2. Costly institutions: the consistency of the Coasian contributions

The substance and the purpose of institutions consist in reducing transaction costs: ‘Markets are institutions that exist to facilitate exchange, that is, they exist in order to reduce the cost of carrying out exchange transactions’ (Coase, Reference Coase and Coase1988a: 7). Subsequently, if the cost of performing the transactions in the market is higher than in firms, then firms ‘substitute’ markets in that task.

And again, Coase maintains that ‘the operation of a market costs something and that, by forming an organization and allowing some authority (an “entrepreneur”) to direct the resources, certain marketing costs are saved’ (Coase, Reference Coase1937: 392).

Thus, the main argument delivered by Coase in ‘The Nature of the Firm’ is that firms exist because there are costs of using the price mechanism and these costs can be reduced by the use of an administrative structure. The firm's ability to substitute administrative fiat for bargaining may resolve concerns linked to dealing with transactions at a lower cost than in the market. In this respect, ‘The most important adaptation to the existence of transaction costs is the emergence of the firm’ (Coase, Reference Coase and Coase1988a: 7). On the other hand, the administrative costs of firm organization may be very high, perhaps exceeding the costs of market organization. For this reason,

Besides market and firm, a further institutional alternative for handling transactions and reducing transaction costs is governmental regulation. Coase (Reference Coase1960) maintained that the government could act as a ‘super-firm’ through its ability to allocate resources by administrative fiat. In particular, the government may impose regulations through, for instance, administrative agencies – in its activity, the government, unlike the firm, also has the monopoly of coercion (by the police) at its disposal. However, nor is the government machine costless. There are governmental failures due, for instance, to incomplete information on benefits and costs, as well as producer or consumer preferences. Then, when the costs of governmental regulation are higher than market costs, the transaction should come back to the market – this was the main argument in regard to broadcasting frequencies put forward by Coase in his article ‘The Federal Communication Commission’ – or carried out within the firm.

Hence, whilst in the case of markets the movement (trans-action) of a resource from an agent Y to an agent X is carried out in a completely decentralized way by price-mechanism, in the firm and in the State this movement or transaction is performed mainly by administrative decisions and hierarchical structures based on authority. Hence, Coase (Reference Coase1937, Reference Coase1960) affirms, there are multiple (not obvious) options for dealing with transactions.Footnote ³ And in regard to the Coasian approach, the economic investigation should explicate the functioning of these different modes of allocation and organization.

Moreover, in 1960, Ronald Coase showed that the essential ingredient of efficient markets – as stated in the First Fundamental Theorem of Welfare Economics – is (the definition and enforcement of) property rights: in the market, the trade of resources relies on alienable rights (that is, property rights) and consists of an exchange of rights on uses deriving from these resources. Like the Old-InstitutionalistsFootnote ⁴ about 90 years ago (e.g., Commons, Reference Commons1924), Coase affirmed that markets require ‘the establishment of legal rules governing the rights and duties of those carrying out transactions in these facilities’ (Coase, Reference Coase and Coase1988a: 10) and the enforcement of these rules in order to secure the agreement of the parties to the exchange. This is the ‘legal Nirvana’ which presupposes clear and enforced property rights, and it leads to the ‘economic Nirvana’ in which the market wholly clears (Nicita and Pagano, Reference Nicita, Pagano, Brosseau and Glachant2008). Furthermore, Coase (Reference Coase1960) demonstrates that there is no (analytical) difference between rights on piece of land and those allowing, for instance, the emission of smoke. If rights are well defined, also the market transaction of rights to smoke could lead to Pareto efficiency. Hence, more precisely defined and easily tradable Footnote ⁵ property rights may improve market exchanges and lead to a Pareto efficient allocation of resources, for every kind of economic resource.

