There Goes the Neighborhood


Review of Saving the Neighborhood: Racially Restrictive Covenants, Law, and Social Norms, by Richard Brooks and Carol Rose

Cambridge: Harvard University Press, 2013

Most law students remember Hansberry v. Lee, a famous case involving a malignantly innovative legal instrument called a racially restrictive covenant. Restrictive covenants were ubiquitous in cities of the segregation-era North and Midwest from the early 1900s to the late 1950s. A restrictive covenant is a contract in which homeowners agree with their neighbors that they will not sell to a person of color. A typical covenant contained the following language:

In consideration of the premises and the sum of five dollars ($5.00) each to the other in hand paid, the parties hereto do hereby mutually covenant, promise, and agree each to the other, and for their respective heirs and assigns, that no part of the land now owned by the parties hereto, a more detailed description of said property, being given after the respective signatures hereto, shall ever be used or occupied by, or sold, conveyed, leased, rented, or given to, Negroes, or any person or persons of the Negro race or blood.

Real estate developers cooperated with white neighbors to spread these covenants – to be effective, a certain percentage of neighbors had to sign up. Restrictive covenants were popular for only a short time, but even now, homebuyers often find a notation of a racially restrictive covenant on their deeds, although the contracts were outlawed in 1948 in the Supreme Court case of Shelley v. Kramer.

Rick Brooks and Carol Rose have written a wonderful book, Saving the Neighborhood (STN), about racially restrictive covenants. Although the book was published in 2013, I think it worth a second look at STN in the wake of Richard McAdams’ new and fabulous book, The Expressive Powers of Law. McAdams’ book is about the expressive role that law can play by informing people about what the law is and by helping them to coordinate their actions. Reading STN through the McAdams frameworks yields fresh insight. If McAdams’ book explains the expressive value of law, STN gives us an in-depth historical case study to apply McAdams’ insights about the relationship between legal and social norms. Both books are must-reads for anyone studying rule of law questions.

To set the framework, let’s briefly review McAdams’ argument on the expressive value of law and then return, fully briefed, to STN. McAdams, who is a renowned scholar of social norms, argues that law plays an important role in connection with informal norms. Beyond legitimacy or deterrence, law adds value to social interactions because law serves two expressive functions in connection with norms. First, law serves an information function because it tells people that particular norms exist and what those norms are. When we learn what the law is with regard to environmental pollution, for example, we learn what the collective rule is with regard to the appropriate way to handle pollution

Second, according to McAdams, law serves a coordinating function, allowing people to coordinate around a particular way of doing things. By putting a spotlight on one particular rule, law enables people to coordinate their actions around that rule or norm. So for example, when choosing whether to drive on the right or left, law tells us that the norm in the US is driving on the right, and enables us to coordinate with each other to avoid accidents.

Armed with a basic understanding of these concepts, we can now return to STN. The book argues that restrictive covenants were valuable not because they could be enforced but because they played important expressive roles, of the sort that McAdams discusses.

In making this argument, STN intervenes in the long-standing debate about what role restrictive covenants played in maintaining residential segregation. Commentators have pointed to what at first glance appears to be an absence of case law enforcing restrictive covenants, to argue that covenants played a very minor role in maintaining residential racial segregation. They also point to the fact that these covenants provided little resistance once neighborhoods began tipping from white to black. Their argument is helped along by the fact that restrictive covenants had a relatively short run, outlawed as they were only thirty years after they became popular.

In the McAdams tradition, Brooks and Rose argue that the real value of these covenants was expressive. STN argues that even in the absence of regular enforcement, and even after covenants were no longer legal, restrictive covenants played two important roles: a warning role (the equivalent of McAdams’ information role) and a reassuring role (McAdams’ coordination role). The book discusses each of those roles in turn.

First, restrictive covenants warned outsiders – black buyers outside the neighborhood – that the whites in a covenanted neighborhood had generated an informal norm to exclude black buyers. In McAdams’ terms, the restrictive covenant provided an information function, taking the temperature of the neighborhood as to exclusion and then communicating that state of affairs to would-be black buyers.

