Pareto Efficiency, the Coase Theorem, and Externalities: A Critical View

Abstract:

We offer a critical review of the Paretian definition of efficiency by noting its correspondence with the description of the logic underpinning market relationships. With reference to Ludwig Wittgenstein’s analysis of the value content of tautologies, we discuss the tautological nature of the propositions regarding the efficiency of exchange and market equilibria when there are externalities. We critically review the debate on the Coase theorem, its treatment of externalities, and the contrast with the Pigouvian approach, taking into account the distinction between propositions that are true based on their formal logical structure and propositions that are true with respect to their correspondence to actual states of affairs. We also reveal and discuss the logical inconsistencies — in particular, the one between the Coase theorem and perfect competition — and the practical consequences of the application of Pareto efficiency to the analysis of externalities.

Keywords::

• Coase theorem
• economic methodology
• efficiency equity
• market failures
• social values

JEL Classification Codes::

• A13
• B41
• D61
• D62
• D63

Notes

¹ The essential elements of the Pigouvian position is defined in Arthur C. Pigou ([1924] 1932), Howard S. Ellis and William Fellner (1943), James E. Meade (1952), Tibor Scitovsky (1954), and Erza J. Mishan (1969, 1971). For the alternative view, see Arrow ([1969] 1983), James M. Buchanan and W. Craig Stubblebine (1962), Coase (1960), Harold Demsetz (1967), Knight (1924), George J. Stigler (1966), and Ralph Turvey (1963).

² Worthy of note is Walter P. Heller and David A. Starret’s (1976) attempt at a systematization of the issue within the neoclassical economics mainstream. Their work could be deemed the final point of the debate that took place in the 1970s, although the difficulties in achieving a consensus are evident in the authors’ own conclusions: “It seems to us that the intuitive concept of externality must remain somewhat imprecise” (Heller and Starret 1976, 20).

³ Terry L. Anderson (2004).

⁴ In the same line of inquiry is Tihomir Ancev and Michael Harris’s (2006) work, in which the difference between the Coasian and Pigouvian approach is identified in the relationship the two approaches maintain with social norms (see also Dragun and O’Connor 1993).

⁵ Paul Samuelson (1995). Among those who stress the tautological character of the Coase theorem, are Robert D. Cooter (1987), Joseph Farrel (1987), and Dan Usher (1998).

⁶ Lionel Robbins (1935).

⁷ On the link between Pareto and the sciences of nature, see for example Roberto Marchionatti and Enrico Gambino (1997).

⁸ A statement that led Hodgson (1992, 756) to note: “The general equilibrium project lies in ruins. It does not ‘deal with’ economic subsystems in any adequate or meaningful sense. It can represent neither true markets, nor money, nor key types of knowledge and uncertainty, nor real time” (see also Blaug’s [2003] assessment of these models). Mark Blaug (2003, 154) assesses these models, concluding: “The best way not to learn how market function and how a competitive economy actually works is to study general equilibrium theory.” The point is that, to general equilibrium theorists, neoclassical economics has lost any link with an empirical starting point. The theory is defined as a “consistent set of definitions, hypotheses, theorems that can be used, when needed, to formally represent various concepts and problems of economic theory” (Ingrao and Israel [1987] 1990, 207). The bases for such axiomatization of general equilibrium have been set by Kenneth J. Arrow and Gerard Debreu (1954) and Gerard Debreu (1959). As noted by Joseph Stiglitz (1994, 29), the major achievement of Arrow and Debreu is that they revealed how exceptional are the conditions required to obtain the correspondence between efficiency and equilibrium — that is, the fact that such conditions are impossible to be realized in actual markets. The truth is that neoclassical economics has utilized these models not to reject the Walrasian theory of equilibrium, but rather to be able to use it. On one hand, the economic system has been described as a simple mathematical structure, “with a radical and uncompromised emptying of the theory from any empirical reference” (Ingrao and Israel [1987] 2006, 270). On the other hand, formal analyses of abstract models have been used as if they could provide indications on how to build an economy that might get close to those ideal conditions. The fallacy of this methodological approach is what we want to expose, with specific reference to the problem of externality.

⁹ Arrow ([1969] 1983, 60) proposed a definition of transaction costs (later used also by Williamson 1985, 18), according to which “transaction costs are the costs of running the economic system.” It is thus clear that to assume absence of transaction costs means to neglect the presence of any cost associated with the functioning of actual markets — that is, of all potential sources of inefficiencies. It follows from this a definition that “market failure is the particular case where transaction costs are so high that the existence of the market is no longer worthwhile” (Arrow [1969] 1983, 68).

¹⁰ Karl Polanyi (1977, ch. 2).

¹¹ In the Universal Declaration of Human Rights, art. 25, para. (1), we read: “Everyone has the right to a standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services, and the right to security in the event of unemployment, sickness, disability, widowhood, old age or other lack of livelihood in circumstances beyond his control.”

¹² A discussion point that, in many ways, parallels the one we make here has been presented by Andrew Halpin (2011). Halpin rejects the idea that the Coasian model sheds light on the world as it is, thus overcoming blackboard economics, and criticizes the legitimacy of its normative consequences.

