DONG Zhiqiang;LI Weicheng
Key Lab for Behavioral Economic Science & Technology, and South China Research Center for Market Economy, South China Normal University;School of Economics and Management, South China Normal University
The existent literature suggests that the endowment effect is essential for the formation of natural property rights. However, where the endowment effect comes from and how it coevolves with natural property rights is not fully understood. This paper develops an agent-based model featuring “preference evolution + individual choice” to examine how the endowment effect could be shaped and how natural property rights could emerge in an evolutionary artificial society. In our agent-based model, each agent has a pair of preference genes that respectively determine its subjective valuations of the resources it owns or does not own. The agent’s subjective evaluation of resources may be biased or deviate from the true value of the resource, which is defined as the survival value, namely, how much benefit the resources objectively bring to the agent for her survival. Each agent can only survive for a single period, during which some agents will randomly become the owner of a resource (that is, the incumbent), and other unlucky agents become wanderers without any resources. The wanderer needs to decide whether to launch a fight to seize resources from the incumbent. Whether or not a fight breaks out, the agents’ interaction will determine how many resources each party can obtain, and thus the number of descendants the agent can produce. At the end of a period, all agents will die, and their descendants will repeat their stories. Each descendant inherits the gene from its parent; in other words, an agent’s subjective valuation of the resource in both owned and unowned states will be the same as that of its parent. When the model is run, the results show that the emergence of the endowment effect is typically evolutionarily stable, namely, a higher valuation of an owned resource and a lower valuation of an unowned resource is an evolutionarily stable result. With massive simulation experiments, we determined that the endowment effect coefficient (measured by the ratio of subjective valuations of resources when owned and not owned) is around 1.6, and in 90% of the experiments it is in the interval [1, 2.5], which is quite close to the endowment effect coefficients found in 169 behavioral experiments in the 39 projects surveyed by Sayman & ?ncüler (2005). The experimental results of our agent-based model are thus consistent with the WTA/WTP deviation observed in a large number of experimental studies. We also found that the relative value of resources affects the evolution rate and the level of the endowment effect, but has no effect on whether the endowment effect occurs. Furthermore, we found that agents’ subjective valuation of unowned resources tends to approach the true value (that is, survival value) of the resources when the evolution process is long enough, while the agents’ subjective valuation of owned resources is always higher than the true value of the resources, even over a very long process of evolution. These results show that the agents’ valuation is closer to the objective value for unowned resources while higher than the objective value for owned resources. The reason is that the excessive valuation of a possession makes an agent more able to defend it from being snatched away, so this seemingly “wrong” preference improves the agent’s situation and thus makes the endowment effect more likely to evolve successfully as a psychological tendency. Once the endowment effect emerges in the evolutionary process, an agent with this psychological tendency has a solid commitment mechanism, which makes the agent credibly “more willing to pay for the defense of possessions,” and thus others are discouraged from stealing them from the agent. When the robbing intentions of other agents are completely inhibited, natural property rights, or spontaneous respect for property rights, will be pervasive in a given society. This paper not only provides an evolutionary explanation for the origin of the endowment effect, but also provides a theoretical logic and computer simulation evidence for the co-evolution of the endowment effect and natural property rights. Our model disposes of the troubles and confusions resulting from the fact that the property rights phenomenon and the anti-property phenomenon always have the same conditions and the same possibilities; this difficulty has long vexed the classic literature such as Smith (1982) and Gintis (2007, 2009). Thus, our approach provides a more solid theoretical foundation for the institutional economic analysis of spontaneous order.