Presenter: Ernest Lai, Lehigh University Title: Communication Design: An Experiment Abstract: We experimentally document the efficiency-enhancing role of communication design. Two communication channels are considered: a noisy channel, in which a noise process maps the sender's message space into itself, and a mediated channel, in which the received messages are framed as action recommendations under the same noise process. We find that both channels improve communication efficiency relative to the benchmark of direct communication. Despite the strategic, theoretical equivalence of the two communication environments, the mediated channel outperforms the noisy channel in the laboratory.
Presenter: Homayoon Moradi, WZB Social Science Center Berlin Title: First Show Me the Money: Moral Wiggle Room Reverted Abstract: Deliberately avoiding information about welfare consequences of self-interested decisions to act selfishly, or willful blindness, is an important source of socially harmful behavior. We study the dictator game with hidden payoffs that can be voluntarily revealed at zero cost, as in Dana, Weber, and Kuang (2007) and vary the timing of revelation. In all treatments the dictator makes a choice between two options: a more selfish option and a more pro-social option. The distribution of payoffs for these options is known, but the actual payoffs can be hidden and then revealed by dictators at zero cost at a particular time depending on the treatment. In the own-payoff-1st treatment, as in Dana, Weber, and Kuang (2007), the dictator observes her actual payoff before deciding whether to reveal the receiver's payoff. In the own-payoff-2nd treatment, the dictator observes her own actual payoff after deciding whether to reveal the receiver's payoff. We replicate their result successfully; 34% of dictators choose the selfish option in the baseline dictator game while 64% do so in the ownpayoff- 1st treatment. But we find that in the own-payoff-2nd treatment only 29% of subjects choose the selfish option. These results contradict the explanation in Dana, Weber, and Kuang (2007): the observed discrepancies cannot be explained by image concerns. We conjecture that dictators wishfully think their self interested action is more likely to be harmless for receivers when information is avoided compared to when it is revealed. We analyze a wishful thinking model of an agent whom avoiding information biases her belief self-servingly. We derive several behavioral predictions that are inconsistent with previous models and conduct experiments to test the wishful thinking bias. Our findings, as well as a number of previous experimental results, offer support for these predictions.
Presenter: Stephen Leider, University of Michigan Title: Promise-keeping Norms and Renegotiation Behavior Abstract: We use a trust game to test the impact of promises on behavior in the presence of uncertainty where the source of the uncertainty stems from acts of nature. Second, we test whether promise-keeping norms affect subsequent decisions about whether to renegotiate the terms of the initial agreement once the uncertainty is resolved. We find that second movers frequently make promises, and that these promises affect their subsequent decisions. We demonstrate that a utility model in which actors have preferences over outcomes as well as social norms is able to account for the pronounced effect of promises on behavior, explain why a second mover would make the promise in the first place and why a first mover would be willing to enter the game. For policies written against the backdrop of strong norms, we address implications and guidelines.
Presenter: Bettina Rockenbach, University of Cologne Title: Using Behavioral Insights to Increase Payment for Water in Rural Namibia Abstract: Defaults on payments for public utilities are a serious constraint to maintaining and expanding infrastructure in developing countries. In a natural field experiment in one of the driest countries in the world, Namibia, we study the power of behavioral interventions for increasing payments for water. In cooperation with the national Namibian water provider, we implement three behavioral interventions, treating about 10,000 customers over a time span of nine months. Our basic treatment provides customers with a regular text message containing simplified invoice information. In the other two treatments we add structural elements to evoke a water paying self-concept ("self-concept" treatment) or implementation intentions about future payments ( "implementation" treatment). In the first month of the intervention average payments increase by 26% in the basic treatment, by 35% in the self-concept treatment and by 45% in the implementation treatment, while water consumption is unaffected. Even after nine months the customers in all three treatments still pay substantially more than in the control group. Our study demonstrates the power of behavioral interventions in a context where standard economic solutions are limited.
Presenter: Valeria Faralla, University of Piemonte Orientale Title: Framing Effects in Intertemporal Choice: A Nudge Experiment Abstract: This paper experimentally investigates the framing effects of intertemporal choice using two different elicitation modes, termed classical and penal. In the classical mode, participants are given the choice between receiving a certain amount of money, smaller and sooner, today and a higher amount, larger and later, delayed (e.g., €55 today vs. €75 in 61 days). This is referred to as the standard mode. In the penalty mode, the participant must give up an explicit amount of money in order to choose the smaller and sooner option (e.g., €75 in 61 days vs. €55 today with a penalty of €20). This is the explicit mode. We find that estimates of individual discount rates are lower in the explicit mode than in the standard mode. This result suggests that even very simple information about the amount of money one must surrender for choosing the earlier option increases delayed consumption. The finding has relevant implications for self-control and long-term planning in intertemporal choice.
Presenter: Erika Gross, Otto-von-Guericke-University Magdeburg Title: Announcing a nudge Abstract: In every decision situation there exists a default solution, which determines what happens if the decider does not take an explicit action. As we know from numerous field and laboratory experiments, defaults do influence decisions. For example, it makes a huge difference if pension systems make use of an opt-in or an opt-out default solution. The fundamental idea of nudging is that policy makers should be able to use such behavioral regularities to manipulate individual behavior without forcing people to do something. For example, if it is necessary to increase the number of employees enrolled in a pension plan a switch from opt-in to opt-out is a very promising reform option. Nobody is forced to enroll but the share of enrolled employees will increase. Nudges of this kind work because they avoid procrastination and overcome the tendency of many people to avoid explicit choices. It is important to note that nudges, like the change of the default option, should be accompanied by a very transparent communication in which people are informed about the fact that they will be nudged and why they will be nudged. The impact of default changes has been documented in a variety of experiments. However, the question how the information about the nudge influences behavior has so far not been investigated experimentally. Three reaction are possible. First people may react to the announcement that they are going to be nudge in a negative way and work against the intention of the nudge because they feel manipulated. Second, they may not care about the announcement and third, they can use the announcement as a kind of a coordination advice. The last reaction would increase the effect of the nudge while the first reaction will decrease the effect. In our experiment, we use a public good design with two different defaults (give and take design). We investigate the introduction of an information shock at different stages of the game. First results show, that information about the use of a nudge and the intention of the nudging increases cooperation as measured by the number of subjects choosing the fully cooperative default under the informational treatment.
Presenter: Hannes Rau, Universität Heidelberg Title: Gender Differences in Negotiations- Experimental Evidence Abstract: We systematically study gender differences in experimental wage negotiations, which are widely seen as one factor influencing the gender wage gap. According to past results, women achieve worse outcomes than men, but only when being in the weaker position of bargaining power. Although some factors such as asymmetry of power, negotiation domain and information about gender are supposed to play a role, there is no comprehensive study of their relative importance. To this end, we conduct an experiment in which subjects negotiate in groups-of-two how to divide a pie with asymmetric outside options. In one treatment the genders of the group members are mutually revealed beforehand. We then compare gender differences between treatments and check if these depend on the position of power. Furthermore, we run the same experiment in a neutral and in a business frame, to check if behavior is influenced by the domain of the negotiation.
Presenter: Elif Demiral, George Mason University Title: Choosing to Compete against Self or Others - Gender Differences Abstract: This paper introduces and examines selection into a new competitive payment scheme - Self- competition- in which individuals compete against their own past performance. Using a lab and an online experiment, we replicate the well- documented gender differences in willingness to compete against others, but find no evidence of a gender difference in willingness to self-compete. We then expand on our previous work using another online experiment and further study the preferences for self-competition from the competitors' point of view. We investigate the decisions on whether to compete against self or others along with whether to compete at all. We report that both men and women prefer self-competitions to competitions against others and when the self-competition option is available, more people choose to compete. We also investigate the role of confidence, risk aversion and self-serving bias in competitive scheme selection and tournament entry.
Presenter: Noemi Peter, University of Groningen Title: Gender, competitiveness and career choices along the whole ability distribution Abstract: Men are generally found to be more willing to compete than women and there is growing evidence that willingness to compete is a predictor of individual and gender differences in career decisions and labor market outcomes. Most existing evidence comes from the top of the education and talent distribution. In this study, we use incentivized choices from more than 1500 Swiss secondary school students to ask how the gender gap in willingness to compete varies with ability and how willingness to compete predicts career choices along the whole ability distribution. Career choices are extremely gendered. In the vocational track, girls are underrepresented in math-intensive apprenticeships which lead to high salaries. In the academic track, girls are less likely to specialize in physics and mathematics. Our main results are: 1. The gender gap in competitiveness is essentially zero among the lowest-ability students, but increases steadily with grades and reaches 30-40 percentage points for the highest-ability students. This is because boys with high grades are substantially more willing to compete than boys with low grades while the relationship between grades and competitiveness is flat for girls. 2. Willingness to compete predicts career choices. In particular, competitive girls are more likely to choose a math-intensive career both in the vocational and the academic track. 3. Results are very similar whether using a stereotypically male numerical task or a stereotypically female letter task to measure willingness to compete.
Presenter: Anders Poulsen, University of East Anglia Title: Are there attraction and compromise effects in bargaining? Experimental evidence Abstract: The Attraction Effect and Compromise Effect (AE and CE) have been studied in individual choice situations. We define and experimentally investigate the AE and CE for bargaining situations. Our data suggest that the AE and CE are significant in bargaining when certain conditions, related to focal equilibrium selection criteria based on payoff equality, efficiency, and symmetry, are met.
Presenter: Magnus Vaage Knutsen, BI Norwegian Business School Title: Obstinate Parties and Non-binding Outside Options in the Lab Abstract: We experimentally investigate the role of reputation building in bargaining, using the standard alternating offer protocol. The opportunity for building reputation is created by the presence of obstinate types, i.e. behavioral types that insist on an out-of-equilibrium demand. In the absence of outside options rational players have an incentive to mimic obstinate types. This leads to significant delay in equilibrium. Introducing outside options in this environment removes the mimicking incentive and forces agreement within at most two periods in equilibrium. This even holds for outside options that are not binding in the absence of obstinate types. In the experiment obstinate types are implemented as automatons. We find strong support for the hypothesis that ("non-binding") outside options are efficiency enhancing in the presence of obstinate types. We argue that our findings cannot readily be explained by preferences for efficiency or equality, or by reciprocity. In an extension we investigate a model in which bargainers are permitted to purchase an outside option prior to the bargaining game. We find that subjects buy such outside options when this is individually rational.
Presenter: Federica Alberti, University of Portsmouth Title: Emotional intelligence in ultimatum and trust games Abstract: Do people make strategic use of their facial expressions? We test the hypothesis that people use their facial expressions to increase own payoffs in two strategic settings: the ultimatum game and the trust game in particular, that a) second-movers choose an angry face in the ultimatum game, and that b) second-movers choose a smiley face in the trust game, anticipating a more favorable response by the first-mover. For each game, we compare choices of facial icons between two treatment in the "information treatment" the second-mover's choice is disclosed to the first-mover prior to the decision phase; in the "non-information treatment" the second-mover's choice is never disclosed to the first-mover. The effect of information is significant only in the trust game. A final questionnaire elicits emotional intelligence scores of the participants. We find no correlation between emotional intelligence and the choice of a facial icon in both games.
Presenter: Mike Farjam, Linnaeus University Title: Does risk communication really decrease cooperation in climate change mitigation? Abstract: Effective communication of risks involved in the climate change discussion is crucial and despite ambitious protection policies, the possibility of irreversible consequences actually occurring can only be diminished but never ruled out completely. We present a laboratory experiment that studies how residual risk of failure affects willingness to contribute to climate protection policies. Contrary to our initial hypothesis, we find that the contributions were higher in treatments with residual risk than in treatments without one. We interpret this as an outcome of a psychological process where residual risk puts participants into an "alarm mode", keeping their contributions high. We discuss the broad practical implications this might have on the real world communication of climate change.
Presenter: Anne Schielke, University of Cologne Title: Cooperation and the prospect of competing for jointly created surplus Abstract: I experimentally study cooperation in a two-stage game where players compete for jointly created surplus. The experiment consists of two stages: in the cooperation stage, players simultaneously decide how many points to invest into a group project. If the group project is successful, players enter the competition stage. In the competition stage, players simultaneously decide how many points to invest into the acquisition of a prize in a generic contest. Holding expected payoffs fixed, the experimental treatments vary the prize allocation mechanism in the contest: In the winner-take-all mechanism, the contest success function defines the probability that the player wins the prize. In the proportional-share mechanism, the contest success function defines the proportion of the prize awarded to the player.
Presenter: Sara le Roux, Oxford Brookes University Title: Climate Change Catastrophes and Insuring Decisions: A Study in the Presence of Ambiguity Abstract: There has been very little research to test whether ambiguity affects individuals' decisions to insure themselves against the catastrophic effects of climate change. This paper attempts to study how individuals respond to the availability of an insurance that would safeguard their interests if a climate change catastrophe occurred. Moreover, if such an insurance is available to them, do they insure themselves sufficiently? Further, the study investigates if increased information regarding the probability of the catastrophic event, leads to an increase in insurance subscriptions. Finally, policy implications for the State are investigated - Can State intervention in the form of a "nudge" ensure a better outcome?
Presenter: Toby Handfield, Monash University Title: When good norms turn bad: Peer punishment used to enforce inefficient social norms Abstract: When norms can be enforced by peer punishment, groups are able to resolve social dilemmas in prosocial, cooperative ways. Little effort has been made, however, to investigate the potential of punishment to promote anti-social or destructive norms. In this paper we show that punishment can easily encourage participation in behaviours that are harmful to group welfare, and that this phenomenon is mediated by a social norm. In a variation of a public goods game, in which the return to investment is negative for both group and contributor, we find that the opportunity to punish led to higher levels of contribution, thereby harming collective payoffs. A second experiment demonstrated that a significant majority of subjects regard non-contribution in this setting as socially inappropriate; thereby confirming that the effect is norm-driven.
Presenter: Robert Schmidt, Heidelberg University Title: Implementing Fair Procedures? Abstract: We experimentally study situations in which a decision maker allocates resources to two agents. The allocation will be unequal, but the allocation procedure may be fair. However, procedural fairness might be violated by favoritism. This feature is common in many practical settings, including governance, medical and educational settings. We study the perception of procedural and outcome fairness in this setting, and the reaction to different allocation procedures by the decision maker and the affected agents. We find that unfair outcomes are acceptable for the agents if procedures are perceived as fair. However, with opaque allocation decisions, it may be difficult to commit to fair allocation procedures. Indeed, we find a very high degree of favoritism by the decision makers when they are forced to allocate unequal outcomes, and have no fair (random) procedure available. Despite the unfair allocation behavior, poor-outcome agents punish the decision maker only modestly; in contrast, good-outcome agents reward substantially. We consider three interventions to get more insights into the underlying decision processes. First, we provide full transparency about the decision by the allocator. This does not reduce favoritism, but reduces rewards and increases punishments, by the affected agents. Second, we provide the allocator with a randomization device. Although agents are not informed about the randomization decision (but know about the possibility), a large share of allocators chooses a fair procedure. Agents do not punish bad outcomes, optimistically expecting allocators to allocate fairly. Third, we allow the allocators to delegate their decision publicly to a fair random device. A large share of allocators is willing to even costly delegate.
Presenter: Ruike Zhang, Nanyang Technological University Title: How Large Should the "Bullets" Be?: Revisiting the Hired-Gun Mechanism in the Provision of Public Goods Abstract: The "hired gun" mechanism which focuses on punishing the lowest contributor so that the person would rather be the second lowest contributor has been shown effective in promoting full cooperation in voluntary contribution settings (Andreoni and Gee, 2012). The suggested punishment involves two components; a unilateral punishment and a tie punishment. The former is imposed to discourage people from wanting to be the lowest contributor and the latter is added on to prevent people from coordinating on a tie at a below full-contribution level. There is essentially a range of values for the relative magnitude of these two components that would sustain full cooperation equilibrium when implemented. In this paper, we aim to examine how severe the unilateral and tie punishment should be to achieve the full-contribution equilibrium. Specifically, we are interested in investigating the size of the "bullets" that the "hired gun" should carry. In our experiments we vary the magnitude of the unilateral and tie punishment in such a way that full cooperation equilibrium will sustain. We also run an experimental treatment wherein we only eliminate the tie punishment but keep the unilateral punishment intact such that the voluntary contribution game is transformed into a coordination game. Our experimental results generally substantiate the theoretical prediction on the full cooperation equilibrium and the coordination outcome, except for the more lenient punishment parameters.
Presenter: Leif Helland, BI Norwegian Business School Title: Electoral competition and political rents in the lab Abstract: We experimentally investigate the canonical model of rent-taking in two-candidate competition. In the absence of popularity shocks rent taking is zero in the equilibrium of the model. This holds independently of the level of exogenous rents (i.e., "ego-rents") going to the winner. In the presence of popularity shocks equilibrium rent taking of the model increases with the variance of the popularity shock, and decreases with the level of exogenous rents. First, we find that rent-taking is not independent of, but decreases with, exogenous rents in the absence of popularity shocks. Second, the view that rent-taking is higher in the presence of popularity shocks, given exogenous rents, is not supported by our data. These two findings are in direct opposition to the Nash prediction. Third, we find that in the presence of popularity shocks, rent-taking is increasing in the variance of the popularity shock for given exogenous rents, and is decreasing in exogenous rents for given variance of the shock. While our third finding is in line with the directional predictions of Nash equilibrium, the deviation from Nash is substantial when the variance of the popularity shock is high and exogenous rents are low. We present a unified explanation of our findings using a Quantal Response Equilibrium while taking risk-attitudes into account.
Presenter: Alexia Gaudeul, University of Göttingen Title: The social preferences of democratically elected decision-makers: fairer, but sometimes at the cost of efficiency. Abstract: We run a laboratory experiment where some participants are selected to make investment decisions on behalf of others. This generates a surplus that must then be shared between decision-makers and others. We vary the way to select decision-makers. We find that the procedure used to select decision-makers influences the allocations they make, their priorities and their preferences. The procedure affects the total amount of social surplus generated and how that surplus is distributed. Decision-makers who are selected democratically are less egoistic than decision-makers who are selected at random or who are selected by ability. They share the social surplus more equally. This leads to more efficient results if the investment generates wealth. However, this can lead to less efficient results if redistributing money away from the decision-maker requires destroying wealth. We discuss possible explanations for this effect of the democratic process.
Presenter: Ginevra Marandola, University of Bologna Title: Democracy, risk, and productivity: a framed field experiment. Abstract: The purpose of this experimental study is to test the effect of democracy on workers' performance in environments characterized by risk. Designing an effort task experiment without a risk component, Dal Bo et al. (2015) have recently shown that allowing subjects to vote for they preferred payment scheme has no significant effect on effort.[1] This opens the path to further research on democratic collective decision-making, as the impact of mechanism design of incentives on intrinsic motivations may largely depend on the saliency of the choice. We design a framed field experiment to provide more evidence on this issue and test if the relevance of self-determination and democracy increases in risky environments. In our experiment, we recruit students who are informed that they are taking part in an experimental session, but as part of it perform a real data-entry job for sixty minutes, in addition to an initial five-minute trial period to test their baseline ability. In addition to the show-up fee and payment for a risk-aversion elicitation method, students were paid either a piece rate per character or a flat hourly wage. Our design varies across two main dimensions: i) the riskiness of the environment ("risk" and "risk-free" environments); ii) whether subjects can vote for the payment scheme or not ("democracy" and "random" treatments). In the democracy treatment, subjects vote for their preferred payment scheme (flat vs. piece rates), which is then assigned according to a majority rule. In the random treatment, which we use as control, subjects are assigned to a random payment scheme. In the risky environment (both in the democracy treatment and in the random treatment), subjects know that they are exposed to the risk of a lower final payment due to a temporary interruption in the data-entry job (which does not occur in the risk-free environment). Using a methodology that allows us to control for self-selection bias, we show that democracy may have opposite effects on productivity depending on payment schemes' characteristics and risk. In particular, democracy may deliver adverse effects when paired with high-powered schemes, and positive effects when paired with flat schemes. Our evidence originally contributes to the literature on workers' non-monetary incentives and productivity.[2] We find that most of the subjects in our sample prefer flat rate payments, even more so in the risky environment. Preferences over the payment method are correlated with risk aversion: more risk averse subjects choose the flat rate. Other observables do not explain preferences over the payment method, with the exception that males seem to be more prone to choose piece rate. The output of the data-entry task can be measured with high precision both in terms of quantity and quality. We measure each subject's speed (number of entries per minute), precision (percentage of correct fields) and efficiency (number of correct fields per minute). The data-entry task exhibits a quality-quantity trade-off within subject. As under piece rates monetary incentives are loaded on a single performance indicator (number of entries), subjects may divert effort from the non-incentivized task (precision), thus strengthening the quality-quantity trade-off. On average, subjects input 34 entries per minute with a precision rate of 78%. Considering that this is a one time job and it is anonymous, the precision rate is relatively high. The flat rate payment scheme implies higher precision and lower speed. On the overall example, the difference with performance under piece rates is statistically significant. The flat rate scheme reduces speed of 8 entries and increases precision by 4%. This result holds both in the risk and risk-free environments. Efficiency also increases under the flat rate, though the difference is not significant in the risk environment. However, when we study performance in the random treatments (no choice of payment scheme), we do not find significant differences in precision between flat rate and piece rate payment schemes, but the sign suggests that flat rate subjects could be less precise. Under the flat rate scheme, subjects are at the same time slower, less efficient and less precise. Combined with the quality-quantity trade-off intrinsic which is in the task, this result suggests that piece rates incentivize speed without hampering quality in random treatments. In the democracy treatments, piece-rate pay increases speed, but reduces precision and efficiency with respect to flat rate schemes. As, by contrast, in the random treatments we observe absence of a significant difference in precision between different payment schemes, a possible self-selection effect may be at work in the choice treatments, which we control for. In the democracy treatment, subjects are randomly matched in groups of three to vote according to a majority rule. To control for self-selection, we exploit the fact that some subjects work under a payment scheme they did not vote for (hence, "unhappy" subjects). We find that subjects who prefer the flat rate scheme, but work under piece rates are faster than flat-rate subjects in the random treatments. We also find that "unhappy" subjects under the flat rate are slower than piece-rate subjects in the random treatments. This suggests that differences in speed across payment schemes are not due to self-selection. Regressing speed on the payment scheme (dummy 1 for piece rate, 0 for flat rate), and controlling for preferences over the payment method, we find that the effect of the payment scheme is positive and significant. Preferences over the payment method are also positive and significant, but only in the risk-free environment. Subjects who did better in the trial are faster, subjects who declare to enjoy the task are faster, and risk averse subjects are faster in the risky environment. The results on precision show that the difference between flat workers in the democracy treatment and piece-rates workers in the random treatment is not significant. We find instead that "happy" piece subjects in the democracy treatment are less precise than flat-rate subject in the random treatment. Evidence on quality is thus ambiguous. It thus seems that the real difference across payment schemes is only within the democratic environment. This is also confirmed by the regressions estimates. Males are less precise in the risk-free environment and subjects who enjoy the task are more precise in the risky environment. One of our key findings is that the effect of democracy differs across treatments and payment schemes. Using Dal Bo et al. (2015) methodology, we find that democracy has a positive effect on speed under piece rates in both risky and risk-free environments. On the contrary, democracy has a negative effect on precision under piece rates in both the risk-free and risky environments (not significant, but expected sign). Democracy also has a negative effect on efficiency in both risk-free and risky environments (not significant, but expected sign). To better understand the effect of democracy we look at happy and unhappy voters and we find that both happy and unhappy voters are faster and less precise. This can be rationalized in two ways: a) the possibility of self-determination through a democratic vote works as reverse gift-exchange, i.e. an entitlement to pursue one's own utility at the expenses of the employer interests; b) happy subjects work faster in the democracy treatment, because of a self-confirmation bias, i.e. they want to prove that they did not make the wrong choice, and lower quality is a consequence. Unhappy voters work faster and less precisely, because they are upset for the forgone opportunity. Under piece rates, the salient aspect of the effort task is speed, therefore democracy has a second-order effect on speed. Efficiency is lower in the democracy treatment. Democracy has no effect on speed under the flat rate scheme. Democracy, however, has a positive effect on precision under the flat rate, but the effect is significant only in the risky environment. Both happy and unhappy subjects work more precisely. Regression analysis show that in the risky environment flat-rate in the democracy treatment speed is lower. This result can be rationalized as gift exchange, i.e. democracy has a positive effect on quality because workers appreciate the opportunity of self-determination. There is no trade-off between pleasing the employer and the subject's own payment. In this case, the salient factor of the effort task is precision, thus democracy affects quality and not speed. Efficiency is also higher in the democracy treatment. Combining our findings, overall our results broadly suggest that democracy may have adverse effects when paired with high-powered incentives (performance pay) and positive effects when paired with flat incentives.
Presenter: Rémi Suchon, GATE CRNS Title: Does upward mobility harm trust ? Abstract: In this paper, we design an economic experiment aiming at capturing the main traits of upward social mobility and measure its impact on behaviors in a trust game (Berg et al., 1995). Upward mobility is a form of social mobility according to which an individual from a low status group is promoted to a high status group. In some developed societies, this upward mobility is still relatively rare and inherited group membership strongly predicts one's final position in the social hierarchy. An example of upward mobility is a child from a blue-collars family who ends up surgeon. While this situation is possible, it remains rare. Upward social mobility has long been considered a cornerstone of modern democracies because preserving some opportunities for every members of the society is pivotal in preserving social stability (de Tocqueville, 1835;Acemoglu et al., 2016), and because it's a fundamental component of a widespread notion of justice (Sen, 1980). In this paper, we focus on the effects of upward mobility on a specific expression of social preferences : interpersonal trust. Trust is a very important dimension of social capital, because in most economic interactions, agents have to trust each others (Arrow, 1972). Interestingly, while trust and social mobility seem both desirable, there is some evidence outside economics that upward mobility disrupts social preferences. For instance, some papers in social psychology suggest based on survey data and qualitative analysis that mobile individuals are often disapproved by left-behinds and not considered as equal by the members of the achieved group (Derks et al., 2015; Kulich et al., 2015). Then, the aim of our paper is to assess whether there is a tension between upward social mobility and trust. Experimental design We design an economic experiment in which we generate upward mobility, and measure its impact on interpersonal trust. We build on social psychology conceptualization of upward mobility, which can be thought of as the discrepancy between one's original and one's achieved group membership. We use a natural group affiliation to generate original group affliation and we induce differentiated status within our experiment. There are two original groups and two status. Every participant is affliated with one original group and achieves one of the two status. Consistently with the definition of upward mobility, the original group affliation is a good predictor for status. In practice, the experiment is divided in two parts : In the first part, we generate social mobility. The original group affiliation is done by recruiting participants from the local business school (EM Lyon) and engineer school (EC Lyon). While being similar in prestige and in educational requirements, those two schools differ in their curriculum : selection and education in the engineer school is more math-oriented. We generate status by assigning participants to one of two roles : experts or agents. This difference in status is reinforced by a task in which experts have to evaluate the quality of agents' work. We assign participants to one of the two status depending on their performances in a math quiz. Participants from the engineer school are expected to, and do perform better, so that they represent the majority of experts. By design, we insure that some participants from the business school who performed well in the math quiz are affiliated to the expert group. In our experiment, participants can have one out of four possible identities : -- Upwardly Mobile : Expert, Business school -- Left Behinds : Agent, Business school -- Downardly Mobile : Agent, Engineer school -- Expected experts : Expert, Engineer school In the second part, every participant takes part in four consecutive trust games as truster and as trustee. Each trust game corresponds to a different match. For each match, both players know both the original group affiliation and the status of his counterpart. We use the strategy method for the decision of the trustee and give no feedback between matches in order to minimize learning. For each trust decision, we elicitate the truster's beliefs about how much the trustee will reciprocate in an incentive-compatible manner. Our main variable of interest are trust and reciprocity for matches that include an upward mobile individual. Preview of the results Data show that upward social mobility does have an impact on trust and reciprocity. First, the upwardly mobile individuals trust left-behinds less than other left-behinds do. Surprisingly, we also show that upwardly mobile individuals are as reciprocal toward left-behinds than other left-behinds. This means that the decrease in trust is not due to a decrease in unconditional kindness. Moreover, our data suggest that the channel of beliefs does not explain this pattern2. Our interpretation is that mobile individuals are intrinsically more reluctant to let their fate in the hands of left-behinds. They anticipate a greater disutility in case left-behinds don't reciprocate their trust. In the wording of Bohnet et al. (2008) or Aimone and Houser (2012), mobile individuals are more averse to betrayal than non mobile individuals . Our data also point to the fact that Upward Mobile individuals experience Nobless oblige : they are more likely to choose transfers that lead to payoff equality as trustees, and less likely to choose to merely reimburse trust. On the other hand, our data shows little difference in transfers toward Upwardly Mobile individuals compared to transfer toward Left-Behinds. This hints to an asymmetrical effect of upward mobility on trust. Acemoglu, D., G. Egorov, and K. Sonin (2016, April). Social mobility and stability of democracy: Re-evaluating de tocqueville. Working Paper 22174, National Bureau of Economic Research. Aimone, J. A. and D. Houser (2012). What you don't know won't hurt you: a laboratory analysis of betrayal aversion. Experimental Economics 15 (4), . Akerlof, G. a. and R. Kranton (2000). Economics and identity. The Quartely Journal Of Economics 115 (3), Arrow, K. J. (1972). Gifts and exchanges. Philosophy and Public Affairs 1 (4), Ashraf, N., I. Bohnet, and N. Piankov (2006). Decomposing trust and trustworthiness. Experimental Economics 9 (3), Berg, J., J. Dickhaut, and K. McCabe (1995). Trust, reciprocity, and social history. Games and Economic Behavior 10 (1), Bohnet, I., F. Greig, B. Herrmann, and R. Zeckhauser (2008, March). Betrayal aversion: Evidence from brazil, china, oman, switzerland, turkey, and the united states. American Economic Review 98 (1), de Tocqueville, A. (1835). De la democratie en Amerique (Democracy in America) Union generale d'editions. Derks, B., C. van Laar, N. Ellemers, and G. Raghoe (2015). Extending the queen bee effect: How hindustani workers cope with disadvantage by distancing the self from the group. Journal of Social Issues 71 (3), Kulich, C., F. Lorenzi-Cioldi, and V. Iacoviello (2015). Moving across status lines: Low concern for the ingroup and group identification. Journal of Social Issues 71 Sen, A. (1980). Equality of What? Cambridge: Cambridge University Press. Reprinted in John Rawls et al., Liberty, Equality and Law (Cambridge: Cambridge University Press, 1987).