¹³ According to Ludwig Wittgenstein ([1921] 1981, Prop. 4.1), “[a] proposition presents the existence and non-existence of atomic facts.”

¹⁴ Although Wittgenstein did not directly state this, and although it might be possible to argue in favor of some propositions of the social and historical sciences (to the extent that they describe facts), we believe that Wittgenstein’s position in Tractatus is basically what we have described.

¹⁵ It is important to keep in mind that, according to Wittgenstein, in order for a proposition to have a cognitive value, it must describe a possible state of the world. Formal disciplines, in this sense, are conceived as rules for the creation of propositions, and not as sets of sensible propositions themselves, given that they cannot indicate facts. Therefore, formalizing a discipline does not exempt it from the need to compare the propositions with the existing facts in order to assess their truthfulness.

¹⁶ Cf. in particular, Wittgenstein ([1921] 1981, Prop. 4.003 and 4.11). Among the many commentaries on the relationship between Wittgenstein and the Vienna Circle, see Friedrich Waismann (1968) and Rudolf Carnap ([1928] 2003, x).

¹⁷ “We feel that, even if all the possible scientific questions are answered, our vital problems have not even been touched” (Wittgenstein [1921] 1981, Prop 6.52, emphasis original).

¹⁸ “Die Phänomenologische Sprache oder ‘primäre Sprache’ wie ich sie nannte schwebt mir jetzt nicht als Ziel vor; ich halte sie jetzt nicht für möglich. Alles was möglich und nötig ist, ist das Wesentliche unserer Sprache von ihrem Unwesentlichen zu sondern.” (“I no longer have the Phenomenological or ‘primary language,’ as I once used to call it, in mind as my goal. I no longer hold it to be necessary. All that is possible and necessary is to separate what is essential from what is inessential in our language.”) (Wittgenstein 2000, 2, 118). We must clarify that the expression “phenomenological or ‘primary language’” must be intended here as an equivalent to the ideal language of Tractatus.

¹⁹ Between 1929 and 1946, Wittgenstein continued his conversations with Piero Sraffa. As Wittgenstein ([1945–1949] 2009) himself recognized in the Philosophical Investigations, his relationship with Sraffa was decisive in Wittgenstein’s decision to abandon the conceptions on language from Tractatus. The theme of the relation between Sraffa and Wittgenstein, and the reasons for Sraffa’s interest on Wittgenstein’s theses on language — although somehow related to the issues we discuss here — would require an in-depth analysis that is beyond the scope of this article.

²⁰ By reviewing this debate, Fabrizio Fornari (2002, 120, authors’ translation) shows how lines of thought, very different from each other, such as those of Edmund Husserl, Hans-Georg Gadamer, Émile Durkheim, or Karl Mannheim, “highlight, first of all, that any form of knowledge, be it commonsensical or scientific, is intimately connected to the socio-cultural context within which it develops, and then that the cognitive and interpretative activities performed by social actors are constitutional elements of society itself.” Our claim that there are valuations hidden behind the Pareto efficiency, and that the latter implies a specific form of social relationships, can be one example of such connections, mostly overlooked by the neoclassical economic literature. The wider implications for economics of these themes are explored by Wade D. Hands (2001).

²¹ As Steven G. Medema (2011) recalls, Stigler (1966) is the first one to present Coase’s (1960) set of propositions as a theorem.

²² Although he initially limited his analysis to the case of a single actor, thus to a partial equilibrium analysis, Coase later accepted the more extensive interpretations of the theorem as proposed, among others, by Stigler (1966), Warren G. Nutter (1968), Harold Demsetz (1972), and Robert Feldman (1974).

²³ On this, see Medema (2011, footnote 29) who states that “[t]he invariance proposition does not hold” and “this is seemingly the only aspect of the Coase theorem milieu that, to this point, had not been subject to any serious debate.” The truth is not that a serious debate has not been attempted, but that those who defend the Coase theorem seem to be diminishing it (see, for example, Medema [1994] in reply to Canterbery and Marvasti [1992]).

²⁴ The invariance proposition has been criticized since the beginning by several authors, including Stanislaw Wellisz (1964), Guido Calabresi (1965), and Erza J. Mishan (1967). Medema (1995, xviii) summarizes the debate on this issue: “The outcomes are not likely to be the same (in the sense of an identical allocation of resources) under different initial rights assignments, owing to effects on income, wealth and tastes.”

