The Diversification of Production
produce more than one product. In this sense their
production is diversified, or horizontally integrated. This paper addresses two questions. First, why have firms become more diversified over the past century? And second, why are diversified firms more oriented toward research and development (R&D) than nondiversified firms?
I tackle these two questions under the assumption that a firm diversifies to maximize its efficiency. Economists have often argued that a firm reaps efficiency gains when it diversifies its production because its managerial and R&D inputs can be shared among its various activities: MOST FIRMS TODAY
The spherein which diversificationis most likely to produceeconomies of scale is researchand develoment. Althoughthe informationthus far gatheredon this question is inconclusive, it is reasonableto say that a for
firmwith a wide rangeof productshas manyopportunities exploiting the results of a programof research. This is because the directions in which researchwill produceresults are to a large extent unpredictable. the
Consequently, greater rangeof activities,the higherarethe chances that a discovery or development in technology will fit into the firms' existing product structure.In this sense, economies are related not so much to size in terms of outputor investmentas to the range of goods I thankthe C. V. StarrCenterfor Applied Economicsat New York Universityfor and
Baily, MichaelGort,Yaw Nyarko,Steve Olley, Peter
Reiss, Mike Scherer,CliffordWinston,andthe discussants usefulcomments,andRay Atje andChungYi Tse for help with the research.Two earlierversionsof this workwere in
grouporganizedby RicardoCaballeroand Andy
presented an economic fluctuations
Caplin,andthis exerciseled to considerable
Brookings Papers: Microeconomics 1993
and services produced.If the level of researchand developmentexpenditures continues to rise rapidly, we can expect an increase in diversiI ficationmotivatedby these considerations.
The idea that know-how can be productively transferred from one activity to another has also been pursued by growth theorists, some of whom argue that spillovers of knowledge among distinct production processes are the engine of growth.2 Nor is R&D the only input that can be shared among the firm's products. Managerial know-how, as well as other indivisible factors, can be used to make, promote, and distribute more than one product at a time.
I shall embed these much-talked-about but still relatively unexplored ideas in two multiproduct versions of Lucas's span-of-control model, which is a general equilibrium model of firm formation.3 Because the trend toward diversification is visible in most sectors despite the technological differences among them, one suspects that these sectors share a reason for this trend, a force that operates at the economy-wide level. The first version of the model isolates the role of the shared managerial input in explaining the trend. The second focuses on the positive relation between diversification and R&D. In both models firms are infinitesimal, they set price of output at marginal cost, and diversification is socially optimal. In fact, firms often have market power, and antitrust policy must then trade off efficiency gains to diversification against its anticompetitive effects. The Justice Department's merger guidelines recognize the possibility that most mergers are not anticompetitive and that they can raise efficiency and benefit the consumer.4
The paper has two main conclusions. First, the secular increase in the capital-labor ratio is a major cause not only of the growth of firm size, but probably also of increased diversification. Second, cross-product spillovers in R&D within the firm seem to be significant. I estimate that a diversified firm gets between 1.025 and 1.3 times as much from 1. The...
Please join StudyMode to read the full document