“We show poverty itself does not affect willingness to cheat. However, our results suggest that poverty negates the impact of interventions to reduce cheating.” 

Cheating such as corruption and tax evasion is prevalent in the developing world; therefore, many interventions have been undertaken to reduce cheating in developing countries. Although some field evidence shows that poverty is correlated with cheating, the causal effect of poverty on cheating in the field and the effectiveness of interventions for financially constrained people remain an open question. 

The goal of this study was therefore twofold. First, we were interested in the causal effect of poverty on cheating behavior in the field, in general. For this purpose, we recruited a unique population of Thai rice farmers to play a simple cheating game. In the game, we exploited the differences in financial constraints the farmers face before harvest (when they are poor) and after harvest (when they are richer). Second, we explored the effectiveness of an intervention to reduce cheating when our participants face different levels of financial constraints. For this, we used a popular instrument—a social-norm reminder. We believe that it is important not only to analyze the effect of poverty on the general tendency of ethical and prosocial behavior but also to consider how interventions can achieve desirable behavior when the target population faces (or does not face) poverty.

Over 550 Thai rice farmers participated in our cheating experiment. Half of them participated in the experiment before harvest (when they were poor) and half after harvest (when they were richer). We used a simple game to measure rice farmers’ cheating behavior. In the first experiment, we gave each participant a sealed envelope containing 10 folded pieces of paper bearing the numbers from 1 to 10. We asked participants to blindly take out one piece of paper and then report the observed number. The payoff was 10 Thai Baht times the number reported. This created an incentive to cheat for monetary benefit. As researchers, we did not know exactly who cheated and by how much. However, in the absence of cheating, we should observe that every number between 1 and 10 occurs approximately 10% of the time. If reported numbers were higher than expected, this would indicate that participants cheated. In the second experiment, we tested for the effectiveness of a social-norm-reminder intervention to reduce cheating. In this experiment, before playing the cheating game, participants were informed that most rice farmers in their province find cheating for one’s own benefit unacceptable.

We found that poverty itself does not affect willingness to cheat—that is, participants cheated to a similar extent before and after harvest. When reminded of the social norm, they did not cheat less when they were poorer (before harvest), but the social-norm-reminder intervention reduced cheating when they were richer (after harvest). This result suggests that the timing of interventions to change behavior matters. Most households experience some more-or-less predictable cycles of financial constraints due to pay cycles or seasonality in income streams. Interventions to change behavior need to take those cycles into account and time their implementation in periods of fewer financial constraints.

Suparee Boonmanunt (Mahidol University)
Agne Kajackaite (WZB Berlin)
Stephan Meier (Columbia Business School)

The full paper “Does poverty negate the impact of social norms on cheating?” is available as a working paper was published in Games and Economic Behavior.

Reference

Boonmanunt, S., Kajackaite, A., Meier, S. (2020). Does poverty negate the impact of social norms on cheating? Games and Economic Behavior, 124, 569-578.