Neither Modularity nor Relational Contracting: Inter-Firm Collaboration in the New Economy

In a series of recent essays, including their contributions to this symposium, Richard N. Langlois and Naomi R. Lamoreaux, Daniel M. G. Raff, and Peter Temin (hereafter LRT) present interesting but contradictory views of the decentralized or vertically disintegrated post-Chandlerian economy from whose vantage point they seek to reconceptualize business history. Starting from an orientation that uneasily combines Adam Smith's ideas about the division of labor with organizational learning, Langlois sees the current situation as dominated by the modularization of production. This modularization and the arm's-length transactions it facilitates create a world reminiscent of the antebellum United States, although today's high-throughput differentiated exchanges are underpinned by a set of market-supporting institutions, notably standard interfaces or design rules. Starting from an orientation toward Oliver Williamson and the minimization of coordination costs, LRT in contrast see a world of collaborators joined by long-term, largely informal relations of a distinct type reducible neither to markets nor hierarchies.¹

In this comment, we will argue that these views are both off the mark. LRT are right to argue that the post-Chandlerian new economy, far from being limited to market exchanges among black-box makers, is deeply collaborative. However, we will show that this collaboration is directed at innovative learning in a way consistent with Langlois's general orientation (though not with his particular arguments about modularization or his specific interpretation of Adam Smith). Such collaboration is also more formalized, and formalized in a different way, than the repeated-game, reputational transacting framework adopted by LRT.

Our argument proceeds in three steps. In the first part, we briefly characterize the contrasting views of Langlois and LRT with an eye toward the internal tensions and gaps in their positions. In the second part, we focus on two salient, complementary features of the post-Chandlerian new economy that figure prominently in contemporary managerial discussions but that are at odds with both accounts on offer. The first of these features counts against the Langlois view on modularization. It has been shown to be impossible to establish standard design interfaces so comprehensive and stable that customers and suppliers can in effect interact as if operating in spot markets for complex components or subassemblies without jeopardizing their own long-term survival. The second feature, which counts against the LRT view, is the profusion of innovative disciplines and practices of co-design such as simultaneous engineering, benchmarking, co-location of personnel, problem-solving teams, processual quality standards, and the like. By routinely obligating collaborators to question and clarify their assumptions about their joint project, these disciplines allow for a corrigible partitioning of tasks within and across firm boundaries in ways that fixed modular interfaces do not. But such practices also formalize cooperation in ways that are inconsistent with the informal, "relational ties" at the heart of the LRT view.

In the third part, we set these controversies about the post-Chandlerian new economy in the context of broader debates about the interpretation and direction of business history. We underscore those common features in the views of Langlois and LRT that give rise to their symmetrical (over- and underformalized) misinterpretations of contemporary developments. Both sets of authors explicitly disavow historicism, understood as the thesis that history is determined by the unfolding of an immanent tendency or deep logic. In that sense, they purport to break with Chandler, who saw the rise of the modern corporation as a functional response to the imperatives of markets and technology. But disclaimers aside, Langlois is openly committed to what might be called multidimensional historicism: the view that historical development is determined by the interaction of two or more independent but simultaneously operating logics or tendencies. Moreover, he is arguably correct in presenting his view as an explication and clarification of the related view of LRT—though given the ambiguity of their...