Transaction Cost in Economics

Transaction cost economics acknowledges that technology and ownership of assets arc both important, but it maintains that neither is determinative of economic organization, nor are both together. Rather, the study of economic organization has to go beyond technology and ownership to include an exam- ination of incentives and governance. Transaction cost economics maintains that the transaction is the basic unit of analysis and gives special emphasis to the study of governance.

Thus, even holding technology constant, three things happen when a transaction is transferred out of the market and is placed under unified ownership: Ownership changes, incentives change, and governance structures change. The first—ownership change—occurs by definition. Even if the formal incentive rules (c.g. transfer pricing) arc held constant between firm and market, the effective incentives change as a consequence of a change in asset ownership. Accordingly, the formal rules are apt to be adapted. New governance structure will appear in either event to support the integrity of the internal exchange relation. All of the above, moreover, will vary as a function of the nature and degree of asset specificity. Plainly, the study of economic organization is a much more complex undertaking than a production function formulation contemplates.

1. Prospective Applications 

As is evident from earlier chapters, transaction dost economics applications have been made in the fields of industrial organization, labor economics, and the study of the modem corporation. Those scarcely require recounting here. Of greater interest are applications made in other areas.

One of the more obvious and natural of them is the application of trans- action cost economics to comparative economic systems. Stephen Sacks’s recent book (1983) on Yugoslav self-management is an illustration. As Hor- vat’s 1972 survey of Yugoslav economic reforms in the postwar period discloses, the links between the microanalytics of capitalism and socialism are numerous and important. Sacks’s treatment confirms that. But as he points out, a great deal remains to be done.

To be sure, linkages can be discovered at several levels. Koopmans, for’ example, regards the “pre-institutional character” of activity analysis as one of its attractions: “Technology and human needs are universal. To start with just these elements has facilitated and intensified professional contacts and interactions between market and socialist countries” (Koopmans, 1977, pp. 264-65). I submit, however, that exclusive reliance on technology and human needs can foreshorten the inquiry. The study of human needs is usefully joined with the study of human nature. Additionally, albeit difficult to orchestrate, contacts between capitalist and socialist countries regarding the study of governance structures—with attention to both similarities and differences—hold considerable promise for deepening our understanding of those matters.

Applications of transaction cost economics to business history also hold out considerable promise. This is not to suggest that successive organizational innovations should be assessed exclusively in this way. But viable modes of economic organization—those that endure, are imitated by rivals, diffuse to other sectors, are successively refined and perfected, and do not depend on the political process for protection against alternative modes—ordinarily possess an efficiency advantage.

Other applications on which headway has already been made but for which further inquiry is warranted include the study of family organization (Ben-Porath, 1980; Poliak, 1983), and nonprofit forms of enterprise (Hans- mann, 1982; Fama and Jensen, 1983).

An area in which Klein and his associates have made significant headway10 and for which more is in prospect is the economics of the middleman: the merchant, the dealer, the franchisee. A vast number of contractual irregularities that occur at that level of organization appear to have the purpose and effect of economizing on transaction costs.

An area to which transaction cost economics has made only limited contributions but that holds out considerable promise is public finance. To be sure, aspects of defense contracting have been examined in terms that are akin to the transaction cost approach (Williamson, 1967a). And the difficult problems of information that have a bearing on R&D policy have also been examined in semi-microanalytic terms (Arrow, 1971; Nelson, 1984). But if the choice and design of institutions is what public finance is all about, then countless applications of transaction cost economics have yet to be made. Problems of incentives and governance are enormously difficult in a political context. Tolerance for greater variance in relation to private sector efficiency assessments is likely to be needed.

