Agent-Based Modeling for Studying the Spontaneous Emergence of Money

A central question in economics is how a society accepts money, defined as a commodity used as a medium of exchange, as an unplanned outcome of the individual interactions. This question has been approached theoretically in the literature and investigated by means of agent-based modeling. While an important aspect of the theory is the individual’s speculative behavior, that is, the acceptance of money despite a potential short-term loss, previous work has been unable to reproduce it with boundedly rational agents. We investigate the reasons for the failure of previous work to have boundedly rational agents learn speculative strategies. Starting with an agent-based model proposed in the literature, where the intelligence of the agents is guided by a learning classifier system that is shown to be capable of learning trade strategies (core strategies) that involve short sequences of trades, we test several modifications of the original model and we come up with a set of assumptions that enable the spontaneous emergence of speculative strategies, which explain the emergence of money even when the agents have bounded rationality.