Abstract: Pre-pandemic private sector of the U.S. economy consisted of more than 120 million employees in some 6 million businesses, generating some $30T of gross output. Using micro-data on all those workers and firms I will describe a computational model realized at full-scale with the actual economy, capable of reproducing dozens of gross features and macro-regularities present in the data on American companies. Both computational and data challenges associated with this work will be discussed
Bio:
Rob Axtell works at the intersection of the computational, social, behavioral and economic sciences. His research group combines agent-based computing with micro-data to build large-scale models having high verisimilitude with the real-world. They have worked on a variety of policy issues, from housing to fisheries, behavioral aspects of retirement and science policy. He has been visiting Professor at the University of Oxford (UK), the New School for Social Research (NY), and Middlebury College (Vermont). His research has been published in leading general interest journals ("Science," "Nature" and "Proceedings of the National Academy of Sciences"), in field journals (e.g., "American Economic Review," "Economic Journal", "Computational and Mathematical Organization Theory," "Journal of Industrial Ecology"), and reprised in newspapers (e.g., "Wall St. Journal," "Washington Post") and technology publications (e.g., "Scientific American," "Technology Review," "Wired").
Forthcoming from MIT Press
Dynamics of Firms from the Bottom Up: Data, Theories and Models
Robert Axtell, Krasnow Institute for Advanced Study, George Mason University
In this new book Professor Axtell uses data on the universe of all U.S. firms to first uncover several dozen gross empirical regularities that characterize American firms, from heavy-tailed distributions of sizes, growth rates, and productivities, to skewed age, job tenure and wage distributions, large labor turnover and high levels of firm entry and exit. He then makes the case that extant economic theories do not explain these many patterns. Such theories tend to focus on static equilibria and represent heterogeneous human behavior in terms of a few representative agents. The book goes on to present a family of simple agent-based computational models that reproduce most of the main empirical features of the U.S. data. Results are reported for 120 million interacting agents who self-organize into 6 million firms, 1-to-1 scale with the U.S. private sector, the first agent model at full-scale with an actual economy. This groundbreaking new work will be of interest to economists working in industrial organization, to financial professionals involved in the valuation of companies, to computer scientists working on group formation in multi-agent systems, to operations researchers interested in agent-based modeling and simulation technologies, and to organization scientists who want to learn how to build empirically-rich, behaviorally-grounded models of business firms.
Robert Axtell is Professor of Computational Social Science at George Mason University and External Faculty Fellow at the Santa Fe Institute. He has been a visiting professor at the University of Oxford, MIT, Johns Hopkins, the New School for Social Research, and Middlebury College, and was for many years Senior Fellow in the Economic Studies Program at the Brookings Institution. His previous MIT Press book, Growing Artificial Societies: Social Science from the Bottom Up, co-authored with Joshua Epstein (NYU) is a citation classic. He earned a Ph.D. from Carnegie Mellon University.