In contrast to mainstream economics, complexity theory conceives the economy as a complex system of heterogeneous interacting agents characterised by limited information and bounded rationality. Agent Based Models (ABMs) are the analytical and computational tools developed by the proponents of this emerging methodology. Aimed at students and scholars of contemporary economics, this book includes a comprehensive toolkit for agent-based computational economics, now quickly becoming the new way to study evolving economic systems. Leading scholars in the field explain how ABMs can be applied fruitfully to many real-world economic examples and represent a great advancement over mainstream approaches. The essays discuss the methodological bases of agent-based approaches and demonstrate step-by-step how to build, simulate and analyse ABMs and how to validate their outputs empirically using the data. They also present a wide set of applications of these models to key economic topics, including the business cycle, labour markets, and economic growth.
This book is a collection of essays which examine how the properties of aggregate variables are influenced by the actions and interactions of heterogenous individuals in different economic contexts. The common denominator of the essays is a critique of the representative agent hypothesis. If this hypothesis were correct, the behaviour of the aggregate variable would simply be the reproduction of individual optimising behaviour. In the methodology of the hard sciences, one of the achievements of the quantum revolution has been the rebuttal of the notion that aggregate behaviour can be explained on the basis of the behaviour of a single unit: the elementary particle does not even exist as a single entity but as a network, a system of interacting units. In this book, new tracks in economics which parallel the developments in physics mentioned above are explored. The essays, in fact are contributions to the analysis of the economy as a complex evolving system of interacting agents.
This book arose from our conviction that the NNS-DSGE approach to the analysis of aggregate market outcomes is fundamentally flawed. The practice of overcoming the SMD result by recurring to a fictitious RA leads to insurmountable methodological problems and lies at the root of DSGE models’ failure to satisfactorily explain real world features, like exchange rate and banking crises, bubbles and herding in financial markets, swings in the sentiment of consumers and entrepreneurs, asymmetries and persistence in aggregate variables, and so on. At odds with this view, our critique rests on the premise that any modern macroeconomy should be modeled instead as a complex system of heterogeneous interacting individuals, acting adaptively and autonomously according to simple and empirically validated rules of thumb. We call our proposed approach Bottom-up Adaptive Macroeconomics (BAM). The reason why we claim that the contents of this book can be inscribed in the realm of macroeconomics is threefold: i) We are looking for a framework that helps us to think coherently about the interrelationships among two or more markets. In what follows, in particular, three markets will be considered: the markets for goods, labor and loanable funds. In this respect, real time matters: what happens in one market depends on what has happened, on what is happening, or on what will happen in other markets. This implies that intertemporal coordination issues cannot be ignored. ii) Eventually, it’s all about prices and quantities. However, we are mostly interested in aggregate prices and quantities, that is indexes built from the dispersed outcomes of the decentralized transactions of a large population of heterogeneous individuals. Each individual acts purposefully, but she knows anything about the levels of prices and quantities which clear markets in the aggregate. iii) In the hope of being allowed to purport scientific claims, BAM relies on the assumption that individual purposeful behaviours aggregates into regularities. Macro behaviour, however, can depart radically from what the individual units are trying to accomplish. It is in this sense that aggregate outcomes emerge from individual actions and interactions.