In the Coasian approach, an externality (i.e., the effect of one agent's action on the welfare of another leading to the divergence between the private and social product of such an action) represents the absence of a market instead of the presence of a market failure. Externality has a reciprocal nature and consists essentially of interferences between rival uses of the resource, e.g., the Rancher's use of the land to herd cattle versus the Farmer's use of it to grow crops. Externality problems are ultimately property rights problems and may be solved by defining property rights clearly and by facilitating the trade of resources. Each property right should specify the relevant attributes (including the enforcement) of each use of the resource and the contingencies that characterize such uses. Accordingly, externality problems do not constitute a prima facie case for public intervention (taxation or regulation) as affirmed by the Pigovian tradition – a policy insight that induced Coase to write, quite accidentally,Footnote ⁶ ‘The Problem of Social Cost’.

In the Coasian approach the relationship between externalities and market failure is reversed. The ‘technical’ problem of the externality is not the source of market failure. The opposite is true. The failure of an institution – in this case the market – to make individuals take into account the effects of their decisions on others is a source of an externality. When the market does not fail (in the strong sense that transaction costs are zero) once property rights have been defined, the parties will exchange them, maximizing the joint value of activities.Footnote ⁷

The market would be substituted by other institutions, such as the firm or state-regulation, when one of these institutions is the cheapest way to internalize the externality (and the benefits of internalizing the externality with the cheapest institution outweigh its institutional costs). Thus, as Coase clarified, in the introduction to his essays (Reference Coase and Coase1988a: 14), his contributions to the understanding of the nature of the firm and the nature of social costs are parts of the same complex but unitary vision:

Similarly to markets and to state-regulation, the firm is another way to deal with what most economists had treated as a technical externality problem but Coase re-framed as a problem of comparative institutional analysis.

From the Coasian point of view, economies of scope and scale can be seen as particular types of externalities arising whenever some agents are not rewarded (or not penalized) for the benefits (damages) that the production of some output has on outputs produced by other agents. When these effects take place among qualitatively different units of output, these externalities identify the case of economies of scope. In the case in which the same effects occur among identical units of output, these externalities can be classified as economies of scale. In a costless market, both economies of scope and of scale would be internalized by complete contracts such that everyone would be fully rewarded (or penalized) for the effects that her production activities have on other individuals. In this case, the economies of scale and scope would be fully exploited by the market economy, and firms’ size would not contribute to reaping the fruits of increasing returns. By contrast, if markets are incomplete, large organizations are likely to be necessary to enjoy the advantages of internalizing the effects that the production of one output has on the other. Similar to (and indeed as a particular case of) externalities a consistent Coasian approach must invert the link between market failure and scope-scale economies, changing the issue from a purely technical problem to one including a rich institutional comparative analysis.Footnote ⁸

Coase's strategy involved two steps. The first was to show that, in a world of zero transaction costs, the traditional sources of market failures, such as externalities (including economies of scope and economies of scale), could not exist. The second was to show that the analysis of these phenomena required a new economic approach where costly resources were allocated in a world of costly institutions.

Most economists ignored the second step of his analysis. They were instead fascinated by the world of zero transaction costs to which the Coase theorem applies. According to Coase (Reference Coase and Coase1988a: 15), this was ‘the world of modern economic analysis’ – a strange institution-free world where we have ‘firms without organization, and even exchange without markets’ (Coase, Reference Coase and Coase1988a: 3).

The first step of Coase's inquiry confused with an analysis of the real world. This misunderstanding transformed most economics into a set of useless exercises:

The central issue of economic analysis should instead have been

Paraphrasing Milton Friedman's popular slogan ‘There is no such thing as a free lunch’, meaning that there is no socially relevant human need whose satisfaction is free, the Coasian message is that the issue of institutional substitution arises because no institution is a ‘free lunch’. This insight marked a dramatic and fertile change in the focus of economic analysis. However, it could not be limited to an extension of Robbins’ logic of substitution among costly production inputs. Costly lunches do not only involve the substitution of the items on the menu but also their complementarities. One item on the menu may be (or not be) substituted by another if the complementary ingredients are available (or lacking). One cannot ignore the historical set-up and the interlocking complementarities where institutional substitution takes place. Otherwise, one is forced to assume the existence of costless meta-institutions. In this case ‘institutional free-lunches’ are not really removed. They are simply moved to a higher level Nirvana. To avoid higher-level Nirvana fallacies, one must make a historical analysis of the complementary institutional factors framing a particular institutional substitution choice. As implied by the title of Hodgson's (Reference Hodgson2001) book, extending the analysis to costly institutions entails that economics cannot forget history.

