Download this chapter as pdf-file


C.II The Behaviorally Informed Design of Institutions for the Provision of Collective Goods

C.II.1 General Outline

C.II.1.1 Motivation

The bulk of our work is experimental. Yet, we deliberately keep the original mission statement: "The Behaviorally Informed Design of Institutions for the Provision of Collective Goods", since our area of interest has not changed. We also see no reason categorically to exclude other empirical methods, or theory, or doctrine, by the very definition of our task.
All research on collective goods asks one of the following three questions: is there a collective-goods problem in the first place? If so, is an existing or a proposed institution able to solve the problem, or at least to improve the situation? Finally, do the normatively appropriate problem definition and the norma-tively preferable institutional response stand a chance of being implemented?
It is natural to address all these three questions by way of rational-choice analysis. Collective-goods problems are then defined as pure public goods, club goods, or common pool resources. In each case, the analysis focuses on incentives and information, and on the way in which institutions shape incentives and channel the information which is required to address the collective-goods problem. Normative analysis deals with the optimal design of incentives, positive analysis with the actual incentives that are generated in a given institutional context. The mechanism design approach summarized above does the former kind of analysis, public choice theory the latter. Here the rational-choice paradigm helps us understand why the political process often fails to harness sovereign powers in the interest of changing incentives such that collective-goods problems disappear.
While evidently fruitful, the rational-choice perspective is also limited. This is due to the very same factor that has made the rational-choice model so visibly successful. The model rests on the strict distinction between objectives and constraints. The objects of study are utility-maximizing individuals reacting to changes in opportunity structures. For methodological reasons, the individual is modeled as Homo Oeconomicus. For sure, these are only assumptions, not claims about reality. They are imposed in order to capture the essence of social phenomena and institutions, and to make predictions for the effect of changing circumstances. However, the scope of this analysis is inherently limited.
An alternative research strategy, which starts with what is known about human behavior, is likely to develop a fairly different depiction of collective goods. Some phenomena that are made visible by behavioral analysis can hardly even be translated back into the world of rational choice. Our work focuses on this alternative approach.
The behavioral analysis of collective goods is not virgin territory. Suffice it to recall a few of the well-known findings: where (simple) rational-choice models would predict the "tragedy of the commons", in practice it is often conspicuously absent. There are various reasons for this, but the fact that they have a more realistic picture of human motivation is part of the explanation. "Public-goods games" are one of the workhorses of experimental economics. Again, contribution rates found in the laboratory by far exceed the prediction of zero contributions made by rational-choice models. If all beneficiaries of a public good agree on a contribution level, in rational-choice terms this is just "cheap talk". At the level of implementing the agreement, the original social dilemma is repeated. However, psychologists have traced a powerful cheater-detection mechanism, effectively exploiting subtle signals. It has bite, since punishing sentiments kick in when cheating seems patent. Emotions thus trump rationality and help solve the social dilemma. It is in this context that our work on the behavioral analysis of collective-goods problems is situated. We are adding new dimensions, exploring new fields of application, and translating the findings into institutional analysis and design.
Likewise, we are of course not the first to be interested in the behavioral analysis of institutions. Behavioral effects have never been fully absent from institutional analysis. An obvious illustration is "moral suasion". But the most prominent force in the area is the growing behavioral law and economics movement. It mainly piggybacks on the Kahneman/Tversky critique of the rational-choice approach. It either interprets legal institutions as remedies to individually or socially detrimental "biases". Or it criticizes the legal community for overlooking that biases prevent the law from being effective. Both have obvious value. Suffice it again to recall two well-known findings. It is much easier to get an appropriate understanding of consumer-protection legislation if one understands the psychological underpinnings of strategies like the "foot-in-the-door technique of salesmen". Environmental policy has long been tempted by torts as a tool for "ex-post regulation", in light of the experiences from concrete cases. This is, however, dubious advice, given the strong "hindsight bias". Once one has seen the evidence of a risk materializing, it is next to impossible to form a proper assessment of its ex-ante likelihood. Consequently, regulation by torts finds itself on a slippery slope towards ever stricter rules.
Some of our work is exactly in this tradition, where it seems helpful to assess the potential of institutions, and of the law in particular, in order to solve collective-goods problems. But in two ways we are going beyond this earlier work. We make a point of not exclusively looking at biases. Related to this, the Kahneman/Tversky literature and the experimental economics literature are not the only sources we are tapping. Rather, we try to purchase directly from psychology. And we are particularly interested in the law as a governance tool. We are convinced that, in a behavioral perspective, one is able to gain a much richer understanding of the law's potential. In these ways, we also hope to bridge the gap between (new) behavioral law and economics and (old) law and psychology. While there has for decades been direct interaction between lawyers and psychologists on issues like lie detection or eyewitness testimony, this strand of research has not thus far been very interested in the law as a governance tool.
Interdisciplinarity is never easy. However, in major US law schools, law and economics has almost become a standard approach. Behavioral law and economics is seen as one of the major strands of this approach, and is itself making headway. The situation in Germany is significantly different. Here, antitrust law notwithstanding, economic analysis is still rare, if not actively combated. The behavioral analysis of law is only just tentatively starting. Against this backdrop, it is inevitable that the widespread skepticism about a closer interaction between law and the social sciences be taken seriously. We are trying to respond at two levels. At one level, we are attempting to determine the proper role of input from the social sciences in both legal doctrine and legal science. At the other level, we are comparing alternative paradigms, starting with rational-choice and behavioral analysis, but not confining ourselves to these.
In earlier reports, we had to admit that the third fundamental question regarding collective goods would lend itself to behavioral analysis no less than the first two, but was largely beyond the scope of our attention. This has changed. A considerable strand of our work now addresses rule generation, from a behavioral, usually experimental angle.

