Last Revised 23:33:48 PST, 2012.11.13
Welcome Reception - Sonoran Rooftop, Westward Look Resort
Free margaritas, chips, and salsa (first hour).
Plenary Session, Sonoran Ballroom
Ulrike Malmendier, Connecting Experimental Work to Theory (with application to Social Preferences and Voting)
Chair: John Wooders
Conference Registration desk (917-2919), graduate student volunteer (220-9015), WWL reception (297-1151)
Plenary Session, Sonoran Ballroom
Guillaume Frechette, Infinitely Repeated Games in the Laboratory
Chair: Daniel Friedman
ABSTRACT:
We report on an experiment designed to investigate theories of portfolio choice under ambiguity. We use the same setting as that used for the basic Capital Asset Pricing Model, with one certain and two uncertain assets, and three states of the world. However, unlike in the CAPM, the three states of the world do not have well-defined probabilities; rather their occurrence is ambiguous, implemented with a Bingo Blower (in which there are balls of three colours û representing the three states û which cannot be counted). Subjects are presented with a series of problems, in each of which they are given a sum of money to be allocated between the three assets, given a dividend payoff matrix for all three assets in all three states of the world. We assume a particular preference functional (the MaxMin model of Gilboa and Schmeidler), assume that subjects' actual allocations depart from optimal allocations with a bivariate normal distribution, estimate the parameters of the model (the risk parameter, the lower bounds on the probabilities and the parameters of the error distribution), and examine how well the model fits the experimental data. Later we plan to explore different preference functionals.
Author(s): Xueqi Dong, John Hey
Topic: Ambiguity, finance
Back to session: Ambiguity 2
ABSTRACT:
In our first-price two-person partnership dissolving auction experiments, the partners with large shares bid less aggressively in an ambiguity environment than in a uniform distribution environment; the opposite holds for small-share partners. We show that this is a result of the pessimistic ambiguity attitude, or so-called ambiguity aversion. In another treatment of subjects bidding against random numbers, the large-share partners bid higher interacting with ambiguous bids than with uniformly-distributed bids: a result of optimistic ambiguity attitude. We attempt to formally characterize ambiguity attitudes by ordering possible probability measures by their 'favorableness', and make inferences based on choices. The instability of ambiguity attitudes could mean new challenges for existing ambiguity theories, which offer no guidance for the choice of ambiguity attitude parameters.
Author(s): Ning Yu, Thorsten Chmura
Topic: Ambiguity
Link: https://sites.google.com/site/neilningyu/working-papers
Back to session: Ambiguity 2
ABSTRACT:
We design and conduct an economic experiment to investigate the learning process of the agents under risk and under ambiguity. We gather data for subjects choosing between lotteries involving risky and ambiguous urns. Decisions are made in conjunction with a sequence of random draws (with replacement), allowing us to estimate the beliefs of the agents at different moments in time. For each of the urn types we estimate three models of learning: (1) Bayesian updating allowing for base rate fallacy; (2) Bayesian updating with multiple priors; and (3) Reinforcement Learning. Our findings suggest an important difference in updating behavior between risky and ambiguous environments. Specifically, when updating under ambiguity subjects significantly underweight the new signal, while when updating under risk subjects are Bayesian. Additionally, estimates of the multiple priors model suggest that when updating in ambiguous environment subjects consider only the most likely prior.
Author(s): Othon Moreno Gonzalez, Yaroslav Rosokha
Topic: Ambiguity, learning
Link: https://webspace.utexas.edu/yr238/www/docs/Learning%20Under%20Ambiguity%20-%20An%20Experiment.pdf
Back to session: Ambiguity 2
ABSTRACT:
Exponential-growth bias (EGB) is the tendency for individuals to underestimate exponential growth due to the neglect of the role of compounding. We develop a theoretical model in which EGB causes consumers to misperceive the budget constraint. Because consumers dynamically misperceive prices, consumers will over-consume early in their lifecycle if the elasticity of intertemporal substitution is sufficiently large. This over-consumption is exacerbated when income is received later in the lifecycle, as biased agents will additionally misperceive the present value of their earnings. These effects can lead to over-indebtedness both for early consumption and for over-investment in human capital. Finally, we show that EGB leads to a form of dynamic inconsistency in which planned future consumption increases when the agent's asset balance is positive and decreases when the balance is negative. This inconsistency can lead to significant welfare losses if the consumers are able to commit lower bounds on future consumption, such as consumption commitments in housing û commitment which other forms of dynamic inconsistency, such as present-biased preferences, would predict to be welfare-enhancing. We directly test our theoretical predictions in a simulated lifecycle-consumption task, and confirm all the main predictions of the theory. We find strong evidence of over-consumption early in the lifecycle and this over-consumption is exacerbated when income is received later in the lifecycle. There is no evidence for learning over the experiment and continual dynamic feedback on one's current balance has virtually no effect. We then run the first incentivized study to measure the extent of exponential-growth bias in a representative sample of the U.S. population. We find substantial EGB, with approximately one-third of subjects estimated to be the fully-biased type. The magnitude of the bias is negatively associated with total asset accumulation, even controlling for income and education. We find that the bias is robust to a simple graphical intervention, and test our model of EGB against potential alternative mistakes. Finally, in a laboratory study with a sample of elite college students we again find systematic bias, but show that subjects are largely unaware and unwilling to pay for aid. This overconfidence suggests that markets will not correct for the bias, as consumers will not seek enough financial advice.
Author(s): Matthew R. Levy and Joshua Tasoff
Topic: Anomalies
Link: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2144358
Back to session: Anomalies 3
ABSTRACT:
We define and experimentally investigate two concepts that are inconsistently treated in previous literature: optimism and overconfidence. These types of ``incorrect' beliefs have potentially major decision-making implications. We are particularly interested in the role optimism and overconfidence may play in an individual's choice to become self-employed despite the high probability of failure in entrepreneurship or to remain in self-employment despite unexpectedly low earnings. Our definition of optimism amounts to a formalization of ``wishful thinking,'' whereby agents' incorrect beliefs are driven by their preferences for certain outcomes. We characterize overconfidence as agent certainty that is greater than what can be warranted by available data. We depart from previous literature in three important ways. (1) We propose definitions that are general, i.e., not context-dependent, and therefore applicable to diverse decision-making scenarios. (2) Our definitions explicitly distinguish optimism from overconfidence, which allows us to separately identify the presence of these beliefs and their interaction. (3) Our experimental approach is designed to eliminate potentially confounding factors in eliciting errors in beliefs. In particular, we can pinpoint optimism by maintaining explicit control over which outcomes are ``favorable.'' Moreover, we first elicit agent beliefs about an objective distribution through a non-performance-based task to ensure that the results do not reflect agent beliefs about their own ability. This design will allow us to subsequently incorporate performance and ``re-elicit' beliefs to then isolate ability-specific optimism and overconfidence.
Author(s): Stephanie A. Heger (Washington University in Saint Louis) & Nick W. Papageorge (Johns Hopkins University)
Topic: Anomalies
Back to session: Anomalies 3
ABSTRACT:
We analyze pre and post-task confidence in an experiment in which subjects bid in multi-unit common value auctions. Subjects return for a second session, so we are able to assess how performance affects the evolution of confidence. Those with low confidence prior to the first session underestimate performance while those with high confidence overestimate performance. Although the change in pre-experiment confidence from session one to session two is close to zero, the dispersion in confidence increases. For those with moderate initial confidence, the change in confidence depends significantly on performance in session one. For those with high initial confidence, the change in confidence does not depend on performance, and the correlation between confidence prior to session two and confidence prior to session one is significantly higher than for those with neutral or low confidence.Subjects with high initial confidence also base their perception of post-experiment relative performance primarily on pre-experiment confidence: an effect not present in the moderate and low confidence groups. Based on a pre-experiment survey, we also find that those with high initial confidence are more likely to have prior experience trading stocks or options.
Author(s): Emmanuel Morales-Camargo, Orly Sade, Charles Schnitzlein, Jaime F. Zender
Topic: Anomalies, auctions
Back to session: Anomalies 3
ABSTRACT:
We investigate whether the Allais paradox is the result of framing, steering subjects toward making a choice that they are likely to reverse if the choice is presented more transparently. The answer would have the following implications: If the Allais paradox is entirely due to the way the choices are presented, then Allais is not about independence of irrelevant alternatives, or about the appropriateness of expected utility theory. To the extent that Allais is limited to decisions by description, it tells us about risk perception, not about risk attitude. Conversely, if the Allais paradox is insensitive to presentation, then the criticism of the von Neumann-Morgenstern approach to expected utility is deep, and decision theoretic research ought to shift its focus away from expected utility and toward one of the many proposed alternative frameworks.
Author(s): Elif Incekara-Hafalir, Jack Stecher
Topic: Anomalies
Back to session: Anomalies 3
ABSTRACT:
We consider an aspiration in a choice situation as the most desirable option within the set of alternatives some of which are possibly unavailable in that choice problem. Even though unavailable at the time of choice, mere presence of an aspiration alternative in the environment may still affect choice among available options by leading the agent to choose alternatives “similar” to her unavailable aspiration. This results in violations of the standard theory if the most similar option is different from the the best item available in the choice problem. Motivated by these, in this paper, we study effects of possibly unavailable aspirations on choice by incorporating into our experimental design two critical aspects ignored in the literature: (1) aspirations are endogenously formed; (2) similarity notion is subjective. Results of our experiment show that (1) a significant number of choice reversals occur when an unavailable alternative deemed as an aspiration is introduced into the environment; (2) odds of choice reversal between two options upon the introduction of an unavailable option into the environment increases significantly when the unavailable option is an aspiration compared the case it is not an aspiration; (3) odds that alternative subjectively closest (most similar) to the unavailable aspiration will be chosen is significantly greater when the unavailable aspiration is present in the environment compared the case it is not present at all; (4) tendency to choose the option 'closest' to unavailable aspiration is explained better by the subjective distance function than the Euclidean one.
Author(s): Begum Guney and Michael Richter
Topic: Anomalies
Back to session: Anomalies 4
ABSTRACT:
In a few provinces in China, applicants for college admissions (still) have to choose the universities they will apply to before they know their scores on the 2-day National College Admissions Exam. The exam is held only once a year, with retake often infeasible, and effectively they can only apply to a very limited number of universities. They have to choose the universities based on self-estimated scores, aided by the official answers and detailed grading guide published immediately after the exam. We collect more than a thousand applicants' estimated and actual scores. We find that most estimates were very close to the actual scores, but consistent with typical findings in the literature, both genders still tend to overestimate their scores, and boys overestimate significantly more than girls. Our study on overconfidence is under real high stakes and the simple environment allows very limited room for alternative interpretations.
Author(s): Huibin Yan & Bei Zhang
Topic: Gender, anomalies
Back to session: Anomalies 4
ABSTRACT:
We conduct intertemporal choice laboratory experiments to test the effectiveness of two incentive schemes designed to encourage saving over current consumption. The first scheme involves a commitment to save from future (unrealized) increases in income while the second incentive scheme provides a choice of two savings plans one of which unambiguously dominates the other. Importantly, these schemes are evaluated in an environment in which reference-dependent preferences can inhibit savings behavior via the endowment effect. With respect to the first scheme, our results indicate that individuals are willing to save money at a much lower rate of interest when money has been earned but is not yet in their possession. With respect to the second scheme, our results indicate that provision of a choice between two savings plans where one plan unambiguously dominates the other, leads to a significant increase in the numbers of individuals who choose to save.
Author(s): William G. Morrison
Topic: endowment effect, saving
Back to session: Anomalies 4
ABSTRACT:
Periods of inactivity have generally been treated as innocuous breaks amid other tasks. Previous experimental studies, however, have observed that subjects are willing to incur effort costs and loss of expected value in order to mitigate inactive periods. Additionally, the psychology literature suggests a possible link between periods of inactivity and changes in behavior. In this experiment, subjects will alternate stages between playing a dictator game with unknown roles and a fixed period of time in which they either are given only a blank screen or are provided with a menu of simple activities in which to partake. Dictatorial decisions following periods of inactivity will be contrasted with those following the more active periods. Additionally, at the conclusion of the experiment, subjects are paid a diminishing wage to remain inactive, with the ability to leave the experiment at any time. Subjects leaving before the rate reaches zero would suggest a utility cost associated with inactivity.
Author(s): John Jensenius III
Topic: Labor, anomalies
Back to session: Anomalies 4
ABSTRACT:
This study compares bidding behavior in a multi-unit uniform-price descending price (Dutch) auction under four different information conditions. Bidders are either informed of the number of bidders in the auction, or know that it is one of two possible sizes; they also either know the number of units remaining for sale or are unaware of how many units have been taken by other bidders. We find that revealing group size decreases bids,and therefore revenue, if units remaining are not shown. When group size is unknown the price also falls if the number of units remaining is revealed. The most efficient and largest revenue outcome occurs when bidders are not provided information on either group size or units remaining.
Author(s): Joy A. Buchanan, Steven Gjerstad, David Porter
Topic: Auctions
Link: http://chapman.edu/research-and-institutions/economic-science-institute/_files/WorkingPapers/buchanan-gjerstad-porter-information-effects-multi-unit-dutch.pdf
Back to session: Auctions
ABSTRACT:
This paper uses a laboratory experiment to show that the revenue premium associated with bidder's choice ('right-to-choose') auctions is dampened when bidders value multiple goods. In a bidder's choice auction, the winner chooses her preferred good from a set of available goods. Economic theory has shown that this auction format results in higher revenue under risk aversion than traditional good-by-good auctions, making it preferable to parties selling multiple similar goods. Most theoretical and experimental work focuses on bidder's choice auctions where bidders value only one of the available goods. However, bidders often value multiple goods in many settings where bidder's choice auctions are used (e.g. the sale of condominiums, antiques, and customized telephone numbers). Our design allows for persistent competition, where bidders have positive values for all goods in the available set and bidders remain in the auction even after their most preferred good has been selected. We find that persistent competition mutes the revenue superiority of bidder's choice auctions, which is consistent with the notion that the perceived risk of losing one's most preferred good is softened when they have multiple chances to win. This result implies that bidder's choice auctions should be used in settings where each bidder is likely to strongly prefer one good over the others. In addition, this result may explain why two field experiments found contrasting results: the persistent competition in Alevy et al. 2009 may have muted the theoretical revenue superiority of the bidder's choice mechanism, while the greater taste diversity that characterized Alevy et al. 2010 may explain why the authors found substantial support for the original theory.
Author(s): Julianna Butler (presenter), Michael K. Price, Jonathan Alevy
Topic: Auctions
Back to session: Auctions
ABSTRACT:
We study first- and second-price sealed-bid private value auctions where the budget (bid cap) of each bidder is endogenously set by a principal. The bidder can only bid up to this bid cap. The bidder is better informed than the principal regarding the value of the auctioned item and she has some additional private benefit from winning the auction. As opposed to the literature on auctions with exogenous budget constraints, our theory predicts that the principals adjust their bid caps based on the auction format and, as a result, will neutralize the choice of auction format. Our experimental findings support the theory and show that the principals are indeed strategic in their bid cap decisions. Taken to the logical conclusion, the revenues may be equivalent and the expected efficiency rates may be the same in the two formats, despite budget constraints. Our results are consistent with the common use of ascending (i.e. second-price) formats in high-stakes auctions where bidders may run up against financial constraints.
Author(s): Lawrence M. Ausubel, Justin E. Burkett, and Emel Filiz-Ozbay
Topic: Auctions
Back to session: Auctions
ABSTRACT:
Auctions of vehicle licenses in Shanghai were reconfigured in 2008 in an attempt to reduce prices by giving bidders the option of modifying initial bids. The modified bids are required to be within a narrow band around the current lowest accepted bid. Bidding constraints limit price competition, so efficiency may be degraded if intended revenue reductions are achieved. The effectiveness of the new procedures, however, is unclear in light of subsequent increases in license prices, even after the license quota was doubled. This paper reports a laboratory experiment to evaluate the revenue and efficiency consequences of the Shanghai auction format.
Author(s): Evan Zuofu Liao and Charles A. Holt
Topic: Auctions
Back to session: Auctions
ABSTRACT:
We present a dynamic, continuous-time bargaining game with asymmetrical private information. Two players must agree on how to split a pie; one of them is informed of the pie size, where the other only knows the probability distribution from which it is drawn. Players' types are fixed for the whole experimental session, and subjects are randomly matched in each period to prevent reputation effects. On each trial, players bargain (using mouse clicks on a computer interface) over the uninformed player's payoff. Subjects must come to an agreement within ten seconds - otherwise both of their payoffs are set to zero (a strike occurs).
Author(s): Gideon Nave, Alec Smith, Colin Camerer
Topic: Bargaining
Link: www.gidinave.com
Back to session: Bargaining 1
ABSTRACT:
Why do legal disputes ever go to trial? Prior research emphasizes the role of mistakes, irrationalities, or asymmetric information because rational litigants with complete or symmetric information should choose pre-trial settlements over the costs and risks of trial. We provide a game-theoretic argument for why going to trial, even though risky and costly, can be both rational and socially efficient when trial can create precedence that deters future costly litigation. Experimental evidence supports these predictions. Our findings provide new insights into our understanding of the incidence of litigation and trial.
Author(s): Michael McBride; Stergios Skaperdas; Pi-Han Tsai
Topic: Bargaining
Back to session: Bargaining 1
ABSTRACT:
Conventional Arbitration (CA) and Final-Offer Arbitration (FOA) are the two methods of arbitration used in the field for solving disputes. Theoretical predictions show that for these approaches final offers may not converge to allow an agreement. Other theoretical arbitration mechanisms have been developed that do lead to agreement, but they do not outperform CA and FOA in laboratory experiments. The most promising results are from Amended Final Offer Arbitration (AFOA) which outperforms FOA and weakly outperforms CA. This paper presents an experiment to evaluate a more general case of AFOA, ?-Final Offer Arbitration (?-FOA). This mechanism is similar to a ôsecond-price auction,ö which punishes the loser with a value proportional (?) to the difference between her final offer and the arbitrator's fair settlement. The experiment furthermore divides the pool of subjects within a session in two groups according to their estimated risk preferences due to the fact that the contract zone depends on the relative risk preferences of the subjects involved in negotiation.
Author(s): Daniel Nedelescu
Topic: Bargaining
Back to session: Bargaining 1
ABSTRACT:
We report results from bilateral bargaining experiments with attention to the effects of a settlement bonus on strategic decision making behavior. In instances with a sufficiently large settlement bonus, truthful revelation emerges as the dominant strategy. However previous work (Parco and Rapoport, 2004) has experimentally tested this ôBonus Effectö and found that although the presence of a settlement bonus improves efficiency, behavior falls drastically short of the normative predictions. This finding illustrates the persistent tendency of decision makers to bid strategically, i.e. shading their bids, even when truthful revelation is a strictly dominant strategy. Herein we investigate the influence of the framing of information and look for ways to nudge decision makers toward making better choices in these strategic environments. Additional results from an adaptive reinforcement-based learning model are discussed as they relate to a potential innate bias for strategic misrepresentation even when contrary to self-interest and collective-interest.
Author(s): James E. Parco and Ryan O. Murphy
Topic: Bargaining
Back to session: Bargaining 1
ABSTRACT:
We study experimentally how the ability to communicate affects the frequency and effectiveness of flexible and inflexible contracts in a bilateral trade context where sellers can adjust trade quality after observing a post-contractual cost shock and a discretionary buyer transfer. In the absence of communication, we find that rigid contracts are more frequent and lead to higher earnings for both buyer and seller. By contrast, in the presence of communication, flexible contracts are much more frequent and considerably more productive, both for buyers and sellers. Also, both buyer and seller earn considerably more from flexible with communication than rigid without communication. Our results show quite strongly that communication, a normal feature in contracting, can remove the potential cost of flexibility (disagreements caused by conflicting perceptions). We offer an explanation based on social norms.
Author(s): Jordi Brandts, Gary Charness, Matthew Ellman
Topic: Bargaining, IO communication
Back to session: Bargaining 2
ABSTRACT:
Kingston (1976) and Anderson (1977) show that the probability that a given contestant wins a best-of-2$k$+1 series of asymmetric, zero-sum, binary-outcome games is, for a large class of assignment rules, independent of which contestant is assigned the advantageous role in each component game. We design a laboratory experiment to test this hypothesis for four simple role-assignment rules. Despite play not uniformly conforming to equilibrium, our results show that the four assignment rules are observationally equivalent at the series level: the fraction of series won by a given contestant and all other series outcomes do not differ across the four rules.
Author(s): Bradley Ruffle and Oscar Volij
Topic: Games, bargaining
Link: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2128225
Back to session: Bargaining 2
ABSTRACT:
We provide a theoretical framework for analyzing when decision makers should rationally follow default options instead of making active choices. We model decision-making in environments with default options as strategic interaction between a policy maker and decision makers. Our model identifies two key dimensions that affect behavior of both parties in equilibrium. First, choices of policy makers and decision makers are influenced by the degree to which their interests are aligned. Second, the strength of default effects depends on the parties' relative level of information about the decision environment, and decision makers' optimal choices. We test the key predictions of our model in a laboratory experiment and find supportive evidence that strategic incentives as well as the level of information affect behavior of policy makers and decision makers.
Author(s): Steffen Altmann, Armin Falk, Andreas Grunewald
Topic: Games, bargaining
Back to session: Bargaining 2
ABSTRACT:
The canonical result in two player ultimatum bargaining is 'fairness', that the responder is willing to forgo payment to reject 'unfair' offers, which in turn leads proposers to make offers more generous than the subgame perfect equilibrium. In this paper we explore the impact of incomplete information on the proposer in making fair offers and the responder whether they are willing to accept offers that are, in expectation, unfair, but the realization need not be. Finally we explore the value of information for either the proposer or the responder about the size of the pie.
Author(s): Judd Kessler and Muriel Niederle
Topic: Bargaining, fairness
Back to session: Bargaining 2
ABSTRACT:
We conduct a laboratory experiment to test if there are differences in behavior when subjects can donate either time or money to charity. Our subjects perform an effort task to earn money. In one treatment they can have their efforts accrue to a charity instead of themselves. Conversely, in other conditions subjects may only earn money for their private account but then donate it to a charity. We vary the timing and availability of donation opportunities in the monetary donation settings to test the impact of subtle solicitation. We find that subjects with a more opportunities to donate will donate more often and in larger amounts. Further, subjects that give time and effort to charity give a larger contribution than subjects who give monetary donations and we posit this difference is driven by different warm-glow from the two gift types.
Author(s): Alexander L. Brown, Jonathan Meer, J. Forrest Williams
Topic: Charitable donations
Back to session: Charitable Giving 1
ABSTRACT:
Blood shortages are common throughout the world and in most developed countries voluntary blood donations do not produce enough blood. However, sufficient pro-sociality in these societies may not be the only (or possibly major) hurdle given massive donation spikes during emergencies such as 9/11 and Hurricane Katrina. Instead, frequent shortages during normal times may be due to difficulties in coordination and noisy demand signals. The current study proposes a donor registry that can solve the coordination and noisy demand signal problems; the registry invites donors to donate only when there is demand. The registry can also organize donors so that those with the highest willingness to donate will be most likely invited to donate, thus further improving social welfare. The current lab study explores five alternative methods to organize the registry beyond testing its viability to tease apart and understand the potential different avenues to efficiently run the registry. The treatments examine three registries: (1) a basic registry that only invites the exact demand needed; (2) a sequential registry that continues to invite registry members until all demand is satisfied; and (3) a learning registry that gives priority to registry members based on past good behavior. We find that all three registries significantly increase 'lives' saved, decrease unneeded donations and increase efficiency. Moreover, disaggregated data shows that the registries work especially well to address the variability in demand better matching overall supply to demand rather than simply increasing overall supply.
Author(s): Robert Slonim, Carmen Wang
Topic: Charitable giving
Back to session: Charitable Giving 1
ABSTRACT:
While in the last two decades the availability of measures of efficiency has increased in the market for charitable giving, little is known about the impact this information and its public visibility has on small donors’ behavior.Using a laboratory experiment where participants play dictator games with charities they freely choose, we study whether receiving new information on actual efficiency modifies previous donation decisions, and how the response to this information depends on that information being public or not. We show that: (1) as long as charities’ efficiency remains private information, individuals reward better-than-expected-charities by increasing their donations;(2) When giving happens under full anonymity, individuals largely disregard bad news about their charities. (3) When efficiency of one’s own charities is revealed to others, more efficient charities receive less, that is, people advertising their gifts reduce their contributions when facing better-than-expected charities, and donate more to unexpectedly inefficient ones. We show that this selection mechanism is driven by image motivated donors for whom information and donation amounts are substitutes in terms of social image payoffs.
Author(s): Luigi Butera, Jeffrey Horn
Topic: Charitable giving
Back to session: Charitable Giving 1
ABSTRACT:
I present the results of a field experiment involving charitable giving. Treatments involve varying language in a non-profit's fundraising letters that alter the economic incentive to give. More specifically, treatments vary the extent to which an individual's donation decision impacts the amount of 'matching money' given to the non-profit by an outside benefactor (in the spirit of Karlan and List 2007, Karlan et al 2011, and Rondeau and List 2008). Unlike those papers, this experiment includes a novel fundraising institution where matching money is provided if a sufficient level of donations is achieved. The effectiveness of this institution is compared to more traditional fundraising alternatives. Results suggest that donors do value matching money that can be triggered by their own donation, but that beliefs about other donors' likelihood to contribute are important in determining their own likelihood to contribute.
Author(s): Mike Schreck
Topic: Charitable giving, field experiment
Back to session: Charitable Giving 1
ABSTRACT:
Rational Inattention provides a compelling framework for modelling the behavior of agents who face costs of gathering and processing of information: The optimizing agent is free to process information in any way they like, but more accurate representations of the world are more costly to obtain. Such models have been used widely in both macroeconomics and game theory. However, little is known about the nature of the cost function that agents face when gathering information. Generally it is assumed that costs are related to Shannon entropy, yet other authors have proposed alternative restrictions (e.g. Woodford [2012], Ellis [2012]). The predictions of rational attention depend crucially on the cost function assumed.
Author(s): Andrew Caplin and Mark Dean
Topic: Cognition
Back to session: Cognition/Inattention
ABSTRACT:
In this paper, I model how buyers attend to information about product quality when prices are easily observable using 'rational inattention' theory (Sims [JME, 2003]). This model predicts how attention varies with prices and the strategies of sellers. I test these predictions by providing experimental subjects with information about product quality that requires cognitive effort to process. I find that rational inattention broadly outperforms alternative theories of information acquisition, which provides the first experimental evidence of rational inattention as an informational constraint.
Author(s): Daniel Martin
Topic: Cognition
Back to session: Cognition/Inattention
ABSTRACT:
Research in psychology suggests that human behavior is the product of two processes in the brain: the 'reasoning' system and the 'intuitive' system. It is thought that when the reasoning system is bogged down, it will not be able to correct mistakes made by the intuitive system. We test this theory by seeing how memorizing numbers of different lengths affects subjects' answers to: (1) arithmetic problems, (2) choices involving varying risk, (3) choices over time, and (4) choices with framing effects. We find that longer number memorization leads to more math mistakes, less risk-taking, more impatience, and more anchoring.
Author(s): Salar Jahedi, Cary Deck
Topic: Cognition, risky choice
Back to session: Cognition/Inattention
ABSTRACT:
This paper reports results of computerized experiments that test for behavioural biases by motorists when facing different road pricing mechanisms. Road pricing or congestion charging has in recent years become a widely used mechanism for managing transportation demand in urban areas. While the economic principle behind congestion charging is well understood, and the associated technologies are mature, the behavioural effects of different mechanisms of charging have hardly been studied. Road pricing system designers have little directly relevant research to draw upon.
Author(s): Qiyan Ong, Yohanes Eko Riyanto, and Walter Theseira
Topic: Cognition, traffic
Back to session: Cognition/Inattention
ABSTRACT:
We test two hypotheses related to self-control and patience as measured by time preference. Are reserves of self-control depleted by heavy use? Can supplying individuals with sugar reinforce patience? To elicit time preferences, we calibrate a frontier technique from the experimental literature to generate extremely precise time preference estimates. We use the results of this elicitation to test for treatment effects relating to the two hypotheses. We find that individuals exposed to a heavy cognitive load immediately prior to the elicitation actually exhibit more patience than those who have not. Both a placebo intervention and a sugar intervention generate patience improvements with respect to no intervention. We find some evidence that the sugar has an additional patience-improving effect beyond the placebo. Lastly, we note that we find evidence for present-bias of monetary payments amongst our subjects, in contrast to many previous studies.
