On the reduction of deception in principal-agent relationships

Resumen

On the reduction of deception in principal-agent relationships Sascha Behnk September 2015 1. Introduction Economic fraud is a prevailing problem and a substantial threat to the well-being of our society. Many recent cases, such as the fraudulent accounting practice of the Enron Corporation with its long-term repercussions for many market participants (Akhigbe et al., 2005) or Bernard Madoff¿s billion dollar pyramid scheme (Krugman, 2008), have attracted much public attention. Beside this fraud in large corporations and markets, deceptive behavior can also appear at the micro-level of economic activity, whenever an economic relationship includes an asymmetric distribution of information among the involved parties. Although examples of economic deception are manifold, we observe that people do not necessarily deceive their fellow man at each and every opportunity. Quite the opposite is true in everyday economic life. The reasons why we mostly refrain from deceiving others are the subject of a variety of studies in behavioral economics literature since a further understanding of the mechanisms behind the decision whether to deceive others or not, are fundamental for the development of efficient measures against fraud. In this thesis, we investigate three different factors that can potentially reduce deception in principal-agent relationships with asymmetric information: transparency, punishment and reputation. Furthermore, we explore the conditions under which they do or do not work efficiently. 2. Methodology In all three studies we make use of the experimental method. Economic experiments not only provide a useful approach to reach important insights into the motivations that drive individual decision-making but are also useful in determining suitable solutions to areas where standard economic theory is yet not able to explain behavior. The nature of the methodology is to obtain and analyze data regarding the decisions of experimental subjects in a controlled environment. In this way, experiments can be designed to filter out the specific effects of the variables being studied. Thus, one can observe and econometrically test changes in the behavior of people in these controlled environments. 3. Conclusions In the first chapter of this thesis, we address an issue regarding an often used measure to increase honesty and trust in economic environments: transparency. Previous studies have shown that ex ante disclosure of conflicts of interest not only fails to improve these relationships but that it even leads to more deception (Cain et al, 2005 & 2011). We assume that the moment in which transparency is provided is important for its efficacy to prevent people from deceiving others and propose that providing ex post transparency could play an important role in reducing deception. We use a sender-receiver game (Crawford and Sobel, 1982) similar to Gneezy (2005) as a basis for the studies in this thesis. In the two-player version of this game, a sender receives information about a variety of options with different payoff distributions for him and his counterpart, the receiver, who is blind regarding the payoffs. The sender¿s task is to recommend one option to the receiver which serves to investigate if he is willing to be dishonest in order to receive a higher payoff. Different scenarios of this sender-receiver game allow us to show that disclosing conflicts of interest ex-post not only does not induce more deception but also that a 50% chance of subsequent disclosure significantly reduces deception in the case in which the sender obtains a small gain at the expense of a comparatively big loss for her counterpart. We conclude that the effect of transparency ex post may be related to people who care not only for their own behavior and its immediate aftermath, but also to try to protect their social image, in the sense that they prefer to be perceived by others as a person with moral integrity. Regarding the receivers, we find that the increase in honest messaging, when it takes place, is not anticipated. In a second study, we investigated whether the use of negative enforcement is able to foster pro-social behavior where ex post transparency alone did not succeed in reducing deception. Therefore, we provided the receiver with the cost-free possibility to severely punish dishonest senders, since a variety of previous studies found that costly and soft punishment does not necessarily lead to less deception (Sánchez-Pagés and Vorsatz, 2007 & 2009, Peeters et al., 2013). Furthermore, we addressed a factor of fundamental importance for the efficiency of a sanctioning system: the probability with which defection is revealed. To shed further light on this issue we used two treatment lines across which we varied the likelihood of being detected from 100% to 50%. Since receivers found out about their counterpart's behavior with same probabilities in the baselines, we were furthermore able to tell apart the deterrence effect of potential punishment from the image concerns caused by ex post disclosure. In contrast to several previous studies, we found that a substantially higher fraction of senders acted honestly with punishment. However, we did not observe a significant difference in the fractions of honest messages in any scenario comparing assured revelation with 50% disclosure probability. We conclude that this stable deterrence effect might enable principals to reduce their monitoring effort in similar settings without risking a decrease in the level