Presenter: Efsan Nas Ozen, University of Bologna Title: Blind Rage: An experiment on the determinants and consequences of self-harming, indiscriminate punishment Abstract: The stark increase in the within-country inequality has exacerbated social conflicts over the last decades. In western societies with occupational and income segregation between the native and immigrant populations, this tension contributes to the frustration and restricted integration of the disadvantaged with the native society by hindering economic interaction of the disadvantaged with the advantaged community. In more extreme forms, this frustration and despair can spark among the have-nots and lead to indiscriminate forms of violence and punishment. In this work, we focus on a particular form of punishment, which targets members of a sub-population indiscriminately, and is not driven by any information on the specific past behavior of the individual victims. In a controlled laboratory experiment, we aim at studying whether this sort of "blind rage" is likely to emerge in highly unequal contexts, and hit the rich in situations when they can or cannot be blamed for the presence of the economic disparities. We develop a novel zero-sum 2-player game - the Inequality Game - and assign participants to two groups. In equilibrium, one (advantaged) group earns 90% of the money, hence leaving the other (disadvantaged) group with just 10% of the pie. Unlike in other experiments, inequality is ingrained in the structure of the game and arises endogenously. We vary the degree of responsibility of the advantaged group in the inequality generation process. In Simultaneous treatments, the members of the advantaged group can shield behind strategic motives and claim that inequality is unavoidable in the game. In Sequential treatments, instead, the members of the advantaged group have the chance to equalize the earnings across groups. To model blind rage within such a context, we introduce the possibility to "exit" the game. The choice to exit can be taken only before playing the game, and hence before knowing the action take by the counterpart. By choosing to "exit", a player destroys all the money at stake in the game, so this action is harmful for both players and extremely costly from the social point of view. We look at the choice to exit as a particular form of punishment that we dub preventive punishment. Differently from other more standard forms of second and third-party punishment, preventive punishment does not provide any incentive to the advantaged players to share resources more equally. For example, while in the Ultimatum Game the proposer has an incentive to increase his offer to the responder if he fears that she will reject, such a strategic incentive is not present in our set-up. We find first that, the potential responsibility of the advantaged group in the creation of inequality does not induce the advantaged group to redistribute, and players converge to the Nash equilibrium. The existence of responsibility (the Sequential game) induces the disadvantaged players to exit more often, however, neither the threat nor the experience of exit has a strong impact on the distribution behavior, causing inequality to increase. Our results help better understand the relationship between inequality, individual responsibility and acts of blind rage in divided societies. Our experiment also contributes to the vast experimental literature on punishment by studying a new form of self-sacrifice unrelated to any strategic motive. The experiment is also linked to the emerging literature on judicial errors in distributive choices.
Presenter: Marcel Preuss, University of Mannheim Title: Why fairness views are income-dependent Abstract: I investigate how fairness views regarding inequality are shaped by individual success and income. Previous research has shown that most people consider inequality as fair if it is due to differences in choices or effort but unfair if it is due to factors beyond individual control, such as luck. To extend these findings, I ask whether people also take into account the likelihood of chance playing a role without knowing the realization of it - as it is often the case in reality. Further, a self-serving bias implies that if there is uncertainty about the probability of (bad) luck affecting the outcome of a task, subjects will underestimate the role of chance if their own outcome is good and overestimate it otherwise. Consequently, high income people might support redistribution the least because they perceive the role of chance and thus the fairness of the earnings procedure differently. In the experiment's control treatment, two players compete in an encryption task and receive fixed earnings which depend on their relative rank. In the chance treatment, there is a 30% probability that the ranking is reverted such that the participant with a lower performance receives the high prize. Differences in a third, noninvolved participant's willingness to redistribute from the richer to the poorer player between the control and the chance treatment suggest that probability of chance - and not just the realization of it - matters for redistribution. In a third, hybrid treatment, subjects know that either the rules from the control treatment or those from the chance treatment hold for all participants of their session. By not fixing beliefs about the role of chance, we are able to attribute larger differences in redistributive choices between rich and poor decision makers relative to the other treatments to heterogeneous beliefs about the role of chance and the level of fairness, respectively.
Presenter: Fabian Paetzel, University of Hamburg Title: Acknowledging needs in a dictator game: Experimental evidence on information-sensitive giving behavior Abstract: We utilize a dictator game to analyze whether the information about a recipient has an effect on transfers. Between four treatments, we vary the information about recipients which is provided to the dictator. In the CONTROL treatment, no information about a recipient is revealed to the dictator. In the treatment INFO, the provided information contains some characteristics of the recipients, like income, social benefits, travel time to the lab and gender. In the treatment NEED, the dictator gets the information about the stated "needed experimental payoff" of recipients. In an additional treatment INFO+NEED, the dictator is informed about both the characteristics and the "needed experimental payoff". It turns out that transfers are sensitive relating to the provided information about recipients. We find that if either the recipient's income or social benefits are relatively high, transfers are significantly lower. If the recipient's travel time to get to the lab is relatively high, transfers are significantly higher. Another finding is that if a recipient states a relatively high "needed experimental payoff" without the provision of additional information which could work as an explanation for the neediness, transfers are significantly lower. Only if recipients state a high neediness and provide additionally information about the reasons for being needy (e.g. high travel time to the lab), transfers are significantly higher. We conclude that (i) dictators are information-sensitive when they decide about transfers and (ii) the acknowledging of stated neediness depends strongly on the provided information.
Presenter: Antonio Filippin, University of Milan Title: Take It or Give It: Measuring Social Preferences Using Mini-Dictator Games Abstract: Research on altruism, egalitarianism, and social preferences has increased exponentially over the last two decades. Experimental and behavioral economists have played an important role in this research program, introducing experimental designs and theoretical models that facilitate the elicitation of pro-social attitudes, their measurement, as well as their formal representation. This remarkable success however has created an embarrassment of riches: a researcher interested in the elicitation and measurement of social preferences today must choose between different methodologies that are liable, at least in principle, to deliver different results. In this paper we set out to remedy, at least in part, this state of affairs. First, we focus on the elicitation of "consequentialist" preferences over monetary distributions focusing on Mini-Dictator Games, a simple design that has been used for years by experimental economists and psychologists. We illustrate the results of a meta-analysis of experimental data that are ideally suited to deliver accurate estimates of social preference parameters, and we offer some comparisons with representative measurements carried out by other economists. Second, we try to reconcile the evidence estimating an encompassing model of social preferences with the aim of testing different models of social preferences put forward in the literature. We show that both the curvature of the utility function and a different behavior under favorable and unfavorable inequality are crucial features of social preferences. Using the classic models in the literature likely delivers biased estimates because one of the two features is missing. We also find that effciency concerns play a significant role and that role uncertainty increases the weight assigned to the recipient's payoff.
Presenter: Charlotte Saucet, GATE CNRS Title: Motivated Memory in Dictator Games: Experimental Evidences Abstract: The desire to see oneself in positive lights is a fundamental feature of humans (Benabou and Tirole, 2002). Yet, this demand for positive self-image can be challenged by the fact that most of us sometimes behave in ways that we would like to think we do not. One way to find consistency between our demand for positive self-image and our prior actions -potentially self-image threatening- is through motivated memory. Time gives us a wiggle room to forget or distort some actions we would rather not remember (Green et al., 2008, 2009). This study aims at understanding how and to what extent individuals distort their memory to sustain their demand for positive self-image. We design a simple laboratory experiment where participants are first asked to play a series of binary dictator games and second to remember the amounts they chose to allocate to the receivers. In this context of social interactions, we hypothesize that dictators memory of the receivers amounts depends on their other-regarding preferences. Our preliminary results seem to show that both, the direction and the intensity of memory biases depend on individual other-regarding preferences. The experiment consists in four parts. In part 1, subjects are asked to perform a series of twelve binary dictator games as in Fehr et al. (2016). This part allows at eliciting, for each individual, ranges for parameters alpha and beta of other-regarding preferences from the Fehr and Schmidt model (1999). In part 2, subjects perform a filler task used to wipe out their immediate visual memory of part 1. In part 3, subjects are asked to recall, for each of the twelve binary dictator games played in part 1, the amounts allocated to the receivers. The key point here is not to observe whether participants are able to remember or not but whether they systematically misremember in one direction and whether the intensity of the bias depends on their other-regarding preferences. In part 4, subjects perform a neutral memory task that is used to control for memory abilities (Hilton, 2006). The experiment consists in two treatments. In the first treatment, the recollection (part 3) is incentivized based on accuracy. In the second treatment, the recollection is not incentivized.
Presenter: Lyla Zhang, Macquarie University Title: Incentives, Environment and Compliance in the Finance Profession Abstract: These days there is considerable interest in whether finance professionals comply with the risk policies of their employers, especially as incentive schemes (designed to encourage high short-term profits) often tend to discourage compliance with risk policy. Workplace environment is a potential tool to increase compliance. We introduce a controlled laboratory experiment, with employees from the Australian financial industry, to analyse the influence of monetary incentives and workplace environment on employees' compliance with risk policy. We find that fewer employees are compliant with risk policy under the piece-rate payments compared to the flat salary. And workplace environment plays an important role in risk compliance. While employees are more likely to comply with the policy in the risk-focused environment, they are more likely to disobey the policy in the profit-oriented environment.
Presenter: Florian Lindner, University of Innsbruck Title: The winner takes it all, the loser’s standing small: Social status and risk-taking in the finance industry Abstract: Rankings are a pervasive feature of the finance industry. These rankings are interrelated with extrinsic (money), intrinsic (self-image), and reputational (status) motives, which might influence investment behavior in different ways. Although status-driven behavior and reputational concerns are relevant in the finance industry, there is little evidence on how financial professionals react to rankings and reputational motives. Therefore, we model a portfolio decision with two assets and investigate how reputational motives (public announcement of the winner/loser) influence risk-taking beyond solely anonymous rankings. To test the behavioral hypotheses derived from this model, we recruit a (i) unique subject pool of 252 financial professionals and (ii) 432 student subjects. We find that reputational motives as an additional extrinsic incentive play a minor role among financial professionals, which suggests that their behavior is mainly explainable by other factors like identity importing or rank incentives. In the student sample, we find markedly higher risk-taking when the winner (loser) is announced; suggesting that rank incentives play a role also with non-financial professionals, depending on the specific implementation, especially with reputational costs and benefits.
Presenter: Michael Kirchler, University of Innsbruck Title: Market Experience and Market Efficiency - Evidence from Experiments With Financial Professionals Abstract: The efficiency of financial markets and their proneness to speculative bubbles are central themes in the academic and the professional debate in finance. Yet, surprisingly, there is no research how the main protagonists, financial professionals, behave and contribute to market efficiency and speculative bubbles. We run 38 lab-in-the-field experimental markets with 294 professionals and 48 classical lab markets with 384 students to investigate how markets populated by professionals behave in the presence of commonly known bubble drivers and bubble moderators. We find qualitatively similar patterns within each subject pool, showing that bubble drivers like money inflow or a high initial monetary supply drive market inefficiency. Moreover, we observe similar qualitative patterns within subject pools for bubble moderators as well, indicating that short selling and a low and constant monetary base increase market efficiency. However, we also show that show that across subject pools markets populated by professionals are significantly more efficient and less prone to bubbles compared to student markets. We conclude that real-world market experience is conducive to the efficiency of financial markets.
Presenter: Frederic Schneider, Yale School of Management Title: Promises over Time Abstract: Economic experiments consistently show some degree of voluntary cooperation, and some degree of trust in others to be cooperative. This cooperation and trust is enhanced when people are given the opportunity to make promises. Trust, cooperation, and the propensity to honor promises have thus been proposed as a means of achieving efficient outcomes in incomplete contracts, social dilemmas, and situations of asymmetric information. Laboratory experiments, however, do not capture a crucial feature of real world interactions, the time gap between the scheduling of an action and its conclusion. For example, lenders transfer money to borrowers, who promise to return the money at a later time. To examine whether the time gap between the scheduling of a cooperative action and its execution, we run an experimental trust game where trustees make their decisions not during the initial laboratory session, but during an online survey, either within 24 hours after the session or three weeks later. We also vary whether trustees can make promises or not. We find that the three-week gap does not decrease the propensity to cooperate; promises increase cooperation, but we again see the no change in cooperation over time, as in the case of no promises. Our results show that the previous body of short-duration laboratory experiments on trust and promises likely translates to longer-term interactions. Thus, in real-world situations, lags between the beginning of the interaction and the time to respond may not substantially alter the trustworthiness of the responder.
Presenter: Sven Simon, Max Planck Institute for Tax Law and Public Finance Title: Deception under Time Pressure: Conscious Decision or a Problem of Awareness? Abstract: Time is a crucial determinant of deception, since some misreporting opportunities come as a surprise and require an intuitive decision while others allow for extensive reflection time. This paper provides experimental evidence on the role of the time dimension for dishonest decision-making. We conduct a laboratory experiment of self-serving deceptive behavior and exogenously vary the level of reflection time. We find that time pressure leads to more honesty compared to sufficient contemplation time. Moreover, dishonest subjects need more response time compared to honest subjects. In addition, we decompose misreporting into two components: first, the cognition process of the misreporting opportunity and, second, the conscious decision to misreport. This decomposition reveals that more reflection time increases awareness of the misreporting opportunity. However, it has no effect on the conscious decision of whether to misreport or not. Due to subjects' lack of awareness under time pressure we conclude that misreporting is not the intuitive response.
Presenter: Wolfgang J. Luhan, University of Portsmouth Title: Lying for others: The impact of agency on misreporting in in an economic laboratory experiment Abstract: We extend two streams of literature, the study of misreporting and lying aversion and the study of agency (acting on behalf of others), and we propose a new experimental method to assess misreporting and lying aversion as an alternative to the prevalent method. In a situation where misreporting of private information (lying) is beneficial, the standard economic theory assumes that behaviour is driven by the trade-off between potential gains and the expected fines associated with detection (e.g., Becker, 1968 on crime, Allingham and Sandmo, 1972, on tax evasion). Without possible detection, people are expected to lie and maximise payoffs. A growing literature in behavioural economics (and psychology) shows that people in general do not lie to the full extent even when detection is impossible, which is partly explained with “lying aversion†(e.g., Benabou and Tirole, 2002; Mazar et al., 2008). When deciding for others, people have been found to behave more conservatively and take fewer risks due to a feeling of accountability (Charness and Jackson, 2009); other studies have found them to act as if there were no consequences to their actions (e.g., Harrison, 2006). The possible behavioural differences between reporting for oneself or for others are manifold. As benefits for others are irrelevant in economic standard theory, lying aversion would predict truthful reporting. Responsibility might lead to more misreporting in order to maximise the beneficiaries' payoff, which might be fostered by accountability. As many economically relevant situations involve both, a demand for revealing private information and some form agency (e.g., tax counselling, financial investments), we need to establish the motivational factors and their impact such situations. We develop a theoretical framework for this decision context, identifying motivational factors and expected behavioural patterns. Using laboratory experiment we test the emerging hypothesis, discriminate behavioural types and identify the models with the highest predictive power. We design an alternative lab procedure to the prevalent die-under-the-cup method (Fischbacher and Follmi-Heusi, 2014) where participants secretly roll a die and are paid according to the (no-verifiable) self-reported outcome. The common analysis for this game is to assume uniform distribution as the true outcome and treat any excess reporting of one number as a lie. Because this distributional assumption is only valid at large sample sizes, we propose a different approach that allows to identify the true range of misreporting and the exact extend of lying aversion. Using lottery tickets we can observe the actual distribution of real draws while maintaining the unobservability of individual draws.
Presenter: Tommaso Reggiani, Masaryk University Brno Title: The (neutral) effect of Transparency on Information Disclosure Nudge Abstract: With the increasing popularity and application of the nudging concept, several ethical objections against it have also emerged due to its subtle architecture. In order to contribute to the debate, this paper examines nudging in the light of transparency. A transparent nudge is when the citizen being nudged knows the intention behind it and the means adopted to pursue the nudging object. A non-transparent nudge works in a way that the citizen cannot reconstruct the object of the intervention and the means by which the behavioral change is pursued. Wide - philosophical - consent claims how nudging can only be regarded as ethically acceptable when the nudge is transparent (Fischer; Lotz 2014; Sunstein 2015). Hausman; Welch (2010) demand for nudging transparency, even if it potentially undercuts the effectiveness of the nudge. However, empirical research on nudging often shies away from incorporating explicit transparency. We design a survey-experiment to examine whether transparency has negative impacts on nudging effectiveness in its configuration of information disclosure nudge (Loewenstein et al. 2015; Wisdom et al. 2010). In partnership with the German Federal Centre for Health (BZgA), we address the very salient and relevant case of the voluntary participation into the German organ donation initiative. We do not find evidence that transparency inhibits the effectiveness of information disclosure nudges. Our finding supports the policy-relevant claim that information disclosure nudges can be transparent and yet effective, reconciling at the same time philosophical and practical issues. This result is in line with recent findings by Bruns et al. (2016): In a lab experiment, focusing on default nudge applied to contributions to carbon emission reduction, they do not find evidence that transparency inhibits the effectiveness of a default nudging strategy. Similarly, Loewenstein et al. (2015) testing the interaction between transparency and defaults, get to similar conclusions according to results from an online-study focused on medical-care related choices.
Presenter: Adriaan Soetevent, University of Groningen Title: Tailored Feedback and Worker Green Behavior: Field Evidence from Bus Drivers Abstract: How to engage workers in conservation efforts when the company pays the bill? We collaborate with a large public transport company and investigate the potential of targeted peer-comparison feedback and on-the-road coaching among 409 bus drivers. In an on-going natural field experiment, drivers are assessed on multiple driving dimensions and randomly assigned to individualized reports with varying numbers of peer-comparison messages. Coaches operate alongside these reports and maintain detailed logs that pinpoint driver-specific coaching moments. Benefiting from the introduction of electronic on-board recorders (EOBR) in the entire fleet of buses, we gather over 800,000 trip-level observations to evaluate both feedback programs. We find that peer-comparison messages are detrimental to driving behavior when drivers received prior coaching. Uncoached drivers, in contrast, show some improvement after being exposed to the messages. Coaching itself positively affects driving on multiple dimensions but these effects diminish over time.
Presenter: Josie Chen, National Taipei University Title: How to Motivate Students to Reduce Dormitory Electricity Use: A Field Experiment Abstract: This study investigates how to motivate students to reduce dormitory electricity use. We ran a field experiment in college dormitory rooms. We grouped three rooms into a team and ran an energy-saving competition for five weeks. At the beginning of the contest, the subjects were honestly told that each team was equally likely to win the competition. We provided the subjects with small monetary incentives to win the competition. We conducted a between-subjects design and randomly assigned subjects to different feedback treatment groups. Each week, we sent messages telling the subjects their team's ranking, their own intra-team ranking, or both.
Presenter: Kohei Nitta, Toyo University Title: Are Females more or less likely to whistleblow? An Experimental Investigation Abstract: We use an experimental game based on the multiplayer extension of the asymmetric information Ultimatum game to study gender differences in whistleblowing behaviours. Our experiment finds no gender differences in whistleblowing willingness even after controls for whistleblowing beneficiaries and the risk of retaliation to whistleblowers.
Presenter: Petra Nieken, Karlsruhe Institute of Technology Title: Bad boys, good girls?! - what causes the gender gap in sabotage Abstract: We study gender differences regarding performance and sabotage in competitions. While we find no systematic gender differences in performance in the real effort task, we observe a strong gender gap in sabotage choices in our experiment. This gap is mainly driven by differences in beliefs about sabotage. Males expect to suffer from sabotage to a higher degree than females and choose higher sabotage levels. Providing subjects with a noisy signal about the sabotage level they might suffer from, leads to an alignment of beliefs and eliminates the gender gap in sabotage. Analyzing the empirical sabotage response function as well as elicited measures for risk aversion and status seeking reveals that differences in those preferences are not a main driver of the gender gap in sabotage. Beside offering an explanation for the gender gap in sabotage, this paper provides valuable insights in how to restore competitive balance between females and males.
Presenter: Janina Kleinknecht, Ulm University Title: A man of his word? An experiment on gender differences in promise keeping Abstract: Cooperation between one or more parties is vital for economic efficiency. Frequently, cooperation requires that the involved parties rely on informal commitments, since not all possible situations can be contracted. Thus, promises as well as expectations about promise keeping play an important role in strategic environments. Numerous experiments have investigated the credibility of informal commitments and shown that they are more than cheap talk. From the literature on social preferences, it is well known that gender differences exist and they may contribute to explaining gender differences in (labor) market outcomes. This paper systematically investigates gender differences in promise keeping by extending the experimental design of Vanberg (2008). I find no gender differences in the likelihood of giving promises but promises raise expectations of women more than those of men. Consistent with this, women break promises less often than men, although both, men and women, know that a promise raises expectations on the part of its receiver. Moreover, the experiment reveals a gender difference in the motivation for promise keeping. For men, a promise seems to be a binding commitment only to a specific subject since they are less likely to fulfill promises by third parties. For women both matter, their own promise as well as the expectations of the receivers. If women make a promise themselves, they are more inclined to keep a promise made by a third party than men. The overall results are strengthened by the fact that without promises there are no differences in cooperative behavior of men and women.
Presenter: Aidas Masiliunas, Aix-Marseille University Title: Payoff risk and foregone payoff information in contests Abstract: We seek to understand the reasons behind overbidding in Tullock contests by explicitly testing a hypothesis that choices deviate from Nash equilibrium predictions because a boundedly-rational learning process converges very slowly or not at all. In an experiment we manipulate two elements that slow down payoff-based learning. The first element is foregone payoff information, which in reinforcement learning models increases the rates of Nash equilibrium play. The second element is payoff risk, which reduces the correlation between realized and expected payoffs and therefore slows down learning. We manipulate the sources of payoff risk using a 2x2 design: payoffs from contest investments are either risky (as in standard contests) or safe (as in proportional contests), and payoffs from the part of endowment not invested in contest can also be either safe (as in standard contests) or risky. Payoff-based learning predicts faster convergence when all payoff risk is eliminated. Other theories proposed in earlier literature ("joy of winning", "fear of losing", risk-seeking preferences, aversion to realized payoff inequality) that can explain over-investments in the standard contest predict no overbidding in at least one of the other three treatments. In contrast, we find significant overbidding in all treatments at the start of the game when foregone payoff information is not available, but in the treatment with no payoff risk all groups converge to the Nash equilibrium once foregone payoff information is introduced. In other three treatments, overbidding persists even with foregone payoff information, although Nash equilibrium play rates increase.
Presenter: Jingcheng Fu, University of Nottingham Title: Gender differences in performance and "bouncebackability": Evidence from a high-stake exam Abstract: This paper investigates gender differences in performance in a highly competitive college entrance examination. The exam system offers candidates the chance to take the exam for some of the subjects twice, scheduled 6 months apart, and only the better result accounts for the admission. We find that the females tend to underperform in the first sit compared to their mock performance, however, they improve more in the second sit. Combining the two sits, there is no significant gender difference in the final result. This is due to that among the candidates who underperform in their first sit, in the second sit the females "bounceback" more from the negative shock, by a larger improvement in their academic abilities over the period after the first sit.
Presenter: Radosveta Ivanova-Stenzel, Technical University Berlin Title: Effort choice in dynamic competitive environment Abstract: High-powered incentives in a competitive environment are usually assumed to lead to higher agents' efforts (provided the competitive pressure does not become too high, such that agents exit the game, shirk or choke). However, performance in dynamic competitive environment, e.g., when agents participate in a series of tournaments, could be impaired, since continual competitive pressure may cause stress and fatigue of decision. We study effort choice behavior in a sequence of three two-player tournaments where we induce fatigue and recovery but also manipulate the severity of competition. We consider two scenarios: (i) all three tournaments feature low and uniform incentives over time in the form of a small prize spread between winner's and loser's prizes; (ii) the pattern of low and uniform incentives is interrupted by a period of high incentives (a high prize spread). We study how this manipulation combined with variation of the fatigue level affects effort choice in the second period (with high incentives) as well as in the periods before and after (with low incentives). We derive theoretical predictions based on the standard, neoclassical approach and test whether these predictions suffice to explain the observed behavior patterns.
Presenter: Eugenio Levi, Sapienza University of Rome Title: I and we: spillovers between individual and social tasks Abstract: Individuals often engage both in individual tasks and in social tasks that require different types of motivation. We investigate with an experiment if engaging in one type of task undermines performancein the other. A related question concerns whether any interaction of this kind is exacerbated or triggered by the use of nudges that are designed to boost performance in each type of activity. To examine these issues we use a combination of games (a real effort task and a public good game). We find that engaging in the public good game first improves performance in the real effort task, because low relative contributors engage in it more effectively. Contrary to much evidence in the literature, a performance nudge and a group identity nudge per se do not significantly affect behaviour in the correspondent decision task. However, when both nudges are implemented, subjects significantly increase their contributions to the PGG and weakly improve learning in the real effort task.
Presenter: Tomas Miklanek, University of Economics, Prague Title: Do Fixed-Prize Lotteries Crowd-Out Public Good Contributions Driven by Social Preferences? Abstract: Fundraising for public goods by private contributions is often undermined by free-riding. One prominent mechanism suggested to alleviate this problem is a fixed-prize lottery with winning probabilities proportional to individual contributions (Morgan, 2000, Morgan and Sefton, 2000). This design increases contributions by coupling financing of the public good with a private monetary incentive to win the prize. Yet, as extensively documented by economic experiments, subjects often contribute even in the absence of incentives of this kind, suggesting that their contributions are driven by social preferences. This raises a question of how the lottery incentive interacts with such preferences. We present an experiment in which we de-couple one's own ability to win the prize from the ability of the others to do so. This way, we de-couple the contribution effect of own prize seeking from the potential crowding out effect due to the perception that the others contribute because of their prize seeking, rather than to benefit the group. Even though the lottery increases contributions relative to the voluntary contribution case, we find that it also significantly crowds out voluntary contributions that are likely driven by social preferences.
Presenter: John Tisdell, University of Tasmania Title: Just Shares Abstract: Social norms and notions of just shares are often cited to explain individual extraction decisions from common pool resources. This paper explores whether and under what conditions individuals extract a self-just share of a common pool resource. The notion of a self-just share is very different from a just share and has different implications for human behaviour. Self-just could imply selfish behaviour on the part of subjects where they try to find justification for their actions. The notion of a just share, on the other hand, is related to issues of fairness and equity. Through a series of take-some public good experiments we look for evidence of self-just behaviour under different endowment and extraction information conditions. We add to the body of knowledge by adding to the understanding of choice in common pool situations, specifically exploring the choice of equality or proportional equality.
Presenter: Lijia Tan, Wang Yanan Institute for Studies in Economics Title: Procurement in the Presence of Supply Disruptions Abstract: We consider a setting with a supply disruption that may occur after a contract is negotiated. If the disruption can be corrected at a cost, and it is more efficient for the supplier, rather than for the buyer, to correct the disruption, a contract that does not include the disruption contingency leads to inefficient outcome. We examine the problem in the laboratory, where we consider the effect of four factors on contract performance: the bargaining protocol, the ability to renegotiate, the long-term relationships, and the ability to communicate. We find that long term relationships and the ability to renegotiate have the predicted effect - inefficiency due to disruptions decreases. The ability to communicate also has a large positive effect by decreasing inefficiency due to disruptions. Contracts negotiated under free bargaining protocol turn out to have more failures than contracts negotiated using a take-it-or-leave-it offer. We propose several behavioral explanations.
Presenter: Christine Grimm, WU Vienna Title: Bargaining, Asymmetric Information and Strategic Communication - An Experiment Abstract: Interactions between suppliers and buyers often face some asymmetric information. This holds potential for efficiency losses in markets that might be overcome by sharing information. In an Ultimatum Game context, we experimentally examines the strategic choice of communication with different reliability and its effects on efficiency and payoff distributions. We replicate previous findings that asymmetric information results in efficiency losses, and show that these losses affect both the informed and the uninformed parties. Communication leads to an improvement of the situation. However, reliable communication does not reach its full potential while cheap-talk communication also has an impact. Communication effects are asymmetric in that mostly proposers benefit from communication, be they informed or uninformed.
Presenter: Ulrike Vollstaedt, University of Duisburg-Essen Title: Voluntary disclosure of product quality Abstract: Often, sellers are better informed about product quality than buyers. This asymmetric information may decrease social welfare (Akerlof, 1970). One potential remedy is to provide credible information about product quality (Viscusi, 1978). This remedy has been institutionalized to some degree in several countries in form of independent product testing organizations, e.g. Consumer Reports (US), Which? (UK) and Stiftung Warentest (Germany). Despite the many advantages that these product testing organizations offer, the provided information may neither be optimal for buyers nor for social welfare due to the procedure how product models are selected for a test. We propose a new procedure which allows sellers to apply for a test and to, in principle, voluntarily and credibly disclose their quality. In a controlled lab experiment, we show that unraveling increases consumers' rent and social welfare when this new procedure is used.
Presenter: Joe Vecci, University of Gothenburg Title: Moving up the ladder: The impact of income mobility on anti-social behaviour Abstract: The positive relationship between inequality and social immobility commonly termed "The Great Gatsby Curve" shows that the greater the distance in a country between rich and poor, the less likely one is to move from being poor to being rich. While research has documented this relationship across countries, less is known about the impact of income mobility on behaviour. In this paper, we use a laboratory experiment to examine the impact of income mobility on anti-social behaviour. In our experiment, subjects participate in a modified version of an investment game where they can reduce others' payoff at a cost to themselves. Subjects are assigned to either the high income or the low income group. A unique feature of the experiment is that low income subjects are given the opportunity to move to the high income group. We find anti-social behaviour decreases two fold when subjects can move up the income ladder. To understand possible mechanisms underlying this result, we design treatments which vary the original income assignment- subjects are assigned to the high income group at random or by exerting high effort and we also vary the mobility mechanism- subjects have the potential to move up the income ladder by chance or by exerting effort. We find that the mechanism that assigns subjects to income groups has strong explanatory power. Subjects attack less often when they are assigned to their income group at random relative to when assignment is based on effort. Results suggest that the lack of income mobility may fuel anti-social behaviour but this depends on how individuals attain their income.
Presenter: Lubomir Cingl, University of Economics in Prague Title: Are juvenile delinquents incorrigible? Experimental evidence from detention centers Abstract: This study explores the differences and similarities in norm-violating behavior between juvenile delinquents and regular adolescents. We conducted a lab-in-the-field experiment in juvenile detention centers and primary schools to study changes in norm-violation in response to two specific contexts: (i) when participants are exposed to a (un)favorable economic situation and (ii) when social values are made more salient. Our results show generally substantial similarities between problematic and non-problematic adolescents. Even though the juvenile delinquents violate norms slightly more, we found no evidence of ingroup favoritism. Moreover, both groups similarly care about their social image and are not very sensitive to positive cues. Looking at the behavioral characteristics of the delinquents, higher norm-violation is correlated with interpersonal problems but not other misbehaviors. Our findings thus show that juvenile delinquents are not inherently different to non-problematic adolescents and highlight the importance of social values as well as interpersonal problems for successful resocialization.
Presenter: Zvonimir Bašić, University of Bonn Title: The Influence of Self and Social Image Concerns on Lying Abstract: We investigate the influence of self and social image concerns on lying behavior in a die-rolling experiment (Fischbacher and Föllmi-Heusi, 2013). We exogenously manipulate self-awareness and observability, which mediate the focus of a person on their private and public self, respectively. We show that the increase of observability decreases average lying compared to a control condition, while the increase of self-awareness does not decrease average lying, but has an influence on partial liars, i.e., subjects who lie but not to the maximal extent. In particular, when exposed to higher self-awareness, partial liars overreport a smaller outcome compared to the control condition. Our results show that models of lying behavior need to include an image concern component in both image dimensions. They also indicate that a crucial motive why some subjects do not lie to the full extent is that they like to view themselves as honest. Finally, our findings offer supportive evidence that intrinsic lying costs are convex and that people care about the relative lie.
