Human Fallibility and Economic Organization

Doctrines concerning what is a good way to organize a society have influenced human societies more deeply than any other set of doctrines. Specifically, beliefs that one way of organizing production and exchange is better than others have inspired a number of socioeconomic experiments leading to modern capitalist and socialist societies, with far reaching implications. Yet surprisingly, the central doctrines, though a source of continuing ideological debate, have been the subject of only limited scientific enquiry. The major proposition, the so-called LangeLerner-Taylor Theorem asserting the equivalence between competitive capitalist economies and decentralized socialist economies which make use of the price system, made a point in stressing that the issue of the ownership of the means of production might not be central in the comparison of economic systems. The theorem, however, was based on models both of capitalism and of market socialism in which the most important differences between the two systems were suppressed. This paper describes a research program attempting to delineate some of the critical differences among alternative forms of economic organization. We contend that central to an understanding of these differences is an understanding of differences in the organization of decision making; of who gathers what information, how it gets communicated and to whom, and how decisions get made, both concerning what actions to take and who should fill decision-making positions. This view should be contrasted with the traditional economic paradigm in which decision making plays no role: the manager, for instance, simply looks up in a book of blueprints what the appropriate technique of production is for the given set of factor prices. In the conventional paradigm, moreover, mistakes are never made, either in gathering or transmitting information, or in making decisions, and indeed, there are no costs associated with these activities. By contrast, the view we take here is that "to err is human," and that different organizational systems differ not only in what kinds of errors individuals make in them, but also in how the systems "aggregate" errors. As a result, organizations differ systematically in the kinds of errors they make, and thus in their overall economic performance. Organizations also differ in the costs associated with information collection, with information communication and processing, and with decision making. Indeed, as we discuss below, perfect decision making can be achieved by arranging enough decision makers in an appropriate manner, no matter how fallible the decision makers are, provided their decisions are not purely random (or worse). What prevents perfect decision making is the cost. We refer to the specification of the structure of information gathering, communication and decision making as an organization's architecture. The objective of our research program has been to construct stylized models of an economic organization within which the consequences of alternative organizational architectures can be examined. Using these models, not only can we compare the performance of particular organizational forms but we can also ask, given a particular set of objectives and circumstances, what is the optimal structure (within tDiscussants: David Kreps, Stanford University; Paul Milgrom, Yale University.