of honesty. Additionally, we found that the deterrence effect of punishment is less pronounced when senders face the temptation of gaining a comparatively high amount from deception, but senders seemed to be unaffected by the financial consequences of deception in terms of the receiver's relative loss when facing the possibility of being sanctioned. Receivers, on the other hand, showed substantially higher trust levels when they were able to sanction dishonest senders. Furthermore, the difference between the high acceptance rates in the punishment treatments was not significant, implying that receivers anticipated that the punishment's deterrence effect is independent of our variation in the disclosure probabilities. We conclude that cost-free and severe punishment in principal-agent relationships with asymmetric information does not only lead to more honesty but enhances trust levels accordingly. This is an important condition since an intact economic relationship can only be established when all involved parties actually agree to interact with each other. In the third study, we investigated the effect of reporting systems on reputation (Rockenbach and Milinski, 2006) and deceptive behavior. This time, we used a repeated sender-receiver game to compare the effects of two different report types, exogenous ones that were generated by the computer in a standardized way and endogenous reports in form of an individual text, written by the receivers. In both cases, the provision of reports led to a significant lower level of deception compared to the baseline. Furthermore, we found that the provision of computer reports decreased deception to a significantly higher extent than receiver reports. With the latter report type, deception even increased over time. Altogether, we conclude that the higher reliability of the exogenous reports has a higher impact on pro-social behavior in our setting than the personal character of endogenous reports. A qualitative analysis of the receiver reports provided a potential explanation for the stronger effect of computer reports. It turned out that a substantial fraction of the receivers did not use the opportunity to report dishonest senders in a proper way. Hence, we suppose that the comparably lower impact of the endogenous reports is due to an anticipation of the high number of missing, incorrect or unrelated reports by the senders. Interestingly, the high number of missing reports after successful deception implies that some receivers experienced an aversion to damage the reputation of others. On the other hand, receivers showed relatively more trust when endogenous reports were provided, whereas computer reports did not have a significant effect on trust. This finding reveals an interesting discrepancy between the agents¿ and theirs principals¿ perception of the two report types. We propose that this discrepancy should be considered regarding the design of reporting systems that affect reputation, for instance in the popular case of consumer feedback on commercial online platforms (Dellarocas, 2003). However, we find that computer reports gradually enhance the level of trust over time and assume that receivers might realize in the long run that the reliability of the exogenous reports leads to comparatively less deception. The studies presented in this thesis reveal both the potential and the limits of the three examined factors in reducing moral hazards. The results also reveal unexpected side-effects and shed light on the circumstances under which these factors can be optimally used. We showed that even a subtle intrinsic motivation can keep agents from deceiving others but also that this effect is only strong enough under specific conditions. Regarding the architecture of markets and organizations, we conclude that this cost-effective tool can be used without negative repercussions but also that stronger incentives for pro-social behavior, such as sanction or reporting systems should be implemented to bridge the gaps in which intrinsic motivation alone is not able to reduce deception. 4. Bibliography Akhigbe, A., Madura, J., Martin, A.D., 2005. Accounting contagion: the case of Enron. Journal of Economics and Finance 29(2), 187-202. Cain, D.M., Loewenstein, G., Moore, D.A., 2005. The dirt on coming clean: Perverse effects of disclosing conflicts of interest. Journal of Legal Studies 34(1), 1-25. Cain, D.M., Loewenstein, G., Moore, D.A., 2011. When sunlight fails to disinfect: Understanding the perverse effects of disclosing conflicts of interest. Journal of Consumer Research 37(5), 836-857. Crawford V.P., Sobel J., 1982. Strategic information transmission. Econometrica 50(6), 1431-1451. Dellarocas, C., 2003. The digitization of word of mouth: Promise and challenges of online feedback mechanisms. Management Science 49(10), 1407-1424. Gneezy, U., 2005. Deception: The role of consequences. American Economic Review 95(1), 384-394. Krugman, P., 2008. The Madoff Economy. The New York Times 2008(19), 45. Peeters, R., Vorsatz, M., Walzl, M., 2013. Truth, trust, and sanctions: On institutional selection in sender-receiver games. Scandinavian Journal of Economics 115(2), 508-548. Rockenbach, B., Milinski, M., 2006. The efficient interaction of indirect reciprocity and costly punishment. Nature 444(7120), 718-723. Sánchez-Pagés, S., Vorsatz, M., 2007. An experimental study of truth-telling in a sender-receiver game. Games and Economic Behavior 61(1), 86-112. Sánchez-Pagés, S., Vorsatz, M., 2009. Enjoy the silence: an experiment on truth telling. Experimental Economics 12(2), 220-241.