Harold Demsetz 1930-2019

Harold Demsetz has passed away. He was a longtime member of the UCLA economics department and one area he wrote on was the theory of the firm. Here he took an anti-Coaseian view of the firm. Ten years ago I wrote on Demsetz's view:

In an provocative and interesting paper written for the Markets, Firms and Property Rights: A Celebration of the Research of Ronald Coase conference, Harold Demsetz articulates a somewhat non-standard view of the firm as it appears in the neo-classical model of the competitive economic system. A common attack, one which I would make, on the neo-classical model is that it has no theory of the firm in it. It is a theory with firms but without a theory of firms. The model has no theory of the structure and methods of firms and no theory of why a firm should even exist in an economy in which prices are treated as the sole guides to the opportunities available for putting resources to work. As Foss, Lando and Thomsen (2000: 632) summarise it:

"The pure analysis of the market institution leaves almost no room for the firm (Debreu 1959). Under the assumption of a perfect set of contingent markets, as well as certain other restrictive assumptions, the model describes how markets may produce efficient outcomes. The question how organizations should be structured does not arise, because market-contracting perfectly solves all incentive and coordination issues. By assumption, firm behaviour (profit maximization) is invariant to institutional form (e.g. ownership structure). The whole economy can operate efficiently as one great system of markets, in which autonomous agents enter into very elaborate contracts with each other. However, by treating the firm itself as a black box, where internal structure, contracts, etc. disappear from the picture, there are many other issues that the theory cannot address. For example, the theory does not tell us why firms exist".

Demsetz takes issue with all of this. He argues that the neo-classical model offers both a definition of the firm and a rationalisation for the existence of firms, but, he notes, these are mostly implicit. For Demsetz the concern of the neo-classical model is with the function of the firm. The function is what defines the neo-classical firm: this function being the production of goods and services for purchase by persons who, in the main, are not involved in the production of what they buy. The firm is an institution specialised in production for use by others. The rationale for the existence of firms in this theory is implicit in this function. They exist because specialisation is productive.

But what form does this institution being called the firm take? Demsetz explains that the internal organisation of the firm is largely irrelevant to the neoclassical theory of a private, decentralized economic system. If the firm has no internal organisation, then how does production take place? The view of the firm Demsetz is arguing for is one in which the theoretical task served by the firm is that of creating a grand coordination problem whose resolution is sought in the price system. He argues this view of the firm is very different from that of Ronald Coase. For Coase the firm stands in contrast to the market, it becomes less vertically integrated and less important in the economic system the smaller are the costs of using the price system, ie lower are the costs of using the market. For Demsetz the firm stands in contrast to self-sufficient production within households. The firm becomes more important the lower is the cost of using the market. This is simply because low transaction costs facilitates the substitution of specialised production destined for others for self-sufficient production for oneself. The lower is the cost of using the price system the more effectively firms can bring their products to consumers and the more effective are households in supplying inputs to specialised producers.

I think this still leaves the question of the institutional structure that production takes place in unanswered. If production does not take place via Coase-type firms, how does it take place? Is it, as the Foss, Lando and Thomsen quote above would suggest, taking place via groups of autonomous agents who enter into very elaborate (complete) contracts with each other. That is, production takes place over the market. If this is so it would, in turn, make the firm look, perhaps unsurprisingly, like a "nexus of contracts".

The nexus of contracts view was developed in papers by Alchian and Demsetz (1972), Jensen and Meckling (1976), Barzel (1997), Fama (1980) and Cheung (1983). The important innovation here was to see that it is difficult to draw a line between firms and markets, firms are seen as a special type of market contracting. What distinguishes firms from other forms of market contract is the continuity of the relationship between input owners.

Most famously in the Alchian and Demsetz version of this approach, they argue that the authority relationship between the employer and employee is in no way the defining characteristic of a firm. The employer has no more authority over an employee than a customer has over his grocer. "Firing", of either the employee or grocer, is the ultimate punishment that either the employer or customer can use in cases of "disobedience". Alchian and Demsetz argue that, in economic terms, the customer "firing" his grocer is no different from the employer firing his employee. In both cases one party stops dealing with the other, terminating the "contract" between them. In this approach the firm is seen as little more than a nexus of contracts, special only in its legal standing and characterised by long term nature of the relationship between the input owners. In this approach it is not generally useful to talk about firms as distinctive entities, a nexus of contracts could be called more firm-like if, for example, the residual claimants belong to a concentrated group but the term "firm" has little meaning beyond this.

Would such a burring of the lines between markets and firms make sense within the neo-classical model? Given the implicit use of complete contracts in the neo-classical approach I would argue that it does. Under complete contracts any institutional form is able to mimic any other institutional form. A well known example of this is the neutrality theorems of the theory of privatisation literature, under whichstate-ownership and private ownership of firms results in the same outcome. In the case of the neo-classical model, production via the market would achieve anything that production via (Coase-type) firms could achieve and thus there is no need for (Coase-type firms in the model.

But I'm not sure we are left with a role for a firm of any type in the neo-classical model. If the firm is defined by its function, production for outsiders, in a zero transaction cost world it is not clear to me that we have any form of firm at all. What we have is production via the market, which doesn't result in a real meaning being given to the term "firm".

In the neo-classical model, then, I think we are left with having to take production for outsiders as the definition of the firm, without knowing how that production takes place, or we have to say that the model doesn't really have firms in it at all. The grand coordination problem is resolved by the price system, by the market, but without recourse to firms. Demsetz's definition of the "firm" and his rationalisation for their existence is, if I understand it correctly, a definition and rationalisation for production, but production via the market, without firms.

If we accept this position, then we also have to ask, What meaning can we give to the "household" in the neo-classical model?

Of course it could all just be semantics.

Update: Thinking about this a bit more, I wonder if perhaps another way to see the point I’m trying to make is to follow the Alchian and Demsetz argument a bit further. As noted above Alchian and Demsetz argue that it is meaningless to try to draw a hard line between markets and firms. They go on to say that the reason that firms are special has to do with the technology of team production., that is, production where the individual production functions are inseparable from one another. This gives raise to the problem of free riding since team production can act as a cover for shirking. Alchian and Demsetz’s answer to this principal-agent problem is to appoint a monitor to oversee the workers. So in this framework team production is the underlying reason that the "nexus of contracts" can be considered a firm. But there are no team production or principal-agent problems within the neo-classical framework. So while the neo-classical firm may be a "nexus of contracts", it is a "nexus of contracts" without team production and thus there is no rationale for firm based production as opposed to market based production.

References:

• Alchian, Armen and Demsetz, Harold (1972 ). 'Production , Information Costs, and Economic Organization', American Economic Review, December, v. 62, iss. 5, pp. 777-95.
• Barzel, Yoram (1997). Economic analysis of property rights Second edition. Political Economy of Institutions and Decisions series. Cambridge; New York and Melbourne: Cambridge University Press.
• Cheung, Steven N.S. (1983). 'The Contractual Nature of the Firm', Journal of Law and Economics, 26(1) April: 1-21.
• Fama, Eugene F. (1980). 'Agency Problems and the Theory of the Firm', Journal of Political Economy, April, v. 88, iss. 2, pp. 288-307.
• Foss, Nicolai J., Henrik Lando and Steen Thomsen (2000). 'The Theory of the Firm'. In Boudewijn Bouckaert and Gerrit De Geest (eds.), Encyclopedia of Law and Economics, Volume III, Cheltenham U.K.: Edward Elgar Publishing Ltd.
• Jensen, Michael C. and Meckling, William H. (1976). 'Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure', Journal of Financial Economics, October, v. 3, iss. 4, pp. 305-60.