Encyclopedia of Law and Economics
Show Less

Encyclopedia of Law and Economics

Edited by Gerrit De Geest

The second, expanded edition of the acclaimed Encyclopedia represents a major update of the most authoritative reference work in the field of law and economics and the nine print volumes are now released online as a single integrated product.The Encyclopedia provides balanced and comprehensive coverage of the major domain in law and economics, including: criminal law, regulation, property law, contract law, tort law, labor and employment law, antitrust law, procedural law, and the production of legal rules. Each theme or volume is overseen by a leading scholar and each of the 166 entries is prepared by an expert in the field, providing an in-depth and authoritative overview of the individual topic, combined with an exhaustive bibliography, allowing users to access and filter the entire corpus of literature in law and economics.As with the print edition, the Encyclopedia is unique in serving both as an entry point and a platform for advanced research. The online edition is enhanced with Elgaronline’s powerful search tools, facilitating the search for key terms across the entire Encyclopedia, whilst the browse function allows users to move seamlessly between the volumes. These elements combine to create a powerful research tool for any researcher or scholar in the field of law and economics.
Show Summary Details
This content is available to you

Chapter 10: The focal point theory of expressive law

Richard H. McAdams

[In: Volume 7, Francesco Parisi (ed) Production of Legal Rules]

1. Introduction

A common claim about law is that it influences behavior “expressively.” At the most general level, the claim is that law’s behavioral influence is not entirely limited to the effect of legal sanctions (deterrence and incapacitation). The residual effect is termed “expressive.” Unfortunately, the expressive effect is often left at this vague level, with no specific theory of how it works. But some theorists offer expressive mechanisms of law’s influence. One possibility is that law communicates information, as by revealing the law-makers’ information about public attitudes or the risks of regulated behavior (see Geisinger 2002; McAdams 2000a; Dharmapala and McAdams 2003). Another theory is that law changes preferences (see Cooter 1998, 2000b; Dau-Schmidt 1990). This chapter addresses a third expressive theory: that in certain circumstances law facilitates coordination by supplying a focal point (see e.g. Cooter 2000a; Garrett and Weingast 1993; Hadfield and Weingast 2010a, 2010b; McAdams 2000b). To some degree, these different theories – sanctions, information, preference change, focal point – may be seen as rival hypotheses for how law influences behavior. But there is no reason why law cannot have all these effects, some being expressive, which work in an additive way.

The focal point theory might also be termed a “coordination theory” of expressive law. Game theory shows how individuals often have a common interest in coordinating their behavior. In these situations, if anything makes mutually salient a particular way of coordinating behavior, we frequently observe self-fulfilling expectations that the salient behavior – the focal point – will occur. One can construct focal points with “cheap talk” and law is one of the ways to do so. Merely by pointing out one way for individuals to coordinate, the law nudges their behavior toward compliance.

An obvious example is a law that tells everyone to drive on the right side of the road. The state’s announcement may be self-enforcing because everyone has an incentive to do what everyone else is doing – to coordinate – so as to avoid a collision. And the announcement makes everyone expect everyone else to drive on the right (below, I discuss whether we know why this is). But this example involves a pure coordination game because there is no conflict over whether to drive on the left or the right side. If law’s focal point power exists p. 168solely in games without conflict, then it is very limited, even trivial, because law almost always addresses conflict. Indeed, this criticism is broader than law. If focal points generally matter only in games of pure coordination, then they are a mere curiosity and not, as Roger Myerson (2004: 92) calls them, “one of the great fundamental ideas of social philosophy.”

Myerson is correct, however, because focal points also influence behavior in what Thomas Schelling (1960) called “mixed motive” games, where there exists both conflict and common interests. Individuals often have a common interest in coordinating their behavior even when some of their other interests conflict. The literature has explored two main categories: (1) games where the players need to coordinate their primary behavior (e.g. McAdams 2000b; McAdams and Nadler 2005) and (2) games where the players need to coordinate their enforcement behavior that sanctions some other individual’s primary behavior (e.g. Hadfield and Weingast 2010a, 2010b; McAdams 2005). Primary behavior might be driving, smoking, recognition of property boundaries, or finance, each of which has an element of coordination despite conflict, modeled by a game like Hawk-Dove, as explained below. Secondary or enforcement behavior might enforce decentralized social norms, private organizational rules, or any other kind of informal order, where the enforcers are more likely to succeed in sustaining the order if they coordinate their enforcement on the same understanding of the rule being enforced, even though they may each prefer a different conception.