Second, covenants reassured insiders – whites inside the neighborhood – that their neighbors were committed to the (relatively risky and potentially violent) norm of residential exclusion. In McAdams’ terms, covenants performed a coordination function. White neighbors only wanted to commit to exclusion if everyone else did it (and on the same terms). Restrictive covenants thus reassured all segregationist whites interested in excluding blacks that their white neighbors would stick to the rule and to the same version of the rule. Brooks and Rose argue that, long after restrictive covenants ceased to be enforceable, covenants continued to play these two important roles – to warn outsiders and to reassure insiders of a continuing commitment to the (same) rule of exclusion.

More broadly, STN argues that this expressive value was more important than the ability to actually enforce the covenants. In terms of enforceability, the authors take the standard line: covenants didn’t have much enforcement value. But like McAdams, they make a more ambitious claim: covenants were more valuable because of their expressive value than because of their enforcement value.

STN lays out this central argument in two ways. First, the book traces the arc of the history of racial covenants, comparing enforceability and expressive value, and stopping along the way to explain the legal intricacies of property law and the reason that property law disfavored restrictive covenants. Second, the book walks through the arguments about warning and reassuring more definitively, using game theoretic analysis to show how covenants performed their expressive function of signaling to outsiders and insiders. I will tease apart and comment on their game theory illustrations in more detail in the last part of this review.

First, let me say a word about the history. STN’s extended discussion about the relationship between property doctrine and covenants is legal history at its best. We come to understand that to fully emerge, racial covenants had to make their way past some legal doctrine in property law that disfavors any constraint on the sale of property. The authors do a terrific job in explaining the convolutions of property law (including a delightful exposition on the Rule Against Perpetuities, the bane of many a law student).

I didn’t much like the case-heavy nature of their legal history in some key parts. Particularly with regard to the period during which covenants emerged, I would have preferred a more comprehensive historical timeline and analysis, of the kind provided by Michael Jones-Correa in his work on covenants.[1] But this is a minor quibble. My bigger quibble is with the second part of the argument: that covenant enforceability mattered less than the expressive informal value of the covenants. The following section details my disagreement.

Is Covenant Enforcement More Important Than or As Important As Expressive Effects?

STN argues that because neighbors had to pay a cost to enforce the covenants, and because only a handful of states developed a significant body of case law around covenants, it is reasonable to conclude that their enforcement played a relatively less important role generally, and in any event less important than their expressive role. But I’ve never really bought into the standard line that enforcement wasn’t important. And I think that both the standard line and STN underestimate the value of low-level covenant enforcement.

Indeed, I think the better argument is that for a key period of time early in a city’s history, low-level enforcement of covenants did much to gain racist whites more time before their little arrangement broke down and whites fled to the suburbs. Enforcing covenants gave residents a far less risky way (compared to cross-burning or violence) to pick off the small number of black buyers who were trying to break into the neighborhood early on.

All the literature on “block busting,” as the collapse of the racial boundaries was called, emphasizes how few black buyers were needed to trigger a cascade that involved whites fleeing and blacks moving in. Early on in a city’s history, pressure from black buyers was relatively light. Few blacks came knocking to try to buy into the neighborhood. During this period, enforcing just one or two contracts was all that was needed, at least in theory, to keep blacks out and the boundaries intact. With few black buyers knocking on the door, contract enforcement could play this “pick-off-the-invader” role quite well.

Indeed, Michael Correa-Jones documents this role, and reviews the details of cases in which covenant enforcement allowed neighbors to more easily police the color line against infrequent line-crossers. Correa-Jones argues, importantly, that although these covenants were expensive to enforce, their use was bolstered by important institutional players. The Federal Housing Association and the Home Owners’ Loan Corporation required that neighborhoods be covenanted before homes were eligible for federal subsidy and homeowners’ associations often footed the bill for litigation.