Presenter: Rostislav Stanek, Masaryk University Brno Title: Tax compliance with endogenous audit selection and heterogeneity of income Abstract: It has been shown in the experimental tax compliance literature that endogenous audit selection mechanism (ASM) increases tax compliance. However, this literature assumes that the tax authority has an unbiased observation of the actual taxpayers' income and consequently the taxpayers with the largest difference between the observed and reported income are most likely to be selected for audit. In reality the tax authority might not have unbiased information about the actual incomes as these might be observed only for taxpayers who have been selected for audit. In this case the ASM can be based only on reported incomes. The aim of the paper is to design an endogenous ASM that uses only the reported incomes and experimentally compare the tax compliance under the endogenous and random ASMs. We develop a theoretical model where taxpayers have heterogeneous income and the ASM is based only on the reported income. We show that in the symmetric Bayes-Nash equilibrium the proposed endogenous ASM entails a higher reported income than the random ASM. We test predictions of the model in an economic experiment. Each experimental session consists of 30 rounds. In each round, subjects may be selected for audit with a certain audit probability. In the random ASM, the probability of audit is exogenous and the same for all taxpayers. In the treatments with endogenous ASM, the subjects are divided into groups of five taxpayers and their audit probability is decreasing in the difference between eac subject's reported income and the average reported income of the other four subjects in their group. We use partner matching in order to increase the learning effect. At the beginning of each round, all taxpayers receive income that is drawn from a uniform distribution. Their task is to choose the reported income. Subjects selected for audit pay a penalty equal to their unreported income. In all situations, the reported income is taxed at a fixed tax rate. The experiment confirms the theoretical prediction that taxpayers report a higher income under the endogenous ASM than under the random ASM.
Presenter: Ondrej Krcal, Masaryk University Brno Title: Killing two birds with one stone: Reducing fiscal and welfare loss of tax evasion Abstract: Tax evasion may cause not only fiscal losses to the tax authority, but also social welfare losses due to taxpayers' investment in the concealment of tax evasion. We develop a theoretical model in which the audit probability depends on two main parameters: basic audit probability and a sensitivity to reported incomes of taxpayers. The model predicts that while a rise in the sensitivity reduces both losses of tax evasion, an increase in the basic audit probability reduces only the fiscal, but increases social losses of tax evasion. We test these predictions using experimental methods and find a pattern consistent with the theoretical model. This result suggests that an endogenous audit selection process where the audit probability depends on taxpayer information leads to better fiscal and social outcomes than increasing audit frequency.
Presenter: Luigi Butera, University of Chicago Title: The Welfare Effects of Social Image: a field experiment Abstract: People respond strongly to the prospect of receiving positive or negative social recognition. As a result, social recognition is often used as a nudge to induce desirable behavioral changes. While social image nudges may be cost-effective compared to other forms of incentives, such as monetary incentives, little is known about their effect on the welfare of those who are nudged. We ran a field experiment with the YMCA where we incentivized people to attend the local YMCA gym for a month through monetary incentives - either benefitting the participants or a charity - and through social recognition of their effort. We elicit people's preferences for social image and observe how it impacts their actual attendance. We find that our social recognition nudge increases average attendance by 24%. However, people largely overestimate ex-ante their ability to benefit from social recognition, resulting in significant average welfare losses. Overall we show that social image nudges, while very cost-effective, they may be a negative sum game if people's biases are not taken into account when nudges are designed.
Presenter: Stefania Bortolotti, University of Cologne Title: Solidarity and Cooperation in an Ageing Society Abstract: One need not look far to realize the importance of solidarity and cooperation in today's world. The "need for solidarity" lurks everywhere, from political issues such as the Greece debt crisis or the global refugee crisis to more mundane solidarity contracts such as all sorts of risk insurances or solidarity taxes. Maintaining the social fabric of modern societies necessarily requires solidarity and cooperation among people with different walks of life and different generations. While a few previous studies show that preferences can change across the life span, there is little or no evidence about solidarity and cooperation among people from different age cohorts. To understand solidarity and cooperation across generations, we run an experiment in two large shopping malls and during a senior fair. Participants played a Dictator Game and a Prisoner's Dilemma and could condition their decisions on the age of the matched partner. We find that behavior change with the age of the matched partner but we do not find any sign of out-group discrimination. Our data suggest that older generations are more generous toward younger partners in a Dictator Game. This positive attitude toward younger matched partners helps sustaining cooperation even though younger participants are expected to be less cooperative than other age cohorts.
Presenter: Michalis Drouvelis, University of Birmingham Title: Information provision about personality types and pro-social behaviour Abstract: We test experimentally the impact of knowing one's own and others' personality type on pro-social behaviour by observing play in a one-shot modified dictator game and a public goods game. Our focus is on the personality trait of agreeableness which has been shown to be a significant determinant of pro-sociality. Our results confirm the positive effect of agreeableness on pro-social behaviour. We also provide robust evidence that knowing one's own and others' low levels of agreeableness has detrimental effects on pro-social behaviour as compared to the baseline no-information benchmark. This is not the case, however, for subjects characterised by high levels of agreeableness. Our results provide important implications relevant for human resource selection processes involving a personality assessment of their employees.
Presenter: Catherine Eckel, Texas A&M University Title: Does How We Measure Altruism Matter? Playing Both Roles in Dictator Games Abstract: Altruism has been measured in the lab with dictator games, sometimes with single decisions and sometimes with subjects making multiple decisions that vary both the endowment and the relative price of transferring resources. In single-decision games, subjects typically are paid for one role, either as dictator or recipient. But in most multi-decision experiments, subjects are matched and paid twice: once in a dictator role, and again as a recipient. In one-game experiments, results show a strong preference for equality. However, in multi-decision, dual-role designs, the results indicate a strong preference for efficiency over equality. In this paper we unpack this result by conducting multi-decision experiments with two payment schemes. In the single role treatment (SR), subjects are assigned as either dictators or recipients, and paid for one randomly-selected decision in one role; in the dual role treatment (DR) all subjects make decisions in the dictator role, and are matched twice and paid for two decisions, once in each role. While preferences for giving in both treatments are rational in the sense of being consistent with GARP, we find subjects in DR display greater sensitivity to the price of giving compared to subjects in SR, where equal division is more prevalent. When estimating individual CES utility functions, we find that the distributions of preference types are quite different between the two treatments: Subjects in DR are substantially more efficiency-focused and more likely to be selfish than the SR subjects. In an additional study we show that the differences in distributions of preferences are not due to the differences in total payoffs between the two treatments. Our results suggest that the DR design distorts elicited preferences, exaggerating the preference for efficiency over equality.
Presenter: Cameron Belton, University of East Anglia Title: Choice Effects and Charitable Giving Abstract: In this paper we implement a novel experimental design which attempts to capture a pure effect of choice on charitable donations. Participants are endowed with an amount of money and may donate as much or as little as they wish to a real charity. In a "choice" treatment we allow participants to choose which charity they wish to donate to. In a "no-choice" treatment, we allocate a charity to participants which they may donate to. In prior experiments testing for choice effects, the preferences of "no-choice" individuals are not accounted for, potentially inflating findings of a choice effect. The novel design in this experiment is that we attempt to match "no-choice" participants with the charity they would have chosen, had they instead been given the opportunity to choose. This is achieved through the use of a series of preference-elicitation tasks. Allowing participants to explicitly choose which charity they wish to donate to causes a significant increase in both the likelihood to donate and actual amount donated. The paper discusses the implications of this in the implementation of donation strategies and public policy.
Presenter: Emanuela Lezzi, University of Genova Title: Charitable giving and donors' profile: an experimental investigation Abstract: A challenge faced by researchers and the philanthropic community is the understanding of how and why donors give. In this study, we experimentally investigate the impact of personality trait (e.g. empathy and self-esteem), self-construal (e.g. a donor's general tendency to identify himself with a group), and social identity (e.g. donor's level of identification with a particular charity) on charitable giving decisions in order to delineate donors' profiles. Our experimental design allows us to classify donors by motivation (altruism, warm glow, and impure altruism): we implement fist a dictator game in which participants are given an opportunity to donate in an environment where donations are fully crowded out (Crumpler and Grossman, 2008). Second, we use another dictator game where participants are given another opportunity to donate to the same charity without a crowding out scenario (Gandadharan et al., 2015). Moreover, we measure personality traits, self-construal and social identity by using psychological scales. Results suggest that altruistic behaviour not only depends on private preference either for others' well-being or for personal satisfaction. The act of giving to a specific charity is correlated to the donor's level of empathy and reflects the extent to which the donor perceives to be connected or identified with a specific charity.
Presenter: Marcus Giamattei, University of Passau Title: Animal Spirits and Foresight in a Price-Setting Experiment Abstract: Calvo (1983) introduced the most common way to model nominal price rigidity. We experimentally test Calvo pricing and the implied level of foresight with a large classroom audience and also across 20 repetitions (lives) in the laboratory. During the first two lives subjects fail to engage in foresight, potentially being excessively optimistic and overestimating probabilities. They engage in excessive pessimism afterwards, failing to learn optimal behavior. In spite of these biases, prices are on average close to the rational solution. Our findings raise some caveats regarding the Calvo model and reveal persistent animal spirits at the individual level.
Presenter: Rosemarie Nagel, Universitat Pompeo Fabra, ICREA, BGSE Title: Noisy Adaptive Learning in an Oligopoly Market with Demand Inertia Abstract: This paper combines the macro- and microeconomic learning literature by showing that the simple adaptive learning model with noise by Mauersberger (2016) that has been developed in a macroeconomic context can be applied to explain the learning dynamics in an oligopoly market game. We use the experimental data of Nagel and Vriend (1999), in which subjects are firms in an oligopoly market with imperfect information about their environment. Subjects need to make two decisions: how much to produce of a perishable good and the amount of signals (advertisement) sent to their consumers. With two simultaneous markets, output and signaling, the setup is comparable to a general equilibrium setting. Our simulations show that the noisy adaptive learning model can explain several features of aggregate and individual data such as the convergence speed to the equilibrium and the level of fluctuations.
Presenter: Jasmina Arifovic, Simon Fraser University Title: Learning to Believe in Simple Equilibria in a Complex OLG Economy - evidence from the lab Abstract: We set up a laboratory experiment to empirically investigate coordination and equilibrium selection in a complex economic environment. We use the overlapping-generations model of Grandmont (1985), which displays multiple perfect-foresight equilibria, including periodic and chaotic dynamics. The equilibrium selection problem is not solved under learning, as there exist learning theories that predict coordination on each of those equilibria. We find that subjects in the lab systematically coordinate on an equilibrium, despite the complexity of the environment. Coordination only happens on simple equilibria, i.e. the steady state or the period-two cycle, which are predicted only if the subjects follow simple learning rules. This suggests that relevant perfect foresight equilibria should be robust to the use of those simple rules.
CANCELED:
Presenter: Thomas Lauer, University of Cologne, Title: Conditional Dishonesty in Teams
Presenter: Adrian Chadi, University of Trier Title: Little Lies and Blind Eyes - Experimental Evidence on the Implications of Cheating for Team Performance Abstract: This experimental study analyses how co-workers are affected in their task performance when members of their team cheat. To this end, we establish a workplace scenario in which three individuals have to do a real-effort task in order to receive pay based on team output. Between two work rounds, it is possible for two of them to obtain a bonus through a dice game, which either allows for cheating or not. By studying worker behavior when cheating takes place, compared to when it does not, we analyze the possible implications of deception on team outcomes, without using deception as experimenters. Our study on cheating in a workplace context sheds light on the economic relevance of information preferences. We find that individuals with low preferences for obtaining inconvenient information do not respond negatively to the event of cheating. It seems that these workers intentionally turn a blind eye to the unethical action of team members in order to maintain motivation and keep task performance high. Those that we identify as information seekers, however, respond to co-worker cheating negatively. We conclude that a key reason why lies at the workplace may not be harmful to team outcomes is the intentional reluctance of a significant share of individuals to actually realize that their colleagues are cheaters.
Presenter: Peter Werner, Maastricht University Title: Evasive Lying in Strategic Communication Abstract: Information asymmetries in economic transactions are omnipresent and a regular source of fraudulent behavior. In a theoretical and an experimental analysis of a sender-receiver game we investigate whether sanctions for lying induce more truth-telling. The novel aspect in our model is that senders may not only choose between truth-telling and (explicit) lying, but may also engage in evasive lying by credibly pretending not to know. While we find that sanctions promote truth-telling when senders cannot engage in evasive lying, this is no longer true when evasive lying is possible. Then, explicit lying is largely substituted by evasive lying, which completely eliminates the otherwise positive effect of sanctions on the rate of truth-telling. As outlined in our model, the necessary prerequisite for such an "erosion" effect is that evasive lying is perceived as sufficiently less psychologically costly than direct lying. Our results clearly demonstrate the limitations of sanctioning lying to counteract the exploitation of informational asymmetries and may explain the empirical evidence from the finance industry that sanctions for financial misconduct eventually appear to be not very efficient.
Presenter: Martin Angerer, University of Liechtenstein Title: Information salience, cognitive load, and trading behavior: Evidence from experimental asset markets Abstract: The availability of information about asset characteristics and the way how this information is presented affect decisions and trading behavior in financial markets. We study experimental asset markets, which differ in the presentation of information concerning the difference between market prices and the asset's fundamental value. In a framework with identical underlying information, we find that prices are more symmetric and more efficient when subjects are provided with dynamic signals of real-time mispricing with respect to estimated fair prices rather than price estimates alone. A combination of numerical and color-coded graphical information has the strongest effect on behavior. Despite identical underlying information, we find that dynamic signals lead to significantly more trading activity than purely static signals.
Presenter: Peter Matthews, Middlebury College Title: Attribute Overload, Consumer Finance and Welfare Abstract: While economists have generally considered it welfare improving to offer an individual more options to choose among, a large literature has documented the phenomenon of "choice overload", where provision of additional choices to individuals can actually lead to suboptimal choices and welfare loss. We investigate a related but distinct phenomenon termed "attribute overload", where increasing the number of characteristics, or attributes, while keeping the size of the choice set constant, may in and of itself lead consumers to make welfare diminishing choices. In a two-stage experiment, we first gather time and risk preference information that is used to structurally estimate the individual-specific parameters of a CRRA utility function. For each participant the observed utility function is then used to inform a choice experiment in stage two over general purpose reloadable prepaid cards, consumer finance products with growing market share. Based on their stage two choices we can measure, for each person, simple dominance violations and the welfare implications of choosing the "wrong" card. Participants are randomized into three information architecture treatments that vary either the presentation or the size of the attribute set in otherwise similar choices. %in order to determine whether dominance violations can be decreased, and, therefore, welfare increased by adopting a simple disclosure policy. Assignment to the few-attribute treatment significantly reduces the chance of choosing dominated credit products, suggesting that indeed consumers experience attribute overload when facing multi-attribute choices.
Presenter: Adrian Hillenbrand, Max Planck Institute for Research on Collective Goods Title: Strategic Rational Inattention? An Experiment on Product Search with Hidden Costs Abstract: Online platforms allow consumers to search for their preferred products out of a plethora of available offers. Consumers profit from this development by saving time and optimally, by receiving better offers. On the other side, consumer search behavior is more easily trackable than ever before. This allows for behavior-based price discrimination of consumers. Thus, consumers face a trade-off: Search more intensively and receive better products at potentially higher prices or restrict their search behavior - be rational inattentive - and receive a worse fit but maybe a better deal. We provide a theoretical and experimental investigation of this strategic environment.
Presenter: Jianjun Tang, Queen's University Belfast Title: A comparison of time preference functional forms: Evidence from a field experiment Abstract: A comprehensive understanding of individual-level discounting behaviours is crucial because of its implications for designing interventions with financial incentives for behavioural change. This paper estimates mixture models to determine probabilistically the discounting functional form which has the best fit of discounting behaviours, on the basis of a series of incentive-compatible time and risk preference field experiments conducted among 176 civil servants in Belfast, Northern Ireland. Time preference was estimated while controlling for risk preference, background consumption, and probability weighting. The results suggested that future monetary incentives were discounted hyperbolically rather than exponentially and that a general-hyperbolic form had the best fit among a series of alternative functional forms. Our results also show that the failure to adopt an appropriate type of functional form lead to significantly different discount rates and misleading associations between time preference and real-world behaviours such as smoking.
Presenter: Sebastian Kruegel, Justus-Liebig-University Giessen Title: Pleasing or Fighting Future Tastes? Projection Bias versus Conflict of Selves Abstract: Many real life choices concern consumption in future periods. Previous studies apparently demonstrate that people systematically mispredict future tastes in such situations. This evidence, however, is also consistent with the idea that people understand, but do not approve of their future tastes. To disentangle both approaches, we conducted a framed field experiment with commitment option. In our experiment, commitment was not a device against weak will. It was a judgment which a planner imposed on another planner. The results suggest that people do not always aim to please future tastes. People may sometimes experience a conflict between two far-sighted selves.
Presenter: Irene Comeig, Universidad de Valencia Title: Do women self-select as good borrowers? Abstract: Banks often cannot observe ex ante failure probabilities of potential borrowers. To deal with this informational asymmetry, banks offer incentive compatible contracts (with collateral) to induce borrowers to disclose their private information. Typically, theoretical models show that private information is fully disclosed in equilibrium, since high risk and low risk borrowers self select by choosing different contracts. Low risk borrowers choose higher collateral at a lower interest rate, while high risk borrowers select contracts without collateral at a higher rate. The key is that the cost of investing in collateral is lower for low risk borrowers as they have a lower probability of failure . However, this self selection is achieved when potential borrowers are identical in every respect other than failure probability. Smart (2000) shows that the addition of the customer's degree of risk aversion can change the nature of equilibrium, and different risk classes may be pooled at a single contract in equilibrium . Specifically, risk averse low-risk borrowers may not be willing to accept higher collateral to self-select. The reason is that the collateral choice is also closely connected to the degree of risk aversion (See Cohen and Einav (2007) and Barseghyan et al. (2011)). The higher the risk aversion, the lower the willingness to accept higher collateral to self select. Low collateral exposes individuals to a lower risk by paying a higher contract price (interest rate). Women are generally found to be more risk averse than men in financial decision making . If this is the case, women might not accept higher collateral to self select. Thus, women's contract choices, even being rational decisions for low risk borrowers, will classify themselves as high risk borrowers. Whether men and women systematically differ in their contract choices in the self-selection mechanism with collateral is an important economic question. If women are particularly averse to financial risk, they may be classified as high-risk borrowers thus not receiving the loan designed for the good borrowers, or even suffering rejections. From the bank's point of view, the women reluctance to accept high collateral generates an adverse selection problem. Particularly risk averse individuals (i.e. women classified here as high risk borrowers) might be also the best borrowers for the bank. In this paper, a laboratory experiment on financial decision making in three different European countries is designed to study systematic gender differences in self selection. The primary interest is in the extent to which the women's patterns of behavior towards risk carry over into the self selection mechanism with collateral. Results show that incentive compatible contracts with collateral fail to disclose women private information, while they disclose men private information. Thus, low risk women consumers do not self select as "theoretical" good borrowers. Besides this contribution, results show that gender differences arise when subjects face downside risk, as defined in Comeig et al. (2015), low failure probabilities (90% success probability).
Presenter: Axel Sonntag, University of Vienna Title: Focality is Intuitive - Experimental Evidence on Time Pressure and Response Times from Coordination Games Abstract: We experimentally examine the effects of time pressure on the likelihood that two players coordinate on a label salient focal point in a coordination game. We consider both payoff-symmetric and payoff-asymmetric coordination games. In symmetric games almost everyone coordinates on the focal point, regardless of how much time they have to decide. In asymmetric games, in contrast, the less time people have the more able they are at coordinating on a focal action, and the higher earnings consequently are. Furthermore, the subjects who prefer coordination on the focal outcome are slower in making their decision than those subjects who prefer coordination on the non-focal action.
Presenter: Odile Poulsen, University of East Anglia Title: Coordination and Focality under Gain-Loss framing: Experimental Evidence Abstract: Are people better at coordinating on a focal point when they coordinate on a division of losses than gains? We keep the net economic payoffs the same and independently vary the gain/loss frame, as well as payoff structure (symmetric versus asymmetric), and stake size (low vs high). We observe that while there are no effects of a gain-loss frame when payoffs are asymmetric (conflict of interest), coordination on the focal point is significantly lower under loss framing with symmetric payoffs.
Presenter: Anita Gantner, University of Innsbruck Title: Learning to Coordinate: Individual Traits, Beliefs and Choices in the Minimum Effort Game Abstract: The minimum-effort game is economists' workhorse model for the study of coordination problems. If the game is played repeatedly, agents can learn about others' choices of effort levels and behavior converges over time. However, existing learning models fail to explain some of the key results in such experiments, such as fast adjustments to payoff changes. We want to understand why some groups converge to the high effort level while others are less successful under the same treatment conditions. This could be due to different initial beliefs or different updating rules or both. In a lab experiment we systematically elicit subjects' beliefs about others' minimum effort, and we further elicit information about subjects' cognitive skills, their trust and risk attitude. We control for the amount of information about others' past choices in order to see how information affects beliefs and choices. We find that higher cognitive skills and trust have significant positive effects on initial beliefs, which also translates into higher effort levels. Risk aversion does not play a major role overall. We observe that contributions systematically exceed beliefs, and we study how the elicited personal traits affect this bias to choose effort levels that are more "optimistic" than beliefs.
Presenter: Ozan Isler, University of Nottingham Title: Is intuition really cooperative? Abstract: Rand, Greene & Novak (2012) found cooperation to increase with time-pressure and hence argued that intuition is cooperative. However, two unresolved issues with evidence accumulated in the past five years hinder verification of this finding. First, compliance with time-limits remains low, and evidence shows intuitive cooperation only when noncompliant participants are excluded: Using pecuniary incentives, we achieve consistently high levels of compliance and provide congruent tests of intuitive cooperation. Second, existing studies show high social dilemma misunderstanding: We compare understanding of the original instructions to two pedagogical variations, and show that the briefness of the Rand et al instructions does not impede understanding; in addition, we assign time-limits on understanding questions and detect no asymmetries of understanding. Our study resolves the issue of noncompliance, shows that misunderstanding does not confound tests of intuitive cooperation, and provides the first independent experimental evidence of intuitive cooperation in a social dilemma using time-pressure.
Presenter: Nina Serdarevic, University of Bergen Title: It Pays to be Nice: The Benefits of Cooperating in the Market Abstract: This paper provides evidence for an apparent paradox - selfish individuals may sometimes do worse in monetary terms than those less concerned about monetary gain. This occurs when there is a market for exchange partners and cooperators may avoid free riders. Using a two-step experimental design, we first measure subjects' types and then assign them to a repeated Prisoner's Dilemma with either random matching or partner choice. In each round, the individual who fails to attain a partner is excluded from the group. This design allows us to study treatment effects on the earnings of different types. Overall results from two experiments indicate that the partner choice treatment increases overall earnings and excludes free riders from the game.
Presenter: Abhijeet Chandra, Indian Institute of Technology Kharagpur Title: Asset allocation under dynamic constraints: role of information and wealth inequality Abstract: Resource allocation is an integral part of the lives of every individual living in today's modern society where capitalism has made available a plethora of choices ranging from choosing your career to choosing your meal. Irrational decision making leads to a cascading effect not on just the well-being of the individuals but also on the society as a whole. Intuitive causality tells us that in decisions made under the absence of information lead to poor results. Today information asymmetry is playing a major rule into the already burgeoning inequality between the people of highest and lowest strata. It might be possible to reduce this inequality by addressing the problem of information asymmetry. We attempt to juxtapose the decisions made in presence and absence of certain information by experiments and try to decipher till what extent the lack of information leads to a poor decision-making and what remedies can be thought of to address the problem. We design a set of experiments carried out with two groups of university students, where one group has access to certain decision inputs while the other group is constrained with limited information. In the dynamic resource allocation game, we observe that the participants with more information were able to take decisions relating to allocation of financial resources that led to better outcomes. This was reflected from the amount of investments made by the group as well as the job choices chosen. These results point towards a dire need for information dissemination especially when it comes to the role of informed resource allocation in reducing inequality in a rapidly growing country like India. When individuals are aware of the opportunities at hand as well as the risks and rewards attributed to the same they are able to make choices that help them lift out of poverty. These results can be particularly useful in conceiving and designing government programmes that are targeted towards the poor. Despite the presence of numerous schemes, the lack of adequate information leads to poor outcomes.
Presenter: Julien Senn, University of Zurich Title: Corruption and cooperation Abstract: Corruption is a widespread phenomenon plaguing entire societies around the globe. While the existing empirical evidence suggests that corruption has a negative effect on aggregate variables such as GDP and growth, causal evidence on the effects of corruption is still lacking. In this paper, we assess whether and how corruption affects cooperation in a public good game experiment in which the punishment authority is centralized. We show that, compared to a control condition in which bribery is impossible, overall contributions to the public good are reduced by 30% when participants have the possibility to bribe the punishment authority. We provide evidence that two channels simultaneously lead to lower levels of cooperation. First, while low contributors are effectively disciplined by large punishment when bribes are not allowed, the punishment authority assigns less severe punishments to low contributors when bribery is possible. Second, the mere possibility of bribery discourages initially high contributors, who gradually decrease their contributions down to the level of initially low contributors.
Presenter: Julien Benistant, CNRS Title: Minimal group identity and misreporting in competition settings Abstract: Using a real-effort laboratory experiment we investigate whether minimal group identity affects the willingness to dishonestly misreport in a competitive environment. We also vary whether individuals have to report on their own output or on the output of their competitor. To assess the effect of group identity on these two behaviors we compare situation in which pairs of competitors share or not the same group identity. We show that subjects misreport as frequently by inflating their output as by decreasing their opponent's one. Furthermore our results indicate no effect of identity sharing within the pair on the decision to misreport or not.
Presenter: Caroline Stein, University of Cologne Title: The effect of autonomy on effort provision in moral dilemmas Abstract: Compliance standards are defined by organizations to ensure fair and legitimate conditions for interactions within and between organizations. Recent corporate scandals at Volkswagen or at Deutsche Bank, for instance, clearly demonstrate the risk for companies to suffer massive reputation losses if employees circumvent these standards. To better ensure compliance in a company, there is a strong demand for better control mechanisms that withdraw decision power from the employees. However, especially in morally loaded situations the withdrawal of decision power might affect the employees' motivation to exert effort. We investigate a situation of a team leader and her worker, where the worker has to do a real effort task and his performance benefits both. Before performing the real effort task it is possible to overstate the designated piece-rate for this task and thereby circumvent predetermined standards to benefit oneself and the team partner. We varied who of the two team-members - team leader or the worker, who finally executes the task - has the decision right whether or not to overstate the designated piece-rate. We find that workers that could decide themselves to state the piece-rate truthfully performed about 20% better than agents of the same moral types who were forced to comply with the truthful behavior by their team leaders. We find that the effect of autonomy on effort for moral type workers is as large as the effect of a 50% increase of monetary incentives.
Presenter: Jana Rollmann, Karlsruhe Institute of Technology Title: Voter Turnout under the Mean vs. the Median Rule Abstract: In a field experiment in the context of a budget allocation problem, we test whether voter turnout varies with the voting rule. While the mean rule is highly manipulable for a small number of voters, the impact of every vote decreases to zero if the group size becomes large (see Ehlers et al. 2004). By contrast, the median rule satisfies strategy-proofness. We analyze the maximal impact on the social outcome for different preference distributions under the mean and the median rule. For some preference distributions, the expected impact on the social outcome is larger under the median rule. This raises the question on the consequences for voter turnout under both voting rules. Given participation of all voters, only the median voter determines the outcome. Therefore, for a group size of 17 potential voters, we expect a higher voter turnout using the mean rule.
Presenter: Yoshio Kamijo, Kochi University of Technology Title: How does proxy vote on behalf of future generation affect the voting behavior of present generation? A laboratory experiment Abstract: This paper reports the experimental evidence of Demeny voting, wherein some people are given additional votes as proxy for the future generation. We hypothesize that the new voting scheme affects the voting behavior of one-ballot voters (i.e., the voters who don't obtain an additional vote under this scheme) because their probability to become a pivot decreases and other proxy voters are expected to behave for the future generation. We compare the three types of one-ballot voter's decision; (i) the one-ballot voter's decision in an ordinary voting where other voters are also one-ballot voter, (ii) one in Demeny voting where other voters have two ballots and the second vote is instructed to use for the future generation, and (iii) one in an asymmetric voting (called non-framed DV) where other voters have two ballots but there is no explanation about the second vote. We found that while the one-ballot voters negatively respond to the new situation in the non-framed DV, such negative reaction is mitigated if the second vote is explained as the vote on behalf of future. Thus, our result indicates that how the second vote is explained and justified is important to the reaction of the voters who obtain no-extra voting power in the new system.
Presenter: Alexander K. Wagner, University of Vienna Title: Bad Memory, Bad Policy? Abstract: We study a dynamic political-agency model in which politicians differ in ability and hold career-concerns. Bounded memory of voters changes the effectiveness of selection and accountability of policy makers, and, in turn, the strategic incentives of politicians to signal their ability through policy choices. Experimental evidence confirms the fundamental predictions of the agency model for voters as well as for politicians. Results highlight the crucial role of feedback for voters with bounded memory, as feedback allows to recover forgotten information required for optimal voting decisions. We also discuss implications for the design of institutions (e.g. media outlets) to promote rational evaluation of policy makers and to maximize social welfare.
Presenter: Pol Campos-Mercade, Lund University Title: Helping Behavior and Group Size Abstract: Will a person in need of help be more likely to be helped when there are one or two potential helpers? Dozens of experiments have led social psychologists to conclude that the answer to this question depends entirely on the situation. This paper uses a simple game theory model to predict in what situations a victim (person in need of help) will be more likely to be helped in a group of one than in a group of two bystanders (potential helpers), and in what situations the opposite is true. We show that, in situations where most bystanders are willing to help, victims are better off when only one bystander can help them. The intuition behind this result is that, despite most bystanders helping when they are the only potential helpers, bystanders have to overcome a coordination problem on who will help when they are two. However, in situations where few bystanders are willing to help, victims are better off when two bystanders can help. This is because the probability that there is at least one bystander that is willing to help is higher in larger groups. We test these predictions by running a small experiment in which groups of one or two bystanders have the possibility to pay a cost to help a victim. If at least one bystander pays the cost, the victim receives a given amount of money. When the victim is placed in a group where all bystanders have previously shown to be prosocial, she is better off when only one bystander can help her than when two can. When the victim is placed in a group where few bystanders have previously been prosocial, the victim is better off when two bystanders can help her than when only one can. This paper suggests a simple and concrete rule to predict whether people will be more likely to be helped when there are one or two potential helpers. This rule does not only help explaining some of the previous experimental results in social psychology, but also can be useful for managers, educators and policy-makers whose objectives are to increase helping behaviors within an organization.
Presenter: Marcela Ibanez Diaz, University of Göttingen Title: Formal insurance, risk sharing, and the dynamics of other-regarding preferences Abstract: In this paper we investigate if formal insurance crowds-out transfers and consider the ex-post effect that this has on other regarding preferences. We test this question using a framed field experiment with rural households in Mexico. In a solidarity game we randomly allocated a fair insurance to two of the three members of the network and find that in case of a shock, the value of transfers to non-insured members is significantly higher than in a control treatment when insurance is not available. This result is observed both when shocks are idiosyncratic and affect only one member of the network at the time or when shocks are collective and two members of the network are affected. This unexpected crowding-in effect leads to an increase in trust among non-insured participants, but a decrease in trust on insured participants. This finding suggests that other regarding preferences evolve as a sense of gratitude for the transfers received and not as a warm-glow for the transfers provided.
Presenter: Ro'i Zultan, Ben-Gurion University of the Negev Title: The Welfare Implications of Social Interactions Abstract: We extend the study of social preferences—preferences over social outcomes—to study preferences for social interactions. While models of social preferences are successful in explaining behavior in social situations, the welfare implications of preferences for social interactions that do not affect behavior remain understudied. We compare different situations that are identical with regard to people’s choices and outcomes, and test whether participating in positive social interactions increases welfare. We consider three dimensions of the social interaction: warm glow (helping others), social warmth (being helped), and positive reciprocity (mutual help). We measure welfare by eliciting the willingness to accept (WTA) for forgoing the social interaction. We find that people indeed have clear preferences over the social content the same outcome reached through the same actions. People attach a higher value to situations involving helping. We do not find evidence that mutual helping increases welfare.