This chapter reviews and extends the literature on focal point theory of law. Section 2 describes the general theory of focal points and the specific theory of law as a focal point. Section 3 describes some of the experimental literature supporting the claim. Section 4 shows how the literature has applied the theory to specific areas of law: constitutional law, international law, everyday disputing, and the enforcement of custom. Section 5 contends that the focal point power is so fundamental that it is necessary to explain the operation of legal sanctions. Section 6 concludes.

2. Theory

2.1. Focal Point Theory – Generally

Game theory attempts to predict behavior by deriving equilibria from the players’ beliefs, preferences, and opportunities. The payoffs for each outcome capture the players’ preferences. Sometimes the payoffs and other variables rationally constrain the possible outcomes to a single equilibrium. In singleequilibrium situations, game theory predicts that rational players will play the strategies that produce that one outcome; here, we can also say that the payoffs create expectations that the sole equilibrium outcome will occur. Yet sometimes the payoffs fail to constrain the outcomes and there are multiple equilibria p. 169consistent with rationality. Here, we cannot say that the payoffs determine a unique set of expectations about how the game will be played. When payoffs do not uniquely determine behavior, then influences other than payoffs can.

Schelling (1960) famously labeled these non-payoff influences “focal points.” He introduced the concept with pure coordination games. If you ask two separated people to name the same positive number as the other without communicating, there are an infinite number of ways to match answers. Yet when Schelling (pp. 55–6) posed the problem to his Yale colleagues and students, 40 percent named the same number – the number one. One might be tempted to say there is a purely rational process that produces this selection, but Schelling’s next example better illustrates the cultural contingency of focal points. Schelling modified the number problem by asking individuals what amount of money they would name if, by matching their answer with someone else trying to match them, they would each win that amount of money. Again, random matching should almost never occur, while matching at the amount of 1 dollar would produce almost no monetary return. Instead, a plurality of 29 percent of his respondents named 1 million dollars (though respondents knew they would not actually receive the money). It is difficult to imagine a culturally neutral explanation of this result. Indeed, psychological research finds that people from different cultures have different perceptual styles, rendering different objects “focal” in their visual field and visual recall. See Nisbett et al. (2001).

An equilibrium is focal because it differs in some salient way from the other equilibria for reasons not captured in the model. As Schelling (p. 96) puts it, some “symbolic or connotative characteristics that transcend the mathematical structure of the game” will tend to draw attention to one solution, making it “stick out” from the others. And human beings have learned that, when faced with just this sort of problem, the best way to proceed is to pick the salient or prominent possibility. Schelling (p. 57) summarizes: “Finding … a key – any key that is mutually recognized as the key becomes the key – may depend on imagination more than on logic; it may depend on analogy, precedent, accidental arrangement, symmetry, aesthetic or geometric configuration, casuistic reasoning, and who the parties are and what they know about each other.”

Schelling asserts that what is true of the pure coordination game is also true of the mixed motive games. We can see his point by introducing a slight degree of conflict in the above examples. Suppose we tell two individuals that they will receive a monetary payoff if they “match” in naming a positive number and nothing if they fail to match. But suppose we tell both individuals that one of them – Player A – will receive $100 if they match on an odd number and $99 if they match on an even number, while the other – Player B – will receive $100 for an even numbered match and $99 for an odd numbered match. The conflict here is trivial compared to the common interest in coordination, so we should not necessarily expect it to prevent coordination at the focal equilibrium. If p. 170the number one is sufficiently salient, B will be better off naming that number and getting a high probability of $99 than naming a non-focal even number and most likely receiving nothing. Although the size of the focal point effect is a contingent and empirical matter (see the experimental discussion below), there is no reason a priori to think that it disappears entirely as the magnitude of the conflict grows.

Where nothing in the feature of a situation makes a particular solution focal, Schelling contended that third party expression can make an outcome focal and thereby influence behavior. When a third party suggests or demands that the individuals coordinate in a particular way, she makes that outcome salient in a way that tends to create self-fulfilling expectations that the recommended or demanded behavior will occur. The third party expression can influence behavior even though it is “cheap talk” and does not change the payoffs or signal pre-existing information.

Schelling offers a paradigmatic example: the coordination of driving by a bystander. Imagine that a failed traffic light causes a traffic jam:

The bystander who jumps into an intersection and begins to direct traffic at an impromptu traffic jam is conceded the power to discriminate among cars by being able to offer a sufficient increase in efficiency to benefit even the cars most discriminated against; his directions have only the power of suggestion, but coordination requires the common acceptance of some source of suggestion.

