The Competitiveness of Nations in a Global Knowledge-Based Economy

Harold Demsetz

The Firm in Economic Theory:  A Quiet Revolution

American Economic Review, 87 (2)

May, 1997, 426-429.



I. The Existence and Importance of Firms
II. Abandoning the Presumptions of Neoclassical Theory

III. Ownership



Neoclassical theory’s objective is to understand price-guided, not management-guided, resource allocation.  The firm does not play a central role in the theory.  It is that well known “black-box” into which resources go and out of which goods come, with little attention paid to how this transformation is accomplished.  In the theory’s core model, perfect competition, the transformation accords with the dictates of known technology and prices.  Management has no real influence.  Nonetheless, the firm in this theory, and the household too, serve an important objective, that of conceptualizing an economy in which there is extreme dependency.  Production takes place in firms.  Consumption and resource supply take place in households.  People must depend on each other, and self-sufficiency is thus barred.  Multi-person firms (and households) are not needed to play this role.  A farm is a firm, whether owned and worked by a single person or by many, if its crop is produced for the market rather than for those who have produced it.

This is quite different from the firm in R. H. Coase’s 1937 classic on the nature of the firm in which managed coordination, presumably involving more than one person, defines the firm.  Prime objectives of his article are to explain the existence of firms and their importance, relative to the price system, in the allocation of resources.  The difference between these views may be examined in regard to these objectives.


I. The Existence and Importance of Firms

In Coase’s (1937) analysis, firms come into existence when the cost of consciously managing inputs (to achieve a given outcome) is less than the cost of using the price system.  The cost of using the price system is not clearly described by Coase, although he refers to the costs of acquiring price information, negotiating, and exchanging.  To interpret Coase in a manner consistent with his objectives, these costs should be interpreted as those of any activity undertaken to use the price system.  Analogously, management cost should be interpreted as those of any activity undertaken to manage consciously the use of resources.  The cost of a telephone call can be a cost of using the market in some circumstances or, in others, a cost of managing; so can the cost of transacting or of contracting.  Negotiating to purchase a good from another firm gives rise to a cost of using the price system; negotiating to set the limits of employee work assignments gives rise to a cost of managing.  This interpretation is consistent with contemporary discussions of governance and contracting, but to avoid the confusion that accompanies terms such as these and “transacting,” I henceforth refer to PSC and MSC as the costs, respectively, of using a price system and of using a management system.

In the perfect-competition model, PSC and MSC are both zero.  This is because prices, goods, and technology are all freely known.  Therefore, the comparison that Coase makes in judging the relative importance of managed and price-guided resource allocation cannot be relevant to firms and markets in the perfect-competition model.  Yet, in that model, there are firms and markets.  They complement each other rather than, as in the views of Coase and contemporary theory, competing with each other.  Firms, the sole producers of goods and services for sale to others, require conceptual “places” (i.e., markets) in which entitlements to these goods and services are exchanged, but even the service of exchanging is produced by firms.  Markets reveal exchange opportunities but do not produce goods or services; they can not substitute for firms in this respect.  The implicit substitute for the firm in this theory is self-sufficient production.

Coase seeks the bright line that separates managed coordination from price-guided coordination.  Neoclassical theory, were it to


search for an analogous bright line, would seek that which separates production for others from self-sufficient production, or in more conventional terminology, that which separates specialization from self-sufficiency.  The relative importance of these turns on the advantages that come from scale economies, economies of repetitive task performance, comparative advantage, and so on.

If one now lets PSC be positive, the neoclassical comparison responds quite differently from the way Coase’s comparison responds.  Neoclassical theory must view PSC as an impediment to specialization and as a cause of diminution in its importance to the economy.  This means diminution in the importance of firms and the price system (i.e., the market).  Specialized production is sold to others across a PSC barrier, so its advantage over self-sufficiency is diminished the larger is PSC; the hypothetical lower limit of PSC, zero, maximizes the range of activities in which specialized production (i.e., the firm) dominates self-sufficiency.  For neoclassical theory, positive MSC, on the other hand, only provides an incentive to favor single-person firms over multi-person firms.

