Economics of Conventions, Social Norms, and Law

Fellow, RIETI

Unfortunately, I have not had the chance to learn "law and economics" in a comprehensive and systematic manner. However, I have devoted myself considerably to the study of "contract theory." The contract theory, as I understand it, though it may be outdated, is classified into two categories - "complete contract theory" and "incomplete contract theory."

In complete contract theory, it is assumed that rules of allocation can be made fully contingent on every possible outcome that is both observable and verifiable. In contrast, incomplete contract theory analyzes the relations between contracting parties in situations where state-contingent contracts cannot be written beforehand and there remains room for renegotiation in the future, either because situations relevant to the contracting parties are not verifiable (though observable), or hard to predict, or because it is too costly to write a contract that covers all the possible states. "Verifiability," as referred to here, means that a certain situation can be determined as fact by a third party, such as a court.

As exemplified by such classification, the concept of verifiability is extremely important in contract theory. Although this theory has been successful to some extent, I have a few doubts about the aforementioned framework of it. First of all, contracts derived from this theory are extremely complex (specialists on the theory recognize the importance of this problem and have been conducting study on it) and it has turned out that the form of real-world contracts (for example, a contract promising a pay rise if performance is favorable) can stand only on the basis of very restricted hypotheses. The second point, profoundly related to the first, is that it is assumed that these real-world contracts are optimal contracts based on the presupposition that contracting parties endeavor to ensure efficiency. Thirdly, analysis conducted under this framework is based on the assumption that any contracts - as long as they are drawn up on verifiable variables - are 100% enforceable. It is this third point that is the main theme of my argument in this column.

Legal Phenomena and Law Enforceability

Not only economists but also many other people seem to assume that the presence of the legal system provides individual laws with substantial enforceability. Simply put, enforceability means that: (1) people in general lead a life that is free from conflicts with laws, and (2) they believe that the legal system is there to protect their legitimate rights even if someone attempts to infringe upon them. When we look at the reality of the Japanese legal system, the first point may be observed to some extent, but it is hard to recognize the second. The reality of Japan's current legal system is dire as described in detail by Yamaguchi and Soejima (1997) based on a number of cases they witnessed as practitioners. Even in very clear-cut cases, such as those regarding cash loans, in which the lender apparently has justice, the debtor who shirked payment, despite being sentenced, still has ways of evading a court order.

Ordinary economic theory tells us that people would stop obeying the law under a legal system that failed to implement penal regulation. In other words, there is even a possibility that the aforementioned first point on enforceability may be undermined. According to Yamaguchi and Soejima (1997), however, this point has actually been maintained to some extent, thanks to the presence of social norms with which many Japanese comply.

What are legal phenomena? This question, which concerns all areas of social science, has been the consistent theme of economics studies ever since this scholastic field was established. As far as I know, it is the theory of the law philosopher Hart as published in his "The Concept of Law" (1961) that provides the clearest distinction between conventions (or tacit agreements), social norms and law. Although conventions and norms are not the exact words Hart used in his argument, I would like to explain his theory by using these terms that are familiar to economists.

According to Hart, when a group of people behave in almost the same or similar manner in certain situations, it is deemed that conventions exist. However, for certain behaviors to be defined as social norms, Hart says, it is necessary that: (1) acts that deviate from generally accepted common practice are regarded as a mistake or a fault, and thus, such acts are subject to criticism; (2) people generally believe that deviations from rules constitute a sufficient reason for criticism; and (3) the behavior considered the norm is regarded as a general standard that should be followed by all. Hart adds that social norms or rules of obligations must, in consideration of an axiom of human nature, include restrictions on the free use of violence, theft and deception; he calls these "primary rules."

When a society grows in complexity, however, it becomes insufficient to rely on primary rules alone, as inefficiencies may arise either because of uncertainty about the existence of rules, the static and entrenched nature of rules, or the lack of authoritative bodies to ascertain violations of rules. To compensate for these shortcomings and enforce the rules, a set of "secondary rules" must be established. Such supplementary rules include "rules of recognition" for disposing of doubts as to the existence of rules, "rules of change" for empowering an individual or body of persons to introduce new rules and eliminate old ones, and "rules of adjudication" for empowering individuals with the authority to determine whether or not a certain rule has been broken. It is around such secondary rules, according to Hart, that there emerges a full-fledged legal system that is composed of statutory law, legislature and courts.

