The Make-or-Buy Decision: Lessons from Empirical Studies

Abstract

The “transaction cost” theory of the firm introduced by Coase (1937) has become a standard framework for the study of institutional arrangements. The Coasian framework helps explain not only the existence of the firm, but also its size and scope. Why, in Coase's (1937, pp. 393-94) words, “does the entrepreneur not organize one less transaction or one more?” Some firms are highly integrated: IBM, for example, produces many of its components and software and maintains its own sales force for mainframe computers. Others are much more specialized: Dell Computer outsources virtually all its hardware and software components, selling directly to end users through its catalog and website, while the shoe company Reebok owns no manufacturing plants, relying on outside suppliers to make its products. U.S. manufacturing and service companies are increasingly contracting with specialized information technology firms for their computing and data warehousing needs, spending $7.2 billion on outsourced computer operations in 1990. Standard and Poor's estimates total worldwide outsourcing for 2003 at $170 billion.1 Why do some firms choose a vertically integrated structure, while others specialize in one stage of production and outsource the remaining stages to other firms? In other words, should a firm make its own inputs, should it buy them on the spot market, or should it maintain an ongoing relationship with a particular supplier? Traditionally, economists viewed vertical integration or vertical control as an attempt to earn monopoly rents by gaining control of input markets or distribution channels. The transaction cost approach, by contrast, emphasizes that vertical coordination can be an efficient means of protecting relationship-specific investments or mitigating other potential conflicts under incomplete contracting. As transaction cost economics was developed in the 1970s and 1980s, a stream of empirical literature emerged explaining the “make-or-buy decision” using transaction cost reasoning. (The traditional approach has generated relatively few empirical applications beyond analyses of particular antitrust cases.) This chapter surveys the empirical literature on vertical boundaries, focusing on the transaction cost approach and emphasizing the most important results, while highlighting the challenges that remain.