C.II.1.2 Summary Report

Over the last three years, on all three overarching questions, the group has been very active. We have also continued to contribute to three obvious meta-questions: the relationship between law and (behavioral) social science, experimental methodology and data analysis. Given the findings are fairly rich and faceted, this summary report is confined to finished work, be it published or available as a working paper. For future directions, please see the portraits of individual researchers.

1. Problem Definition

Most of our work on defining social problems the law actually does or might want to address starts from game theory. We thus predominantly work in the paradigm of behavioral economics. We are of course aware of the inevitable limitations inherent in any paradigm, and inherent in this specific paradigm. Some of our contributions therefore directly purchase from psychology. We always carefully discuss external validity. Since the legal community has had less exposure to experiments in the tradition of experimental economics, in our papers – the ones that target legal outlets – a discussion of the status of the argument always features prominently. We make a point of the many facets of the real-world problem that triggered our research, facets that are not captured by the design of the experiment; yet, at the same time, we stress why causal inference should be critical for the proper interpretation of the law in force, and for legal policy.
That said, game theory nonetheless provides the most natural and the most appropriate way of organizing our findings. Given that the mission of the institute is deepening our knowledge of collective goods, this report starts with cooperation problems (a), including markets (where non-cooperation of suppliers would be socially desirable) (b). We then turn to bargaining (c) and (intellectual) property (d) as two major institutions for enabling decentral coordination in markets. While principal-agent relationships are not confined to the interior of the firm, this is the most prominent and the best study application (e). We have begun to work on a different class of games where the problem is coordination, not cooperation (f). Further contributions are in the tradition of the heuristics and biases paradigm, and address non-interactive choice (g). There are also a number of contributions to general behavioral theory (h), and to the translation of these findings into legal doctrine (i).

