The Role of Law in the New Institutional Economics in Comparison With the Economic Analysis of Law
(1)
Department of Economics, Saarland University, Saarbrücken, Germany
In slightly changed form reprinted from Washington University Journal of Law & Policy, Vol. 26: 13; 2008, pp. 13–36 (Wahington University, St. Louis, MO, USA). The manuscript was finalized during my stay at the Hoover Institution, Stanford University, Fall 2006. I wish to thank Ken Arrow, Eirik Furubotn, Scott Kieff, Mitchell Polinsky and Oliver Williamson for their critical comments – and especially Ken Scott who advised me patiently on my legal presentations. Any remaining mistakes are the sole responsibility of the author.
1 Introductory Remarks
To describe the role of law in the new institutional economics (NIE), we compare this approach with the economic analysis of law (EAL) of the 1970s when the NIE evolved. At that time, the EAL was dominated by the “Chicago” or “market-based” approach that builds on the theory of perfect competition. Contracts are complete and Pareto efficient (allocative efficient). Ten years on, informational economic models that are briefly touched upon here extended this approach. After a few methodological considerations, this chapter concentrates on the Williamsonian branch of the NIE, i.e., the transaction cost approach (TCA). This theory argues that, in the real world of positive transaction costs and limited foresight, adaptation to the unforeseen becomes a central issue. Incomplete contracts can at most be “adaptive efficient” (North 1990, p. 80). The governance structure of contracts matters and becomes a bargaining point. Court ordering has to be complemented or substituted by private ordering. Attentive actors come to terms on a governance structure that protects them against ex post opportunistic manoeuvres of their opponents. Generic governance structures are (according to Williamson) markets, hybrids, and hierarchies. Court ordering works best for market governance. In the case of hybrid modes (franchising, leasing, etc.), courts would mainly be supplanted by private ordering between the parties. As for hierarchies, courts would stay out of conflict resolution (fraud and conflict of interest excepted). While the objects of research in NIE and EAL remained different, the latter’s methodology appears to move closer to that of NIE.
Both, NIE and EAL, result from early attempts to apply the methods of economic analysis not only to economic life proper but also to its institutional framework. Outstanding early contributors to this economic analysis of institutions are Coase (1937, 1960), Alchian (1961), Calabresi (1961), Buchanan and Tullock (1962), Olson (1965). What became known as the NIE gained momentum by the work of Williamson (1971, 1975)—who introduced the term NIE1—and North & Thomas (1973). The field of EAL had its take-off at the University of Chicago Law School in its “Chicago approach” to legal issues.2 It was Posner (1972) who brought EAL—under this name—to the attention of the general legal academy.3
NIE is more general than EAL regarding both its object of research and its methodology. The special object of EAL—at any rate during the 1970s4—is the economic analysis of legal rules under the assumption of perfect law enforcement.5 In contrast, the special object of the NIE is the economic analysis of organizational design6 of a wider set of transactions, by assuming that legal enforcement is impeded or impossible. As far as their methodology is concerned both, EAL and NIE pay attention to positive transaction costs. However, while EAL—at least in its “Chicago approach”—assumes perfect individual rationality and perfect foresight,7 the NIE from its very beginnings proceeded in a more general fashion by taking account of the imperfections of individual rationality and limited foresight. In short, the NIE and EAL differ in their specific objects of research with the NIE applying a more general method than EAL—at least during those early years.
Both fields expanded, and moved on, considerably during the past 30 years, and so a complete comparison of the two approaches would go beyond the scope of this chapter. We’ll restrict ourselves to a comparison of the two fields at their common period of origin. EAL was the precursor; NIE followed and criticized its methodology. That was certainly true for the Williamson branch of NIE on which we are focusing in this chapter.8 Williamson extended the Coasian criticism of the realism of the assumptions of neoclassical microeconomics by questioning the efficacy of court orderings. Due to transaction costs, legal arrangements are generally incomplete and would have to be complemented or substituted by forms of “private ordering”.9 As a consequence, Williamsonian NIE blurs the crystal clear neoclassical theories of (early) EAL. Not surprisingly, Williamson (1985) was heavily criticized by Posner (1993).