(p. 144)

The situation involves mixed motives: there is conflict because each driver wishes to proceed before the other, but a mutual interest in coordinating to avoid a collision. Yet it seems likely that there will be some, even substantial, compliance with the Bystander’s hand signals. The Bystander does not threaten to chase down and punish drivers who ignore her hand signals, nor does she possess the indicia of legitimacy that sociologists and psychologists attribute to law. Nonetheless, given a need to coordinate, the Bystander’s ability to make salient one of the multiple equilibria gives her, as Schelling puts it, a genuine “power of suggestion.”

2.2. The Focal Point Theory of Expressive Law – Generally

The focal point theory of expressive law claims that law can work like Schelling’s bystander in the intersection, by using cheap talk to make salient a particular outcome. The theory claims that economists err in treating sanctions as strictly necessary to change behavior in the presence of multiple equilibria. Legal expression is one of the factors other than payoffs that can influence behavior.

Yet the scope of the theory is limited. McAdams and Nadler (2005) specify four conditions: (1) The situation the law addresses includes an element of coordination – the most important limit, addressed extensively below; (2) the law is sufficiently public; (3) the law is sufficiently clear, and; (4) there are p. 171no other competing focal points. The necessary conditions for the focal effect obviously do not always hold. First, law may address situations of pure conflict, where there is no element of coordination. For example, there is no element of coordination in a one-shot Prisoners’ Dilemma. Second, the publicity of the law usually depends on media coverage or advertisement. Law cannot create a focal point if the content of the law is generally unknown. Third, the content of the law is often unclear, especially to non-lawyers. Law cannot align expectations unless it is sufficiently clear that most individuals have the same interpretation of it, that is, believe it “points to” the same outcome. Finally, even if the rule is clear, law may face competition from factors that make another outcome salient. Most commonly, the law might attempt to change an existing norm that, as precedent for past behavior, continues to make salient the behavior that adheres to the norm.

Nevertheless, the necessary conditions frequently do hold. Indeed, law might be the form of third party expression for which these conditions are most likely to hold, making law a society’s general purpose mechanism for constructing a focal point. First, law frequently addresses disputes that often contain an element of coordination, as addressed extensively below in section 4. Second, there is often great publicity for legal rules from either media coverage or direct government advertising, as by public service announcements or the posting of signs. In any event, while it is true that law will have no focal point effect if it is not generally known, it is also true that law will have no sanction (or legitimacy) effect among those who are not aware of its existence. Third, though many laws are opaque, some are fairly simple, for example, the right of way goes to the driver with the green light or “no smoking” in restaurants. Finally, law often avoids having to compete with other stronger focal points, as where expectations are not fully settled.

3. The Experimental Literature on Legal Focal Points

An experimental literature generally supports the focal point theory. Several papers demonstrate the basic idea that focal points influence behavior in games of coordination (see e.g. Bacharach and Bernasconi 1997; Bosch-Domènech and Vriend 2008; Mehta et al. 1992, 1994a, 1994b). The Mehta et al. papers show that individuals do not just thoughtlessly choose the salient solution, but reason about what is likely to be mutually understood as the salient solution. Bosch-Domènech and Vriend (2008) demonstrate that focal points can even cause players to reach non-equilibrium and Pareto-dominated outcomes.

Considerable research demonstrates that cheap talk can create a focal point. When two players need to coordinate and one can send a message to the other identifying one of multiple equilibria, the expressed equilibrium tends to result. For reviews, see Crawford (1998); Valley et al. (2002). Other experiments establish specifically that third party expression can influence behavior in p. 172coordination situations. See Brandts and Holt (1992); Brandts and MacLeod (1995); Croson and Marks (2001); McAdams and Nadler (2005; 2008); Schotter and Sopher (2003); Van Huyck, Gillette, and Battalio (1992); Wilson and Rhodes (1997).

Most of these experiments involve pure coordination games, which heightens the possible importance of Crawford et al. (2008), who find that focal point influence can collapse as one moves from a pure coordination game to an asymmetric coordination game. Crawford et al. make an equilibrium focal by using a more salient label for the strategies that produce that equilibrium. They identify a pair of labels (e.g., among Chicago subjects, the Sears Tower versus AT&T Tower) that successfully produces coordination in the pure form of the game and then show that the same labels fail to produce coordination when they introduced even slightly asymmetric payoffs (creating a slight degree of conflict between the players over which equilibrium is best). Yet the Crawford et al. results are not uniform. In two conditions of their second experiment (AM1 and AL1 of the “pie games” results reported in Table 5 on p. 1455), the focal point continues to influence behavior despite payoff asymmetries. As the authors state (p. 1454): “label salience remains powerful, even in the presence of very large payoff asymmetries.” Thus, on the evidence presented, the provocative title of the article – “The Power of Focal Points is Limited” – is not generalizable. Apparently, some focal points work in the face of conflict and some don’t. More data are needed.

