An agent-based model of repeated decision making under risk: modeling the role of alternate reference points and risk behavior on long-run outcomes
Arpan Jani
DOI: https://doi.org/10.1007/s11573-021-01048-7
2021-05-31
Journal of Business Economics
Abstract:Prospect theory suggests that decision makers facing uncertain outcomes tend to be risk-seeking when choosing between losses but risk-averse when choosing between gains. Whether an outcome is framed as a gain or a loss depends on the reference point considered. While multiple reference points could be relevant to decision makers in a particular situation, prospect theory does not specify which reference point will be salient other than the status-quo. Social comparison theory proposes that individuals often compare their performance with relevant others, and hence this could become a reference point. An agent-based model (ABM) was developed drawing upon these two theories. In each time step of the simulation, agents invested a proportion of their current wealth in one of three investment options with varied risks, resulting in increase or decrease of their wealth. Agents then compared the outcome of their investment, i.e., their new wealth, with a reference point to determine whether the outcome was a gain or a loss. The ABM investigated the consequences of adopting five alternate reference points for evaluating an outcome as a gain or a loss (comparison rules), seven alternate strategies for selecting an investment option (investment rules), and three levels of proportion of wealth (low, medium, high) that was reinvested on the prevalence of investment options (riskiest, safest, or medium risk) chosen by agents, the mean wealth acquired, and the median-to-mean ratio of wealth. Results revealed conditions under which the riskiest or the safest investment option emerged as the most prevalent.