Author(s): Michael Kuhn
Topic: Self control, patience, time preference
Back to session: Cognition
ABSTRACT:
We provide experimental evidence that cognitive skills, social skills and personality traits all affect strategic behaviour in an ultimatum game played by children. Subjects were 450 seven-year-old children in 21 classes in 8 Italian elementary schools, playing a three-round ultimatum game. The cognitive abilities of the experimental subjects are measured by accessing school records and by using a set of the visual matrices resulting from part of the brief intelligence test developed by Kaufman (KBIT). The personality traits of the children are measured both with reports from their teachers and self-reports from the children themselves. The children completed a 30-item Big Five questionnaire and an equivalent 65-item questionnaire was administered to their teachers. Peer relations and social skills were quantified by asking the children to choose five peers in their class who they consider to be best friends. They were also asked to identify peers with whom they usually play and those they usually ask for help during class hours. Negative nominations were also elicited. Each child was given two scores: (i) a popularity score, indicating the number of times the child was chosen by a classmate in one of the fields investigated; and (ii) a friendship score, showing how many times the childs relationships were reciprocated. These scores were adjusted for unequal class size using a Z transformation. A final questionnaire was also administered with the aim of recording demographic information: country of origin of parents, size and the social status of family, and also whether the child practises any sports, etc. Our first result is that cognitive abilities, as recorded by school grades, are positively correlated not only with childrens personality traits , as expected, but also with popularity and number of friendships. The second is that the first offer in the ultimatum game is mainly explained by cognitive abilities as measured by school marks and by the Mental Openness personality trait, together with the composition of the class. The other personality traits such as introversion, consciousness and agreeableness together with the child's social skills, explain behaviour along the game. In other words, the fact that the same choice is repeated in the second or third round by the same subject is well explained by his/her personality and social skills. This is true regardless if the child is playing as proposer or receiver.
Author(s): Chiara Rapallini, Aldo Rustichini
Topic: Cognition, bargaining, social preferences
Back to session: Cognition
ABSTRACT:
Our goal in this paper is to examine whether human factors in terms of cognitive capacity can explain the differences in subjects market performance. To do this, we measure subjects' working memory capacity and let them trade in double auctions. We had two treatments. In treatment 1, subjects have to maximize their profits in a market where other traders are truthful bidding agents. In treatment 2, subjects have to compete with adaptive software agents. Because the behavior of subjects' opponents are controlled, there is no room for other factors such as the complex interaction among human's strategies to get involved. Our auctions thus serves as a good environment to answer our question.
Author(s): Shu-Heng Chen, Chung-Ching Tai, Lee-Xieng Yang
Topic: Cognition, markets
Back to session: Cognition
ABSTRACT:
We conduct experiments in which we attempt to affect choices in simple economic games using cognitive dissonance. Specifically, we have subjects answer a question about fairness and then we use two manipulations to create dissonance between their answers and self-interest in dictator and public good games. When the mechanism for creating dissonance has a social component, subjects reduce dissonance by choosing pro-social actions. When the dissonance mechanism is private, we observe no significant change in behavior.
Author(s): Robert J Oxoby and Alexander Smith
Topic: Social preferences, cognition
Back to session: Cognition
ABSTRACT:
Established theoretical results predict a focal equilibrium for an internet auction with many simultaneous traders. In particular, it is a weakly dominant strategy for each bidder to bid on the lowest price unit and submit a bid equal to the minimum of the price of the second lowest unit and the bidder's valuation for the object. This project uses experimental methodologies to test how bidders actually formulate their bidding strategies in such a competitive auction market. We design a laboratory experiment that tests the predictions of an 'alternating recognition' model where bidders move in a randomly-determined order when choosing with which seller and how much to bid. Although the above equilibrium is in weakly dominant strategies, it is not unique. We test if this strategy is in fact focal to uncover the nature of bidding behavior in online marketplaces such as eBay. Further, we incorporate psychological factors into our experimental setting. In particular, we use the personality measurement system known as the Big-Five Trait Taxonomy to quantify the degree to which our experimental subjects vary along the following personality dimensions: extraversion/introversion, agreeableness/antagonism, conscientiousness/lack of direction, neuroticism/emotional stability, and openness/closedness to experience. Our hypotheses draw from psychological literature to connect personality traits to a competitive setting where economic agents are required to engage in an advanced decision-making process under time constraints while competing against similarly situated agents whose personality traits differ from their own.
Author(s): Robert G. Hammond and Thayer Morrill
Topic: Auctions, homegrown
Back to session: Contests 1
ABSTRACT:
We experimentally study how the outcome from the previous round affects subsequent effort provision in sequential best-of-three contests. In particular, we compare individual and team contest, the latter in which a team wins if two members from that team exert higher effort than their respective opponents in two rounds. Economic theory predicts that albeit ex-ante homogeneous in ability, the first-round winners will exert higher effort strategically than the first-round losers and hence are more likely to win the second round. But this effect does not carry over in team contests. Behavior in the experiment supports the prediction for team contests but rejects that for individual contests. More specifically, outcome of the previous round does not affect effort provision for players from either side in a team contest. In an individual contest, however, such feedback reduces effort for the first-round winners but increases effort for the first-round losers to compensate their losses. The results suggest that being in a team makes a player evaluate feedback differently and hence guarantees stable high effort provision.
Author(s): Changxia Ke, Fangfang Tan, Qiang Fu and Jingfeng Lu
Topic: Contests
Back to session: Contests 1
ABSTRACT:
We run experiments on second price auctions with resale opportunities, where a speculator (a bidder with a value of zero) is commonly known to exist. Garratt and Troeger (2006) show that there is an asymmetric equilibrium, apart from the standard bid your value one. In this equilibrium the speculator gets the good in the first stage auction for a price of zero, and resells it in the second stage. We run two treatments, with fixed/random matching. We find that bidders do play something resembling the speculation equilibrium, however they tend to split the rents more fairly than predicted by theory. A fixed matching yields more speculation than random matching, as expected.
Author(s): Rod Garratt & Sotiris Georganas
Topic: Auctions, finance
Back to session: Contests 1
ABSTRACT:
Contest experiments find (i) overbidding relative to the standard Nash equilibrium prediction and (ii) heterogeneous behavior of ex-ante symmetric contestants. We provide different explanations for the overbidding phenomenon, including bounded rationality and mistakes, utility of winning, other-regarding preferences, probability distortion, and the shape of the payoff function. We also provide explanations for heterogeneous behavior in contests, including differences in preferences towards winning, inequality, risk and losses, and demographic differences. Furthermore, we suggest mechanisms that can reduce overbidding and induce more homogeneous behavior. Finally, we discuss directions for future research.
Author(s): Roman M. Sheremeta
Topic: Auctions, overbidding
Back to session: Contests 1
ABSTRACT:
This paper examines the effect of electoral delegation on providing global public goods when the elected delegate has the decision power on contributions at the subgroup level. People mostly vote for pro-social delegates who assign an equally high contribution for everyone in the subgroup, even though these delegates face the temptation to free ride on contributions of other subgroups and to exploit members in their own subgroup. Moreover, the delegation mechanism protects members of a subgroup from being exploited by other subgroups or by a selfish delegate. The results highlight the importance of the delegation mechanism in solving the conflict between equality and efficiency.
Author(s): Martin Kocher (University of Munich), Fangfang Tan (Max Planck Institute for Tax Law and Public Finance) and Jing Yu (University of Munich)
Topic: 2nd paper
Back to session: Cooperation 1
ABSTRACT:
This experiment tests theoretical predictions for a stochastic game that is a variant of the infinitely repeated prisoner's dilemma. There are two prisoner's dilemmas, differing in their respective payoffs. After playing one of the dilemmas in period 1, the period 2 dilemma is determined stochastically and then this dilemma is repeated infinitely thereafter. The main treatment variable is the stochastic distribution, and theoretically is crucial to determining when cooperation is possible. Of particular interest, in contrast to the standard infinitely repeated prisoner's dilemma, is that the theoretical prediction for the most cooperative equilibrium can include both cooperation and defection on the equilibrium path.
Author(s): Andrew Kloosterman
Topic: Cooperation, games
Back to session: Cooperation 1
ABSTRACT:
This paper presents experimental evidence concerned with behavior in Cournot duopolies where players indulge in a form of pre-play communication termed as revision phase before playing the 'one-shot' game. During this revision phase individuals announce their tentative quantities that are publicly observed and revisions are costless. The payoffs are determined only by the quantities selected at the end in a real time revision game while in a Poisson revision game, opportunities to revise arrive according to a synchronous Poisson process and the tentative quantity corresponding to the last revision opportunity are implemented. Contrasting results emerge. While real time revision of quantities results in choices that are more competitive than the static Cournot-Nash, significantly lower quantities are implemented in the Poisson revision games. This shows that partial cooperation can be sustained even when individuals interact only once.The dynamics of revisions gives rise to the following two observations. First, while a real time revision game is characterized only by late upward quantity adjustments, the Poisson revision games are characterized primarily by initial downward adjustments and sometimes also by the late upward revisions. Second, the quantity adjustments during the initial period of the revision phase show that individuals imitate their opponent’s desired quantity choices, whereas, behavior towards the end of the revision phase can be explained by the best response to the competitor’s desired output.
Author(s): Nilanjan Roy
Topic: Cooperation, games
Back to session: Cooperation 1
ABSTRACT:
Punishment institutions may encourage pro-social behavior, as demonstrated by the increased cooperation often observed in public goods settings when punishment is made available. On the other hand, punishment institutions may spur a behavioral backlash, as with the reduced cooperation observed in gift-exchange games when principals may punish agents. Why does punishment cause increased cooperation in one setting but reduced cooperation in a similar setting? A major difference between the common implementations of punishment in the two kinds of games is that punishment is generally parameterized to be cheaper and more powerful in public goods games as compared to gift exchange games. We test this hypothesis in the gift exchange setting. We replicate the result that punishment is at best ineffective when punishment is parameterized as it often is in gift exchange experiments. We then show that when we change parameters to render punishment as cheap and powerful as it often is in a public goods game, the ability to punish increases cooperation. These results are observed with ex ante punishment, in which the punishment decision occurs before the cooperation decision, and are consistent with payoff-maximizing behavior on the part of agents and principals. Thus, the 'size of the stick' explains some differences between results in public goods and gift exchange games. However, our results highlight a new puzzle. In public goods games, punishment decisions are often made after the cooperation decision. In our results, such ex post punishment is rarely used and an institution allowing this kind of punishment has no effect on cooperation. Why, then, does ex post punishment seem successful in public goods games?
Author(s): Patrick Aquino, Robert Gazzale, Sarah Jacobson
Topic: Cooperation,
Back to session: Cooperation 1
ABSTRACT:
The threshold public goods game, often referred to as the provision point mechanism, is a voluntary contribution mechanism with great potential for efficient outcomes. Bagnoli and Lipman (1989) demonstrate that, under complete information, the undominated, trembling hand perfect refinement of Nash equilibrium is sufficient to guarantee efficient provision of a unitary public good. However, this theory does not select among the multiplicity of remaining equilibria, which experimental evidence suggests leads to a significant coordination problem. Theory also implies an important role for payoff uncertainty, while the experimental work of Marks and Croson (1999) appears to contradict this prediction. I propose a solution to this puzzle using logit quantal response equilibrium. The behavioral model will parse the effects of strategic and informational uncertainty on provision rates. Using the maximum likelihood estimator admitted by the quantal response framework, I test this new explanation with a two-by-two experiment.
Author(s): Daniel K. Saunders
Topic: Coordination
Link: http://econdan.files.wordpress.com/2012/02/saunders_abstract1.pdf
Back to session: Coordination 1
ABSTRACT:
In a pure coordination game with multiple Nash Equilibrium, subjects can only randomly make their decision when there is no focal point. When a payoff focal point exist, the focal point itself can attract subjects to choose, and hence, subjects can coordinate more often when focal points are present even if the payoff of the focal point is lower. As the payoff of the focal point increases, the chances of achieving coordination also increases. In contrast, when the payoff of the focal point is low, subjects would attempt to coordinate on locations around the focal point. We construct a mixed strategy equilibrium that involves choosing focal points as part of the mixed strategy and present some experimental support for the roles of focal points.
Author(s): Wei-Ting Liao and Joseph Tao-Yi Wang
Topic: Coordination
Back to session: Coordination 1
ABSTRACT:
In a period that is marked by extensive cross-national interactions, nationality may present one important focal point that individuals can coordinate on. This paper uses an experimental approach to study whether nationality serves as a coordination device. We let subjects from Japan, Korea, and China play stag-hunt coordination games in which we vary information about their partner. The results show that nationality serves as a coordination device if common nationality is the only piece of information available to the subjects. However, providing participants with additional information about their partner diminishes this effect. We also find that subjects are likely to coordinate on the Pareto-dominant equilibrium at about the same rate if the partner has a different nationality than if nationality is unknown.
Author(s): Olga Bogach, Andreas Liebbrandt
Topic: Coordination, norms
Back to session: Coordination 1
ABSTRACT:
Can cheap-talk always help achieve maximum efficiency in situations that require coordination? Utilizing a class of two-person two-action coordination games, we show experimentally that cheap-talk may or may not be informative, rending it helpful for or detrimental to efficiency (relative to no cheap-talk condition) and that the nature of coordination outcome crucially hinges upon two key equilibrium-selection criteria: Payoff- and Risk-dominance. To explain our data, we develop a level-k model of thinking in which players tremble while choosing their actions and incur lying costs. Econometric estimations confirm the superior explanatory power of our model over Ellingsen and Östling (2010) model.
Author(s): Quazi Shahriar, Subhasish Dugar
Topic: Coordination, games
Back to session: Coordination 1
ABSTRACT:
Individuals make decisions for others could be different from what they do for themselves. In this paper, we investigate the differences between the decisions made for spouses and subjects' own decisions and the predictions of spouses' decisions regarding intertemporal choices. We also have examined whether and to what extent the decisions made for spouses are related to both own decisions and the predictions of spouses' decisions. We conduct a high stake field experiment with 122 married couples in rural China, and use the Convex Time Budget experimental method to elicit subjects' time preferences over own money and their spouses' money and the predictions of spouses' time preferences. We find that husbands have made significantly more patient but less time consistent decisions for wives than their predictions of wives' decisions. However, neither husbands nor wives are more (im)patient or time (in)consistency in the decisions made for spouses than their own decisions at the household level. More patient husbands/wives tend to make more patient decisions for wives/husbands than their predictions of wives/husbands' decisions, yet they tend to make more impatient decisions for wives/husbands than their own decisions. Furthermore, we find that both own decisions and the predictions of spouses' decisions have a significant and positive influence on the decisions made for spouses.
Author(s): Xiaojun Yang
Topic: Field experiment, anomalies
Back to session: Field 1
ABSTRACT:
We investigate how a known, successful behavioral nudge fares in a high stakes context the decision to accept a job. The job offers come from Teach for America (TFA), a nationwide non-profit that, every year, admits around 8,000 teachers and makes arrangements for them to work in at-risk public schools for a period of two years. We find a substantial, significant effect on the decision to join TFA from a small, one-line change to the offer letter. From survey data, we provide some insight as to the mechanism driving the change.
Author(s): Lucas Coffman, with Judd Kessler (Wharton) and Clayton Featherstone (Wharton)
Topic: Field experiment, charitable giving
Back to session: Field 1
ABSTRACT:
In low income countries with incomplete credit and insurance markets idiosyncratic risk is often pooled through personal transfers. Using a modified social dilemma game conducted in rural Kamchatka Russia and Western Alaska we find that players make large unconditional transfers using a simple sharing rule that reflects a subsistence food-sharing institution familiar to study participants; evidence suggests that risk can be pooled without an external commitment device or an opportunity for direct reciprocity.In this paper, we present results from a second-stage dictator game to test the robustness of sharing observed in the social dilemma game. In this almost double-blind dictator game, subjects use own earnings from the social dilemma to give to anonymous recipients at varying levels of social distance and worthiness. Preliminary results suggest that dictator game giving is sensitive to both social distance and the need of the recipient. Also there is a strong relationship between giving in the dictator game and sharing in the social dilemma; more generous dictators also share more with players experiencing a shock in the social dilemma.
Author(s): Lance Howe, James Murphy, and Andrew Gerkey
Topic: Field experiment, charitable giving
Back to session: Field 1
ABSTRACT:
Blood shortages are common throughout the world. High seasonal variation in donations and massive spikes during disasters (e.g., 9/11, Hurricane Katrina and the Australian Bushfires) reveal enormous potential blood supply, suggesting that frequent shortages during normal times may be due to poor market signals (i.e., no price mechanism). The current study examines the introduction of a donor registry that will proxy for a clearinghouse, aggregating information on needs for blood and potential donors' willingness-to-donate to more efficiently coordinate demand and supply. We hypothesize that people will be more likely to donate because the registry can credibly signal demand to registered donors and due to an implicit commitment to donate in the future when donors join the registry.
Author(s): Ellen Garbarino, Robert Slonim and Carmen Wang
Topic: Field experiment, charitable giving
Back to session: Field 1
ABSTRACT:
The decision to immigrate is a risky and ambiguous one. However, very little is understood about how individuals make decisions like this one. We draw on competing theories of the transmission of information in social networks in immigration, positive externalities and herding. We use data from an artefactual field experiment to test these theories, and to identify the role of information and its transmission in social networks in current and potential immigrants. Our results demonstrate the ways in which individuals exchange information about risky and ambiguous decisions with others in their social network, and rely on that information in informing their own decisions. The results from this study have the potential to enrich the explanations and predictions on how information affects individuals' immigration decision, which will provide an additional policy lever to affect the intensity of immigration.
Author(s): NATALIA CANDELO, RACHEL T. A. CROSON, CATHERINE ECKEL
Topic: Field experiment, labor
Back to session: Field 2
ABSTRACT:
This paper presents evidence from a field experiment in Mozambique designed to increase the productivity of health workers, and subsequently improve the treatment of HIV/AIDS, through the use of team financial incentives. A group of public health facilities are paid a per-unit fee for delivering certain HIV/AIDS treatments and procedures and the fees are distributed proportionally by seniority among the facility workers. A second group of facilities is monitored and serves as a control.
Author(s): Ben Fitch-Fleischmann and Alfredo Burlando
Topic: Field experiment, labor
Back to session: Field 2
ABSTRACT:
Local authorities may play a crucial role in voluntary provision of public goods in developing countries. We conduct a controlled field experiment in 52 communities in rural Bolivia, where community members pool resources to provide environmental education material for local schools. We find that democratically elected local authorities increase public good provision when they lead by example. The results are driven by two margins: (1) authorities increase their contributions when they are called upon to lead, and (2) they crowd-in the contributions of others. Both effects are stronger when authorities, as compared to random community members, lead by example. We explore two mechanisms underlying the result. First, authority influence is driven, in part, by the types of individuals elected to office in our study communities. Second, it cannot be explained solely by information signaling about the quality of the public good.
Author(s): B. Kelsey Jack and Maria P. Recalde
Topic: Field experiment, public goods
Back to session: Field 2
ABSTRACT:
This paper uses a field experiment in Sierra Leone to investigate elite capture of community projects. In 56 rural communities dependent on slash and burn agriculture we collaborated with a conservation NGO to vary the conditions under which an aid program is managed. In half the villages, the local chief was responsible for the management of the project. In the other half three randomly selected villagers (heads of household) managed the project. Project management entailed storing supplies and managing work crews for a community construction project. We use several indicators, including experimental games and audits of project value by local engineers, to measure elite capture and villager support for the NGO. We find evidence of significantly more elite capture in projects organized by chiefs, but despite that chiefs deliver equally valuable projects and, surprisingly, villages with a committee are much less satisfied with the project. This provides evidence that common villagers prefer their traditional social hierarchy to a more egalitarian alternative.
Author(s): Erwin Bulte, Andreas Kontoleon, John List, Ty Turley and Maarten Voors
Topic: Field experiment, public goods
Back to session: Field 2
ABSTRACT:
Examining how people share risk is an important question to ask when looking at specific communities of individuals. Examining risk pooling behavior may offer important insight into how communities handle unexpected shocks such as weather related disasters or other unexpected hardships. We examine risk pooling of subjects by using data from a field study of two small towns. Using the solidarity game introduced by Selten and Ockenfels (1998), we implement a simple, visual representation of the game. Two modified versions of the baseline treatment test potential policies.
Author(s): Haley Harwell and Catherine Eckel
Topic: Field experiment, risky choice, finance
Back to session: Field 3
ABSTRACT:
This paper studies whether the act of regularly saving money with a financial institution fundamentally changes one's time preferences. We take advantage of a field experiment in Nepal that randomized access to savings accounts for a population with limited access to formal savings products, thus generating random variation in the ability to practice delaying gratification in the savings dimension. We study time preferences using both a hypothetical survey and an incentivized intertemporal choice-task run a year after the start of the intervention. We find no evidence that experience with savings causes any changes to how people discount the future. We do find, however, that the treatment group with access to savings shows a greater sensitivity to changes in interest rates and is more willing to accept risk in a lottery task. Those findings indicate that that experience with savings may increase the intertemporal elasticity of substitution. We use a structural model to estimate time preferences with our sample and find present bias coefficients near 1 (no present bias) for both groups, annual discount rates in the neighborhood of 30% for both groups, and an intertemporal elasticity of substitution about 60% higher for those given access to savings accounts.
Author(s): Leandro Carvalho; Silvia Prina; Justin Sydnor
Topic: Field experiment, time preference
Back to session: Field 3
ABSTRACT:
Our main intervention pairs two complementary randomized interventions (an unconditional cash grant and a psychosocial transformation program) to test how binding are poverty traps (economic and behavioral, respectively) on urban street youth in Monrovia. In this context we measure both time preferences and executive function (e.g. cognitive self-control) of the subjects before and after the intervention. Initial baseline analysis correlates these measures to each other; to demographic variables; and to economic outcomes of interest. Further analysis seeks to test if either intervention was able to alter such measures, which are typically assumed by economists and psychologists to be fixed, in adults.
Author(s): Chris Blattman; Julian Jamison; Margaret Sheridan
Topic: Field experiment, time preference
Link: http://www.poverty-action.org/project/0166
Back to session: Field 3
ABSTRACT:
We conduct a randomized field experiment in Peruvian schools to test the impact health information given to children has on adult members of the household. Our results suggest that children are effective transmitters of relevant information related not only to their own utility but to the welfare of all household members. We combine survey data with our experimental data to further explore the mechanisms by which these impacts occur. We find that this effect is stronger as the number of children in the household exposed to the treatment increases as well as with the child's age (additional tests are being conducted). In rural areas where roads are poor and farmers have distant plots, the concentration of children at schools offers an enormous opportunity to spread information and knowledge on health and rural development policies to the rest of the household in a cost (and outcome) efficient manner. One of the main policy implications of our findings is thus that children can be (effective) drivers of socialization and development.
Author(s): Eduardo Maruyama, Maximo Torero, Angelino Viceisza
Topic: Field Experiments
Back to session: Field 3
ABSTRACT:
Many studies have examined the under-saving phenomenon, but few have addresed the closely related question of why individuals might fail to meet household spending targets. We are motivated by the observation that individuals are routinely negatively surprised by how much they spend when they get bills, look at receipts, or view their credit card statements. We conduct a field experiment at two grocery stores in a large chain in which subjects are approached directly before checking out of the store. Subjects are asked, among other things, how much they believe they are about to spend. We find that a disproportionate number of subjects underestimate how much they are about to spend. Moreover, subjects who underestimate guess amounts that are, on average, about 20% below the amounts they actually spend. Standard theory prescribes that individuals make decisions by maximizing their utility subject to a budget constraint. In contrast, we observe that consumers spend more than they believe they do. Possible reasons for this fallacy are explored through additional experimental evidence.
Author(s): Uri Gneezy, Adam Eric Greenberg
Topic: Field experiments, anomalies
Back to session: Field 4
ABSTRACT:
Previous studies that test one class of household model against another have focused almost exclusively on the average behavior in a population, hiding possible heterogeneity across households. Using varieties of dictator games played between spouses in rural Kenya, I test for heterogeneity explicitly and measure what fraction of households behave efficiently and what fraction respond to asymmetric information. I find that 97% of choices are inefficient, suggesting that the Collective model is inappropriate for the vast majority of this sample. I find that 37% of people give more when their choice will be revealed to their spouse, while 50% act as if there is perfect information in their household. These results together suggest that theorists should write non-cooperative models of the household that explain heterogeneous responses to asymmetric information, and policy makers should design different policies for different types of household.
Author(s): Jessica Hoel
Topic: Field experiments, bargaining, fairness
Back to session: Field 4
ABSTRACT:
In a randomized field experiment, we investigate the connection between work goals, monetary incentives, and work performance. Employees are observed in a natural work environment where they have to do a simple, but effort-intense task. Output is perfectly observable and workers are paid for performance. While a regular piece-rate contract serves as a benchmark, in some treatments workers are paid a bonus conditional on reaching a pre-specified goal. We observe that the use of personal work goals leads to a significant output increase. The positive effect of goals not only prevails if they are self-chosen by the workers, but also if goals are set exogenously by the principal -- although in the latter case, the exact size of the goal plays a crucial role. Strikingly, the positive effect of self-chosen goals persists even if the goal is not backed up by monetary incentives. We propose a novel incentive contract where -- through the choice of a personal work goal -- workers themselves determine the risk and the size of their bonus payment at the same time.
Author(s): Sebastian J. Goerg, Sebastian Kube
Topic: Field Experiments, Labor
Back to session: Field 4
ABSTRACT:
Based on the observation that performance monitoring in government is weak, contracts between principals and agents incomplete, and incentives generally low-powered, James Q. Wilson (1989) concluded that what is surprising is that bureaucrats work at all (p. 156). So, why do bureaucrats work? One explanationhas to do with the intrinsic motivation of bureaucrats for providing public goods. In this study, over 1000 subjects from the public and privatesectorsof Indonesia participated in a variation of the dictator game: they received a sum of money and were told that they could give as much of it as they wanted to the Indonesian Red Cross, keeping the rest for themselves. Those from the public sector donated significantly more to the Indonesian Red Cross than others. The study design enables us to address two central theoretical questions in the literature concerning the mechanism through which government employees come to be more pro-social. Two mechanisms are immediately apparent: (1) government employees are more pro-social because pro-social individuals select into the government sector (selection), or (2) service in the government sector reinforces pro-social behavior (socialization).Altogether, the evidence indicates that both socialization and selection lead to more pro-social behavior among government employees.
Author(s): Sheheryar Banuri; Phil Keefer
Topic: Field experiments, public goods, norms
Back to session: Field 5
ABSTRACT:
Index insurance has recently emerged as a potential alternative to conventional crop insurance. With relatively few exceptions, however, uptake rates have been disappointingly low across the developing world. One hypothesis for low index insurance uptake is that people do not fully behave as predicted by Expected Utility Theory (EUT,) with implications for their decisions not only about index insurance, but also about individual crop insurance. We tested the link between Cumulative Prospect Theory (CPT) and insurance choices by conducting two artefactual experimental games with 480 cotton farmers in Peru. Both games consisted of decisions which involved, first prospects over gains only, and then mixed prospect (gains and losses.) The first series of decisions were framed as insurance purchase over a stochastic farm income. The second series of decisions were unframed decisions over lotteries, so as to determine each farmer's three individual CPT parameters. We find heterogeneity in CPT parameters, indicative of probability weighing and loss aversion. Furthermore, over 20% of farmers changed their insurance choice when the game was framed over losses. Results over losses are consistent with risk seeking behavior over losses, which is predicted by CPT. However, results over gains contradict both CPT and EUT standard predictions: On average, the more risk averse a farmer the less likely they were to chose insurance. Similar results have been observed empirically elsewhere and one possible behavioral explanation is that insurance constitutes an unfamiliar, therefore more risky proposition.
Author(s): Jean Paul Petraud with Steve Boucher and Michael Carter, UC Davis.
Topic: Field experiments, risky choice
Back to session: Field 5
ABSTRACT:
This paper analyzes deterrence schemes and their impact on stealing from victims by a series of laboratory and field experiments with students and inmates. Results confirm Becker's deterrence theory and substitutability of punishment size and probability. Analysis of regime change shows that incentive history matters: when deterrence incentives are removed subjects steal more than without this history. Crowding out of pro-social behavior is a common explanation for this result. This paper offers evidence that (at least part of this) crowding out takes place via change of emotionally motivated behavior. Without deterrence incentives in place, a variant of the dictator game, criminals with pro-social emotions steal less from victims. When criminals face expected punishment pro-social emotions do not decrease stealing; in this case self-centered emotions motivate more stealing. These findings support theory that incentives both directly and indirectly impact behavior via rational and emotional motivation channels. Hence, this paper provides support for recent theory on behavioral criminal law and economics and offers new insights for deterrence policy.
Author(s): Menusch Khadjavi
Topic: Field Experiments, social preferences, IO
Back to session: Field 5
ABSTRACT:
To what extent do international interventions have transformative effects on cooperative norms in ethnically divided societies? This research considers the sustaining impact of international peacekeeping forces on monitoring and enforcement of basic norms of inter-ethnic cooperation following NATO's intervention and subsequent decade-long peacekeeping mission in Kosovo. Our approach involves a novel experimental technique for measuring effects of third-party monitoring and enforcement on inter-ethnic trust. We then consider a natural experiment of the effects of different peacekeeping intervention regimes (French KFOR monitoring vs British KFOR enforcement) on inter-ethnic norms of fairness. Finally, we address external validity with cross-sectional and longitudinal observational data from the past two decades. Our results show that while intervention can have short-term positive effects on social cooperation, trust quickly deteriorates once monitoring and sanctioning mechanisms are removed. This leads us to speculate that international peacekeeping has only limited positive effects on the ability of adversaries to reconcile after violence, and that those marginal effects dissipate when peacekeepers withdraw.
