Agent-Based Computational Economics

Agent-Based Computational Economics

C. Bruun (Aalborg University, Denmark)
DOI: 10.4018/978-1-59140-984-7.ch014
OnDemand PDF Download:
No Current Special Offers


This chapter argues that the economic system is best perceived as a complex adaptive system, and as such, the traditional analytical methods of economics are not optimal for its study. Agent-based computational economics (ACE) studies the economic system from the bottom up and recognizes interaction between autonomous agents as the central mechanism in generating the self-organizing features of economic systems. Besides a discussion of this new economic methodology, a short how-to introduction is given, and the problem of constraining economics as a science within the ACE approach is raised. It is argued that ACE should be perceived as a new methodological approach to the study of economic systems rather than a new approach to economics, and that the use of ACE should be anchored in existing economic theory.

Key Terms in this Chapter

Macrofoundation: If for example methodological individualism is violated in a composite system, simple aggregation is no longer possible, and studying macroproperties directly may be the only way to learn about certain aspects of the system. This may be characterized as the macrofoundation of the system.

Agent-Based Computational Economics: Acknowledging the economy as a complex adaptive system, agent-based computational economics (ACE) is the computational study of economies as collections of interacting agents. By letting autonomous agents interact in silico , the overall aim is to represent self-organizing features of real-world economic systems in a more realistic way than has hitherto been possible. But ACE may be applied to many different aspects of economics and should be perceived as a new methodological approach to the study of economic systems rather than a new approach to economics. Focus may be placed on the individual agent and its interaction with its surroundings (e.g., learning by the agent), or on the system resulting from the interaction of many agents (e.g., learning by the system).

Methodological Individualism: Methodological individualism’s purpose is to explain and understand macro phenomena as the aggregation of decisions by individuals—the whole is nothing but the sum of its parts. This may also be described as reductionism since methodological individualism allows the whole to be reduced to the sum of its parts, and thus the parts that make up the system may be studied in isolation. Methodological individualism is an essential part of general equilibrium theory and in the demand for microfoundation in economics.

General Equilibrium Theory: As agent-based computational economics (ACE), general equilibrium theory is a bottom-up approach to the study of economic systems. But in opposition to ACE, general equilibrium theory makes use of methodological individualism and reductionism. Taking as its starting point endowments, preferences, and technologies of economic agents, general equilibrium theory uses an auctioneer to calculate a vector of prices that will clear all markets. Methodological individualism allows general equilibrium theory to regard preferences as given, and the existence of an auctioneer prevents trading at false prices, since agents are only allowed to interact through the auctioneer. A number of fundamental problems arise when the ideal general equilibrium model meets the real world. Among these is the problem of making room for money in a setting where there is no need for a medium of exchange and where any good can take the role as unit of account.

Keynesian Macroeconomics: Keynesian macroeconomics takes as its starting point aggregate magnitudes and their interrelations. Of pivotal importance to the ideas of Keynes is his definition of national (money) income as the earnings of the factors of production, which must be equal to the cost of production. From this follows the equivalence of saving and investment; both are defined as the excess of income over consumption. This definition implies that consumers will always hold enough money to buy the product. However, this does not imply that supply creates its own demand in the sense that lack of aggregate demand cannot cause unemployment (Say’s law). Entrepreneurs decide the level of production, and they will not produce beyond their own investment demand plus expected consumption demand. If, at the level of full employment, consumers in the aggregate are expected to save more than entrepreneurs wish to invest, entrepreneurs will reduce production until expected savings equals expected investment. By, in this way, arguing in terms of aggregate magnitudes, Keynes depicts a different economy than the one known from neoclassical or general equilibrium theory—an economy where unemployment due to a lack of aggregate demand is not only possible but also likely.

Rational Expectations Hypothesis: Rational expectations hypothesis is the idea that economic agents use all available information, including information on economic relations, in forming expectations about the future. Rational expectations may also be denoted “model consistent expectations” since consistency between the economic model and the model used by agents in expectation formation is required. Economic agents are just as well informed as the economist (or even econometrician) building the model.

Microoundation: Making use of methodological individualism, the claim for microfoundation states that any macroeconomic phenomena must be shown to result from individual (rational) actions.

Complete Chapter List

Search this Book: