ICC Advance Access originally published online on August 21, 2009
Industrial and Corporate Change 2009 18(5):785-806; doi:10.1093/icc/dtp013
| ||||||||||||||||||||||||||||||||||||||||||||||||||
Is the division of labor limited by the extent of the market?: evidence from the chemical industry
Correspondence: Ashish Arora, Fuqua School of Business, Duke University and NBER, Cambridge, MA, USA. email: ashish.arora{at}duke.edu
Correspondence: William B. Vogt, RAND Corp, Pittsburgh, PA, USA and NBER, Cambridge, MA, USA. email: william.b.vogt{at}gmail.com
Correspondence: Ji Woong Yoon, Department of Public Administration, School of Politics and Economics, Kyung Hee University, Korea. email: jiwoongy{at}khu.ac.kr
In the age of outsourcing, it is easy to forget that outsourcing is simply one manifestation of the division of labor. Adam Smith's dictum that the division of labor is limited by the extent of the market has created difficulties when applied to a division of labor among firms (rather than within a firm). The problems are both for analytical attempts to formalize it and for empirical attempts to test it. Bresnahan and Gambardella show that the Smith–Stigler theorem holds when the extent of the market is defined in terms of the number of users instead of simply the total size of demand; therefore, division of labor is increasing in the number of users and decreasing in the average size of users. This article provides an empirical test of Bresnahan and Gambardella's theoretical argument, using data from the chemical industry. The chemical industry shows systematic variation across technologies and countries in the extent of the division of labor in plant design and engineering. We develop an empirical model in which large firms decide whether to build plants using in-house resources or to contract out, and small chemical firms also decide whether to invest in a plant. The number of specialized suppliers of plant design and engineering services (SEFs) vary with the demand for their services. The empirical results support the predictions of Bresnahan and Gambardella. We find that the number of SEFs increases when the market expands through an increase in the number of potential buyers but not when the market expansion is due to an increase in the average size of buyers. Moreover, an increase in the share of large-firm investment decreases small-firm investment, which decreases the number of SEFs. In turn, this further depresses small firm investment.