Toward a Computational Understanding of Bribe-taking Behavior

Understanding how corrupt behavior occurs is a critical issue at the intersection of behavioral ethics, social psychology, and other related social sciences, which lays the foundation for establishing effective policies combating corruption. Previous studies in Behavioral Economics and Social Psychology have primarily focused on developing lab-based research paradigms and identifying various situational and personality factors that modulate the bribe-taking behavior, a major form of corruption. However, a theoretical framework that quantitatively accounts for the psychological processes underpinning bribe-taking behaviors is still lacking. Inspired by recent literature on neuroeconomics and moral decision-making, here we provide a computational framework which assumes that the decision of whether to take or refuse a bribe involves a value-based computational process. We showcase how this framework can advance our theoretical understanding of bribe-taking behavior 1) by helping clarify how a power-holder weighs benefits against costs and integrates them into a value signal that allows the power-holder to make the decision of taking or refuse the bribe, 2) by enhancing our understanding of the potential mechanisms through which various contextual modulators impact the bribe-taking behavior, and 3) by improving the prediction of bribe-taking behaviors in individuals through the association between bribe-taking and personality traits. Moreover, we showcase how this framework can be potentially extended to explain the computational processes underlying more intricate forms of corrupt behaviors. We also highlight the practical application of this framework in offering mechanistic insights into effective anti-corruption intervention policies and design of morally-aligned AI systems.