Presenter: Mengling Li, Xiamen University Title: Organ Donation Decisions under Different Allocation Rules Abstract: A major source of transplanted organs around the world is from deceased donors. A well-designed organ allocation rule helps to moderate the scarcity of organs by improving efficiencies in allocation as well as by providing stronger incentives for donation. We study the incentives to register as a deceased organ donor under different organ allocation rules in both theory and laboratory. The organ waiting lists may prioritize registered organ donors and/or agents with high value in organ allocation. We consistently find that the dual-incentive allocation rules taking advantage of both donor status and value in organ allocation have negative impacts on both the donor registration rates and the aggregate welfare as compared to the simple donor priority rule. The dual-incentive rules distort the incentives of low value agents the most especially when the aggregate organ supply-demand ratio is low.
Presenter: Jordi Brandts, Instituto de Analisis Economico (CSIC) and Barcelona GSE Title: Supply Function Competition, Private Information and Market Power: A Laboratory Study Abstract: In the context of supply function competition with private information, we test in the laboratory whether - as predicted in Bayesian equilibrium - costs that are positively correlated lead to steeper supply functions and less competitive outcomes than do uncorrelated costs. We find that the majority of subjects bid in accordance with the equilibrium prediction when the environment is simple (uncorrelated costs treatment) but fail to do so in a more complex environment (positively correlated costs treatment). Although we find no statistically significant differences between treatments in average behaviour and outcomes, there are significant differences in the distribution of supply functions. Our results are consistent with the presence of sophisticated agents that on average best respond to a large proportion of subjects who ignore the correlation among costs. Even though we do not find evidence of greater market power on average, experimental welfare losses are higher than the equilibrium prediction owing to a substantial degree of productive inefficiency.
Presenter: Lisa Bruttel, University of Potsdam Title: Buyer power in large buyer groups? Abstract: This paper studies the exertion of market power in large buyer groups confronting an incumbent monopolist and a potential market entrant in a repeated trade situation. In the experiment, buyer power can either occur as demand withholding when only the incumbent is present in the market, or it can take the form of buying at higher prices from the entrant in order to foster future re-entry. Comparing markets with groups of two and eight buyers, we find that both these forms of buyer behavior are prevalent and occur less often when the number of buyers is large. However, a control treatment shows that such seemingly strategic behavior is better explained by inequality aversion of the buyers towards the two different sellers than by the strategic exertion of buyer power.
Presenter: Yu-Chin (Annie) Hsiao, University of Canterbury, Macquarie University Title: The effect of time cost on search behavior Abstract: We experimentally investigate the effect of time search cost on behavior in a decision task framed as selling houses. Sequential search in such a scenario is often costly and time-consuming. Previous literature has shown that if the search cost is monetary, people search less, but despite the unavoidable nature of time search cost, its effect on behavior has previously been unexplored. In our experiment, we implement two treatments: no time delay and a five seconds delay before a new price offer is available. The results show that, if each additional search involves time delay, people do indeed search less.
Presenter: Eli Spiegelman, BSB Université de Bourgogne Franche Comté Title: Too much for me! Limited awareness and preference reversals Abstract: We present a model of attention-based awareness. States are described by attributes, some subset of which may be forgotten at the moment of choice. A fixed attention budget is allocated to the attributes to reduce the probability of this forgetting. The model predicts (1) people focus on attributes that essentially agree with their vision of the world; (2) people focus on attributes that are unfamiliar; (3) those with a larger attention budget will make fewer mistakes. These predictions are applied to a classical preference reversal experiment. The data from the experiment include eye-tracking information that measures attention directly, allowing us to test the model's predictions. Results support the model: (1) people who choose a bet with a relatively high prize (probability) focus relatively more on that attribute; (2) overall, people focus more on probabilities than on prizes; (3) those who make reversals pay significantly less attention to the problem than those who are consistent.
Presenter: Cary Deck, University of Alabama Title: The Effects of Different Cognitive Manipulations on Decision Making Abstract: Cognitive load is known to impact economic decision making. This paper experimentally tests how that impact differs based on four commonly used techniques for manipulating cognitive capacity: a number memorization task, a visual pattern task, an auditory recall task, and time pressure. In a within-subject design, subjects complete a series of risk taking decisions, allocation decisions, pattern recognition logic problems, and math problems under each load manipulation. The results indicate that number memorization and auditory recall have comparable effects: behavior in allocation tasks are similar, both lead to poorer performance on math and logic problems as well as more risk aversion relative to a baseline with no load. Time pressure tends to impact decision making, but does so by increasing the frequency of errors. The visual pattern technique yields behavior broadly similar to the baseline setting. These results suggest that number memorization and auditory recall are the most reliable techniques for inducing cognitive load among those considered.
Presenter: Martin Kocher, Ludwig-Maximilian-University Munich Title: Birthright Citizenship and Discrimination in a Migration Society: Combining a Natural Experiment with a Large-Scale Trust Experiment in Schools Abstract: A fundamental aspect of migrant integration pertains inter-pers onal interactions between natives and immigrants. Ingrained in such interactions are issues of group identity which may give rise to phenomena such as favoritism and discrimination. We have (i) run an artefactual field experiment based on the trust game with a large, representative sample of German adolescents; (ii) allowed participants to condition their strategies on the migration background of their opponents; and (iii) matched the experimental data with individual background information from an extensive, self-conducted survey. As our first main result, we document a pattern of discrimination frequently neglected in public discourse: children with migrational backgrounds strongly discriminate in their trust decisions against their native peers who, in turn, themselves are much less inclined to discriminate against immigrants. On inspection, this discriminatory behavior turns out to be statistically unjustified, not driven by wrong stereotypes, and it involves a sacrifice of money, which points to a preference-based explanation. For our second main result, we connect our artefactual field experiment with a natural policy experiment which saw the introduction of a widely debated integration policy: birthright citizenship, which automatically grants children born to foreign parents the nationality of the host country. We find that the policy substantially reduced the degree of discrimination among male, but not among female, immigrants. This effect is accompanied by an improved educational integration of immigrant males, but not by a greater sense of affiliation with the host nation.
Presenter: J Michelle Brock, European Bank for Reconstruction and Development Title: Gender Bias in Bank Lending: Experimental Evidence from Turkey Abstract: Evidence from a broad range of countries shows that female borrowers tend to receive worse loan terms than male borrowers. This may simply reflect that female-owned enterprises are typically smaller and riskier. Legislation in many countries restricts women's access to collateralizable household assets, such as real estate, and this may make female borrowers more risky as well. However, disadvantageous loan terms for female borrowers may also result from social or cultural norms that lead to an unjustified bias among loan officers against female loan applicants. We use a lab-in-the-field experiment to assess the extent of bias against women among staff of a major Turkish bank. Participants reviewed a series of credit applications in the lab, where each application was viewed by some participants as if it came from a female-owned firm and by others as if it came from a male-owned firm. The applications were real applications that the bank had processed in the recent past. Subjects earned rewards for each lending decision they made, depending on whether the application was declined, performing or non-performing in real life. We can thus directly compare how a business owner is treated by bank staff when gender of the business owner is randomized. The lab is especially useful for this kind of research because even with very high-quality administrative data it is difficult to precisely identify gender bias in lending. This is mainly because female and male entrepreneurs may operate in different sectors or at a different operational scale, such that the overlap in distributions may be relatively small. There are also likely to be unobservables that cannot be controlled for in a regression but that matter for the credit allocation decision. Consistent with some earlier non-experimental literature, we find that while rejection rates are similar for male and female loan applicants, women face different collateral requirements. We also find evidence of differences in the rejection reasons. These results vary depending on the job title of the participant, highlighting the importance of organizational hierarchies in the expression of gender bias in an organization. Finally, we show the extent to which any bias is due to implicit beliefs about the value of women as business owners, using data from an implicit association test (IAT), and due to participant risk preferences.
Presenter: Redzo Mujcic, WU Vienna Title: The Colour of a Free Ride Abstract: We present a natural field experiment on discriminatory gifts in a unique public setting where formal rules and social norms oppose the decision maker from granting any favours to others. Testers from distinct racial and ethnic backgrounds are assigned to board a public bus without any money for purchasing a fare, leaving the bus driver to voluntarily decide whether or not to provide a service free of monetary charge. Based on 1,552 social interactions, we find strong evidence of favouritism toward light-skinned customer groups: white testers are 45 percentage points more likely to be favoured than black testers. There is no evidence of same-race preferences. Increased signals of socioeconomic status and selflessness improve minority testers' outcomes, while busier periods lead to more bias. A complementary survey of subjects reveals contrasting findings to the field, indicating the presence of a social desirability bias.
Presenter: Dorothea Kuebler, WZB Social Science Center Berlin Title: Self-confidence and unraveling in matching markets Abstract: We document experimentally how biased self-assessments affect the outcome of matching markets. In the experiments, we exogenously manipulate the self-confidence of participants regarding their relative performance by employing hard and easy real-effort tasks. We give participants the option to accept early offers when information about their performance has not been revealed, or to wait for the assortative matching based on their actual relative performance. Early offers are accepted more often when the task is hard than when it is easy. We show that the treatment effect works through a shift in beliefs, i.e., underconfident agents are more likely to accept early offers than overconfident agents. The experiment identifies a behavioral determinant of unraveling, namely biased self-assessments, which can lead to penalties for underconfident individuals as well as efficiency losses.
Presenter: Levent Yilmaz, University of Innsbruck Title: Multiple applications and probability of being invited Abstract: During successful careers, many people apply for positions in different functions (e.g., marketing and finance) at the same organization. We study whether sending multiple applications to different functional areas increases applicants' chance to be invited to a job interview. Human capital theory would predict an increase in the chance, while the categorical imperative would predict a decrease in the chance. Our results from data of a large US-American multinational company support for both. We show that sending multiple applications to different functions is penalized more than sending multiple applications to the same functions is rewarded.
Presenter: Johannes Hoelzemann, University of New South Wales Title: Organizational Design and Coordination Failure in the Network Minimum Game Abstract: We study, both theoretically and experimentally, organizational design and its role in coordination failure. To do so, we introduce the network minimum game, a version of the minimum-eff ort game where dependencies between players are captured by a directed network. (The standard minimum game corresponds to the special case of a complete network.) We show theoretically that acyclic networks are most conducive to successful coordination. These findings are borne out in our laboratory experiments: acyclic networks foster more coordination than their cyclic counterparts. Further, acyclic networks facilitate resilient coordination: teams that start with low actions successfully coordinate after repeated play in acyclic networks, but not in cyclic networks. Our fi ndings provide a novel perspective on the near-ubiquity of hierarchical (i.e. acyclic) structures in organizations.
Presenter: Corina Haita-Falah, University of Kassel Title: Sunk Contributions and Beliefs in a Dynamic Threshold Public Good Game Abstract: Duffy et al. [Duffy, John; Ochs, Jack; Vesterlund, Lise, "Giving little by little: Dynamic voluntary contribution games" Journal of Public Economics, 91, 2007, 1708, 1730] find that contributing to a public good over multiple contributions stages (dynamic setting) increases the overall contributions compared to one-shot contributions (static setting). Moreover, they find that dynamic contributions without feedback about the contributions of the other group members between stages yield the same contributions as with feedback. Most importantly, they also yield higher contributions than in the static setting. Because the dynamic-contribution setting without feedback is strategically equivalent to the static setting, this finding is puzzling. In this paper, we test whether the sunk character of the stage-by-stage contributions is what explains the higher contributions in the dynamic setting without feedback relative to the statics setting. For this, we have symmetric players contributing over two stages to a threshold public good. Between the first and the second stage, players learn the first-stage individual and group contributions. The treatment differences consist of the options available in the second stage, i.e. increase-only contributions (the NoTake option) or both increase and decrease contributions (the Take option). Thus, when the second stage contributions are decided, the first stage contributions are sunk in the NoTake option, while they are not sunk in the Take option. Moreover, the game with the Take option is theoretically equivalent with the static contributions setting. In the baseline conditions, subjects know their second stage option before they make their first stage contributions. In the treatment conditions, before the first stage contributions they are informed that with 50% probability either the NoTake or the Take option is available in the second stage. They learn the available option after the first stage and before the second stage. We find that the beliefs about the contributions of the other group members have a significantly positive effect over individual contributions. At the group level, the total contributions and the completion rates are lower in the treatment conditions than in the baseline conditions. However, both within the baseline conditions and within the treatment conditions, the share of the groups that reach or are close to the completion threshold is larger with the NoTake option than with the Take option. Moreover, based on linear regressions we find that the differences in the second stage contributions between the NoTake and the Take options are increasing in the first stage contributions, both in the baseline and the treatment condition. These differences are larger in the treatment condition than in the baseline condition, with the NoTake-option contributions being larger than the Take-option contributions for large values of the first-stage contributions. In fact, the NoTake-option contributions in the second stage of the treatment condition are increasing in the first stage contributions, while in the other three conditions the second-stage contributions are decreasing in the first-stage contributions. These finding indicate that sunk contributions indeed play a role in the superior performance of the dynamic setting over the static one.
Presenter: Karen Hauge, Ragnar Frisch Centre for Economic Research Title: The Good, the Bad, and the Conditional: Sorting and Dynamics in a Public Good Game with Endogenous Group Formation Abstract: Previous research has shown that charitable commitments can serve as a screening device in public good games (Brekke et al., 2011), enabling cooperative subjects to endogenously self-select into groups which manage to sustain cooperation over time. We present a public good game experiment in which subjects choose between two group types: in blue groups, subjects receive a fixed extra payoff; in red groups, this extra payoff is donated, instead, to the Red Cross. We use the strategy method to elicit the conditional contribution preferences of the subjects, using this to classify subjects as unconditional altruists, free-riders, conditional cooperators or others. In the current paper we investigate further why red groups manage to sustain cooperation over time while blue groups do not. Is it simply because red groups start off at higher contribution levels? Could it be that the behavior of subjects who select into blue groups is more conditional on others' behavior? Our experiment starts with a one-shot standard public good game with no group choice and no feedback. The preliminary results show that subjects who later choose red groups contribute significantly higher amounts in this first one-shot game compared to subjects who later choose blue groups. We find (as in Fischbacher et al. (2001)) that a small group of subjects can be classified as unconditional altruists and a small group as free-riders, while the largest group are conditional cooperators. While the altruists overwhelmingly select into red groups, the free-riders select into blue groups. Although these groups are relatively small, their very systematic selection into different groups is important for the dynamics of play.
Presenter: Britta Butz, RWTH Aachen University Title: Donations as incentive for prosocial behaviour in public good games Abstract: In repeated public good games, cooperation typically declines over time. Evidence suggests that the implementation of charitable donations leads to more pro-social behaviour. In this paper, we investigate if donations can serve as incentives for a higher level of cooperation in public good games. For this purpose, we include a twenty-percent donation share which is dependent on participants' contributions. Donations are either financed internally (by group members' contributions) or externally (by the experimenter). We also explore possible spill-over effects on donation behaviour. We have observed that a donation which is externally provided raises contribution significantly, whereas an internally provided donation leads to a decrease in contribution.
Presenter: Simeon Schudy, Ludwig-Maximilian-University Munich Title: Incentivizing Complex Problem Solving in Teams - Evidence from a Field Experiment Abstract: We document the causal effect of simple bonus incentives on performance in a non-routine, cognitively demanding, interactive team task. These tasks are more and more important in the economy and at the same time understudied. We conduct a field experiment and show a causal positive effect of incentives on the completion probability and the overall completion time of the task. Using several experimental treatment variations we shed light on the importance of different bonus components. We study the framing of bonuses (as gains or losses) and investigate whether bonus incentives work due to i)the monetary reward or ii) the reference performance bonus incentives provide. We also investigate the robustness of the effect in an additional sample and study the reactions to bonus incentives by differently composed teams. Finally, we shed light on how bonus incentives affect teams' willingness to explore in the non-routine task.
Presenter: Sebastian Schaube, University of Bonn Title: Self-Selected Peers and Their Effects on Individual Performance Abstract: Peers impact behavior in various dimensions of life. However, for many people not randomly assigned peers might matter, but rather they choose their peers carefully. In this paper, we present how self-selected peers affect individual performance in contrast to randomly assigned ones. For this purpose, we conduct a framed field experiment in secondary schools. Students participate in suicide runs twice. First, students run alone, then with a peer. Before the second run, we elicit preferences for peers both based on names and on performance. We experimentally vary the matching in the second run and form pairs either randomly or based on elicited preferences. We find that endogenously selected peers improve individual performance by an additional .16-.21 SD relative to randomly assigned peers. While the aggregate effect is similar for both treatments, selection on Performance tends to improve the performance of faster students - in contrast to selection on names which benefits slower students the most. These improvements do not only result from a changed composition of pairs, but rather suggest that contextual effects such as self-selection per se and changed social interactions in the treatment groups play a crucial role.
Presenter: Muruvvet Buyukboyaci, Middle East Technical University Title: Collaboration and Free-Riding in Team Contests Abstract: The organization of team contests can enhance productivity if teammates with complementary skills are able to allocate the team's tasks efficiently, but can also suffer from free-riding incentives. We report the results of a real effort experiment in which production requires the completion of two complementary tasks, at which workers have heterogeneous skills. We vary whether participants: compete individually; compete in teams where each member must complete each task; or compete in teams where the agents can divide tasks between them and potentially specialize in the task they do best. We report three main results. First, individuals who must work alone divide their work time in a way that is qualitatively consistent with the theoretical predictions, but allocate too little time to their weaker task. Second, there is no difference in productivity or free-riding behavior between individual contests and team contests where teammates cannot specialize. Finally, and most notably, when teammates can divide work tasks, they allocate more time to the tasks they are best at and experience a strong productivity gain. This is true even among teams that cannot communicate despite the potential for coordination failure or coordination on Pareto dominated equilibria but the effect is strongest when communication is available.
Presenter: David Hugh-Jones, University of East Anglia Title: Where do fairness preferences come from? An experiment on norm transmission in an adolescent social network Abstract: We ran an experiment on transmission of fairness norms in a UK school. Participants chose an allocation of money between two others, which could be meritocratic (based on task performance) or egalitarian; they then observed another person's allocation choice, before making a second allocation themselves. Participants' behaviour, and their stated attitudes about fairness, were influenced by the choice they observed. Our results show that children and adolescents can learn social norms from each other.
Presenter: Daniele Nosenzo, University of Nottingham Title: Norm fragility in the presence of image-motivated agents: the role of norm legitimacy Abstract: We study, theoretically and experimentally, the relation between norm fragility (i.e. the extent to which a norm is vulnerable to exploitation of moral wiggle room) and norm strength (i.e. the extent to which a norm is perceived as legitimate and is internalized by the agents). Using a simple model where compliance with norms of fair sharing is driven both by intrinsic and image motivation, we show that strengthening weak norms increases compliance and can reduce norm fragility. We corroborate these theoretical results in a dictator game experiment by showing that the extent to which subjects are willing to exploit moral wiggle room depends on the legitimacy of the norm.
Presenter: Alexandra Seidel, Otto-von-Guericke-University Magdeburg Title: Kick the ball back? Abstract: A well-known situation occurring in a football match is a player being injured and lying on the ground unable to go on. The respecting team then usually kicks the ball out of the game in order to interrupt the match. After the break, the opposing team receives the ball and according to the rules starts the next move. However, it is common to kick the ball back to the team with the injured footballer and the match continues. Following the standard economic model, subjects will always maximize their own utility but in this case, an advantage that came to one team by chance is actively not taken, or in other words the disadvantage the opposing team faces is actively not utilised. This act of fairness and sportsmanship is an unwritten rule in football and it works in most cases although it is not sure that both teams will face this situation with interchanged roles within the match. Since the situation the football teams are in is a very special one, we want to investigate whether this kind of fairness is an exclusive good of football matches or whether people could also decide that way under laboratory conditions. In our experiment, we examine whether this credit of trust will be granted if it is not guaranteed that the other one will even have the chance to reciprocate the favour (and doing it if possible). We are basically using a matching pennies design and vary the cost of being fair and the level of observability of the individual's actions between the treatments. First results show that the higher the cost of being nice, the lower the share of subjects playing fair. Furthermore, the amount of reciprocity observed in the lab seems to be remarkably lower than on the football field.
Presenter: Sibilla Di Guida, University of Southern Denmark Title: Taking Care of High-Need Patients in Capitation-Based Payment Schemes - An experimental investigation into the importance of market conditions Abstract: A capitation-based payment scheme is known to lead to underprovision of health care especially to patients in high need of care. This paper contributes to the literature by testing whether two alternative versions of the capitation-based payment scheme make physicians prioritise high-need patients: 1) ring-fencing part of the capitation payment to a fixed physician salary, and 2) making the per capita payment higher for patients expected to be of high-need of care than for patients expected to be of low-need of care. We test the schemes using a controlled, incentivised laboratory experiment where participants take the role of physicians and decide on the supply of health care services. A total of 55 prospective physicians (medical students) participated in three experimental sessions conducted in Denmark in 2015. The results show that high-need patients gain the most from a fixed physician salary when there is resource abundance, whereas there is no difference in gains between high-need and low-need patients when physicians are resource constrained. We also find that differentiation of payments make physicians take relatively better care of high-payment patients than low-payment patients. This result holds even though physicians' total payment is fixed. We conclude that capitation-based payment schemes can be designed to incentivise physicians to prioritise high-need patients. However, policymakers should be aware that the presence of resource constraints significantly impact how the different schemes affect allocation of care.
Presenter: Franziska Brendel, University of Duisburg-Essen Title: Scarce resources, allocation decisions, and prioritization in medical care - A laboratory experiment Abstract: Medical resources are limited in many instances. Their scarcity may for example stem from timely limitations physicians face every day, from limited infrastructure such as not enough intensive care unit beds, or simply from a lack of budgetary funds. The allocation of resources among patients is therefore a ubiquitous feature of health care systems around the world and the prioritization of medical services has become an important topic on many political agendas. In experimental health economics, previous studies mainly investigate the weight that physicians attach to patient benefit relative to their own profit. However, physicians do not only face trade-offs between their own profit and the medical care assigned to one single patient. In fact, they usually treat several patients and, therefore, also face tough decisions regarding the allocation of resources between patients. How do physicians allocate scarce resources among several patients? How do they prioritize? Finding an answer to these questions is important for designing health systems, especially when physicians have a strong degree of discretionary power with respect to the allocation rule they choose when treating several patients. We conduct a laboratory experiment with subjects in the role of physicians and patients. In particular, one physician faces two patients. She decides how much of a given budget is spent on medical services for each of the two patients, with patients differing in their need and urgency of treatment. With our experiment, we shed light on the questions, (1) how the size of a physician's budget (capitation fee) influences allocation decisions of physicians, (2) how patients' health characteristics (urgency, need) influence the allocation of scarce medical resources, and (3) how efficiency concerns and prioritization are integrated into decision making.
Presenter: Jeannette Brosig-Koch, University of Duisburg-Essen Title: Who benefits from physician competition? A laboratory experiment on the relevance of patients' mobility and state of health Abstract: We examine how competition between physicians affects the quality of medical care given that patients differ in both, their mobility, i.e. their ability to freely choose a physician, and their health characteristics. Previous research has shown that, without competition, physicians provide too many services with fee-for-service payment. While competition might reduce these distortions, only a fraction of patients switches providers at all and physicians usually interact with each other repeatedly over time. Both patterns might limit the positive effects of competition. In our experiment, we find that, on average, the quality of medical care provided with competition differs significantly from patient optimal treatment, but observed deviations are lower than that without competition. We further observe that the quality of medical care increases for mobile patients with repetition, while it decreases for immobile ones. Thus, in line with our theoretical model, competition reduces the distortive impact of fee-for-service payment, but only for mobile patients. Observed deviations from patient-optimal treatment are also sensible to patient health characteristics: deviations are particularly pronounced for immobile patients in need of a low number of medical services. Overall, our study provides some support for the view that competition may have positive effects on the quality of medical care. In case that some patients are not willing or are not able to freely choose a physician, competition seems to evoke inequalities regarding the medical treatment of the different patient groups, however.
Presenter: Jana Peliova, University of Economics in Bratislava Title: Portfolio allocation under the loss limitation Abstract: Retirement savings portfolio allocation has been under ongoing theoretical and empirical consideration in economic literature. Investment strategy of retirement savings represents a long-term decision with substantial effect on the retirement wealth. However, empirical findings conclude that individuals fail in such decisions. Therefore, policy makers in many countries try to avoid insufficient retirement income and offer a guaranteed retirement income. These policies introduce a substantial level of moral hazard into individuals' portfolio allocation decisions by the limitation of the potential loss. We use a laboratory experiment to investigate the reaction to presence of a 50 percent loss limitation on the net return to a fund of three risky assets. The experimental design follows the "low-frequency" design of the investment decision task pioneered by Gneezy and Potters (1997). Subjects allocate their portfolio between a fund of three risky assets and cash. We conduct four separate treatments. In the nolimit/limit (NL) treatment, subjects participate in three (investment) decisions without limited losses, followed by three decisions with limited losses. In the no-limit/nolimit (NN) treatment, subjects participate in six decisions without limited losses. In the limit/no-limit (LN) treatment, subjects participate in three decisions with limited losses, followed by three decisions without limited losses. Finally, in the limit/limit (LL) treatment, subjects participate in six decisions with limited losses. We found different, however counterintuitive investment decisions among subjects.
Presenter: Florian Schneider, University of Zurich Title: Collective Self-Deception and Markets Abstract: There is widespread evidence for motivated beliefs and self-delusion in individual judgment. People are overconfident (e.g. Svenson, 1981; Quattrone and Tversky, 1984), engage in wishful thinking (e.g. Irwin, 1953; Mayraz, 2013), and update their beliefs about their own ability in a manner inconsistent with Bayes' rule that attends more to positive feedback than to negative feedback (Eil & Rao, 2011; Maebius, et al., 2016). However, there is little work on whether beliefs are similarly biased when they relate to group outcomes. Understanding motivated beliefs in non-individual choice situations is particularly important in markets, as motivated beliefs are believed to play an important role for market outcomes, for example in bubble formation (Shiller, 2002; Cheng, Raina & Xiong, 2014), and can potentially explain the equity home bias puzzle (Strong & Xu, 2003) and the disposition effect (Chang, Solomon & Westerfield, 2016). This paper addresses two primary research questions. First, we use experiments to explore whether over optimism and biased information processing exist at the group level in a manner similar to how they occur for individuals. Second, we investigate how motivated beliefs are influenced by institutions that aggregate and shape information, like markets. Markets often mitigate individual biases (Camerer, Loewenstein & Weber, 1989; Camerer, 1987), but, there are also arguments that the collective judgment produced in markets can exacerbate biased beliefs. In our experiment, subjects first solve a task alone (individual treatment) or in a group (group treatment). The subjects in the individual treatment are paired with another subject that does the same task. Each subject receives points according to how close her solution is to the optimal solution. At the end of the experiment, the points of the paired subjects are compared, and some money from the subject that has fewer points is transferred to the subject with more points. Before subjects receive any feedback about their performance, we elicit their beliefs about the probability that they have more points than the paired subject. Then, subjects receive imperfect feedback about their own relative performance and are asked again to report their beliefs. A group treatment uses similar procedures, except that the task is performed in groups, and the relevant outcome for determining the monetary transfer and beliefs is whether a subject's group outperformed the other group. In a third treatment, the market treatment, subjects first participate in the group treatment, but after they receive feedback, they participate in an asset market. In the market, they trade assets with the other members of their group. Each asset pays a positive dividend if their group has more points, but no dividend if their group has fewer points. After they have participated in the market, they also report their belief about having more points. We find that groups and individuals are similarly overconfident and that they exhibit similar degrees of biased information processing. Moreover, overconfidence and information processing biases are similarly strong in the presence of a market institution. We conclude that self-deception at the individual level is robust to beliefs formed over group outcomes and to at least some kinds of collective information aggregation mechanisms. For market outcomes, we find that overconfidence and self deception exacerbate departures from fundamentals.
CANCELED:
Presenter: Natalia Shestakova, University of Vienna, Title: Role of natural disasters and information in the housing market
Presenter: Jin Di Zheng, University of Amsterdam Title: Eyes on You: Observer Identity in a Dictator Game Abstract: It is a well-established result in the literature that being observed by others already affects pro-social behavior. Less research, however, has shed light on the heterogeneity of such effect. Does it matter whether someone is observed by an in-group member, or by an out-group member? The answer is important for charitable fundraising strategy, for instance. In a laboratory experiment, we test the effect of the observer identity in a dictator game. We do so by creating in-group and out-group in the laboratory, using the minimum-group paradigm. Varying the group composition, we conduct a one-shot dictator game with a passive observer. There are four scenarios (treatments) among the dictator, the recipient and the observer: 1) three of them are in-group, 2) the observer is out-group, 3) the recipient is out-group, and 4) the dictator is out-group. The results show that dictators in scenario 4 give less than those in the other three scenarios, whereas there are insignificant differences in the dictators' giving in the other three scenarios. This illustrates that dictators give more to the recipient when at least either recipient or the observer is in-group. When the dictator is matched with an in-group recipient and an in-group observer, however, no additional dictator giving is observed.
Presenter: Tobias Regner, Friedrich-Schiller-University Jena Title: Actions and the self: I give, therefore I am? Abstract: Self-signaling models predict less selfish behavior in a probabilistic giving setting as individuals are predicted to invest in a pro-social identity. However, there is also substantial evidence that people tend to exploit situational excuses for selfish choices (for instance, uncertainty) and behave more selfishly. We contrast these two motivations experimentally in order to test which one is more prevalent in a reciprocal giving setting. Trustees' back transfer choices are elicited for five different transfer levels of the trustor. Moreover, we ask trustees to provide their back transfer schedule for different scenarios that vary the implementation probability of the back transfer. This design allows us to identify subjects who reciprocate and analyze how these reciprocators respond when self-image relevant factors are varied. Our results indicate that self-deception is prevalent when subjects make the back transfer choice. Twice as many subjects seem to exploit situational excuses than subjects who appear to invest in a pro-social identity.
Presenter: Danijela Vuletić, CERGE-EI Title: How Effective are Reminders and Frames in Incentivizing Blood Donations? Abstract: This paper studies the effects of reminders, and frames used to invoke higher levels of empathy and altruistic motives on the willingness to donate blood. We have conducted a randomized field experiment with 3236 blood donors from Bosnia and Herzegovina, in order to test how effective frames were when used in letter soliciting blood donations. Further, we tested the effectiveness of the letter itself which served as a specific reminder, making the need for blood more salient. Our baseline group did not receive any letter. Another seven groups received letters which differed in terms of goal framing; whether a specific victim was identified; and the gender of a victim. We found that a reminder of the need for blood in the form of a simple letter increases the probability of coming to donate blood by 63% relative to the baseline group, suggesting that reminder letters may serve as a cost-effective policy tool. At the same time, we found that the framing of the letter had relatively little effect when donors are allowed longer period to make their donation decision.
Presenter: Peter Lewisch, University of Vienna Title: From spontaneous cooperation to spontaneous punishment – Distinguishing the underlying motives driving spontaneous behavior in first and second order public goods Abstract: Recent findings indicate that at least some individuals use prosocial options by default in social dilemmas, known as ‘spontaneous cooperation’. In two studies, we test whether this effect generalizes to second order public goods in the form of punishment behavior in one-shot and iterated public good games, and investigate the underlying motivations. In line with spontaneous cooperation, punishment decreases with increasing decision time. Negative affect moderates this spontaneous punishment effect in one-shot public goods, in that punishment decisions are taken more quickly by persons who are more upset about the contribution behavior of their group members. Unlike spontaneous cooperation, spontaneous punishment is not driven by dispositional pro-sociality, but by above-average contributions. An overall analysis indicates a three-way interaction in that the spontaneous punishment effect is mainly valid for above-average, highly upset contributors. Our results highlight the uniqueness of spontaneous punishment, being in contrast to spontaneous cooperation independent of social preferences but similar to a revenge-orientated phenomenon.