3. Beyond Coase: institutional complementarities

In his pivotal book Markets and Hierarchies, Oliver Williamson succinctly summarized the Transaction Cost Economics generated by Coase's argument as follows:

These two statements recapitulate the novelty of the Coasian view, namely, that economic analysis should include the study of substitution among costly institutions; but they also indicate its limitations: the exclusion of institutional complementarities and history from economic analysis. Indeed, the fact that institutions are costly implies not only that institutional arrangements are substitutes intended to reduce transaction costs, but also that a certain degree of complementarity (or bundling) among institutions may be needed.Footnote ⁹

When the ‘performance’ of one institution is conditioned by and benefits from the presence of another institutions and vice-versa, we speak of institutional complementarity;Footnote ¹⁰ when, instead, the presence of one institution undermines the functioning of another, we speak of institutional crowding out.Footnote ¹¹ For instance, the more a market of production factors clears, the more efficiently a firm produces; the more efficiently a firm produces, the more tax-revenue the State may collect and use to provide better courts for the protection of rights; and the more property rights are protected, the more the market clears; and so on (see Vatiero, Reference Vatiero2009: chapter 4). Put briefly, private ownership, competitive markets and the rule of law often implement highly efficient solutions to allocation problems, but only if all three components are present (see Bowles, Reference Bowles2004: 12).

The notion of institutional complementarity (Aoki, Reference Aoki2001; Milgrom and Roberts, Reference Milgrom and Roberts1990; Pagano and Rowthorn, Reference Pagano and Rowthorn1994) relies on the idea that, in a given institutional framework, economic agents operate in different institutional domains. As a consequence, choices in one domain act as exogenous parameters in other domains and constitute the institutional environment in which institutional choices are made. According to Aoki (Reference Aoki2001), institutional complementarities are situations of synchronic interdependence across distinct institutional domains. In this setting ‘one type of institution rather than another becomes viable in one domain, when a fitting institution is present in another domain and vice-versa’ (Aoki, Reference Aoki2001: 225).

The term ‘fitting’ used by Aoki invokes an evolutionary approach to institutional contexts. Accordingly, a growing body of literature shows that the evolution of economic systems may share some of the complicated intellectual challenges that characterize the Darwinian evolution of natural species in biologyFootnote ¹² (Hodgson, Reference Hodgson1993). As long as institutional complementarities are deep and strong, they can affect, or conceivably determine, the best-fitting institutional arrangements, and that system will differ depending on which local complement dominates.

Analytically, suppose two institutional domains, σ and φ, with sets of agents, m and n, that do not directly interact. Assume that the agents in domain σ face the choice between a rule from either σ^A or σ^B, while agents in domain φ face the choice of a rule from φ^A and φ^B. Supposing a payoff function u, let us suppose the following conditions, for all agents m and n:

(1)\begin{equation} u(\sigma ^A ,\varphi ^A ) - u(\sigma ^B ,\varphi ^A ) \ge u(\sigma ^A ,\varphi ^B ) - u(\sigma ^B ,\varphi ^B )\end{equation}

(2)\begin{equation} u(\varphi^A, \sigma^A ) - u(\varphi^B, \sigma^A ) \ge u(\varphi^A, \sigma^B ) - u(\varphi^B, \sigma^B )\end{equation}

