Project group A01: Experiments

Conducted by Si Chen (former member of A01)

In decisions over actions that might do harm to others, how does a selfish incentive to favor one of the actions change decision makers’ information acquisition on the consequences of their actions for others? Backed up by a theoretical model, we predict that the presence of a selfishly incentivized action will cause acquisition of sequential imperfect information on the consequences of this action to be asymmetric: termination of information acquisition is more likely after evidence indicating no harm and equally likely after one indicating harm. This way, individuals exploit information to justify their selfish action. We test the predictions in a laboratory experiment.

The experimental insights are relevant for the incentive design of physicians, the procedural design of the publication process of experimental studies, and to understand how information acquisition enables discrimination, for example in hiring decisions.

Conducted by Jana Hofmeier (former member of A01) and Thomas Neuber, A01

People care about others. But how are benefits which are not experienced by individuals themselves transformed into motives for behavior? We conjecture that people assess consequences arising to others based on a combination of their own and of receivers’ preferences. This phenomenon is muted when choices are about money and we call it imperfect empathy. In a laboratory experiment, we elicit subjects’ willingness to pay (WTP) for not having to eat food items containing dried insects. We then study the determinants of senders’ WTP for sparing randomly determined receiving subjects having to eat those same items. We find that own and senders’ WTP are complements in generating transfers. As a consequence, transfers between subjects with dissimilar preferences are lower than between subjects who are more alike. Using subjects’ personal preferences as the normative benchmark, we can further show that heterogeneity in preferences reduces welfare. Our findings are directly relevant for individual behaviors such as charitable giving or volunteering. Further, they allow for alternative perspectives on the phenomenon of in-group bias and have implications on the aggregate level. If preferences of people who receive welfare benefits systematically differ from those of net payers, this could reduce support for redistributive policies.

Conducted by Simone Quercia (former member of A01, in cooperation with Francesco Fallucchi (LISER - Luxembourg)

The project aims to investigate experimentally the effects of an affirmative action policy that reserves a share of the prize to subjects of a disadvantaged category in competitive environments. We test three potential pitfalls of the affirmative action policy: (i) whether the introduction of the policy distorts effort and selection in the contest, (ii) whether it leads to reverse discrimination, that is, discourages entry from the advantaged category and (iii) whether the possibility of ex-post retaliatory actions undermines the effectiveness of the policy. We find that the affirmative action contest increases entry of players from the disadvantaged category without affecting entry of advantaged players. This increases overall effort in the contest. However, we find that the possibility of retaliation can undermine the benefits of the affirmative action policy reducing contest participation. This suggests that retaliation is an important aspect to consider when implementing affirmative action policies.

Conducted by Thomas Dohmen, A01 and Elena Shvartsman (University of Basel)

The goal of the project is to analyze how work arrangements that couple employee autonomy and performance incentives affect workplace stress. According to the Job Demand-Control (Karasek, 1979) individuals may experience detrimental health effects when job demands are not met by a sufficient level of decision latitude. Yet, it is not clear, whether granting autonomy (decision latitude) to employees who face incentivized performance targets may impair their well-being. In such arrangements, autonomy, for instance, in terms of self-determined working time, could backfire, because workers may strive to reach their performance goals and potentially exhibit self-harming behavioral patterns such as excessive overtime work. This detrimental effects may be particularly pronounced if individuals underestimate the stressfulness of their working conditions and therefore provide sub-optimal effort. It is therefore of great interest to society and firms to elicit how modern work arrangements affect job stress and thereby to illuminate the particular role of combinations of managerial practices, such as autonomy and incentives.

Conducted by Louis Strang, (former member of A01A01) and Sebastian Schaube (University of Bonn)

In this project, we explore the importance of accountability for others’ payoffs. Own high outcomes either directly result in low outcomes for another individual, or do not affect the success of others. We investigate the fairness differences between these two opposing systems, using laboratory experiments, by causally establishing that a direct payoff interdependence increases demand for redistribution. Our results highlight the importance of (perceived) interrelations of success within societies for inequality acceptance. Furthermore, our results might inform optimal wage setting within firms, as forced rankings or promotion tournaments are often used as incentives. Allowing bonuses that can be gained by every employee simultaneously might enhance work satisfaction and, hence, production in the long run.

Conducted by Si Chen, Thomas Dohmen, A01 and Elena Shvartsman (University of Basel)

From fund managers who allocate assets worth millions to small business owners in developing countries investing in new products, people often make risky economic decisions under stress. Understanding how stress changes risk preferences would not only help decision makers to adjust their risk taking behaviour and improve their ex-ante welfare, but also enhance our comprehension of stress-induced risk seeking or aversion in stressful environments such as financial market or impoverishment. Existing evidence on the effect of stress on risk preferences is inconclusive. The goal of this project is to reconcile the previous conflicting findings by testing the hypothesis that stress amplifies the underlying risk  taking tendency, i.e., more risk averse individuals become more risk averse under stress, while less risk averse individuals become less risk averse.