Author(s): Vera Mironova ; Sam Whitt
Topic: Field Experiments, social preferences, norms and fairness
Back to session: Field 5
ABSTRACT:
We study a two-period experimental market, in which traders receive private information about the dividend in the first period and a public announcement in the second. We vary the public signal quality by changing its precision and the information asymmetry among traders by varying the distribution of private signals. We find that high quality public signal increases price efficiency and reduces underpricing in the pre-disclosure period in our high asymmetry treatment. The impending disclosure increases the competition among informed traders, which leads prices to impound more private information. As the uninformed traders extract more information from prices, they are more willing to hold shares.
Author(s): Orie Barron and Hong Qu
Topic: Finance
Back to session: Finance 1
ABSTRACT:
In order to understand the mechanisms behind bank run contagions, we conduct a bank run experiment in a modified Diamond-Dybvig setup with two banks (Left and Right). We consider two treatments where the banks' liquidity levels are either linked or independent. Left Bank depositors see their bank's liquidity level and make their withdrawal decisions first. Right Bank depositors only see Left Bank withdrawals before deciding. The Right Bank depositors' expectations about their bank's liquidity level affect the likelihood of a run on the Right Bank, more so when bank liquidities are independent of each other. However, actions taken by depositors in Left Bank significantly affect depositor behavior in the Right Bank, even when both bank liquidities are independent. Furthermore, a panic may be a one-way street: an increase in Left Bank withdrawals can cause a panic run on the Right Bank, but a decrease cannot calm markets.
Author(s): Surajeet Chakravarty, Miguel A. Fonseca and Todd R. Kaplan
Topic: Finance
Back to session: Finance 1
ABSTRACT:
Saving more in the face of the ability to spend is an unpleasant but not conceptually difficult task. A popular “debt snowball” approach to debt repayment contradicts traditional economic theory by suggesting that those in debt should pay off their debts from smallest size to largest regardless of interest rate, in order to realize quick motivational gains from eliminating debts. Using a laboratory computer task, we break down this “debt snowball” approach into component parts and examine their efficacy. We find that when a mildly unpleasant computer task is broken down into parts of unequal size, subjects complete these tasks faster than when they are arranged in ascending order rather than descending order. Participants behave similarly to the baseline task when forced to complete tasks in a “minimum payment” condition. Adding “interest” to these conditions weakens these effects. We discuss and model alternative behavioral theories that can explain these results.
Author(s): Alexander L. Brown, Joanna N. Lahey
Topic: Finance, cognition
Back to session: Finance 1
ABSTRACT:
The prevalence, causes, and appropriate policy response to price bubbles in asset markets are widely debated. We conducted experiments in which subjects traded an asset with a constant fundamental value. Bubbles are common and have some regularity, reaching 3-7 times the fundamental value before reliably crashing. Bubbles seem to emerge from a mixture of trader types. We fit a structural model that classifies subjects according to their estimated beliefs about the likely termination time of the bubble, and we explore connections between this model and existing approaches to modeling strategic sophistication in static games.
Author(s): Alec Smith
Topic: Finance, markets, elicitation
Back to session: Finance 1
ABSTRACT:
Experimental beauty contests are a test of subjects' ability to think iteratively about others' beliefs and actions. In a typical version of the game, each player in a group of participants must choose a number between 0 and 100, and whoever comes closest to 2/3 of the average choice wins a prize. In theory, iterated reasoning should lead players to choose 0, and participants usually approach this equilibrium after playing the game multiple times. Though based on Keynes' famous metaphor likening the stock market to a newspaper beauty contest, this task bears little resemblance to an actual stock market scenario. We examine whether framing the game explicitly in a stock context changes subjects' decisions. We find that a simple rephrasing of instructions to describe the number decision as the choice of a stock price increases the frequency of dominated choices and hinders convergence to 0 by the end of a 10-round session. Particularly in the stock context, convergence is further hurt by a payment scheme designed to create bubbles while leaving the equilibrium unchanged.
Author(s): Lauren Feiler and Sam Kalar
Topic: Finance, games
Back to session: Finance 2
ABSTRACT:
We revisit the effect of traders' experience on price bubbles by introducing either one-third or two-thirds steady inflow of new traders in the repeated experimental asset markets. We find that bubbles are not significantly abated by the third repetition of the market with the inflow of new traders. The relative importance of experience to the formation of bubbles depends on the proportion of new traders in the market. Our findings identify a market environment where increased experience is not sufficient to eliminate price bubbles.
Author(s): Huan Xie and Jipeng Zhang
Topic: Finance, learning
Back to session: Finance 2
ABSTRACT:
We report the results of an experiment designed to study the effect of asset-holdings caps on the formation of bubbles and crashes in laboratory asset markets. Bubbles and crashes are a quite robust phenomenon in experimental settings and the only factor appearing to reduce bubbles is experience across markets. Motivated by concentration control policies employed in the Chinese real-estate market, we explore the effects of permanent and short-term caps on individual asset holdings. We find that permanent caps greatly reduce positive bubbles, but tend to generate negative bubbles in later periods. Under short-term cap, on the other hand, we observe no negative bubble in later periods. Furthermore, there are no positive bubbles in six out of six sessions. Our results indicate that concentration control policies can be effective in eliminating bubbles if properly designed.
Author(s): Volodymyr Lugovskyy, Daniela Puzzello, Steven Tucker, Arlington Williams
Topic: Finance, markets
Back to session: Finance 2
ABSTRACT:
Short-term credit was at the center of the recent crisis, yet very little work exists on this subject. Inspired on He and Xiong (2012), we present a web-based continuous time experiment in which groups of subjects extend (asset-backed) credits to a firm, whose fundamental value follows a stochastic process. The only decision that subjects make is whether or not to roll over their credit. Due to the asynchronous maturity of the credits in each group, results converge to an equilibrium where there are runs on firms whose fundamentals are sound (i.e.they can pay back all subjects); a result that could never occur in the classic static bank run models. In addition, we test the effect that the length of contracts might on different threshold strategies and whether shorter contract lengths might help prevent credit freezes. Finally we analyze the learning dynamics in our experiment, discovering what we call boomerang learning.
Author(s): Ciril Bosch-Rosa
Topic: Finance, markets
Back to session: Finance 2
ABSTRACT:
We present an agent-based model of experimental asset markets in which, in addition to momentum-noise traders as in Duffy and Unver (2006), we add adaptive fundamental traders. We use experimental data to identify players' types: 20-35% of our subjects behave like adaptive fundamental traders. We also elicit traders predictions of the next period's price and collect their responses to the standard Frederick (2005) cognitive reflection test. We find that adaptive fundamental traders perform significantly better in terms of price forecasts, cognitive reflection test, and terminal wealth. More importantly, we find that fundamental traders initially set the trend for bubble formation, and later initiate the bubble crash.
Author(s): Sascha Baghestanian, Volodymyr Lugovskyy and Daniela Puzzello
Topic: Finance, markets
Back to session: Finance 3
ABSTRACT:
This paper presents an experimental investigation of a financing-investment environment under asymmetric information. It examines the underpricing-signaling hypothesis. Importantly, the paper tests and compares this hypothesis under the two institutions for financing offers that are commonly observed in corporate financial markets: take-it-or-leave-it offer (TLO) and the competitive bidding offer (CBO). The results suggest that underpricing can serve as a credible signal of quality under certain parameters. The underpricing is lower under CBO than under TLO institution. The use of experimental methods contributes an empirical perspective on the role of underpricing as signaling and the choice of institution in financial markets.
Author(s): Angelina Christie and Daniel Houser
Topic: Finance, markets
Back to session: Finance 3
ABSTRACT:
This computerized experiment immerses students in a financial market. Students learn the basic elements of bank balance sheets, the tradeoffs a bank makes when it hedges against liquidity risk, and the macroeconomic implications of the network aspects of the banking system. Student teams act as bank managers. Banks are distributed over islands, and each island is divided into regions. Banks within the same island are linked by an interbank overnight funds market. Each bank attempts to maximize cumulative profit (bank equity) by managing their balance sheets over multiple rounds. Each round is divided into three periods as in the standard Diamond and Dybvig model. In period one banks allocate assets between cash, deposits with other banks, and corporate loans. Banks are faced with the exogenous risk in period two of depositors withdrawing a larger than anticipated percentage of their funds in response to a local regional recession. Because banks that are forced to liquidate their corporate loans incur a very large discount there is an incentive for banks in adjacent regions to coordinate and provide each other with overnight funds. These funds act as a form of mutual insurance. In the event that banks in one region experience a run by depositors since the overnight deposits by banks in the neighboring region can serve as a reserve buffer of liquidity. This experiment differs from previous classroom bank run experiments because students act as banks, not depositors, and also because banks are not isolated but connected via the interbank funds market. Early trials show that a team's performance is positively correlated with its success in coordinating interbank deposits with other banks. Another result is that teams that have incurred serious losses tend to double down, making bigger bets as the experiment proceeds. The experiment is written in the open source Python language, parameters are chosen by the instructor, and all results can be saved as a spreadsheet data file.
Author(s): John Alevy and Paul Johnson
Topic: Finance, markets
Back to session: Finance 3
ABSTRACT:
Most theories of risky choice postulate that a decision maker maximizes the expectation of a Bernoulli (or utility or similar) function. We tour 60 years of empirical search and conclude that no such functions have yet been found that are useful for out-of-sample prediction. Nor do we find practical applications of Bernoulli functions in major risk-based industries such as finance, insurance and gambling. We sketch an alternative approach to modeling risky choice that focuses on potentially observable opportunities rather than on unobservable Bernoulli functions.
Author(s): Daniel Friedman, Shyam Sunder
Topic: expected utility, risk aversion, St. Petersburg Paradox, decisions under uncertainty, option theory
Back to session: Finance 3
ABSTRACT:
Research in psychology, behavioral economics and neuro-economics suggests that emotions such as fear and anger can play an important role in decision-making. These emotions are likely to be a major factor in the public response to terrorist events. In our study, different techniques including a financial loss and video with a 110 dB explosion were employed in a lab setting to induce emotional responses. Participants were asked to vote for or against a tax on all participants to reduce the probability of an event. In some treatments, participants were also provided a chance to purchase private insurance. In all treatments, if a majority of participants vote in favor of the tax, all pay the tax and the probability of an event is reduced by 50%. Those who purchased private insurance prevented themselves from suffering the financial loss but they were still exposed to the other stimuli depending on specific treatments.
Author(s): Jubo Yan, Kevin Kniffen, William Schulze, Brian Wansink (Cornell University)
Topic: Emotions and choice
Back to session: Games 2
ABSTRACT:
Cheap talk is defined as costless communication that is nonbinding and unverifiable. In the context of a game with complete information, cheap talk should have no effect on the players' decisions. Yet, experimental data show that cheap talk can affect equilibrium selection in coordination games, and motivate cooperation in prisoner dilemmas' games. In a sense, there seems to be a consensus that cheap talk is a necessary condition for selecting socially efficient equilibria in single shot games. In this study, we utilized fMRI technology to identify the mechanisms whereby cheap talk affects decision-making. Our experiments consisted of a series of simultaneous games with one-way communication to analyze which kinds of messages were more likely to resolve coordination failures and social dilemmas. We hypothesize that the messages sent encompasses more information than just the intentions of the sender. Along with sender's intentions, messages contain information of sender's social type, which allows socially efficient equilibria. Preliminary behavioral and neuroimaging data is presented to support our hypothesis.
Author(s): W. Gavin Ekins, Monica Capra, Gregory S. Berns
Topic: Games
Back to session: Games 2
ABSTRACT:
This paper studies pre-commitment investment strategy in a one shot game, where agents seek to form matches in the presence of an assortative matching rule. We show that a bimodal distribution of investment arises in equilibrium where most players select high levels of investment — achieving a pareto superior solution — meanwhile few stay at the lower bound — the trivial NE. The experimental evidence obtained supports our predictions and shows that the median investment levels are remarkably high, close to 91 percent of the initial endowment. This result is novel as one shot games have generally been unable to produce such high levels of cooperation (ef?ciency).
Author(s): Jean Paul Rabanal and Olga A. Rabanal
Topic: Games, bargaining, IO
Back to session: Games 2
ABSTRACT:
Understanding the problem of strategic communication inherent in various political decision-making environments (e.g. between committees and chambers, legislatures and executives, voters and politicians) has provided key insights about institutions and policy-making. Much of the theoretical framework, however, is built upon models of the bilateral interaction between a single expert and a single decision-maker. The effects of competition between multiple information sources or opposing interests have largely been ignored. To remedy this, we conduct a series of experiments to investigate the effects of competition on strategic information transmission. Our experiment tests the predictions of Krishna and Morgan's (2001) theoretical analysis of an extension of the seminal Crawford and Sobel (1982) cheap talk game with two senders. We vary two features of the experimental environment that the theory predicts should affect the amount of information transmitted: the number of senders and the degree (and direction) of preference divergence between the sender(s) and receiver. Our design also features a richer action/state/message space than previous strategic communication experiments.
Author(s): Jonathan Woon
Topic: Games
Back to session: Games 2
ABSTRACT:
This paper studies endogenous termination within the laboratory, examining a repeated prisoner's dilemma with imperfect public monitoring. By allowing for the possibility of termination we examine how contracts that specify how the partnership surplus is divided affect cooperation within the partnership. We vary both how much of the partnership value is returned to the players and how this amount is divided between them. Our experiment answers the following questions: i) Does the threat to walk away from a joint project lead to higher or lower cooperation within the partnership? ii) Do human subjects prefer cooperative strategies supported by termination punishments as opposed to within partnership punishment? iii) In what situations does the possibility for termination increase/decrease welfare? Preliminary results indicate that contracts that destroy partnership value might be welfare enhancing through greater cooperation, as termination becomes inefficiently relied on as the partnership value returned increases.
Author(s): Hong Wu/Alistair J.Wilson
Topic: Games, cooperation
Back to session: Games 3
ABSTRACT:
Backward induction in the finitely repeated prisoner's dilemma has been previously studied in the laboratory. However, the answer to whether subjects perform backward induction remains inconclusive. Some studies find increases in unraveling of cooperation with experience, while others find that first defection occurs later on in the interaction as subjects gain experience. Understanding this phenomenon is made difficult by the fact that studies vary in the number of repetitions and horizon of the finitely repeated game, as well as in the parameters of the stage game. An additional problem is that most payoff parameters that have been studied are in a relatively small range. Furthermore, many of the prior studies use a within subject design when varying the horizon and the stage game, making the impact of experience unclear. In this paper, we conduct a meta-analysis of prior experimental research on the topic as well as present results of a new experiment to elucidate how the ability to perform backward induction varies with the environment in this canonical game.
Author(s): Matthew Embrey, Guillaume Frechette, Sevgi Yuksel
Topic: Games, cooperation, cognition
Back to session: Games 3
ABSTRACT:
Entities often engage in interdependent activities that may have an effect on other entities (e.g., airport security, health information exchanges). We modify the prisoner's dilemma for our interdependent setting in the lab, with two or three players in each interdependent group. We implement a payoff structure using expected values and test several transfer pricing mechanisms: no transfer pricing; endogenously-determined transfer pricing; and exogenously-determined transfer pricing. We draw conclusions about the ability for subjects to achieve the socially-optimal level of collaboration, and also conduct a behavioral analysis on the consequences of expected values on decision-making.
Author(s): Matthew J. Hashim
Topic: Games, IO
Back to session: Games 3
ABSTRACT:
One of the fundamental results in repeated games is the Folk theorem, which predicts a plethora of equilibrium outcomes. Many have argued that this extreme number of equilibria is a virtue, as it can explain a variety of different behaviors. However, this result leaves us with almost no predictive power. This paper provides measures for evaluating the predictive power of a theory given experimental data. After running experiments with human subjects in the experimental laboratory, we use these measures to compare a variety of equilibrium theory models as well as several learning models. The models examined include Mathevet (2012)'s axiomatic approach, and Ioannou and Romero (2012)'s learning model.
Author(s): Laurent Mathevet and Julian Romero
Topic: Games, learning
Back to session: Games 3
ABSTRACT:
Laboratory experiments have found that that women are less likely to choose competitive payoffs, less confident, more risk averse, and more inequity averse than men. However, there is no evidence that these characteristics measured in the lab are related to real life economic outcomes. This project links laboratory results to actual labor market experiences. We test the hypothesis that competitiveness, confidence, risk aversion, social preferences, and gendered personality characteristics affect labor market outcomes and contribute to the gender gap in labor market performance. We run laboratory experiments designed to measure these characteristics and then survey students on employment, salary, and other labor market outcomes both at graduation and 7 months after.From these data we can determine whether personality attributes and behaviors measured in the lab affect early labor market success and if they explain differences in outcomes for men and women. We extend the literature by developing a continuous measure of willingness to compete and, because people choose into professions, allowing the subjects to choose the type of task performed. The experiments identify three types of confidence: estimation, placement, and precision. Gendered personality characteristics are measured by the Bem Sex Role Inventory.Obtaining a clearer picture of the causes of gender differences in labor market outcomes is important because policy prescriptions to address gender inequities must be based on a correct understanding of the source of such differences.
Author(s): Linda Kamas and Anne Preston
Topic: Gender
Back to session: Gender 1
ABSTRACT:
We use a laboratory experiment to test Du & Wei (2010)'s Sexually Unbalanced Model of Current Account Imbalancesö. We control for the sex ratio of a group and let subjects decides how much to save before entering the marriage market, where one to one matches can be made and matched couples share their savings as public goods with congestion. We do observe higher savings rate on the long side when there is a bigger surplus of one gender, until the sex ratio reaches a very high level, when some subjects on the long side decide to give up competition. Using two other treatments with partial or no information on saving before matching, we try to distinguish competition effect and hedging effect.
Author(s): Hanming Fang, Binglin Gong, Shangjin Wei
Topic: Gender
Back to session: Gender 1
ABSTRACT:
Following the seminal work of Niederle and Vesterlund (2007), a large literature has evolved documenting gender differences in competitiveness, with women being less likely to compete in a tournament setting in the lab. How these gender differences in competitiveness map into differences in educational and career choices are, however, less understood. To address this question, we combine rich survey data on undergraduate students' choice of college major and expectations of future career-specific outcomes, with a modified version of the setup in Niederle and Vesterlund (2007).
Author(s): Ernesto Reuben, Matthew Wiswall, and Basit Zafar
Topic: Gender
Back to session: Gender 1
ABSTRACT:
There is a consensus that there is a gap between male and female wages. This paper investigates whether this is due to a negotiation gap, and what mechanisms can alleviate it, by examining whether females ask for less in a controlled bargaining setting that accurately models relevant aspects of negotiating over a starting salary. The results are stark: females ask for less, and earn less than males. Providing social information eliminates the negotiation gap, and more importantly, the wage gap.
Author(s): Mary L. Rigdon
Topic: Gender
Back to session: Gender 1
ABSTRACT:
We examine a new intervention to overcome gender biases in hiring, promotion, and job assignments: an evaluation nudge, in which people are evaluated jointly rather than separately regarding their future performance. Evaluators are more likely to focus on individual performance in joint than in separate evaluation and on group stereotypes in separate than in joint evaluation, making joint evaluation the money-maximizing evaluation procedure. Our findings are compatible with a behavioral model of information processing and with the System 1/System 2 distinction in behavioral decision research where people have two distinct modes of thinking that are activated under certain conditions.
Author(s): Iris Bohnet; Alexandra van Geen; Max Bazerman
Topic: Gender, labor, cognition
Link: http://www.google.com/url?sa=t&rct=j&q=bohnet%20van%20geen&source=web&cd=1&ved=0CCIQFjAA&url=http%3A%2F%2Fweb.hks.harvard.edu%2Fpublications%2FgetFile.aspx%3FId%3D769&ei=XIdcUI7pCKft0gGisIDYCg&usg=AFQjCNG0R6vWevpKIldTZN7KbI7MLR603Q
Back to session: Gender 2
ABSTRACT:
In many high income countries, a substantial gender gap in favor of women is emerging in human capital investment. Potential explanations include gender differences in non-cognitive ability in youth and gender segregation across occupations and industries. We present experimental evidence of an additional explanation: women may place a greater value on formal training as a means of learning. In an online labor market, we pay subjects a piece rate for completing an unfamiliar task and elicit willingness to pay for task-specific training. We find that when subjects have limited experience with the task, women are more willing to pay for training than men, and, ex-post, women estimate that the returns to training are higher than men do. These gender differences are significant despite there being no difference between men and women in task performance without training, and randomly-allocated mandatory training increasing men's performance more than women's. In a follow-up experiment, we explore how training may impact preferences for a challenge.
Author(s): Melissa Adelman, Katherine Baldiga
Topic: Gender, labor, field experiment, internet
Back to session: Gender 2
ABSTRACT:
As part of a longitudinal study of Peruvian children, an incentivized Binswanger (1981)-type risk question was asked to a panel of 1,943 seven and eight year-old boys and girls. Consistent with results from previous studies on older children and adults, girls are significantly more risk averse than boys. These differences are persistent even with the inclusion of controls for weight and height and a variety of measures of cognitive ability at different ages. They also persist with controls for family structure, parental background and socio-economic status. While welfare and intellectual development do differ by gender, these do not explain why boys and girls have different experimental risk measurements. Gender is not the only factor that affects risk attitudes. Children of better-off families are more risk averse, and the behavior of boys and girls responds to different factors. In the first incentivized representative study of children's risk preferences, gender differences in risk attitudes appear at an early age.
Author(s): Marco Castillo
Topic: Gender, risky choice
Back to session: Gender 2
ABSTRACT:
In a gift exchange experiment, higher wages are offered by men compared to women, by agents with greater cognitive ability, and by subjects who score higher on agreeableness on the Big Five personality scale. In turn men provide greater effort than women and a one standard deviation increase in agreeableness generates almost the same increase in effort as a comparable increase in wages for both genders. Further, conscientiousness plays a statistically and economically significant role in wage rates offered and effort levels provided, but the sign of this effect differs between men and women.
Author(s): Emel Feliz-Ozbay, John Ham, John Kagel, Erkut Ozbay
Topic: Labor, Gender
Back to session: Gender 2
ABSTRACT:
While experiments have been an integral part of teaching the natural sciences for centuries, they haven't gained similar traction in social science classrooms, largely due to lack of accessible, scalable, and robust platforms, both hardware and software, for such real time interactions. However, with the exponential growth of Internet and wireless technologies for the masses, time is ripe to overcome the previously insurmountable barriers. At MobLab, we are developing a full suite of popular interactive games, such as many variations on bargaining, markets, auctions, voting, matrix games, etc., for use in economics, political science, finance, and management classes. In addition to web browsers, all of the games can be run natively and concurrently on iPad, iPhone, and Android phones in either standalone or networked mode. A preliminary suite has been developed and are being piloted at several universities this Fall. Supplementary materials for instructors and real-time analysis of data are included features. We will demonstrate the intuitive and flexible console for instructors to configure, run, and analyze the data in a matter of minutes. Preliminary classroom pilot results will be discussed.
Author(s): Walter Yuan, Stephanie Wang, and Thomas Palfrey
Topic: Methodology, internet
Link: http://www.moblab.com
Back to session: Internet
ABSTRACT:
This experiment investigates how groups form institutional rules governing access to and the exploitation of a set of common-pool goods. By using the virtual world of Second Life, we designed a three-dimensional spatial setting where subjects in command of representative avatars were free to explore, gather resources, communicate, monitor, and punish one another. Within a session, subjects are free to use these capabilities to generate institutional mechanisms that aim to prevent the over-exploitation of common resources. This experiment is particularly concerned with investigating how different sorts of environmental parameters such as having goods that subjects can specialize in collecting or having goods that are spatially-concentrated in a more-focal manner modulate both the propensity for groups to establish different institutional rules and the effectiveness of those rules in staving off the tragedy of the commons. Preliminary data analysis shows that there is a large amount of variance in these variables of interest within our different environmental treatment conditions.
Author(s): Kevin McCabe, Peter Twieg
Topic: Public goods, internet
Back to session: Internet
ABSTRACT:
The Internet is becoming increasingly appealing for experimentalists, both as a medium through which to target more diverse samples and run experiments at a very large scale, but also as a field of social and economic research in its own right. Given those appealing features, it is surprising that the online laboratory has not yet been developed. This paper reports on the construction of an online platform specifically dedicated to conducting social experiments over the Internet. We use this online experimental economics platform to conduct an evaluation of the Internet-based methodology. Using the same subject pool, the same monetary stakes and the same decision interface over the Internet and in a traditional University laboratory, we randomly assign two groups of subjects between both conditions and compare their behavior in a set of social preferences games. Our results contradict the predictions of social distance theory, as we find that subjects allocated to the Internet treatment behave as if they were more altruistic, more trusting, more trustworthy and less risk averse than laboratory subjects. Those findings have practical importance for the growing community of researchers interested in using the Internet as a vehicle for social experiments and bear interesting methodological lessons for social scientists interested in using experiments to research the Internet as a field.
Author(s): Algan, Y., Hergueux, J. & Jacquemet, N.
Topic: Social preferences, internet
Back to session: Internet
ABSTRACT:
We investigate the effects of lending team competition on pro-social lending on Kiva.org, the first peer-to-peer microlending site which matches citizen lenders with entrepreneurs in developing countries. We find that lenders are more similar to their team members than to those outside the team. When choosing a team, lenders pay more attention to team rank than to member characteristics. Furthermore, joining a lending team increases lending, but mostly for those who are not already engaged in the site. Lenders who join teams make 1.8 more loans per month than their counterparts who do not join teams, which translates to at least $45 per month. Next, using a random sample of team forum data, we find that teams who post many links to specific loans increase their lending level. Finally, we find that team competition plays a significant role in lending behavior.
Author(s): Roy Chen, Yan Chen, Yang Liu, Qiaozhu Mei
Topic: Internet, finance
Back to session: Internet
ABSTRACT:
Affirmative action is a subject of intense and polarized debate (Cohen and Sterba, 2003). Supporters point to the opportunities to address historical and statistical discrimination, and to the advantages of diversity both in the workplace and in the classroom. Critics contend that affirmative action is reverse discrimination Newton (1973)and violates the principle of merit Walzer (1983, : 143-154). Quite apart from moral concerns, affirmative action can lead to economic inefficiencies. For instance, if employers are forced to lower their hiring standards then the intended beneficiaries of affirmative action may be unable to compete with their peers, and may be worse off as a result Sowell (1990).
Author(s): Marcela Ibanez, Gerhard Riener, Ashok Rai
Topic: Labor
Back to session: Labor 1
ABSTRACT:
Using a gift exchange experiment, we show that the ability of reciprocity to overcome incentive problems inherent in the principal-agent settings is greatly reduced when the agent's effort is distorted by random shocks and transmitted imperfectly to the principal. Specifically, we find that gift exchange contracts without shocks encourage effort and wages well above standard predictions. However, the introduction of random shocks reduces wages and effort, regardless of whether the shocks can be observed by the principal. Moreover, the introduction of shocks significantly reduces the probability of fulfilling the contract by the agent, the payoff of the principal, as well as the total welfare.
Author(s): Jared Rubin and Roman Sheremeta
Topic: Reciprocity, labor
Back to session: Labor 1
ABSTRACT:
We propose a novel approach to the analysis of organizations by developing a computerized platform that reproduces relevant features of existing organizations such as real-effort tasks and real-leisure alternative activities (Internet). In this environment, we find strong incentives effects as organizations using individual incentives significantly outperform those relying on team incentives. Combining real-time peer monitoring with team incentives, we report striking evidence of positive peer effects as production increases by 50% and Internet usage decreases by 54% compared with organizations using team incentives alone. Peer monitoring allows virtual organizations using team incentives to perform as well as those using individual incentives. However, the positive effect of peer monitoring does not apply to low performers.
Author(s): Brice Corgnet, Roberto Hernan and Stephen Rassenti
Topic: Labor
Link: http://ideas.repec.org/p/chu/wpaper/11-05.html
Back to session: Labor 1
ABSTRACT:
This paper investigates how compensation structure affects the performance of people that are engaged in a simple but effort-consuming task. In this experimental study the subjects, recruited from a pool of Undergraduate students, were asked to multiply 2-digit numbers for 40 minutes without the use of a calculator. In total, 231 have participated in 6 different treatments. In three treatments subjects payments linearly depend on the number of correctly computed multiplications. In the first treatment subjects were paid $5 show-up fee plus $0.15 for each correctly computed multiplication. In the second treatment they we paid $15 +$0.05 per multiplication. In the third - $18+$0.02. In the forth treatment subjects were paid a flat wage of $20. In the fifth treatment the subject were paid using option-based compensation scheme: they were paid $5 show-up fee plus $0.60 for each correct multiplication in excess of 90. In the final treatment subjects were able to chose between the option-based and linear ($0.15 per multiplication) payment structure. The parameters of the model were chooses such that on average subjects were paid about $20 and in the option-based treatment both the average and median expected payment was $20.