Coase’s view, and contemporary theory’s also, must be the opposite of this.  Zero PSC eliminates the need for managed coordination, at least if there is positive MSC (or if management provides weaker incentives than do market prices), and price-guided organization dominates managed organization; the importance of firms is diminished.  Positive PSC, on the other hand, increases the relative importance of firms (i.e., of managed coordination).  These conclusions clash with those of neoclassical theory because managed coordination increases as the vertical depth of the firm increases.  The firm manages the production of more of what it uses.  But this can also be accompanied by horizontal shrinkage in the size of the firm.  This will normally happen if PSC increases.  This horizontal shrinkage means less specialized production for others, and it is this that is interpreted by neoclassical theory as a diminution in the importance of firms.  And so it is. Coase’s theory correctly deduces that an increase in PSC leads to vertical integration and correctly sees this as a reduction in the scope of market activity in terms of the “make or buy” decision, but it neglects the impact of a larger PSC on the importance of specialization in the economy.  If PSC is made prohibitively high, ending all transactions with others, the firm will have vertically integrated with the household and become self-sufficient.  For neoclassical theory, the self-sufficient unit is the “non-firm,” whereas Coase and contemporary theory see reliance on prices as the non-firm or as the market-based contractual form.

But what does use of the price system mean?  It would seem to mean self-interested owner allocation of resources in accordance with market prices offered and not in accordance with directions given by other persons.  Direction by others would constitute managed coordination.  So the non-firm for Coase and contemporary theory is really self-direction of one’s resources in response to prices.

However, in the view of neoclassical theory, self-direction in response to prices does not create self-sufficiency, which is its non-firm.  Self-sufficiency means production obtained from one’s own resources but not offered to others and, therefore, not undertaken in response to price signals.  The one-person firm entails no management of the activities of others, so it is a non-firm for Coase and contemporary theory, but the one-person production unit can sell to others.  This it does to the fullest possible extent if PSC is zero, and if it does this, it qualifies as a firm in the view of neoclassical theory.  If PSC is prohibitively high, self-sufficiency reigns supreme, and firms and prices have no role to play in the economy.

The different views I have been discussing are directed toward understanding different phenomena.  Neoclassical theory is focused on specialization, not on managed coordination.  Coase’s theory is focused on managed coordination, not on specialization.  Contemporary theory has a still different emphasis.  Its concern is mainly with agency problems, but, it is more closely related to Coase’ s theory than to neoclassical theory because its focus is on optimal mixtures of market-based incentives and management-based controls.  Nonetheless, both have propositions concerning the existence and importance of firms and firm-like contracts.  In regard to these propositions, the neoclassical view seems to me to be superior to the views of Coase and contemporary theory.


II. Abandoning the Presumptions of Neoclassical Theory

If contemporary theorizing about firms shares the weakness of Coase’s views on the existence and importance of firms, it has nonetheless improved knowledge of institutional arrangements in other respects.  This it has done by rejecting the assumptions used by neoclassical theory in its core model of the competitive firm: (i) markets function freely, (ii) prices and technology are known by all interested parties, and (iii) owners are effective in controlling the use of their assets.  In their place, contemporary theory substitutes positive information cost.  This creates a productive role for management where none exists in neoclassical theory.  The firm in neoclassical theory reflects the imperatives of the price system, not those of management; if the price system works well, resources are allocated well.  Imperfect information, in contrast, makes the judgment of managers and owners a source of productivity enhancement.  The main source of management’s productivity in contemporary theory has been in its response to agency problems.  Shirking, opportunism, and reputation are brought to the fore.  Understanding of these, and of business practices responding to them, has improved rapidly during the last quarter century.  This is properly a source of pride for those who have worked on these topics.

However, the focus in this effort has led to the neglect of information problems that do not involve agency relationships.  These are associated with planning in a world in which the future is highly uncertain, and they include problems of product choice, investment and marketing policies, and scope of operations.  Agency problems often intertwine with these but are not central to them.  Neglect of this class of problems is unfortunate.  Their solutions offer different explanations of business practices than those derived from agency relationships.  Should agency problems be the primary guide to understanding vertical integration?  Perhaps, but consider, for example, the role of “expertise.”  Specialization is productive in regard to acquiring knowledge.  Forsaking reliance on a specialized body of knowledge and using multiple kinds of knowledge instead raises the cost per unit of knowledge acquired.  The manufacturer of commercial aircraft possesses specialized knowledge that can be extended at little cost into the business of maintaining older aircraft.  Vertical integration of these two businesses is, therefore, a practical possibility.  Vertical integration of aircraft manufacturing and airline transportation, on the other hand, is more difficult.  Effective vending of airline transportation to the general public requires operational and marketing know-how not normally part of the knowledge needed to manufacture aircraft.  The difficulty this poses does not depend on agency problems.  This is readily seen by supposing that the firm comprises but a single person, or if composed of many persons, by supposing that no one engages in opportunistic behavior.  Agency problems are absent, but the advantages of avoiding operations requiring different kinds of knowledge remain.

