114 Moreover, economists Kerwin Charles and Erik Hurst present evidence showing that parents’ passing on their attitudes about wealth (that is, saving and investment preferences) was a significant predictor of a child’s later wealth.115

In line with these empirical studies, attitude formation and transmission, particularly in educational settings, are central parts of Roithmayr’s thesis, which she places under the heading of institutional rules. “Institutional rules are the rules of the game that govern the day-to-day operation of an institution.”116 These rules influence not only what we do, but also how we think about and justify our choices and those of others.117 “Perhaps most importantly,” she observes, “institutional rules play an important role in shaping the mental models that people use to interpret the environment.”118 As Roithmayr writes, “these mental models and their accompanying institutional rules can get stuck in a particular niche, unable to move toward potentially more efficient rules and mental models.”119 Mental models and institutional rules often interact to create a perceived neutrality or naturalness to institutional practices, making them even more secure.

Roithmayr illustrates these claims effectively by looking at law school admissions rules, particularly the early development of the Law School Admission Test (LSAT). For instance, she observes, “remarkably, law schools wanted something that correlated specifically to first year grades but not to bar passage scores.”120 The law schools’ admission committees had their reasons for preferring a test more aligned with first-year grades, but it is far from obvious that these grades are a better predictor of success in legal practice than bar exam scores. Bar passage rates and first-year grades, of course, remain highly controversial in law school affirmative action debates.121

Roithmayr also usefully recalls the historical collusion, as told by Jerold Auerbach, between the American Bar Association and the American Association of Law Schools to exclude racial and ethnic minorities, along with foreigners, from legal education and practice.122 But active and conscious collusion to exclude on the basis of race or ethnicity is not the story that she is telling in this part of the book. The argument is subtler. “At a more abstract level, law schools retain an admissions test that disproportionately excludes students of color because they think it measures something real about the merit of applicants.”123 Law school admissions officers are working with a “mental model” of merit and deservingness that evaluates students based on criteria that were questionable from the start.124 She continues, “[i]n this way, the mental models that structure our understanding of merit are tied closely to the country’s history of white privilege.”125 Her legal treatment of education here is reminiscent of Pierre Bourdieu’s Homo Academicus.126 She effectively introduces power explicitly into the discussion, rendering a narrative that is both thoughtful and revealing.

D. Employment and Housing

“To be sure,” Rotihmayr concedes, “this [mental model] narrative risks an overly static view of the world, as is true of the lock-in model more generally.”127 But this concession does not deter her from extending the framework to employment. She argues that the mental model of merit may be even more distorted when it comes to jobs. “Contrary to some narratives about merit and hard work,” says Roithmayr, “the reality is that a person’s chance of getting a good job depends quite heavily on her social network.”128 Social links reproduce racial employment inequality though formal and informal job networks.129 Network size and density is influenced by one’s racial context.130 Class, again, is a driving factor,131 as is the early history of discrimination;132 Roithmayr argues, “even if all intentional discrimination were to end tomorrow, social networks would continue to reproduce gaps in wages and employment status.”133 Racial homophily, operative for whites as well as blacks, continues to separate networks.134

Long-established institutional rules and practices continue to disfavor minorities, but, explains Roithmayr, these are not the main factors securing modern-day racial inequality. Segregation perpetuates our present inequality. “If social networks were to be integrated,” Roithmayr observes, “then the resources contained in those networks might be more evenly shared. Job referrals would be doled out more evenly. Family help for housing and college would be shared among nonwhite children as fairly as among white children.”135 With this line of argument, Roithmayr shifts the inquiry from inequality to segregation, with residential segregation, in particular, underwriting many of the differences in institutional access and wealth discussed throughout the book.

Roithmayr effectively enlists both theoretical and empirical literatures on racial housing patterns, and she confronts class issues most explicitly in the chapter focusing on how neighborhood effects reproduce racial segregation. While observing that wealth and class differences are maintained though contemporary housing and neighborhood feedback loops,136 the history of residential and racial discrimination is never far removed from her analysis. “Restrictive covenants paved the way to residential segregation, as homeowners’ associations policed against the entry of blacks and Latinos into all-white neighborhoods. Residential segregation, in turn, structured racial differences in wealth, social networks, and institutional networks of distribution.”137


Two lingering questions are suggested by the discussion of historical segregation in this context. First, why start with Jim Crow and not, for instance, slavery? Where to begin—the so-called starting gate problem—is a challenge faced by all equality theorists.138 Roithmayr is aware of the issue and some of the knotty questions it raises.139 “What would have happened if the government had made good on its promise of forty acres and a mule to American former slaves?”140 Clear answers to such questions are, of course, elusive. However, because the starting points are so central to her lock-in theory, Roithmayr might have done a little more to explain what are her relevant points of departure, for which groups, and why.