Presenter: Maroš Servátka, Macquarie University Title: Nice to You, Even Nicer to Me: Does Self-Serving Generosity Diminish Reciprocal Behavior? Abstract: We propose a conjecture that self-serving but generous actions diminish the positively reciprocal response, compared to selfless generous actions. We embed our conjecture in Cox, Friedman & Sadiraja's (2008) model of Revealed Altruism. According to Revealed Altruism reciprocal responses are influenced by a "more generous than" (MGT) ordering. The MGT ordering is defined by two conditions. Condition A states that an action that increases one's opportunity set is MGT an action that decreases, does not change, or increases the opportunity set by less. Condition B states that the action cannot increase the "givers" opportunity set by more than the "recipients" opportunity set. We focus on Condition B, and classify actions that satisfy Condition B as selfless generous actions, and actions that violate Condition B as self-serving generous actions. We hypothesize that selfless generous actions are MGT self-serving generous actions, and that self-serving generous actions will result in a diminished reciprocal response. We test this conjecture using two novel experimental designs. We find evidence that subjects perceive self-serving generous actions as being less generous than selfless generous actions, but no empirical support for our conjecture on the diminished reciprocal response, suggesting a refinement for the MGT ordering that does not include Condition B.
Presenter: Wendelin Schnedler, University of Paderborn Title: Will you shout at your children when your boss cuts your wage? An experimental study on generalized reciprocity Abstract: Individuals return favors and treat those badly who treated them badly. If a supervisor reduces the wage, the employee may reciprocate by no longer putting in extra hours. For the concept of reciprocity, it only matters that the supervisor negatively affects the employee and vice versa. How the supervisor does that and how the employee pays her back is irrelevant. Indeed, actions may belong to very different domains. The wage cut, for example, is monetary and the reduced work hours non-monetary. Suppose the employee has no means to payback the supervisor. Then, the employee may channel his anger to an innocent third person. Like the notion of reciprocity itself, this generalized notion of reciprocity should cross domains. Although his wage was cut (monetary domain), the employee may, for example, shout at the children (non-monetary domain). This paper reports on an experiment to check whether such type of generalized reciprocity exists. We find that is does not. People who were badly treated do not pass their bad experiences on. This calls into question the legendary "chain of evil."
Presenter: Stefan Traub, Helmut-Schmidt-University, Hamburg Title: Power illusion in coalitional bargaining: An experimental analysis Abstract: One feature of legislative bargaining in the real world is that the distribution of seats or voting weights often does not accurately re flect real bargaining power. Game-theoretic predictions about payoffs and coalition formation are insensitive to such purely nominal differences. The converse idea that nominal differences might matter is referred to as power illusion's. We conduct an experimental test of the classical Baron-Ferejohn model with five-player groups. We compare treatments with differences in nominal power holding real power constant. We find strong evidence for the existence of power illusion in almost all aspects of bargaining behavior even if subjects got familiar with the game. This implies that attention needs to be paid to nominal power differentials in the design of weighted voting systems.
Presenter: Róbert Veszteg, Waseda University Title: (New) experimental results on unconstrained bargaining from Tokyo and Paris Abstract: We report results from a series of laboratory experiments on two-person unconstrained bargaining. Participants considered different situations in randomly formed pairs and had to decide in a fixed amount of time how to share a scarce resource. The main goal of our novel experimental design is to systematically study the empirical relevance of the axioms behind the bargaining solutions proposed by Nash (Nash, 1950) and Kalai and Smorodinsky (Kalai and Smorodinsky, 1975) and to test a number of alternative solution concepts. In line with the existing literature, we find evidence for violations of both the axiom of independence of irrelevant alternative and the axiom of scale covariance. The best-performing solution concept turns out to be "deal-me-out" (Binmore et al., 1989) followed by the "maximum" solution (Roth and Malouf, 1982). Although prescriptions by axiomatic bargaining theory are independent from culture, our experiment and the post-experimental survey reveal noticeable differences in individual attitudes and bargaining outcomes between the two metropoles (Tokyo and Paris) where we collected data.
Presenter: Matej Lorko, Macquarie University Title: The role of numerical anchors in performance time estimation process Abstract: In order to perform a cost-benefit analysis of a new project, it is necessary to estimate its completion time. It is naturally appealing to expect that the accuracy of project estimates increases with estimators' professional experience. However, contrary to the intuition, many companies keep estimating their projects over-optimistically. As a result, the projects are late and over budget. In this paper, I test one potential source of bias in the estimation process' numerical anchors. In everyday life, these anchors can appear, for example, in the form of initial wild guess, suggestion, customer expectation or, perhaps a tentative management deadline. In addition, I test whether the influence of anchor can persist over time and affect not only the immediate estimate, but also subsequent estimates of the same or similar task. This may happen especially in the absence of feedback regarding the actual duration of the first task. Even when there is no arbitrary anchor before the first estimate available, the first estimate itself can serve as an anchor for future estimates and cause a systematic bias. The persistence of the anchoring effect is experimentally tested by repeating the task estimation and execution process in multiple rounds. Thus, the anchor is directly pitted against subject's task experience.
Presenter: Vojtěch Zíka, Charles-University Prague Title: Overestimation and Competition: Evidence From The Lab Abstract: Overestimation is one's impression that his ability, performance or level of control is greater than it actually is. Together with overprecision and overplacement, this concept constitutes one of the most robust biases known in behavioral economics - overconfidence. This bias oftentimes prevents decision-makers from making "rational" choices as judged by the standard economic theory, however, from the point of view of the evolutionary rationality, the overconfidence might be advantageous as it encourages individuals to claim resources they could not otherwise win in a competition. If the overestimation, as a variety of overconfidence bias, is a kind of a signaling or bluffing strategy, then it should follow that as the competition gets more salient, the individual's ability to evaluate her own performance realistically declines. This article reports on laboratory experiments based on slider effort task that test this assumption.
Presenter: Baiba Renerte, University of Konstanz Title: Confidence spillover effects in risky group decisions Abstract: The focus of overconfidence studies has mostly been on individual decision making, even though important economic decisions are often made by groups rather than individuals. We address this gap and examine whether groups make less optimal investments in risky prospects when (more) overconfident individuals are involved in the decision. We examine two channels for this effect. We firstly use a novel laboratory experiment with objective investment situations as compared to ability-dependent investment situations (within-subjects comparison) to disentangle overconfident group members' perceived ability to "beat the odds". Secondly, we add a treatment with pre-decision communication (between-subjects comparison) to disentangle the "behavioral signature" of overconfident group members. We intend to show that group decisions tend to gravitate toward overconfident members' decisions and pinpoint behavioral reasons explaining this result.
Presenter: Stefan Grimm, Ludwig-Maximilian-University Munich Title: Blaming the Refugees? Experimental Evidence on Responsibility Attribution Abstract: Do people blame refugees for negative events? We propose a novel experimental paradigm to measure responsibility attribution. Participants experience an income shock which is with equal probability caused by either a random draw or another participant's performance in a real effort task. We measure responsibility attribution by eliciting beliefs about which mechanism was responsible. We find evidence for costly reverse discrimination: Participants attribute responsibility more favorably to refugees compared to behavior towards other local participants. Beliefs about performance in the real effort task and hence statistical discrimination cannot explain these findings. We relate our findings to individual political preferences and measured implicit associations towards Arabic names.
Presenter: Ahrash Dianat, University of Essex Title: Statistical Discrimination and Affirmative Action in the Lab Abstract: We investigate experimentally the conditions that foster statistical discrimination in labor markets, as well as the effectiveness of several policies designed to combat discriminatory practices. In our experimental labor markets, firms and workers interact repeatedly. Workers are associated with one of two colors and, in each period, have to decide whether or not to invest in costly training. In each period, firms observe only a worker's associated color and decide whether or not to hire the worker. We induce a pattern of statistical discrimination in which workers associated with one color do not invest in costly training, correctly foreseeing firms not hiring them. We then implement an array of affirmative action policies, in which firms are given subsidies for hiring workers of one color. While all such policies help equalize the playing field when they are in place, discrimination reappears once they are lifted. However, we do find that a longer duration of affirmative action mitigates the reversal to previous patterns of discrimination. Our results shed light on the efficacy of commonly used methods to combat discrimination in labor markets.
Presenter: Nikhil Masters, University of Manchester Title: Economic Insecurity, Racial Bias and Trust Abstract: The issue of economic insecurity features heavily in debates surrounding immigration and ethnic integration. According to realistic group conflict theory, perceived threats to a group's resources manifest in negative stereotyping and increased in-group identification. We investigate this experimentally by increasing the salience of economic insecurity and then measuring racial bias in the form of implicit attitudes and behaviour in a trust game. We find that although implicit attitudes are not malleable to increased insecurity threat, trust between ethnic groups is reduced.
Presenter: Felix Klimm, Ludwig-Maximilian-University Munich Title: Too Good to Be True? Cheating and Inequality Acceptance Abstract: When is inequality considered to be fair? Fong (2001) finds that people are more willing to accept inequality if they believe income distribution is due to factors within individual control. But what if these factors involve cheating? This paper experimentally examines the effect of cheating opportunities on inequality acceptance. Participants perform a real effort task and earn money according to their stated performance. Thereafter, a third party can redistribute earnings among two other participants. My main finding is that payoffs are equalized more often when cheating is possible, although cheating cannot be detected on an individual level. This suggests that if cheating opportunities exist, income inequalities are considered to be caused at least partly by cheating which in turn decreases acceptance of inequality.
Presenter: Shaun Grimshaw, University of Exeter Title: An experimental test of a simple behavioural model of lying aversion Abstract: A number of experiments show that while prone to lying, many do not do so in a way that maximises their financial payoff. This suggests that dishonest behaviour induces some cost in many individuals which is inconsistent with payoff maximisation. In this paper we present a behavioural model of lying aversion with two cost elements. The first is a function of the size the lie to reflect the act of telling a lie; the second is defined in terms of the deviation above a reference point that represents a concern for being seen to tell a non-credible lie. We test the predictions of our model in an experiment where subjects have to report their performance in a real-effort task. The results demonstrate that the reporting behaviour of subjects differs with different reference points reflecting the performance of others doing the same task in a manner consistent with our model.
Presenter: Fabio Galeotti, GATE CRNS Title: When Foul Play Seems Fair: Exploring the Link between Just Deserts and Honesty Abstract: The distributive justice norm of just deserts---i.e. the notion that one gets what one deserves---is an essential norm in a market society, and honesty is an important factor in economic and social exchange. We experimentally investigate the effect of violations of the distributive justice norm of just deserts on honesty in a setting where behaving dishonestly entails income redistribution. We find that the violation of the just deserts norm results in a greater propensity toward dishonesty. We then test a more general proposition that violations of just deserts induce dishonesty, even in cases where dishonesty does not have redistributive consequences. Our results confirm this proposition but only for cases in which the violation of just deserts also entails income inequality.
Presenter: Thomas de Haan, NHH Norwegian School of Economics Title: Fairness preferences in the face of limited information Abstract: In this project we investigate, theoretically and experimentally, the role of limited information about performance in a collective production task on redistributional preferences. In our study we focus both on distribution decisions by one of two matched workers, as well as distribution decisions made by a monetarily unaffected spectator. For both the setting with a stakeholder and a spectator redistribution decision, we will compare two situations. One, where decision makers have full information regarding what influence worker performance and what influence "random factor" luck had in determining the workers' contribution to the total income. Two, where decision makers have limited information regarding the worker's actual performance. Using the fairness preferences model from Cappelen et al. (2007, AER) as a basis, we predict that stakeholders who are ex ante pessimistic about other workers' performance, should become less prone to redistribute in the face of limited information if they hold a meritocratic fairness ideal. In contrast meritocratic spectators are predicted to become more egalitarian in their redistribution choices when facing limited information. We report the results of a laboratory experiment conducted in Bergen in 2016 and recently in May 2017.
Presenter: Ali Seyhun Saral, University of Trento Title: Conditional Cooperation, Beliefs, and Learning Abstract: In this study we experimentally investigate learning and behavioral dynamics on conditional cooperation. We use a minimal setting where we can categorize subjects into conditional types that are defined in the literature. Specifically, we use an extended version of prisoners' dilemma game in which both players have three options: no cooperation, moderate cooperation, and full cooperation. Each round subjects were matched to a random participant in the lab and they stated their strategies, both as the first-mover and as the second-mover. Moreover we elicited beliefs on the other players' strategies. Two major types we observe are conditional cooperators and selfish players, while a notable number of hump-shaped cooperators are present. The data exhibit a similar pattern to the well-documented decline of cooperation in public good game experiments and we show that players learn to play dominant strategy, thus the conditional cooperation tend to decrease over time.
Presenter: Astrid Gamba, University of Milan Title: Frustration and Anger in the Ultimatum Game: An Experiment Abstract: In social dilemmas choices may depend on belief-dependent preferences. We address this issue theoretically and experimentally in the context of the Ultimatum Minigame, assuming that the responders' choice of accepting or rejecting a greedy offer is affected by a combination of frustration, due to unfulfilled expectations, and inequity aversion. We derive our behavioral predictions by assuming that players perform two steps of elimination of non-best replies given plausible beliefs restrictions. Our experimental manipulation is twofold, affecting both the payoff structure and the method of play in a 2X2 design. With the purpose of increasing frustration due to the unfair offer, we increase the responder's payoff from the fair allocation and test whether rejections increase and so do the fair allocations. In addition, we implement both the direct response and the strategy method with the purpose of switching on (direct response method) and off (strategy method) the responder's experience of anger. We find that while rejections are more frequent when the game is played sequentially, on average the payoff increase does not have a significant effect on the rejection rate and on the frequency of fair offers. Yet, the higher the initial expected payoff, the more likely it is that a greedy offer is rejected and this effect is stronger in the direct response method. In addition, we find relevant gender differences. While for males the increase of the responder's fair allocation payoff decreases the acceptance rate and accordingly the frequency of greedy offers, in line with our predictions, for females the payoff increase has the opposite effect.
Presenter: Xiaocheng Hu, University of Southampton Title: Dynamic Incentive Effects of Team Formation: Experimental Evidence Abstract: Optimal team composition has been the focus of exhaustive analysis. Much of it has ignored possible dynamic effects: anticipating that team formation is based on prior performance may affect prior performance. We test this hypothesis in a lab experiment. Participants first work individually without monetary incentives and are then assigned to teams of two where compensation is based on team performance. Our results are consistent with theory: pairing the worst performing individuals with the best yields 20% lower first stage effort than random matching. Pairing the best with the best, however, yields 5% higher first stage effort than random matching.
Presenter: Francesco Falucchi, Luxembourg Institute of Socio-Economic Research (LISER) Title: Leadership and Group Formation in Team Production Abstract: Team production is known to suffer from the free-riding problem. There is evidence that leaders who can "lead by example" or decide on the distribution of team output can "encourage" higher cooperation from team members. We investigate another function of leaders - assembling teams of productive employees. In a lab experiment using a public goods game, we explore the ability of leaders to identify productive workers and assemble productive teams, while relegating unproductive workers to "internal exile" in teams that generate little revenue for the firm. We compare performance of "firms" with leaders to those with random assignment of workers. We also check if making the leader contributing to the public good of one of the teams has any effect. Our results show that leaders manage successfully to separate "workers" into productive and unproductive teams. The latter are frequently composed by a single "worker", and used as a threat to induce higher cooperation towards the public good of the former. Finally, an even higher level of cooperation is achieved when the manager actively contributes to one of the teams.
Presenter: Ada Kovaliukaite, Texas A&M University Title: Can Friendship Paradox Generate Biased Inference in Social Networks? Abstract: Literature in Economics and Sociology suggests that in certain social contexts individuals may possess biased beliefs regarding the behavior of the population. Jackson (2016) proposes that, in the context of socially complementary behaviors, such misperceptions may arise due to the correlation between individual preferences and social network connectivity. In this paper, we present a laboratory experiment, where we exogenously manipulate the correlation between induced individual preferences and social network connectivity. We demonstrate that subjects exhibit overconsumption of an abstract, socially-complementary good and upwards-biased beliefs regarding the consumption of others when this correlation is positive. However, participants choose optimal levels of consumption and possess accurate beliefs when this correlation is set to zero. Evidence from a continuous-time belief elicitation suggests that subjects use a naive averaging rule to form beliefs about the behavior of the population from their observation of social network neighbors.
Presenter: Melanie Dunger, University of Göttingen Title: Reporting my peer? The likelihood to blow the whistle as a function of relative closeness Abstract: We conducted a laboratory experiment that captures the effect of relative closeness of subjects on the likelihood to denounce stealing. First results indicate that participants are less likely to denounce an in-group member than an out-group member, especially if the amount stolen is higher. An extensive survey including a vignette study, SVO test, personality questions etc. is combined to the experiment with the aim to describe personal factors that influence the likelihood to blow the whistle.
Presenter: Franziska Heinicke, Utrecht University Title: Reputation concerns and hypothetical bias in willingness to pay elicitation: The case of a private good Abstract: Stated willingness-to-pay from hypothetical surveys generally exceeds revealed willingness-to-pay from real purchase experiments. One explanation for this hypothetical bias that has been put forward is that participants use high willingness-to-pay statements for public goods to improve their moral self-image. We formalize this argument by building on the reputation model of Banabou and Tirole (2006) and extend it to private goods by applying the concept of ideal self-congruity from consumer research. In the resulting theory people enhance their self-image by showing support for a desired product-image. The comparative statics effects for changes in the intrinsic valuation of a good and reputation concerns derived from the formal model are tested in a 2x2 between subject design with 395 participants and a private good. The results show that greater valuation for the good, measured by familiarity with the good, affects hypothetical and actual willingness-to-pay similarly but higher reputation concerns as measured by ideal self-congruity have explanatory relevance for the size of the bias.
Presenter: Lisa Spantig, Ludwig-Maximilian-University Munich Title: The Endowment Effect and Savings Decisions of the Poor Abstract: The relative reluctance of individuals to give up things they have been endowed with (the "endowment effect") has been documented in a large body of experimental studies and might have important consequences for the design of contracts and institutions. One area in which the endowment effect may play an important role are savings decisions of the poor. In many developing countries, poor households receive their income in cash. This cash "endowment" may increase the psychological costs of saving, contributing to the widely observed low savings rates of the poor. I study the causal effect of cash endowment on savings decisions of 300 Filipino microfinance clients in an experimental setting, controlling for potential confounds such as transaction costs and time preferences. During an individual interview, study participants decide whether they want to save some of their compensation for survey participation in their existing savings account. Treated individuals already hold on to the cash for about 20 minutes when making the savings decision whereas the control group knows that they will receive the money (less the amount saved) in cash just after deciding. I find that cash in hand makes individuals feel richer. However, this feeling of endowment does not translate into smaller savings amounts. The distribution of savings in treatment and control is merely identical. Using administrative and survey data as well as economic preference measures, I classify savers and investigate treatment effect heterogeneity while controlling for multiple inference.
Presenter: Andreas Nicklisch, HTW Chur Title: Risk-taking, Altruism, Trust, and Traumatization of Refugees: A Field Experiment Abstract: An unprecedented number of refugees from the MENA region has sought refuge in Europe over the last couple of years. They do not only bring with them the - possibly traumatic - flight experience, but also values and norms they have acquired over their childhood and adolescence. Among those, trust, altruism, and risk-taking are of major importance. We ran a number of experiments among refugees in Jordan and Germany. We show that post-traumatic stress disorder crucially influences risk aversion (in comparison to the typical population of Jordan), while risk influences trust in a systematic way. In a second step, we analyze the association between various possibly traumatic experiences before the flight, during the flight, and in their current life, and a trauma index developed by psychiatrists.
Presenter: Kristina Czura, Ludwig-Maximilian-University Munich Title: Group dynamics in microfinance - Leadership qualities and group performance in rural Bangladesh Abstract: We seek to understand which organizational elements within groups such as monitoring, punishment and leadership quality are important for positive economic outcomes. Motivated by the fact that group lending structures can improve repayment rates in microcredit lending even when borrowers are individually liable, we study this question in the context of a large scale lab-in-the-field experiment with more than 1500 members of lending groups in rural Bangladesh. These groups meet regularly, have distinct group leaders and rely strongly on social capital for loan repayment enforcement in the absence of physical collateral and legal institutions. We elicit experimental measures of the importance of monitoring, punishment and leadership and collect survey data on additional group characteristics. Linking the elicited measures to individual repayment data allows us to disentangle which elements within lending groups are crucial for positive outcomes.
Presenter: Susanna Grundmann, University of Passau Title: Money Illusion in an Experimental Labour Market Abstract: Money illusion is commonly investigated by observing how subjects interact in response to an exogenous inflation shock (set by the experimenter). Evidence on money illusion has been related to limited reasoning, where players may temporarily consider others to disregard the inflation shock and then adjust insufficiently. In a more realistic experimental design, we observe how subjects in the role of employees react to nominal and real wages (and thus inflation shocks) set by the employers. We implement this as a standard gift exchange game, with an endogenous exchange rate between the experimental currency unit and the payoff in euros. In each period, this exchange rate depends on the level of inflation, modelled by the average wage paid by other employers, such that average real wages are constant. Controlling for real wages, we find a short-term influence of nominal wages across all treatments, including those in which real wages are also reported to employees. Tentative evidence implies that a long-term influence of nominal wages, that is, money illusion, persists in treatments in which the real wage is not immediately reported. We argue that this is not explained by limited reasoning but by employees' reciprocity towards employers' good intentions, which might persist in the long run.
Presenter: Knut-Eric Joslin, BI Norwegian Business School Title: Market Entry with Frictional Matching and Bargaining: Labor Search in the Lab Abstract: This paper studies an experimental labor market that incorporates elements from the standard theory of equilibrium unemployment. Specifically, we test in a controlled lab setting a novel market entry game that includes matching frictions and wage bargaining. The model predicts that firms will enter up to the point at which stochastic rationing of workers equalizes the value of a vacancy with its costs. Between treatments, we vary productivity and bargaining strength. Consistent with theory, we find that increases in productivity increase job creation and thereby reduce unemployment. We also reproduce the expected outcomes associated with different forms of wage negotiation. When wages are determined by bargaining after match, firms face a hold up problem. As a consequence, job creation collapses when workers have excessive bargaining power. In contrast, when wages are determined prior to entry, workers moderate their wage claims to induce vacancy creation. Although our findings tend to align with theory, we observe some deviations. In particular, there is a systematic bias in the aggregate level of entry: There is too much vacancy creation when productivity is low and too little vacancy creation when productivity is high. To explain this bias and to account for heterogeneity at the individual level, we estimate a quantal response equilibrium in which we allow for idiosyncratic preferences for entry.
Presenter: Tibor Neugebauer, University of Luxembourg Title: An experimental comparison of two market institutions Abstract: We conducted experiments with asset markets and money markets to address questions of optimal resource allocation, consumption smoothing and equilibrium pricing.
Presenter: Aneeque Javaid, University of Osnabrück Title: Playing with beliefs: why strategy is not independent of beliefs about the state of the world Abstract: According to the literature on cooperation in social dilemma, an individual's (positive and normative) beliefs about a given situation are important in determining her response. This is especially true with respect to an individual's (positive) beliefs about what others are going to do in the same situation. For instance, individuals are more likely to cooperate in a social dilemma situation if they have a general belief that others are also likely to cooperate. In this context, an individual's beliefs represent her subjective probability of other players cooperating. In experiments conducted with students in Zanzibar, Tanzania (n=120), we examine what happens when this probability (of other player cooperating) is manipulated exogenously. We use a modified version of simple 2-person prisoner's dilemma game where participants know the probability of defection/cooperation by the other player. More specifically, we are interested in two conditions; first where the probability of cooperation is high (High-C) and secondly where the probability of cooperation is low (Low-C). We conduct these experiment using strategy method where participants have to decide for all possible scenarios, thereby (potentially) decreasing the importance of beliefs about other player's action. We find that a majority of the participants change their strategy based on the probability of (non) cooperation. We observe three major types; (i) those who do not change their strategy, (ii) those who are conditional (or unconditional) cooperators in High-C situation and free-riders in Low-C situation, and lastly (iii) those who are conditional (or unconditional) cooperators in Low-C situations and free-riders in High-C situation. Furthermore, we observe that this finding holds even after controlling for the cognitive ability of the participants as well as their personality type. Overall, our findings indicate that even contingent choices (i.e. strategies which theoretically cover response to all possibilities) are not immune to overall beliefs regarding the underlying situation and are indeed influenced by these beliefs.
Presenter: Krzysztof Drachal, University of Warsaw Title: An application of some Bayesian model combination scheme with model uncertainty to spot oil price Abstract: Forecasting oil price is a difficult task. The oil market structure is quite complex, including nonlinearities, chaotic behaviours, etc. Therefore, it has been found that in various periods, different factors are the important oil price drivers. In particular, in some period a model build from certain variables can work well, whereas in the other period, another set of variables can constitute a better model. As a result, there is an uncertainty about the true set of oil price drivers. Moreover, there are arguments favouring time-varying parameters in the considered models. For example, in a regression model to allow regression coefficients to vary in time. Naturally, these features can be implemented with Bayesian methodology. Indeed, Dynamic Model Averaging (shortly: DMA) allows for both time-varying: the state space of the model and its parameters [Raftery, A.E., Karny, M., Etler, P., 2010. Online prediction under model uncertainty via Dynamic Model Averaging: Application to a cold rolling mill. Technometrics 52, 52-66.]. The final DMA forecast is computed as a weighted forecast from the predictions of all models possible to construct from the given set of variables. The weights are updated recursively with respect to the quality of forecast produced by the models in preceding periods. Its variation, Dynamic Model Selection, instead of averaging chooses one model (i.e, the one that receives the maximum support from the data). However, there is also possibility to choose the median probability model. Except that, the data about Internet search terms can also be reasonably included into the above Bayesian modelling schemes. These methods can help to answer: How the most important oil price drivers vary in time?
Presenter: Ciril Bosch-Rosa, Colegio Universitario de Estudios Financieros & TU Berlin Title: A diagnosis on the relationship between equilibrium play, stated beliefs, and best responses. Abstract: Strategic games have two dimensions of difficulty for subjects in the laboratory. One is understanding the rules of the game and forming a best response to whatever beliefs they hold. The other is forming these beliefs correctly. Typically, these two dimensions cannot be disentangled as belief formation crucially depends on the understanding of the game. We present a variation of the Two Player Guessing Game (Grosskopf and Nagel, 2008) which turns an otherwise strategic game into an individual decision-making task. This allows us to perform a within subject analysis of the decisions made for the same "game" with and without strategic uncertainty. The results show that subjects with a better score at the individual decision making task form more accurate beliefs of other player's choices, and better-respond to these beliefs. Additionally, we show that those who score higher at our new task modify their beliefs based on the population they play against. This suggests that out of equilibrium play is mostly driven by a limited understanding of the game mechanics.
Presenter: Sergio Mittlaender, Max Planck Institute for Social Law and Social Policy Title: Winners don't cheat Abstract: Humans can commit to a future course of action by the making of a promise. In real-life social dilemmas, we rely on non-enforceable agreements in which we promise to each other to cooperate, imposing upon ourselves a moral obligation to comply. While promises without sanctions are often broken, it is our human tendency to punish violations of moral norms more fiercely than mere defections. Promises induce cooperation not only because they signal willingness to cooperate, but mainly because they trigger higher punishment in case they are broken, which in turn induces higher cooperation. The institutions of promises and punishment, largely unique to humans, might hence lead to the preponderance of cooperation even in single, anonymous interactions. This article presents experimental results that confirm this hypothesis, and reveals that while the possibility to promise or to punish each increases cooperation, it is only the efficient interaction between moral commitment and punishment of wrongdoers that leads most people to cooperate. The profit-maximizing strategy was to cooperate, and to suppress the temptation to defect after a successful agreement. Cheaters, in contrast, were harshly punished, and their earnings were almost zero. Those that made and kept promises avoided punishment, and achieved the highest gains.
Presenter: Urs Fischbacher, University of Konstanz Title: Incentives for conformity and disconformity Abstract: There is abundant evidence for conformity but there are also situations in which people try to set themselves apart. We investigate how punishment and reward affect these behaviors. We rely on a 3 (punishment vs. no incentive vs. reward) x 2 (arts vs. quiz) experiment design. First, two subjects make a binary choice. In the arts treatment, they choose one out of two arts postcards and in the quiz treatment, they choose one out of two answers to a difficult knowledge question. Then, a third person makes the same choice, knowing the decision of the first two subjects. Since this third person also makes unconditional decisions in which the two options each are compared with a third option, we can infer whether subjects make conform, disconform, or autonomous choices. Finally, evaluators are shown the three choices of a group and, depending on the treatment, assign a bonus or a deduction to one of the three subjects. We find that punishment leads to more conformity and reward leads to more disconformity, and we find more conformity in the quiz treatments than in the arts treatments. Disconformity is rare. It exists only in the reward treatment and it is stronger in the arts than in the quiz treatment.
Presenter: Ann-Kathrin Koessler, University of Osnabrück Title: Exploiting conformity concerns to create commitment for public good provision Abstract: Sustainability pledges for companies are en vogue, but also in climate negations pledges on emission reductions are commonly observed. First experimental studies investigated the effectiveness of such public promises in the controlled environment of the laboratory. Contrary to this research, pledges in reality are not taken simultaneously, but in succession. In this study we examine how conformity concerns impact the decision to make a pledge and consequent behaviour. In a public good game, subjects have prior play the possibility to pledge publicly to provide social optimal contributions. These pledges are made voluntarily and taken simultaneously or sequentially. We find sequential elicitation increases pledge-making, particularly when individuals' inherent social orientation is considered in the elicitation order. If initial high-contributors are asked first, also previously low contributors make the pledge due conformity concerns. This commitment nudge howsoever does not come negative behavioural effect. On the contrary, previous low contributors increase their contributions significantly and the pledge helps groups to coordinate on better social outcomes.
Presenter: Tobias Gesche, University College London Title: Reference Prices and Customer Antagonism in Auctions Abstract: Using data from a large scale auction on Ebay, I show that buyers get antagonized towards the seller when they learn that the item they acquired was later sold for a lower fixed price. They punish the seller by giving a hostile, publicly visible seller rating. The probability that they do so is increasing in the price difference to the fixed price even though buyers determined a cap on the price they paid through their bid. These price effects on ratings cannot be explained by seller or good characteristics as these were identical across auctions. The findings are consistent with a proposed model which combines reference-dependent with reciprocal preferences and in which downward shifts to consumers' reference prices have a disproportionally large negative effect on customers' utility and the resulting seller ratings.
Presenter: Alexander Heczko, RWTH Aachen University Title: Core-Selecting Auctions in the Laboratory Abstract: We experimentally analyze core-selecting auctions and find that they perform better than Vickrey auctions. The proportions of efficient allocations are similar in both types of auctions, but the proportions of stable (core) allocations and the revenue are higher in the core-selecting auctions. We observe this for two different informational setups, one in which theory predicts the better performance of the core-selecting auction and one in which it does not. We trace the causes of the performance differences back to patterns in bids. This shows that in the core-selecting auctions, bidders react to incentives to deviate from reporting their valuations truthfully in the predicted direction, though less pronounced than predicted.