The condition [1] implies that the ‘incremental’ benefit for the agents in σ from choosing σ^A rather than σ^B increases as their institutional environment in φ is φ^A rather than φ^B; the condition [2] implies that the ‘incremental’ benefit for the agents in φ from choosing φ^A rather than φ^B increases as their institutional environment in σ is σ^A rather than σ^B. These two conditions express the idea of complementarity between two different domains. It can be proved that the two strategy profiles (σ^A, φ^A) and (σ^B, φ^B) may be Nash equilibrium profiles (Aoki, Reference Aoki2001, see also Milgrom and Roberts, Reference Milgrom and Roberts1990). When these two equilibria exist, σ^A and φ^A as well as σ^B and φ^B are called ‘institutional complements’. Note that these two equilibria may be ranked in terms of efficiency: for instance, equilibrium (σ^A, φ^A) may be Pareto superior to (σ^B, φ^B). Hence, institutional complementarities may engender, not a tendency towards systemic efficiency, but the emergence of different and (in)efficient economic equilibria.

Moreover, the emergence of multiple equilibria, namely (σ^A, φ^A) and (σ^B, φ^B), means that the convergence towards a certain Pareto-efficient equilibrium will be more complicated in both theory and practice than a simple smooth and expectable change. Indeed, institutional interdependences lead to patterns that Gunnar Myrdal and institutionalists such as William K. Kapp termed ‘circular and cumulative causation’ and are now called ‘path-dependency’. Because of institutional complementarities, small changes may have durable consequences on ‘hybrid’ situations, e.g., disequilibrium among institutional complements such as (σ^A, φ^B) and (σ^B, φ^A). They may set off a circular and cumulative causation process leading rapidly to an equilibrium. On the other hand, big changes may not produce institutional or economic readjustments because of the costs of switching from one equilibrium to another; in this case the initial condition may have persistent ‘lock-in’ effectsFootnote ¹³ – this is one of the (unfortunately implicit and untested) conditions of the so-called legal origins theory (e.g., La Porta et al., Reference La Porta, Lopez-de-Silanes and Shleifer1999).

Institutional complementarity entails that economics, similar to evolutionary biology, is a historical discipline. When institutions are costly, one should consider not only their possible substitutes but also how complementary institutions affect their costs and the costs of the possible institutional substitutes. Whereas in the Coasian approach optimal institutional substitution implies the existence and the convergence to a single optimal equilibrium, institutional complementarities entail that multiple (and possibly inefficient or non-comparable) equilibria can well exist.Footnote ¹⁴

4. The market as a meta-institution: difficulties of the Coasian approach

Also in this section a quotation from Oliver Williamson provides a useful starting point. In his book Markets and Hierarchies, Williamson writes: ‘I assume, for expositional convenience, that “in the beginning there was the market”’ (Williamson, Reference Williamson1975: 20). As this statement shows, a recurrent condition of transaction cost economics is that the market, like a product of Immaculate Conception (Dugger, Reference Dugger, Dugger and Waller1992: 89), is assumed to pre-exist other institutional substitutes (see also Hodgson, Reference Hodgson1988: 177–182). This assumption that in the beginning there was the market limits the analysis of the consequences of costly institutions and involves the puzzling idea that costless meta-markets are available to select other institutions, including ordinary markets. (Meta)-markets are vested with the contradictory status of being both among the costly institutions to be selected and, at the same time, the only costless institutions by which all the institutions are selected. The assumption that in the beginning there was the market means losing large part of the analysis stemming from Coase's insight of costly marketsFootnote ¹⁵ and, in general, of non-free-lunch institutions.

