For best experience please turn on javascript and use a modern browser!

Substituting Complements

Event details of June 13, 2005: Giuseppe Dari-Mattiacci (Universiteit van Amsterdam)
Date 13 June 2005
Time 11:45 -13:15


The recent case brought against Microsoft has drawn attention to a problem seldom arising: competition in the provision of complementary goods may be worse than monopoly. Some proposed the break-up of the firm as a way to reduce its market dominance and hence enhance competitiveness hindered by it. The proposed break-up would have taken place along vertical lines, and the resulting smaller-scaled firms would have produced the operating system and software applications.

The result would have been similar to cases discussed by Cournot in 1838 and, somewhat differently, by Charles Ellet in 1839 in his study on internal improvements in the United States. Divided ownership on a single line of transportation (a railway for example) causes the owners of individual segments of it to price the segment higher than a monopoly owner would do, thus making quantity demanded shrink and partially appropriating, partially destroying consumer surplus. This result is driven by the fact that owners of complements cause a negative externality to the remaining owners when they increase their price, as the quantity sold by everyone plums. Since they earn the full benefit while bearing only part of the cost of their pricing decisions, suppliers of complements will overprice their products. A monopolist, or a cartel of individual producers, will instead bear both the above cost and benefit and hence choose a lower price.

This externality problem, which is similar to a prisoners’ dilemma, has been recently re-discovered and re-labeled as the anticommons problem and has shown to be particularly powerful a tool in analyzing issues related to the fragmentation of physical and intellectual property rights. From this perspective, fragmenting a single resource into complementary portions causes externality in exclusion (the one mentioned above), while a traditional commons problem is affected by externality in use. The legal problems to which the anticommons theory had been applied have the advantage of relying on uniqueness or quasi-uniqueness of the complements.

Unique and not-substitutable are the votes of the 5 members of the United Nations Security Council (if unanimity is the rule); (quasi)unique are the patents that a developer of a derivative innovation needs to hold in order to proceed with his research; unique are the property fragments that a developer need to by to put the land to a different use. The point that the Microsoft break-up solution raises is indeed that when the complements are not unique, there could be competition in the production of each of them, thus counteracting (or solving) the anticommons problem. Several firms may produce competing applications compatible with the operating or the many interchangeable operating system or, to put it in Ellet’s rhetoric, the could be several alternative routes to each segment of a fragmented line of transportation.

The question that still needs to be asked is: How much competition in the supply of each complement do we need in order to solve the anticommons problem completely?


Oudemanhuispoort 4-6
1012 CN Amsterdam