The Transactions Cost Approach to Vertical Integration: An Empirical Examination

Abstract

The transactions cost approach developed by Coase and Williamson provides a coherent framework for investigating determinants of vertical integration in different industries. Empirical implications are developed and then tested using a cross section of firm level data pooled over different time periods. The results tend to confirm hypotheses regarding internal costs of management, small numbers bargaining problems and notion of firm as suited to adaptive sequential decision making under conditions of uncertainty. IT rHILE there exists an extensive literature on VY theoretical rationales for vertical integration, surprisingly little is known about importance of different theories.' The empirical literature includes two types of studies. Case studies test applicability of a particular theory to a single firm or industry (e.g., Armour and Teece (1980), Perry (1980) or Monteverde and Teece (1982a)). They are susceptible to bias in their choice of industry and generality of their results cannot be established. Another category of studies examines a broad base of industries to determine relationship between vertical integration and such variables as industry concentration, average firm size and sales growth (e.g., Adelman (1955), Gort (1962) and Tucker and Wilder (1977)). The link between these studies and theories of vertical integration is generally left unclear.2 This study will adopt strategy of investigating a broad base of industries, but, unlike earlier studies of this type, will test empirical implications following from transactions cost approach.3 This approach as developed by Coase (1937) and Williamson (1975, 1979) provides a coherent framework for investigating determinants of vertical integration over different industries. Furthermore, Williamson has persuasively argued that transactions cost considerations underlie such prominent reasons for vertical integration as elimination of monopoly distortions, technical complementarities, supply reliability and economies in acquisition of information. The transactions cost approach is summarized and implications are developed in first section of this paper. The implications are tested using firm level data pooled over time. The empirical model is presented in second section and results are presented in third section. The model incorporates variables not included in previous studies of vertical integration, such as measures of market risk and organizational structure. The results tend to confirm hypotheses regarding internal costs of management, small numbers bargaining problems and notion of firm as an institution suited to adaptive sequential decision making. A summary of results and conclusions are contained in final section. I. The Transactions Cost Approach Arrow (1969) has defined transactions costs as the cost of organizing economic system. In attempting to translate this idea into a framework for explaining organization of economic activity, Williamson and others have focused on role of opportunism and limited capability of individuals in processing information. The choice of institutional alternative depends on minimizing costs which arise in presence of transaction-specific investments and uncertainty. While a wide variety of institutional alternatives exist, a simple dichotomy is adopted here for sake of tractability. Following Coase, transactions are classified according to whether they take place in firm or across markets. Market alternatives become hazardous in recurring exchanges involving transaction-specific capital and efficient information processing. The firm then provides a Received for publication November 21, 1983. Revision accepted for publication December 3, 1984. *Rutgers University. This paper is based on my dissertation. I would like to thank my committee, Armen Alchian, Harold Demsetz, David Mayers, Fred Weston and my chairman, Benjamin Klein, for their helpful comments. Recent versions of paper also benefited from comments made by K. Chung, D. Carlton, D. Kaserman and anonymous referees. I would also like to thank participants of Transactions Cost Workshop at University of Pennsylvania. The usual disclaimer applies. 1 Recent summaries of theoretical literature on vertical integration are in Warren-Boulton (1978) and Kaserman (1978). See Levy (1981, 1984) for discussion of this problem. 3Earlier studies by Armour and Teece (1980) and Monteverde and Teece (1982(a, b)) have successfully employed transactions cost approach to examine vertical integration in particular industries. This study differs in that it adopts transactions cost approach to test implications across industries.