These limitations are particularly evident in the case of institutional rearrangements deriving from technological shocks. In 1937 Coase maintained that technological innovation can change the institutional context; ‘[c]hanges like the telephone and the telegraph, which tend to reduce the cost of organising spatially, will tend to increase the size of the firm’ (Coase, Reference Coase1937: 397). The idea is that ‘[i]nventions which tend to bring factors of production nearer together, by lessening spatial distribution’, that is, by exploiting economies of scale, ‘tend to increase the size of the firm’ (Coase, Reference Coase1937: 397). Furthermore, in a footnote he also underlined that:

In other words, the optimal institutional bundle, consisting of features from different institutions, is recalculated and implemented each time that the technological data change.Footnote ¹⁶ However, such an institutional rearrangement seems to be cost-free, as if some costless meta-markets take care of the transition from markets and firms (and vice-versa). The transition from a market price-mechanism to a hierarchy such as the firm, due to the development of new technologies, occurs without significant costs. Thus, Coasian markets seem to have a double role in the reaction to technological innovations: on the one hand, they represent one of the possible costly institutional substitutes; on the other, they serve as a cost-free meta-institution within which the other institutions are efficiently substituted. The functioning costs of markets limit the size of markets and explain the emergence of firms. At the same time, (meta-)markets justify the emergence of an efficient mix of markets and firms. Their functioning costs do not affect the achievement of an efficient institutional bundle, and they do not constrain the transition, required by technological change, from one organization to the other.

If we take serious account of Coase's intuition that neither markets nor firms can be first best solutions (in the sense that they are both constrained by their own costs of functioning), then the transition from market-type to firm-type organization (and vice-versa) must also be grounded on the analysis of transaction costs. Costly institutions have to be substituted in a costly institutional context. The existence of transaction costs implies the existence of transition costs: in a world of positive transaction costs, transition, or institutional switching, costs must also be positive. With significant transition costs, the institutional bundle may not simply tend to optimally correspond to the level of technological development, but must be heavily influenced by the pre-existing institutional structure of the economy. In other words, high costs of transition may inhibit or reduce potential (efficient) institutional rearrangements even in the presence of significant technological shocks. For this reason, the assumption that in the beginning there was the market neglects the fact that the substitution of institutions may not occur because of transition costs. It fails to understand that each transition occurs in a pre-existing institutional context, which is often characterized by numerous interlocking complementarities.

If one wants to avoid the ‘Nirvana fallacy’ of a zero transaction world, one has to specify the relations of the initial institutional set-up and the costs of the potential institutional changes. In other words, one must move towards a history-dependent economics. A consistent Coasian approach must necessarily lead to some form of ‘evolutionary’ economics where the initial institutional features are necessarily relevant to the explanation and the prediction of the final outcomes. Unfortunately, Coase did not consider these logical consequences of his analysis and failed to consider economics as a historical discipline.

A historical perspective should start from the fact that costly institutions (including the market), where costly resources are allocated, must evolve in a costly manner. Institutional substitution is not determined in a meta-market Nirvana but it occurs within a pre-existing institutional bundle constraining the set and the costs of institutional substitution. New technologies such as the telephone (analyzed by Coase in his Reference Coase1937 article) cannot be the only factors determining the future institutional bundle. Each transition (as well as each stasis of an economic system) is always influenced also by the past mix of institutions and by their interlocking complementarities. As a result, the development of markets and of other institutions is a historical outcome influenced by other complementary domains such as the legal context and politics. A further consequence is that ‘institutions are not necessarily or even usually created to be socially efficient, rather they, or at least the formal rules, are created to serve the interests of those with the bargaining power to devise new rules’ (North, Reference North1990: 16).Footnote ¹⁷

Linking institutions and institutional change only to economic performance and technological changes sets serious limitations on economic analysis. One should analyze institutions with respect to the establishment and evolution of social and political compromises. Institutional change is often the outcome of strategies aimed at improving the situation of some or all components of the dominant elite. Mancur Olson's early contribution ‘The Rise and the Decline of Nations’ and Daron Acemoglu's recent book (with J.A. Robinson) ‘Why Nations Fail’ consider numerous historical examples of how the logic of groups’ collective behavior may favor or damage the growth of nations by changing the institutions of society. The main result of this stream of literature is that historical specificity matters because ‘past institutional choices open up some paths and foreclose others for future institutional development’ (Ostrom, Reference Ostrom1990: 202).