This paper develops a baseline agent-based macroeconomic model and contrasts it with the common dynamic stochastic general equilibrium approach. Although simple, the model can reproduce a lot of the stylized facts of business cycles. The author argues that agent-based modeling is an adequate response to the recently expressed criticism of macroeconomic methodology because it allows for aggregate behavior that is more than simply a replication of microeconomic optimization decisions in equilibrium. At the same time it allows for absolutely consistent microfoundations, including the structure and properties of markets. Most importantly, it does not depend on equilibrium assumptions or fictitious auctioneers and does therefore not rule out coordination failures, instability and crisis by definition. A situation that is very close to a general equilibrium can instead be shown to result endogenously from non-rational micro interaction.

/pdf/agent-based-macroeconomics-a-baseline-model-4aw4mmky59.pdf

Agent-based macroeconomics: A baseline model

Macroeconomic modelling has been under intense scrutiny since the Great Financial Crisis, when serious shortcomings were exposed in the methodology used to understand the economy as a whole. Criticism has been levelled at the assumptions employed in the dominant models, particularly that economic agents are homogeneous and optimizing and that the economy is equilibrating. This paper seeks to explore an interdisciplinary approach to macroeconomic modelling, with techniques drawn from other (natural and social) sciences. Specifically, it discusses agent-based modelling, which is used across a wide range of disciplines, as an example of such a technique. Agent-based models are complementary to existing approaches and are suited to answering macroeconomic questions where complexity, heterogeneity, networks, and heuristics play an important role.

An interdisciplinary model for macroeconomics

Macroeconomic modelling has been under intense scrutiny since the Great Financial Crisis, when serious shortcomings were exposed in the methodology used to understand the economy as a whole. Criticism has been levelled at the assumptions employed in the dominant models, particularly that economic agents are homogenous and optimising and that the economy is equilibrating. In a related paper (Haldane and Turrell Oxford Rev Econ Polic 34(1–2):219–251 2018), we argue that an interdisciplinary approach to modelling in macroeconomics is beneficial. Here we focus on what one such approach - agent-based modelling, which has been extensively used across a wide range of disciplines - could do for macroeconomics. Agent-based models are complementary to existing approaches to macroeconomics and are particularly well-suited to answering questions where complexity, heterogeneity, networks, and heuristics play an important role.

Drawing on different disciplines: macroeconomic agent-based models

There has been much recent interest in studying the implications of economic behavior by simulating markets. For example, the group at the University of Arizona used the PLATO computer network to conduct laboratory experiments with double auctions using human participants. More recent results from a double auction tournament with computer trading programs is reported by Rust et al. Smith gives an overview of the literature on experimental economics. Work at Xerox PARC uses market-based ideas to allocate computer resources (see Huberman and Hogg, Waldspurger et al., for example). The term "computational ecology" has been used to describe this emerging field. Finally, recent work at Cal Tech by Ledyard et al. and Porter and Rangel uses computerized exchange mechanisms to allocate resources in a space exploration project.

A Computational Market Model Based on Individual Action: A Paradigm for Distributed Resource Allocation

Analysing the eThekwini green economy according to its dynamic and complex components for identifying leverage points.

We present an agent-based model of a minimal economy containing households, retail banks, and producers of consumer and capital goods. Household behavior is based on the buffer-stock savings model by Deaton (1961), while the profit-maximizing firms employ reinforcement learning to determine pricing and production. Competitive retail banks facilitate the flow of funds between households and producers through a fractional-reserve system. Stability of the simulated markets depends only on the self-adjusting, boundedly rational behavior of the agents in the completely closed economic system.

/pdf/an-agent-based-model-of-a-minimal-economy-2nhk6y4rwc.pdf

Christopher K. Chan

Papers

An Agent-Based Model of a Minimal Economy