The following is an attempt to explain the role of law in the NIE as compared with what Posner would call “second generation” EAL. This chapter starts with an outline of the basic idea of the seemingly clear “Chicago” or “market-based” approach10 of the EAL which dominated in the 1960s and 1970s, illustrated by a legal interpretation of the theory of perfect competition. To better understand Williamson’s criticism, this chapter then briefly describes the basic ideas of mathematical contract theory. It is an attempt to model efficient contracting under transaction costs that started in the 1970s and was soon applied by economic analysts of law.11 A few methodological considerations follow. The chapter concludes with a summary and assessment of our remarks on the role of law in the new institutional economics.
2 Basic Ideas of the Market-based Approach to the Economic Analysis of Law
The economic analysis of law deals with a positive and a normative problem: The effects of legal rules on individual behaviour, and the social evaluation of these effects.12 As mentioned above, the assumed human behaviour was that of perfect individual rationality, i.e., individuals possess consistent and stable preferences, and maximize their utility.13 The social evaluation of the effects of legal rules followed certain ideas of welfare economics. We shall first discuss above-mentioned problems of EAL in the language of general equilibrium theory (information is perfect) and then of mathematical contract theory (information is imperfect).
2.1 The Institutional Framework of General Equilibrium Theory
Assume an economy consisting of a large number of perfectly rational individuals each endowed with a well-defined preference order and a bundle of goods. Individuals know that they can improve their lot by exchange. Given zero transaction costs,14 they will bargain with each other until they reach a Pareto efficient exchange equilibrium, i. e., a state of the economy in which no one can improve his individual position without hurting someone else.15
This is not the place to survey general equilibrium theory (GET). We content ourselves with a brief description of its assumptions and hypotheses to illustrate some of the basic arguments of the market-based approach to EAL. Its attractiveness lies in its (seemingly) clear predictions and clear valuations of the impacts of law,16 i.e., changes in legal norms and judgments.
The institutional framework of GET can be interpreted as the order of a private ownership economy, i.e., an economy whose elementary constitutional rules are based on the principle of inviolability of individual property rights. This demands an elementary legal order, plus its enforcement mechanism, regulating the property rights of individuals according to the general principles of private property and the transfer of these rights by consent according to the principle of freedom of contract.17 The enforcement of property and contract rights are supplied by an imaginary state or government, which otherwise remains perfectly passive.18
Individual property rights are embodied in GET by the assumption that individuals have full ownership in their endowments before and after trade.19 In the zero transaction-cost world, the definition and protection of individual property rights are problem free, and so are external effects. This is what was argued by Coase (1960), on the grounds that unlimited bargaining for rights (including the right to do something which has a harmful effect, such as the creation of smoke) will or may lead to a Pareto efficient general equilibrium.20 In other words, under the assumption of zero transaction costs, Pareto efficiency of general equilibrium is independent of the regulation of legal liability for damages.
Contractual obligations play a central role in GET. They are based on purchasing agreements, and consist of voluntarily assumed obligations like the obligation of the seller of a commodity to deliver the appropriate merchandise to the purchaser at the agreed upon time and location and the obligation of the buyer to pay the purchase price in timely fashion. GET describes both instantaneous as well as deferred exchange that involves the passage of time for its completion under otherwise given conditions. As for the latter, the future is involved and there are risks in the sense of the influence of other factors (hail, sunshine) on the contractual outcome. GET takes this into account by assuming that both the kinds of “other factors” and their statistical properties are known by everybody (hail with probability π = 0.01; not-hail π = 0.99)—and by extending the model in a manner that allows for the allocation of risk between contractual parties. Finally, GET can be used to describe all kinds of commercial contracts such as sales, lease, employment, loan, or insurance contracts.
In terms of efficiency and equity—general equilibrium under perfect competition is Pareto efficient but not necessarily socially fair. A “fair” distribution may require some redistribution of income. Since this is rather an issue of public law (the tax and transfer system) rather than of private law, representatives of EAL argue that “efficiency generally should be the primary criterion for evaluating legal rules.” (Polinsky 2003, p. 158)21
For illustration, we’ll have a brief look at the ingenious, though quite strong, assumptions of the Arrow-Debreu model, the most exact version of GET.22
Commodities are characterized not only by their physical nature and the date and location at which they are available, but also by uncertain events (“states of the world”) on the occurrence of which their agreed upon transfer is contingent (Debreu 1959, pp. 28, 98).23
Information of individuals is assumed to be complete in the sense that individuals possess “full information about the nature and consequences of their choice.”24 That is, consumers are perfectly informed about all commodities (goods and services) that exist at any given time and location; in addition, they know all possible states of the world and the probability of their occurrence25; and they are fully informed about all quoted and traded market prices.
