Applications of a Model of Endogenous Technical Change to US Industry Data

Abstract

Economists have generally accepted the proposition that technical change is partially endogenous to an economic system. Increases in the productivity of inputs, particularly labour, are effected by the responses of firms and individuals to the incentive of possible profits created by allocating resources to the knowledge-generating and knowledge-using activities of invention and innovation. It is hoped that insight into the determinants of the rate of technical change can be gained here by focusing on the links between inventive activity and on the character of the incentives which induce a firm to engage in inventive activity. An intertemporal optimization model incorporating links between the firm's level of research and development activity and increases in its productive capacity is presented. This model posits a production function for the firm consisting of the three arguments: capital services, labour services, and the (firm's) stock of technological knowledge.3 Increases in the stock of technological knowledge are derived from the flow of research services in the firm's employ. The reasons for neglecting non-research inventive activities are essentially pragmatic. First, none of the comparative statics implications with respect to research and development presented subsequently are altered if the relationship between changes in the firm's stock of technological knowledge and inventive activities is reasonably well behaved.4 Second, measures of other inventive activities and their prices are not generally available but data on research and development have been collected for a number of US industries. These data can be used to examine some of the implications of the model. Assuming that the firm's objective is the minimization of discounted future costs of production and research a number of implications with respect to the response of the firm's research and development spending to movements in input prices, research costs, and product demand are derived. These implications are examined empirically with a body of data for eleven United States industries for the period 1959-1966. This presentation is organized as follows. In the first section the essential features of the model and an exposition of the implications of the model are presented. Section II contains the empirical results of applying the model to industry data. A brief summary concludes the investigation.