Sourcing by design : product architecture and the supply chain

Abstract

The first paper presents a model of the supply organization problem in the automotive industry that characterizes both ownership of and access to productive assets. Contracting between a manufacturing firm and a parts supplier is modeled using an incomplete contracting framework that synthesizes those of Grossman-Hart (1986) and Rajan-Zingales (1998). Optimal assignments of ownership and access at key milestones in the product development process are analyzed with respect to their incentive effects on non contractible investment. The merits of this approach are argued using case study evidence from an original five year study of parts contracting in the automotive industry. By incorporating quantifiable measures of design complexity from the product development literature, this model also lends itself to more refined testing of vertical integration behavior in a broad range of industries. The second paper focuses on the coupling of product architecture and sourcing decisions using empirical evidence from the auto industry. The emphasis is on two decisions that a firm must make: whether to manufacture products in-house or to outsource - the make/buy decision, and on product complexity. There has been much research into product architecture and the make/buy decision in the auto industry, yet to the best of the authors' knowledge these decisions have never been studied jointly. The central hypothesis is that increasing complexity in product architecture will drive vertical integration, in that a firm seeking to minimize the coordination costs associated with developing a complex part will internalize production. This proposition is consistent with transaction cost theory regarding asset specificity. Product architectural complexity is proposed as a better proxy for asset specificity than those used previously in empirical testing of transaction costs. Original data is used to test an empirical model of the link between product architectural complexity and make/buy decisions within the context of automobile systems development. Results provide evidence of complementarity between product architectural complexity and vertical integration, as well as evidence of clustering within the auto industry around high performance combinations of the two choice variables. This has implications for optimal incentive structures within firms as well as for interpreting the performance outcomes of firms.