a) Cooperation
In its most basic form, a cooperation problem is characterized by a difference between individual and social rationality. The individual is best off if she pursues her own goals, knowing or at least taking into account that others suffer, or do not benefit to the ideally possible extent. If one assumes "standard preferences", game theory predicts that the social optimum is not reached. This analysis provides a natural justification for legal intervention. The sovereign state has the power to overrule individual selfishness. Sovereignty is the technology to avoid the trap of the social dilemma.
Many of our contributions have revolved around testing and refining the behavioral assumptions underly-ing this prediction. Are individuals indeed as selfish as predicted by standard theory? The natural tool for exploring this question is the dictator game (aa). Are individuals even willing to forego a higher income for themselves if this means tolerating exploitation? The natural tool for exploring this question is the ultimatum game (bb). Do groups find a way to coordinate on an outcome closer to the social optimum, even if each member could exploit the willingness of others to cooperate? The natural tool for exploring this question is the prisoner's dilemma, possibly specified as a linear public good (cc).

aa) Dictator Games
As one of the founding fathers of the empirical legal movement once put it, the legislator should not believe individual empirical papers, but it should believe entire empirical literatures. Therefore summariz-ing an experimental literature is not only interesting in its own right. It has also value for the legal con-sumers of that evidence. In that spirit, the meta-study by Engel (2011b) demonstrates how rare it is for entire experimental populations to disregard completely the distributional imbalance characteristic for the standard dictator game. While this is a very robust finding, the amount given is also highly sensitive to subtle manipulations of the design and, even more importantly for policymakers, there is pronounced heterogeneity. If the legal order does not intervene, some help for the needy is to be expected, but there will be a pronounced amount of free-riding.
Chmura, Engel et al. (2013) use this meta-study to provide them with a benchmark against which they cast new light on one of the definitional questions of criminology. While much of that discipline is interested in explaining crime, from a rational choice perspective the opposite question is more challenging. Why is there not more crime, given that the expected value of the sanction is very often below the expected benefit from crime? One reason might be a prominent explanation for giving in dictator games: aversion against advantageous inequity. If that was true, prison inmates should give significantly less in the dictator game. For the theory treats inequity aversion as a personality trait that is stable across situations. Yet, the prediction is clearly refuted. Prison inmates give almost exactly as much as the general population. A further meta-study organizes the evidence on a competing explanation of crime, a lack of self-control (Engel 2012a).
In the typical design of the dictator game, the dictator's endowment is manna from heaven. If dictators have to earn their endowment in a real-effort task, they give considerably less. Nisvan, Gangadharan et al. (2011) refine this finding. In the first stage of their experiment, participants compete in a tournament that defines their earnings. If the winner is chosen to be the dictator, she gives considerably less than if this role is assigned to the second in line.
In a natural field experiment, Aretz and Kube (2013) show that a plausible other determinant does not influence how much individuals give: if they are given a choice between multiple recipients, they do not give more.
Two papers by psychologists unpack the underlying mental process. Dickert, Sagara et al. (2011) show that individuals actually take two decisions: whether to give anything and, if so, how much. While the former decision is influenced by mood management, for the latter decision empathy is critical. Dickert, Kleber et al. (2011) investigate in which ways the presentation of numerical information about the needs of potential recipients influences the amount given.

bb) Ultimatum Game
The standard ultimatum game pairs one proposer with one responder. Two papers investigate in which ways standard results change if this two-person interaction is put into a social context. Fischer and Güth (2012) pair one proposer with two responders. The proposer is free to exclude one responder from the game, which leaves this player with a minimum income. The pie to be distributed among the remaining two players increases. They find no signs of indirect reciprocity. The remaining responder does not increase her acceptance threshold if the other responder has been excluded intentionally, compared to a baseline where exclusion was at random. McDonald, Nikiforakis et al. (2013) have participants compete in a real-effort task for the role of proposer. The remaining participants are randomly assigned the role of responder, or they receive a fixed payment and remain passive. Responders ignore the outsider if she is poor, but raise their acceptance threshold if the fixed payment to the outsider is high (also see Nicklisch and Wolff 2012).