Regarding individual behaviour, individuals are assumed to act perfectly rational in the sense that they maximize their individual utility, subject to their given endowments.26 Their individual utility functions are based on stable, well-ordered preferences with respect to all possible consumption plans (bundles of commodities), and weighted by their individual state preferences (their attitude towards risk).
Competition is perfect in the sense that there are enough potential buyers and sellers for each commodity that individual actors (households and firms) are unable to influence market prices.27
There are no frictional losses, i.e., no transaction costs: This is an old and well-known assumption of classical economic theory.28 When Debreu wrote his book, the neglect of frictional losses was a matter of course that needn’t be mentioned.
Given above conditions, exchange contracts are “complete”29 and, in general equilibrium, Pareto efficient. They consequently fulfill a necessary, but not sufficient, condition of a socially optimal income or wealth distribution (Samuelson 1947, Ch. VIII). However, its legal and institutional economic interpretation by use of a social preference function would cause considerable headaches.30 The EAL literature avoids these and the measurement problems related to Pareto efficiency that is based on ordinal individual utility measures. It employs instead, as a concept of efficiency, the maximization of aggregate wealth principle31 (Kaldor-Hicks criterion32), which uses a cardinal utility measure, and leaves distributional issues aside. At the same time, EAL avoids the general equilibrium approach and makes do with partial equilibrium considerations. Though, transaction costs are a recurrent theme in the EAL literature, it is argued that they are either low enough that their impeding influence on efficiency could be eased by use of freely concluded standard contracts (Posner 1977, pp. 69 and 292), or if transaction costs are prohibitively high the state would have to employ suitable methods of public control—either by court ordering or direct public control (Posner 1977, p. 271).33 In those cases, the government or the court should simulate the result of an unlimited private bargaining; that is, it should aim at a regulation or judgment that imitates a complete and, thus, an (at least bilaterally) efficient contract. “The court should correct an imperfect contract by restructuring its terms as the parties would have wanted it if the contract were perfect. Thus, the model of perfect contracts provides the organizing principles for the theory of contract enforcement.” (Cooter and Ulen 1988, p. 233. Italics in the original)
2.2 Institutional Perspectives of “Principal-Agent Theory”34
Information costs are an important part of transaction costs. A simple way to model them is to assume that information costs are either zero or indefinitely high—as assumed in principal-agent theory. Principal-agent theory deals with bilateral exchange. One actor, the “principal,” makes a take it or leave it offer; his opponent—the “agent”—accepts or rejects. Both are fully informed except the principal—who has a complete “blind spot” towards his (potential) agent. For example, the principal is unable to observe: (1) before contract conclusion his potential agent’s (e.g., his job applicants’) quality35; or (2) after contract conclusion, his agent’s effort level.36
However, to enable the principal to calculate his constrained individual utility maximum, the model builder gave him some extra pieces of information that he didn’t have (or needed to have) in GET. Examples of this follow:
For case (1): The principal now knows his applicants’ quality distribution (say, the numbers of high and low-quality job applicants). As a consequence, the principal is able to avoid an “adverse selection” among applicants by calculating a menu of contracts that brings the individual applicant37 to reveal his quality and, at the same time, helps to realize a second-best utility optimum for the principal.
For case (2): The principal knows, in addition to his agent’s reservation price, his agent’s utility function. He thus can calculate how his prospective agent will react to his offer. Given this information, the principal is able to offer his agent a compensation scheme that makes the agent maximize the principal’s utility together with his own. In case of uncertain results, the principal obtains a second-best optimum.
In both cases, no bargaining is allowed—only a take it or leave it offer by the principal. Efficiency measure is in both cases the maximization of aggregate wealth principle.38
The assumed institutional framework and individual behaviour of actors are the same as in GET: The elementary legal order of a private ownership economy, a state that is strong enough to enforce this legal order, perfectly rational acting individuals, and zero transaction costs—apart from the assumed specific “blind spot” of the principal. Agency contracts are complete and perfectly enforceable by the courts at no cost. As Williamson (1985, p. 27) puts it: “The future…holds no surprises; all of the relevant contracting action is packed into ex ante incentive alignments.” It is this lack of realism of the assumptions of GET and principal agent theory that gave rise to Williamsonian NIE.