In any event, there are additional experiments finding focal point power in mixed motive games. One example is McAdams and Nadler (2005), who randomly assigned subjects to play the game in Figure 10.1. The subjects were not given the labels “hawk” or “dove,” but that is the game (Hawk-Dove) these monetary payoffs produce (where strategy 1 is Hawk and strategy 2 is Dove). The two equilibria are C1/R2 and C2/R1 (where payoffs are underlined). The game creates conflict because player 1 prefers the R2/C1 equilibrium, while player 2 prefers the R1/C2 equilibrium. Yet the players also share an interest in avoiding the R2/C2 outcome.

Figure 10.1
A Hawk-Dove Game

McAdams and Nadler compared the participants’ interactions with and without third party cheap talk. Thus, in the control, there was no expression. The expressive condition consisted of one of two messages describing the two p. 173equilibria: either “R1/C2” or “C1/R2,” delivered in one of two ways: (1) by engaging a mechanical device (spinning a spinner) that would randomly point to a space containing one message or the other (where each message took up exactly half of the space to which the spinner could point); or (2) by keeping one subject – a “leader” – out of the Hawk-Dove interaction and assigning him or her instead the task of writing one message or the other on a blackboard. All messages were in plain view of all the participants in the decision-making process, so there was common knowledge that each individual observed the message.

Because of the conflict, a subject who receives a message to play Dove (and receive 0) would prefer that both parties ignore the message and instead play the mixed strategy equilibrium (and receive 0.5). Nonetheless, expression influenced the behavior of the subjects in the expected direction. The R player was significantly more likely to play strategy 2 (Dove) when the message included R2, compared to the no-expression control, while the C player was significantly more likely to play strategy 1 (Hawk) when the message included C1, compared to the control. The opposite message had the opposite effects (also significant). McAdams and Nadler found a stronger effect when a human leader delivered the message, but even the mechanical device produced a significant effect on behavior.

Other experiments more explicitly introduce law. One way is to test whether behavior changes when the experimenters introduce a rule enforced by non-deterrent sanctions (see Bohnet and Cooter 2001 and Tyran and Feld 2006). If the rule’s only effect is the deterrence generated by the fines enforcing the rule, then non-deterrent sanctions should have no effect on behavior. But these studies find that the expression of a rule affects behavior. For example, Tyran and Feld (2006) have players in a three-person voluntary contributions game vote for or against a rule requiring a certain level of contribution, and find that the creation of a rule increases contributions even though the sanctions enforcing the rule are non-deterrent. Based on the material payoffs, there is only one equilibrium to the game, but one can explain the results as a focal point by noting that non-material payoffs (which typically produce results other than the sole Nash equilibrium) can create multiple equilibria and a need to coordinate among them.

Similarly, Galbiati and Vertova (2008) run experiments with public goods games (where each subject decides whether to contribute some part of his or her monetary endowment, provided by the experimenter, to the group, where it will be multiplied and then allocated back in equal shares to each member of the group). The efficient outcome is for everyone to allocate all of their endowment, but the only Nash equilibrium is to give nothing. Despite this prediction, public goods experiments typically find that some individuals will contribute. Usually, however, if the game is repeated, the contribution rates fall p. 174towards zero. One possibility for explaining why there are any contributions is that, because of non-self-interested or non-material motivations the experiment cannot exclude, some people act as reciprocators. For reciprocators, the public goods game involves coordination. A reciprocator may be torn between returning the high contribution levels from other reciprocators (who start out expecting high contributions) and the zero contribution levels from selfish players taking advantage of the reciprocators. Anticipating the non-contributors, and given risk aversion, reciprocators may begin with modest contributions. So reciprocators struggle to coordinate with each other on single contribution level, while selfish players drag that level towards zero.