2. Research Needs 


Given the requisite preconditions, quasi-market and internal modes of organization realize governance benefits in relation to autonomous market trading. The high-powered incentives of markets are unavoidably compromised, however, when a transaction is placed under unified ownership. As between the two, this book gives disproportionate attention to the governance as compared with the incentive features of capitalist modes of economic organization.11

Attention to both, however, is essential if the economic institutions of capitalism are to be accurately assessed. Among other things, ways by which to enhance the incentive efficacy of internal organization—by effecting semi- decomposability (which is a leading purpose of the M-form structure), by surrounding operating rules and procedures with credible commitments, thereby to enhance reliance—require concerted study. Comparative institutional assessments of proposals to enhance incentive .efficacy will be realized only when the relevant microanalytic details of market, quasi-market, and administrative modes are set out. Partly that is a conceptual exercise. Partly it is empirical. An enormous amount of work plainly needs to be done.


Reputation effects will deter defection from the letter and the spirit of an agreement in the degree to which (1) defections can be made public knowledge, (2) the consequences of defection can be fully ascertained (which will permit, among other things, real versus contrived claims of defection to be distinguished), and (3) parties who experience or observe defection penalize the offender and/or his successors in “full measure.”

None of those conditions is easy to satisfy. With respect to the first, it is costly to advertise defections. Even if a simple announcement could be costlessly made, moreover, it is further necessary to supply the details. Is it a bogus claim? What is the magnitude of the damages? Did the plaintiff take appropriate steps to mitigate the damages? What were the alternatives and when did they become known? Information at this second level becomes enormously costly to supply and evaluate.

Fly-by-night operators aside, the third condition guarantees that full penalties will be extracted from offending parties. A firm under one ownership/management cannot escape the penalties assigned to an earlier ownership/management by asking for forgiveness. Instead, the sins of the fathers are assuredly visited on the sons—in which event the asset valuations of an enterprise will continuously reflect prior behavior {Kreps, 1984).

The issues here go to the behavioral attributes of human actors and are discussed further in section 4. Suffice it to observe here that reputation effects are no contracting panacea. The limits as well as the powers of those effects need to be studied (Carmichael, 1984).


Any discussion of the economic institutions of capitalism that does not deal with final product markets is egregiously incomplete. This is a very large and complicated topic. An informed assessment of final product market practices will require a great deal of detailed knowledge of those practices. Although 1 am confident that the approach herein developed has considerable generality and am furthermore encouraged that my opinion is shared by others,12 an application to final product markets is beyond the scope of this book.

Strong commonalities notwithstanding, the correspondence between in- termediate product markets and final product markets is inexact. Some of the differences are attributable to the differential ease with which hierarchical organizations can relieve bounds on rationality as compared with small groups (families) and individuals.

Purchasing and contracting functions in large organizations can be and commonly are assigned to specialists who are deeply knowledgeable of the technical, market, and contractual features of each of the many goods and services bought and sold. Information asymmetries between the parties are greatly relieved as a consequence. Individual consumers, by contrast, are unable to delegate in the same way and therefore rely much more on market signals to infer product attributes.

Branding and advertising serve signaling purposes. But the signals can be and sometimes are used strategically, which complicates the welfare assessment. The possibility that consumers can be provided with more reliable, compact, economical signals warrants sympathetic study. Truth in lending is in that spirit. Can egregious cases be identified for which truth in advertising efforts are warranted?

A further feature of the consumer market that warrants comment is the evident incapacity of the average individual to make probabilistic choices in a consistent way. The biases to which large numbers of individuals are subject in dealing with probabilistic matters have been documented repeatedly (Tversky and Kahneman, 1974). They are especially evident when low-probability events are being evaluated (Kunreuther et aL, 1978). That most individuals possess those biases and limitations does not, however, imply that most organizations will also display them. If more and less cornpelent proba- bilists can be distinguished, and if responsibility for procesjsihg and displaying the consequences of probabilistic choices is concentrated on the more competent types, then economies of specialization will be realized. But whereas organizations (corporations) can effect such specialization easily, individual consumers are much more limited. Again, the possibility of public policy intervention to yield (on average) improved consumer decisions when con- fronted with probabilistic choice suggests itself. Insurance is an obvious candidate.

Source: Williamson Oliver E. (1998), The Economic Institutions of Capitalism, Free Press; Illustrated edition.

Leave a Reply

Your email address will not be published. Required fields are marked *