cc) Prisoner's Dilemma
In line with the mission of the institute, a considerable amount of our work directly tests social dilemmas. Glöckner, Irlenbusch et al. (2011) introduce heterogeneity into a linear public good. One group member is privileged by the fact that her contributions yield a higher return for themselves and all other group members. They compare a setting where contributing the entire endowment is a dominant strategy for the privileged player with another setting where the dominant strategy remains to contribute nothing. In the former setting, privileged players contribute significantly more, but non-privileged players contribute significantly less. The authors interpret their result as saying that leadership requires sacrifice.
Engel and Zhurakhovska (2012a) add a passive third player to a symmetric two-person prisoner's dilemma with discrete action space. Using the strategy method, they manipulate the degree by which making a cooperative move inflicts harm on the passive outsider. However severe the harm, active players do not become less willing to cooperate. They do, however, become more pessimistic about the willingness of their anonymous counterpart to cooperate. Conditional on their more pessimistic beliefs, they are more willing to cooperate if this is to the detriment of the outsider. Harm on third parties serves as a lubricant for cooperation (see also Ding (2012) on another context effect).
One of the deepest puzzles of the social sciences is the emergence of cooperation, in particular in large-scale societies. How can it be that there is so much non-selfishness, even in the absence of powerful institutions who discipline free-riders? Fischer, Frid et al. (2013) propose a novel explanation. When deciding how to behave, agents engage in "mimicry". The critical component of the strategy is a similarity index. If the other player in question has been observed or is expected to act such that my cooperation promises to be more profitable than my defection, I cooperate myself; I defect otherwise. The authors use simulation to demonstrate how well this strategy performs, and that it outperforms all alternative strate-gies over time, in particular in environments where initially hostile and unpredictable agents are frequent (on a related topic, see also Traxler and Spichtig 2011).
Several papers cast light on decision-making in dilemma situations. Previous attempts in the literature at explaining choices in dilemma games with personality measures have been rather frustrating. It has even been claimed that, given that there seems to be so little consistency of behavior across games, that prominent motivational models are not helpful. Engel and Zhurakhovska (2012b) show this skeptical stance to be premature. Provided one controls for beliefs, it can be shown that a prisoner's dilemma is a game of multiple motives. Possible gains from co-operation matter, as do risk and loss aversion, and the degree of greed.
Glöckner and Hilbig (2012) demonstrate another moderating variable. While risk aversion predicts defec-tion in a hostile environment, it predicts cooperation in a cooperation-friendly environment. Now, the risk participants dread is missing gains from cooperation.
Using eye-tracking, Fiedler, Glöckner et al. (2013) show that differences in social value orientation are associated with characteristic differences in information search. The more participants are concerned with the well-being of others, the more intensely they search for information (see also Fiedler, Glöckner et al. 2012).
Deck and Nikiforakis (2012) modify the standard design of a public good and implement a minimum-effort game. While effort is costly, payoff only depends on the lowest effort in the group of six. This, of course, transforms the public good into a coordination game with multiple equilibria. They manipulate the information set. While in the baseline participants have no information about provisional choices of other group members, in the first treatment they see a sample, and in the second treatment they see all provi-sional choices. In the second design, participants almost perfectly coordinate on the payoff-dominant equilibrium, while seeing a sample does not make coordination on this equilibrium more likely than in the baseline.