If reciprocators need to coordinate around a particular contribution level, then they would be influenced by the third party construction of a focal point. The Galbiati and Vertova (2008) experiment captures this dynamic. When the experimenters included in the instructions a statement that a specific “minimum contribution” of 80 percent of the per round endowment “is required,” contributions rose significantly compared to the control that imposed no such obligation, even though there was still no monetary incentive to contribute in either condition. Galbiati and Vertova also created a monetary incentive for correctly predicting what amounts the other players would contribute. They determined that the effect of the high “minimum contribution” requirement was to increase the subjects’ estimate of how much others would give. For reciprocators, this belief would motivate increased contributions. Galbiati and Vertova infer that the 80 percent obligation operated as a focal point for such players, creating self-fulfilling expectations that others would give close to 80 percent. See also McAdams and Nadler (2008).

In sum, the experimental evidence supports the idea that legal expression influences behavior by creating a focal point for coordination.

4. Applications: The Scope of Focal Point Theory

Theorists have discussed several important examples of the coordinating power of legal focal points, primarily in legal contexts where the threat of sanctions does not exist to explain compliance. Most of the existing discussion concerns the need to coordinate primary behavior, where the underlying game is one of coordination. Sections 4.1, 4.2, and 4.3 discuss constitutional law, international law, and the possibility of law’s margin of focal point influence in everyday disputes. By contrast, section 4.4 discusses the enforcement of informal order, where the underlying primary conduct may not present a game of coordination, but those enforcing the informal rules may need to coordinate their sanctions.

4.1. The Focal Power of Constitutional Law

Some constitutional law is enforceable like other law, as where a private party resists a private claim of defamation by asserting free speech limits to the tort. p. 175But when disputes arise directly between private citizens and the government or between different branches of government, it is artificial to speak of legal sanctions enforcing constitutional rights. In these contexts, one must wonder why the executive or legislative branches would heed judicial declarations of unconstitutionality, given that the courts have no legal means of enforcing their judgments except through agents of the executive and no way of funding itself except through agents of the legislature. Stephenson (2003) frames the general question by asking why the parts of government with “the money and guns” listen to the part with neither? One can go a step further and ask, in what sense the legislature has any money to allocate or the chief executive has any arms to wield except for the expectation that lower level actors will carry out legislative directives on the budget and executive directive on coercion.

The answer is that legal actors, as salient speakers, can provide a focal point around which the parties coordinate. The first step is to see that constitutions address coordination problems, as a significant political science literature suggests. Hardin (1989, 1999, 2006) characterizes constitutions as arising from a bargain between powerful interest groups who wish to form a state. He models the situation as a multi-player Battle of the Sexes game because the players are always better off “matching” their strategies, by agreeing to the same structure of government, than they are if they fail to agree. In Elster’s (2007) model, political parties expect to alternate control of government and thus find themselves in an iterated Prisoners’ Dilemma (PD) game where each party will benefit if both parties adhere to certain restraints of power when in office. Despite the conflict between parties, there is a strong element of coordination because, to achieve cooperation in the iterated PD game, the parties must agree on the boundaries of governmental powers.

Weingast (1997) focuses on democratic constitutions. He claims that the stability of democracy depends on “the people” being willing to challenge official action that transgresses democratic principles, as by a politician attempting to stay in office after being defeated in a lawful election. He models the problem as a complex game where citizen groups can maintain democratic rule only if they challenge the official jointly. Each group prefers to challenge the official if the other group does the same, because only joint action will work, but would rather acquiesce if the other group acquiesces, because unilateral action is costly as well as ineffective. It is therefore essential that citizen groups coordinate their efforts to challenge government officials around a “social consensus” of state authority. In a large diverse society, that consensus is unlikely to arise in a decentralized fashion, so the constitution is the means by which some actors create a consensus. See also Ordeshook (1992, 2002); Posner (2001); Posner and Vermeule (2008).

In these theories, a constitutional text works by making focal the forms of governmental behavior the constitution demands. Writing down the allocation p. 176of power in a particular structure of government makes that allocation salient and creates self-fulfilling expectations that the various players will demand at least as much power as granted in the writing, forcing players who wish to avoid chaos to cede that much power. Strauss (1996) observes that the theory directs courts to give serious weight to constitutional text. We see the tool of writing used not only in nation states, but by organizations that enact by-laws that constrain the powers of different actors in the organization. Even pirates used written constitutions to limit the power of their captains and quartermasters (see Leeson 2007, 2009).

4.2 The Focal Power of International Law

Because there is no overarching sovereign among nations, international relations are understood as a form of anarchy, a state of nature. Various scholars, notably Goldsmith and Posner (1999, 2005), have made the same points about international treaties as those just covered regarding constitutions: First, the interaction of states frequently presents a game involving coordination. Second, a written arrangement or adjudication between states may influence their behavior by virtue of creating a focal point and thereby creating self-fulfilling expectations of how to behave.