Later on, as Southern blacks migrated to cities of the north by the thousands, covenants became unable to withstand the increased pressure. During the Great Migration, black neighborhoods swelled, and property values for the white “border” neighborhoods nearby began to drop. At that point, the number of blacks knocking on the door increased, and whites were unable to hold the line.

The reason that covenants could not hold the line against increased pressure is simple. Whites who stayed behind now had to pay two costs that increased over time: (i) white neighbors had to pay to litigate the contracts against an increasing number of line-crossers, and/or (ii) for those who stayed but were going to flee to the suburbs eventually, residents had to pay in lost property value, if they waited too long. Owing to these two costs, once whites began to flee in large numbers, the cost of remaining to fight the rising tide became prohibitive, and flight was the only option for the rest.

I am persuaded that covenants played the expressive roles that Brooks and Rose describe, both before and after neighborhoods were breached. But I think that the authors undervalue those long peaceful periods early in the game, in which covenant enforcement offered a much easier way to keep out the occasional black buyer who tried to cross the line. In cities of the North, as the Great Migration was just beginning, covenants bought segregationist whites more time.

The Real Meaning of Covenant Ghosts

Brooks and Rose overestimate the expressive value of covenants in part because they undervalue the importance of playing for time. One indication of this is that they focus too much on the existence of covenant ghosts – covenants that continued to be written after the Supreme Court had struck them down as illegal. Even after the Supreme Court declared the covenants illegal, the FHA continued to insure newly covenanted subdivisions for over a year and real estate professionals continued to write covenants into new deeds and to refer to covenants when selling real estate. The authors make a big deal about these post-Shelley ghost covenants:

The fact that racial covenants continued to be written after Shelley suggests that even before that case, a major function of racial covenants was to allow white neighbors to identify themselves as allies in a preference for segregation and an intention to maintain it, and to signal the same to outsiders. Actual legal enforceability had never mattered as much as civil rights advocates thought. And hence Shelley’s denial of legal enforceability did not matter so much either. The case certainly weakened the larger imprimatur on covenants, but it left intact covenants’ ability to create among all the relevant parties a common knowledge of the local racial attitudes.

Thus, in the authors’ view, covenants continued after they had been outlawed because they were still able to play their warning and reassuring role even without being enforceable. Legal enforcement, and its demise, did not matter so much.

But a simpler explanation for covenant ghosts is available: these ghosts persisted because whites were playing for more time before they had to move on to some other way of excluding black buyers legally. In many areas of regulation – for example, tax shelters, payday lending – before letting go of the prohibited behavior, many regulated actors will first try to find loopholes that allow them to continue to engage in the behavior that has formally been prohibited.

This transitional period of loophole-searching is present in the authors’ own history of residential segregation.[2] For example, Brooks and Rose document this mode of loophole tweaking for over a decade in the case of the transition from segregation ordinances to restrictive covenants:

Just as Alabama’s legislatures had evaded federal peonage decisions with ever-shifting variations to give white planters control over black labor, so did southern cities and towns continually tweak the Buchanan decision, coming up with one new version of racial zoning after another, in order to help white neighborhoods hold on to their racial composition. These variations on the zoning theme were a drain on the slender litigation resources of the NAACP for years. But the pattern of judicial rulings, both in the federal and the state courts, clearly illustrated that those local laws could not last forever (p. 44-46). 

The existence of covenant ghosts really proves only that covenants were one of several legal instruments, each costlier than the last, that segregationist whites tried to tweak to find a loophole before moving to the next mode of segregation. These covenant ghosts were predominantly tweaked versions of the earlier instruments. For example, Zorita Mikva notes that post-Shelley, covenants were often modified slightly to avoid regulation, written with vague terms like “undesirable residents” to replace explicit racial terms.[3]

Evolutionary Game Theory: Public Goods Game with Punishment

My final comment on Saving the Neighborhood has to do with its use of game theory to illustrate the way in which the expressive function of law operates. This section of the book (and of my review) descends into the technical explanations of game theory, so readers might wish to skip down to the conclusion if the going gets too tough.