Presenter: Dirk Engelmann, Humboldt University Berlin Title: Winner's and Loser's Curse in Single-Unit and Multi-Unit Auctions Abstract: We show theoretically that naive bidding, ignoring the implications of winning the auction, leads not only to a winner's curse in a single-unit second-price auction, but also to a corresponding loser's curse in a multi-unit auction where n-1 units are sold to n bidders at a uniform price equal to the lowest bid if the distribution of valuations is symmetric. We experimentally study an application to a wallet auction. Average bidding behavior in the single-unit auction is remarkably close to naive bidding, but average bidding in the multi-unit auction is even more timid than naive bidding would suggest.
Presenter: Oktay Sürücü, Bielefeld University Title: Asymmetric Dominance Effect with Multiple Decoys for Low- and High-Variance Lotteries Abstract: The asymmetric dominance effect refers to the phenomenon according to which the choice probability of an alternative increases when an inferior alternative - the decoy - is included into the choice set. The objective of this experimental study is twofold. First, we investigate the asymmetric dominance effect on two-outcome lotteries with almost equal expected values. We find that the impact of a decoy on low-variance lotteries (LVLs) is much higher than on high-variance lotteries (HVLs). Second, we examine the asymmetric dominance effect in the presence of two decoys. While the asymmetric dominance effect persists when the choice set includes two decoys, the effect is not always further enhanced compared to the setting with one decoy and again much stronger for LVLs than for HVLs. Controlling for subjects' degrees of risk aversion, we find support for consistency between individual risk preferences and choice behavior among the lotteries. However, we observe decoy effects of equal strength irrespective of the subjects' degree of risk aversion. Thus, our analysis indicates that to a substantial extent the presence of decoys subtly makes decision-makers choose against their risk preferences by favoring lotteries that entail risks contrary to their elicited individual risk-taking profile.
Presenter: Kjell Arne Brekke, University of Oslo Title: Does competence reduce ambiguity aversion? Abstract: Heath and Tversky (1991) suggested that participants will be more willing to gamble on an ambiguous natural event over a lottery with well-defined risk, the more competent they felt evaluating the possible outcome of the natural event. In a series of experiments, we ask participants to first state the probability of each of five comprehensive and mutually exclusive outcomes of a natural event. As previous studies found, the stated probabilities added to more than 150% on average. We then randomly picked one outcome "the given alternative" out of the five possible outcomes. Participants where than asked to choose between (a) a gamble where they would win if the given alternative occurred, or (b) a random lottery where they win with a given probability. We find that participants have a strong preference for lotteries over gambles, in one experiment even when the stated probability for the given alternative was three times as high as the given probability of the lottery. We argue that this can be interpreted as an indication of ambiguity aversion. We find however that the degree of ambiguity aversion varies across different experiments. In the first experiment, the natural event was a possible change of the key policy rate set by the Central bank of Norway, and we recruited participants from medicine, mathematics and language classes. In this case, we found a very strong preference for the lottery - indicating strong ambiguity aversion. It seems reasonable that these participants had low confidence in their assessment of Central bank policy. Experiment 2 used the Republican Iowa caucuses as a natural event, with Donald Trump as one candidates and with participants from Mechanical Turk. Here we observed much less ambiguity aversion. This event had such a huge media coverage that it seems reasonable to think that participants felt much more competent. The results thus seems consistent with Heath and Tversy's competence hypothesis, applied to ambiguity aversion. However, we cannot make a deduction from a comparison between experiments. Moreover, other indicators of competence had no impact on the likelihood of choosing the gamble. To test further we ran a third experiment, this time using the Oscar awards as a natural event. While Central bank policy was very unfamiliar to almost all participants in the first experiment and few in the second experiment had avoided news about the Trump campaign, we believe that the knowledge about movies is much more dispersed among Amazon Mechanical Turk participants. In addition to the choice between gamble and lottery as above, we also elicited knowledge about movies and the actors in the set. We predict that participants with less knowledge about movies and the actors will exhibit more ambiguity aversion. References Heath and Tversky (1991): Preference and Belief: Ambiguity and Competence in Choice under Uncertainty, Journal of Risk and Uncertainty, 4:5-28 (1991)
Presenter: Ange Weinrabe, University of Sydney Title: Economic Rationality in Youth with Emerging Mood Disorders; Evaluating the Impact of Dis-regulated Affect and Emotion in Adolescents using Economic Decision-Making tools Abstract: Anxious and depressed people are often characterized as irrational in common language. Underlying and contributing to "irrational" behaviour are biological and psychological substrates, such as motivation and emotion. However, in mental health care a clear definition of what it constitutes to be irrational is missing. In economics, on the contrary, irrationality in behavior is precisely defined as making intransitive choices or, violating the Generalized Axiom of Revealed Preference (GARP). Previous research has shown that young children, older adults and people with ventromedial prefrontal cortex damage show intransitivity in their choices. In this study, we extend these results by investigating whether people with severe mood disorders make economically rational choices. At two time points, separated by eight weeks, we measured subjects' rationality using a widely used economic paradigm and we quantified the severity of subject's mood via widely used psychological self-report scales for youth disorders (Kessler Psychological Distress K10), Quick Inventory of Depressive Symptomology QIDS-A17 and SPHERE-12). We found that help-seekers, rated as severely anxious and depressed, are more likely than those who scored lower on these scales, to make choices that violate GARP. Importantly, we found that help seekers became less irrational at round two, but that little within subject variation in mood occurred.
Presenter: Özgür Gürerk, RWTH Aachen University Title: Understanding the Behavioral Drivers of Execution Failures in Retail Supply Chains: An Experimental Study Using Virtual Reality Abstract: We conduct a real-effort experiment in an immersive virtual environment in order to quantify the role of product similarity in execution failures in a retail setting. In our experiments, subjects must identify and sort two types of products based on their observable characteristics. When the two products are very similar, performance is substantially lower (with both more sorting errors and more products left unsorted) than when the products are more dissimilar. Introducing a clear visual cue to distinguish the products improves execution when the products are dissimilar (by lowering sorting mistakes) and, even more so, when they are similar (both by reducing sorting mistakes and the number of products unsorted). In the latter case the average increase in overall performance is over 22 percentage points. Our results suggest that there may be important gains from taking ease of execution into account in the design of products, product packaging, and labeling.
Presenter: J. Frederik Graff, RWTH Aachen University Title: Competing on the Holodeck - On the effect of virtual peers and heterogeneity in dynamic tournaments Abstract: For obvious reasons tournaments in real organizations usually involve heterogeneous contestants which has often been modeled for example by different cost functions. In a real- effort environment one naturally captures different abilities of subjects but the experimenter loses control in the sense that one cannot easily match pairs of specific abilities as necessary to analyze a research question. Moreover, tournaments typically involve some dynamic with participants' reaction towards midterm results which is not modeled in most experimental settings. We offer a new approach and simulate a realistic working situation in a highly immersive environment. Implementing a tournament in virtual reality (VR) allows us to ceteris paribus control for the performance of the humanoid avatar and, thus, to get an understanding of the reaction of the subject to the avatar in a really dynamic setting as the subject is constantly able to observe the avatar's performance. Additionally, using VR can solve the "reflection problem" (Manski 1993), which states that in experiments where participants interact, it can never be disentangled whether one participant i influences participant j or vice versa. In VR, avatars can be programmed to execute all kinds of behavior - always in exactly the same way and are never influenced by the participant's reactions (Gürerk et al. 2014). We compare three treatments: (i) a piece-rate scheme, (ii) a symmetric tournament against an equally able avatar, and (iii) an asymmetric tournament against an avatar with higher ability. First results indicate that participants perform best when matched in a dyadic tournament with an opponent of the same ability.
Presenter: Özgür Gürerk, RWTH Aachen University Title: Presentation of Virtual Reality Technology
Presenter: Florian Kerzenmacher, Frankfurt School of Finance & Management Title: Lying opportunities and incentives to lie: Reference dependence versus reputation Abstract: Recent experiments on lying behavior show that the lying frequency in case of low outcomes increases in the ex ante probability of high outcomes; a finding that can well be explained by (external) reputation costs, but also by models combining loss aversion with internal lying costs. To compare the explanatory power of these two approaches, we design an experiment in which we manipulate the ex ante probability that lying is possible at all (q). We show theoretically that the reputation model then predicts that the lying frequency decreases in q, while the loss aversion model suggests the opposite. Our experimental results strongly support the reputation model. From an applied perspective, our results suggest establishing safeguards for reducing the ex ante probability that lying is possible may (partially) backfire.
Presenter: Jarid Zimmermann, University of Cologne Title: On the motivated acquisition of social information Abstract: We suggest that information avoidance may not only be motivated by the desire to appear prosocial to oneself, but also by the desire to appear prosocial to others. We design an experimental study to test this idea. An individual ("dictator") decides on a transfer to a recipient. Prior to deciding, dictators choose among multiple sources of information about the transfer decisions of previous dictators. Our treatments vary whether or not this information becomes public and whether or not recipients can punish their dictators. We provide compelling evidence for information acquisition supporting prosocial appearance, in particular when acquired information becomes public information. Thus, we identify the desire to appear prosocial to others as a key driver in information acquisition, operating independent from payoff concerns resulting from punishment possibilities. We further provide first clean evidence that information choices predict behavior: Irrespective of the received information, we find that individuals who choose information supporting prosocial appearance display lower prosocial behavior than individuals choosing other information sources. This suggests that heterogeneity in prosocial motives is an important determinant in explaining information avoidance. Our results further indicate important limitations of policies making information public: First, aggregate behavior may be less prosocial when social information is made public. Second, individuals willing to avoid certain information may be less likely to follow its implications, and change their behavior, when receiving it.
Presenter: Astri Drange Hole, NHH Norwegian School of Economics Title: Social preferences and honesty Abstract: The project studies if there are any systematic differences in social preferences between managers and people in general. More specifically, the project examines if the sources of income and the cost of redistribution matter for people in the two groups when they make distribution decisions. Given this information, we can also study the prevalence of different fairness ideals in the two groups: if the two groups are motivated by different moral norms when they decide on important distributive issues. We then test for any differences between the two groups on these matters. The project also measures existence of dishonesty. We map any pattern of lying behaviour amongst managers and people in general, and test for any difference in this pattern between the two groups, and we test if a high degree of honesty in a group correlates with social preferences in that group. Furthermore, we examine if there is a relationship between distributive choices and attitudes towards materialistic values and willingness to compete, respectively. We also explore if there is a correlation on group level between existence of dishonesty and focus on materialistic values and willingness to compete, respectively. Finally, we examine if demographic and/or socio-economic variables can explain any difference between the two groups. The web experiment and the surveys are designed by the author. The designs are programmed and carried out by Norstat. The panel of participants recruited by Norstat are offered a choice between gift cards or donation to charities as show-up fees. The experiment is incentivised by points added to a panel account from which the participants can buy items online. The data analysis is ongoing. A pre-analysis plan including the data sources, the design of the experiments and the surveys, the empirical strategy and the experimetrics are registered at the AEA RCT Registry.
Presenter: Adam Zylbersztejn, GATE CNRS, University of Lyon Title: Cooperation in a risky world Abstract: We offer a novel investigation of the effect of environmental risk on cooperation in the Voluntary Contribution Mechanism. Our baseline is the standard setting in which the personal return from the public good is deterministic, homogeneous, and publicly known. Our experimental treatments alter this classic design by making the marginal per capita return from the public good probabilistic. In the homogeneous risk treatment (HomR), the random draw is made for the whole group, while in the heterogeneous risk treatment (HetR) this happens independently for each group member. Our hypothesis is that different environmental risks may differently affect the ex post payoff inequalities, so that other-regarding preferences (inequality aversion) may generate higher contributions in HomR than in HetR. Our main result is that the environmental risk does not affect the patterns of cooperation either in the one-shot or in the finitely repeated version of the game. This suggests that the standard experimental methodology provides a robust and conservative measure of human cooperation.
Presenter: Arjun Sengupta, Nanyang Technological University Title: Contribution and Punishment Norms in Public Good Games Under Income Inequality from Nominal Choice Abstract: Behavioral economics models of inequity aversion and altruism predict that more high income individuals would contribute more to public goods than low income individuals, but this prediction has not held in past research (Buckley and Croson, 2006). We tested the hypothesis that high income individuals make smaller proportional contributions to public goods, and low income individuals allow them to do so, because of the underlying belief that high income individuals made good choices and thus deserve to keep a bigger proportion of their income. We use a lab experiment to investigate whether income inequality generated through nominal choice can influence people's perceptions of what is a fair contribution in a public good game with the opportunity to punish. In our baseline treatment, the income inequality was generated randomly within a group. In our nominal choice treatment, subjects make a nominal choice but were aware that the income inequality was still randomly generated, as in the baseline condition. We found that the overall contribution towards public goods decreased when subjects make a nominal choice. Although high income individuals contributed more than low income individuals across both treatments, the high income individuals contributed less in the choice condition compared to the baseline condition. Moreover, those who have high income are punished less in the choice treatment than in the baseline condition, suggesting that individuals expect lower contribution from high income individuals when people's income is based on a nominal choice. We conjectured that such a change in perception of fair contribution in the choice treatment will be because of belief that good choice leads to good outcome. We find that this is not the case, suggesting that an act of choice itself can change preference for fairness.
Presenter: Jia Liu, Newcastle University Title: Information defaults in repeated public good provision Abstract: We present an experiment on information defaults and information seeking in a repeated public goods provision setting. In our experiment the default is one either with or without information about others’ contributions, and having information comes either at a financial cost or at a financial benefit. Subjects mostly follow the money in deciding whether to have the information or not, but around a third seek or stick to information even when this is costly. However, a default of not having information about the others’ contributions leads to a slower unraveling of cooperation, whether or not the information is financially costly or beneficial. This slower unraveling is explained by the dynamic of beliefs about others’ contributions in these treatments. A secondary informational default effect appears to take place. When the default is no information, subjects do not seek information more but, conditional on taking financial incentives into account, they tend to believe that more other subjects seek information.
Presenter: Chiara Nardi, Frankfurt School of Finance & Management Title: Gender, self-selection and moral hazard in teams Abstract: Motivated by previous research on incentive design in organizations, we conduct a laboratory experiment that investigates the role of specific contract features in the effort choice, and examines if and how these contract features interact with gender. We compare an individual payment scheme to a team payment scheme in which partners equally share the produced output and bear effort costs individually. We distinguish between random assignment and self-selection to payment schemes. Our analysis reveals interesting results. First, females self-select into team payment more often than males. Second, females who are randomly assigned to team payment are more likely to choose the first-best (cooperative) effort level than males. Third, males (but not females) who prefer to be rewarded with team payment are more likely to choose the cooperative effort level than those who are randomly assigned to it. Finally, both the payment choice and the effort choice are driven by beliefs on the partner's behavior, irrespectively of gender.
Presenter: Ofer Azar, Ben-Gurion University of the Negev Title: Does relative thinking exist in mixed compensation schemes? Abstract: Several earlier studies have shown that people exhibit "relative thinking" they consider relative price differences even when only absolute price differences are relevant. The article examines whether relative thinking exists when people face mixed compensation schemes that include both fixed and pay-for-performance components. Such compensation schemes are prevalent in many occupations (e.g., salespeople and managers) and therefore are an important practical issue. Surprisingly, the ratio between the pay-for-performance and the fixed compensation does not affect effort, meaning that no relative thinking is found. Another experiment shows that this is not due to reciprocity that cancels out relative thinking. In a third experiment subjects make similar decisions without incentives and the results suggest that the different context (compensation schemes instead of price comparisons) and not the introduction of financial incentives (which were not used in previous studies) is the reason that relative thinking disappears. The results have implications for designing incentive schemes in firms and for designing experiments.
Presenter: Heike Hennig-Schmidt, Bonn University Title: Performance Appraisal Systems and Cooperation in Teams – Evidence from a Lab Experiment in Germany and China Abstract: In the present study, we investigate the application and incentive effect of two classical (Western) performance appraisal systems - free attribution of ratings and predetermined attribution of ratings (forced ranking) - in two culturally different subject pools (Bonn, Germany; Chengdu, China) by applying a controlled real-effort laboratory experiment. We find a leniency and a centrality bias in both countries, yet the effects are stronger in the Chinese subject pool. Introducing a forced ranking system can pay off as output per Euro increases by 6%. If supervisors differentiate it is effective in the German subject pool: undeserved high rankings demotivate whereas low rankings motivate workers whereas in China we find the reverse effect; also workers want to avoid the worst ranking. We also find that forced ranking affects team cooperation: workers tend to contribute less, yet supervisors contribute more. Ramifications for the design and application of human resources practices applied in Non-Western work environments like China are discussed.
Presenter: Felix Holzmeister, University of Innsbruck Title: Ethnical Discrimination in Europe: Field Evidence from the Finance Industry Abstract: The integration of ethnical minorities has been one of the hotly debated issues in politics in the last two years and it is safe to say that it will remain one of the big challenges for societies over the next years. One factor that can hinder successful integration is the persistent discrimination of ethnical minorities. This is particularly relevant in the finance sector, where unequal access to investment and financing opportunities can cause social and economic exclusion as well as disparities due to inferior economic prospects. We provide the first crosscountry study investigating ethnical discrimination in the finance sector in a controlled field experiment with 1,218 banks in seven European countries. In particular, we contacted banks via e-mail with domestic sounding and Arabic sounding names as senders, asking for contact details. We find pronounced discrimination in terms of a substantially lower response rate to e-mails with Arabic sounding names compared to domestic sounding names. Remarkably, the observed discrimination effect is robust over several dimensions, in particular for loan- and for investment-related requests, across countries, and across banks located in rural as well as urban areas.
CANCELED:
Presenter: Jana Freundt, University of Hamburg, Title: Green Assets on Financial Markets
Presenter: Lukas Wenner, University of Cologne Title: Disclosure of Socially Responsible Product Information in Competitive Markets Abstract: We study competitive markets for products with a social externality in production. We experimentally vary the degree to which this externality is private information. Our main question of interest is whether market forces lead to a proliferation of “high qualityâ€, i.e. generate socially responsible products, when disclosure of the social impact of the product is voluntary. Standard theory predicts that this will be the case, through a channel similar to a simple unraveling logic. However, there is evidence that in many real-world markets firms do not disclose attributes of their products such as working conditions or the environmental impact, casting doubt on the theoretical argument. In our experiment, sellers can invest into products that are costlier to produce but generate a higher payoff to a third party (in our case a charity). We show that if there is forced disclosure, market forces lead to much higher positive externalities than if the social impact is not visible to consumers. However, when sellers are given the choice whether to reveal the social dimension, we find that disclosure is imperfect and market outcomes do not replicate the full information condition. We discuss the implications of our findings for the effectiveness of voluntary disclosure regimes.
Presenter: Ferdinand Langnickel, University of Zurich Title: Private Investors and Public Information Abstract: Private investors trade too much. The common explanation for this is overconfidence which assumes that people have private information about a stock and overestimate the precision of that information. But in reality, private investors typically only have access to public information. Trading on public information, however, should not be profitable for private investors since professional investors are likely to have used the information before them. In this study, I analyze whether private investors trade too much because they neglect other market participants when processing (public) information. This study is based on two experiments in which participants trade an artificial asset that pays out a random dividend to its owners. Prior to trading, participants may receive a noisy, unbiased signal about the dividend. In Experiment 1, I elicit supply and demand curves from participants as in Biais et al. (2017) to obtain market prices of the asset for different signal values. In Experiment 2, participants can buy or sell shares of the asset at the market price from Experiment 1. Participants cannot observe the price, but they know that it is the market-clearing price from the the round of Experiment 1 where all participants received the same dividend signal. Thus, participants should base their trading decision on their updated beliefs about the dividend and the behavior of the market participants. The main result is that participants in Experiment 2 (N=27) fail to anticipate the behavior of participants in Experiment 1 (N=22). As a result, the number of shares bought by participants significantly increases with the dividend signal. On average participants own 1.75 times more shares at the best signal value compared to the worst signal value (t = 2.72, p = 0.01) even though the market prices lead to almost identical return distributions in both cases. Further sessions are scheduled to be conducted in the next months. Bibliography: Bruno Biais, Thomas Mariotti, Sophie Moinas, and Sebastien Pouget, "Asset pricing and risk sharing in a complete market: An experimental investigation", TSE Working Paper, n. 17-798, April 2017.
Presenter: Philipp Marquardt, University of Mannheim Title: Asymmetric Reactions to News under Risk and Ambiguity Abstract: Two stylized facts in experimental research on individual decision behavior under ambiguity are that attitudes towards ambiguity are heterogeneous and that attitudes are domain dependent: individuals tend to be ambiguity averse in the gain domain and ambiguity seeking in the loss domain. We propose and experimentally test a heterogeneous agent model, to study how this individual-level behavior aggregates into prices in an experimental asset market. Specifically, we compare market price reactions to good and bad news in information environments that are risky (known probability distributions) or ambiguous (unknown probability distributions). On the individual level, we find that individuals are more ambiguity averse in the gain than in the loss domain. This pattern also translates into prices, where we find an asymmetry in the reaction to good and bad ambiguous news. We find that the ambiguous asset is overvalued relative to the risky asset following bad news and undervalued following good news. Prices also converge more slowly towards rational expectations prices under ambiguity than under risk. Convergence under ambiguity is slower, because sophisticated agents (proxied for by cognitive ability) engage less in arbitrage under ambiguity.
Presenter: Olivier Armantier, Federal Reserve Bank of New York Title: Discount Window Stigma: An Experimental Investigation Abstract: A core function of central banks is to act as a "lender of last resort" to the banking system. In the U.S., the Federal Reserve (the Fed) uses the Discount Window (DW) to fulfill this task. Historically, however, the DW has been little used, even when banks faced acute liquidity shortages. This lack of DW borrowing is commonly attributed to stigma, as illustrated by Chairman Bernanke's quote above. The economic consequences of DW stigma may be severe for the financial system (e.g. fire-sales, excessive self-insurance, failures). Further, DW stigma may prevent central banks from effectively providing emergency liquidity and implementing monetary policy. Despite its relevance, DW stigma is not a well-understood phenomenon, in large part because of the lack of data. In this paper, we use lab experiments to generate data in order to better understand DW stigma and how to eliminate it. To the best of our knowledge, this is the first paper that relies on experimental methods to address DW stigma and more generally the provision of emergency liquidity by central banks.
Presenter: Stephan Müller, University of Göttingen Title: Decisions under Uncertainty in Social Contexts Abstract: This paper theoretically and experimentally studies decision-making in risky and social environments. We explore the interdependence of individual risk attitudes and social preferences in form of inequality aversion as two decisive behavioral determinants in such contexts. Our model and the data demonstrate that individual risk aversion is attenuated when lagging behind peers, whereas it is amplified under favorable income inequality. Moreover, people's choices are not only contextdependent, but are sensitive to their degree of inequality aversion. The majority of our experimental findings cannot be rationalized by rank-dependent utility models or cumulative prospect theory. Our results contribute to the basic understanding of the underlying principles of decision-making in environments like trading at stock markets, the saving patterns of households or charitable giving under uncertainty.
Presenter: Helga Fehr-Duda, University of Zurich Title: A Tale of Two Tales: The Impact of Interactions of Risk and Time on the Valuation of Tail Events Abstract: The majority of economic decisions involve consequences that are both risky and delayed. While the bulk of previous research has focused on measuring atemporal risk preferences, a growing body of evidence shows that risk tolerance is delay dependent, usually increasing with the delay to the resolution of uncertainty. I will argue that the driving force of this effect is the inherent uncertainty of future events which generates an additional layer of uncertainty not accounted for by the objectively given probabilities. This future uncertainty conjointly with people's proneness to nonlinear probability weighting provides an explanation for the co-existence of underweighting and overweighting of tail events, manifest, for example, in the low take up of disaster insurance and the high willingness to purchase extended warranties, respectively.
Presenter: Christoph Huber, University of Innsbruck Title: Scaling, Volatility, and Risk Perception Abstract: We investigate whether the perception of risk and return expecations depend on the presentation format of past return and past price charts with regard to di erent scalings of the y-axes. For this, we run an experiment in which we expose subjects to return and price charts of hypothetical low and high volatility assets and ask them to rate and compare them regarding their risk, their return expectations, and the probability with which they want to invest. We nd that a narrow scaling of the y-axis leads to a signi cantly higher perceived risk of an asset across di erent price trends and across price and return presentations { even if the underlying volatility is the same.
Presenter: Irenaeus Wolff, TWI/University of Konstanz Title: Belief Uncertainty and Stochastic Choice Abstract: People often cannot assign a clear probability to an event but face uncertainty about their subjective probabilities. We model belief uncertainty by assuming that agents cannot access their inner belief-probability distribution but have to draw a probability from that distribution. Our model produces stochastic choice because each decision-relevant belief is but one realization out of a distribution of many possible beliefs. The model offers an explanation for experiment participants not playing a best-response to their stated beliefs: participants are uncertain what belief to report or base their decision on. In an experiment, we manipulate participants' belief uncertainty by exogenously manipulating their strategic uncertainty, providing varying levels of information about historical choice data. In situations where belief uncertainty is low, observed best response rates are high and increasing in the amount of information we provide. Accordingly, high belief uncertainty leads to lower consistency.
Presenter: Ye Jin, New York University Shanghai Title: What Determines Non-equilibrium Behavior in Games: Wrong Belief or Limited Ability? Abstract: Evidence from the literature suggests that people use very limited reasoning in games, which is usually far from the requirement to reach Nash equilibrium. I design an experiment to identify the decisive factor that prevents people from using more depth of reasoning: their incorrect belief of others' rationality or their limited reasoning ability. The experiment first classifies subjects into different Lk types by the reasoning steps they use in the games. It then distinguishes between the "Lkb" players, who have high ability and best respond to Lk belief, and the "Lka" players, who could use, at most, k steps of reasoning, and thus could not respond to L(k+1) or higher-order belief. The separation utilizes a combination of simultaneous and sequential ring games. In the sequential games it requires more than k reasoning steps to respond to Lk belief, so Lkb players still best respond but Lka would fail. I find that around half of the L2 and L3 subjects are best responding to L2 or L3 belief, while the rest have reached their upper boundaries of reasoning. Additionally, subjects' CRT scores, a measure of their cognitive ability, support the separation of the two types. The findings suggest that both belief and reasoning ability could be the decisive factors of players' observed levels.
Presenter: Kinga Posadzy, Linköping University Title: Should misfortunes come singly? Investigation of individuals' outcome editing in financial and social contexts Abstract: Originating from prospect theory, hedonic editing hypothesis posits that individuals integrate or segregate multiple outcomes in order to simplify prospects and simultaneously maximize the subjective value derived from them. While majority of previous studies focus on testing the predictions of hedonic editing in the financial context, little is known how the outcome editing process looks like in the social context, where the outcomes concern social interactions. The following study uses an online survey with a large representative sample of Swedish population (n=2064) to test whether the preference to combine or segregate outcomes is stable across social and financial contexts. Our findings show that editing of outcomes is context-dependent. While in financial context individuals prefer to segregate two outcomes in the same domain (gain/loss) and integrate mixed outcomes, their preferences reverse in social context once losses are involved. Furthermore, we find differences in what role individual differences play in explaining the editing of outcomes between these contexts. Our findings can help improve the understanding on how to design the choice architecture or present news to individuals to make them better off.
Presenter: Wladislaw Mill, Friedrich-Schiller-University Jena Title: Hillary Clinton supporters and their spiteful behavior towards Donald Trump supporters Abstract: In a large scale study with over 1.000 subjects we investigate the attitudes of supporters of Donald Trump and Hillary Clinton towards each other and how these attitudes affect spiteful behavior. We find that both Trump and Clinton supporters have less positive attitudes towards the opposing supporters. Interestingly, we show that Clinton supporters feel significantly more negative towards Trump supporters. This effect even increased after the inauguration of Donald Trump. Moreover, we show that Clinton supporters behave spiteful towards Trump supporters, while Trump supporters do not differentiate between Trump and Clinton voters. Using a path analysis we show that attitudes mediate spiteful behavior.
Presenter: Karim Sadrieh, Otto-von-Guericke-University Magdeburg Title: Motivational Foundations of Expressive Voting Abstract: Most political and managerial decisions are made using more or less formal voting procedures. Shareholders, management teams, or committee members may either support available options based on instrumental motives (e.g. profit or efficiency goals) or based on expressive motives (e.g. group identity or ethical concerns). We present the first experimental study of expressive voting (i.e. supporting options due to expressive motives), in which both the cost of and the utility gain from expressive votes are systematically varied in a framed field experiment. We find evidence that both dimensions affect the frequency of expressive voting in the theoretically expected directions. The frequency of expressive voting (i) increases with the size of the audience (i.e. when increasing the benefit), but (ii) decreases with the probability that an individual vote is pivotal (i.e. when increasing the expected cost). Interestingly, however, these results mainly concern individuals with a low involvement in the issue at hand. The individuals with a strong emotional attachment to the issue (in our study, to the professional soccer league) tend to behave expressively at any level of cost or audience size. Hence, when high involvement individuals are included, designing a group decision institution without an expressive vote bias may be difficult to achieve.
Presenter: Anita Zednik, WU Vienna Title: "None of the above" as an alternative to protest voting Abstract: In an election, apart from voting sincerely or voting strategically, a voter can decide to use his vote as a signal of dissatisfaction to his most preferred party or the political system. Protest voting is a world wide phenomenon and has increased over the past decades, however the reasons why people protest vote and the strategic interactions of protest voting are not well understood. We investigate the expressive and instrumental role of protest voting in general and of the option "None of these candidates" in particular as an alternative to abstention and blank voting. We present results from two field surveys and lab experiments.
Presenter: Nicolas Fugger, Centre for European Economic Research (ZEW) Title: Exploiting uncertainty about the number of competitors in procurement auctions Abstract: In procurement practice first-price auctions are used if the number of potential suppliers is small and second-price auctions if it is large. This observation cannot easily be explained by standard economic theory as suppliers should anticipate little competition whenever they participate in a first-price auction. We test this setup experimentally and find that buyers employ this strategy. Suppliers on average interpret the buyer's selection of a first-price auction as a signal of low competition. However, most suppliers still overestimate the degree of competition in first-price auctions. As a consequence, they bid too aggressive in first-price auctions, which rationalizes buyer's format choice.
Presenter: Philippe Gillen, Centre for European Economic Research (ZEW) Title: Preferences and Decision Support in Competitive Bidding Abstract: We examine bidding behavior in first-price sealed-bid and Dutch auctions, which are strategically equivalent under standard preferences. We investigate whether the empirical breakdown of this equivalence is due to (non-standard) preferences or due to the different complexity of the two formats (i.e., a different level of mathematical/individual sophistication needed to derive the optimal bidding strategy). We first elicit measures of individual preferences and then manipulate the degree of complexity by offering various levels of decision support. Our results show that the equivalence of the two auction formats only breaks down in the absence of decision support. This indicates that the empirical breakdown is caused by differing complexity between the two formats rather than non-standard preferences.
Presenter: Daniel Wärtenberger, Ulm University Title: Index-Based Insurance in Developing Countries: Rational Neglect? Abstract: Microinsurance adoption in developing countries is low, despite its potential to foster economic growth. Recent research is not able to explain the low take-up rates within the neoclassical framework. I contribute to this stream of research by proposing rational as well as boundedly rational explanations for the low attractiveness of microinsurance within a stochastic framework. More precisely, I analyse weather index insurance that guarantees farmers an indemnity payment contingent on the realized rainfall at the nearest weather station. My model makes separate predictions for "close farmers", whose location is near a weather station, and "distant farmers". Results show that the latter ask for less than 50% insurance coverage even under perfect rationality. I extend the model by integrating two possible cognitive biases that reflect a lack of trust and a lack of knowledge about the insurance. I can show that a lack of trust reduces insurance demand most for close farmers, while a lack of understanding negatively affects the demand of distant farmers. Moreover, subsidies are more effective for close than for distant farmers.
