Williamson, TCE and Open Source

Shortcomings in Transaction Cost Economics: Success of Open-Source and Commons-Based Peer Production

Borrowing from Ronald Coase, “we should be cautious about using stylized but incomplete models of the firm to make empirical generalizations” (Freeland, 63). The problem, however, with literature on transaction cost economics and organizational theory is it does just that. The models are broadly applicable and their tensions provoke thought within organizational theory (Williamson 1996). However, the organizational ideas in transaction cost economics appear to rise from problematic assumptions about the shape of firms, the issues they face, and the possibilities for technology to mollify some of these issues. In other words, transaction cost economics is insufficient to fully explain organizational behavior. To demonstrate this insufficiency, I offer the existence and effectiveness of open-source software and commons-based peer production. The organizational behaviors behind these practices serve as a counterexample to transaction cost economics ideas of firm efficiency, hierarchy and hold-ups. If the hypotheses of transaction cost economics held true in all situations, we should not expect open-source software to exist or operate the way it does. But it does, and it therefore provides a powerful counterexample to the transaction cost economics account of organizational behavior. First: a brief account of open-source and commons-based peer production.

The term open-source, most commonly used to describe software (but also to describe similar forms of production in other industries), or software code created via collaborative effort in which programmers improve the product through personal contributions. They then share these contributions with the community of software developers also working on the code. Open source software developed in response to privately owned software in a burgeoning computer technology industry where privately developed code was seen as an impediment to development (it put obstacles in the way of other developers who had to start from scratch to advance existing technologies if they weren’t able to obtain the proprietary software). Similarly, commons-based peer production, a term coined by Yochai Benkler, is a social mode of production where the effort of large numbers of people is coordinated through the internet toward productive ends without hierarchical organization and with no remuneration for contributors. Wikipedia is probably the best-known example. These two cases—open-source software and commons-based peer production—share a great deal, and most interestingly are situations where the organizational theory thrust of transaction cost economics appears to fall short.

Starting from Coase, we may wonder how such forms of organization may emerge. Coase argues that the firm emerges when there is a cost to using the market’s price mechanisms and where short-term contracts would be unsatisfactory. Under these conditions, the firm organizes around an entrepreneur, a coordinator of productive activity and grows and shrinks corresponding to returns to the “entrepreneur function.” In this light, the firm, which arises out of a desire to save on transaction costs, is also faced with a decision to make or buy. If it decides to produce something on its own, it does so because the costs of acquiring it on the market are higher than producing it within the firm (Coase 1937). Buying on the market indicates the costs of producing do not make production worthwhile. Williamson takes this analysis further in his discussion of firm forms (which appears to be limited to firms with capitalist aims) showing how firms may take different forms based on their increased complexity and how such increased complexity necessitates both a systematic control mechanism and distance between the general management and operations (Williamson 1975). Williamson furthers this analysis of organizational form by arguing that the least hierarchical forms of organization generally have the lowest level of efficiency (Williamson 1985, 231).

It seems, according to the arguments of Coase and Williamson, that open-source and commons-based peer production shouldn’t exist. Regarding Coase’s view on organizational emergence, at least in the short-run, developments of major software platforms would be more costly than contracting for access to existing platforms. There may be two views on why this is. One, the hypothetical decision maker deciding to buy or produce may not truly be in the situation to buy something—it may not be affordable and production may be the only remaining option. Two, this decision maker may be considering more than the immediate instance of wanting a particular software platform from which to build. Perhaps he is also considering future instances. I did not read Coase as accounting for these possibilities. I see the first possibility as necessitating greater consideration of the entrepreneur—who is he? How is he socially embedded? In regard to possibility two, on what time horizon is the decision maker operating? Transaction costs are likely to present themselves differently over different time periods (and this seems is central in considerations of contracts and hold ups). In these ways, Granovetter’s account of structural embeddedness offers a more realistic picture of the economic actors Coase and Williamson describe.

The operation of open-source and commons-based peer production also offer a counterexample to Williamson’s accounts of the form of complex organizations and the efficiencies related to hierarchy. Both are complex operations and neither is governed as an M-Form organization. Both open-source and commons-based peer production rely on a wide net of contributors and have very little hierarchy. Using Williamson’s framework, these systems appear a hybrid of the entrepreneurial and collective forms of organization, specifically a combination of putting out and peer group production. Neither of these, in Williamson’s estimation, is more efficient than the authority relation capitalist form of organization, but in reality these forms of organization can be immensely productive—more so than capitalist authority relation organizations. The fact that private firms are willing to pay employees to work on open-source software from which they collect no direct revenue is indicative of this fact[1]. Here, the case for strong links between productivity and hierarchy seems thin. That is not to say hierarchy isn’t a factor, but it suggests there is more—possibly regarding the social structure and purpose of a particular organization—that drives its efficiency and productivity.