Author(s): Anna Dodonova and Yuri Khoroshilov
Topic: Labor
Back to session: Labor 1
ABSTRACT:
Economic outcomes, especially wages, differ substantially with individual personality traits. However, the influence of other people's personality on one's economic decision remains unclear. We designed a worker-employer experiment where similarly skilled workers performed the same task and earnings were determined solely by employers. Our results suggest that personality of the worker influences decisions of the manager, and personality of the manager influences decisions of the worker. Moreover, we find that such effects are not completely conscious.
Author(s): Piotr Evdokimov and David Rahman
Topic: Labor
Link: http://www.econ.umn.edu/~evdok003/evdokimov.pdf
Back to session: Labor 2
ABSTRACT:
Jovanovic (1979) studies the problem of optimal assignment of workers to jobs. Jovanovic recognizes the existence of imperfect information on both sides of the market about optimal assignment. Following an initial assignmen
Author(s): Marina Agranov, Chloe Tergiman
Topic: Labor
Back to session: Labor 2
ABSTRACT:
As the workforce becomes increasingly diverse, motivating individuals from different backgrounds to work together effectively is a major challenge facing organizations. In an experiment conducted at a large public university in the U.S., we manipulate the salience of participants' multidimensional natural identities and investigate the effects of identity on coordination and cooperation in a series of coordination games and prisoner's dilemma games. We find that in the coordination games priming a fragmenting (ethnic) identity results in a significant decrease in the choices of the Pareto dominant Nash outcome. In the prisoner's dilemma games, ethnic prime makes Asian participants exhibit significantly more ingroup cooperation and outgroup discrimination. School priming enhances the negative effects of the competitiveness stereotype on cooperation among Asian participants.
Author(s): Sherry Li
Topic: Labor, cooperation, norms
Back to session: Labor 2
ABSTRACT:
Many Internet users reveal online personal information that employers are not supposed to use during the hiring process. Self-reported surveys of employers suggest that US firms have, in fact, started using popular online social networks to seek information about prospective hires.. However, no randomized experiment has yet measured the extent to which firms use online social networks to screen job applicants, and how their hiring activities are influenced by the information they find. We use two IRB-approved randomized experiments to investigate the effects of job candidates' personal information, posted on an online social network, on the search activities of employers.
Author(s): Alessandro Acquisti and Christina Fong
Topic: Labor, field experiments, internet
Back to session: Labor 2
ABSTRACT:
We predict threshold strategies in a 2x2 global coordination game by analyzing the response times associated to signals and actions. In a global game subjects observe noisy private signals about a state variable that, together with the action of the other player, determines payoffs. An optimal threshold is defined as the signal for which a subject is indifferent between taking any one of the two actions. In terms of response times, this would imply longer consideration periods for signals closer to individual thresholds. We estimate 'choice based' thresholds by using standard methods with the observed choices in the last rounds of the experiment, once behavior is stable. When analyzing response times, we look at the initial rounds of the experiment, when subjects are building a strategy, and find that the signal associated to the highest or second highest response time is a better predictor of the threshold that subjects exhibit in later periods than the theoretical equilibrium. This result also allows us to characterize two different patterns of learning across subjects.
Author(s): Andrew Schotter and Isabel Trevino
Topic: Cognition, learning
Back to session: Learning
ABSTRACT:
The formation of beliefs is a key element in virtually all areas of economic decision-making. A common feature of many information structures is that incoming signals from different sources are not independent. Using two different experimental designs, we present systematic evidence for correlation neglect in belief formation. The propensity to fall prey to the bias is highly heterogeneous and significantly associated with cognitive ability. Neglecting correlations in the belief formation process has significant economic implications. In particular, it systematically produces overshooting beliefs, i.e. overly optimistic and overly pessimistic beliefs, respectively. In an additional study we show that in a double auction setting, correlation neglect predictably drives price bubbles.
Author(s): Benjamin Enke and Florian Zimmermann
Topic: Learning, anomalies, finance
Back to session: Learning
ABSTRACT:
This paper discusses an experiment designed to test which model of error-prone best-responses most accurately predicts long run behaviour in large populations, i.e. what is the correct noisy population dynamic to use when invoking the equilibrium selection technique of stochastic stability (Foster and Young, 1990). The game the subjects play is the Language Game of Neary (2012a), in which different error models can predict different long run outcomes. We find that the best-reply dynamic with uniform-errors outperforms any dynamic with one randomly selected player best-responding each period, whether errors are either logit or uniform.
Author(s): Wooyoung Lim, Philip R. Neary
Topic: Learning, games
Back to session: Learning
ABSTRACT:
We characterize and identify properties of individual opinion formation and revision in social networks via an experiment. The experiment investigates social learning within a finite set of agents organized in a network and involved in a prediction task over a finite set of alternatives. The opinion-revision behavior of subjects is characterized in terms of the following properties: structural and identity independence, monotonicity and consistency. The properties link the experimental evidence to a model of non-Bayesian learning in social networks and to its implications in terms of achievement of consensus. The experimental design that we propose is especially suitable for studying the role of lack of common knowledge of subjects' actions and of network structure.
Author(s): Claudia Neri, Manuel Mueller-Frank
Topic: Learning, networks
Back to session: Learning
ABSTRACT:
Coordination games with multiple Pareto-ranked equilibria feature player beliefs on other players’ strategies as crucial in determining outcomes. If players are to coordinate on efficient investment, they must place high subjective probability on others investing. While equilibrium analysis enforces common knowledge of beliefs, many settings are best modeled with uncertainty about beliefs. For example, heterogeneous societies may lack a shared understand- ing of the prevailing social norm. I design an experimental coordination game where higher-order beliefs and uncertainty about higher-order beliefs are induced. I find that players prefer to invest especially when common knowledge of an “optimistic” belief that players will invest prevails.
Author(s): Steven Bosworth
Topic: coordination
Link: http://www.pitt.edu/~sjb74/paper_2.pdf
Back to session: Level-K 2
ABSTRACT:
Our goal in this paper is to examine whether human factors in terms of cognitive capacity can explain the differences in subjects market performance. To do this, we measure subjects’ working memory capacity and let them trade in double auctions. We had two treatments. In treatment 1, subjects have to maximize their profits in a market where other traders are truthful bidding agents. In treatment 2, subjects have to compete with adaptive software agents. Because the “behavior” of subjects’opponents are controlled, there is no room for other factors such as the complex interaction among human’s strategies to get involved. Our auctions thus serves as a good environment to answer our question. Can cognitive capacity explain the differences in people’s market performance? Our results show that WMC is an important factor in predicting subjects’ performance in almost every market experiment. However, the only exception reminds us that WMC’s influence may not be universal. If making transaction is easy in the market, WMC’s contribution is either less in magnitude or even insignificant.Is cognitive capacity, measured as WMC, is really a decisive factor even when people can learn and gain experiences? We observe that learning does happen, but we don’t have evidence strong enough to indicate that it can eliminate the differences in performance resulting from the heterogeneity in innate capability. However, experiences in financial markets seem to be another crucial element when we try to figure out why some subjects were more successful than the others in a more realistic environment.
Author(s): Shu-Heng Chen, Ye-Rong Du, and Lee-Xieng Yang
Topic: Cognition, games
Back to session: Level-K 2
ABSTRACT:
Non-equilibrium behavioral game theory models have been useful in organizing experimental data in strategic decision-making studies. However, how can we know whether or not we are too stringent or too lenient when identifying subjects? In our experiment, we control subjects' beliefs by having subjects play against random participants, and then against their past selves as in Ivanov et al (2010). We adapt their paradigm to simple two-player guessing games of the form in Costa-Gomes and Crawford. We argue that players that best respond to their past selves indeed used deliberate decision-making rules in the first phase, and seek to describe all such subjects' initial behavior using existing (and possibly novel) behavioral game theory models. Additionally, for subjects with behavior we can purely describe from initial play against other subjects, we investigate whether or not we can correctly predict their behavior when they play against their past selves. We find that most subjects that best respond to their past behavior are those that we 'classify' (using the methods in Costa-Gomes and Crawford) in phase I, and that the classified subjects in phase I are correctly predicted in phase II. In addition, we investigate various methods of classifying subjects and find that as we classify more subjects, the fraction of classified subjects that are correctly predicted in phase II gets worse.
Author(s): Daniel E. Fragiadakis, Asen Ivanov, Daniel T. Knoepfle, Muriel Niederle
Topic: Games, level-k
Back to session: Level-K 2
ABSTRACT:
We report an experimental test of level-k theory, applied to three simple games with non-neutral frames – Coordination, Discoordination and Hide and Seek. Using the same frame for all three games, we derive hypotheses that apply across the games and are independent of prior assumptions about salience. Those hypotheses are not confirmed by our experimental results. Our findings contrast with previous research which has fitted parameterised level-k models to Hide and Seek data. We show that, as a theory-testing criterion, the existence of a plausible model that replicates the main patterns in these data has a high probability of false positives.
Author(s): Shaun Hargreaves, David Rojo Arjona, Robert Sugden
Topic: Games, cognition
Back to session: Level-K 2
ABSTRACT:
There is serious information asymmetry between brokers and buyers in resale housing market. Buyers often entrust brokers to search for housing. However, the brokers not provide agent services, but the intermediary services. Brokers' benefits objectives not completely consistent with the buyers'. Buyers feel the brokers are not completely reliable. In the case of disadvantageous information and not completely reliable brokers, buyers face the problem of how to game with brokers in order to protect their own interests. Usually, buyers are not transmitting personal information directly to the brokers and entrust several brokers to search for housing at the same time. However, it is still unclear that whether buyers' measures can really play a role and also the mechanism of action. This study sought to explore the effects of buyers' reserve price and number of brokers on housing search.
Author(s): Yang Zhang and Hong Zhang
Topic: Markets
Back to session: Markets 1
ABSTRACT:
In many online marketplaces, buyers and sellers can negotiate a product transaction. With the advent of socially embedded marketplaces (e.g., Facebook Marketplace) it was believed that such platforms may outperform traditional marketplaces that are not socially embedded and often anonymous (e.g. Craigslist). While the buyers and sellers may be more likely to complete a transaction through trusted relationships in a socially embedded marketplace, they may not be in a better position to bargain compared to anonymous networks. Our research primarily deals with endogenous formation of coalition structures through negotiations based on varied resources of individuals who may or may not know their social relationships with other negotiators. While the dilemma of a player in the public goods game is how much to contribute, the members of our experiment face the tradeoff between the optimal choice of coalition members versus incorporating members with whom they may have a social tie in the real world. The past literature on the impact of social ties on marketplace efficiency is sparse. Does information on social ties aid in the success of a negotiation? Is social information detrimental to marketplace efficiency? This research attempts to answer such questions using a variation of the ultimatum game and the weighted voting game with different social information revelation policies in an experimental setup.
Author(s): Rajiv Mukherjee (The University of Texas at Austin), Karthik Kannan (Purdue University), Anitesh Barua (The University of Texas at Austin)
Topic: Networks, bargaining, voting
Back to session: Markets 1
ABSTRACT:
We study how the design of market institutions influences price dynamics and equilibrium selection in a laboratory economy with multiple competitive equilibria. We use Gale's (Naval Research Logistics Quarterly, 1963) two good exchange economy.First, we replicate the experiments of Crockett, Oprea, and Plott (American Economic Review, 2011). When agents trade using a continuous double auction, as predicted by Tatonnement theory, prices diverge from an equitable interior equilibrium to one of two exterior equilibria. At these exterior equilibria, agents on one side of the market capture nearly all the gains from trade. Second, we introduce and test a market mechanism where agents submit demand schedules. Using this new mechanism, prices quickly converge to the interior equilibrium and the gains from trade are shared approximately equally.
Author(s): Jacob Goeree and Luke Lindsay
Topic: Markets
Back to session: Markets 1
ABSTRACT:
Attainment of rational expectations equilibria in asset markets calls for the price system to disseminate traders' private information to others. It is known that markets populated by asymmetrically-informed profit-motivated human traders can converge to rational expectations equilibria. This paper reports comparable market outcomes when human traders are replaced by boundedly-rational algorithmic agents who use a simple means- end heuristic. These algorithmic agents lack the capability to optimize; yet outcomes of markets populated by them converge near the equilibrium derived from optimization assumptions. These findings suggest that market structure is an important determinant of efficient aggregate level outcomes, and that care is necessary not to overstate the importance of human cognition and conscious optimization in such contexts
Author(s): Karim Jamal, Michael Maier, Shyam Sunder
Topic: Markets, Agent based modeling
Link: http://cowles.econ.yale.edu/P/cd/d18b/d1868.pdf
Back to session: Markets 1
ABSTRACT:
Performance and behavioral properties of markets in patents are investigated using a dynamic economic system design with inventive, intermediary trading and innovating activities in a laboratory environment. The study is a dual study of contract and mechanisms (rules of trade) treating the patent system as a trade system in technical ideas (technology). Prices and dynamic gains are recorded and analyzed as linear contracts (two-part tariff with a fixed fee and a royalty on revenues) are traded using three mechanisms and two levels of patent validity (probability that the rights are upheld in court) with broad uncertain dual patent values (for investment and blocking) for the three activities.
Author(s): Eskil Ullberg
Topic: Markets, IO
Link: http://gmu.academia.edu/EskilUllberg
Back to session: Markets 2
ABSTRACT:
We re-examine the findings of Fehr and Tyran (2008) that the speed of price adjustment to a nominal shock is dependent on whether firms' prices are strategic complements or strategic substitutes. We demonstrate that their original findings are well explained by focal points in their subjects' payoff space. In our revised experiments that control for focal points, prices in our complements and substitutes treatments adjust to the nominal shock symmetrically.
Author(s): Heng Sok and Abel Winn
Topic: price setting game, focal points
Back to session: Markets 2
ABSTRACT:
Monopolistic competition (MC) is widely used in international and macroeconomic frameworks. Experimental methods are increasingly being applied in these areas, thus making it important to establish a benchmark experimental design. The limited experimental evidence suggests firms interacting in MC environments set too high of a markup relative to theoretical predictions. This experiment compares different parameters of a MC market (number of firms, degree of substitution between varieties, and market information) to identify a design that leads to quicker convergence to the MC equilibrium price.
Author(s): Luba Petersen
Topic: Markets, IO, methodology
Back to session: Markets 2
ABSTRACT:
This paper reports findings from an experiment that implements the Lagos-Wright (2005) model of monetary exchange. We find that subjects generally avoid the autar- kic equilibrium of that model and make trading decisions consistent with the model’s monetary equilibrium. Aliprantis, Camera and Puzzello (ACP, 2007) show that pro- viding periodic access to centralized markets as in the Lagos and Wright framework may facilitate the sustainability of social norms of gift exchange, thus rendering money inessential in decentralized exchange. We also explore this hypothesis by replacing the centralized market of the Lagos-Wright model with a version of the centralized market of ACP’s model. We find that the essentiality of money is not threatened by the pres- ence of centralized meetings. Indeed, the efficiency of allocations is significantly higher in the environment with money than without money, suggesting that money plays a role as an efficiency enhancing coordination device.
Author(s): John Duffy and Daniela Puzzello
Topic: Markets, bargaining
Back to session: Markets 2
ABSTRACT:
We study a model with a positive production externality in which the productivity is a non- decreasing function of average level of employment of other firms. Each subject represents a firm making its employment decision. This decision depends on firm’s expectation of average employ- ment of other firms, so the subjects make forecasts of the average employment of other firms. This model has 3 steady states: low and high steady states are E-stable, middle steady state is not E-stable. There also exists E-stable sunspot equilibrium that fluctuates between values near two steady states that are E-stable. These two steady states are payoff ranked: low value gives lower profit than higher value. We have performed two treatments: one in which payoff is based on the firm’s profits, and the other in which payoff is based on the forecast squared error. We observe coordination on the extrinsic announcements in both treatments. In the treatments with forecast squared error, the average employment and average forecasts of subjects are closer to the equilibrium corresponding to the announcement.
Author(s): Arifovic, J, G. Evans and Ol. Kostyshyna
Topic: Markets, learning
Back to session: Markets 3
ABSTRACT:
We present an experimental design that explores the interaction between group identity formation and collusion in an oligopoly supergame. Subjects are randomly assigned to sell one of two commodities in an oligopoly market. Our treatments vary along two dimensions. First, group identity among sellers may or may not be induced before the market stage. Second, communication possibilities in the market stage may not exist, allow for cheap talk, or permit binding agreements. We elicit social preferences after the market stage of the experiment to capture existence and strength of group identity among sellers. Our design addresses the following questions: does induced group identity facilitate collusion among sellers? Does group identity emerge endogenously as a result of 'trade conspiracies'? Does market competition erode or amplify pre-existing social identity among sellers?
Author(s): Viktor Brech; Daniel Houser
Topic: Markets, norms and fairness, bargaining
Back to session: Markets 3
ABSTRACT:
Goods and services are often transacted in settings where the buyer faces an explicit choice of whether to purchase at the stated posted price, or engage in bilateral price negotiations with the seller. In this paper, I develop a theoretical framework where the negotiation behavior of buyers in such settings depends on their perceptions regarding the fairness of the posted price. Specifically, the model posits that buyers dislike engaging in negotiations, and aggressively negotiating, when the posted price is perceived to be fair. As a result, buyers who are sensitive enough to these posted price fairness concerns may even forgo engaging in profitable negotiations, and simply purchase the good at the stated posted price. I then present the results from a series of price negotiation experiments, which are intended to shed light empirically on the influence of posted price fairness perceptions on the negotiation behavior of buyers.
Author(s): Eric Cardella
Topic: Markets, norms and fairness, bargaining
Back to session: Markets 3
ABSTRACT:
We extend the results of Bartling and Fischbacher (Rev. Econ. Stud. 79(1):6787, 2012) by showing that, by delegating to an intermediary, a dictator facing an allocation decision can effectively shift blame onto the delegee even when doing so necessarily eliminates the possibility of a fair outcome. Dictators choosing selfishly via an intermediary are punished less and earn greater profits than those who do so directly. Despite being powerless to influence the fairness of the outcome, an intermediary given the choice between two unfair outcomes is punished more than when the dictator chooses one directly. This is not the case when the intermediary merely can initiate the random selection of one of the outcomes. Our findings reinforce and clarify the usefulness of agency as a tool to evade perceived culpability.
Author(s): Zachary Grossman, Regine Oexl
Topic: Reciprocity, norms and fairness
Link: http://www.springerlink.com/content/x42342725q213372/
Back to session: Markets 3
ABSTRACT:
The scale of experimental school choice studies is usually significantly smaller than its real world counterpart. In this study, we expand the scale of school choice experiment in the laboratory setting by pairing human subjects with many computer agents who either reveal their true preferences, or draw their strategies from those of the human subjects in similarly designed experiments. We find that the truth-telling property of both the Boston and the DA mechanisms holds when the scale becomes larger. In addition, the Boston mechanism is more efficient than the DA mechanism under larger scale. However, we do not observe that scale has significant effect on truth-telling behavior across both mechanisms.
Author(s): Yan Chen, Ming Jiang, Onur Kesten, St?phane Robin, Min Zhu
Topic: Matching markets
Back to session: Matching Markets
ABSTRACT:
We study in the laboratory a variant of the house allocation with existing tenants problem where (i) subjects are partitioned into tiers with hierarchical privileges, (ii) they play multiple matches, and (iii) they know their position in the priority queue before making their decision. We evaluate the performance of the modified versions of three well-known mechanisms: Top Trading Cycle, Gale-Shapley and Random Serial Dictatorship with Squatting Rights. For all three mechanisms, we find low rates of participation (around 40%), high rates of truth-telling conditional on participation (around 90%), high proportions of fair allocations (above 90%) and significant efficiency losses. We also observe differences across mechanisms: Random Serial Dictatorship is ranked highest in efficiency and Top Trading Cycle is ranked lowest in fairness. We then show that position in the queue has a positive and significant impact on participation whereas experience and tier has little effect on behavior. Finally, the individual analysis reveals that the majority of subjects who do not play according to the theory still follow discernible patterns of participation and preference revelation.
Author(s): Juan Carrillo & Saurabh Singhal
Topic: Matching markets
Link: http://www-bcf.usc.edu/~juandc/PDFpapers/wp-matching.pdf
Back to session: Matching Markets
ABSTRACT:
We experimentally vary the participants' level of uncertainty about his/her rank (no uncertainty or some uncertainty) when stating the preference ranking or the matching mechanism (Boston or Gale-Shapley) in the centralised college choice problem. We examine how strategic behaviour and efficiency differ across the four matching regimes. The study also explores the relationship between various preference, demographic, and experience variables to ranking decisions. We discuss the potential policy implications of our results for college matching in China.
Author(s): Stephanie W. Wang; Xiaohan Zhong
Topic: Matching markets
Back to session: Matching Markets
ABSTRACT:
We conduct a laboratory analysis of the econometric procedure proposed by Aguirregabiria and Magesan's (2012, A&M hereafter). This approach enables one to estimate the structural parameters of dynamic games without assuming players' beliefs are in equilibrium. This econometric approach is valuable in that, despite its analytical convenience, assuming rational expectations (that players' beliefs about their opponents' strategies represents their opponents' actual strategies) can be both restrictive and misleading. Using a laboratory experiment with human participants who make decisions in a market entry game, we (i) discover the conditions under which A&M's procedure provides reliable inference with respect to our game's structural parameters, and (ii) draw inferences about the nature of human players' beliefs in these environments.
Author(s): Li Hao, Daniel Houser, and Joachim Winter
Topic: Games, methods
Back to session: Methods 2
ABSTRACT:
The aim of this paper is to test the effectiveness of wage-irrelevant goal setting policies in the lab. In our setting, subjects who are assigned the role of managers can assign a goal to their workers which specify a certain level of performance on the work task. We find evidence that managers set goals that are challenging but attainable for an average-ability worker. Workers respond to these goals by increasing effort, performance and by decreasing on-the-job leisure activities with respect to the no-goal setting baseline. We study the interaction between goal setting and monetary rewards by considering different values for the monetary incentives involved in completing the work task. Interestingly, we find that goal setting is especially effective when monetary incentives are strong because it eliminates the adverse effect associated with high monetary rewards which is observed in the baseline. These results suggest that goal setting may foster a workers' intrinsic motivation and increase their level of performance beyond what is achieved using solely monetary incentives.
Author(s): Joaquin Gomez-Minambres, Brice Corgnet and Roberto Hernn-Gonzalez.
Topic: Labor, methods
Back to session: Methods 2
ABSTRACT:
Despite the rising profile of experimental economic methods for both research and teaching, the incidence of usage in college and high school economics classrooms remains limited. The goal of the Kiviq' is to address the limitations of currently available software and provide instructors with a tool that is simple, flexible, fast, and fun. Kiviq' allows economic experiments to be run in class with large numbers of students via mobile devices. The current software features a variety of double auction institutions with additional experiments in the works.
Author(s): Kyle Hampton
Topic: Methodology
Link: http://auction.akresearch.org
Back to session: Methods 2
ABSTRACT:
Laboratory experiments employing an induced-values methodology report on allocative efficiencies observed. That methodology requires experimenters know subjects' motivations, impossible in field experiments. Allocative efficiency implies a hypothetical costless aftermarket would be inactive. An allocation mechanism's outcome is defined to be behaviorally efficient if an appropriate aftermarket is actually appended to the mechanism and at most a negligible size of remaining mutually beneficial gains identified. Methodological requirements for behavioral efficiency observation are provided. A first demonstration observes significantly greater behavioral inefficiencies in second- than in first-price auctions. A simple field demonstration indicates when a public good increase can be observed to mutually beneficially cover marginal cost, without knowing valuations. Several empirical issues that arise are noted.
Author(s): Ronald M. Harstad
Topic: Methodology
Link: http://harstad.missouri.edu/Getfiles/BEfJuly.pdf
Back to session: Methods 2
ABSTRACT:
Experimental Evidence from Japan, China, the Netherlands and Italy. This paper studies ingroup bias in a simulated unenforceable bribery context. Subjects were paired and played a game in which one subject could cheat for the other (but not the other way around). Before they played this cheating game, however, the potential beneficiary of cheating had the opportunity to bribe the potential cheater and the latter had the opportunity to solicit a bribe from the former. Note that the opportunity to bribe and cheat was not explicitly offered by the experimenter but had to be discovered by the subjects. Before the game started, the two parties were given five minutes of bilateral online chat time to negotiate a side payment, if any. They both knew that the side payment would only be implemented if they agreed on the amount and zero otherwise. Since they inputted this amount in private, however, the briber could defect on the agreement by inputting a different amount. Since players only interacted once and the briber had no possibility to punish the bribee if the bribee did not cheat for him or her, the bribee could also defect on the briber by not cheating or even cheating to the disadvantage of the briber. This cheating game with bribe was played with an ingroup member and with an outgroup member. The results show that the amount of the bribe significantly correlates with the amount of cheating in ingroup bribery deals, but no correlation is found for the outgroup bribery deals. Moreover, while Japan has the lowest success rate (inputting the same positive bribe amount), the Netherlands has the highest. In addition, the ingroup bias is stronger among Dutch students than among the Japanese, Chinese or Italian students.
Author(s): Ting Jiang, Jan Willem Lindemans
Topic: Games, morals
Back to session: Morals 1
ABSTRACT:
This paper examines the effectiveness of using asymmetric liability to combat harassment bribes. Basu (2011) advocates legal immunity for bribe-givers, while retaining culpability for bribe-takers. Results from our experiment indicate that this policy has the potential to significantly reduce corrupt practices; however, weak economic incentives for the bribe-giver, or retaliation by bribe-takers can mitigate the positive disciplining effect of such an implementation. As a result, asymmetric liability on its own may face challenges in the field.
Author(s): Klaus Abbink, Utteeyo Dasgupta, Lata Gangadharan, Tarun Jain
Topic: Games, Norms
Back to session: Morals 1
ABSTRACT:
We argue that the willingness to follow costly rules reveals information about an individual's type and facilitates the solution of social dilemmas via assortative matching. To illustrate this point, we study a novel, repeated common pool resource game in which current resource stocks depend on resource extraction in previous periods. We show that for a sufficiently high regrowth rate, there is no commons dilemma: the resource will be preserved indefinitely in equilibrium. Behavioral tests of the model indicate that favorable ecological characteristics are necessary but insufficient to encourage effective CPR governance. However, by screening and sorting individuals according to their willingness to follow costly rules in an individual choice task, we show that CPR groups composed of rule-followers are less likely to exhaust the resource than both groups of rule-breakers and mixed-type groups. Rules function as screening mechanisms for the identification of cooperative types.
Author(s): Erik O Kimbrough and Alexander Vostroknutov
Topic: Homegrown preferences, cooperation
Back to session: Morals 1
ABSTRACT:
Cheating is sometime used to gain competitive advantage. It is conjectured that in a competitive environment, the less advantageous people would cheat more to compensate for their deficiency. We use a lab experiments to test the hypothesis that smart people cheat less. Subjects take an incentive compatible IQ test to reveal their overall intelligence. Then subjects with the similar IQ scores are randomly matched into groups of four. A depository device is used to elicit subjects' cheating levels. The overall finding confirms the hypothesis, that is, compared to subjects with low IQ scores, subjects with high IQ scores cheat less in a competitive environment. We also alter the environment to eliminate competition, we find that the difference in cheating levels re-emerges, but in a different direction. Furthermore, the risk neutral assumption is relaxed in this study. We use the lottery game to assess each subject's risk preference, and then incorporate the degree of risk preference into analyzing the subject's cheating level.
Author(s): Xiangdong Qin and Shuwen Li
Topic: Morals, cognition
Back to session: Morals 2
ABSTRACT:
Abstract: Anecdotal evidence suggests that after people first engage in immoral behavior,it becomes easier to continue to do. This may imply that the cost of additional moral transgressions is decreasing in the number of previous transgressions. Here we explore this phenomenon by providing participants in a laboratory environment the opportunity to lie repeatedly. Our treatments vary in terms of the psychic difficulty of the first lie: in one treatment it is a harder lie to tell (a selfish hurtful lie), while in the other it is an easy lie (Pareto improving). We hypothesize that the act of telling the first lie will make subsequent lies more common.
Author(s): Siyu Wang and Daniel Houser
Topic: Morals, norms
Back to session: Morals 2
ABSTRACT:
A growing literature in experimental economics is studying the interaction between work-place culture and worker selection. Dohmen and Falk (2011) show that compensation schemes induce sorting of workers by productivity, their degree of risk aversion, extent of self-confidence and their concern for behaving altruistically. Niederle and Vesterlund (2007) find that women are less likely than men to enter tournaments when offered a choice between tournaments and piece rates. We want to build on this literature and study the effects of labor market sorting in the choice of profession between public vs private sector in India based on cheating behavior. We answer the question whether labor market sorting in the choice between public and private sector can be explained in part by unethical behavior in the laboratory experiment. We conducted this experiment in India and through a unique design we captured the cheating behavior of agents. We found that there is a significant difference in cheating behavior between public and private sector aspirants. The amount of cheating of public sector aspirants is significantly more than private sector aspirants.