Presenter: Jan Jozwik, Oxford University Title: Behaviour Factors and Adoption of New Technologies bundled with Index Insurance Schemes Abstract: This paper investigates the effect of an ambiguous environment on fertiliser investments under index insurance. This behavioural factor was studied by means of a framed field experiment conducted with Ghanaian cocoa farmers. The subjects had an option to invest in a package of fertiliser bundled with index insurance with a positive level of basis risk. The returns depended both on the subjects' investment choices and a stochastic weather realisation. The key ingredient of the study was that for different subjects, the exact probability of the basis risk was unknown. A large negative effect on fertiliser investments was found in the treatment with large ambiguity regarding the exact level of basis risk. The effect was also negative in the treatment with small ambiguity, but this result was not statistically significant. The results suggest that technologies with which farmers are relatively more experienced may be more likely to be adopted under index insurance schemes. The overall experimental findings provide evidence that ambiguity may be a significant factor other than basis risk, limiting the effectiveness of index insurance in promoting agricultural innovation.
Presenter: Douadia Bougherara, INRA Title: Production Risk and Input Use: Impact of Subsidies and Crop Insurance Abstract: We study the impact of subsidies and crop insurance on the use of two types of inputs: risk decreasing and risk increasing. This research question is not new in agricultural economics. However, although theoretical predictions are established, the empirical evidence is not clear-cut. We argue that advances in experimental economics can give new insights on that issue. Our experimental approach allows dealing with three severe issues that face agricultural economists when using production data. First, the separate identification of risk preferences and technologies is possible in the lab since the experimenter controls for the parameters of the production function. Second, we overcome simultaneity issues such as the simultaneous choice of input use and crop insurance using a careful design of treatments in the lab. Third, we escape the selection bias issue by imposing mandatory insurance in some treatments as compared to others without insurance. We carry out 8 sessions where subjects choose the level of risk decreasing or risk increasing input to maximize profits under three types of treatments: benchmark, subsidy and actuarially fair crop insurance. We also elicit subjects' risk preferences using binary lottery choices. First results show that (i) risk averse subjects use more risk decreasing input than risk increasing input, (ii) that the subsidy has a positive impact on the use of both types of inputs, and (ii) that crop insurance appears to be a complement to the use of risk decreasing input and a substitute to the use of risk increasing input.
Presenter: Eva Poen, University of Exeter Title: The challenges of pooling experimental data for analysis: a case for a curated repository Abstract: Meta-analysis (where the results of several studies are combined) is a very attractive tool for researchers. Analysing existing data in this way means an increase in statistical power and more reliable estimates of treatment effect sizes. Meta-analyses can also help in uncovering effects related to individual characteristics of participants, which were recorded as part of the original experiments but were not part of the experimental design. At present, a researcher wishing to conduct such an analysis is faced with multiple obstacles: First, she needs gain access to data and obtain permission to use it (including the meta-data for each session). The decision which data to include is the first big hurdle; lack of availability or access to data can lead to seriously misleading results (similar to the effects of sample selection bias). Next, she needs to understand exactly how comparable the different data sources are, how every variable is coded in order to decide if pooling the data is an option, and which statistical methods will be appropriate. Following all this is the laborious task of re-coding the data to a common standard, followed by the analysis itself. Based on the example of public goods data, we will show the difficulties associated with meta-analysis in the absence of a repository, and how the existence of a curated repository for experimental data could help address many of the problems, paving the way for more high quality meta-analyses using data from economic experiments.
Presenter: Elisabet Rutström, Örebro University and Georgia State University Title: Experiences from past attempts at building sharing sites: ExLab Abstract: ExLab was built in 1993 at the University of South Carolina and moved to the University of Central Florida in 2004. It was closed in 2010. It housed a software for recruiting participants that serviced several labs. It also housed a repository for data and working papers. The talk will share my experiences with building and managing ExLab and the ideas that guided my efforts.
Presenter: Joachim Weimann, Otto-von-Guericke-University Magdeburg Title: x-econ.org as part of x-hub: a repository specialized for experimental economics Abstract: x-econ.org is a new repository hosted by Leibniz-Institute GESIS Data Archive in cooperation with the Universities of Magdeburg, Passau and Vienna. The x-econ repository aims to meet the needs of researchers in the field of experimental economics to have a convenient and effortless way to upload and archive their data. In addition, by assigning a DOI the dataset can be cited and searched. The description of experimental datasets and associated files is made possible by metadata tailored to the needs of experimental economists including the specification of game, topic, or design and the description of the participant pool. x-econ is part of the DFG funded x-hub project which addresses experimental data in economics and social sciences. x-hub aims to establish an infrastructure that fosters easy findability, transparency and reuse of experimental data also across disciplinary borders. Diverging methods and paradigms applicable to primary data sets shall become visible at first sight, such that experimental researchers from economics, sociology and political sciences can find and understand the data and research results under their discipline-specific perspective.
Presenter: Mehmet Yigit Gurdal, Bogazici University Title: Cheap Talk with Two Senders Abstract: Asymmetric information and its consequences are ubiquitous in economic and social relationships, which give rise to limited transmission of information, causing inefficiencies. In this paper, we consider a 3-player game with 2 senders and 1 receiver. Each sender is randomly assigned a specific quality which is known to both senders, but not to the receiver. After a round of communication, the receiver chooses one of the two senders as the winner. The receiver's payoff is the chosen sender's quality, while a sender's payoff is set to 50 if chosen, and 0 otherwise. Thus the receiver would prefer to choose the higher-quality sender, while both the higher- and the lower-quality sender just want to be chosen. We have three treatments such that communication in each consecutive treatment is richer than the one before. In T1 ("Byte"), each sender can only send an "own quality" message, which is essentially a number chosen from the set of possible quality levels (0, 5, 10, 50). In T2 ("Rich"), each sender can send one free-form message up to 500 characters. In T3 ("Chat"), however, the receiver can chat - send and receive messages - with each sender separately up to 4 minutes. Our results show that receivers do slightly better than chance in picking senders, yet with varying degrees of success across treatments. This success rate is also highly sensitive to quality differences across senders and high quality difference doesn't necessarily yield high success rates for receivers. Truth-telling and informative messages are more common when the communication is richer, yet, senders often distort their private information and distortion behavior systematically differs when lying about self vs. others.
Presenter: Gerhard Riener, Heinrich-Heine-University Düsseldorf Title: Ambiguous communication: an experimental analysis Abstract: We test experimentally potential benefits of ambiguous communication. Specifically, we interpret ambiguous communication as messaging rules that depend on Ellsbergian randomization, were the message sent depends not only about the true state of the world but also about independent draw from an Ellsberg urn, an urn of unknown composition. We find that indeed most participants are sensitive to Ellsbergian randomization in this context, and this is the case most markedly for ambiguity averse individuals. The observed reaction of receivers to Ellsbergian strategies is such that it would allow a strategic sender to communicate more effciently. Additionally, we also find that similar randomization based on a non-ambiguous urn appears to have a similar effect. We conjecture possible explanations and discuss robustness of these finding, for instance by providing help with the involved updating.
Presenter: Jan Hausfeld, University of Konstanz Title: Strategic Gazing: An Interactive Eye-Tracking Study Abstract: We present the first interactive eye tracking study that explores the strategic use of gaze. We analyze gaze behavior in an experiment with four simple simultaneous-move games. The game can either be a competitive (Hide & Seek) game, in which players want to be unpredictable, or a game of common interest, in which players want to be predictable. Gaze is either transmitted in real time to another subject, or it is not transmitted and therefore non-strategic. We find that subjects are able to interpret non-strategic gaze, obtaining substantially higher payoffs than subjects who did not see gaze patterns. If gaze is transmitted in real time, eye-movements become more informative in the common interest games and players predominantly succeed to coordinate on efficient outcomes. In contrast, eye-movements become less informative in the competitive game. In fact, signal jamming by randomizing gaze across or within trials makes the players’ choices completely unpredictable for the opponents.
Presenter: Sonja Zitzelsberger, Kassel University Title: Voting on the Threat of Exclusion in a Public Goods Experiment Abstract: The possibility to exclude members from a group has been shown to help cooperation in public goods experiments. In this paper, we investigate experimentally if groups decide to implement an exclusion institution when they have the choice. We consider institutional choices in finitely repeated games as well as indefinitely repeated games which maintain a "shadow of the future". We furthermore distinguish between a costless exclusion option and a costly exclusion option that, if chosen, reduces the endowment of all players. Behavior in the experiment is surprisingly similar under the finite and the indefinite time horizon, suggesting that subjects treat the finitely repeated game more like a indefinitely repeated game, except the last round. Overall, groups favored the game with exclusion option over the game without this option. The institutional cost of the exclusion option, however, reduces the support, especially at the beginning of the experiment. Contribution levels are significantly higher when the exclusion option has been implemented. Subjects who vote in favor of the exclusion option contribute more than subjects who vote against it if the exclusion option is implemented. Groups exclude on average one group member (out of five) and exclusion is exclusively targeted at the lowest contributor. The higher contribution rates among the cooperators often (but not always) compensate them for the costs of the exclusion option in form of smaller groups and institutional costs. When there is no institutional costs, cooperators benefit from the exclusion option.
Presenter: Jinkwon Lee, Sogang University Title: Would a happy society cooperate more than an angry one in the long run? An experimental study on the effect of incidental emotions on cooperation and punishment Abstract: We, in this study, experimentally investigate the effect of incidental emotions (i.e., incidental happiness vs. incidental anger) on cooperation and punishment in a repeated public good game wherein random matching is used. This study finds that contributions in the happiness treatment significantly decay with each round while those in the anger treatment are stable, resulting in significantly more contributions in the anger treatment in the long run. This happens even in the absence of significant differences in average prosocial and antisocial punishment behaviours between the treatments. This study shows that the difference in the cooperators' contribution reaction to antisocial punishment between the treatments is the main reason for this: in response to antisocial punishment, cooperators in the happiness treatment tend to reduce their contributions, while cooperators in the anger treatment do not reduce their contributions. In fact, this result is consistent with the findings of many previous studies investigating the effect of incidental emotions. Our study suggests that the effect of incidental emotions on the dynamics between punishment and cooperation may not be ignorable and that a happy society may in the long-run possibly cooperate less than an angry one in a particular task where a costly voluntary punishment is allowed. The nonnegligible impact of incidental emotions on decision-making that this study has found, also suggests that manipulating incidental emotion as well as integral emotion could be an effective policy tool.
Presenter: Astrid Dannenberg, University of Kassel Title: "Naming and shaming" of individuals and groups in a public goods experiment Abstract: Previous research has shown that revealing players' identities increases cooperation in public goods games. In this paper we experimentally investigate whether this finding holds true when decisions are made by groups rather than individuals. We find that groups contribute more than individuals in the anonymous public goods game when no information about contributors is revealed. Especially at the beginning of the game, when it is not yet clear how the other players will behave, groups are more willing to risk a high contribution. However, while we confirm the positive effect of identification on cooperation among individuals, we do not find such an effect for groups. This indicates that "naming and shaming" works for individuals but not for groups.
Presenter: Martin Vollmann, Heidelberg University Title: Under Pressure: Depression and Group Performance Abstract: Stress in the workplace puts a high burden on employees' mental and physical health. Consequently, many of them develop a severe illness. One of the most prevalent is depression. The purpose of this study is to see how the presence of a depressed person affects the healthy individuals within a group and vice versa. To reach the highest degree of control and comparability the experiment was conducted in a laboratory setting. The contribution of this study consists of two major findings. First, it could be shown that the healthy members of a group including a depressed participant compensate the underperformance of the depressive member. Therefore, no performance differences on a group level can be detected, compared with a group consisting only of healthy participants. Secondly, it can be observed that depressed group members experience a mood deterioration over time, which is not in accordance with their increased performance and the encouraging feedback from the other team members. Those acquired findings give rise to multiple further research questions concerning the interaction between workplace and depression and form the basis for potential field studies.
Presenter: Steven Bosworth, University of Reading Title: Cross-task spillovers in workplace teams: The role of incentives Abstract: Our study seeks to isolate and examine non-strategic spillovers in multi-agent teams. In particular, we study a two-task setting with a principal and two agents in which the optimal effort choice in each task is independent of the effort exerted in the other task. The principal can observe individual efforts in one task, but can only observe team effort in the other. We vary between subjects the availability of individual piece rate, rank-order tournament, team piece rate, and fixed wage contracts for the individually-observable task while holding fixed the use of a team piece rate contract for the task where only team output is observable. We find that team effort levels are surprisingly highest under the rank-order tournament contracts for individual effort, followed by individual piece rate, then team piece rate, with fixed wage contracts performing worst. We use econometric analysis corroborated with incentivized Social Value Orientations and comprehension questions collected post-game in order to identify the channels of motivational crowding, path dependence, anchoring, and learning as driving factors of the spillovers.
Presenter: Simone Haeckl, WU Vienna Title: Teams Don't Work Well for the Motivated Abstract: What type of workers respond to team incentives and peer pressure? To address this question, we develop a novel technique to measure a worker's motivation to work hard as a deviation from optimal behavior in a real effort experiment. While we find that average output increases in response to team incentives and peer pressure, we find that highly motivated workers do not respond. The reason is that the highly motivated already work hard, and increasing effort even further is very costly to them. Our research sheds new light on sources of worker heterogeneity by providing a clean measure of work motivation and by showing that it moderates team effects.
Presenter: Federica D'Isanto, University of Birmingham Title: Homo Oeconomicus at the Caffe: A Field Experiment on "Suspended Coffee" Abstract: Individuals engage in daily behaviours that are often at issue with the self-interest and rational model of homo oeconomicus. The paper is the first of its kind to showcase the findings of a field experiment - that is a real life experiment carried out outside the laboratory setting - conducted on the behaviour of suspended coffee consumers. The suspended coffee tradition was born in Naples by people purchasing two coffees, one to drink on the spot and one to be left "suspended" for someone else to enjoy for free. A total of 230 people (113 men and 117 women) took part in our study and more than one third of the sample decided to buy a suspended coffee. This results shows that, amongst the extensive stream of evidence rejecting the model of homo oeconomicus, the suspended coffee stands out as a prime example of human beings capacity to act pro-socially and altruistically. The study also focused on analysing the main characteristics and motivations in terms of demographics, consumer choices, and adherence to social norms, underlying such a distinguishing act of generosity. Results from a latent Path Analysis carried out within the framework of Structural Equation Modeling, which is a statistical set of regression-based techniques that takes into account the complex relationships between both observed and latent variables reveal that the knowledge of the suspended coffee tradition has a direct effect on the purchase of this product and also it mediates the effect of variables such as social norms, cafe frequented, nationality, and age of the respondents. These results challenge the assumptions underlying the homo oeconomicus paradigm and suggest that, contrary to the mainstream economic theory, individual preferences are modifiable and are constantly affected by environmental factors of the surrounding context. Keywords: suspended coffee; homo oeconimicus; prosocial behaviour; consumer behaviour; altruism; generosity.
Presenter: Igor Asanov, University of Kassel Title: Folktale Narratives and Economic Behavior Abstract: Narratives - stories - prevail in social life but little is known about their relation to economic behavior (Akerlof and Snower, 2016; Hoff and Stiglitz, 2016; Schiller, 2017). We shed light on this issue by studying the association between folktales motifs and behavior in economic experiments across the world. We use a mythology/folklore database that contains information about the geographical distribution of more than 1500 motifs across the world (Berezkin, 2014). We match information about the presence of a motif in a country with (1) individual choices in standard dictator game from the comprehensive meta-study by Engel (2011 and (2) individual behavior in the "die-in-a-cup" task (deception game) across societies (Gächter and Schulz, 2016). To study the association between motifs and decision in the games in case of large-scale hypothesis testing, we use random forest algorithm (Breiman, 2001). In two distinct cases and among a large number of motifs (ranging from the sun to hungry fingers), narratives not only strongly correlate with in-game behavior but describe it. We find the strongest positive association of giving behavior in standard dictator game and the "Girl-helper" motif. In this narrative, to help a young man get an object from a remote place, the girl tells him to dismember her, and then collect her parts. In the case of the deception game, we find that "Brides for first men/Artificial wife" motif has the highest positive association with deceptive behavior. This motif depicts the situation where the main character transforms animals or pretends that animals are girls and that they can be married. However, after marriage, the truth about the wives' animal nature is revealed. Finally, to assess the causal link between motifs and economic outcomes we use genetic distance (Spolaore and Wacziarg, 2009) as an instrument for motif distance - difference between countries in the motifs repertoire.
Presenter: Jana Vyrastekova, Radboud University Nijmegen Title: Professional norms as incentives: experiments with professionals and students Abstract: Do professional norms affect behavior and even override monetary incentives? We run incentivized experiments and provide evidence that this is the case. Purchasing professionals make decisions in an incentivized economic experiment that favor the passive recipients, Internal customers, more when the decision situation is framed to appeal to the professional norm of the Purchasing professionals than when making the same decision in the absence of the framing. Professionals sacrifice more money for the passive receiver when it is described as offering higher quality for the internal customer. As a robustness check, we find that the decision of student subjects is not affected by such framing. We also find that the length of the exposure to the profession explains the impact of the framing. The novices to the profession are not affected significantly, in contrast to the professionals with longer professional life. This is consistent with internalization of professional norms to be a long-term process.
Presenter: Carlos Cueva, University of Alicante Title: Optimistic and Stubborn: An Experimental Analysis of the Disposition Effect Abstract: The disposition effect (DE) is a common bias by which investors tend to sell profitable assets too soon and hold losing assets too long. We investigate psychological correlates of the DE in a standard experimental environment and find that subjects scoring high in optimism and stubbornness (difficulty recognizing one's errors) tend to sell fewer losers and consequently exhibit a larger DE. We then test a beliefs-based explanation for these findings by conducting two novel treatments, called allocate (A) and choose (CH), in which we elicit price expectations after subjects are either randomly allocated, or forced to choose some assets, respectively. We find that subjects' beliefs are more optimistic about assets that they own than about other assets, regardless of the treatment. However, while beliefs in A about owned assets adjust in line with Bayesian inference, beliefs in CH remain stubbornly optimistic even when an asset is more likely to be a loser.
Presenter: Alexander Vostroknutov, University of Trento Title: Dynamic Regret Avoidance Abstract: In a stock market experiment we examine how regret avoidance influences the decision to sell an asset while its price changes over time. Participants know beforehand whether they will observe the future prices after they sell the asset or not. The estimation of a structural model shows that without future prices participants are affected only by past regret, but when future prices are available, they also avoid future regret. Moreover, as the relative sizes of past and future regret change with the price, participants dynamically switch between them, with the larger regret term dominating the decision to sell.
Presenter: Thomas Alexander Stephens, WU Vienna Title: Stock Holdings and Money Illusion Abstract: We investigate the relationship between money illusion and investment in stocks. We elicit money illusion, risk aversion, and other relevant measures from a quasi-representative sample of adult Danes, and match these measures with asset allocation data and controls from official registers. We find that the elicited money illusion and risk aversion measures are strongly related to portfolio composition. Among stockholders, an increase of one standard deviation in measured money illusion or risk aversion is associated, respectively, with a reduction in the share of financial assets allocated to stocks of about 30 and 35 percent.
Presenter: Ayse G. Mermer, University of Manchester Title: Cooperation in Indefinitely Repeated Games of Strategic Complements and Substitutes Abstract: In this paper we experimentally study the effect of strategic substitutability versus strategic complementarity on cooperative behavior in indefinitely repeated two-player games. We allow for learning by letting subjects play several repeated games and randomly re-match them after each repeated game. We find that behavior in the experiment is driven by two countervailing forces. On the one hand, the percentage of joint-payoff maximizing choices is significantly higher under strategic substitutes than under strategic complements. We argue that this difference is driven by the fact that it is less risky to cooperate under substitutes than under complements. On the other hand, we find that for those subjects who do not succeed in reaching full cooperation with their partner or do not attempt to reach it, the average choice is lower - less cooperative - under substitutes than under complements. We relate the latter result to non-equilibrium forces stemming from differences in the slope of the best-response curve between substitutes and complements.
Presenter: Hugh Sibly, University of Tasmania Title: Cooperation and Turn Taking in Finitely-Repeated Prisoners' Dilemmas: An experimental analysis Abstract: In the standard prisoners' dilemma (SPD) cooperation is efficient. In the repeated, modified prisoners' dilemma (MPD) turn taking is efficient. It might be expected that the cooperative preferences exhibited by some agents, which gives rise to cooperation in the finitely repeated SPD, also give rise to turn taking in the finitely repeated MPD. We consider this question theoretically and experimentally. We find experimentally that 50% of participant pairs undertake turn-taking in the finitely repeated MPD. Turn taking continues to the final round for 37.5% of participant pairs, indicating that at least 18% of participant pairs exhibit cooperative preferences. Further, the prevalence of turn taking in the MPD is less than cooperation in the SPD, a finding that is consistent with theoretical predictions.
Presenter: Elisabeth Gsottbauer, University of Innsbruck Title: Peer-Monitoring and Cooperation in Repeated Social Dilemmas: Experimental Evidence on a Folk Theorem Abstract: Voluntary cooperation is an important prerequisite for realizing the gains from cooperating in the many circumstances in which contractibility is limited. One important limitation in such situations is the perfect observability of others actions, which in many real-life interactions is far from guaranteed. We present a laboratory experiment designed to investigate, under controlled conditions, how such limited observability of actions and the availability of a monitoring technology, i.e. costly or costless action to learn about co-players'; actions, impact on cooperation and efficiency in indefinitely repeated social dilemmas. A recent Folk Theorem for games with costly monitoring gives reason to expect a positive relationship between information constraints and cooperation levels. We contrast a baseline condition with automatic and perfect feedback on co-player's actions with three conditions in which monitoring requires a deliberate action at zero, low and high cost, respectively. Counter-intuitively, but consistent with the theorem we find that cooperation and efficiency is indeed quite robust to the level of monitoring costs. Specifically, subjects behave consistent with the strategies used to construct the equilibrium in that under high monitoring costs they perform very infrequent inspections, yet in a way to sustain cooperation rates not significantly smaller than under low monitoring costs. The results provide new insights with respect to the puzzle of how people manage to cooperate in very opaque settings.
Presenter: Arno Apffelstaedt, University of Hamburg Title: Corrupted Votes and Rule Compliance Abstract: This paper investigates how undermining democratic voting procedures affects redistribution choices of individuals in a society. We use an online experiment to establish a causal effect of electoral malpractice in a referendum on compliance with rules that were implemented by this referendum. The rules are codes of conduct that ask people to share or not to share their income with unlucky agents in modified dictator games. Treatments vary between subjects whether a rule has been selected by a democratic procedure to aggregate votes (i.e., by majority vote), or by a corrupt voting procedure where the majority vote is undermined either by the possibility to accept bribes, by introducing a fee for voting, or by excluding poor voters. The results show that a democratic majority vote has the power to convince people to follow the selected rule irrespective of their own initial preferences. We rely on an instrumental variable to isolate the causal effect of beliefs about others'; behavior on individuals'; compliance decisions and find that the decision to follow a prosocial rule (prescribing to redistribute one's income) is driven by intrinsic motivation rather than strategic concerns. Corruption has a strong and significant adverse effect on individuals'; intrinsic motivation to follow a prosocial rule. In particular, a subject is less likely to follow the rule when corruption prevents her from casting her vote or when it leads her to believe that the outcome of the referendum will be biased. This results in overall lower compliance levels, thus eroding the power of the democratic vote to unite people on the question whether to redistribute. Taking advantage of the heterogeneous online subject pool, we conduct a post-experimental questionnaire to relate behavior in the experiment to real-world political attitudes and characteristics. We find that the detrimental effects of corruption are largely driven by participants of Western nationalities, i.e., those participants who live in relatively stable and well-functioning democracies.
Presenter: Dmitri Bershadskyy, Leibniz Institute for Economic Research Halle Title: The Endogenous Formation of Institutions in Public Goods Provision Abstract: In a laboratory experiment we analyze the second-order incentive problem that may arise when individuals decentrally build up an institution that helps to overcome a typical public good problem (first-order incentive problem). A focus of the paper is on behavioral effects that stem from economic education. In the multi-stage experiment individuals firstly learn about the strong positive impact of an institution (pre-play communication device) on cooperation rates compared to a simple public good game without any further arrangements. After a randomization of the group composition individuals subsequently decide on the building-up of the costly communication platform based on their experience. We confirm the finding of the literature that economists tend to free ride more intensively in a simple public good game than non-economists. However, we make clear that first-order results cannot be simply applied for second-order incentive problems. When it comes to the decentralized building-up of efficiency-enhancing institution, there is no significant difference between economists and non-economists. Instead, the individual efficiency gain induced by pre-play communication and the time taken to decide are the driving forces for the contributions to the institution. JEL Codes C91; C92, H41
Presenter: Christian Koch, New York University Abu Dhabi Title: The Ghost of Institutions Past: An Experiment on Institutional Reform and Tax Compliance Abstract: We study how past institutions affect contemporaneous tax compliance using a novel experiment in which both full and no compliance are equilibria. The treatments vary (i) the extent to which tax revenue aimed for public good provision is embezzled by a politician, and (ii) the history of embezzlement. We find evidence that tax compliance is asymmetrically path-dependent. A history of high embezzlement depresses compliance even after institutions have been exogenously improved. A history of low embezzlement, resulting in high compliance levels originally, however, cannot prevent tax compliance from waning after institutions have been exogenously deteriorated. We conduct an analysis based on individual-level heterogeneity which can account for path dependence and its asymmetry and show that a society-wide poll could assist in overcoming the "ghost" of past institutions. Our analysis therefore suggests institutional reforms are necessary but not sufficient for combating widespread tax evasion.
Presenter: Alasdair Brown, University of East Anglia Title: Anchoring and Manipulation in Speculative Markets: A Field Experiment Abstract: A common finding in laboratory studies is that subjects anchor on irrelevant initial cues when valuing assets. We run a field experiment to examine whether this heuristic can be exploited to manipulate prices in real markets. We provide early quotes in a series of horse race betting markets, and randomly vary whether these quotes are high or low. We find that subsequent prices are indeed distorted in the direction of our anchor, and that the effect spills over into correlated markets. Furthermore, we show that there are positive returns to some market manipulation strategies predicated on exploiting the anchoring heuristic.
Presenter: Kenan Kalayci, University of Queensland Title: Algorithmic third-party advice in markets with complex goods Abstract: We conduct a duopoly market experiment in which evaluating the benefits of a particular good requires complex arithmetic calculations by the buyers. We study the effects of the availability of cheap and unbiased algorithmic (robot) recommendations in this market. We employ a 2 x 2 between-subjects design with simple vs complex goods and with robot advice vs. no robot advice available to the buyers. Our results show that complexity leads to suboptimal buyer choices and lower surplus for buyers when no robot advice is available. Availability of robot advice does not affect buyer surplus when the goods are complex but leads to even lower buyer surplus, and hence benefits the sellers, when goods are simple.
Presenter: Tobias Aufenanger, University of Erlangen-Nuremberg Title: Machine Learning to Improve Experimental Design Abstract: This paper proposes a way of using covariate information in the design of experiment. In particular, this paper suggest a way of defining strata based on pretest data. The idea is to use machine learning algorithms to build a predictive model on the pretest data and use this model to stratify on predicted outcomes in the main experiment. This approach reduces much of the arbitrariness involved in defining strata directly on the basis of covariates. A simulation based on 7 different data sets shows that this algorithm is extremely effective in increasing power compared to random allocation and to traditional ways of stratification.
Presenter: Sebastian O. Schneider, University of Göttingen Title: A New Approach to Treatment Assignment for One and Multiple Treatment Groups Abstract: We present a new approach to treatment assignment in (field) experiments for the case of one or multiple treatment groups. This approach - which we call minimizing MSE approach - uses sample characteristics to obtain balanced treatment groups. Compared to other methods, the min MSE procedure is attrition tolerant, more flexible and very fast, it can conveniently be implemented and balances different moments of the distribution of the treatment groups. Additionally, it has a clear theoretical foundation, which bases on the idea by Kasy (2016), but works without any parameter to be specified by the researcher and is extended to multiple treatments. The information used for treatment assignment can be multivariate and continuous and is not limited to only a few variables. In this paper, we theoretically derive the underlying selection criteria which we then apply to various simulated treatment effect scenarios and datasets, comparing it to established approaches. Our proposed method performs superior or comparable to competing approaches such as matching in most measures of balance commonly used. We provide software to apply the min MSE approach as an ado-package for Stata.
CANCELED:
Presenter: Olaf Bock, University of Hamburg, Title: Integrating Eye-Tracking into Economic Experiment Settings
Presenter: Johannes Leutgeb, WZB Social Science Center Berlin Title: Games Played Through Agents in the Laboratory - A Test of Prat & Rustichini's Model Abstract: From the regulation of sports to lawmaking in parliament, in many situations one group of people ("agents") make decisions that affect payoffs of others ("principals") who may offer action-contingent transfers in order to sway the agents' decisions. Prat and Rustichini (2003) characterize pure-strategy equilibria of such Games Played Through Agents. Specifically, they predict the equilibrium outcome in pure strategies to be efficient. We test the theory in a series of experimental treatments with human principals and computerized agents. The theory predicts remarkably well which actions and outcomes are implemented but subjects' transfer offers deviate systematically from equilibrium. We show how quantal response equilibrium accounts for the deviations and test its predictions out of sample. Our results show that quantal response equilibrium is particularly well suited for explaining behavior in such games.
Presenter: Yadi Yang, Tilburg University Title: Oligopoly with strategic delegation Abstract: In a two-stage oligopoly game with delegation, where the profit-maximizing principals first set incentive schemes for the agents as a weighted sum of his own principal's profit and the rival principal's profit and then the agents compete in strategic complements or strategic substitutes, the principals may choose different incentive schemes for their agents at their own best interest, depending on the strategic feature of the game. We present a model with predictions that when decision variables are strategic complements, principals choose more cooperative weights, while they choose more competitive weights with strategic substitutes. A lab experiment only finds that principals choose more competitive incentive schemes for their agents in the case with strategic substitutes, while principals do not set more cooperative incentive schemes in the case with strategic complements.
Presenter: João Ferreira, Aix-Marseille University Title: On the Roots of the Intrinsic Value of Decision Rights: Evidence from France and Japan Abstract: Bartling et al. (2014, ECMA) found that Swiss individuals attach an economically meaningful intrinsic value to make a decision by themselves rather than delegating it to another person. We refine their analysis in order to disentangle how much of such value stems from (i) a preference for independence from others, (ii) a desire for power, or (iii) other motives such as a preference for self-reliance, and conduct a cross-cultural comparison between France and Japan. Our findings suggest that (i) Japanese and French individuals intrinsically value decision rights beyond their instrumental benefit, that (ii) self-reliance is the main rationale behind this intrinsic value in both France and Japan, that (iii) independence is differently valued in the two countries, and that (iv) power is not a motivation in neither of the countries. These results bring new insights into the roots of the preference for being in control, which can be relevant for institutional design.