3 An Interjection on Methodology
In his famous article on “The Methodology of Positive Economics”, Friedman argues against “…the belief that a theory can be tested by the realism of its assumptions independently of the accuracy of its predictions…” (Friedman 1953, p. 4139). Instead
…, the relevant question to ask about the “assumptions” of a theory is not whether they are descriptively “realistic,” for they never are, but whether they are sufficiently good approximations for the purpose in hand. And this question can be answered only by seeing whether the theory works, which means whether it yields sufficiently accurate predictions (Friedman 1953, p. 15; italics added).
For these reasons, Friedman prefers Marshall’s pure theory of perfect competition and pure monopoly to Chamberlin’s or Robinson’s theory of monopolistic or imperfect competition. He concedes, however, that his rejection of Chamberlin’s and Robinson’s criticism of GET does not imply that GET deserves any high degree of confidence. Any theory would be necessarily provisional and subject to change with the advance of knowledge (loc. cit. 41). Friedman (1953, 42f.) admits that
[p]rogress in positive economics will require not only testing and elaborating of existing hypotheses but also the construction of new hypotheses. On this problem there is little to say on a formal level. The construction of a hypothesis is a creative act of inspiration, intuition; its essence is the vision of something new in familiar material.
It would be highly desirable to have a more general theory than Marshall’s hypothesis of atomistically competitive firms, grouped into industries, and monopolies. Such a more general theory would “…enable us to handle problems we now cannot…”. However, “[t]o perform this function, the more general theory must have content and substance; it must have implications susceptible to empirical contradiction and of substantive interest and importance.” (Friedman loc. cit. 38)
The theory of imperfect or monopolistic competition, unfortunately, would possess none of the attributes that would make it a truly useful general theory (ibid., similarly Stigler 1968a, 311 ff.). “Its contribution has been limited largely to improving the exposition of the economics of the individual firm and thereby the derivation of implications of the Marshallian model, refining Marshall’s monopoly analysis, and enriching the vocabulary available for describing industrial experience.” (Friedman, ibid.)
Friedman concludes: “The theory of monopolistic competition offers no tools for the analysis of an industry and so no stopping place between the firm at one extreme and general equilibrium at the other.” (loc. cit., 39)
Would principal-agent theory meet Friedman’s attributes of a truly useful general theory? Does it have content and substance? Are its implications susceptible to empirical contradiction and of substantive interest and importance?40 Stiglitz (1985) understood informational economics (of which principal-agent theory is a special case) to be a more promising approach to economic life than GET. It certainly helped to rationalize actual business relationships like sharing contracts. But, as for its predictive power, doubts may be in place. As Hart and Holmström (1987, p. 105) admit, “… the extreme sensitivity to informational variables that comes across from this type of modelling is at odds with reality. Real-world schemes are simpler than the theory would dictate and surprisingly uniform across a wide range of circumstances…” We won’t discuss this question here any further. In this paper we are interested in the question: Does Williamsonian new institutional economics—the methodology of his transaction cost approach (TCA)—meet Friedman’s demanding criteria? To answer this question we’ll first have a brief look at Williamson’s TCA.
4 Williamson’s Transaction Cost Approach: Its Claims and Its Analytical Core
4.1 Williamson
Williamson claims that his transaction cost approach offers a more general theory of contracts than market-based EAL. The latter’s view of contracting would suggest “that commercial transactions are greatly dependent on and governed by legal forms and rules, while TCA addresses itself to a wider set of transactions.”41
The absence of frictions (=zero transaction costs) would certainly provide a useful standard, but there would remain numerous circumstances where the departure from ideal conditions is sufficient to warrant sacrificing the frictionless assumption when studying actual phenomena. Milton Friedman’s (1953, pp. 16–19) example of estimating the velocity of falling bodies by a simple formula (s = ½ gt2) is illustrative. Whether this ‘oversimplifies’ or is quite adequate depends principally on what is being dropped (e.g., a steel ball or a feather), how close to a perfect vacuum has been attained and the precision of measurement required (Williamson 1981, 56 f).
