The influence of experimental and computational economics: Economics back to the future of social sciences

Economics has been a most puzzling science, namely since the neoclassical revolution defined the legitimate procedures for theorisation and quantification. Its epistemology has based on farce: decisive tests are not applied on dare predictions. As a consequence, estimation has finally been replaced by simulation, and empirical tests have been substituted by non-disciplined exercises of comparison of models with reality. Furthermore, the core concepts of economics defy the normally accepted semantics and tend to establish meanings of their own. One of the obvious instances is the notion of rationality, which has been generally equated with the apt use of formal logic or the ability to apply econometric estimation as a rule of thumb for daily life. In that sense, rationality is defined devoid of content, as alien to the construction of significance and reference by reason and social communication. The contradictory use of simulacra and automata, by John von Neumann and Herbert Simon, was a response to this escape of economic models from reality, suggesting that markets could be conceived of as complex institutions. But most mainstream economists did not understand or did not accept these novelties, and the empirical inquiry or the realistic representation of the action of agents and of their social interaction remained a minor domain of economics, and was essentially ignored by canonical theorizing. The argument of the current paper is based on a survey and discussion of the twin contributions of experimental and computational economics to these issues. Although mainly arising out of the mainstream, these emergent fields of economics generate challenging heuristics as well as new empirical results that defy orthodoxy. Their contributions both to the definition of the social meanings of rationality and to the definition of a new brand of inductive economics are discussed. JEL: B41, B52, C71, C72