In line with the key points described below, implications of the above argument can be summarized as follows. Firstly, conventions and social norms differ on one extremely crucial point. To make a certain behavior enforceable, there must be foundations to make people believe that the behavior needs to be acknowledged as a social norm rather than simply being convention. Secondly, even the enforceability of statutory law is still greatly affected or restricted by the existence of social norms that constitute primary rules. In the case of the Japanese legal system, which is weak in areas involving secondary rules, the role of social norms as primary rules becomes all the more important.

From Game Theory of Conventions to Game Theory of Norms

Conventions and norms are both characterized by a strategic complementarity, in other words, the property that each member of a society gains more by following the rules that other members follow. As a tool for analyzing more general strategic relationships, game theory has been rapidly developed with much research being published on its application to the study of conventions and norms. These research publications should be highly valued as an attempt to explore normative approaches that differ from the traditional framework of social choice theory. To name but a few, such works include those by: Lewis (1969), Sugden (1986), Binmore (1994, 1998), Aoki (2001), Young (2001) and Matsui (2003).

As a rough generalization, many of these arguments are based on the assumption that a self-enforcing property - which is characteristic of the Nash equilibrium of game (i.e. no player has an incentive to unilaterally deviate from the equilibrium as long as others stay there) - is the essential nature of conventions. Meanwhile, some try to incorporate the Hobbesian idea of social contract.

These research works are extremely diverse both in content and underlying ideas, and they cannot be reviewed against any coherent standard. As a common drawback, however, it can be pointed out that most of these works are devoted to analyzing conventions and fall short of analyzing social norms in the aforementioned context.

For instance, Sugden (1986), who attempted to analyze by using game theory the arguments David Hume made on conventions in his "A Treatise of Human Nature" provides the following explanation for the emergence of property rights in a society:

Suppose there are two people and a physical property that both of them want, but which can only be used by one person in an exclusive manner. Conflicts of interests between the two people over the ownership of the property can be formulated into the hawk-dove game as described below. (Both the row and the column players have to select either a dovish strategy or a hawkish strategy. In the "payoff" matrix provided below, each cell contains two numbers, with the left number representing the payoff to the row player, and the right number that to the column player following the respective combinations of their strategies.)

Dove1, 10, 2
Hawk2, 0-2, -2

This game has three equilibria, one symmetrical and two asymmetrical. In the symmetrical equilibrium, where both players adopt the same strategy, the players choose "dove" with a probability of 2/3 and "hawk" with a probability of 1/3. This equilibrium would emerge when nothing differentiates between the players. The sum of the players' expected payoffs in this equilibrium is 4/3, which means this is a socially inefficient result. If either one of the players already possessed the property, however, this fact becomes the focal presupposition, and breaks the symmetry of the game. The one who already possesses the property would play the "hawk", while the other would play the "dove," producing a socially efficient result with the sum of both players' payoffs being 2. Based on this observation, Sugden says that property rights emerge as a convention in the form of ownership and possession as a package. In a likewise manner, he explains various conventions as equilibria of games.

Meanwhile, Sugden has devoted a substantial number of pages to explaining the difference between social norms and conventions. Specifically, he emphasizes the nature of social norms, that people want others to act in accordance with them, and then checks whether each of the conventions he identified can be a social norm by examining the payoff structure of the corresponding games. However, there has still been insufficient analysis made of social norms themselves.

Today's arguments on "law and economics" are also not paying sufficient attention to the issue of social norms. Kaplow and Shavell (2002), for instance, insist that it is more desirable to base arguments regarding legal policies on social welfare criteria rather than on justice. The argument is extremely stimulating and worth listening to, but it does not give sufficient consideration to social norms, a critical foundation that constitute the law.