Author(s): Ritwik Banerjee, Tushi Baul and Tanya Rosenblat
Topic: Morals, norms, labor
Back to session: Morals 2
ABSTRACT:
We report results from an experiment that investigates truthfulness in self-reporting under different reporting regimes. The experiment involves a production task with self-reporting of accidents, with reporting compulsory for some participants, but only voluntary for others. We find that dishonesty is prevalent, but accident reporting is more frequent with compulsory reporting compared with voluntary. This suggests that lie aversion is a stronger force than the intrinsic motivation to voluntarily report, and that careful design of self-reporting regimes is necessary by enforcement agencies to achieve satisfactory compliance outcomes. Our results are relevant for several areas beyond regulatory compliance, including dishonesty in social security claims, insurance claims, workplace expense claims, income tax returns, and financial reporting.
Author(s): Lana Friesen (University of Queensland) and Lata Gangadharan (Monash University)
Topic: Morals, norms
Back to session: Morals 2
ABSTRACT:
In a budget reporting experiment in which lying about privately known costs increases personal monetary payoffs, we examine how an individual's honesty is affected by information about their peers who face similar reporting dilemma. We find that subjects decrease their honesty in response to less honest peer and increase their honesty in response to more honest peer. The response to less honest peer is much stronger than the response to more honest peer. Furthermore, the reaction to more honest peer is significant predominantly for subjects who are identified as the conventional type by a psychometric test. The conventional type is sensitive to social norms according to the Kohlberg's theory of moral development (Kohlberg 1958). All types react to less honest peer, and the magnitude of the reaction is correlated with a measure of disadvantageous payoff aversion in the post experiment questionnaire.
Author(s): Steve Huddart and Hong Qu
Topic: Networks, norms
Back to session: Networks 1
ABSTRACT:
The University of California schools have experienced a drastic change in racial makeup over the last 20 years. The top ranked UC's (Berkeley and Los Angeles) experienced a sharp drop-off in the share of black and Hispanic enrollees in 1998, following Prop 209, the ban on affirmative action. While the lower ranked UC's, such as Riverside, have experienced an increase in the share of black and Hispanic students. I exploit this variation in racial make-up at University of California schools, and find that when a race makes up a larger share of a cohort in a school, that race has a higher graduation rate in spite of a lower college GPA. Since an increase in race-share has a negative effect on GPA, I argue that the mechanism is not through selection of better students, but through the stronger social groups that arise when students have more own-race peers.
Author(s): Daryl Fairweather
Topic: Labor, networks
Back to session: Networks 1
ABSTRACT:
This experiment focuses on a contest played over a star network of 6 nodes. By placing targets in a network, the value of a target is dependent on its connectivity to other targets. The experiment compares the cases where the defense is centrally planned by a defense planner with a case where the defense is coordinated by six individuals each placed at one node in the network. In treatments allowing centrally planned defense, both attackers and defenders generally recognize that periphery nodes have symmetric value and the center node is more valuable, however too much effort is placed on the center node by both types. With the coordinated defense treatments, not enough effort is placed on the center node which is caused by a combination of self-interest and coordination errors. This behavior by the attackers and defenders leads to networks with a centrally planned defense earning higher than equilibrium profits whereas a network with a coordinated defense earning lower than equilibrium profits.
Author(s): Kory Garner
Topic: Network, coordination, games
Back to session: Networks 1
ABSTRACT:
Field experiments on real-world social networks typically impose a binary link structure: two agents are either classified as friends or strangers. We present three measures of link strength that can be easily elicited through a survey or a baseline experiment: (1) the order in which subjects list their friends during network elicitation; (2) time spent together elicited through a coordination game; (3) the extent to which subjects make correct guesses about friends habits through a trivia game. We present evidence from 3 separate field experiments on directed altruism, flu vaccinations and healthy eating habits in an elementary school to show that these measures strongly predict the strength of peer effects. In each case, we estimate a decay function and the size of one's circle of friends that generates the majority of peer effects. Our findings have two main implications for the design of field experiments. First of all, there is an optimal cutoff point for eliciting network connections. Second, network surveys often contain underused paradata that can help estimate interaction decay in social networks.
Author(s): Tanya Rosenblat and Markus Mobius
Topic: Networks, field experiment
Back to session: Networks 1
ABSTRACT:
We conducted a Bertrand and Mullainathan AER(2004) type audit study to test for ethnic discrimination in China. We sent resumes with identical distributions of qualifications to advertised internship positions for business majors in Shanghai and Beijing. Our preliminary results show that resumes with Han names are marginally significantly more likely than Tibetan names to get a callback. However, more interestingly, the effect becomes more significant when the resume is local and completely insignificant when the resume is non-local. Similarly, CCP membership only has a significant effect on Han callbacks when local. Our result seems to suggest that people are really discriminating by usability of social network, and race is a proxy for usable network in China.
Author(s): David Ong, Sophie Zeng, Zachary Zhong
Topic: Networks, field experiment
Back to session: Networks 2
ABSTRACT:
This paper experimentally studies the formation of social networks when individual decision-makers imperfectly monitor others' network ties. As predicted by the theory, we find that the networks will sometimes converge to inefficient stationary configurations. However, consistent with prior experimental research on network formation, we find that convergence to a stationary configuration is not guaranteed. We compare network formation under neighbor sight, in which the decision maker's own ties and neighbors' ties are visible, to network formation under full observation, in which the entire network is visible. We also vary the relative cost of ties across treatments and compare the observed networks to the equilibrium prediction. For some values of the parameters, any minimally connected network is predicted be stable, and for other values, only the empty network is predicted to be stable.
Author(s): Michael Caldara; Michael McBride
Topic: Networks, IO
Back to session: Networks 2
ABSTRACT:
The Braess Paradox is a counterintuitive finding that upgrading a network may increase users' cost. Our goal is to determine theoretically and empirically if users may benefit when the population consists of groups that operate as unitary players. We have players choose routes in a network game under two conditions. In one, each user controls a single unit. In the second, users control multiple units which may be split. Our results show that in both conditions behavior converges to equilibrium.
Author(s): Eyran Gisches & Amnon Rapoport
Topic: Networks, IO
Back to session: Networks 2
ABSTRACT:
Extending Crawford and Sobel (1982), theory suggests (1) small incentives can reduce information transmitted between groups and (2) larger groups can be less likely to lie than smaller groups (Galeotti et al, 2012). These results demonstrate how monetary incentives change the information transmitted in a strategic environment. Information transmission in natural environments, however, includes both monetary and social incentives. Unfortunately, little is known about how these two effects interact to determine truth-telling. We investigate how money and social identity impact truth-telling in a laboratory experiment. Absent social identity, players' choices are highly consistent with the theoretical predictions. In the presence of social identity, people are willing to sacrifice monetary gains in order to avoid lying to ingroup members. On the other hand, they lie more to out-group members. Our results are consistent with parochial altruism, a behavioral phenomenon that has been found in many social settings (Bernhard et al, 2006). Our results that parochial altruism impacts information transmission networks may have important implications for many social environments, especially those where players in different monetary alliances shares the same social identity.
Author(s): Rong Rong and Daniel Houser
Topic: Networks, norms
Back to session: Networks 3
ABSTRACT:
Behavioral experiments have shown that individuals vary in their cooperative preferences in public goods environments. Such preference heterogeneity is also observed to be robust across cultures and over time. Besides preference heterogeneity, real world social groups are also characterized by the network structure that constrains the interaction among constituent individuals. In this paper, using data generated by subjects in behavioral experiments and computational simulations, we explore the interaction between network structure and behavioral heterogeneity in networked public goods environments. Agent based models are populated with heterogeneous agents whose behavioral specification is derived from the experimental data. We then systematically investigate aggregate contribution dynamics in random regular, small world, and scale-free networks. In contrast to the evolutionary models of networked public goods games, we find that the population of heterogeneous agents reverts to near full free riding irrespective of the underlying network topology. These results suggest that the population structure alone may be inadequate to promote cooperation.
Author(s): Chenna Reddy Cotla and Ragan Petrie
Topic: Networks, public goods, cooperation
Back to session: Networks 3
ABSTRACT:
The aim of this paper is to ascertain whether trust is affected by contagion and herding in small groups of trustors who can observe each other's choices over time. We account for three important factors of trustors' preferences, namely: risk attitude, generosity and expected trustworthiness. Using our data, we test the basic hypothesis that an individual's propensity to trust recipients in the Trust Game may be affected by the observed behavior of other trustors. Our results confirm that trust is affected by contagion effects. Furthermore, we find that specific types of agents (generous or untrusting) frequently imitate the same type when placed in the same group
Author(s): L. Luini, A. Nese, P. Sbriglia
Topic: Networks, social preferences
Back to session: Networks 3
ABSTRACT:
Since the 9-11 attacks, there has been increased interest in understanding how terrorist organizations emerge, function, and succeed, and also in understanding how they adapt in the face of counterterrorist operations. Important empirical studies have mapped the organizational structures of terrorist groups; however, the extreme difficulty in developing accurate case studies and in conducting field experiments makes theoretical generalizations difficult. Recognizing these difficulties, we develop a set of theoretical models of terrorist organization formation and test these models using laboratory experiments. Drawing insights from game theory, we developed a stylized model of terrorist network formation, derived predictions about how the structure of terrorist networks change in response to changes in random counterterrorist interventions, and obtained experimental evidence that qualitatively supports our model’s predictions. We conducted these experiments with the help of a seed grant awarded at the NSF-funded conference Bridging Areas of Expertise: Funding Research on Terrorism. In our conference paper, we present our model and initial findings regarding the impact of counterterrorism tactics on the emergence and organizational structures of terrorist networks.
Author(s): Natalia Candelo, Sherry Forbes, Michael McBride, Susanne Martin
Topic: networks, games
Back to session: Networks 3
ABSTRACT:
Enforcing the law is a primary function of governments, but regimes vary markedly in the extent to which their activities are viewed as legitimate by the general population. These perceptions, in turn, affect the willingness of citizens either to cooperate with the regimes' enforcement activities or to hinder their operation. We devise a laboratory experiment that assesses the dynamics of legitimacy and citizen-regime cooperation and conflict. In our experimental scenario, players can assist or hinder another player--the enforcer--when she attempts to punish others' failure to contribute to a public good. We take players' propensities to assist or hinder the enforcer as a behavioral measure of legitimacy, and exogenously manipulate features of the institutional and informational environment across experimental treatments as a means of exploring potential causes of variation in this measure. We find that the legitimacy of the enforcer varies systematically, depending on the means through which the enforcer is compensated, the degree of transparency with which the enforcer's decisions are observed, and an interaction between these quantities.
Author(s): Eric Dickson, Sanford Gordon, and Gregory Huber
Topic: Law, enforcement, government
Back to session: Norms and Fairness 1
ABSTRACT:
I examine the selection of individuals for a team and signaling behavior of potential teammates in an experiment involving team competitions. Individuals choose team members based on randomized information treatments providing actual performance, stated performance or a portrait photograph. I find that generally individuals inflate performance scores when sending a signal of quality, but the magnitude of inflation is greatest after viewing pictures of all candidates in a session. In a later treatment, subjects observe stated performance scores jointly with photographs to test for the effects of potential stereotypes. I find little discrimination between genders, but race effects make it less likely for individuals of the same race to believe each other's quality signals. Using external ratings of photographs for perceived attractiveness and intelligence, I find that perception based on photographs significantly impacts selection. These results suggest that job applicants are more honest if they have no information regarding (performance and physical) qualities of other applicants while job screeners may apply racial stereotypes to applicants that have been applied to themselves. Furthermore, when no performance information is provided screeners focus on attractiveness or other perceived qualities from appearance to choose candidates.
Author(s): David Wozniak
Topic: Labor, norms and fairness
Back to session: Norms and Fairness 1
ABSTRACT:
Wage differentials can arise to compensate workers when their jobs are seen as more arduous or risky than other jobs. Firms need to pay workers more to entice them to do the less attractive jobs, such as working at night or working in industries with greater health risks. We ask whether the antisocial behavior of fellow workers might also lead to compensating differentials. We report the results of a real effort experiment in which workers can move between firms and firm managers compete for workers by posting piece rates. We find that wage differentials evolve to compensate workers in firms that permit theft or sabotage of their output by coworkers.
Author(s): Jeffrey Carpenter, Peter Hans Matthews, and Andrea Robbett
Topic: Labor, social preferences
Back to session: Norms and Fairness 1
ABSTRACT:
We use an online real-effort experiment to investigate how bonus-based pay and worker productivity interact with workplace cheating. Firms often use bonus-based compensation plans, such as group bonuses and firm-wide profit sharing, that induce considerable uncertainty in how much workers are paid. Exposing workers to a compensation scheme based on random bonuses makes them cheat more but has no effect on their productivity. We also find that more productive workers behave more dishonestly. We explain how these results suggest that workers' cheating behavior responds to the perceived fairness of their employer's compensation scheme.
Author(s): David Gill, Victoria Prowse, Michael Vlassopoulos
Topic: Labor, norms and fairness
Link: https://faculty.cit.cornell.edu/vlp33/Bonuses.pdf
Back to session: Norms and Fairness 1
ABSTRACT:
Economists typically advocate price or quantity based solutions to deal with the problem of externalities. Quantity solutions are typically Coasean in nature and involve allocating property rights to produce an externality. The Coase theorem states that with low bargaining costs, perfect competition, perfect information and the absence of wealth and income effects, externalities will be internalized regardless of who property rights are assigned to. All property rights are to some degree incomplete; they rely on a mix of both external penalties and social norms in order to function. Thus, Coasean solutions may depend heavily on social norms, in ways that have not been apparent to many economists. We clearly demonstrate in a controlled laboratory setting that is possible to either undermine or enhance Coasean solutions by crowding out or crowding in social norms. This has important implications for the use of property rights solutions in public policy.
Author(s): Andreas Leibbrandt and John Lynham
Topic: Norms
Back to session: Norms and Fairness 2
ABSTRACT:
This paper studies the effects of voicing one's opinion towards an independent decision-maker on generosity towards him in subsequent situations. We find that voicing one's opinion before a decision is made positively influences generosity towards the decision-maker. Two players complete real effort tasks with asymmetric workload and asymmetric earnings per task in order to earn money which is then distributed by an independent decision maker among the two players. Treatments differ in the possibility for the player with the higher workload to voice his opinion about a fair distribution towards the independent decision-maker. In the baseline no messages can be sent. In the restricted-voice treatment the desired distribution is sent, while in the full-voice treatment a the desired distribution plus an additional free from message is sent. In an unannounced second stage this player plays a dictator-game with the decision-maker. Applying the strategy method the player indicates how much money he would transfer conditional on every possible distribution which could have been chosen by the decision-maker in the first stage. We find that the players are more generous towards independent decision-makers in the voice treatments than in the baseline treatment. Generosity is not significantly different across the voice treatments, indicating that the extent to which the opinion can be stated is not decisive for the result to hold. The higher generosity in the voice treatments is a general result, in that the players are more generous for any possible distribution implemented by the independent decision-maker.
Author(s): Kleine, Marco; Langenbach, Pascal; Zhurakhovska, Lilia
Topic: Norms and Fairness
Link: http://www.coll.mpg.de/team/page/lilia_zhurakhovska
Back to session: Norms and Fairness 2
ABSTRACT:
In threshold public goods games, there typically exists an efficient and fair equilibrium at equal contributions, as well as many efficient but unfair equilibria and the inefficient equilibrium of full free riding. Nonetheless, subjects in experiment often fail to reach any of the efficient equilibra, even in the case in which contributions are rebated if the threshold is not met. Small groups will often create a rule to punish free riding behavior, but these rules may require heavy monitoring. We present a new mechanism, the hired-gun-mechanism, which only needs to identify the least compliant group members, and punish only the least compliant individual, such that this person would have rather been the second least compliant. We test this hired-gun-mechanism under two conditions. First, we use the mechanism to remove the inefficient equilibrium only, and second we remove all but the fair equilibrium. We find that the mechanism makes only modest improvements in efficiency in the first condition, but makes huge improvements when it targets only the fair equilibrium. This suggests that mechanisms may also need to focus on fairness, not simply efficiency, to be successfully employed.
Author(s): James Andreoni and Laura K. Gee
Topic: Norms and fairness, Public goods
Back to session: Norms and Fairness 2
ABSTRACT:
There is a large literature looking at the impact of various incentive schemes in general and tournaments in particular on performance. The basic insight coming out of this literature is that tournaments in general improve performance. We explore aspects of these questions using a multiple cue probabilistic learning task requiring real effort. We compare the impact of extrinsic incentive schemes, where pay is explicitly tied to performance such as piece rates and tournaments, with intrinsic schemes, where players earn a flat salary regardless of performance. In order to separate out the role of psychological competition as opposed to the incentive effects, within our tournament treatments we vary the information that our subjects receive where in some treatment they are paid using a winner-take-all tournament scheme and learn about their relative standing while in others they are unaware of how they are performing vis-à-vis a pair member. We also use psychological questionnaires to control for trait anxiety and intrinsic motivation. We find that after controlling for levels of anxiety during the task, overall we get better performance (as measured by lower forecasting errors) in the piece-rate and salary treatments, where participants face less competitive pressure. We also find that women perform worse than men in all treatments except for salary. We vary the degree of task difficulty and find interesting differences in performance under various schemes. Our results have implications for the design of incentive schemes and suggest that we need to take better account of intrinsic motivation.
Author(s): Paul Brown, Linda Cameron, Ananish Chaudhuri and Tony So
Topic: Labor, emotion and choice
Back to session: Norms and Fairness 2
ABSTRACT:
A trend in fishery management allocates to self-identifying subgroups of harvesters in a fishery, or ôsectors,ö their collective total share of the harvest of each species as a group right, that they may manage in any way they wish. This means a single fishery with a single set of overall total allowable catches can be managed by multiple management systems concurrently. We ask whether common pool or individual quota type management systems are more likely to emerge as groups gain experience with their collective rights. In a novel quasi-continuous time experimental environment with a contemporaneous price externality, harvesters can choose to affiliate with a common pool managed group, or an individual quota managed group. We find that the common pool group engages in a fishing derby and receives lower prices, whereas the individual quota group achieves stable harvest levels throughout each season to minimize the price externality. Through successive fishing seasons, the frequency of subjects' choosing individual quota rises from about half to over 85% of subjects. This suggests that the efficiencies associated with strong individual fishing rights may emerge endogenously from the sectorization process, even without imposing them through regulation. We use three supplemental instruments to measure individual preferences to explain which subjects remain in common pool management, and find enjoyment of competition is significant, but risk attitudes and other-regarding preferences are not.
Author(s): Mihoko Tegawa and Christopher M. Anderson
Topic: Norms and fairness, Public goods
Back to session: Norms and Fairness 3
ABSTRACT:
We investigate experimentally whether the choice of an incentive scheme can signal a social norm, when the principal is better informed about past actions of a broader number of other agents and the agent is aware of this. The principal can choose between a fixed wage ('trust contract') and performance pay ('contingent contract'). She is matched to an agent who then determines his effort. We find that indeed agents react differently to identical contract choices when they know that the proposing principal is informed about prior efforts of other agents in an identical situation. Agents exert significantly higher effort under a fixed wage contract when knowing that the principal proposed this contract being aware of others' reaction to it.
Author(s): Anastasia Danilov and Dirk Sliwka
Topic: Norms, labor
Back to session: Norms and Fairness 3
ABSTRACT:
This paper studies inequality due to luck. Our experiment allows for two types of luck; option luck and brute luck. Option luck involves risk that partipants can insure against, while brute luck is a risk beyond the participants' control. In a risk-taking experiment where both types of luck are present, we study distribution decisions made by spectators. Our design allows for an ex-post distinction of the kind of luck that caused the inequality. Using the distribution decisions, we are able to categorize the spectators according to the main four fairness norms discussed in the philosophical literature: strict-, luck-, choice- and ex-ante egalitarianism.
Author(s): Johanna Möllerström and Bjorn-Atle Reme
Topic: Social preferences, anomalies
Back to session: Norms and Fairness 3
ABSTRACT:
We designed an experiment that partially mimics Niederle and Vesterlund (2007), but we use an alternative consinuous measure of competitiveness. Our design allows us to explore the links between competitiveness, risk, cognitive abilities and fetal exposure to testosterone (2D:4D) by gender. We find no gender differences in competitiveness and no effect of fetal exposure to testosterone. Other factors do seem to explain differences in competitiveness.
Author(s): Diego Aycinena, Rimvyidas Baltaduonis, Lucas Rentschler
Topic: Gender, physiology
Back to session: Physiology
ABSTRACT:
There is evidence that testosterone (T) affects risk-taking behavior, aggression and cognitive abilities, all of which are essential in financial decision-making. Given the large proportion of male traders in financial markets, the impact of the potent androgen can be formidable. We externally administer T and Placebo (P) on male participants in a double-auction asset trading experiment to examine the effect of T on behavior that relates to the formation and characteristics of asset price bubbles. We perform analyses at two distinct yet interconnected ways: market level and individual level. At the market level, we measure bubbles in each market and find that T leads to bubbles with larger amplitude and non-declining duration in repeated markets, but T does not significantly increase turnover. At the individual level, we find that compared to P subjects, T subjects are more willing to buy as price increases and as bubble forms, exhibiting a tendency to participate and perpetuate the bubble; they incorporate the constantly declining fundamental value at much slower rates; these lead to a negative correlation between subjects' level of T and their performance; and higher T subjects attribute their performance more to luck and less to talent. The paper concludes with a discussion of the possible underlying mechanism for the effects of T on trading behavior and asset bubbles.
Author(s): Amos Nadler, Peiran Jiao, Veronika Alexander, Cameron Johnson, Paul Zak
Topic: Physiology, finance, markets
Back to session: Physiology
ABSTRACT:
Decision making is a central component of human behavior. People make numerous decisions every day in a variety of situations including perceptual decisions that guide walking through a crowd or driving a car, decisions over primary rewards like what to eat or drink, and social decisions that require balancing benefits for oneself and other people. The ability to explain and predict such choices is a goal aspired to across many scientific fields. Mathematical models that describe choice behavior in specific contexts have provided important insights into the computations that may underlie decision making in the brain. However, a critical and largely unanswered question is how well these models can generalize from one choice context to another. Here we show that a model adapted from the perceptual decision-making domain and estimated on choices over food rewards can be used to accurately predict choices and reaction times in three independent sets of subjects making social decisions. The robustness of the model provides strong evidence for a common decision-making process in perceptual, primary reward, and social decision making.
Author(s): Ian Krajbich, Todd Hare, Bjoern Bartling, Yosuke Morishima, Ernst Fehr
Topic: Physiology, social choice
Back to session: Physiology
ABSTRACT:
Exercise is known to improve health along many dimensions. Decision making is an understudied dimension of one's (behavioral) health where exercise effects are not well-known. Because certain physiological changes are known to impact decision making, exercise may modulate decisions via its effect on physiological or psychological variables. We examine how moderate aerobic exercise affects outcomes and the decision process in an incentivized Bayesian choice task. Twenty-six adult subjects (30-60 years old, 14 female) are administered the decision task under both exercise and no-exercise conditions. Our results indicate that the estimated decision model changes post-exercise such that exercise increases the decision weight subjects place on new evidence in making their choices. This same effect is found among those with higher fitness levels, which is a long-term result of regular exercise. We also show this effect is not due to repeat administration of the task. Nevertheless, the accuracy rate of subjects' choices is not significantly affected by the exercise treatment. The fact that post-exercise decisions, as well as decisions from more fit individuals, rely relatively more on evidence compared to base rate information is important in terms of identifying when exercise may lead to improved versus harmed decision making in more complex environments.
Author(s): David L. Dickinson and Scott R. Collier
Topic: Physiology, cognition
Back to session: Physiology
ABSTRACT:
In this paper I study a simple model of binary excludable public goods with interdependent values. Previous research on excludable and non-excludable public goods has largely focused on private values or complete information. Allowing for interdependent values (including the special case of pure common value) leads to a potentially important selection effect. Under a simple threshold or provision point mechanism, provision of the public good in this environment is good news about its value in Bayesian Nash equilibrium. Because of this favorable selection effect, naive agents who condition their expectations only on their own private information will contribute too little. I experimentally examine whether such a 'curse' exists in this environment by testing Bayesian Nash equilibrium against naive strategies. I also consider a similar 'anti-threshold' treatment exhibiting adverse selection and an uncertain private values treatment with no selection effect to examine how subjects respond to and learn to account for favorable and adverse selection. Behavior is quite similar across these treatments, despite sharp differences in equilibrium predictions. I find little evidence of learning over multiple rounds of play.
Author(s): Caleb A. Cox
Topic: Public Goods
Back to session: Public Goods 1
ABSTRACT:
Recent influential research in Political Economy has emphasized the need to explicitly model dynamic incentives However, whether and how much dynamic incentives matter is ultimately an empirical question. In this paper we present an experimental design that allows us to disentangle between static and dynamic incentives to under invest in a public good. Experimental sessions have been conducted, we are currently processing the data and writing the working paper.
Author(s): Marina Agranov, Guillaume Frechette, Thomas Palfrey, Emanuel Vespa
Topic: Public goods
Back to session: Public Goods 1
ABSTRACT:
Our study takes an individual perspective on receiver credulity in a public good setting with deceptive messages. In a laboratory experiment, subjects play a public good game with punishment in which feedback on actual contributions is obscured. Instead, subjects can communicate what they have contributed through a post-hoc announcement mechanism. Using subject's social value orientation, we show that those highest on the measure are too optimistic towards announcements of their fellow group members. This, in turn, influences payoff-relevant decisions: those high on social value orientation contribute more to the public good and punish their fellow group members less.
Author(s): Bernd Irlenbusch & Janna Ter Meer
Topic: Public goods, communication
Back to session: Public Goods 1
ABSTRACT:
We study the effects of fiscal equalization between different regions or groups in a public goods experiment. Groups can differ with respect to their marginal per capita return (MPCR) or capital endowment. We find strong negative effects of fiscal equalization on public goods contributions, in particular when groups with small MPCRs are involved.
Author(s): Jens Grosser (FSU and IAS, Princeton), Ernesto Reuben (Columbia University and IZA, Bonn), and Michael Seebauer (University of Erlangen-Nuremberg)
Topic: Public goods
Back to session: Public Goods 1
ABSTRACT:
In this paper we report the results of experiments designed to investigate the effects of public revelation of individual behavior on voluntary contributions to a public good. This work is motivated by recent studies that provide mixed evidence on the effects of public revelation. Lopez et al. (2010) found in a framed field experiment that the random revelation of the contribution of a single individual led to a significant and lasting increase in the provision of a public good. On the other hand, in a laboratory setting, Noussair and Tucker (2007) found that revealing the contributions of all subjects in the group initially resulted in higher contributions but this gain dissipated over time. To investigate the sources of these differences, we conduct the experiment of Lopez et al. (2010) in a laboratory setting. Moreover, to test the effects of the degree of public revelation we vary the number of subjects whose contributions are made public from zero of five in a group (our baseline), to one of five, three of five, and finally full revelation of all subjects' contributions. We find that public revelation always leads to higher contributions. However, the impact of the random revelation of one individual's choice is not as great as in Lopez et al. (2010). Revealing the choices of three and five individuals leads to higher contributions than revealing one individual's contribution, but there is not a significant difference between revealing the contributions of three or five subjects. We also observe decay in contributions for all levels of public revelation as in Noussair and Tucker (2007).
Author(s): John M. Spraggon, John K. Stranland, Lucia Andrea Vergara Sobarzo
Topic: Public goods, communication
Back to session: Public Goods 2
ABSTRACT:
In a linear appropriation game setting, this study examines behavior across treatments where the marginal damage from appropriation and the marginal private benefit of appropriation are varied in both symmetric and asymmetric treatments. In summary, subjects respond systematically to changes in marginal incentives. Individual appropriation levels are inversely correlated with the ratio of marginal external damage from appropriation to the marginal private benefit of appropriation. Moreover, no significant differences are observed in individual appropriation levels across treatments where the ratio of marginal external damage to the marginal private benefit is equal, independent of whether the games are symmetric or asymmetric. These findings are consistent with previous evidence for public goods games supporting the relevance of the marginal per-capita return in explaining variations in cooperation levels.