This brings us to the second lingering matter in this section: is there anything distinctive about black or Latino inequality, and, if so, how do racial covenants and the like help us to appreciate this distinction? Racial restrictions were notoriously inclusive in their list of excluded groups. While “Whites Only” or “Caucasians Only” were common in covenants, just as often there was an enumerated list of groups who were unwelcome in the neighborhood. “Armenians, Jews, Hebrews, Persians[,] and Syrians, or anyone else of the ‘Semitic race’” were identified for exclusion in some covenants, as were, of course, persons of “Negro, African, or Asiatic race,” in addition to Mongolians, Mexicans, and numerous other nationalities.141 Since all of these groups, and unnamed others, suffered from unfair cartel-like conduct facilitated through covenants, by which criterion are blacks and Latinos selected for redress? Roithmayr, no doubt, has an answer to this question, but it is not clear from the text. One is left wishing for more discussion on how she selected the starting points and groups for her analysis.142

V. theory and implications

In the first chapter of Reproducing Racism, Roithmayr quickly casts off an admittedly “[i]ncomplete and [u]nsatisfying”143 set of alternative theoretical explanations for persistent racial inequality, beginning with conventional economic accounts144 and then moving swiftly through various psychological,145 biological,146 cultural,147 and structural arguments.148 Her lock-in logic is, of course, itself a kind of structural argument, building on earlier economic models.149 To illustrate the operation of lock-in, Roithmayr makes use of several path-dependency stories.150 Readers will find some of these accounts familiar, such as the QWERTY keyboard and the VHS versus Betamax stories. Her discussion of the Polya urn experiment, however, will likely be new to many lawyers and some legal scholars, although old hat to mathematicians.151 The experiment begins with an urn containing an equal number of red and white balls.152 Next, the experimenter selects a single ball at random from the urn and returns it along with a new ball of the same color. At this point there is no longer an equal number of red and white balls in the urn. There is exactly one more ball of one color than the other. A second ball is then randomly selected from the urn and returned with another ball of the some color. Then a third ball is selected and returned with another of the same color, and so on.

The relevant feature of this experiment for Roithmayr is that while there may be large fluctuations in the percentage of red and white balls following the initial rounds of draws, “at some key threshold point in the drawing process, as the urn continues to fill,” she writes, “the percentage of reds and whites settles at a particular proportion, and remains very stable for all later draws. Amazingly, later events don’t change the final percentages.”153

As if by magic, the urn suddenly and dramatically settles down to some fixed and quite durable proportion of red and white balls[,] no matter how many more draws the experimenter completes. . . . Even more interesting, this ending proportion will vary from experiment to experiment. In one run, the urn might settle into a proportion of 22 [sic] percent red and 68 percent white. On a subsequent run, the proportion might instead end up at 78 percent red.154

Figure 1 shows five simulations of the Polya urn experiment that I ran.155

Figure 1.

basic urn (traditional polya urn experiment)


Roithmayr’s observation concerning the highly variable ending proportions is clearly apparent in the figure. Two simulations settle with less than twenty percent red balls, two above eighty percent, and one with a little more than thirty percent. Moreover, notwithstanding the high variance at the start, the tail of each distribution ends up fairly flat. In other words, “the urn’s early history matters far more than later history, because the early draws chart the path that subsequent developments will reinforce. By analogy,” writes Roithmayr, “the early history of competition among racial groups can explain contemporary outcomes.”156 Exactly how does this analogy work? “In much the same way that the early draws of the urn determined the ultimate composition of the urn,” Roithmayr tells us, “those early rounds of economic, social, and political competition among the races were rigged anticompetitively by racial cartels.”157

But that framing doesn’t quite fit. Early selection in the urn isn’t “rigged.” It’s random. Selection is determined by the relative proportions of red and white balls. Proximity, connectedness, and neighborhood effects play no role in the selection process. There is no color segregation within the urn. The balls are all always mixed up. The Polya urn experiment describes a process that locks in certain proportions based on distributions established by early randomness. At each round, each ball, regardless of its color, has fair odds of being selected when the experimenter reaches in the urn and picks one. If red balls are disfavored, it is merely because they are numerical minorities. But Roithmayr isn’t telling a story about numerical minorities. She is concerned with racial and ethnic minorities.

Non-Hispanic whites in the United States are projected to fall below the threshold of numerical majority in 2042, with continued declining relative numbers thereafter.158 As whites increasingly become numerical minorities, their chances of being selected for society’s prizes and privileges should fall—if selection is made as it is in the urn. Roithmayr’s argument, however, implies that white advantage will persist even as their numbers fall. This doesn’t mean the Polya urn is a poor analogy for her argument. For instance, perhaps it is not how balls in the urn are selected that’s telling, but rather how new balls are brought in.