b) Markets
It is straightforward to model the relationship between competitors as a prisoner's dilemma. Collectively, they are best off if they collude. This gives them a share of the monopoly profit. Yet, individually, all are best off if the remaining competitors are faithful cartelists, while the firm in question undercuts price or, for that matter, exceeds its quota. Now there is a rich experimental literature on prisoner's dilemma games, and an equally rich experimental literature on collusion. Engel and Zhurakhovska (2011) run an experiment to bridge the gap between both literatures. Arguably, the difference is social embeddedness. Collusion is socially undesirable and exposes firms to a risk of legal sanctions. Moreover, collusion inflicts harm on the opposite market site. The experiment crosses both qualifications. The risk of sanctions has the expected dampening effect, even if the sanction is not a deterrent. By contrast, the degree of harm is immaterial and, conditional on beliefs, even fosters collusion. Since the latter finding seems most important, and has ramifications unrelated to collusion, the current version of the paper focuses on this aspect. A small theoretical contribution shows that mapping collusion to a linear public good is also technically feasible, but requires more challenging definitions of the payoff function (Engel 2011a).
While the experimental literature on oligopoly is rich, it has almost exclusively been written by economists. Institutional detail is not routinely taken into account by this literature. This gives room for fertile cooperation between antitrust lawyers and economists. One application is a graphic term from merger control. On both sides of the Atlantic, antitrust authorities are concerned if a merger has the effect that a particularly aggressive firm leaves the market. Aggressiveness may, of course, have many sources. Yet, one plausible source is behavioral. Some individuals are just particularly competitive. Such individuals are not unlikely to select themselves into sales or higher management. They may help create a culture of com-petitiveness in their firms. Engel and Ockenfels (2013) exploit a well-worn tool from social psychology, the ring measure of social value orientation, to pre-classify participants by their individual degree of competi-tiveness. They have rivalistic participants enter a symmetric Cournot duopoly. After 10 periods, either duopolist may negotiate with the entrant over market exit. Rivalistic entrants do make markets more competitive, compared with benchmark markets where entrants just maximize profit or are socially minded. Yet, the effect is only present when, pre-entry, the market was collusive. This is, of course, also the situation when antitrust policy finds mavericks most desirable. The authors observe no buyouts, though, most likely because the one firm that orchestrates the merger gives her remaining competitor a free lunch.
Several contributions directly speak to the legal audience, based on behavioral insights. In his monograph, Morell (2011) does so for the abuse of a dominant position, with a special eye to the normative assess-ment of rollback rebates. Upon the invitation of GD-COMP, Engel (2011e) summarizes the experimental insights that are relevant for the treatment of research and development agreements. Relying on his meta-study of oligopoly experiments, Engel (2011c) sketches the insights to be gained from experiments for antitrust law more generally.

c) Bargaining
Money is a powerful facilitator. Money makes everything comparable and therefore tradable. Yet, even in markets, this assumption is sometimes too strong. The well-established firm does not want to sell the premises in the city center, since this is where the firm was once founded. But if this same firm was offered to become the leading partner in a joint venture, this firm may want to reconsider letting another partner use the precious piece of land. This is one illustration of the problems Hortala-Vallve, Llorente-Saguer et al. (2013) are interested in. They experimentally compare two negotiation protocols. The first one is free bargaining. In the alternative protocol, each negotiator assigns points that express the comparative valuation of each of the negotiated issues; of course, this introduces the functional equivalent of a currency. They find that unconstrained negotiations do better if preferences are common knowledge, whereas the point allotment mechanism does better if there is preference uncertainty.