Author(s): Esther Blanco, University of Innsbruck; Maria Claudia Lopez, Pontificia Universidad Javeriana;James M. Walker, Indiana University
Topic: Public goods, cooperation
Back to session: Public Goods 2
ABSTRACT:
It has been shown recently that funding public goods through lottery mechanisms can secure higher levels of provision and greater social welfare than traditional voluntary contribution mechanisms. It has been previously been assumed that the prize is taken out of the raised funds, leaving the potential for such lotteries to either risk cancellation or actually lose money in the event of a deficit. This encourages fundraisers to secure prizes from corporate sponsors before using this strategy, offering advertising in return for prize donations. I theoretically examine the expected total prize offers from two competing firms under three fundraising solicitation strategies: an exclusive fundraiser with fixed advertising, where only the most generous offer is accepted; an non-exclusive fundraiser with fixed advertising, where all offers are accepted but the most generous firm is rewarded disproportionately; and a non-exclusive fundraiser with proportional advertising, where all offers are accepted and the firms share benefits in accordance with their relative contributions. After solving for the equilibria, I replicate each environment in the laboratory, with subjects playing the roles of both firms, who decide if and how much to contribute to the prize pool, and consumers, who observe prize amounts and make contribution decisions. While theoretical predictions suggest that exclusivity can be beneficial to the fundraiser, the experimental evidence contradicts this claim. In addition, both exclusive and non-exclusive fixed advertising environments generate smaller fixed prizes than does the proportional mechanism, regardless of the size of the advertising benefits.
Author(s): Michael LeGower
Topic: Public goods, games, IO
Back to session: Public Goods 2
ABSTRACT:
This paper presents an experimental study of two mechanisms that may influence incentives to manage common property resources. A decentralized peer punishment mechanism (swords) has been shown to increase cooperation in related social dilemmas, but only with linear private benefits and costs of public goods provision. We compare the effectiveness of this mechanism for a more realistic nonlinear public goods environment, in isolation and in combination with nonbinding communication about intended actions (covenants). The results suggest that swords alone do not increase cooperation or efficiency, although with covenants also possible peer punishment does marginally improve public goods provision. Covenants are much more effective in solving this social dilemma, and peer punishment is unnecessary if communication is possible.
Author(s): Timothy N. Cason and Lata Gangadharan
Topic: Public goods, cooperation
Back to session: Public Goods 2
ABSTRACT:
Many economic and personal interactions rely on trust and reciprocity. While this reliance often works well, it sometimes breaks down. The fallout observed in cases of business fraud and other malfeasance (by financial institutions in particular), shows serious economic consequences of failed trust and trustworthiness. Simple rules mandating minimum standards on reciprocation are attractive responses because they prevent the most egregious trust violations. However, minimum standard rules may also undermine behavior that otherwise would have produced higher overall economic welfare. As such, it is unclear whether stronger rules provide best responses to trust failures, or should be avoided. We use an experimental investment/trust game to test the efficacy of exogenously imposed minimum standard rules. We find that, in our experiments, rules fail to increase trust and trustworthiness. As a result, minimum standards can decrease economic welfare. Although sufficiently restrictive rules restore welfare, trust and trustworthiness never return.
Author(s): Thomas A. Rietz, Eric Schniter, Roman M. Sheremeta and Timothy W. Shields
Topic: Reciprocity
Link: "http://www.biz.uiowa.edu/faculty/trietz/papers/Trust,%20Reciprocity%20and%20Rules.pdf"
Back to session: Reciprocity
ABSTRACT:
Field studies of conflict report cycles of mutual revenge between groups, often linked to perceptions of intergroup injustice. We test the hypothesis that people are predisposed to reciprocate against groups. In a computerized laboratory experiment, subjects who were harmed by a partner's uncooperative action reacted by harming other members of the partner's group. This group reciprocity was only observed when one group was seen to be unfairly advantaged. Our results support a behavioral mechanism leading from perceived injustice to intergroup conflict. We discuss the relevance of group reciprocity to economic and political phenomena including conflict, discrimination and team competition.
Author(s): David Hugh-Jones, Martin Leroch
Topic: Reciprocity
Link: https://sites.google.com/site/davidhughjones/files/AVcomment.pdf?attredirects=0
Back to session: Reciprocity
ABSTRACT:
In finitely repeated games it is hard to distinguish true pro-social behavior from strategic instrumental behavior. In this paper we run repeated game versions of Berg at al (1995)'s well known trust game that enable us to study the relationship between instrumental (forward-looking) reciprocity and intrinsic (backward-looking) reciprocity. We find that i) people anticipate instrumental reciprocity and ii) that instrumental reciprocity reduce the level of intrinsic reciprocal behavior.
Author(s): Åshild Johnsen, Ola Kvaloy
Topic: Reciprocity
Back to session: Reciprocity
ABSTRACT:
This paper explores the implications of self-serving generosity for reciprocal behavior. Cox, Friedman & Sadiraj (2008) model of revealed altruism includes a partial ordering of opportunity sets, specifying that an opportunity set chosen by a first mover is 'more generous than' another opportunity set if it (a) increases the second mover's maximum payoff and (b) if the maximum payoff of the first mover is not increased by more than that of the second mover. The latter condition distinguishes between actions that clearly benefit the second mover and those that might involve ôself-serving generosityö of the first mover. To test whether self-serving generosity reduces the extent of reciprocal responses by the second movers we design an experiment that holds condition (a) constant and varies condition (b). We find that second movers' responses are unaffected by such a treatment change.
Author(s): Maros Servatka and Daniel Woods
Topic: Reciprocity
Back to session: Reciprocity
ABSTRACT:
Multiple Price List (MPL) elicitation method (Anderson et al. (Econometrica, 2008), and Holt and Laury (AER, 2002)) is commonly used in measuring individual risk aversion. MPL consists of pairwise lottery questions presented in a table which allows subjects to switch from safe to risky options. We show that MPL has a bias possibly related to individual loss aversion. To document this bias we use within-subject design in which participants answer both a single question chosen from one of the traditional MPL choices and the entire MPL. We find that distributions of the two responses to the identical question are different, which changes rankings of individual risk aversion as measured by the single question and the whole MPL. In addition, we explore loss aversion as a possible source of the bias and attempt to distinguish this explanation from other decision errors.
Author(s): Younjun Kim and Tanya Rosenblat
Topic: Risky choice
Back to session: Risk 1
ABSTRACT:
While the exhibition of risk aversion when the gamble involves small stakes (i.e., an amount that an individual would consider to be insignificant with respect to his or her wealth) is inconsistent with expected-utility theory, economists have seemingly found evidence of this behavior in numerous experiments. This inconsistency between the small stakes risk aversion researchers observe in the laboratory and the expected-utility theory model of risk aversion can be explained by either: (i) failure of the expected-utility theory to adequately describe behavior, or (ii) ascription of observed behavior to risk aversion when the behavior is actually motivated by some other factor. In this paper we will test the hypothesis that the small-stakes risk aversion observed in previous laboratory experiments has been primarily a manifestation of reference-dependant preferences. As experience in a market has been shown to moderate a number of behavioral regularities, including the endowment effect, a secondary goal of the present research is to identify whether small stakes risk aversion subsides in the presence of market experience in a controlled laboratory setting.
Author(s): Sean Crockett and B. Erin Fairweather
Topic: Risky Choice
Back to session: Risk 1
ABSTRACT:
We use probability list elicitation to experimentally study choices between lotteries in the Marschak-Machina triangle for a population of online workers. While probability list data illuminates aspects of preferences that are difficult to observe with binary choice data, we document that the former method significantly diminishes risk aversion. We show that this observation is consistent with a decision maker who has non-expected utility preferences, but when list elicitation is employed, reduces the compound lottery induced by her choices and the external randomization device used to determine payment. As a result, list elicitation distorts the inferences that can be drawn about non-expected utility preferences.
Author(s): David Freeman, Yoram Halevy, Terri Kneeland
Topic: Risky choice
Back to session: Risk 1
ABSTRACT:
One of the most consistent finding of human decision making is that subjects often make different choice from the same set of options -- a phenomenon referred to as random choice. Several theories have been proposed to account for such behavior. Amongst them: random utility, indifference, mistakes, and hedging/incomplete preferences. We design a controlled laboratory experiment to distinguish between these theories in a standard framework of choice over lotteries.
Author(s): Marina Agranov and Pietro Ortoleva
Topic: Risky choice
Back to session: Risk 1
ABSTRACT:
We experimentally examine how attitudes towards risk and ambiguity as well as other properties of the decision making process such as choice consistency and frequency of dominance violations change over the lifespan. We assess all of these characteristics of the decision process both in the gain and loss domain in a sample of 165 subjects that range in age from 12 to 90 years old. We also collected extensive demographic information and psychological measures that we use as co-variates in our analysis. Our results indicate that there is substantial and domain specific variability across subjects in all parameters with indications of age-related differences in both risk and ambiguity preferences as well as choice consistency and frequency of dominance violations.
Author(s): Agnieszka Tymula, Lior A. Rosenberg Belmaker, Lital Ruderman, Kirk Mason, Paul Glimcher and Ifat Levy
Topic: Risky choice
Back to session: Risk 2
ABSTRACT:
We modify the multiple price list design of Holt and Laury (2002) to study the trade-off between expected payoff and payoff’s skewness. We find that our subjects are equally split between those that are and those that are not averse to negative skewness, despite the fact that they all display risk aversion in the standard Holt and Laury (2002) treatment. We estimate the parameters of a power-expo utility function only on the subset of subjects, whose behavior across treatments is consistent with this functional form and find that relative risk aversion can be twice as large as what is typically found in these experiments.
Author(s): Anna Bassi, Riccardo Colacito, and Paolo Fulghieri
Topic: Risky choice
Back to session: Risk 2
ABSTRACT:
This paper shows that breaking down a monetary payoff to a sure and a risky component may induce an arbitrary reference point by creation of an endowment perceived by the subjects. We find that decisions are not neutral to this framing effect, suggesting possible policies for selling risky assets.
Author(s): Irene Comeig, Charloes A. Holt, Ainhoa Jaramillo-Gutierrez
Topic: Risky choice, anomalies
Back to session: Risk 2
ABSTRACT:
We examine how almost winning affects subsequent betting decisions. Using a unique data set from the game rapid roulette, we find that gamblers who place bets on numbers that almost won are less likely to place bets on those numbers on the next spin of the wheel. The representativeness heuristic provides an explanation for roulette player beliefs regarding near misses.
Author(s): James Sundali, Amanda Safford, Rachel Croson (presenting)
Topic: Risky choice
Back to session: Risk 2
ABSTRACT:
I consider whether subjects appear to acquire information in sequences consistent with expected utility when presented with a series of choices between risky options in a commonly-used method of eliciting risk attitudes (Holt and Laury, 2002). I collect data on subjects' information acquisition during real and hypothetical risky choices using process-tracing software called Mouselab. I find evidence to suggest that subjects acquire information in a manner consistent with the implicit calculation of expected utility. In other words, individuals do not merely make choices as if they are integrating probabilities with their outcomes, it appears that they actually are. Moreover, as they progress through the series of choices in the Holt & Laury Multiple Price List (HL MPL), their information acquisition becomes progressively more consistent with integration models. Finally, on average, individuals appear to acquire information in real and hypothetical settings in similar ways.
Author(s): Matthew P. Taylor
Topic: Risky choice
Back to session: Risk 3
ABSTRACT:
Financial economists have presented evidence that there are systematic differences in returns in financial markets that are linked to seasonal changes in the number of daylight hours. One hypothesis is that the cyclical returns are the result of variation in individual risk preferences. This paper explores whether risk preferences vary with the length of days while controlling for a large number of alternative explanatory variables.We find economically and statistically significant effects of daylight hours.
Author(s): Jonathan E. Alevy and Audrey Tanaka
Topic: Risky choice, anomalies
Back to session: Risk 3
ABSTRACT:
Between-subjects experiments on myopic loss aversion find that willingness to invest in a risky asset is higher for subjects who receive less frequent feedback. Researchers have therefore proposed changes in feedback institutions as a policy measure to affect risk taking, e.g., in retirement saving. We provide controlled evidence on the effects of such changes in feedback institutions. Our main finding is an asymmetric effect of changing feedback institutions. While subjects who go from a high to a low frequency of feedback increase their investment as expected, those who go from low to high frequency of feedback do not reduce their investment. Our results suggest that policy changes based on between-subjects findings may not always have the intended consequences because decision frames may only be malleable in one direction.
Author(s): Daniel Hilgers and Matthias Wibral
Topic: Risky choice, learning
Back to session: Risk 3
ABSTRACT:
We take Cumulative Prospect Theory (CPT) seriously, and rigorously estimate structural models from substantial laboratory experiments with undergraduate students and MBA students. We also estimate structural models from Expected Utility Theory (EUT) and Rank-Dependent Utility (RDU) for comparison. Our major finding is that a clear majority of individuals in our sample locally asset integrate. That is, they see a loss frame for what it is, a frame, and behave as if they evaluate the net payment rather than the gross loss when one is presented to them. This finding is devastating to the direct application of CPT to these data. We also find very little evidence for loss aversion in estimated models of CPT, but considerable evidence for loss seeking. Finally, we find that the preferences estimated under EUT and RDU models do show a significant shift when subjects are confronted with framed losses. It is not the case that preferences are invariant to losses, even if subjects appear to integrate the losses with their endowments. In effect, one has to move to a 'state dependent' specification of EUT and RDU to properly characterize these data. So loss frames matter, but not at all in the way assumed by CPT. Defenders of the CPT model claim, correctly, that the CPT model exists 'because the data says it should.' In other words, the CPT model was borne from a wide range of stylized facts culled from parts of the cognitive psychology literature. If one is to take the CPT model seriously and rigorously then it needs to do a much better job of explaining the data than we see here.
Author(s): Glenn W. Harrison and J. Todd Swarthout
Topic: Risky choice, anomalies
Back to session: Risk 3
ABSTRACT:
This paper examines experimentally whether an individual's desire to take risk is affected by the decision of others, his 'peers', and, if so, through which channels peer effects occur. We examine the two main channels of peer effects: social utility and social learning. Within social utility, we distinguish between the desire to obtain the same payoff as the peer, social regret, and that to make the same choice as the peer, conformism, by varying whether the peer is randomly allocated a lottery or actively chooses one. Both are in turn compared to anticipated social utility, whereby individuals are only informed ex-post about the choice of the peer. Further, social learning is investigated by introducing incomplete information about the lotteries. Our results show that individuals are significantly influenced by the choices of their peer: lottery choices change significantly when peers are present. Peer effects are driven by both social utility and social learning, and are strongly anticipated. Interestingly, a safe choice by the peer has the strongest effect: individuals are more likely to imitate a safe choice than a risky one.
Author(s): Amrei Lahno and Marta Serra-Garcia
Topic: Risky choice, learning,
Back to session: Risk 4
ABSTRACT:
In a lab experiment, we test consumption and portfolio choice predictions of expectations-based reference-dependent preferences (Köszegi and Rabin, AER 2009). The subjects make actual consumption choices in four consecutive periods. In the first period, subjects are endowed with initial experimental wealth. In turn, they decide each period how much experimental wealth to consume in that period or to save for consumption in future periods. Subjects can save wealth by storing it safely or investing it in a risky lottery. Consumption corresponds to time subjects can surf the web instead of doing an alternative boring task. Treatments include decisions under pre-commitment and under certainty.
Author(s): Michaela Pagel, Christopher Zeppenfeld
Topic: Risky choice, learning,
Back to session: Risk 4
ABSTRACT:
Dual process theory from psychology argues that both slow conscious reasoning and fast unconscious feelings affect decisions. Attempts to formalize dual process theory as an economic model in which people maximize the weighted sum of expected utility and feelings based utility (that is hypothesized to ignore probability) have proven controversial. This paper presents experiments that test dual process theory where the emotional content of decisions made under uncertainty varies.
Author(s): Dawn Schrader, Jubo Yan, William Schulze (Presenter)
Topic: Risky choice, learning,
Back to session: Risk 4
ABSTRACT:
This study uses a laboratory experiment to determine how an individual's willingness to take risks changes when other people are impacted by their decision. While many of the decisions we make primarily affect our own well-being (like saving for retirement) many of the risky behaviors generate externalities, which can either be positive (like starting a new business) or negative (like smoking). Understanding how these external effects impact decision making is difficult outside the lab since so many aspects of the situation change. I therefore use a common setting (a financial asset) and examine decisions in settings where there are positive spillovers from the investment, no effect, or negative spillovers.
Author(s): Angela C. M. de Oliveira
Topic: Risky choice, public goods
Back to session: Risk 4
ABSTRACT:
People prefer resource allocations that remove inequality, yet at the same time they have a strong tendency to maintain the status quo. These two motivations often conflict with each other. Here we designed an economic game (i.e., the redistribution game) to address a particular form of status quo motivation the motivation to prevent reversing the current socioeconomic hierarchy. In the redistribution game, participants were shown an unequal distribution of resources between two other people. Participants were given an option to transfer a fixed amount of resources from the advantageous side to the disadvantageous side. Choosing to transfer resources would invariably lead to the reduction of inequality. However, in some of trials, transferring resources would also lead to the reversal of the hierarchy. For example, the initial distribution is (4,1), that is, 4 for one person, and 1 for another person. The redistributor need to decide whether they want to transfer 2 from the advantageous side to the disadvantageous side or not to make the transfer. Thus, if participants just want to reduce inequality, they should choose to transfer, making the final distribution (2,3). In contrast, if participants are driven to maintain the hierarchy, they should choose not to transfer because making the transfer would lead to the reversal of the hierarchy.
Author(s): Xinyue Zhou, Wenwen Xie, Qing Yang
Topic: Social Preferences
Back to session: Social Preferences 1
ABSTRACT:
A large and growing literature has demonstrated that explicit incentives imposed by principals such as enforceable contracts, fines, and monitoring regimes can have detrimental effects on agent behavior. This literature, however, focuses on what can happen and provides little guidance about when detrimental effects will arise and when they should be incorporated into economic models. We investigate the hidden cost of control, a result in which adding enforceability to a principal-agent contract decreases agent effort. We show that principals are only harmed by imposing control when: (1) they have previously established a prosocial norm with their agent; (2) they impose control unilaterally on the agent; (3) the control is weak, in that it cannot induce significant effort from the agent; and (4) the agent is of a type that does not use control himself when acting as a principal.
Author(s): Judd B. Kessler and Stephen Leider
Topic: Social Preferences
Back to session: Social Preferences 1
ABSTRACT:
Some evidence suggests that people behave more pro-socially in small groups than in market-like situations. We construct an experiment in which people choose between allocations that affect their payoff and that of others. The choices of some participants are randomly selected to determine payoffs. We test whether people exhibit different other-regarding preferences depending on how the choice is framed. To mimic a market-like environment, we ask subjects to select a type of partner, either high or low. Selecting a partner of a given type effectively removes this pairing from other players. We compare this treatment to two alternatives where people are first assigned to groups of two high and two low participants. In one treatment, they are then asked to choose between a high and low partner. In the other, they are asked to choose between two payoff allocations for the four individuals. These two treatments make the implication of one's choice on others more salient. We find that most subjects pursue their self-interest, but high payoff participants behave more altruistically in small groups while low payoff participants display more invidious choices in the market-like environment. The implication is that while some efficiency can be achieved in small groups thanks to altruism, a market-like environment reduces good samaritan tendencies, possibly because the negative effect of one's choice on others is less salient.
Author(s): Michele Belot and Marcel Fafchamps
Topic: Social preferences
Link: http://ideas.repec.org/p/cex/dpaper/2012001.html
Back to session: Social Preferences 1
ABSTRACT:
Policymakers have a limited set of mechanisms available to enhance the private provision of public goods. One increasingly popular mechanism is to make public who contributes to the public good and who does not. An important question for policymakers hoping to turn up social pressure by making behavior public is whether such a policy too has a uniformly positive impact on behavior, or whether making behavior public might be ineffective or even backfire in low cooperation environments. In a laboratory experiment, we demonstrate that even groups with very low initial cooperation can benefit substantially from the social pressure that accompanies public display of behavior.
Author(s): Raj Chetty, Judd B. Kessler, and Monica Singhal
Topic: Social preferences, cooperation
Back to session: Social Preferences 1
ABSTRACT:
A handful of recent studies suggest that historically disadvantaged groups hold less sanguine beliefs about their ability than other groups. Such belief patterns may contribute to the persistence of inequality by, e.g., leading disadvantaged groups to shy away from competitive environments. If disadvantage is the result of factors unambiguously unrelated to ability, it is not known why, or if, such beliefs obtain. In particular, an important question not addressed by the existing literature is whether a history of disadvantage is necessary to generate such self-discriminatory belief patterns or, alternatively, whether current inequality is sufficient to undermine relative ability beliefs. This paper presents the results of a series of experiments addressing these questions. The first experiment randomly assigns earnings inequality in the lab and finds strong evidence supporting the notion that current inequality undermines relative ability beliefs. A second experiment separates ability beliefs from performance beliefs by randomly assigning both earnings inequality and whether participants' tasks are ability-intensive or effort-intensive, finding that beliefs in ability-intensive, but not the effort-intensive, task are sensitive to inequality. A third experiment investigates directly the implications for competition, with results differing by gender: disadvantaged male participants are less likely to select pay structures based on competition, while female participants' preferences for competition are unaffected by inequality. The results from all three experiments are largely consistent with separate, but closely related, theories in sociology and social psychology arguing that individuals have a deep-seated need and tendency to ex-post rationalize outcome inequality as being fair (e.g., 'blaming the victim.')
Author(s): Jeffrey V. Butler
Topic: Social preferences, anomalies
Back to session: Social Preferences 2
ABSTRACT:
Recent advances in economic thought, largely motivated by experimental findings, have led to the adoption of models of human behavior that not only regard one's own payoff (homo oeconomicus) but also the consequences for others, and their opinion of oneself. Investigations of deontological motivations, where an agent chooses among actions based on the action per se rather than its consequences, seem to be largely absent from the literature. In an experiment conducted in a setting between field and laboratory, we find that neither purely consequentialist nor purely deontological preferences can explain behavior. Rather a mixture of both seems to motivate behavior. Our experiment uses a global, computer-mediated labor market intermediary in order to mimic mod- ern economic contexts and investigate the effect of disintermediation on moral behavior. We identify consequentialism by varying the probability of individuals' decision being consequential. Probabilistic disintermediation increases selfish behavior, which counter- intuitively provides evidence of deontological preferences. Any social preference that is outcomes-based would predict no change in generosity with increasing disintermediation: when the donor's split is not chosen, donor's decisions have no effect on outcomes. If a mix of consequentialist and deontological preferences determine altruism, people become more generous as the likelihood of actually having to pay for the altruism decreases. If utility is additive, then only deontological preferences that are convex can explain our findings. Our results suggest psychological priming may underpin at least some of the patterns previously attributed to outcomes-based incentives.
Author(s): Daniel L. Chen and Martin Schonger
Topic: Social preferences, internet
Link: http://users.nber.org/~dlchen/papers/EgoismAltruism.pdf
Back to session: Social Preferences 2
ABSTRACT:
This research explores the relative influence of social image concerns and four widely discussed norms of distributive justice on sharing rules in tightly knit communities of a developing country. Specifically, I investigate how divisions of an effort-generated pie depend on whether the receiver in a dictator game with anonymously matched players knows the contribution and share taken by the dictator. To this end, I estimate the weight on selfishness relative to that on strict egalitarianism, inequality aversion, liberal egalitarianism, and libertarianism for a representative sample of Malawian villagers using data from a unique lab-in-the-field experiment.
Author(s): Annika Mueller
Topic: Social preferences,norms and fairness
Back to session: Social Preferences 2
ABSTRACT:
We present experimental evidence on policies that leaders choose depending on whether they were elected or appointed. Consistent with previous studies of the dictator game, we find that unitary policymakers do not always act selfishly, that is, choose a policy that maximizes their own payoffs. However, the way in which one became the leader matters. Leaders who are elected are significantly more likely to choose a policy not equal to their type than leaders who are appointed. Elected leaders who act non-selfishly will favor the voter rather than the losing candidate, while appointed leaders show no tendency to favor the voter over the losing candidate. Our results provide support for the view that non-selfish behavior of leaders reflects a reciprocity motive. They also show that candidates do not simply implement their own preferences once in office, as suggested by the basic citizen-candidate model.
Author(s): Allan Drazen; Erkut Y. Ozbay
Topic: Social preferences, voting
Back to session: Social Preferences 2
ABSTRACT:
In a laboratory experiment based in a career-concerns setting, we analyze how sharing a membership in a social group affects principals' assessments of the causes of agent performance. The treatments vary whether (artificially induced) group identification is specifically primed, whether effort is observable, as well as the in-group vs. out-group matching of the principals and agents. We elicit from the principals their beliefs about agents' type and effort. We find that principals' explanations of agents' performance vary depending on whether they share the social identity with the agents, and the variation is sometimes justified by the expectations of the effects of sharing group identification on agents' choices. However, we find that principals overestimate agents' willingness to exert effort when they share a group membership with the agent; this is especially the case when principals face low-competence agents.
Author(s): Dominik Duell (NYU) and Dimitri Landa (NYU)
Topic: Social preferences,norms and fairness
Back to session: Social Preferences 3
ABSTRACT:
This paper studies the effect of uncertainty on altruistic preferences. In our two-by-two design, dictators allocate either cash or a chance to win a larger prize between themselves and a receiver. We vary the price of giving across fifty allocation decisions for each participant, one of which is executed for payment. We find giving to be highly price-sensitive across treatments, but that altruism decreases significantly when dictators give up a chance at a larger prize, rather than cash. Our results suggest, among other things, that the deterministic relationship between decisions and outcomes in laboratory settings may be one reason for the high level of altruism found in experimental settings.
Author(s): David Owens
Topic: Social preferences,norms and fairness
Back to session: Social Preferences 3
ABSTRACT:
This paper uses a new theoretical approach to analyze how players form their preferences in 2x2 games. Using this approac
Author(s): Ryan Kendall and Daniel Jessie
Topic: Games, social preferences
Back to session: Social Preferences 3
ABSTRACT:
We conduct an experiment to examine the role of retribution and deterrence in motivating third party punishment and, in particular, how the role of these two motives may differ according to whether the third party is a group or an individual. In the experiment, third parties are empowered with the right to uphold or negate punishments proposed by players in a PD game. We vary the timing with which third party punishment decisions are displayed across treatments: some are imposed ex- post (after players decisions) and thus can have a deterrence effect while others are threatened ex-ante (before players decisions) and thus can be only retributive. In all cases, third parties must specify punishment amounts when approving punishment proposals. We find the timing of punishment not to impact individual third party decisions. On the other hand, we find third party groups to be significantly more likely to approve punishment that includes a deterrent effect. We also find that, in general, third party punishment is weaker when it is levied by a group than by an individual.
Author(s): Fangfang Tan (Max Planck Institute for Tax Law and Public Finance) and Erte Xiao (Carnegie Mellon University)
Topic: Social preferences,norms and fairness
Back to session: Social Preferences 3
ABSTRACT:
We introduce the mobbing game. A group of players coordinates on selecting a victim among them. Each player simultaneously nominates a player. Players can also choose not to nominate a player. If all players (except the victim) nominate the same player, then the victim's payoff is taken and distributed among the other players. We vary the efficiency of mobbing and the distribution of initial endowments. We find that groups of four players easily coordinate on a victim even in absence of focal points and efficiency gains from mobbing. Asymmetries facilitate coordination even further.
Author(s): Klaus Abbink, Gonul Dogan
Topic: games
Back to session: Tournaments
ABSTRACT:
Incentive schemes resulting in tournament competition have been criticized for inducing excessive risk-taking among investment managers. Another feature of tournament incentives is that they induce a particular kind of risk-taking: when underperforming it becomes optimal to invest in positively skewed idiosyncratic assets, whereas when outperforming a negatively skewed portfolio that correlates with the market would be optimal. We test these propositions with a lab experiment where experimental participants choose portfolios from among correlated, idiosyncratic, and positively, negatively or zero skewed assets. We show that tournament incentives indeed lead to a preference for positively skewed idiosyncratic portfolios when lagging in performance, and negatively skewed correlated portfolios when leading. However this effect shows only in the final periods of the treatment. With standard linear incentives coupled with relative performance feedback, we find the same effect persistent throughout the treatment. We conclude that myopic competitive preferences are enough to induce portfolio biases.
Author(s): Oege Dijk, Martin Holmen, Michael Kirchler
Topic: Tournaments, finance
Link: https://sites.google.com/site/oegedijk/TournamentIncentivesAug2012.pdf
Back to session: Tournaments
ABSTRACT:
Tournaments are widely used in organizations, explicitly or implicitly, to reward the best-performing employees, e.g., through promotion or bonuses, and to punish the worst-performing employees, e.g., through .ring or unfavorable job assignments. Although both types of tournament mechanisms are used in organizations, research to date has focused primarily on reward mechanisms. We set up optimal contracts in a principal-agent framework and experimentally compare the efficiency of three tournament incentive schemes reward tournament (R), punishment tournament (P), and reward & punishment tournament (RP) - which, respectively, reward the best performer, punish the worst performer, and do both. We also vary the size of the tournaments. The data do not support the theoretical prediction that the three mechanisms are efficiency-equivalent when agents are symmetric in their ability. Specifically, we find that the RP mechanism produces more aggregate effort than the R and P mechanisms when the group size is relatively large (n=6), and both the RP and P mechanisms perform better than R when the group size is small (n=3). We propose an explanation for these differences based on a reinforcement learning story. Lastly, we find that the variance of effort is lowest in the punishment tournament which should be of particular interest to managers who are concerned about the efficiency loss due to the high variance or the bifurcation
Author(s): Loukas Balafoutas, Glenn Dutcher, Florian Lindner, Dmitry Ryvkin, Matthias Sutter
Topic: Tournaments, learning
Back to session: Tournaments
ABSTRACT:
We investigate experimentally to which extent people trust and honor trust when they are playing with other people's money. We adopt the well-known trust game by Berg, Dickhaut and McCabe (1995), with the important difference that the trustor (sender) who sends money to the trustee (receiver) does this on behalf of a third party. We find that senders who make decisions on behalf of others do not behave significantly different from senders who manage their own money. But we find a gender specific treatment effect among the receivers. Women return significantly less money when senders send a third party's money than when senders send their own money, while there are no such treatment effects among men. Moreover, women return significantly less than men when the senders are managing a third party's money.