Conducted by Marta Kozakiewicz, A01 and Lorenz Götte, B07

In the project we experimentally investigate the mechanism behind misguided learning. It is closely related to our recent work, Götte and Kozakiewicz (2018), in which we corroborate the theory by Heidhues et al. (2018) and show that, under certain conditions, persistent overconfidence can lead individuals to taking suboptimal actions, misreading the output and forming erroneous beliefs about the state of the world. In line with the model's predictions, the learning process of overconfident subjects is "misguided", as it drives the agents away from the correct belief and, since the agents themselves generate observation that lead them astray, one can describe it as "self-defeating".

In the current project we conducted an additional control treatment in which the output is based on a parameter that is not relevant to agent's self-esteem. The results demonstrate that the ego-relevance of the unknown parameter exacerbates overconfident agents' tendency to mislearn. Pooling the data from our main study and the additional treatment, we find that for the overconfident agents the effect of the treatment is large and significant. The effect persists even if we control for the initial bias, and is not significant for underconfident of unbiased agents. Our interpretation of the results is that when the output is based on a parameter that is not important to one's self-esteem, overconfident subjects are more willing to admit that they are wrong and abandon their model of the world. The experiment enabled us to shed a light on the driving forces behind the misguided learning, and better understand the behavior of overconfident individuals. The topic is of great relevance as the accurate assessment of the state of the world is essential in virtually all economic decisions.

Conducted by Thomas Dohmen, A01, and Tom Stolp

We conduct a laboratory experiment to investigate basic predictions from principal-agent theory, namely that risk-averse workers are willing to trade-off expected earnings for greater certainty in their compensation. This tradeoff lies at the heart of principal-agent theory and has two implications. First, more risk averse workers prefer weaker incentives for a given tradeoff between risk and expected total compensation. Second, risk averse workers prefer weaker incentives when the riskiness of output increases. By choosing weaker incentives, workers insure themselves against the negative payoff consequences of receiving negative output shocks. In a first laboratoy experiment, we show that risk averse workers do in fact prefer lower piece rates in environments with low risk. However, in risky environment workers do not choose lower piece rates on average. Remarkably, there is heterogeneity in workers' responses: while high productive workers - in line with predictions of the standards model - choose a lower piece rate in more risky environments, the least productive workers react to a more risky environment by choosing higher piece rates. We hypothesize that this behavior is driven by reference-dependent preferences.

In order to test this hypothesis, we conduct a second experiment, in which individuals perform the same real-effort task and choose how they want to be rewarded from the same menu of linear piece-rate contracts, but in which we manipulate reference points and the riskiness of the environment using a 2x2 between-subjects design (no-risk vs high-risk and low vs high reference point treatments). We manipulate reference points by introducing salient counterfactural earnings. In line with our hypothesis, we find that individuals in the low reference point treatment choose significantly lower piece rates in risky environments while individuals in the high reference point treatment choose higher piece rates.

Conducted by Thomas Dohmen, A01 and Simone Quercia (former member of A01)

In this project, we show that the disposition to focus on favorable or unfavorable outcomes of risky situations affects willingness to take risk as measured by the general risk question. We demonstrate that this disposition, which we call risk conception, is strongly associated with optimism, a stable facet of personality, and that it predicts real-life risk taking. We do so using (i) an incentivized measure of risk taking contained in our experimental dataset and (ii) self-reported real-life behaviors from the German Socio-Economic Panel (henceforth SOEP). For both datasets, we find a significant association between risk taking behavior and dispositional optimism. Finally, we investigate the channels through which this association operates and show that risk conception and dispositional optimism are not related to probability weighting but rather to cognitive aspects related to focusing on the outcomes of risk. We conclude that, in addition to being a proxy for a latent risk preference parameter, the general risk question captures important personality characteristics relevant for risk taking behavior, thereby providing a broader representation of the factors that should be taken into account when studying decision making under risk.

From a policy perspective, our project sheds light on the determinants of risk taking and in particular on the importance of underlying personality characteristics on top of the curvature of the utility function. Moreover, it suggests that policy makers interested in increasing (decreasing) risk taking may want to nudge people to focus on the positive (negative) aspects of risk. Finally, our results highlight one additional advantage of using the general risk question, on top of its simplicity and easiness of implementation; that is, the fact that the general risk question includes some elements of personality.