Abstract. Alongside increasing returns, network externalities, and information cascades, information contagion has been presented in the literature as an explanation for particular patterns of macrobehavior that may seem at odds with the underlying micromotives. We present an Agent-based Computational Economics model that could provide a microfoundation for information contagion. Our model exhibits self-organization through information contagious behavior, and the emergence of spontaneous orders, in which typically most agents choose the same, superior item. But it turns out that this is not a simple monotonic process from disorder to order. Instead, the system continually moves back and forth between order and disorder as the self-organization is a continuing story in which the emerging order unravels time and again. In other words, information contagion is an inherently complex phenomenon.

J.E.L. classification codes. D11, D83, O33

Keywords. Agent-based Computational Economics, Decentralized interaction, Reinforcement learning, Information aggregation, Self-organization