Author(s): Ola Kvaloy and Miguel Luzuriaga
Topic: Trust, gender
Back to session: Trust 2
ABSTRACT:
We experimentally investigate how the country origins of supply chain members China versus the U.S. impact trust, trustworthiness, and strategic information sharing behavior in a cross-country supply chain. We consider a two-tier supply chain in which the upstream supplier solicits demand forecast information from the retailer to plan production, yet the retailer has an incentive to manipulate her forecast to ensure abundant supply. Both the levels of trust and trustworthiness in the supply chain and the supplier's capability to solve for the optimal production quantity affect the efficacy of forecast sharing and the resulting profits. We develop a novel and rigorous experimental framework to disentangle these two aspects and to allow for real-time cross-country interactions. We determine that Chinese consistently exhibit lower trust and trustworthiness than their U.S. counterparts, regardless of the partners' country origin and the supply chain cost structure. Chinese individuals on average inflate forecast information 100% more than do U.S. ones, and Chinese also tend to rely less on the forecast report when determining the production quantity. In addition, both Chinese and U.S. individuals trust U.S. partners more than they trust Chinese ones, due to their perception that U.S. individuals are more trusting and trustworthy in general. Further, we show that Chinese individuals value long-term relationships to a greater extent than do U.S. ones. Our conclusions highlight the importance for companies to devote more time and effort to maintaining a trusting relationship with their Chinese partners than they do with U.S. ones.
Author(s): Ozalp Ozer, Yanchong Zheng, Yufei Ren
Topic: Trust, norms, IO
Back to session: Trust 2
ABSTRACT:
Using trust games, we study how promises and messages are used to build new trust where it did not previously exist and to rebuild damaged trust. In these games, trustees made non-binding promises of investment-contingent returns, then investors decided whether to invest, and finally trustees decided how much to return. After an unexpected second game was announced, but before it commenced, trustees could send a one-way message. This design allowed us to observe the endogenous emergence and natural distribution of trust-relevant behaviors and focus on naturally occurring remedial strategies used by promise-breakers and distrusted trustees, their effects on investors, and subsequent outcomes. In the first game 16.6% of trustees were distrusted and 18.8% of trusted trustees broke promises. Trustees distrusted in the first game used long messages and promises closer to equal splits to encourage trust in the second game. To restore damaged trust, promise-breakers used apologies and upgraded promises. On average, investments in each game paid off for investors and trustees, suggesting that effective use of cheap signals fosters profitable trust-based exchange in these economies.
Author(s): Eric Schniter, Roman Sheremeta, Daniel Sznycer
Topic: Trust, social preferences, reciprocity
Link: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2144021
Back to session: Trust 2
ABSTRACT:
In this paper, we show that a rather light modification to unanimity rule rules out the behavior that account for its adverse properties. The rule in question is unanimity with abstention and majority quorum. We show that, regardless of the sophistication of voters, this voting system weakly dominates unanimity in the setup of the Condorcet jury. We test these predictions in the lab. Additionally, we test the influence of framing effects, given that this rule is equivalent to majority rule with veto power for one of the alternatives.
Author(s): Laurent Bouton, Aniol Llorente-Saguer and Frederic Malherbe
Topic: Voting
Back to session: Voting
ABSTRACT:
This paper empirically tests theoretical predictions of voting equilibria derived from formal political economy models of mixed finance under different types of mixed public/private financing environments. The environments vary with respect to the relationship between the publicly and privately financed private good. The two relationships considered are `top-up' in which individuals can purchase the good privately to supplement their consumption of the publicly provided good and continue to pay proportional income taxes; and `opt-out' in which individuals can consume the publicly provided good or purchase a private alternative instead of the publicly provided good while continuing to pay income taxes. By using experimental methods the paper provides an explicit test of the conjecture that private finance reduces public support for the publicly financed private good. Using non-parametric local linear regression avoids issues of model misspecification and allows insights not possible using common parametric methods.
Author(s): Neil Buckley, David Cameron, Katherine Cuff , Jeremiah Hurley,Stuart Mestelman, Stephanie Thomas
Topic: Voting
Back to session: Voting
ABSTRACT:
We analyze the informational properties of debates preceding binary agenda voting in games of persuasion. Agents vary with respect to what arguments they find convincing. The key prediction is that debate is fully revealing without delay (i.e., before voting on the agenda commences) regardless of debate rules and voting agendas if and only if speakers have the ability to restart the debate before every vote on the voting agenda (the condition we call ôopen-debate voting). We evaluate this prediction in a laboratory experiment. The treatments compare, across different agendas and debate rules, open-debate voting with voting in which all debates must conclude before the first binary vote on the agenda commences. While subjects' tendency to over-speak weakens the behavioral effect of open-debate voting, the aggregate pattern of results provides a broad support for the equilibrium prediction.
Author(s): Catherine Hafer (NYU) and Dimitri Landa (NYU)
Topic: Voting, communication
Back to session: Voting
ABSTRACT:
Political campaigns spend millions of dollars each voting cycle on persuading voters, and it is well established that these campaigns do affect voting decisions. What is less understood is what element of campaigningùthe content of the message or the delivery method itselfùsways voters, a question that relates back to how advertising works generally. We use a field experiment in a 2010 general election for local office to identify the persuasive mechanism behind a particular form of campaigning: candidate door-to-door canvassing. In the experiment, the candidate either canvassed a household or left literature without meeting the voters. In addition, the literature either contained information on the candidate or on how to vote. Our main results are that voters are persuaded by personal contact (the delivery method), but we find no evidence supporting the importance of messages in political persuasion. In our setting, personal contact seems to work, not through social pressure, as has been found in other research on persuasion, but by providing a costly or verifiable signal of quality.
Author(s): Jared Barton, Ragan Petrie, Marco Castillo
Topic: Voting, field experiment
Back to session: Voting
ABSTRACT:
The attraction effect (or asymmetric dominance effect) occurs when adding an asymmetrically dominated alternative to a choice set increases the share of people who choose the alternative that dominates it. Previous studies have provided strong evidence suggesting that the attraction effect can play an important role in individual decision making. With this project, we extend the attraction effect to strategic decision making. In particular, we experimentally examine several classes of 2x2 games and investigate whether adding an asymmetrically dominated third alternative affects behavior in the direction predicted by the attraction effect.
Author(s): Jona Linde, Joep Sonnemans, Roel van Veldhuizen
Topic: Anomalies, games
Back to session: Anomalies 1
ABSTRACT:
This paper reports the results of online experiments designed to test whether and to what extent the incidence of cognitive fallacies increases when people interact on social networks. In particular, we hypothesize that imitation, herding, and pressure to conform, likely to arise on social networks, may amplify judgmental errors with respect to more traditional lab environments. For this purpose, we use the very well known Linda problem first studied by Kahneman and Tversky (1983) and we set up an experiment on Facebook: in the “Private” treatment of the experiment, participants must send their answers to the Linda problem in the form of a private message to the experimenter, while in the “Public” treatment the answers must be posted on the experiment public “wall” so that each participant can see previous answers before posting his own. We then replicate the same experiment on Mechanical Turk (an online labor market) adding monetary incentives. Our results can be summarized as follows: in the Facebook experiment, contrary to our hypothesis, there is no significant difference in the percentage of correct responses between the Private and the Public treatment, and in both the Private and the Public treatment the majority of respondents identify the correct answer, thus failing to exhibit the conjunction fallacy. On Mechanical Turk, on the contrary, roughly 70% of respondents commit the conjunction fallacy, a datum which closely replicates previous conjunction fallacy experiments conducted on Mechanical Turk and on Internet boards (Paolacci et al. 2010), although the violation rate is significantly different from that reported in Kaheman and Tversky (1983). Introducing a bonus for the correct answer improves performance by more than 50% in the Private treatment, but not in the Public treatment, suggesting that imitation may have a detrimental effect on MT. We explain the apparently anomalous findings from the Fecbook experiment with a combination of self-selection and willingness to signal one’s ability to his peers.
Author(s): Giovanna Devetag, Francesca Ceccacci, Paola De Salvo
Topic: anomalies, internet, networks
Link: none available
Back to session: Anomalies 1
ABSTRACT:
Equilibrium Selection, Similarity Judgments and the Nothing to Gain/Nothing to Lose Effect Rubinstein (1988, 2003) and Leland (1994, 1998, 2001, 2002) have shown that many observed violations of expected and discounted utility can be explained if it is the case that people employ similarity judgments to guide risky and intertemporal choices. In this paper, I show that this decision process also explains which equilibria will be selected in single-shot games with multiple equilibria and implies that play in games will be associated with anomalies in risky choice. Data supporting these predictions is presented.
Author(s): Jonathan W. Leland
Topic: Anomalies, coordination
Back to session: Anomalies 1
ABSTRACT:
Experimental tests of dynamically inconsistent time preferences have largely relied on choices over time-dated monetary rewards. We investigate choices over effort in a longitudinal experiment, eliminating often discussed confounds present in monetary studies. We find subjects make initial allocations of effort to a future date that are significantly higher than their subsequent reallocations. Such present bias in allocation of effort has predictive power for demand of a laboratory-offered commitment device, validating a key implication of models of dynamic inconsistency. A companion monetary discounting study shows limited present bias.
Author(s): Ned Augenblick,Muriel Niederle,Charles Sprenger
Topic: Anomalies, time preference
Back to session: Anomalies 2
ABSTRACT:
This paper investigates the effect of past prices on investors' behavior in the lab, where participants buy or sell stock shares according to time series of stock prices (displayed in both the form of graph and table) and real-time estimates of stock's fundamental value (presented in the form of table) over 30 periods. Unlike most assets market experiment, prices here were not determined by the trading actions of subjects. Instead, the prices, along with the estimates of stock's fundamental value provided by financial analysts, were sampled from historical records of real market, because it is of interest to isolate the impact of price history on investors' behavior from the process of price formation. In this ôatomisticö partial-equilibrium scenario described above, I apply between-subjects design, in which subjects were randomly assigned to two treatments, and the only difference between treatments is the certain feature of past price that is of interest, such as preceding price trend, or the previous highest or lowest price. Comparison of trading behaviors between the treatments can provide conclusive results on how these features of past price influence investors' behavior. It is found that, if stock's fundamental value stays constant, past upward trend causes investors to sell stock, acting as ôcontrariansö, and investors sell or buy stock when price approaches previous highest or lowest price. If the fundamental value of stock has been changed, selling or buying when price returns to previous peak or trough becomes weaker. According to subjects' answers to the post-experiment questionnaire, and individual decisions at certain prices, it is proposed that, the finding of contrarian selling following upward trend can be explained by individual's tendency to sell stocks that have gained profit (disposition effect). In addition, selling or buying stocks at recent highest or lowest price is caused by subjects' heuristics of anchoring on these price levels, which tends to be attenuated given changed fundamental value. Above results offers innovative behavioral explanations for the phenomena of trend and trend reversal in financial market, and the predictive power of technical analysis.
Author(s): Chao Jiang
Topic: Finance, anomalies
Back to session: Anomalies 2
ABSTRACT:
Normalization is a canonical feature of neural computation required for the efficient coding of information, and has been demonstrated in value signals that are believed to implement decision-making. We demonstrate that a neuroeconomic choice model employing normalization predicts that the composition and size of the choice set influences choice behaviour, including specific patterns violating the Independence of Irrelevant Alternatives axiom. We then show behavioural data from both monkey and human choosers demonstrating the contextually-dependent choice behaviour predicted by this neuroeconomic normalization model. Finally, we argue that choice inconsistencies predicted by the model may well be normatively optimal if the resources required to implement choice behaviour in finite neural systems is considered.
Author(s): Ryan Webb, Kenway Louie, Paul W. Glimcher
Topic: Neuro, anomalies
Back to session: Anomalies 2
ABSTRACT:
Evolutionary dynamics provides an iconic relationship — the periodic frequencies of a game are determined by the payoff matrix of the game. This letter reports the first experimental evidence confirming such a relationship. In this investigation, we use two-population random-matched 2 x 2 games with 12 different payoff matrixes (Selten and Chmura, 2008). We find, in different games, the observed periodic frequencies of cycles are different. Interestingly, the experimental frequencies have a significant (larger than 5 sigma) linear positive correlation with theoretical frequencies which are calculated explicitly from payoff-normalized replicator dynamics—one of the simplest evolutionary dynamics.
Author(s): Bin Xu, Shuang Wang, Zhijian Wang
Topic: Cycles, games
Link: http://arxiv.org/abs/1208.6469
Back to session: Cycles
ABSTRACT:
The rotations introduced by Shapley (1964) as never ending cycles are long time expected. To investigate the rotations, we use the experimental data from 7 laboratory games with each consists 13 groups, and 825 rounds. The experimental games are of random matching two-person zero-sum extended 2x2 games (Binmore, Swierzbinski and Proulx, 2001 [1]). We find, the directions and strengths of social rotations are changed simultaneously when the payoff matrix is switched. The directions of the rotations can be captured by evolutionary dynamics. Meanwhile, the observed rotation strengths are related to the eigenvalues calculated from the Jacobian of normalized replicator dynamics. In addition, we find that, every group has its own distinguishable coefficients. (http://arxiv.org/abs/1203.2591)
Author(s): Zhijian Wang and Bin Xu
Topic: Games
Link: http://arxiv.org/abs/1203.2591
Back to session: Cycles
ABSTRACT:
We report laboratory experiments that use new, visually oriented software to explore the dynamics of $3\times 3$ games with intransitive best responses. Each moment, each player is matched against the entire population, here 8 human subjects. A ``heat map'' offers instantaneous feedback on current profit opportunities. In the continuous slow adjustment treatment, we see distinct cycles in the population mix. The cycle amplitude, frequency and direction are consistent with standard learning models. Cycles are more erratic and higher frequency in the instantaneous adjustment treatment. Control treatments (using simultaneous matching in discrete time) replicate previous results that exhibit weak or no cycles. Average play is approximated fairly well by Nash equilibrium, and an alternative point prediction, ``TASP'' (Time Average of the Shapley Polygon), captures some regularities that NE misses.
Author(s): Timothy N. Cason, Daniel Friedman, Ed Hopkins
Topic: Games, learning
Link: http://www.krannert.purdue.edu/faculty/cason/papers/rps-continuous.pdf
Back to session: Cycles
ABSTRACT:
We examine, both through a formal game theoretic model and through a laboratory experiment, four institutions for information exchange. Our paper's aim is to examine how consumers provide information on products they have experience with to ratings. Using as baseline a decentralized institution in which provision of information by senders incurs a private cost, we examine three relevant transfer mechanisms that incentivize rating provision: a transfer from receivers; from the commerce platforms conditional on a sale; and from individual manufacturers conditional on a sale of their product. Each transfer mechanism has differing effects on both the number and accuracy of sent ratings.
Author(s): Jonathan Lafky and Alistair J. Wilson
Topic: Games
Back to session: Games 1
ABSTRACT:
In society, power is often transferred to another person or group. Nash (2008) studied the evolution of cooperation among robot players through a coalition formation game with a non-cooperative procedure of acceptance of an agency of another player. Motivated by this paper, we conduct a laboratory experiment on finitely repeated three person coalition formation games. Human players with different strength according to the coalition payoffs can accept a transfer of power to another player, the agent, who then distributes the coalition payoffs. We find that the agencies method for coalition formation is quite successful in promoting efficiency. However, the agent faces a tension between short term incentives of not equally distributing the coalition payoff and the long term concern to keep cooperation going. In a given period, the strong player in our experiment often resolves this tension approximately in line with the Shapley value and the nucleolus. Yet aggregated over all periods, the payoff differences between players are rather small, and the equal division of payoffs predicts about 80% of all groups best. One reason is that the voting procedure appears to induce a balance of power, independent of the individual player's strength: Selfish subjects tend to be voted out of their agency, and are further disciplined by reciprocal behaviors.
Author(s): John Nash, Rosemarie Nagel, Axel Ockenfels, and Reinhard Selten
Topic: Cooperative games
Back to session: Games 1
ABSTRACT:
Abstract: From transportation to Internet communications, modern life is enveloped in network traffic in which multiple independent users make strategic choices over congestible routes. While there has been phenomenal progress in the theoretical study of routing in traffic networks in computer science, economics, and transportation, most developments are based on game-theoretic analysis that assumes selfish behavior by rational agents. We report a laboratory experiment in which large groups of subjects repeatedly and independently made route choices in a computerized traffic network with linear latency functions for payoff contingent on performance. In one condition (ROUTE), on each round of play each player commits herself to a single route; in a second condition (SEGMENT), she only commits herself to the next road segment. We construct and then test the equilibrium predictions of traffic flow in the two conditions. Despite the complexity of the network, observed traffic in both conditions rapidly approaches equilibrium.
Author(s): Amnon Rapoport, Eyran J. Gisches, Vincent Mak
Topic: IO, traffic
Back to session: Games 1
ABSTRACT:
In a linear appropriation game setting, this study examines individual and group behavior across treatments, where the opportunity cost of conservation is varied in both symmetric and asymmetric treatments. These effects are explored in situations where there is certainty over the opportunity costs, as well as uncertainty, where probabilistic degradation of the commons is linked to the degree of appropriation. In summary, subjects respond systematically to changes in the marginal incentives, as well as to the possibility of degradation. These responses are shown to be related to a direct effect of changes in marginal monetary incentives, as well as to an indirect effect associated with changes in subject's forecasts of the appropriation decisions of others.
Author(s): Esther Blanco, Maria Claudia Lopez, James M. Walker
Topic: Games, cooperation
Back to session: IEAs and the Commons
ABSTRACT:
In this paper, we examine the formation of international environmental agreements (IEAs) based on theoretical analysis and experimental evidences. Our theoretical results are as follows: In a framework based on heterogeneous players (and with certain particular conditions), we find that a unique stable coalition can exist. Specifically, we find that countries with a high marginal benefit of abatement are more willing to join an IEA, whereas those with a low marginal benefit have little incentives to join. Our experimental results show that the size of coalition is often larger than the equilibrium size predicted by our theory. Hence, the overall payoff and the abatement levels in our experiment are often higher than our theoretical prediction. Furthermore, we find that some players with a low marginal benefit join the coalition even though not joining is their weakly dominant strategy. We therefore argue that economic incentives are not the only factors that decide whether an individual wants to join a coalition or not. Several other hypotheses about players' behaviour (e.g. the issue on fairness; beliefs about cooperation and punishment threat) are also proposed and tested.
Author(s): Yu-Hsuan Lin, Bipasa Datta
Topic: IO, games, IEA
Back to session: IEAs and the Commons
ABSTRACT:
During the past round of negotiations of the UN Climate Change Conference in Durban, all parties committed to a legally binding International Environmental Agreement for Climate Change (IEACC) that will limit the emissions of green-house gases (GHG) of all countries. The concentration of GHG is a perfect case of a pure global public good, whereby parties' pledges for GHG emission reductions when negotiating the details of an IEACC can be thought as voluntary contributions to provide such a public good.
Author(s): J. Cristobal Ruiz-Tagle
Topic: Public goods, IEA
Back to session: IEAs and the Commons
ABSTRACT:
I develop a model of herding with capacity constraints called the Restaurant Game and study play of this game in a lab experiment. The model predicts and the experiment confirms that capacity constraints attenuate herding, with the strength of the effect dependent on the cost of choosing an option after its capacity has been reached. The difference between the rational and Level-k predictions in the game depend on the size of this cost, with a maximum possible difference between rational and Level-1 benchmarks when the cost is high. Hence, I can more easily distinguish rational from boundedly-rational behavior in the high-cost setting than in the standard no-cost setting, and the experimental design allows within-subject comparisons across these settings. I find that subjects using strategies closer to Level-1 than rational in the high-cost setting are more likely to use strategies closer to Level-1 in the standard setting. Furthermore, I find evidence that proximity of behavior to the rational prediction is positively correlated with measures of a subject's cognitive ability. I also find that reducing strategic uncertainty by substituting rational computer players for human subjects increases observational learning.
Author(s): Matthew T. Jones
Topic: Games, coordination
Back to session: Level-K 1
ABSTRACT:
We employ a novel experimental design that allows us to manipulate higher- and lower-orders beliefs about payoffs in a complete information framework. This allows us to cleanly estimate subjects' levels of rationality and features of the level-k solution concept without imposing restrictions on the anchoring L0 behavior. We classify subjects into three epistemic types: (1) boundedly rational (2) highly rational and (3) highly rational with mutually consistent beliefs. We find that mutually consistent beliefs are not a significant feature of individual reasoning, even when subjects are highly rational. In addition, we provide insight into the level-k solution concept in terms of the type distribution, stability of levels and the L0 behavior.
Author(s): Terri Kneeland
Topic: Games, learning
Back to session: Level-K 1
ABSTRACT:
In this paper we investigate how cognitive ability influences behavior, success and the evolution of play towards Nash equilibrium in repeated strategic interactions. We study behavior in a p-beauty contest experiment and find striking differences according to cognitive ability: more cognitively able subjects choose numbers closer to equilibrium, converge more frequently to equilibrium play and earn more even as behavior approaches the equilibrium prediction. To understand better how subjects with different cognitive abilities learn differently, we estimate a structural model of learning based on level-k reasoning. We find a systematic positive relationship between cognitive ability and levels; furthermore, the average level of more cognitively able subjects responds positively to the cognitive ability of their opponents, while the average level of less cognitively able subjects does not respond at all. Our results suggest that, in strategic environments, higher cognitive ability translates into better analytic reasoning and a better `theory of mind'.
Author(s): David Gill (Oxford); Victoria Prowse (Cornell)
Topic: Cognitive, behavior
Link: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2043336
Back to session: Level-K 1
ABSTRACT:
This paper reports results from a laboratory experiment that investigates the Meltzer-Richards model of equilibrium tax rates, income redistribution, and the growth of government. In our experiment, subjects operate in two environments: political environment, in which the level of taxation is determined, and the labor market (economical environment), in which given the level of redistribution and the assigned productivities citizens choose labor supply that generates pre-tax income. We implement the model under two different competitive political institutions: direct democracy and representative democracy, and compare the level of taxation that emerges in these two institutions. In a different exercise, we fix the political environment and vary the inequality level, by varying the initial productivities of the citizens, to study the link between inequality and the redistribution.
Author(s): Marina Agranov and Thomas Palfrey
Topic: Voting, social preferences
Back to session: Political Economy
ABSTRACT:
Electoral clientelism and vote buying phenomena where leaders exchange favors or money for political allegiance are widely perceived as major obstacles to economic development. In well developed democracies, the selection of leaders who contribute to the public good, rather than focus only on narrowly targeted transfers to a few supporters, may play part in coordinating voters around good policies. On the other hand, in less developed democracies, political norms may revolve around reciprocity, where voters may disproportionately reward private transfers and disregard a politician's performance in delivering public goods. This paper studies connections between the emergence of clientelism and features of the socio-political environment within which leadership decisions occur. Our findings may provide insight on types of political-economic interventions useful in promoting economic development by orienting leadership towards broadly socially beneficial policies.
Author(s): James Habyarimana, Daniel Houser, Stuti Khemani and Seung Ginny Choi
Topic: Political economy
Back to session: Political Economy
ABSTRACT:
Authority, and the behavioral response to authority, is central to many important questions in public economics, but has received insufficient attention from economists. In particular, research has not differentiated between legitimate coercive power and the presumption of expert knowledge, what we call ôauthority to and authority in. In this paper we report on the results of a series of lab experiments designed to distinguish the effects of the two sources of authority on contributions to a public project. The results suggest that authority to and authority in interact in ways not heretofore understood. Coercion without expert explanation does not increase voluntary contributions, nor does explanation without coercion, but together they induce more contributions than any other combination of policies. We interpret these findings to indicate that the reaction to an authority depends on whether that authority is perceived to be legitimate.
Author(s): Dan Silverman, Joel Slemrod, Neslihan Uler
Topic: Public Goods
Link: http://www-personal.umich.edu/~neslihan/Uler_Authority_taxevasion.pdf
Back to session: Political Economy
ABSTRACT:
It is well-known that efficient planning for groups, organizations or individuals requires time consistency and stationarity of preferences across time. Due to the stakes and con- sequences implied by group decision making, stability of preferences and the existence of future reversed decisions are particularly important. When an efficient set of finan- cial market exists, the invisible hand provides the right incentives to insure stability, consistency and social efficiency of intertemporal choices. Things are less clear cut for intertemporal decisions that are to be done in the absence of a coordinating financial market. In particular, the basic determinants of intertemporal group decision making in the absence of formal financial markets, such as the collective degree of impatience or the decision processes, remain to a large extent unknown. As it is emphasized in Char- ness and Sutter (2012), a review of the literature regarding the difference between group decision-making and choices at the individual level tend to show that individuals are more likely to be influenced by various biases or cognitive limitations, whereas groups are more likely to make choices that follow standard economic theory predictions. A huge number of experiments investigate time-preferences for individuals (among others, see Frederick et al., 2002 ; Percoco and Nijkamp, 2009 ; Noor, 2010), but very few exist about collective time preference. A recent debate at the theoretical level concerns the issue of collective dynamic choice (Jackson and Yariv, 2011). In this paper, we propose an experimental study of three elements of collective decision making over time: the pattern of valuation of futures outcomes (discounting), the stability of collective preferences (time consistency) and the informational exchanges between individuals (decision process). To this aim, we build an experimental design based on mixed longitudinal and cross-sectional procedures in order to elicit the pattern of time preferences. Subjects subsequently participate to an individual part and to a collective part by using the same elicitation method for both parts. After the individual part, participants were matched randomly in a group of five members and have to reach a common decision regarding a temporal prospect. Major- ity voting is the basic decision process for group decision making. In order to get some information in the way individuals interact with others preferences, we used a 5-stage voting procedure. This procedure is based on a message game where individual reveal some information about their preferences over collective outcomes to the other members of the group before the final vote. Sixty subjects took part in the experiment, each being informed that they would have to participate to three successive experimental sessions, scheduled at regular intervals (4 weeks). Before entering the lab, participants had to confirm their will to engage in the whole set of three experimental sessions.
Author(s): Laurent Denant-Boemont, Enrico Diecidue and Olivier l'Haridon
Topic: Time preferences
Back to session: Preferences
ABSTRACT:
Entrepreneurs are often thought to be risk-taking individuals, who are willing to risk present resources for a chance of a future reward. Entrepreneurs are also thought to be extraverts, optimistic, and overconfident about their relative skills. Are these assumptions true? To test these assumptions, we designed a survey-basedö experiment with self-identified entrepreneurs to measure their risk attitudes, motivation, optimism, overconfidence and other relevant characteristics. Most of the existing knowledge regarding traits and characteristics of entrepreneurs has been qualitative and inconclusive. This study would be the first to scientifically establish a correlation between individual characteristics and entrepreneurship using experimental methods.
Author(s): Bing Jiang and C. Monica Capra
Topic: Homegrown preferences, cooperation
Back to session: Preferences
ABSTRACT:
There is much anecdotal evidence for conspicuous consumption - the willingness to pay for certain goods because they are expensive and therefore signal wealth, social status, or some other desirable attribute. However, formal empirical investigation of this phenomenon has been limited. While some studies have attempted to gauge the importance of conspicuous consumption using field data, that evidence is difficult to interpret, and not sufficiently rich to permit tests of specific theories. An alternative is to induce and explore conspicuous consumption in a controlled laboratory experiment. To our knowledge, this is the first study that pursues that strategy. We report the results of an experiment in which subjects were provided with opportunities to purchase a publicly observable good (entrance to a lottery game that was played out at the end of the session), and where the prices charged for that good depended on the private results of a dictator game, in which the subjects divided cash between themselves and the American Red Cross. Consistent with the implications of signaling explanations for conspicuous consumption, we find that high givers are willing to pay more for the publicly observed good when the good is more costly for low givers (and hence more exclusive). In contrast, the willingness-to-pay among low givers is not sensitive to the price paid by high givers. The latter finding implies that, unlike high givers, low givers are not concerned about reputations for generosity. We also estimate a structural model of preferences that include a desire to signal generosity, which allows us to evaluate welfare effects. The positive correlation between generosity and concern about reputation for generosity implies that the public observability of consumption is more likely to be welfare-enhancing.