The literature on transaction cost economics is suggestive of a number of issues the firm is bound to face. These issues fall along the lines of dealing with opportunism and malfeasance—particularly with the associated costs of policing and enforcing against these. For open-source and commons-based peer production, there is evidence that the community is self-policing and sabotage is quickly weeded out. Technology enables the “final” product to not be damaged by malevolent opportunists (for example, Wikipedia pages have stored version histories to which the community can revert). Normativity and informal relations, rules and expectations also play a role, possibly in ways that go beyond the control and governance mechanisms suggested by Williamson. In policing and enforcing against opportunism, these forms of organization are faced with transaction costs—but in these specific instances the costs of policing and enforcing appear to be lowered by technology. Even more interesting however is how these particular forms of organization deal with the issues of hold ups obsessed over in the transaction cost literature.

Klein wrote extensively about the problems with hold-ups. When a business relies on assets from another company, it may become dependent on that company’s cooperation in the future. In such a situation, the party holding the key asset in question may be able to hold up the other party by demanding a possibly usurious rate (to use that asset). This is especially a problem when a business considers a large capital investment that presumes future cooperation from the owner of a complementary asset. This problem can be especially bad in the information technology sector where perhaps one software company needs to build using other companies’ proprietary software. The buyer can be held up by the proprietor. If they want to update in the future, this can only be accomplished through coordination with the original software vendor (in order to make changes to their foundation). Thus, it is much less of a risk to rely on open-source software than privately sourced software. Open-source provides a powerful economic alternative. Otherwise, building software on a platform of another company’s software is like building a house on land that you don’t own (as described by Klein).

We can see a divergence between possible hold-up solutions offered in the literature and those that have occurred in reality. From the literature, we see a florid debate around the GM-Fisher Body integration that sheds light on dealing with hold up problems. Coase sees long-term contracts to be the solution. Klein argues that Coase is wrong and long-term contracts are too inflexible and encourage hold-ups and thus vertical integration is the solution. Freeland argues that vertical integration is not a necessarily a solution to hold ups (after all, according to his more detailed historical account, hold ups didn’t occur until after Fisher Body was integrated). What seems evident from this range of views on the same event is that neither vertical integration, nor long-term contracting is sufficient to prevent hold ups. Open-source software manages to skirt all these issues. The software industry does present plenty of cases of contracting and vertical integration, but also exhibits a widespread dependence on open-source solutions. The failure to clearly resolve the argument over whether vertical integration or long-term contracts can prevent hold ups seems to indicate that transaction cost economics takes the wrong approach to understanding socially complex phenomena. Organizations vary greatly and what governs the relations between them and others is a difficult task without full consideration of each party’s interests and needs. What makes the difference between one course of action or another may be a whole host of considerations beyond mere transaction costs. In this sense, Freeland succeeds in exposing this through his account of the GM-Fisher Body integration.


It is unfair in some ways for me to compare contemporary examples—largely rooted in technological advancements—to the less recent examples provided in transaction cost economics. This may serve as evidence of the fact that transaction cost economics fails to fully consider how technological revolutions can change the very nature of how firms and markets interact and provide new solutions to seemingly perennial economic problems. But perhaps it also serves as evidence that transaction cost economics still applies. One way to read the successes of open-source and commons-based peer production is that technology lowers the transaction costs of information and search. It could be said that these particular forms of organization exist due to the lowered transaction costs the present. I don’t believe this is sufficient, however, to justify transaction cost economics as a complete theoretical approach to organizations. The approach’s logic seems to stem from the consideration of transaction costs to the decision to make or buy to decisions surrounding contracting, integration and growth, and to a seemingly inevitable assumption that hierarchy and complexity is an optimal solution. Open-source and commons-based peer production appear to indicate otherwise. What they demonstrate is that individual actors, voluntarily, with no real coordinator stronger than the technological medium through which they interact (the internet) are able to create massive, complicated products, with no direct pay. Transaction cost economics is insufficient to explain this type of organizing. For it to be explained, it’s necessary to go beyond transactions. How are people motivated to contribute to these products? How does trust operate in this system? How are decisions made about directions products should take?

The inability for transaction cost economics to explain all organizational behavior does not consign it to a theoretical dustbin. For one, Coase and Williamson provide a strong framework through which to empirically examine capitalist organizations. This seems a novel step within the history of organizational theory. The transaction cost economics approach to understanding firms, however, is incomplete in its ability to objectively capture reality. Granovetter’s work on embeddedness provides a useful lens through which to view the organization and the many factors beyond transaction costs that impact its behavior. When the social embeddedness of individual and organizational economic actors is brought to bear on the organization, we can better conceptualize organizational forms such as open-source and commons-based peer production. Embeddedness enables us to theorize the organization in a neither under- or over-socialized view, but as an entity subject to both social and economic interests and variations. So, while transaction costs matter, there is more—germane to the specific organizations being analyzed—that must be considered.

[1] This also seems indicative of Cohen and Levinthal’s (1990) point about absorptive capacity.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s