Garrett and Weingast (1993) provide perhaps the first model of this sort, viewing international law as arising in an iterated PD game. They make the same analytic move as Elster does with constitutional law, emphasizing an element of coordination within the PD game because there is more than one way to cooperate. All states are better off when everyone complies with one form of cooperation than when all states defect or try to cooperate in different ways. Thus, they view a treaty as a potentially self-enforcing agreement that specifies what form their cooperation will take.

McAdams (2005) uses coordination games and focal points to explain the puzzle of successful international adjudication despite the substantial lack of sanctions. If disputing states have an interest in coordinating – as where the worst outcome for each is a Hawk/Hawk conflict – then the arbiter can influence their behavior by making one equilibrium a focal point (the one where the party the arbiter declares to be the loser gives in to the arbiter’s declared winner). Ginsburg and McAdams (2004) use this theory to explain international adjudication, including the puzzling fact that 68 percent of the decisions of the International Court of Justice (“ICJ”) produce prompt compliance, despite the fact that the ICJ cannot credibly threaten sanctions. Ginsburg and McAdams claim that the compliance rate is high because nations tend to bring to the ICJ those disputes for which a focal point will resolve their dispute. They also claim to explain the observed low compliance in cases lacking mutual consent to the adjudication, i.e. where the ICJ overrules one party’s jurisdictional objection and later rules against the same party on the merits. If one party seeks to avoid p. 177IC jurisdiction, that party may not regard the dispute as a game involving coordination, in which case there is no room for a focal point.

Another international example is standard setting (see Drezner 2007). When international parties seek to standardize weights and measures, communications protocols for air traffic control, the right of way among vessels in international waters, the international exchange of mailed and telephonic communications, or the exchange of fingerprints by police departments, they are solving a coordination problem. If the states were indifferent between different standards, the problem would be purely one of coordination. More likely, the states disagree over which standard is best, each preferring its own national standard. But despite this disagreement, the great benefit from “matching” standards with others means that the worst outcome is often to fail to coordinate. A legal agreement setting a standard works by creating a focal point.

Finally, the effort to harmonize domestic regulations and legal rules is itself a type of standard setting. Here the “standard” is the private domestic law of say, contracts, securities, or antitrust. The greater the divergences between nations, the more difficult it is to transact across national borders. Policy convergence and harmonization save transaction costs when private firms seek to do business internationally. At the same time, a state incurs costs in switching from one legal standard to another. When two or more nations would benefit from policy convergence or harmonization, but each would prefer that other nations shift to their legal standard, the result is a Battle of the Sexes game. Again, once the legal agreement is made, it works as a focal point and there is no incentive to deviate.

4.3. The Focal Power of Law in Everyday Disputes

The above examples are relatively clear because much of the compliance we observe with international or constitutional law does not appear to be the result of sanctions. McAdams (2005) and McAdams and Nadler (2005, 2008) argue that, even where sanctions exist, law’s focal point power may have some independent influence on behavior. They focus on disputes, which can frequently be modeled by the Hawk-Dove game in those cases where the worst outcome for both sides is the “total conflict” that results when the parties fail to resolve the dispute.

One example is traffic, which involves priority “disputes” between drivers. For example, two stopped drivers on perpendicular streets seek to make turns that cut across the path of the other. Or where drivers are in merging lanes, each seeks to get ahead of the other. Or drivers traveling in opposite directions approach a one-lane bridge that each wishes to cross first. In all these cases, each driver wants to proceed ahead of the other, but there is a common interest in coordinating to avoid a collision (and to avoid the outcome where both wait for the other to proceed). In traffic situations, the state may generate much of the compliance that it achieves merely by expression. Traffic signals, for example, are much like the randomizing device (a spinner) that McAdams and p. 178Nadler (2005) used experimentally to express one or the other equilibrium to a Hawk-Dove game. If a driver approaches an intersection for the first time, sees a “yield” sign, and knows there is no such sign for the other road or lane, the driver is more likely to expect a collision if he fails to yield. If the drivers in the other lane can see that the first driver’s yield sign applies only to his road or lane, the inference that the other drivers will not yield is particularly strong. Even without the fear of state sanctions or respect for state legitimacy, the driver complies with the yield sign to avoid a collision.

