Peter Klein‘s post over at the always excellent Organizations and Markets reminded me that I have been wanting to blog about the most recent exchange between Ben Klein and Ronald Coase over the asset specificity, vertical integration, and the famous Fisher Body – General Motors example which has become a classic example of hold up in the literature. While this example from the original Klein, Crawford, and Alchian (1978) piece is almost 30 years old, and has been the subject of literally thousands of pages of debate and an entire JLE issue (April 2000), there is still some disagreement over the facts and what they tell us about the relationship between asset specificity and vertical integration.
In some ways, this newest part of the exchange is much ado about nothing in so far as it will not add much to our fundamental understanding of the theory of the firm. The holdup theory and the relationship between asset specificity and vertical integration is perhaps the most empirically tested economic propositions of modern industrial organization. On the other hand, there is probably something to learn from the most recent (and most heated in terms of tone) exchange in terms of both (1) what actually happened with the Fisher Body-General Motors relationship, and (2) the process of academic discourse in economics.
Recently, however, the debate (which had been dormant for awhile) got a tad bit uglier. Apparently, Klein had obtained a copy of the original 1919 Fisher Body-General Motors contract which had been previously unavailable and sent the contract and an early version draft of his new paper to Coase who responded with the publication of “The Conduct of Economics: The Example of Fisher Body and General Motors,” which is essentially an attack on Klein’s account and April 2000 JLE response. Coase essentially alleges that Klein made up the Fisher Body story to fit the theory but that a hold up “never happened” because Fisher Body did not actually mislocate plants or adopt an inefficient low-capital production technology which were the two mechanisms for holdup Klein had previously discussed (more on what actually did happen later). Coase goes on to criticize economists more generally for a failure to check facts and propensity to rely on theory alone.
Klein has now responded with “The Economic Lessons of Fisher Body-General Motors“, which has been posted on SSRN and is forthcoming in International Journal of Law and Business. Here is the abstract:
The costs of using rigid, inherently imperfect, long-term contracts to solve potential holdup problems, and the corresponding flexibility advantages of vertical integration, are illustrated by the Fisher Body-General Motors case. The holdup of General Motors by Fisher Body is shown to have involved Fisher renegotiating its body supply contract with G.M. so that, contrary to the original understanding, G.M. made half of the required investments in new body plants. Under the unchanged cost-plus contract terms designed to provide Fisher with a return on its equity capital investments, the decline in Fisher Body’s capital to sales ratio led to a substantial wealth transfer from G.M. to Fisher. G.M. was forced to accept this unfavorable contract adjustment because it desired co located body plants and was operating under a long-term exclusive dealing arrangement designed to protect Fisher Body’s original G.M.-specific capacity investments. The contract adjustment demonstrates the importance of distinguishing between inefficient threatened holdup behavior and the efficient way it is in both transactors’ interests to actually accomplish a holdup. Contrary to Ronald Coase’s recent criticism, this analysis reconciles all the available evidence.
Klein’s new analysis incorporates the previously unavailable 1919 contract to the existing Dupont case record which showed clear evidence (never disputed to my knowledge) of Fisher Body’s expressed refusal to locate plants adjacent to G.M. facilities and evidence of a dramatic increase in Fisher Body’s measured capital to sales ratio after 1922 and demonstrates that Fisher Body leveraged its bargaining position created by GM’s exclusive purchasing commitment to renegotiate a favorable contract adjustment in 1922 prior to agreeing to co-locate plants.
Klein’s analysis adds new evidence (the contract and 1922 contract adjustment) to the factual record and corrects the details of his previous account: it was not mis-locating plants but a threat to mis-locate plants combined with a reduction in capital to sales which resulted in a substantial increase in Fisher Body profits from 1922-26. As Klein writes:
Focusing on the 1922 contract adjustment as the way Fisher Body held up General Motors reconciles all the available evidence. Specifically, the fact that no plants were mislocated and that the contract required Fisher Body to use efficient production technology is fully consistent with Fisher Body’s initial refusal to locate body plants adjacent to GM assembly facilities, the reduction in Fisher Body’s actual measured capital to sales ratio and GM’s complaints about the extra costs it was bearing due to Fisher Body’s reduced capital intensity. Moreover, the contractual adjustment is consistent with the economic theory of holdup behavior, where transactors will attempt to hold up their transacting partner in the most efficient manner.
