The Incentive to Invent When a New Input is Involved

Abstract

In a recent paper, Dasgupta and Stiglitz (1980a) have suggested that there is a paucity of microeconomic theory to explain the rate and direction of technical progress and, more particularly, the links between invention, innovation and market structure. There and elsewhere (e.g. 1980b) they have sought to remedy this situation. They have focused their attention on rivalry among firms producing a given product in the R and D activity itself, given various market preconditions, and have analysed situations where research expenditure reduces the expected unit costs or brings forward the date of invention of the given product in a regular manner. In this paper we examine another aspect of the general question. We hark back to an earlier debate (Arrow, 1962; Demsetz, 1969; Kamien and Schwartz, 1970), which looked at an inventor outside the final product industry who has sole property rights in a new process for producing the given product, which is a discrete rather than marginal improvement over the old process, and who may sell to a monopolist or a competitive industry. In our chosen framework the inventor has monopoly control through an input involved in the production process.1 The earlier debate centred around what was meant by similarity of demand conditions between industries of differing market structures. However, we feel that the most policy-relevant comparison is Arrow's original one between two industries with the same 2 demand curve. Theoretically, it is easiest to make consistent welfare comparisons when the area under the demand curve is the same in monopoly as in competition; and practically, assessments of merger proposals in the United Kingdom, for example, involve criteria such as whether a more monopolistic structure will give more or less technical innovation.3 On the other hand, we believe that the assumption of a pure (marginal and average) cost-reducing invention adopted in the original debate was rather naive. Progress often occurs through a particular input, and as relative factor prices change, fixed proportions are unlikely to obtain. The situation we have in mind is as follows. At present, production takes place using input 1 and some others. The invention is a new set of processes (an isoquant) combining a new input, 2, rather than 1, with the other inputs in variable proportions, which enables a Hicks-neutral improvement over the old production function. The question is whether the inventor, in control of the new input, will make greater profits from its use in an industry that is monopolistic or one that is perfectly competitive. For example, if monopoly provides the inventor with more profit we would expect, all other things equal, more invention in areas where user industries are monopolistic and swifter innovation of potentially widely applicable input inventions such as microprocessors into monopolistic industries.