Many disputes have the same structure because disputants may share an interest in avoiding the outcome of “all-out” conflict (unlike a Prisoners’ Dilemma). The most pervasive reason to avoid winning at all costs is the potential for violence. However unlikely, violence is a background risk of disputing. Besides American news reports of a shooting over something trivial, like breaking in line or talking in the movie theatre, much of the violence in ostensibly ordered societies involves individuals engaged in “self-help” remedies against someone whom they regard as having infringed on their rights (see Black 1983; Merry 1991; Nisbett and Cohen 1996). Less dramatically, a second result of continued conflict is a heated shouting match or exchange of profane insults. The parties may regard this outcome as the worst possible result of a dispute, particularly when people know each other socially – such as neighbors or co-workers – because the exchange may terminate their relationship.

Thus, for example, suppose two individuals want to sit in the same public area for a time and one wishes to smoke a cigarette and the other wishes not to be exposed to cigarette smoke. Or two neighbors dispute the exact location of their property line and one wants to plant a tree on the contested land while the other wants that no tree be planted. It is likely that the smoker and non-smoker will each think it better to give in to the other side that to have a physical brawl. Thus, the smoker wants to give in if the non-smoker will insist, but to insist if the non-smoker will give in; so too for the non-smoker. In these cases, there are two equilibria: where the smoker defers and non-smoker insists and vice versa. Similarly, the neighbors may not care enough about a sliver of property to think it worth committing or suffering an assault. Here, too, there are multiple equilibria, each being where one party gives in to the other.

Whether the concern is avoiding violence or embarrassment, these are situations where a legally constructed focal point will generate compliance. A law that is sufficiently publicized – say, by a “smoking prohibited” sign – can influence expectations by making focal the outcome where the non-smoker insists and the smoker gives in. In the designated no-smoking area, the law “waves on” the non-smoker and the smoker believes it is more likely he will insist. In a designated smoking area, the law (or other designator) creates the opposite expectations. The same is true of a legal rule or ruling that clearly awards the disputed property to one of the claimants. Law can create self-fulfillingp. 179 expectations of what the players will do, generating compliance independently of sanctions or legitimacy.

4.4 The Focal Point Power to Clarify Customary Rules

Hadfield and Weingast (2010a, 2010b) emphasize an entirely different coordination problem, not the one that arises between branches of government, nations, drivers, or disputants, but one that arises between mutual enforcers of informal order. Informal order arises without law and may determine how people solve coordination problems. But ambiguity in informal order can prevent coordination in some cases As Postema (1982: 178) explains, while “the standard situations needing coordination are provided for,” “[t]here may … be substantial differences of opinion in the community regarding what the convention requires in some specific instances.” The inevitability of ambiguity is an old idea in language and law. One example is the inevitable incompleteness of contracts, which Garrett and Weingast (1993) use to explain why international tribunals are needed to supplement the meaning of treaties. McAdams (2005) discusses two forms of ambiguity in informal order: “fuzziness” – when the conditions for conventional rules are not precisely defined – and “potential incompleteness” – where there is uncertainty whether the precisely defined convention is subject to some exception for novel circumstances.

Hadfield and Weingast (2010a, 2010b) discuss the example of trade. A trading culture may informally sanction contractual cheating, but the ban on cheating may be effective only because third parties – not the cheater or the cheater’s victim – boycott the cheater. (Note this model is similar to Weingast’s model of democratic government, where the citizens can hold politicians in check only if they coordinate their sanctioning.) McAdams (2005) discusses the example of property. For example, Sugden (1986) describes property as a convention that emerges informally when individuals play strategies based on who currently possesses a valued resource. But McAdams (2005) argues that, without centralized articulation of rules, individuals will encounter some cases where it is difficult to decide who is currently in possession and other cases with uncertainty whether something other than current possession matters.

Without common expectations, players fail to coordinate. Legal expression is one salient means of clarifying an existing custom, removing ambiguity in how it applies to a particular situation. Clarifying the custom means providing a supplement to the customary rule. If the legal supplement is salient, it is a focal point. Even without possessing sanctions (or legitimacy), the legal clarification can align the disputant’s expectations and change behavior.

5. The Focal Point Theory of Legal Sanctions

In mature legal systems, one may lapse into thinking that sanctions are independent and exogenous. But why does it change the expected payoffs of p. 180a behavior when a legislature or administrative agency promulgates penalties for the behavior? Why should one fear sanctions when the court announces a judgment of damages or a sentence of imprisonment? To ask these questions is really to ask why enforcement agents who actually impose sanctions obey officials who pronounce the law. Here, there is a potential for infinite regress. One can say that the sheriff fears that, if he refuses to carry out a court order to seize property, then the court will hold him in contempt. One could then say that the sheriff fears being held in contempt because, if he is, the bailiff will seize him. But why will the bailiff do that? Who does he fear? Obviously, the bailiff does not comply because the individual who issues the legal pronouncement – the judge – will personally, physically enforce it.

So the problem arises at each level – the legal pronouncement that Y is to sanction Z does not ensure that Y fears sanctions for failing to sanction Z. The pronouncement that X is to sanction Y for failing to sanction Z does not ensure that X fears sanctions for failing to sanction Y. And so on. As Mailath et al. (2001: 3) put it: “Words written on a piece of paper don’t alter the laws of physics.” Thus, the law against theft doesn’t “affect what physical activities (such as attempts to lock an individual in a cell) individuals are capable of…. [M]y struggle to keep from being locked in a cage by you will result in an outcome that is determined by our relative strengths and martial arts abilities, but not by the existence or nonexistence of a law” (p. 3).

The path out of this puzzle is the power of expectations in a coordination situation. A mob boss (or guerilla leader or pirate captain) may not be the best physical fighter in his organization, but has power by virtue of everyone expecting everyone else to obey his orders. Any one person expects other mob members to kill him if he defies the boss’s orders just because he would expect the boss to respond by ordering his death and would expect others will follow the boss’s order. Wherever there is effective leadership, these interlocking expectations exist that one must obey the leader because the leader controls the group and the group has more power than any one individual.

Why does this leadership situation occur? Again, the basic answer is coordination. As Schelling (1960: 91) says, “The coordination game probably lies behind the stability of institutions and traditions and perhaps the phenomenon of leadership itself.” The first level of coordination is merely that there is some advantage for the group to have different people doing different tasks at different times. Carrying out illegal activities, fighting a war, sailing a ship, all require careful coordination of the labor of multiple individuals. Frequently, the best action to take depends on circumstances that are changing rapidly, making time of the essence. For this reason, coordination is unlikely to occur in a decentralized, bottoms-up fashion. Even though everyone in the mob or guerilla army benefits from coordinating an attack on rival forces, it will not naturally occur that everyone decides on his own to implement the individual parts of p. 181the best overall strategy of attack. Instead, an individual needs to give orders. The group requires a leader.

Now we reach the second level of coordination: who will the leader be? If everyone has an opinion he is willing to express about how the group should proceed, upon which expression will everyone coordinate? To solve this problem, the group needs to designate a particular speaker whose expression, being the focus of everyone’s attention, will be salient. But there are too many volunteers to lead. Here the situation is like a Battle of the Sexes game: possibly every individual would like to be the boss (commander, captain, etc.), but if each holds out to be the leader, then the result is chaos. Because there are gains from coordination – for example, the profits of criminal organization – each person gains more from being in a mob where someone else is boss than from the chaos where everyone claims to be the boss and no one is.

The critical point is that, once the group selects a boss, this person is like the Bystander Example in the introduction: his instructions create self-fulfilling expectations of compliance. The same is true of legal actors. First, society has basic needs for coordination that it accomplishes through law. Building roads, fighting crime, regulating pollution, waging war, all require coordination. Second, to solve the basic coordination problem, societies will in some way select a political and legal leader or, more likely, an array of them. This selection will render focal the orders of that individual. Once in place, the orders of legislators or judges are obeyed merely because everyone expects everyone else to obey them (and everyone expects everyone to expect everyone to obey them, etc.). Given those expectations, the legal actor can use expression to influence the enforcers who impose sanctions. Like the mob boss, everyone comes to expect that everyone else (or enough to make it matter) will obey the executive’s order or the judge’s decree. The legal actors have the power, by expression, to create self-fulfilling expectations that their demanded behavior will occur.

To return to game theory, where legal sanctions work, it is not by changing the payoffs for given actions, but by changing expectations about what actions will be taken. The act of stealing always pays x if no one subsequently seizes the thief’s property and always pays y when someone later seizes the thief’s property. Law does not affect these functions. Instead, because the political and legal leaders – legislators and judges – declare that the thief’s property will be seized, the law changes the expectations about whether one’s property will be seized if one steals. This is the foundation of legal sanctions.

6. Conclusion

The need for coordination is ubiquitous. When coordination is required, law can influence behavior merely by making the outcome it demands focal. The articulation of legal rules can also facilitate coordination among enforcers of p. 182informal customs and norms. Ultimately, the coordination provided by leaders underlies the actual infliction of legal sanctions.