To illustrate their argument, the authors focus on three games, to suggest that over time, the relevant game changed along with the relevant Nash equilibrium, and to evaluate the role that covenants played in achieving the Nash equilibrium.

First, in terms of the warning function, STN suggests that restrictive covenants operated to signal to outside buyers that whites were committed to playing hawk in what is called a hawk-dove game.[4] That is, covenants signaled that those whites who had signed contracts were committed to playing the riskier and potentially more costly strategy of aggressively excluding black buyers and were willing to pursue exclusion through legal and perhaps also illegal means.  In this way, covenants permitted actors to converge on signaling a commitment to play hawk, to limit the cost of violence on both whites and blacks.

Second, in terms of the reassuring function, the authors argue that whites used covenants as a salient focal point around which to coordinate contributions in a Stag Hunt coordination game.[5] In the Stag Hunt game, the trick is to coordinate each player who wants to cooperate but only insofar as others do, and to coordinate on the same cooperative strategy. The restrictive covenant allowed whites to bind each other in their commitment to enforce the color line and cooperate by foregoing the potentially more attractive black buyer, attractive especially in a cascade after the breach.

Third, Brooks and Rose argue that once the wall had been breached, the game shifted from a Stag Hunt to a Prisoner’s Dilemma,[6] as whites now had the incentive to be the first to sell while property values were still high, and to sell to the first available buyer without regard to race. The authors also make reference to the Tragedy of the Commons for the same idea – that each person in a multi-player game has the incentive to take the goodies (fish, trees, profits from polluting or from selling across the color line) for herself even though she would do better if everyone continued to cooperate.[7]

My suggestion here is that making the game dynamic would have illuminated the importance of time that I discuss above – the role that enforceability played in giving whites more time, and the loophole efforts in which whites engaged to extract a bit more time from old legal instruments before moving to new ones. In addition, a more dynamic and up-to-date version of game theory could have illustrated most of their arguments, without the need for three games and the shift in between them, which seemed to me a bit clunky.  

In particular, the authors might have used a dynamic version of the Public Goods Game with Punishment (PGGP) with legal punishment. A dynamic PGGP is an n-player game of cooperation that mimics taxpaying—people contribute, everyone enjoys the benefits of those contributions regardless of whether they contribute, and non-contributors are punished. The game is dynamic because in each round, players compare notes with others, and then switch their strategy if they observe that another strategy produces a higher payoff. [8]

A dynamic model far more easily incorporates the elements of restrictive covenants that are of interest to Brooks and Rose. Notably, a dynamic evolutionary game permits us to easily see the effect of punishment – for example, punishing whites for defecting to sell across racial lines by litigating or violence – because the PGGP has punishment built in as a separate stage. As well, the dynamic game permits us to see the effect of time, and the way in which playing for more time can motivate strategic choices.

Finally, and in the same vein, a dynamic game also allows us to see the transition from a well-functioning cooperative equilibrium (where whites are holding the line) to a white-flight equilibrium where the covenants are now too expensive to enforce, and whites are fleeing and defecting en masse. No need to switch games mid-stream, as Rose and Brooks do.

To be sure, the PGGP game with signaling is far more complicated than the three games the authors use (and they do make reference to the Tragedy of the Commons, which is similar). Additionally, in a book written for the sophisticated but not necessarily technical reader, the PGGP game might well have proved too technical. Even so, the authors might have said much more about the transition from well-functioning covenant to white flight, to illustrate their warning and reassurance points.


Restrictive covenants are a terrific example of a social phenomenon in which social norms and law both play important roles. Does it matter if the law can’t be enforced? Is the more important role that law plays an expressive one, informing some about what the law is and reassuring others that people will coordinate to enable the kind of tasks that can only be accomplished via coordinated cooperation? In a conversation about that relationship, Brooks and Rose prove that a really great legal history can join game theory to explore that question. 

McAdams’ book extends the expressive value analysis more generally. He explores a range of examples to try to assess what expressive role law plays, and how much that expressive role matters relative to the role of enforcement. I look forward to reading more about this line of work from all these authors, and to exploring in my own work the role that time – and playing for more time – plays.

