By MITU GULATI
Review of THE EXPRESSIVE POWERS OF LAW: Theories and Limits, by Richard McAdams
Harvard University Press, 2015
The question why people obey the law is foundational in legal scholarship. Yet, as Richard McAdams illustrates in The Expressive Powers of Law, we do not ask it enough. On the occasions when I teach first-year law students, I sometimes begin by asking them why they think people obey the law. The first answer inevitably is “because there are penalties for disobeying.” If pushed further, answers such as “it is the right thing to do” or “it is part of the social compact” emerge. McAdams suggests that there may be an additional set of explanations for legal compliance – having to do with law’s expressive powers. He makes the case that law, through its ability to communicate information, is often able to induce people to comply regardless of whether there are penalties or whether people think that the legal system is legitimate.
It is tempting to conclude at the outset that the observation that law has expressive effects is trivial. Of course, law is an expression of the views of those who run the state. Indeed, it is a particular type of expression of those in power because something can only become “law” if it goes through certain processes such as vote of the legislature or a decision of a court. McAdams argues that the question of when this particular kind of expression of the views of the state impacts behavior is a complex one; and one worthy of more study. In particular, what seems to have excited McAdams is the possibility that law, by expression alone, can sometimes induce changes in behavior. If that is the case in certain contexts, then that means that resources do not have to be wasted in these contexts on the traditional methods of compliance inducement (such as the threat of penalties) that are expensive. McAdams’ project then is about rigorously (a) making the case that there are indeed some circumstances where law can work to change behavior through its expressive effects alone and (b) identifying those conditions.
The focus of the book, at least implicitly, is on the socially beneficial ways in which law, via expression alone, can impact behavior. But if McAdams is right about these expressive powers – and I buy his arguments -- there is a negative side as well that is worth considering. In particular, I’m thinking of the circumstance where the government miscalculates what is optimal for the public and puts out, via law, a preference for some new standard (for example, a type of cellphone technology or a boilerplate contract provision). Let us also say that the product here is characterized by network effects; that is, the users have an interest in purchasing the product that attracts the most other users because the value of the product is higher if it is the one that becomes the market standard. Government endorsement in such a circumstance could result in people coordinating around the suboptimal product, which in turn might result in the suboptimal product or custom getting entrenched. The point is that whether we want governments to try to alter behavior using their expressive powers will depend on how good governments are calculating what the socially optimal alterations of behavior are.
While McAdams is a true believer in the expressive value of law, however, he is also skeptical of many of the claims being made these days about this precise topic. McAdams thinks that too many people play fast and loose with claims of law’s expressive value. For law to have expressive value, he argues, it has to meet certain sets of conditions. He identifies two broad categories as circumstances where this expressive value can show up. First, where law helps to coordinate the expectations of people – traffic signs are a classic example in that people can use the signs to generate predictions about what other people are likely to do. Second, where the law provides people with credible information, particularly about risks – and here the classic example is smoking, where many governments appear to have been successful in conveying information about the dangers of smoking and, as a result, have altered behavior.
McAdams does of beautiful job of explaining, via the use of simple game theory, how law’s expressive effects work. Along the way, he also criticizes legal scholars for being unduly obsessed with the prisoner’s dilemma game and paying inadequate attention to the multiple equilibria coordination games that he says are far more applicable to legal analysis. I’m not sure whether all of the game theory detail about prisoners having dilemmas, hawks beating up on doves, and battling sexes was necessary to persuade the reader. But that is a matter of personal preference; more than a couple of pages of acronyms and game theory notation and I find myself getting confused. That said, the pages turned easily and the book moves fluidly between theory and application (I particularly enjoyed the discussions of Thomas Schelling’s work). One of the most appealing features of this book is that it is clear that the author is genuinely curious and excited about the question he is exploring; and he wants to share that with the reader. Put that together with the elegant and lucid prose, and the end product is a fun read. For me though, the best thing about the book was that it kept me thinking about its ideas many days after I had turned the last page. Two issues, in particular, struck me and I briefly sketch them out below.
McAdams tells us that law can impact people’s behavior through at least three different mechanisms: (i) fear of penalties from the state; (ii) belief in the legitimacy of the state; (iii) providing information that enables coordination among citizens or updates them credibly about risks. McAdams focuses on making the argument that law, in certain circumstances, can impact behavior via expressive effects alone. Fair enough; that is where the big theoretical payoff is. But if the goal, for the social planner, is to maximize compliance, the bigger question is how to use a combination of the three tools to maximize compliance. That means asking about interaction effects. These three methods of inducing compliance are surely related to each other. A state that imposes draconian penalties for small violations probably isn’t going to be perceived as all that legitimate and the expressions of an illegitimate state are unlikely to be trusted by the public.
Let us take traffic laws. The cross sectional variation across different states in terms of perceptions of legitimacy and in terms of the penalties imposed for basic traffic violations, could be exploited to get traction on the residual question that McAdams is most interested in – law’s expressive effects. Analogous, if not identical, traffic devices, such as stop and yield signs exist around the world. But they do not work with the same level of effectiveness in different settings. If it is correct that the same traffic laws, which are primarily coordination devices, work with different levels of effectiveness across different legal contexts, the question is what explains the differences in compliance. I’m writing this review while visiting my hometown in India. Traffic in my hometown is chaotic. More to the point, most drivers here do not seem to understand a Stop (or any other traffic sign) sign as anything more than suggestive. If someone decided to comply with the signs (for example, making a full stop at a Stop sign), that would probably cause a accident because no one else would be expecting that. But why? Is it because the drivers in my hometown do not trust the state to have properly estimated the need for coordination? Is it because there is just a general lack of belief that anyone else will comply, which results in an utter lack of coordination? Is it because the police are corrupt and rarely enforce the traffic laws in circumstances where they can obtain a bribe? In India alone, there is immense variation in terms of across the various states in terms of the quality of the government and efficacy of traditional deterrence based enforcement. It should be possible to exploit that variation to get better answers on many of the questions that McAdams raises. Until empirical work has been done, skeptics might wonder whether McAdams’ theory has general validity.
McAdams is persuasive in terms of the theoretical case that governments might be able to make law work by communicating risks (remember the paradigmatic example of smoking). But I’m not so convinced that law-making is how they do that, particularly in cases where there is an urgent need to alter behavior (as compared to, for example, via reports from government-sponsored, albeit independent and expert, health research institutes). Law, as McAdams acknowledges, is probably some combination of a reflection of the preferences of the median voter and interest group politics. In places where the median voter expects his or her government to perform the function of aggregating information about certain types of risks and of informing people of those risks and helping them make choices, one might expect the government to perform the role McAdams assigns to it. But do governments really ever do this? In the case of smoking, for example, is it possible that most legislatures around the world only acted after the median voter was so fully informed of the risks of smoking that a failure of the legislature to override the preferences of the tobacco industry would have resulted in job loss? And more important, how far beyond the smoking example does the McAdams claim go? Consider all the controversial cases—involving GMOs, vaccines, evolution, the efficacy of herbal remedies, climate change, and so on—where many people don’t pay attention to the government. Is it possible that people will stop believing the government just when the government tells them true things that they don’t believe are true?
To put this in testable terms, the claim that McAdams makes about risk disclosure should rise and fall as a function of the quality of the state’s credibility in (a) performing high quality research and (b) resisting the pressures of interest groups. Both (a) and (b) are measurable quantities. And so is the thing that McAdams cares about – legal compliance. So, to go back to smoking, we should be able to evaluate the anti-smoking legal expressions of governments around the world as a function of their effectiveness. Under the McAdams theory, that effectiveness should show up to be a function of (a) and (b) mentioned above. Of course there are issues that would have to be corrected for, such as the possible informational spillover effects from a jurisdiction with strong informational credibility (and weak interest groups) to one with weak credibility (and strong interest groups). But correcting for such complications is the bread and butter of empirical work. As mentioned earlier, I buy the theoretical case. Law can work in the way that McAdams suggests. But does it? And, if so, how often does this magic occur? In other words, are the incentives and constraints that operate on lawmakers such that the McAdams magic actually happens? The empirics, I am optimistic, will tell us a more nuanced and interesting story; they always do.
MITU GULATI is a professor of law at Duke University.