Limitations on the Perspective of Representative Economic Agent: Agent Based Model’s Alternative

Please cite the paper as:
João Victor Souza da Silva & Solange Regina Marin, (2017), Limitations on the Perspective of Representative Economic Agent: Agent Based Model’s Alternative, World Economics Association (WEA) Conferences, No. 2 2017, Economic Philosophy, Complexities in Economics


Economic Science emerges from the questions about the behavior human, conflict and social order. Smith, preoccupied with the complexity of social relations and economic phenomena, point out the individual as basis of Political Economics. For the author, social and economic order are maintained in function of the singular behavior of subjects in internal conflict between the search for self-interest and the necessity of social acceptance. A superficial lecture of his work, however, gave the basis of the development of neoclassical economic approach and the simplification of the individual and economic order, with objective to provide predictions about the phenomena, ordered by general and simplistic laws of operation. The homo economicus, caricature of the selfish subject of Smith, with unbounded rationality, homogeneous and self-interested, is the central of the economic order, static, in equilibrium and universal, like the perfect world of Newton, symbol of modern science. However, the simplistic model of individual as central to the neoclassical theory and economics mainstream implicates in difficulties to apprehender the economic phenomena, on a complexity and real world. The agent-based model consider the Economy as complex and dynamic system, in constant evolution, based on the mutative and interactive economic behavior. Through computer simulations, this model generate scenarios based on the change of rules in function of the adaptation and evolution of interactive individuals. Furthermore, consider the individual bounded-rationality and heterogeneity, unbalanced and instable environment, with maintenance of mathematical instruments, search precision, objectivity and robustness. This work aims to present the formation and evolution of Economic Science and the figure of representative economic agent, basis of the Neoclassic Economic. Next, is exposed the agent-based model as methodological alternative for Economic Sciences. Its recognizes the limitations of the economic representative economic model and the incipience of the complexity perspective, however, is fundamental the dialogue and review of economic basis, in reason to best comprehend the economic phenomena on a complexity and real world.

Keywords: , , ,

Recent comments


2 comment

  • johndavis says:

    I believe this entry point, the nature of the agent in a complex world, is an important focus, unfortunately somewhat overlooked by many commentators in this conference. The essential intuition that the authors investigate is how complex systems, as contrasted with equilibrium systems, require agents constantly able to adjust their behavior. That idea is at odds with the idea that agents are rational optimizers. How can you optimize if the world is highly dynamic? One thing that worries me, then, is that some researchers who take complexity seriously think that you can simply insert rational optimizers into a complex world and explain economic behavior. I think this is a mistaken view, but until researchers make some effort to develop an alternative agent conception appropriate to a complex world, it seems complexity researchers will still operate with one foot in an old neoclassical world that they otherwise reject. This paper is very helpful in redirecting attention to the issue of what agents are in a complex world.

  • João Victor Souza da Silva says:

    Dear Davis,
    Thank you very much for your consideration. This is a problem that greatly disturbs me, about the need for a rational maximizing agent. Linear models of equilibrium are based on centrality in the rational, static and universal subject. What then grounds complex systems if not the conception of another type of agent, not maximizing, adaptive? What is the need to seek the optimization of the agent in a model that is characteristically far from a static, macro-level optimal point?

    My great quest for understanding complex systems is to seek to understand the role of the agent in this process. But is another conception of subject enough to guide understanding of economics as a complex system? How are ABM / ACE models useful for methodologically and epistemologically reshaping Economic Science?