Presenter: Alexander Coutts, New York University Title: The Impact of Relative Performance Feedback on Student Achievement: Experimental Evidence from the Lab and Field Abstract: Both within and across educational institutions there is substantial variation in the type of feedback that students receive regarding their performance. Media attention has highlighted examples where parents and students have protested policies of public posting of grades. Yet little is known about the effects of such feedback policies. To answer this question we combine a lab and a field experiment. The lab experiment is designed to identify and disentangle the effects of private versus public feedback, and separately identify the effects of past feedback and the effects of the anticipation of feedback. In a field experiment in the classroom, we study the effects of public versus private feedback on the formation of study networks, as well as quiz and exam performance. Our results show that providing students with information about their relative performance significantly increases future effort if this information is provided publicly. This effect is significantly different for female and male students. Whereas male students perform worse, female students increase performance under the public feedback scheme. The announcement of a public feedback rule further seems to benefit mostly the top performing students of a group, as they increase both effort and accuracy. On the contrary, after receiving feedback, low performing students show significantly worse results after having received either private or public feedback.
Presenter: Marina Chugunova, University of Hamburg Title: Redistribution and Production with the Subsistence Income Constraint: a Real Effort Experiment. Abstract: A large body of literature demonstrates that redistribution leads to inefficiencies: taxation distorts incentives in favor of leisure activities and against effort provision. Yet, these results are obtained under the assumption that, firstly, people are absolutely free in their labor-leisure allocation decisions and, secondly, that redistribution is nothing but a wage cut, i.e. nobody benefits from the tax revenues in any way. We challenge these assumptions and study labor supply decisions in a richer framework. We conduct a real effort experiment with a subsistence level of income and functioning redistribution system.On the side of the taxpayers, we find that even under a very low subsistence threshold, taxpayers increase their effort in the redistributive treatments with a slight additional increase if tax returns are transferred to actual recipients, and not wasted unproductively. As for recipients, the transfer significantly boosts their productivity and spurs the overall efficiency leading to a self-sorting of recipients' performance in the real effort task according to their skills.
Presenter: Marco Faravelli, University of Queensland Title: Feedback, Learning, and Performance: Evidence from a Field Experiment Abstract: We conduct a large scale field experiment to investigate the impact of feedback on students' performance in a research intensive selective university. We provide students with information on the relative ranking in a semester long computerized assignment. We run five treatments varying the timing and nature of the feedback. We find evidence that feedback is more effective if it is provided only when relative position is affected. In particular, we find that the provision of this type of feedback led to an increase of around 4% in students' grades. We also find that the effect was significant for the whole distribution and both types of exams; midterms and final. Feedback biased toward positive or negative variations in ranking is less effective.
Presenter: Florian Engl, University of Cologne Title: Coordinated Unethical Behavior Abstract: We study unethical behavior in groups when the implementation of an individually beneficial but morally questionable outcome requires coordination between group members. Specifically, we compare unethical behavior in small and large groups in a simple lab coordination game. Participants can choose between an unethical and an ethical option, where the unethical option is a donation to a negatively perceived charity. The donation and high individual payoffs are implemented only if all group members decide to donate. We hypothesize less unethical behavior will prevail in large groups because large groups are more likely to entail an "ethical type" who is willing to forego private benefits to prevent unethical outcomes. We find that, indeed, fewer unethical outcomes result in large groups, since more large groups than small groups entail an ethical type. We relate our findings to those of several other experiments, which suggest the opposite relationship between group size and unethical behavior, and show causal responsibility theory can reconcile the divergence.
Presenter: Kei Tsutsui, University of Bath Title: A two-roles bribery experiment with unstructured bargaining Abstract: This research analyzes bribery behavior using a repeated corruption game in the laboratory. This study contributes in two ways. Firstly, we investigate if the amount of bribery is larger in a society where a person experiences both roles of bribe-giver and bribe-receiver. Secondly, we propose a new experimental design where the bribe-receiver and the bribe-giver negotiate on the amount of bribery to be transferred between them using an unstructured bargaining protocol. The results show that more bribery is paid when the bribe-receiver experienced the role of the bribe-giver in the previous round.
Presenter: Sven Fischer, Newcastle University Title: Consuming copyrighted media without paying - A controlled experiment with a representative sample Abstract: We report the results of controlled experiments on unlawful consumption of non-rivalry consumer goods such as film and music, with a sample of 1,010 participants, representative for the adult population aged 18 to 65 in the United Kingdom. Our study allows to identify how consumer behaviour is affected by considerations towards the rights holder, deterrence, intrinsic norms, and normative and empirical expectations. We are furthermore able to correlate experimental and real-world behaviour with reported norms, personality measures and socio-demographic characteristics. This main study with a representative sample is a follow up on an experiment with a student population. More specifically, the experiments replicate a situation in which a consumer can decide whether to pay a small price for the consumption of a non-rivalry consumer good (such as media) he likes very much, or obtain it without paying the seller or rights holder. The good is linked to a "rights holder" (represented by another participant), who is the recipient of the price paid, whenever the consumer decides to pay. Obtaining it without paying, on the other hand bears the risk of being detected, which results in a fine in excess of the gain from consumption. We systematically analyse the effects of several aspects on consumer behaviour in such an environment. i) We manipulate deterrence, by varying the detection probability from 0%, to 1%, 10%, or 20%. This tests standard financial risk considerations. ii) We compare consumption of goods the rights holder worked to produce, to those she herself received for free. Equity considerations would predict a higher willingness to pay from consumers in the first case. iii) We force some consumers to reflect on their own norm, that of others, and their expectations of how other consumers behave, before they make their own purchasing decision. This enables us to get an understanding of consumers' social norms and whether reflecting on them nudges behaviour. We elicit and compare perceptions and attitudes towards consumption of music to that of films, and of downloading such media versus consuming it by streaming. In a follow up survey, two months after the first one, we test whether real world consumption of either films or music, was related to the choices our participants made in our main study. Our results confirm a strong reaction to deterrence with an overweighting of small probabilities. Deservingness of the seller has only a limited, and mostly insignificant effect on behaviour. More specifically, we only observe a significant decrease in unlawful consumption due to deservingness of the seller, in treatments where we do not elicit norms and expectations before decisions are made. Overall, eliciting norms and expectations before decisions decreases unlawful consumption throughout. This effect is strongest among women and older participants, and results in significantly less unlawful consumptions among these demographics. Reported intrinsic norms (what someone should do) are highly positively correlated with decisions, but only weakly, though significantly, with normative (what you think others think one should do) and empirical expectations. Finally, behaviour in our experiment is significantly correlated with reported downloading behaviour in the real world: Participants who engaged in more unlawful consumption in our experiment, reported a significantly higher share of downloads coming from unlawful sources, especially of Music files. No such correlation could be established for streams.
Presenter: Sebastian Fehrler, University of Konstanz Title: Honesty and Self-Selection into Cheap Talk Abstract: In some situations some people lie strategically - political campaigns and sales promotions being obvious examples. However, recent empirical studies suggest that people are very heterogeneous with respect to their willingness to lie strategically for their own advantage. Hence, the level of lying and thus the credibility of statements will crucially depend on who self-selects into the cheap talk situation. We study this broad idea in the concrete setting of a two-stage political competition model and test its key predictions in the lab. At the entry stage, potential candidates compete in a contest to become their party's nominee. At the election stage, the nominated candidates campaign by making non-binding (cheap talk) promises to voters. Confirming the model's key predictions, we find in the experiment that dishonest people over-proportionally self-select into the political race; and that this adverse selection effect can be prevented if the entry stage is made transparent.
Presenter: Thanee Chaiwat, Chulalongkorn University Title: Default Option of Corruption Behavior on Bribery and Patronage: An Experiments from Online Game "the Corrupt" Abstract: Corruption is one of the big problems for many countries. It is difficult to detect individual corruption behavior in the real world. We, in collaborating with a big online-game production house of Thailand "opendream", designed visual novel game which is freely downloaded on iOS and android platforms "the CORRUPT" to study decision on bribery and patronage of Thai people. Almost 100,000 observations are collected during 3 months, both choice and time. Most of them are male, 18-30 years old and familiar with technology. Results showed that 25% of the players decided to bribe and 37% helped their friend even by wrong way. They spent longer time to pay bribe than no pay, but shorter time to help friend than ignore. So, people should help their friends as a default option, but bribery is not. However, there is no significant different behavior among gender. Conclusion indicates the different cost of anti-corruption policy between bribery and patronage problems of Thai society.
Presenter: Milos Fisar, Masaryk University Brno Title: Transparent lobbying in a lab Abstract: Transparency and lobbying, two terms that have direct influences on the behavior of policymakers, politicians, and bureaucrats. Two terms that are modeled for decades but have been barely studied. We suggest an experimental laboratory approach to observing effects of transparency on rent-seeking oriented decisions of the policy maker. In our game is a policy maker adjusts policy to existing environment and also is facing an offer from the endogenous lobbyist. The citizen observes the decisions with different levels of transparency (no, mandatory or voluntary transparency) and can punish the policy maker. The preliminary results suggest, that with the larger level of transparency, the policy makers are more likely to chose lobbed alternative if it is the one in line with the environment. Also in transparent treatment, the citizens are more likely to punish the "misbehaving" policymaker.
Presenter: Huanren Zhang, New York University Abu Dhabi Title: Egalitarianism and turn taking in repeated coordination games Abstract: In asymmetric repeated battle of the sexes, there are two repeated game strategies that are potentially focal, one leads to equal payoff (egalitarian outcome), and the other lead to equal opportunity. Players must play complex strategies to obtain the egalitarian outcome. We design an experiment to study whether subjects can discover and converge complex behavioral patterns in order to achieve egalitarian outcomes. We also gradually change the payoff structure and the degree of conflict to investigate whether there are behavioral spillovers over a class of strategies. The results show that players discover and converge to complex behavior patterns in order to achieve egalitarian outcomes. Egalitarian outcomes are more common when there is a gradual decrease in the degree of conflict compared to the case where the degree of conflict increases. Behavioral spillovers are commonly observed across games. Subjects with previous experience of turn taking or egalitarian patterns are more likely to converge to the same behavioral patterns. We also find evidence that social preferences influence the likelihood of Turn Taking and Egalitarian plays.
Presenter: Konstantinos Georgalos, Lancaster University Title: Comparing Behavioural Models Using Data from Experimental Centipede Games Abstract: The Centipede game (Rosenthal, 1982) consists one of the most well-known paradoxes of backward induction in the literature of experimental game theory. Given that deviations from the unique subgame perfect Nash predicted play usually generate a Pareto improvement, several theoretical models have been employed in order to rationalise this kind of behaviour ranging from fairness considerations (McKelvey and Palfrey, 1992), noisy choice (Quantal Response Equilibrium, Fey et al., 1996), learning (Nagel and Tang, 1998), level-k cognitive hierarchy models (Kawagoe and Takizawa) or decision theory probabilistic models (Blavatsky, 2015). Nevertheless, up to now, there is no empirical study to perform a horse-race between all these alternative models. In this paper, we use Maximum Likelihood Estimation techniques to evaluate the descriptive (in-sample fit) and predictive (out-of-sample fit) performance of the most prominent models, using data from various experimental studies.
Presenter: Peter Moffatt, University of East Anglia Title: Parametric Estimation of Social Preference Models Abstract: We analyse data on dictator game giving from an experiment in which the price of giving is varied. We use this data set to estimate a number of different models, each based on a utility function with own-payoff and other's payoff as arguments. We are particularly interested in the well-known Fehr-Schmidt utility function. A fundamental problem with this utility function is non-differentiability, which leads to the solution to the constrained optimisation problem being either one of extreme egalitarianism or extreme selfishness. We overcome this problem by introducing a parameter to the utility function that allows strict quasi-concavity, and therefore allows non-corner solutions. The Fehr-Schmidt function is a limiting case of our extended function. We estimate the extended model by MSL, allowing between-subject heterogeneity in both of the Fehr-Schmidt parameters (aversion to disadvantageous and advantageous inequality). We find substantial heterogeneity in both of these.
Presenter: Roel van Veldhuizen, WZB Social Science Center Berlin Title: Corrupt Communication Abstract: We use a laboratory experiment to investigate the role of bribes and free-form communication in distorting a referee's judgment. Participants in the role of workers take part in a real effort task and are entitled to a bonus if they meet a certain performance threshold. Participants in the role of referee are tasked with judging the worker's performance and decide whether to award the bonus. We vary whether workers are able to bribe and/or engage in free-form communication to persuade the referee. We find that bribes and communication are both distortionary, but in different ways: whereas communication increases the overall number of winners, bribes change the type of worker who wins. Specifically, bribes weaken the relationship between performance and winning, with bonuses mostly being awarded to workers who are willing to bribe. We discuss implications for anti-corruption interventions.
Presenter: Ivan Soraperra, University of Amsterdam Title: Too Lucky to be True. Fairness views under the shadow of cheating Abstract: The steady increase in inequality over the past decades has revived a lively debate about what can be considered a fair distribution of income. Public support for redistribution and its extent typically depends on the perceived causes of income inequality, such as differences in effort, luck or opportunities. We study how fairness views and the extent of redistribution are affected by a hitherto overlooked, but relevant factor: immoral self-serving behavior that can lead to increased inequality. We focus on situations in which the rich have potentially acquired their fortunes by means of cheating. In an experiment, we let third parties redistribute resources between two stakeholders who could earn money either by choosing a safe amount or by engaging in a risky, but potentially more profitable, investment. In one treatment, the outcome of the risky investment is determined by a random move, while in another treatment stakeholders can cheat to obtain the more profitable outcome. Although third parties cannot verify cheating, we find that the mere suspicion of cheating changes fairness views of third parties considerably and leads to a strong polarization. When cheating opportunities are present, the share of subjects redistributing money from rich to poor stakeholders triples and becomes as large as the fraction of libertarian subjects who never redistribute. Without cheating opportunities, libertarian fairness views dominate, while egalitarian views are much less prevalent. These results indicate that fairness views and attitudes towards redistribution change significantly when people believe that income inequality is the result of cheating by the rich.
Presenter: Elina Khachatryan, University of Kassel Title: Competitive Pressure to Cheat: Experimental Evidence of Individuals and Groups Abstract: We investigate experimentally if individuals and groups cheat more when they are in competition with rivals compared to a noncompetitive environment. Our results show that individuals do not significantly change their behavior when they compete against other individuals or groups. Groups generally cheat more than individuals. Furthermore, they increase the degree of cheating in competition but only when they compete against other groups and not when they compete against individuals. Keywords: lying; honesty; competition; individual behavior; group behavior
Presenter: Aaron Kamm, New York University Abu Dhabi Title: What Causes Gamson's Law? Experimental Evidence on Coalitional Bargaining and Commitment Abstract: One of the most robust empirical findings on coalition governments is that government portfolios are allocated proportionally to the seat share of the parties forming the government. This empirical observation (known as Gamson's Law) is at odds with the economic theory on bargaining and previous experimental evidence. We find strong support for the theoretical predictions which suggests that Gamson's Law might be caused by parties first forming the coalition and only in a second step bargaining over portfolios.
Presenter: Matt Wiser, University of South Alabama Title: Electoral College: A Multibattle Contest with Complementarities Abstract: The Electoral College system used to elect the President of the United States is an example of a multibattle contest with complementarities. Given the differences in the values of states comprising the Electoral College, many different combinations of state victories can lead to an overall victory, which impacts contestant strategy. This paper develops a theoretical model of such a multibattle contest with complementarities, and tests its predictions in the lab for a simplified four state Electoral College map. To focus on purely on the strategic aspects of the Electoral College game, in the lab we use three different Electoral College maps while always keeping the total number of delegates the same. Our experimental results show that subjects behave quite differently from the theoretical predictions. They seem to be using a mental accounting rule according to which they choose to spend half the value of the prize as their budget in the lottery success function. Then, depending on the map, they follow different bidding patterns the most predominant pattern being one that implicitly takes into account the marginal importance of each state for winning the election.
Presenter: James Tremewan, University of Vienna Title: An experiment in (almost) unstructured multilateral bargaining Abstract: We conduct a three-person bargaining experiment in which the timing of proposals and votes is unrestricted. Players interact in real time, and players may propose divisions or agree to others' proposals at any point in time. The game ends when a sufficient number of players agree on a division. The available surplus diminishes over time, so that delay in reaching agreement is costly. We compare behavior under majority and unanimity rule. In contrast to earlier work we find that those who make proposals often demand a smaller rather than larger share of the surplus. Other than that, our results are very similar to those obtained under more structured experimental designs such as those which implement a Baron-Ferejohn protocol. This suggests that the main results of prior experiments are robust to less stringent implementation of procedural assumptions.
Presenter: Marco Mantovani, University of Milan - Bicocca Title: Beliefs, risk aversion and herding in experimental prediction markets Abstract: Prediction markets promise a mechanism for aggregating dispersed information that provides proper incentives while being virtually costless. Nowadays, they are commonly used to forecast the probability of a wide range of outcomes: elections, sport events, innovations etc. Predictions markets have grown and diffused in the absence of a solid empirical validation of the underlying theory. The question of whether we can build and use markets to accurately predict uncertain events remains unanswered. We provide the first experimental test of prediction markets where all the elements that are needed to assess the accuracy of market prices as well as the sources of inaccuracies are either induced by the design or controlled for. In particular, our design allows us to answer to the following questions: (a) Are market prices accurate in predicting uncertain events? (b) If not, are inaccuracies explained by risk aversion as predicted by the theory? Do incorrect beliefs add to the inaccuracy of the observed prices? (c) Does the market activity make traders beliefs more or less accurate? Does this feedback effect contribute to price inaccuracy? Our preliminary results show that prediction markets provide relatively accurate predictions on aggregate. Inaccuracies, which are substantial in many markets, cannot be explained by incorrect beliefs and risk preferences. On the other hand market activity substantially distorts beliefs, which we interpret as an instance of herding. As a consequence, the average individual holds more inaccurate beliefs at the end than at the beginning of the market.
Presenter: Daniel Grosshans, University of Zurich Title: How Investment Performance Affects the Formation and Use of Beliefs Abstract: Research Questions: In this study we use a novel experimental design to address two unresolved aspects of the trading behavior of individuals: First, we analyze if investors maintain overly optimistic beliefs in the face of losses. And second, we test whether buying is more forward-looking than selling. Experimental Design We invited 63 advanced economics students to participate in our experiment. Each participant had to make investment decisions for an artificial stock over 40 rounds. The stock price dynamics were driven by a two-state Markov process that created positive (negative) short-term momentum of stock prices in the good (bad) state. In every round participants had to decide whether they want to buy or sell the stock, or not trade at all. In addition, participants estimated the probability of a price increase to the next round. The final earnings were based on the trading performance and the accuracy of the probability estimates. Main Results Beliefs in the loss domain are significantly more optimistic than in the gain domain when controlling for the objective Bayesian benchmark. This finding is in line with a belief channel that helps to alleviate cognitive dissonances of investors who are facing a paper loss. In addition, we find that selling is less sensitive to beliefs than buying. We further show that this difference is driven by selling decisions in face of a paper loss. We therefore conclude that selling at a loss but not selling as such is less forward-looking than buying.
Presenter: Doron Sonsino, The College of Management Title: The conflicting links between forecast confidence and trading propensity Abstract: Confidence in personal forecasts may act as motivator of trading. If two comparable investors, for example, agree that the excepted return on some stock is 10% but investor A assigns 90% probability to an interval of plus or minus 5% around this point estimate, while investor B assigns only 30% probability to the same interval, A should reveal stronger willingness to purchase the stock compared to B. The current study, however, shows that at the more general trait-level, personal tendency for forecast-confidence associates with smaller inclination to trade stocks. Individual tendency for forecast confidence was estimated using 8 distinct field-based return forecasting assignments, while personal trading inclination was measured using a 6 items survey questionnaire. By asking subjects to decide on the amount they choose to buy or sell from 4 of the prediction targets, we parallely illustrate that forecast-confidence indeed boosts trading at the case-specific level in spite of the broader negative confidence-trading link. Simultaneous estimations of the confidence and trading variables show that the trait level correlation follows from more general personality traits, while confidence in profitability plays direct, causal role in case-specific purchase and sell decisions.
Presenter: Fanny Brun, University of Zurich Title: Immoral labor markets Abstract: Corporate scandals e.g., in the banking industry - regularly bring to light the existence of immoral practices conducted by professionals. Moreover, some jobs e.g., marketing tobacco products or predatory lending likely involve inherently immoral acts by their nature. Conventional wisdom posits that workers performing these kinds of jobs are compensated for their unethical conduct - a form of compensating differential for the aversive nature of acting unethically. However, little empirical evidence exists regarding heterogeneity in willingness to perform immoral jobs and the resulting differential labor market outcomes. We attempt to provide such evidence. In a first step, we document, using descriptive statistics for Switzerland, that individuals working in immoral industries receive a wage premium, confirming patterns observed in previous research that suggest workers are financially compensated for performing an immoral job. However, as with previous research, there remains the possibility that unobserved worker and job characteristics unrelated to morality account for the relationship. Therefore, as a second step, we study a laboratory labor market in which we exogenously vary the moral nature of the work performed, while holding constant all other aspects of the job and labor markets. We also measure, separately from market behavior, an individual-level willingness to act immorally. By comparing markets with varying immorality of work tasks, we can test for the existence of an immorality premium and other labor market outcomes. We observe labor-market outcomes that are consistent with the predictions of a simple model with heterogeneity in moral worker types and moral job characteristics. For example, we find that those who exhibit less distaste for immorality are more likely to be employed and that they command a wage premium in an immoral labor market. Moreover, we find that immoral types benefit the most in terms of earnings and welfare in the immoral job market, and that these gains are even higher if more moral types are present in the market. Finally, as a third step, we use responses provided by participants to several questions about moral judgments and behaviors "collected outside of the laboratory experiment" to construct a separate individual-level measure of morality. We show that this type classification is predictive both for labor-market outcomes in our experiment, as well as for expected career trajectories in real labor markets, such as the willingness to work in industries or firms with lower perceived morality. We conclude that the moral dimensions of both workers and jobs are relevant for understanding sorting behavior and outcomes in labor markets, and that workers' moral characteristics are measurable both by simple laboratory tasks and by questionnaire responses.
CANCELED:
Presenter: Linda Kamas, Santa Clara University, Title: Does Empathy Pay?
Presenter: Leonie Kühl, Karlsruhe Institute of Technology Title: What Is the Adequate Wage? The Causal Role of Gender, Greed and Own Work Experience Abstract: The increasing wage gap between CEOs and workers has been a matter of public debate and political discourse repeatedly and it remains a major topic in many societies. This paper explores in a laboratory experiment what decision-makers consider an adequate wage for workers. We investigate the casual influence of gender, experience with the workers' task, and potential trade-offs with own financial interests in decision-makers. If there is no conflict with own earnings, males tend to be more generous with wages than females. Yet as soon as own financial interests conflict, this picture changes significantly. While female decision-makers display a rather consistent choice of wages, male decision-makers become greedy in the sense that they choose wages significantly below what workers would consider adequate. This happens though decision-makers earn much more than workers, and though decision-makers know that workers have to work on a tedious task: the assembly of 100 pens and the complete disassembly of 100 pens (which takes about one hour). Experience with the task does not reduce greed in male decision-makers. Therefore, we find that the "adequate" wage strongly hinges on gender and circumstances.
Presenter: Christoph Siemroth, University of Essex Title: How much information is incorporated in financial asset prices? Experimental Evidence Abstract: Rather than testing whether financial market prices are informationally efficient or not, we derive a method to estimate how much information is incorporated in financial asset prices. Using data from an experimental asset market trading an Arrow security, the method quantifies which share of information available in the market is consistent with the observed market prices. The estimates indicate that a surprisingly small share of information is incorporated in asset prices. We confirm this finding using data from other experimental studies in the literature. We also find that a larger share of information is incorporated if at least part of it is public. Our results are inconsistent with a strong form of informational efficiency, but potentially consistent with a semi-strong form.
Presenter: Stefan Palan, University of Graz Title: Does Investor Risk Perception Drive Asset Prices? Experimental Evidence Abstract: What people perceive as risk clearly goes beyond variance. Several papers have shown that e.g. probability of loss plays a more prominent role in perceived risk than does variance. We are the first to explore how individual risk perception influences prices and trading behavior in a market setting by exposing subjects to a number of differently shaped return distributions. We elicit subjects' individual risk perception and then let them trade assets with these return distributions on a continuous double auction market. We find individual stated risk perception in line with earlier papers. In the markets we observe quite active trading and prices strongly driven by average risk perception. While standard finance theory predicts identical prices for all assets we find average prices to vary by up to 20 percent, with assets perceived as less risky trading at lower prices.
Presenter: Owen Powell, WU Vienna Title: Group size in experimental asset markets Abstract: In this study we compare asset market outcomes produced by a small sample of large markets and a large sample of small markets, where market size is determined by the number of traders. The market outcomes that we consider include overpricing and absolute mispricing in markets with inexperienced and experienced traders, as well as the effect of experience on these measures of market efficiency. We find that even though a large sample of small markets produce noisier measures, they neither stochastically dominate the measures produced by a small sample of large markets, nor are they dominated by the latter.
Presenter: Sander Onderstal, University of Amsterdam Title: Can Collusion Promote Corporate Social Responsibility? Evidence from the Lab Abstract: Recent experimental studies show that competition may erode moral values on markets. This suggests that cartel agreements among firms might promote public interest objectives. We test this intuition in a laboratory experiment. In the experiment, we explore variants of Bartling et al.'s (2015) experimental protocol. Participants playing the role of firms can choose between offering a "fair"and an "unfair" product to a consumer after which they post a price. Subsequently, the consumer can choose between the two offers. We vary between two experimental protocols modelling the externality imposed when an unfair product was traded. The first protocol is closest to Bartling et al.'s (2015) in that the "unfair" product imposes an externality on a third party in the lab. In the second protocol, when a "unfair" product is traded, we refrain from buying carbon reduction certificates on behalf of the participants. Within both protocols, we vary whether or not the firms can coordinate on the type of product they sell. Our findings under both protocols are qualitatively similar. The opportunity to coordinate on the kinds of products sold has no significant impact on the fraction of fair products offered to nor traded on the market. It does have a positive effect on the fraction of markets where only fair goods are offered. For both firms and consumers, third-party preferences are a powerful predictor for behavior. Our results have implications for competition policy.
Presenter: Daniel Chen, Toulouse School of Economics Title: The Impact of Economics on Moral Decision-Making and Legal Thought Abstract: This paper provides a quantitative analysis of the effects of economics on moral decision-making and legal thought using the universe of opinions in U.S. Circuit Courts from 1891 and 1 million District Court criminal sentencing deci- sions linked to judge identity. We use attendance in a controversial economics training program that 40% of federal judges attended by 1990, a policy change giving judges more sentencing discretion, random assignment of judges to control for court- and case-level factors, an exogenous seating network from random panel composition to trace the spread and impact of new ways of moral reasoning, and ordering of cases within Circuit to identify memetic phrases that move across legal topics. We find that judges who use law and economics language vote for and author conservative verdicts in economics cases and are more opposed to government regulation and criminal appeals. After attending Henry Manne's economics training program, judges use economics language and render conservative verdicts in economics cases and reject criminal appeals. Manne economics training is more predictive of these decisions than political party. Manne judges render 20% harsher criminal sentences after U.S. v. Booker allowed more sentenc- ing discretion. They also impact criminal appeals verdicts when not authoring the opinion. Judges exposed to Manne peers increase their use of economics language in subsequent opinions. "Deterrence", "capital", and "law and economics" move across legal topics.
Presenter: Pietro Guarnieri, University of Pisa Title: The effect of ethics meetings on risk-taking behaviour: an experiment Abstract: Many financial decisions imply taking risk for others. Relying on an experimental setting, we investigate the effect of bank ethics meetings on individual risk-taking behaviour when there are monetary conflicts of interest between the decision maker and the client. In our decision setting, participants will be either assigned to investor or client roles. Investors choose between two risky gambles, one of which always implies a better outcome for the investor but exposes the client to higher risk. In the main treatments (i.e. Ethics meeting), in contrast to the Baseline treatment, participants also discuss - before their investment decision - the consequences of their choice within a group of decision makers (i.e. peers). Our results shows that, in treatments with ethics meetings, investors tend to choose more often the gamble less risky for clients.
Presenter: Maximilian Germann, University of Mannheim Title: Trust and Delegated Investing Abstract: A recent theory by Gennaioli, Shleifer, and Vishny (2015) proposes that trust is an important component for delegated investing. As a result, investors are willing to invest more risky with trustworthy agents, but more trustworthy agents are able to charge higher fees for generic services. This paper tests the theory in a laboratory experiment. Participants first play a standard trust game. Participants then act as investors who have to make two separate delegated investment decisions. Using the amount returned in the trust game as measure of trustworthiness, we show that investors are willing to take substantially more risk when the agent is more trustworthy, even if the agent charges higher costs. The willingness to take more risk is increasing in the difference in trustworthiness of the two agents. This finding is robust to different specifications of the difference in trustworthiness.
Presenter: Menghan Xu, Xiamen University Title: Monopolistic Screening with Limited Enforcement Abstract: The paper studies a optimal mechanism design and screening problem, a la Mussa & Rosen(1978) in the presence of limited enforcement. Namely, the bank (principal) cannot prevent borrowing firm (agent) from consuming acquired fund without producing. The impediment of forming contract creates an endogenous outside option to all agent borrowers. We show that in the optimal mechanism, loan sizes for higher types are shrunk by ironing, i.e. pooling on the top. Meanwhile, the lower types enjoy production at the same levels as second best environment. Moreover, we show that firms' participation is independent of enforcement level.
Presenter: Florian Spitzer, WU Vienna Title: Product Testing in Markets for Experience Goods Abstract: We examine the effectiveness of a product testing institution in a market for experience goods where reputation formation is not feasible. In our setup, sellers commit ex ante to a quality level (high or low). Buyers can - in addition to buying immediately and leaving the market - pay a fee to get a noisy signal about the quality chosen by the matched seller. We show the existence of a mixed perfect Bayesian equilibrium in which sellers provide high quality with positive probability, while buyers mix between buying immediately and acquiring information. We vary the cost of the signal such that the mixed equilibrium exists in a low-cost treatment but not in a high-cost treatment. The experimental results show that the low-cost signal improves efficiency, but not to the extent as predicted by theory. Surprisingly, the high-cost signal - which is predicted to have no effect - also improves efficiency compared to a control treatment where no signal is available.
Neuroeconomics – where science and economics meet For years, while building their mathematical predictive models of choice, economists treated the brain as a “black box”. The economic models of choice were thus mathematically very elegant; but have been shown to systematically fail under numerous circumstances. Neuroeconomists now understand the basic architecture of the brain and how it actually makes choices. To demonstrate these advances, I will present a descriptive model of choice with normative foundations based on how the brain is thought to represent value. The model captures similar phenomena as Prospect Theory but employs fewer parameters, allows for individual heterogeneity, and retains neurobiological plausibility, as well as making a series of novel predictions. In addition, I will present recent research on the relation of neuroanatomy, age, risk-taking and choice rationality, and give some practical advice on how to go about neuroeconomics research, sharing some personal successes and pitfalls.
Behavioral Operations: Research advances and future directions Behavioral Operations Management is a relatively new field of business research that applies methods of behavioral and experimental economics to such applied business topics as revenue management and pricing, inventory management, and competitive sourcing. I will discuss some recent work that has been done on those topics, including future directions, and how behavioral and experimental economics is contributing to this work. I will also discuss advantages as well as roadblocks that are involved in interdisciplinary work.
Can We Trust Scientific Results? Replications, Reproducibility, and Prediction Markets Why are there so many false positive results in the published scientific literature? And what is the actual share in different literatures? We will discuss these questions, see what prediction markets might add to the understanding of the reproducibility of science and what we can do to increase the reproducibility. We will also discuss the hormonal literature in economics and specific power problems there.