The focus on agency relationships has also led to neglect of some useful neoclassical theorizing.  The successive-monopoly problem, price discrimination, and price controls, as examples, offer motives for vertical integration not involving agency problems.  Of more general importance to business organization, is the neglect of the impact of prices themselves.  No firm can long survive a management that persists in using capital-intensive methods of production and of monitoring behavior if the price of capital keeps rising relative to the price of labor.  How much of the variation in management techniques or contractual form is explained by agency problems and how much by relative prices?  The work of contemporary theorists suggests that they believe agency problems are the important source of variation, but I wonder if this is so.


III. Ownership

Neoclassical theory has no serious treatment of business ownership.  This is not surprising in a theory whose core model presumes full knowledge and, therefore, no risk.  Risk becomes relevant if information is imperfect.  A problem arises for would-be controlling owners, that of having “too many eggs in one basket,” or of bearing too much firm-specific risk.  The problem is more severe if a firm must be large to compete effectively, but it can be ameliorated by acquiring equity capital from many persons rather than from a few.  Joint ownership, with its panoply of control problems, re-


sults.  The ownership structure of the firm and the ownership of assets by the firm become phenomena in search of explanation.

The relationship between diffuseness in the ownership of the firm and variables such as firm size and regulation has been and is being studied.  There is no need to refer to these studies here, but behind their logic lies productive roles for both management and ownership.  There are two primary sources of ownership’s productivity.  One is to provide and guide the flow of equity capital.  The other is to control the activities of management.  These involve a control-risk trade-off.  A “tighter” ownership structure increases control of professional management but raises the firm-specific risk borne by suppliers of equity.  This trade-off depends not only on the variables currently being studied, but also on a society’s wealth and the distribution of same.  A wealthy person can take a large position in a firm’s ownership structure, yet enjoy diversification of his total wealth.  An egalitarian distribution of wealth, precisely because it deprives a society of wealthy people, undermines even a wealthy society’s ability to put owners in effective control of professional management.  No single family or group of, say, five families would possess enough wealth to own a substantial fraction of the equity of a firm that must be large to compete effectively.  An extremely unequal distribution of wealth also exacerbates the entrenched management problem.  The time and expertise available to those few who are very wealthy are too limited to allow them to control effectively the many enterprises they own.  Between these extremes is a wealth distribution that is best suited to accommodate the control-risk trade-off problem; the wealthier is a society, the more moderate can be the inequality in this distribution.

Contemporary theorizing about the firm indicates that wealth and its distribution matter to society’s productivity because of the constraints these put on the ownership structure of firms; in this new way, they have an impact on the effectiveness with which business assets are managed.  Barring foreign equity capital, a poor nation that follows an egalitarian income policy cannot have firms that are effectively controlled if they operate in markets where large size is important to competitive success.  Equity for such firms can be supplied only through a diffuse private-ownership structure or through an even more diffuse public-ownership structure.  Both make the control problem more severe than it would be if resources had gone into markets in which large size is not a competitive advantage.  The implications of this for economic growth are already reflected in recent work (D. Lal and H.Myint, 1996).

Consider now the firm’s ownership of assets.  If a firm can be small and compete, its assets can be owned by a single person or by very few.  In this case, the firm’s operations impinge directly on the reputation of the owner(s).  The ownership structure of a large firm, however, must be more diffuse in order to ameliorate the firm-specific risk problem.  If each of many owners of a small share of equity could insist on the return of his pro rata share of the firm’s assets, as he can, for example, in dealing with an open-end mutual fund, the firm could be forced to disgorge assets.  This would create a great deal of uncertainty for customers and suppliers if these were to find that long-term relationships with the firm are important, for the overhang of the threat to the firm’s assets makes the firm’s ability to honor such relationships problematic.  Shareholder reputation cannot play as large a role here in assuring customers and suppliers as it does in the small firm, because no single shareholder is undermining the firm’s ability to honor long-term agreements when asking the firm to repurchase his relatively small share of the firm’s assets.  To prevent disgorgement of assets, it is desirable for the firm itself to own the assets once they are acquired with equity provided by shareholders.  Accordingly, under present corporate law, shareholders in the general case have a claim on dividends, can sell their shares to others, and can vote on important corporate matters, but they cannot insist on a pro rata return of the value of a going firm’s assets.



Coase, R. H. “The Nature of the Firm.” Economica, November 1937, 4(16), pp. 386-405.

Lal, D. and Myint, H. The political economy of poverty, equity, and growth. Oxford: Clarendon, 1996.