Williamson answers Friedman’s demanding criteria by, first, requiring that the transaction cost approach (TCA) has to be able to determine when the concept of “complete contract”42 applies and “when it needs to be augmented by a richer conception of [imperfect] contract.” Second, regarding the empirical support of the TCA, he requires that “[t]he critical transaction cost dimensions for describing transactions need to be identified, and the resulting mix of transactions needs to be matched with governance structures in a discriminating (economizing) way.” (Williamson 1981, p. 57)
How does Williamson fulfill these two claims? We’ll try to couch our answer in an outline of the analytical core of his transaction cost approach.
4.2 The Analytical Core of Williamson’s Transaction Cost Approach
EAL, as explored in this chapter,43 distinguishes between private contracting under conditions of low transaction costs with coordination of individual plans by complete contracts, and public intervention or court ordering if the transaction costs of private contracting are prohibitively high, with nothing in between. Williamson’s TCA, on the other hand, questions court orderings or state interventions as the only means to protect one’s contractual agreements against an opportunistic opening of renegotiations by the other side.44 TCA concentrates, therefore, on various means of private ordering.45 The two extreme poles of TCA are, accordingly, low transaction costs on the one hand (Williamson’s “market” case) and prohibitively high transaction costs on the other (coordination through privately agreed upon hierarchical “governance structures” like corporations—Williamson’s “hierarchy” case). The zone between of the two poles is filled by “hybrid modes of contracting”46 as represented by various types of “governance structures”47 of incomplete private contracts.48 Thus, the concept of “governance structure” or “organizational design” is not only used for the control of individual activities within hierarchies but also within incomplete contracts. Consequently, Williamson (1993, p. 38) occasionally calls his approach the “new economics of organization.”
However, neither EAL nor TCA offers a quantitative measure of transaction costs to facilitate empirical verification or refutation. Yet Williamson applies a substitute measure for transaction cost levels: frequency of transactions and transaction specific investments.49
Interestingly, neither Coase nor North or Williamson negates completely neoclassical microeconomics.50 They criticize51 the general assumptions of individual rationality, perfect information, and zero transaction costs. Their more general assumptions are bounded rationality, imperfect information (incl. uncertainty in the sense of not knowing what the future will bring—hail, rain, sunshine or what else?52) and positive transaction costs.53 If honesty is not considered to be an inherent personality trait (and it is not), opportunistic behaviour has to be expected. In fact, it is part of the central assumptions of Williamsonian NIE.54 As a consequence, the time after contract conclusion may cause “troublesome” problems in the world of NIE. Contract conclusion may lead to what Williamson calls the Fundamental Transformation: Because of transaction-specific investments, the contractual parties may find themselves locked into a bilateral monopoly situation.55 Due to their bounded individual rationality, the existence of transaction costs and imperfect foresight, the parties are unable to write a strictly enforceable complete contract.56 As a consequence, opportunistic behaviour of the other side becomes a problem.57
Preventing ex post opportunistic behaviour is difficult if not impossible.58 Due to information costs (as part of transaction costs) it is expensive, or even impossible, to verify one’s case to a third party such as a court. Court orderings may need to be supplemented or substituted for by private orderings if the parties wish to effectively protect themselves against ex post opportunism of their trading partners. “An important element in designing contracts becomes economizing on the costs associated with resolving disputes and governing exchange.” (Crocker and Masten 1991, p. 70). Williamson distinguishes in this context between three generic forms of governance, which also differ in contract law terms: markets, hybrids59 and hierarchies (Williamson 1996, p. 101).60
Empirical examples of contractual governance structures to limit opportunist renegotiations are described for long-term contractual relationships with highly idiosyncratic investments by the parties, for cases of vertical integration and instances of franchise bidding for public monopolies. For such examples, see papers by B. Klein, P. L. Joskow, Th. M. Palay, V. P. Goldberg & J. R. Erickson, O. E. Williamson, V. P. Goldberg and others reprinted in Williamson and Masten (1995).61 They support the hypothesis that, depending on circumstances, the governance structures of incomplete contracts can be seen as “efficient” adaptations—with “efficient” not being understood as Pareto efficiency. Efficiency in this sense “is not that of replicating ideal market results, but procedural efficiency in adjusting to an uncertain and changing environment” (Burrows and Veljanovsky 1981, p. 24). North (1990, p. 80) speaks of “adaptive efficiency”. Certainly, given the basic assumptions of the NIE, attempts at adjustment may well bring about improvements but they are not the same as “efficiency” in the sense of optimal procedural or adaptive means.62 For want of a better term, we’ll speak of “NIE-efficiency.”63