Setting aside the Kaplow-Shavell argument, reasons behind the tendency to not give enough consideration to social norms may be attributable to the fact that economists have been too preoccupied with finding correspondence between Hume's notion of conventions and the Nash equilibrium of games. From now on, it is necessary to conduct research along the framework set up by Hart, as pointed out at the beginning of this column. Meanwhile, as suggested by Aoki (2001), recent research on game theory seems to be going beyond the realm of pure mathematical theories, moving in the direction of placing more emphasis on interaction with knowledge of institutional analysis. Research activities are expanding with an unprecedented diversity and are not yet showing signs of converging. Any attempts to include social norms in such research may be fragmented, given the dynamic trends of this area, but activation of this field of study which has hitherto only been subjected to modest attention should nevertheless deserve appreciation.

In the Case of Intellectual Property Rights

Some readers may think the argument of this column lacks relevance to any policy but that is not necessarily the case.

For example, let's take a look at "intellectual property rights," a topical issue receiving increasing attention. Earlier on in this article, I introduced Sugden's argument on how the concept of property rights emerged as a "convention" of society. The argument was based on the assumption that the property in question cannot be simultaneously owned by two or more people. Today, however, there is a growing move to establish and protect the "property rights" of the fruits of human creative activities, a property that does not have the characteristics assumed by Sugden, going under the name of "intellectual property rights." This can be interpreted as an attempt to analogically expand the concept of rights, which used to acknowledge only tangible properties, to include ideas, information and other intangible properties that have otherwise not been regarded as subject to exclusive ownership.

Predicting the fate of "intellectual property rights" is beyond my ability. However, the series of conceptual frameworks presented above are applicable when looking into its future development.

First of all, although the aforementioned idea concerning the conceptual expansion of rights - from a concept acknowledging only tangible properties to the one that is also applicable to intangible properties - has surely become more common than before, such a viewpoint will not likely expand linearly. In China, for instance, social norms respecting intellectual property rights are said to be scarce. In such a country, simply strengthening law enforcement to prevent the infringement of rights may not lead to the actual strengthening of intellectual property rights. Secondly, even if the idea concerning the analogical expansion becomes widespread and the rights to intangible properties become acknowledged just the same as those of tangible properties, this does not necessarily guarantee the enhancement of social efficiency. Rather, recent findings of empirical studies suggest the possibility that further strengthening of intellectual property rights may lower the level of social welfare. Thirdly, even if legal protection over intellectual property rights is reinforced, the question remains unanswered as to whether a more draconian system is actually enforceable. This third point, which is also related to the first point, involves questions of whether the protection of intellectual property rights can take root as a norm that is generally accepted in a society, and to what extent violators will be punished in the event that they don't take root.

As such, there is no denying that the situation is still in primitive stages. However, the enforceability of law is now being required to take an approach different from that under conventional contract theory, and it is believed that such an attempt has quite a good chance of success.

July 22, 2003
  • Aoki, M. (2001), Toward a Comparative Institutional Analysis, MIT Press.
  • Binmore, K. (1994), Game Theory and the Social Contract: Playing Fair, MIT Press.
  • Binmore, K. (1998), Game Theory and the Social Contract: Just Playing, MIT Press.
  • Hart, H. L. A. (1961), The Concept of Law, Oxford University Press.
  • Kaplow, L. and S. Shavell (2002), Fairness versus Welfare, Harvard University Press.
  • Lewis, D. (1969), Convention: A Philosophical Study, Basil Blackwell.
  • Sugden, R. (1986), The Economics of Rights, Co-operation and Welfare, Basil Blackwell.
  • Young, P. (2001), Individual Strategy and Social Structure: A Evolutionary Theory of Institutions, Princeton University Press.
  • Yamaguchi, H. and Soejima, T. (1997) , Saiban no Himitsu (Secret of Trials), Yosensha.
  • Matsui, A. (2003), Kanshu to Kihan no Keizaigaku (Economics of Conventions and Norms), Toyo Keizai Shimposha.

July 22, 2003

Article(s) by this author