ABSTRACT
Industrial structure is really about two interrelated but conceptually distinct systems: the technology of production and the organizational structure that directs production. These systems jointly must solve the problem of value: how to deliver the most utility to ultimate consumers at the lowest cost. Industrial structure is an evolutionary design problem. It is also a continually changing problem, one continually posed in new ways by factors like population, real income, and the changing technology of production and transaction. It was one of the founding insights of transaction-cost economics that the technological system does not fully determine the organizational system (Williamson 1975). Organizations – governance structures – bring with them their own costs, which need to be taken into account. But technology clearly affects organization. This is essentially Chandler’s claim. The large-scale, high-throughput technology of the nineteenth century “required” vertical integration and conscious managerial attention. In order to explicate this claim, we need to explore the nature of the evolutionary design problem that industrial structure must solve. Like a biological organism, an organization confronts an environment
that is changing, variable, and uncertain. To survive and prosper, the organization must perceive and interpret a variety of signals from the environment and adjust its conduct in light of those signals. In short, organizations are information processing systems. This is no less true of early nineteenth century production networks than it is of an Internet-enabled firm of today: in a real sense, the economy has long been a knowledge economy. Also like biological organisms, business organizations differ in the mechanisms they use to process information and to deal with variation and uncertainty. Nonetheless, as James Thompson (1967, p. 20) argued, all organizations respond to a changing environment by seeking to “buffer
environmental influences by surrounding their technical cores with input and output components.” Understanding the ways in which organizations buffer uncertainty is thus crucial to understanding organizational structure. In Thompson’s discussion, buffers seem to take many forms. The “input and output components” he refers to are various kinds of shock absorbers mediating between a highly variable environment and a more predictable production process. Inventories are a classic example: they can ebb and swell with changes in demand or supply while allowing a smooth flow of product. But Thompson also mentions preventive maintenance, which reduces the number of unplanned outages, as well as the training or indoctrination of personnel, which reduces variability in human performance.1