Posted on 8 June 2016

DARIA ROITHMAYR writes about the dynamics of racial inequality at the USC Gould School of Law, where she is the George T. and Harriet E. Pfleger Professor of Law. Her most recent book is Reproducing Racism: How Everyday Choices Lock In White Advantage, (NYU Press, 2014).

[1] See Michael Correa-Jones, “The Origins and Diffusion of Racially Restrictive  Covenants,” 115 Poli Sci Q. 541 (2001).

[2] Early on, the Black Codes in the South were designed to keep blacks on the plantation as agricultural workers. When middle-class Southern blacks began to double up or could afford on their own to buy into white neighborhoods, whites tried to find a loophole by enacting city ordinances to prohibit such moves. When ordinances were declared illegal in Buchanan v. Warley, whites in the South and now the North (owing to the Great Migration) turned to restrictive covenants to hold the line. And once Shelley v. Kramer and then the Fair Housing Act delivered the death blow to covenants, whites then turned to exclusionary zoning, a move that has held the line at least for many of the cities of the North, like Chicago, St. Louis and Cleveland.

[3] Zorita Mikva, The Neighborhood Improvement Association: A Counterforce to the Expansion of Chicago’s Negro Population (1951) (unpublished dissertation on homeowners’ associations, which were instrumental in both drafting and enforcing covenants).

[4] In a hawk-dove game, players have to decide whether to play an aggressive hawk (and risk getting beaten up if the other player is also playing hawk but get a bigger portion of the prize), or play a deferential dove (which avoids violence but gets less of the prize if the other player is playing hawk). The best response is for both players either to mix their strategy by probabilistically choosing hawk or dove depending on the payoffs, or to signal the other player that they are committed to playing hawk and will not back down.

[5] In a stag hunt game, players have to decide whether to play stag and pay a cost to cooperatively join others in hunting the stag (which produces more meat but only if other people join the hunt), or stay home and trap a hare alone (which is less costly but produces less meat). The best response here is for both players to find some way to coordinate, perhaps by agreeing on days of the week that are hare hunting days or stag hunting days. This need for coordination is at the center of the coordination function that McAdams and Brooks and Rose claim for law.

[6] In a prisoner’s dilemma game, players have to choose whether to pay a cost to cooperate with their partner and increases everyone’s payoff, or defect but still obtain some payoff, without paying to cooperate. Here, the best response is to defect.

[7] This Tragedy of the Commons is very much like the Public Goods Game for which I advocate. STN does mention it, but in my view, it should have been the central game, not just a casual reference. In addition, I argue that the game should be a dynamic one.

[8] In a dynamic public goods game with punishment, play takes place in stages. Players have to first decide whether to pay a cost to contribute to the public good or contribute nothing. The “state” then collects all the contributions, multiplies them by some term, and distributes a dividend to all players regardless of whether they contributed. In the final event of the sequence, players must decide whether to punish those who did not contribute. After punishment occurs and payoffs are calculated, players decide whether to switch their strategy for the next round, by comparing their playoffs to a random neighbor’s (or a neighbor on a lattice), and then switching to the neighbor’s strategy with some probability that depends on the difference between their payoffs. For examples, see Sigmund, K., de Silva, H., Traulsen, A., Hauert, C., "Social learning promotes institutions for governing the commons," Nature 2010, 466, 861–863; Boyd, R., Gintis, H., Bowles, S., "Coordinated punishment of defectors sustains cooperation and can proliferate when rare," Science 2010, 328, 617–620; Isakov, A., Rand, D.G., "The evolution of coercive institutional punishment," Dyn. Games. Appl. 2012, 2, 97–109; and Schoenmakers, S., Hilbe, C., Blasiusa, B., Traulsen, A., "Sanctions as honest signals -- The evolution of pool punishment by public sanctioning institutions," J. Theor. Biol. 2014, 356, 36–46.