Among other things, the traditional Polya urn experiment describes a process of inclusion that is entirely biased. Exclusion is not a salient frame of the Polya urn. Disfavored-color balls are not removed or prohibited from the urn. All the bias results from the method of inclusion. Recall that each ball selected from the urn is returned with a ball of the same color. We might say that the selected ball invites a same-colored partner-ball back to the urn in a manner akin to the kind of racial patronage that Roithmayr described as operating in job networks. Nothing is happening to the numerically disfavored balls in the urn—or at least that’s one way of looking at the process. This too is a feature of Rotihmayr’s narrative account, an important but subtle feature, which the Polya urn makes exceedingly clear: everybody knows that it is wrong to exclude based on color, but including those “like you” is perfectly natural and good, just like family and friends. That, at least, is one way of looking at things.159

Consider another way—that is, another process of adding balls to the urn. What if the selected ball were made to return with a ball of a different color? The result of that experiment, which I call the “anti-urn,” is shown in Figure 2.160 Again, there is significant fluctuation in the early draws, but very quickly the proportions of red and white balls settle down at near equality. An experiment is hardly needed to appreciate this result. By including balls of the non-selected color, the experimenter constantly adds countervailing probability mass to achieve a balance. Roithmayr, it is fair to say, may likely endorse this inclusion criterion. Even though it is as color conscious as the original Polya experiment, this inclusion criterion brings balance to the urn, which in turn makes selection of red and white balls equally likely.

Figure 2.

anti-urn (adding ball of other color)


An alternative method to approach balance in the urn, and therefore an alternative method to move selection toward equality, would be for the experimenter to choose at random the color of the new ball that is to be placed in the urn along with the selected ball. No doubt such an approach would lead to equality, although perhaps not with the same efficiency and speed as would be achieved by color-conscious inclusion of balls different in color from the ones selected.161 One may read into Roithmayr’s discussion that it is too late in the game for that kind of neutral selection. It is a tempting leap, to suggest, for instance, that if we are already in the settled tail of a traditional Polya urn process, it would not matter much if the experimenter suddenly started to randomly choose the color ball to partner with the selected ball that is going back to the urn.

Figure 3.

random urn (adding ball of random color)


Caution would be warranted in taking that leap. The analogy breaks down in attempting to resist random inclusion by arguing that nothing would change after the early history has set its course. The course can be altered by a random process. Consider, for example, a traditional Polya urn process where after the first one hundred draws, the experimenter begins to add balls of a randomly chosen color instead of the same color as the selected ball. The result from several runs of this experiment is shown in Figure 4. Variable proportions are kicked off by the early history of arbitrary draws, but eventually convergence toward equality is reached by including balls of randomly chosen colors.

Figure 4.

mixed urn (100 basic, then random)


What insights are gained from the traditional Polya urn experiment and its variations? Path dependency established by early advantage and followed by mechanisms of inclusion based on likeness can lock-in long-term patterns of inequality. But if the mechanisms of inclusion are changed, long-term patterns can also, in time, change. Hence, one may be optimistic, as Roithmayr ultimately is, about the possibility of change and the possibility of some equality. Mundane matters of selection and inclusion must, however, be carefully considered. Ostensibly fair contests, such as random draws from an urn or color-blind tests of formal equality—won’t necessarily lead to real equality if biased mechanisms interact with the contests.

Consider, for example, Bernard Williams’s fable of the warrior society, where after a prolonged period hereditary caste-based discrimination in favor of warrior families, the society changed to fair and formal contests for the award of privileged positions.162 Children of the old warrior families continued to secure privilege, not because the contests were rigged (for they were now fair), but because their family advantage had better prepared them for the contest.Families biased the contest. It’s not hard to understand. Who wouldn’t give their child every advantage possible?

Warrior families, in the end, are just like all other families. It is this fact that makes Williams’s fable so broadly accessible. In much the same way that Williams makes use of the warrior society, Roithmayr deploys Polya urns—along with QWERTY keyboards, VHS tapes, operating programs, racial covenants, wildcat strikes, and market cartelization, among other descriptions—to offer graspable narratives on an elusive topic. There is something for everyone to grab hold of here. And it is this diversity that makes Roithmayr’s account so broadly accessible and effective.


In 2007 Chief Justice Roberts wrote, “[t]he way to stop discrimination on the basis of race is to stop discriminating on the basis of race.”163 A provocative tautology, borrowed without attribution from the lower court,164 its purpose was entirely rhetorical. Roberts was urging that the time had come for the Court, if not the nation, to look beyond race. Race no longer mattered, he implied, except among racist integrationists and the like. Barack Obama’s presidential election the following year gave fleeting support to this view of a post-racial America. Yet by 2014, when racial discourse hit a fever pitch with protests and riots throughout the country,165 it was implausible to claim that race no longer mattered. The only question was how race mattered. “Race matters,” Justice Sotomayor wrote in that fitful year, “in part because of the long history of racial minorities being denied access to the political process. . . . Race also matters because of persistent racial inequality in society—inequality that cannot be ignored and that has produced stark socioeconomic disparities.”166These are the essential challenges advanced by Roithmayr. What is to be done? Sotomayor, again, offers an answer: “The way to stop discrimination on the basis of race is to speak openly and candidly on the subject of race, and to apply the Constitution with eyes open to the unfortunate effects of centuries of racial discrimination.”167 Readers of Reproducing Racism will be well primed for the difficult yet crucial conversation that awaits.