انگیزش، پول، اعتبار و تقلب ها

This paper investigates the effects of supervision and incentives on subjects’ performance and cheating behavior in a real effort task. With a sample of 540 participants in three different experiments, we investigated the interaction between motivation and monetary and social rewards, with and without supervision. Our results suggest: (1) lack of supervision promotes cheating, though workers tend to cheat moderately; (2) both economic and social incentives increase motivation but only when workers like their jobs; (3) workers do not increase their band of acceptable dishonest behavior for possible economic rewards, but they do increase dishonest behavior for possible social rewards, like prestige.

The effects of supervision and incentives on subjects’ performance and cheating have been studied in last years and results are controversial. On one hand, regarding to supervision-performance relationship, Falk and Kosfeld (2006) have suggested that close supervision of workers might undermine intrinsic motivation. Ariely et al. (2008) have proposed that the way in which monitoring is framed crucially influences its effect on motivation. On the other hand, supervision is expensive, especially for small firms. In order to reduce the cost of supervision and encourage workers to actively contribute, organizations often distribute incentives (Clark and Wilson, 1961). Financial and economic incentives are commonly used in the labor market, and the effect of those incentives varies depending on the type of task (Camerer and Hogarth, 1999). In fact, although according to standard economic reasoning an increase in the financial incentives provided for an activity will improve performance, there are some notable exceptions. Gneezy and Rustichini (2000) show that people who are unpaid tend to exert greater effort than those paid only a small amount. Ariely et al. (2009), meanwhile, demonstrate that when incentives are too high, people may “choke under pressure,” resulting in very bad performance. However, supervision and incentives not only have effects on subjects’ motivation and performance but also on cheating. Without supervision people could cheat. According to Nagin et al. (2002) employees are “rational cheaters”. They anticipate the consequences and firms respond with monitoring and incentive systems. In the labor market, employees have many incentives for being dishonest, whether they are ordinary workers or top executives. One clear example are bank sales employees. Their job should be to advise customers how to best invest their money. However, in reality, bank employees have sales targets that will affect the quality of their advice in order to increase sales and, in turn, their salary. Another good example are brokers. Although brokers are supposed to act in their clients’ best interests, the commissions system can tempt them to choose personal gains over their clients’ interests (Davis, 2004 and McDonald, 2002). We could also find dishonest behavior of CEOs and high executives in big companies. While money markets are characterized by a monotonic relationship between payment and effort, in social markets effort is independent of economic compensation levels (Heyman and Ariely, 2004). Social motivation, like prestige, may play a large role in promoting their dishonest (and sometimes irrational) behavior. Especially in the case of experienced CEOs, for whom economic incentives are not as important, prestige could be a good argument to increase their motivation but also their dishonest behavior: “…you do not understand anything. It is not the money. It is the game, the game between people.” (Gordon Grekko in Wall Street: Money never sleeps). Moreover, the effect could be bigger if competition appears. Under competition individual increase their effort (Gneezy and Rustichini, 2004) but if individuals recognize a possibility to cheat, their incentives to increase effort under competition are muted (Schwieren and Weichselbaumer, 2010). Thus, we have the following hypotheses: Hypothesis 1. Supervision is necessary not only because supervision affects motivation, but also because lack of supervision promotes dishonest behavior. Hypothesis 2. Although incentives increase motivation and, therefore, performance, without supervision incentives will also increase cheating behavior, especially in the case of social incentives like prestige, which are obtained through a competitive process. With respect to the level of cheating and taking into account that there are different types of cheaters (Gneezy et al., in press), the Theory of Self-Concept Maintenance (Mazar et al., 2008) suggests that workers typically solve this motivational dilemma adaptively by finding a balance or equilibrium between two motivating forces: financial incentive and positive self-concept. In this equilibrium, workers can derive some financial benefit from behaving dishonestly (but not too dishonestly) and still maintain their positive self-concept in terms of being honest individuals. Following this reasoning, (1) the equilibrium between incentives and positive-self concept should be independent of supervision, and (2) workers should be less likely to cheat in a highly supervised setting, because the chance of “getting caught” is significantly higher. In addition, the cheating could be affected by lack of motivation. To control for this effect, we included a measure of intrinsic motivation in our study. If we take into account intrinsic motivation, unmotivated employees are likely to extend little effort in their jobs, avoiding the workplace as much as possible, leaving the organization if given the opportunity, and producing low quality work. Employees who feel motivated toward their work, however, are likely to be persistent, creative and productive, turning out high quality work that they willingly undertake (Amabile, 1993). Intrinsic motivation is the extent to which an individual is interested in a task and engages in it for the sake of the task itself (Utman, 1997). Therefore: Hypothesis 3. People with high intrinsic motivation in their job not only have greater motivation and become more productive, but also cheat less and need less supervision. Hypothesis 4. Employees still maintain the same positive self-concept, in terms of being honest individuals, with different grades of supervision. As such, with high levels of supervision, where workers are more likely to “get caught,” cheating behavior will decrease. Given these points, we conducted three experiments to examine the relationship between supervision and incentives and their implication on performance and cheating. In the first experiment, in order to analyze performance and grade of cheating in different monitoring situations, we replicated experiment 1 of Ariely et al. (2008), but including a control measure of efficiency. In the second experiment, we added economic and social incentives to the task in order to analyze how they modify levels of both motivation and cheating. In the third experiment, our main objective was to control if the levels of motivation and cheating change with intrinsic motivation, with and without incentives. Our results suggest that workers’ supervision is necessary, not only because supervision may increase motivation, but also because without supervision, workers cheat. Consistent with the literature, we found that incentives increase motivation; however, this only occurs for participants with high intrinsic motivation. At the same time, certain incentives (in particular social incentives) seem to increase participants’ band of acceptable dishonest behavior.