I will leave to readers to work through both the Coase and Klein articles on their own, but I thought I might share a few reflections on this exchange below the fold.
My biases is this debate should be fully disclosed before I proceed: Ben Klein was my dissertation chair at UCLA, my former employer as a research assistant, and has been a co-author. In addition, as the author’s note to the JLE (2000) paper or this 2007 paper also reveal, I did some research for these papers (ask me anything about cars in the 1920s at your own risk …).
- This debate demonstrates what is good about developing economic theory. The seminal Klein, Crawford, Alchian (1978) work on asset specificity and holdup mentioned the Fisher-Body case as an example to motivate the discussion of transaction costs associated with long term contracts. Subsequent attacks by Coase and others challenged Klein’s account of the facts and how these facts should be interpreted for the theory of the firm. The Klein JLE (2000) response and this 2007 article present a detailed account of the contractual relationship between Fisher Body and GM and increase our understanding of what actually happened. The debate also spurred a voluminous empirical literature testing the association between asset specificity and vertical integration that has confirmed the theory. And while I suspect that Klein’s recent contribution will spur more thought about the Fisher Body-GM example and the theory of holdup more generally, in many ways, this is a battle that has continued to rage on well after the war has ended.
- This debate also demonstrates the bad and the ugly about academic discourse. Accusing another scholar of making up facts is a pretty serious charge. As serious as it gets, really. Perhaps even my bit part in this exchange renders me too close to this debate, having spent thousands of hours searching archives for documents and data from the 1920s, but I cannot figure out how one can read the 2000 or 2007 piece, both which present new and critically important evidence to interpreting the Fisher Body-GM relationship, and conclude that they amount to “a tale.” It is especially unfortunate, sad even, that this exchange has come to that in light of the fact that the asset specificity based theories were designed to build upon the theoretical foundations of Coase’s own transaction cost analysis. Klein himself notes the irony of Coase’s attack on economists who rely excessively on theory: “It is ironic that Coase criticizes economists who accept Fisher Body-General Motors as an example of holdup behaviour for relying too heavily on theoretical analysis independent of the facts when his landmark article on “The Nature of the Firm†(1937) was essentially pure theory. Coase provided us with an important theoretical insight — that organizational arrangements are indeterminate in a world of zero transaction costs — but very little in the way of an empirical answer to the question posed by this insight, namely the type of transaction costs that determine vertical integration and the conditions under which we are likely to observe such transaction costs. I, along with many other economists, have used asset specificity considerations in an attempt to build upon Coase’s theoretical insight. The expanded analysis of the Fisher Body-General Motors case presented here continues in this tradition, placing empirical flesh on Coase’s theoretical skeleton by examining the costs transactors bear when using long-term contracts to solve potential holdup problems.”
- Detailed case studies are empirical work too. I like quantitative empirical work as much as much as the next guy, but with apologies to a few of my economist friends who disdain case studies (you know who you are), the Fisher Body-GM example illustrates what case studies can contribute to empirical work. By working through the Fisher Body-GM example in a number of exchanges, our understanding of what occurred in this case has increased and over time, the theory of asset specificity and vertical integration refined. The theory was developed in this way and created testable implications which were subsequently confirmed in a ton of studies (see, e.g. this survey by Peter Klein, Joskow (1988), Shelanski & Klein (1995), Crocker & Masten (1996) ). A detailed understanding of a contractual relationship such as this one can uncover details that help in the theory building process and to create testable implications which can later be used to run more rigorous quantitative tests.
For those of you who have made it this far, I assume you are interested in contract theory and economic organization or might be, and so highly recommend Gordon Smith’s new work with Brayden King on Contracts as Organizations which summarizes the existing empirical literature and suggests that organizational theories provide a way to understand the functions of contracts other than holdup.