Author(s): James Andreoni, B. Douglas Bernheim, Christine Exley, Jeffrey Naecker, and Paul Wong
Topic: Social preferences, anomalies
Back to session: Preferences
ABSTRACT:
Using incentivized economic experiments we examine the factors that affect rewards and punishments in a trust relationship. Furthermore, the impact of reward and punishment schemes in fostering trust will also be measured. We use a modified two person trust game to understand how individuals make punishment and reward decisions in trust relationship, how these decisions change when (1) expectations are violated; (2) expectations are common knowledge, and (3) moral hazard exists, and how these decisions subsequently impact trust and reciprocity. Furthermore, the experimental design allows for investigation into rewards behavior, and the extent which rewards are conditional upon adherence to the same two factors. Finally, the magnitude of reward/punishment behavior is observed, to test the hypothesis that rewards have smaller magnitudes than punishment.
Author(s): Quoc H. Tran and Sheheryar Banuri
Topic: Communication, trust
Back to session: Trust 1
ABSTRACT:
The effect of market on people's behaviors has been widely studied, questions remain as to what mechanisms are at play. In this paper, we hypothesize that thinking about money focuses people's attention on self-interest thereby increases defection and decreases trustworthiness. We implemented a modified trust game with communication, and the preliminary results support our hypothesis.
Author(s): Jingnan (Cecilia) Chen; Daniel Houser
Topic: Trust, cognition
Back to session: Trust 1
ABSTRACT:
Empirical data has shown that those with high levels of anxiety suffer from not only health problems but also tend to have impaired economic outcomes. While it seems likely that these poor outcomes are the result of differences in decision-making by individuals with anxiety, the mechanisms by which anxiety affects economic decisions remain relatively unexplored. For example, successful working relationships rely on some amount of trust in one's co-workers: how does anxiety affect an economic agent's willingness/ability to build economic relationships requiring trust? Such environments are influenced by factors like betrayal aversion - expectations about the negative emotions associated with the knowledge being betrayed in a trust game - and anxious people could be more likely to be betrayal averse. However, just as anxiety's relationship to trust remains unknown, no studies have looked at whether this interaction between anxiety and betrayal aversion exists.
Author(s): Jason A. Aimone, Sheryl Ball, and Brooks King-Casas
Topic: Emotions and Choice
Back to session: Trust 1
ABSTRACT:
The identification of compound risk attitudes and ambiguity attitudes has recently received experimental support (Halevy, 2007) and been incorporated in decision models (Seo, 2009; Halevy and Ozdenoren, 2008; Segal, 1987). Non reduction of compound lotteries is this literature's explanation of Ellsberg type behavior. We conduct an experiment measuring individual behavior under simple risk, under various types of compound risk and under ambiguity. We examine how each of these behaviors changes as the probability (or size) of the winning event varies. We find that attitudes towards all three types of uncertainties move from seeking to aversion as the probability level increases. Controlling for probability level, we find that the link between ambiguity and compound risk attitudes is partial and sensitive to the type of compound risk considered. We do not support the equivalence between reduction of these compound risks and ambiguity neutrality.
Author(s): Mohammed Abdellaoui, Peter Klibanoff & Laetitia Placido
Topic: Ambiguity, risky choice, learning
Link: http://www.kellogg.northwestern.edu/faculty/klibanof/ftp/workpap/AKP2011.pdf
Back to session: Ambiguity 1
ABSTRACT:
We use a mechanical randomization device with an unknown distribution (to both subjects and experimenters) to elicit choices in an ambiguous environment. We use this device in order to eliminate possible 'strategic' ambiguity related to subjects' beliefs about the experimenters' motivations. What remains should be closer to the 'mechanical' ambiguity of unknown natural phenomena.
Author(s): Joerg Oechssler and Alex Roomets
Topic: Ambiguity, risky choice, learning
Back to session: Ambiguity 1
ABSTRACT:
Risk equivalents to uncertain environments are elicited in an experiment that controls the instruction level subjects receive. We find much higher frequency of probabilistically sophisticated behavior than in previous experiments that elicited certainty equivalents, and establish a causal relation between reduction of compound objective lotteries and ambiguity attitude.
Author(s): Yoram Halevy
Topic: Ambiguity, risky choice, learning
Back to session: Ambiguity 1
ABSTRACT:
We conduct a field experiment to establish the extent to which gifts to different charitable organizations are substitute goods. To do so, we solicit individuals to contribute to three charities A, B and C. The fundraising campaigns for A and B are separated by less than a week, while the fundraising campaign for C is conducted three months later. The subject pool consists of people approached to contribute to charity A. It is partitioned in three groups: regular donors, occasional donors and prospects. Unlike subjects in the control treatment, subjects in the matching treatment are offered a 1:1 match for charity A. The question addressed in the paper is therefore: How does lowering the price of A, affect giving to charity A, B and C?
Author(s): Olivier Armantier
Topic: Charitable giving, field experiment
Back to session: Charitable Giving 2
ABSTRACT:
Motivated by the success of a recent multi-stage charity raffle, we theoretically and experimentally evaluate these raffles and compare them to traditional ones. In a multi-stage raffle, individuals can buy tickets in the first round for a low price. In later rounds, all but a few tickets are eliminated at which point the same individuals can buy more tickets at a higher price. We show theoretically that charitable individuals should buy tickets in the early round, despite the low probability of winning, because this primes the pump of raffle competition, even among those who are purely self-interested, in later rounds. In our experiment, we investigate whether priming occurs, the role of learning, and compare the revenue generated in multi-stage raffles to traditional single-stage ones.
Author(s): Sebastian Goerg, John Lightle, Dmitry Ryvkin
Topic: Charitable giving, learning
Back to session: Charitable Giving 2
ABSTRACT:
Charity ratings organizations gather information about charity quality and provide that information to otherwise uninformed donors. There is empirical evidence that these organizations impact giving, but how do they impact welfare more generally? In this paper, I suggest that the mechanism through which informed third-party organizations impact welfare in a public goods game -- as charity raters might in a charitable giving setting -- is through their ability to restrict information. In particular, by pooling the messages they send donors, donors may contribute when they would otherwise free-ride, as they are unable to distinguish between 'the best' charities and other charities that are also socially efficient. However, this suggestion depends on rather narrow assumptions about donor preferences and behavior. Thus, I conduct a laboratory experiment to test whether the success of this mechanism is robust to actual behavior. The results illustrate that the appropriate level of information revelation depends heavily on the 'donor' population. Despite this, the existence of an informer with the ability to filter information -- conditional on the behavior of donors -- is shown to have a significantly positive effect on average.
Author(s): Daniel Jones
Topic: Public goods, charitable giving
Back to session: Charitable Giving 2
ABSTRACT:
Selection by quotas is an important policy measure in the affirmative action tool box. Despite quotas being effective for achieving certain numerical representations, they can come with unwanted side effects, for example in the form of uncooperative behavior in the group created by the quota. In the laboratory, we investigate the extent to which preferential selection in the form of an unrepresentative quota impacts group cooperation and examine the underlying mechanisms. In the experiment we create two groups by randomly assigning subjects colors. In the unrepresentative quota treatment, subjects of one color are selected into a privileged group by performance in an unrelated task whereas subjects of the other color are selected solely based on the quota. We compare contributions in a public good game in this treatment to behavior in a control treatment, where the two colors are treated symmetrically and all members of the privileged group are selected by performance. Our results show significantly less cooperation in the former treatment and we furthermore find that all players in the group are similarly affected. This effect is persistent even when subjects are given a rationale for the unrepresentative performance quota, e.g., by appealing to efficiency or fairness arguments. The negative effect on cooperation from the unrepresentative quota disappears when selection is made completely randomly instead of by performance.
Author(s): Johanna Mollerstrom
Topic: Cooperation, norms and fairness
Back to session: Cooperation 2
ABSTRACT:
Recent theoretical advances in the study of human cooperation and quasi-experimental findings from modern conflict areas suggest that intergroup conflict and ostracism may contribute to the evolution of cooperative behaviors. However, relatively little is known about the forces that shape group formation and migration between human groups. Here, we show that intergroup competition under endogenous group formation may restrict migration to larger and more efficient groups, decreasing the individuals' net returns from cooperation. At the same time, we find that ostracism by majority vote leads to significantly larger groups and increased cooperation. In addition, we observe that the competitive pressure between groups decreases the likelihood of being ostracized from a group due to free-riding. Our findings are in stark contrast to the prevailing empirical evidence suggesting increased levels of cooperation and net pay-offs in public goods games with group competition. Our results suggest that the practices of group formation have a significant effect on the evolution of human cooperation and outcomes of collective decision-making processes.
Author(s): Lauri Sääksvuori
Topic: Cooperation, norms and fairness
Back to session: Cooperation 2
ABSTRACT:
The recent literature on in-group/out-group bias has demonstrated that subjects who participate in group-identity building games are more likely to cooperate in subsequent experiments (for example, Chen and Chen (2011), Chen and Li (2009), Eckel Meier (2006), Goette, Huffman and Meier (2006), Ruffle and Sosis (2009)). We show that in-group/out-group bias can be mitigated by observing in-group members cooperate with out-group members.
Author(s): Qiqi Wang
Topic: Cooperation, norms, reciprocity
Back to session: Cooperation 2
ABSTRACT:
Anger towards an out-group is believed to increase an individual's willingness to engage in in-group cooperation. Social-psychological theories about anger indicate that individuals become angry when they are blocked from a goal. In response, individuals are likely to take actions to remove the barrier. In a series experiments, I induce anger towards an out-group and examine the levels of within group cooperation. Subjects are made angry towards an out-group and then asked to participate in a series of repeated public good games that examine two conditions (1) group competition or (2) group threat. I find that subjects angry towards the out-group contribute more when they are facing group competition than threat.
Author(s): Ngoc T. Phan
Topic: Social preferences, cooperation
Back to session: Group Norms
ABSTRACT:
People often cooperate with members of their own group, and discriminate against members of other groups. Research demonstrates that cultural groups can form endogenously, and that these groups demonstrate in-group favoritism. Given the presence of cultural groups, evolutionary theory then argues that cultural evolution selects for groups that demonstrate parochial altruism. Nonetheless, the source of initial variation in these traits remains un-illuminated. We show here that the group production environment may substantially influence parochial tendencies, with groups formed around more cooperative production (CP) displaying less parochialism than groups formed around more independent production (IP) processes. Participants randomized into CP and IP production tasks formed cultural groups, and subsequently played a modified trust game with in-group and out-group trustees. We found CP to be associated with significantly greater sharing and exchanging behaviors than IP. In the trust game, significant parochial altruism (in-group favoritism combined with out-group discrimination) was displayed by members of IP groups. In contrast, members of CP groups did not engage in either in-group favoritism or out-group discrimination. Further, we found the absence of out-group discrimination in CP to persist even following betrayal. Finally, data collected on beliefs suggest that members of CP are not more generous than members of IP, but rather more likely to believe that out-group trustees will positively reciprocate. Our results have important implications for anyone interested in building cooperative teams, and shed new light on the connection between culture and cooperation.
Author(s): Xiaofei (Sophia) Pan and Daniel Houser
Topic: Social preferences, cooperation
Back to session: Group Norms
ABSTRACT:
Laboratory and field experiments show that conflict with another group leads people to contribute in in-group social dilemmas. In particular, higher contributions in an intragroup Prisoner's Dilemma game were observed when the game was embedded in intergroup conflict. This phenomenon is typically attributed to enhanced group identity due to the common fate of the in-group. We study different social motivations that come into play with intergroup conflict by (a) manipulating the framing of the game as either a competition between groups or a game of externalities and (b) studying an asymmetric game, in which contributions made in one group harm the other group, but not vice versa.
Author(s): Ori Weisel and Ro'i Zultan
Topic: Social preferences, cooperation
Back to session: Group Norms
ABSTRACT:
Based on the serial supply chain model of Anand, Anupindi, and Bassok (2008) with 2-periods and price-sensitive demand, we present the first experimental test of the effect of strategic inventories on supply chain performance. In theory, if holding costs are low enough, the buyer builds up a strategic inventory (even if no operational reasons for stock-holding exist) to limit the supplier's market power, to increase the own profit share, and to enhance the overall supply chain performance. The supplier anticipates the effect of the strategic inventory and differentiates prices to capture a part of the increased supply chain profits. Our results show that the positive effects of strategic inventories are even more pronounced than theoretically predicted, because strategic inventories empower buyers to reduce payoff inequalities and suppliers exhibit a willingness to reduce inequalities as long as their payoff remains above a certain threshold. Overall, strategic inventories have a double positive effect, a strategic and a behavioral, both reducing the average wholesale prices and damping the double marginalization effect and the latter leading to more equitable payoffs.
Author(s): Robin Hartwig, Karl Inderfurth, Abdolkarim Sadrieh, Guido Voigt
Topic: IO, games
Back to session: IO
ABSTRACT:
An important discussion concerns the question whether Western performance appraisal practices can be transferred to and applied in countries quite different in culture like for instance China or India. This issue is particularly important for multi-national joint ventures and multi-national companies tending to transfer their Western human resource management systems to their foreign sub-companies without adequate modification and adaptation. This policy frequently results in large acceptance problems by the local workforce (e.g. Schulz, 2008). A central question, therefore, is how performance appraisal systems have to be designed to meet both the requirements of the organization as well as the local environment.
Author(s): Christine Harbring, Heike Hennig-Schmidt, Gari Walkowitz
Topic: Labor. IO
Back to session: IO
ABSTRACT:
We challenge the recent claim that mispricing in the experimental asset markets introduced by Smith, Suchanek, and Williams (1988) is merely an artefact of confusion over declining fundamental value, and can be eliminated through appropriate training. We instead propose that when training is public knowledge, it reduces uncertainty over the behavior of others and facilitates the formation of common expectations. We disentangle the effect of training from the effect of its public knowledge, and find that when all subjects are trained to understand fundamental value, but this is not public knowledge, mispricing is as great as when training is absent.
Author(s): Stephen L. Cheung, Morten Hedegaard, and Stefan Palan
Topic: Markets, finance
Back to session: IO
ABSTRACT:
This paper shows that the conventional econometric estimation methods such as nonparametric rank sum test, censored Tobit regression and dynamic panel regression with random effects fail to measure the overall treatment effects. After providing the econometric explanations of such failures, this paper proposes a simple but efficient method to test treatment effects, and shows that the suggested methods work well in practice.
Author(s): Donggyu Sul
Topic: Methodology
Link: http://www.utdallas.edu/~dxs093000/papers/exp24.pdf
Back to session: Methods 1
ABSTRACT:
This paper demonstrates that laboratory experiments often produce biased results because of several recurrent problems with the way in which behavioral games are designed and analyzed statistically. The evidence comes in the form of sample schedules of potential outcomes under the Rubin causal model, Monte Carlo simulations, and examples of recently published and widely cited studies in the field of experimental economics. First, we point to the practices of laboratory research that can introduce bias because researches can no longer asses the crucial assumption that potential outcomes are independent of treatment assignment. Second, we show that the first mover's action almost inevitably biases the results for the second mover in two-stage games, such as the ultimatum game and the trust game. Despite their use in the literature, the usual statistical techniques to correct this bias are inadequate. Third, we discuss the limits of causal inference in repeated games, such as the linear public goods game. Fourth, we describe the danger of uncontrolled stimuli including face-to-face interaction during the experiment, photographs or video recordings of another player, and the use of real first names. Finally, we outline several concerns about the clustered random assignment, mediation, endogenous stopping rules, and unwarranted causal interpretation of covariates. The paper suggests that by adjusting laboratory procedures and estimation methods, researchers may lessen their reliance on substantive assumptions not grounded in experimental design.
Author(s): Donald P. Green, Andrej Tusicisny
Topic: Methodology
Back to session: Methods 1
ABSTRACT:
When there exist multiple equilibriums in a repeated experimental game, this paper shows that the conventional measures for treatment effects -- including standard z-score, tobit, nonparametric rank-sum, and Sul's trend regressions -- fail. We develop a new pre-test for the existence of single equilibrium, and suggest the estimation method to measure consistent treatment effects under multiple equilibriums. The finite sample properties of the suggested methods are performed reasonably well. We apply the new test and find that as the return to the public account (MPCR) increases for given group size, the contribution to the public account increases and more importantly subjects become more altruists and less free riders.
Author(s): Jianning Kong, Donggyu Sul
Topic: Public goods, methodology
Link: http://www.utdallas.edu/~dxs093000/papers/conv19.pdf
Back to session: Methods 1
ABSTRACT:
Jurors face highly complex decision making problems with large consequences to third parties. Deciding whether to go to trial, or engage in settlement bargaining is largely based upon how jurors behave. However, there has been little economic literature on this topic. I design a laboratory experiment examining jury decision making in a legal context to investigate how decisions of guilt are made. Subjects who are not jurors play a dictator game with asymmetric information followed by a trial mechanism. I design an incentive compatible belief elicitation mechanism to measure guilt beliefs of jurors prior to the trail. I also investigate subjects' behavior with regards to both the legal institution and willingness to lie when punishment may be possible. The dynamics between those subjects who play the dictator game and subjects who are jurors are also of note. Treatments include changing the legal institution and the cost of a trial.
Author(s): Brent Davis
Topic: Law, morals, voting
Back to session: Mix and Match 1
ABSTRACT:
Bonuses are commonly used as a method of employee compensation. This introduces significant ambiguity into employees’ year-end pay when there is not a set formula used to determine the amount of the bonus. Using the online labor market, Amazon Mechanical Turk, we examine ambiguity in contracts by manipulating the amount of information workers are given as to how the bonus is assigned and their ability at the task. Preliminary results suggest contracts with comparatively high levels of ambiguity attract significantly more subjects than contracts with lower levels of ambiguity. We also find performance feedback is negatively correlated with a subject’s willingness to take a contract.
Author(s): David Cooper, David Johnson
Topic: Labor, ambiguity
Back to session: Mix and Match 1
ABSTRACT:
In public settings, the impact of monetary incentives on prosocial behavior is empirically mixed. Existing theory substantiates this by showing how incentives can introduce a 'greedy' signal that may or may not crowd out image motivation to appear 'prosocial.' However, it remains unclear as to when significant crowding out is likely to occur and render incentives ineffective. To overcome this limitation, I propose a potentially unifying explanation: monetary incentives are unlikely to crowd out the image motivation of individuals with public reputations relative to those with private reputations since the former have little to gain or loose in terms of image motivations. In a laboratory experiment on volunteer behavior, I confirm this interactions effect as well as other image-related effects among female participants.
Author(s): Christine Exley
Topic: Social preferences, cooperation
Back to session: Mix and Match 1
ABSTRACT:
We examine the role of higher order beliefs in markets where coordination can lead to gains to trade. The scenarios are modeled such that trader's strategies do not only depend upon their beliefs of underlying economic phenomena, but also upon the others' beliefs regarding the beliefs of themselves. Under certain parameters the breakdown of coordination is predicted--even when both traders are certain trade is advantageous. We demonstrate the equilibrium predictions can be constructed via a small number of iterated thought exercises. The experimental design allows us to control for various behavioral phenomena and examine subjects' decisions across different accounting regimes as to tease out strategic uncertainty due solely to information asymmetry. In this setting we find evidence supporting higher order beliefs. An implication is that the lack of uniformity leads to lack of common knowledge of the beliefs of others, which in turn leads to the spreading of inefficient outcomes.
Author(s): Timothy Shields & Baohua Xin
Topic: Beliefs, markets
Back to session: Mix and Match 1
ABSTRACT:
We explore the impact of reference dependence on trade in dynamic economies, both theoretically and experimentally. We first show that under plausible axioms, dynamic economies generate a momentum effect capable of reducing or eliminating the initial trade-dampening impact of the endowment effect over time. We then introduce a novel experiment designed to test this prediction. Our results confirm that subjects' willingness to trade dramatically increases between the first and second period of our experiment as predicted by the theory. The data also suggest that subjects become increasingly sophisticated over the course of the experiment leading to stronger erosion of the endowment effect than our theory predicts.
Author(s): Sean Crockett and Ryan Oprea
Topic: Anomalies, markets, homegrown preferences
Link: http://faculty.arts.ubc.ca/roprea/momentumEX.pdf
Back to session: Financial Markets
ABSTRACT:
This is the first paper to test the asset pricing implication of leverage in a laboratory. We show that as theory predicts, leverage increases asset prices: When an asset can be used as collateral (that is, when the asset can be bought on margin), its price goes up. This increase is significant, and quantitatively close to what theory predicts. However, important deviations from the theory arise in the laboratory. First, the demand for the asset shifts when it can be used as a collateral, even though agents do not exhaust their purchasing power when collateralized borrowing is not allowed. Second, the spread between collateralizable and noncollateralizable assets does not increase during crises, in contrast to what theory predicts.
Author(s): Marco Cipriani, Ana Fostel, Daniel Houser
Topic: finance, markets
Link: http://www.newyorkfed.org/research/staff_reports/sr548.pdf
Back to session: Financial Markets
ABSTRACT:
A revealed preference test of disappointment aversion and cumulative prospect theory is developed in the context of asset markets. The test is non parametric and can be applied to experimental as well as consumer data. The test is an extension of VarianÆs (1988) test of expected utility theory for assets markets. Using data from Choi et al. (2007), behavioral theories are found to be no better than expected utility theory at rationalizing the data.
Author(s): Marco Castillo
Topic: finance
Back to session: Financial Markets
ABSTRACT:
This paper explores the effect of restricted pre-play communication on offer and acceptance behavior in Ultimatum bargaining with asymmetric information. In particular, we study an Ultimatum Game in which the total amount to be bargained over (the pie) is either $10 or $20. We systematically vary whether both proposer and responder are informed about the pie size (baseline), or only the proposer, or only the responder. In addition, we vary whether there is no pre-play communication, or whether the informed party can send message about the pie size communication before decisions are made. In one communication condition, the message sender is free to choose the correct or the wrong message (cheap talk), while in the second communication condition the sender can only choose whether to reveal or not to reveal the true pie size ('true talk'). We find that contrary to the theoretical prediction cheap talk has a significant positive effect on efficiency, while true talk is less effective than theoretically predicted.
Author(s): Olga Bystrova and Ben Greiner
Topic: communication, social preferences
Back to session: Beliefs
ABSTRACT:
We study how trust between strangers evolves in a setting with noisy feedback on actions of others. We first examine how players choose to optimally manage their trust in others in a theoretical model where trustors process noisy feedback about the “trustworthiness” type of trustees. We discuss whether trustors use Bayesian updating to calculate their degree of trust or whether they process information in a biased manner. To study this empirically, we design a laboratory experiment using a two player sequential trust game with several rounds of subsequent noisy feedback on the action of the trustee. We track the evolution of trustors’ individual beliefs about trustworthiness types of trustees to document that subjects process information in an asymmetric way compared to a perfect Bayesian: they react more to negative feedback rather than positive. Trust is therefore much easier to destroy than to build. We then use this setting to study how social identity information affects information processing. We match participants from two different universities (in Hong Kong and Beijing, respectively) and prime them on the social identity of their counterpart. We first document that players exhibit in-group favoritism by having higher degree of trust in the in-group members. We then show that trustors are more asymmetric compared to anonymous matching baseline in reacting to feedback about in-group trustees: they overreact to negative feedback and underreact to positive feedback of in-group members.
Author(s): Fanzheng Yang
Topic: trust, learning
Back to session: Beliefs
ABSTRACT:
This paper investigates attitude towards partial ambiguity in a laboratory setting using different decks of 100 cards. We assess certainty equivalents on three symmetric variants of the fully ambiguous urn in the classical Ellsberg 2-urn paradox by limiting the range of the possible number of red cards with the rest black. In the interval variants, the possible number of red cards ranges from n to 100-n. In the disjoint variants, the possible number of red cards can range from 0 to n and from 100-n to 100. In the third symmetric variant called two-point ambiguity, the number of red cards is limited to either n or 100-n. For both interval and disjoint ambiguity, subjects tend to value betting on a deck with a smaller set of ambiguous states more. For two- point ambiguity, subjects exhibit greater aversion as n goes from 50 (no ambiguity) to 10 in four steps. Paradoxically, there is a reversal from n = 10 to n = 0 which turns out to be valued similarly as the no-ambiguity lottery with n = 50. We further examine attitude towards skewed partial ambiguity by eliciting subjects' preference between betting on a deck with a known number of n red cards versus betting on a deck whose possible number of red cards can range from 0 to 2n. Here, subjects tend to go from being averse to ambiguity at n = 50 and n = 40 to being ambiguity seeking for n less than 30. We also discuss the implications of our findings for existing models of decision making under uncertainty in the literature.
Author(s): Soo Hong Chew, Bin Miao, and Songfa Zhong
Topic: ambiguity, risky choice
Back to session: Beliefs
ABSTRACT:
We propose that individuals use emotions strategically in interactions. In our experiments, we first show that in some environments angering people makes them more effective in competitions, and, in others, anger makes them less effective. We then show that participants in our experiments strategically used the option to anger their opponents. In particular, they were more likely to anger their opponents when anger would have a negative effect on the opponents’ performances. This finding suggests people understand the effects of emotions on behavior and exploit them to their advantage, providing insight into strategic interactions between emotional agents.
Author(s): Uri Gneezy and Alex Imas
Topic: emotions and games
Back to session: Emotion and Choice
ABSTRACT:
We examine in the laboratory how having the opportunity to behave prosocially in the future affects the likelihood of currently engaging in antisocial behavior. Subjects first complete a task which determines their payoff. They then self-report their performance, which provides them with an opportunity for undetected cheating. In the second stage they are given the opportunity to donate some of the money earned in the first stage to a charity. Only subjects in the treatment group know about the opportunity to donate in the second stage. We find that subjects cheat more if they know they can donate some of the money to charity. Surprisingly, they also end up donating less to charity. The first result is consistent with charitable giving alleviating guilt. The second result can be explained by models of identity, time-inconsistent preferences, or self-signaling.
Author(s): Doru Cojoc and Adrian Stoian
Topic: charitable giving, social preferences
Link: https://dl.dropbox.com/u/11212329/Charity%20and%20Dishonesty%20-%20April%2016.pdf
Back to session: Emotion and Choice
ABSTRACT:
I explore how optimism bias affects the behavior and trading decisions of the human subjects in experimental entrepreneurial prediction markets. I designed two prediction markets (original and control) and a measure of optimism bias which I call optimism coefficient. In the experiment, I first established the presence of the bias, then measured its extent and finally corrected the predictions for it. Results show that control market predictions were more accurate than original peer market predictions across the board, attesting to the presence of the bias. When peer predictions were adjusted for the bias, they became more accurate than the original unadjusted ones, but were still less accurate than control ones. These results imply that efficiency of entrepreneurial peer markets depends on the interplay of effects of the psychological biases and information asymmetry, which in turn depend on their magnitude and direction. Entrepreneurs know more about their own projects than public at large, but at the same time are more biased. Public traders do not have biases, but are disadvantaged in terms of information.
Author(s): Momo Deretic
Topic: markets, beliefs, prediction mkts
Back to session: Emotion and Choice
ABSTRACT:
This paper investigates the role of stress for competition attitudes confirming the predictive role of self-reported stress levels for competition choices. Cortisol samples are collected during a Niederle/Vesterlund-style experiment to assess physical stress levels in addition to self-reported stress levels. Although, physical and self-reported stress levels are somewhat correlated, perceived stress levels play a larger role than physical stress levels for individual choices.
Author(s): Dorothea Herreiner, Heather Watts
Topic: physio, elicitation, homegrown
Back to session: Emotion and Choice
ABSTRACT:
Public goods such as climate change mitigation affect people in multiple locations. This paper studies the allocation behavior using a voluntary contribution experiment, where subjects can invest not only into a local public good provision but also into a global public good provision. Moreover, localities may differ in their wealth. In this experiment, we compare a setting where two localities have the same wealth (homogenous treatment) with a setting where the localities differ in wealth (heterogeneous treatment). This paper finds that: (1) The social efficiency is higher in the treatment with homogeneous localities than when the localities are heterogeneous; and 2) This efficiency difference comes from differences in contribution choices: Participants in the homogeneous treatment contribute more to the global public good, which is the most efficient, whereas participants in the heterogeneous treatment are more likely to contribute to the local public good.
Author(s): Kohei Nitta
Topic: public goods, games
Back to session: Mix and Match 2
ABSTRACT:
We present the results of two controlled behavioral studies on the effects of online recommendations on consumers’ economic behavior. In the first study, we found strong evidence that participants’ willingness to pay was significantly affected by randomly assigned song recommendations, even when controlling for participants’ preferences and demographics. In the second study, we presented participants with actual system-generated recommendations that were intentionally perturbed (i.e., significant error was introduced) and observed similar effects on willingness to pay. The results have significant implications for the design and application of recommender systems as well as for e-commerce practice.
Author(s): Gedas Adomavicius, Jesse Bockstedt, Shawn Curley, Jingjing Zhang
Topic: internet, recommendations, wtp
Back to session: Mix and Match 2