²⁵ We must note here that bilateral contracting is very different from a market process. Contrary to what happens in perfectly competitive markets, in a bilateral bargaining the possibility remains that gains from trade would be dissipated by the strategic behavior of the involved parties. Cento Veljanovski (1982, 60) draws attention to this issue and, in a lucid discussion on the theorem and on the definition of transaction cost, concludes: “[T]he assumption of zero transaction costs is not sufficient to ensure that the parties settle at the gain maximizing outcome unless it is explicitly assumed that bargaining is cooperative, and this, I suggest, is to assume rather than establish the validity of the Coase theorem. … Indeed if there is any theorem in such a world it is the exact opposite of the Coase theorem. The appropriate theorem in bargaining context is what I shall term the Johansen theorem [Johansen 1979]: Direct bargaining has an inherent tendency to dissipate the gains-from-trade through strategic behavior.” Similarly, Robin Hahnel and Kristeen A. Sheeran (2009) underline the difference between game theorists’ “complete information,” and the traditional assumption of “perfect information,” concluding that the latter is insufficient to avoid. Also, in a bilateral bargaining, the incentive to hide one’s own preferences would destroy the possibility of reaching an efficient outcome.

²⁶ According to Coase (1960, 41), if a polluter is forced to pay a Pigouvian tax equal to the damage caused, the symmetry between those who produce damages and those who suffer from them requires that a “polluted tax” should also be imposed to any additional cost of pollution-reducing measures taken by those who create the externality.

²⁷ It may be worth noting that such “social omniscience” hidden in the “[l]ack of precision regarding ‘transaction costs’” (Hahnel and Sheeran 2009, 234), necessary for the Coase theorem, could also be used to defend the Pigouvian scheme.

²⁸ It is worth recalling that to overcome the logical impossibility of reconciling perfect competition — which implies perfectly elastic demand functions for individual firms — with positions that are off equilibrium (where, given a discrepancy between the amount supplied and the amount demanded, some of the agents would be “rationed,” i.e., either consumers would not be able to acquire, or producers would not be allowed to sell all the goods they want at the equilibrium price), Arrow (1959) stated that any departure from the equilibrium would mean an exit from perfectly competitive conditions. Thus, already in 1959, Arrow associated perfect competition exclusively with equilibrium positions, while, in 1969, he claimed that a situation where there is no tendency toward equilibrium could be defined as a “competitive equilibrium.” A critique similar to ours has been proposed by Joseph Stiglitz (1994, 34), where he notes that Arrow and Debreu’s idea of treating commodities at different times and in different states of nature as different commodities is at odds with the fact that there could never be competitive markets for them.

²⁹ See footnote 25.

³⁰ Stigler (1966, 113; see also Conley and Smith [2005] and Cerin [2006]). Medema (2011, 21) also considers this point, arguing that “under Stigler’s version of perfect competition, perfect knowledge preclude the sort of bilateral monopoly situation that could interfere with the bargaining process contemplated by the Coase theorem.” The statement stems from the proposition by which, essentially, the hypotheses that define perfect competition necessarily include the absence of transactions costs (in its broadest definition, inclusive of those relative to imperfect information). As we demonstrated, such an instrumental use of the concept of transaction cost is intrinsically contradictory.

³¹ We must, however, emphasize that, although the Pigouvian position implies such a possibility, the prevailing attitude is still that of trying to evaluate any externality in monetary terms by making direct or indirect reference to market valuation methods (see, for example, Freeman 2003).

³² We think that the expression “market failure” is nevertheless ambiguous. If we consider that virtually any element of reality introduced in the neoclassical model determines a discrepancy between market outcomes and efficiency, it would be proper to say that the “failure” is not of the market, but rather of the neoclassical economic model of the market. Particularly valuable is, in this respect, Julia Black’s (2013) suggestion that, to be meaningful, economic analysis should borrow from other social disciplines and abandon the absolutely abstract conception of the “market” typical of neoclassical economics.

³³ Although not directly linked to the theorem, one logical conclusion of treating externalities this way has been the idea that it is possible to arrive at a decision to export toxic waste to poor countries in compliance with economic efficiency. The Economist (February 15, 1992), for example, published a memorandum by the then chief economist of the World Bank, Lawrence Summers, stating: “The measurement of the costs of health-impairing pollution depends on the foregone earnings from increasing morbidity and mortality. From this point of view a given amount of health-impairing pollution should be done in the country with the lowest cost, which will be the country with the lowest wage. I think the economic logic behind dumping a load of toxic waste in the lowest-wage country is impeccable and we should face up to that. … I’ve always thought that under populated countries in Africa are vastly underpolluted” (cited in Peter 2004, 2).

³⁴ An essential element of logical positivism is to maintain that only propositions that are empirically verifiable have a cognitive value since they can affirm the existence or non-existence of the phenomena they describe. By contrast, a purely logical inference that does not refer to any real things necessarily ends up in a “tautological transformation.” “The second basic error of metaphysics consists in the notion that thinking can either lead to knowledge out of its own resources without using any empirical material, or at least arrive at new contents by an inference from given states of affair. Logical investigation, however, leads to the result that all thought and inference consists of nothing but a transition from statements to other statements that contain nothing that was not already in the former (tautological transformation)” (Neurath [1929] 1973, 308).

Additional information

Notes on contributors

Andrea Ventura

Andrea Ventura is a lecturer at the Università di Firenze.

Carlo Cafiero

Carlo Cafiero is a senior statistician and economist at the United Nations Food and Agriculture Organization.

Marcello Montibeller

Marcello Montibeller is a doctoral fellow at the Università di Sassari.