Using Agent-Based Models to Understand the Surprising Complexity of Global Markets

Abstract

Why is the economy “complex”? For all the recent advances in the management sciences, businesses today face markets full of unexpected events, from cascading failures of lending institutions to speculative investment and the explosive growth of tech firms. In this talk, I argue that the complexity sciences, particularly a simulation methodology known as agent-based modeling, will help key decision-makers anticipate the global dynamics of modern consumers, firms and markets. The talk demonstrates that many of the “surprises” we observe in today’s economy arise from our poor understanding of the relationship between what Nobel laureate Thomas Schelling called “micro motives and macro behavior”—that is, how individual decisions produce social structural outcomes. Our thinking tends to suffer from two fallacies: of composition (inferring structural properties from observing the behavior of individuals) and the ecological fallacy (inferring individual attributes from observing social aggregates). The talk illustrates how agent-based modeling helps us bridge the divide between individual agency and social structure. Using examples from economics and politics, I illustrate the four principal causes of complexity in today’s business climate: interaction effects, strategic complexity, ecological complexity, and reflexive complexity.