Our results suggest that the supervision of workers is necessary not only because of its effects on motivation, but also because lack of supervision promotes dishonest behavior. Without supervision, we have found that workers only complete half of the work they purport to have finished and therefore spend less time “on the job”. As we can see in experiment 1, the fact that workers cheat when they are not supervised decreases their economic profits (they earn less money than if they work honestly for the same time), while at the same time decreasing the profits of the organization (because more than 50% of the work is not completed). Regarding incentives, both economic and social incentives motivate participants who enjoy their job to be more efficient. At the same time, it seems that people can increase their band of acceptable dishonest behavior if presented with the right incentive. In the case of economic rewards, we did not see a significant effect on dishonest behavior. However, our economic incentive was small, 10€. It is possible that increasing the reward may cause the level of cheating to become significant. Nonetheless, the amount of money offered to the participants was about two times what participants considered a “good deal”, and three times the amount of money they earned for the task they completed during the initial experiment. Moreover, higher reward amounts are not realistic incentives for regular workers in real firms, because economic incentives are rarely bigger than three times regular salary. In the case of social reward, we find that even prestige in a small social group (in this case the participants’ class) increases participants’ motivation, but at the same time also increases cheating to win the reward. Our results show the importance of a firms’ hiring procedure. In particular, companies should place a high value on hiring employees who display a preference for the tasks they would need to complete during their jobs. Along the same line of reasoning, Collins (2010) concludes that because the performance of intrinsically motivated employees ultimately benefits the organization, human resource professionals should understand and support antecedents that empower their most capable employees. According to our results, workers who did not like the task worked less and spent more time completing the task than workers who enjoyed the task. Moreover, under low supervision, intrinsically motivated employees had greater quality of work and cheated less. While participants cheated, their level of cheating was not as high as it might have been. Participants could have reported they completed all sheets, especially in the Shredded conditions where there was no way for the experimenter to monitor if subjects cheated. However, it seems participants tempered their level of cheating, potentially to fall within a band of “acceptable dishonesty,” which is limited by internal reward considerations. These results support the Theory of Self-Concept Maintenance (Mazar et al., 2008). In reference to the labor market, workers without supervision could have a dilemma about whether or not to cheat. Finally, Ariely et al. (2008) found significant differences between supervised and non-supervised participants. In their case, American participants in the non-supervised conditions allege to have completed less work. We do not see this trend with Spanish participants. Spanish students in the current study cheat when unsupervised while MIT students in the original simply work less. While this could be explained by cultural differences between Spanish and Americans in their concept of right and wrong (Spanish participants may accept a higher level of cheating without tipping the moral/immoral scale (Kuehn et al., 1990)), it would be appropriate to take into account other possibilities. For instance, Charness and Dufwenberg (2006) assume that people feel guilty if they harm others by lying to them, and Gneezy (2005) found that lying behavior depended on the costs it imposed on both the liar and on the one who is lied to. We can link those ideas with the CuPS (Culture X Person X Situation) approach (Leung and Cohen, 2011) and its conclusions about American and Latin people. In this approach, one variable may predict a given behavior in one culture, and it may predict the opposite type of behavior in another culture. According to Leung and Cohen (2011), it is necessary to understand the particular logics of a cultural system, mainly dignity and honor. It is very important not to forget that US Universities like MIT where Ariely et al. (2008) run the experiments have a very important Honor Code that students in Spanish Universities do not. The distinction between honor, dignity, glory and respect as values and principles that may be used to read value structures and cultural codes (Orit, 2002) could explain the difference found. We consider this line very interesting for future research. Furthermore, as in our experiment each participant did the task alone, it would be very interesting analyze in future research the role of others in individual behavior in line with recent works about peer effects (Fosgaard et al., in press) or Robin Hood effect (Ploner and Regner, in press and Gino et al., in press). In summary, our results suggest that the supervision of workers is necessary, not only because it may moderate motivation, but in particular because lack of supervision promotes cheating. In this article, we show that economic and social incentives can increase peoples’ motivation (when participants like the task), but they can also incentivize workers to cheat. People can choose to be honest or not. They often have two competing motivations: gaining from cheating versus maintaining their positive self-image as honest individuals (Aronson, 1969 and Harris et al., 1976). It seems to be a win–lose situation from which we have to choose. However, Mazar et al. (2008) proposed in their Theory of Self-Concept Maintenance that people typically solve this motivational dilemma adaptively by finding a balance or equilibrium between the two motivating forces. People tend to behave dishonestly to derive some financial benefit, but only to a small degree in order to maintain their positive self-concept as honest individuals. Our results support this theory and go beyond, demonstrating that not only must workers be supervised to discourage cheating, but also that monetary or social incentives can modulate the amount of cheating and in some cases increase dishonesty. In fact, when we analyze the level of cheating, people who like their job cheat less than people who dislike their job, even with economic incentives. Importantly for future research, when social incentives appear, the level of cheating becomes the same for all workers regardless of intrinsic motivation.