d) Intellectual Property
In the debate over intellectual property, the starting point is frequently the claim that knowledge is a pure public good. Once a new technology has been invented, there is neither rivalry in consumption (as long as one brackets the possibility of a monopoly profit from being ahead of others, the invention does not become less valuable if it is used by others), nor is there exclusion (intellectual property legislation notwithstanding, everybody can appropriate the new idea). This starting point has often been criticized as being overly simplistic. Engel and Kleine (2013) test one of the criticisms in the lab. Out there in the field, appropriating a foreign innovation is usually an investment decision on its own. Appropriation is at least costly, and usually also risky. They translate this into a game where two competitors may invest an endowment in either risky innovation or risky appropriation. Unlike the impression frequently conveyed in the policy debate, the prospect of appropriation does not deter innovation. Participants split their en-dowments into either purpose in the qualitative way suggested by standard theory. Actually they invest even considerably more into innovation than predicted. If there is a policy problem, it results from too much investment into either purpose, in particular if appropriation has low risk (see also the theory paper by Engel 2011f).
Privacy is not routinely regarded as a public good. Yet, with the advent of social networks like Facebook, this is an appropriate model. Social networks do not hide their business model. They provide their services for free, but on the condition that they may use the information customers introduce into the network as they, the social networks, deem fit. The most important use is pattern recognition. The more data the network has collected, the better it is able to predict future choices based on less and less information. Actually this information about cues need not even have been given to the network by the person affected. If I know three of your friends, this may be enough to predict how you will react to a stimulus. The resulting public-good nature of social networks has not only largely escaped public attention. It additionally has an important twist. The ability of the network to inflict harm on its official users, and even on non-users, is not developing linearly in time. For quite some time, a lot of information in the hands of the network may still be quite harmless. But at some point, the information becomes good enough to be used. It is this feature of the policy problem that Engel and Fairfield (2013) model and test experimentally. While holding the expected value constant, they manipulate the factor by which contributions to a linear public good made in an earlier round lose their negative impact on today's profit. If this effect is backloaded, participants contribute significantly more to a public bad, and make significantly less profit (for an application to German criminal procedure, see Englerth and Hermstrüwer (forthcoming)).

e) Principal Agent
A considerable strand of our work is concerned with the determinants of behavior in principal-agent relationships. Kube, Puppe et al. (2012) investigate in a field experiment in which ways giving employees a bonus increases performance. The effect turns out substantially stronger if the bonus comes as a nicely wrapped gift, compared with its money equivalent. The effect is no longer distinguishable, and equally strong, if cash is also wrapped as a gift.
Using the same natural setting of a university library, Kube, Puppe et al. (2013) demonstrate that wage cuts have a sizeable negative effect on performance.
For another field experiment, Goerg and Kube (2012) exploit the fact that our library underwent labor-intense reorganization. They invite student helpers and compare a piece rate with several bonus schemes. If the bonus is paid for a goal chosen by the employee herself, this has a strong positive effect. Yet, the effect is equally strong if the goal is chosen by the employer (provided the goal is sensible) and, most interestingly, if workers are just induced to set a goal, without receiving an extra bonus (for a further contribution to this line of research, see Goerg, Abeler et al. (2011)).
Several papers deal with the legal framework of relations within a firm. Bünstorf, Engel et al. (2013) experimentally test the effects of a non-compete clause on workers' effort. If participants hold standard preferences, the fact that the worker cannot leave the firm and competes with her former employer upon making an innovation, or increasing the customer base for that matter, should reduce effort, provided the compensation paid when exercising the clause is not very high. In the lab, this prediction is rejected. Effort is not significantly lower, even if compensation leaves most of gains from innovation with the employer. In such a setting, employers motivate employees by paying substantial wages upfront.
Many legal orders aim at inducing firms not to serve the monetary interests of their shareholders exclu-sively, but to take the effects of their action on multiple stakeholders into account. However, stakeholders other than the workforce do not usually have a formal say on corporate decision-making. Legal orders content themselves with admonishing firms to take the effects on the broader social context into account. In a lab experiment, this manipulation fails to be effective. Actually it even is counter-productive (Fischer, Goerg et al. 2013). See also Hamann (2012), discussing deliberate choice vs. intuitive decision-making at the board; Han and Jekel (2011), investigating the role of job satisfaction in the interaction between leaders and members of teams; Himmler, Cornelissen et al. (2011), using field data to see the repercus-sions of perceived unfairness in CEO compensation on worker effort; and Weinschenk (2011), modeling procrastination in a team.

f) Coordination and Anti-coordination
While many social problems arguably result from a dilemma, coordination problems are not infrequent either. Fehr, Heinemann et al. (2011) study the power of arbitrary, minimal interventions, which in the game-theoretic literature are called sunspots, as coordination devices. They not only show that public signals serve that purpose. Even salient private signals may shift the equilibrium away from risk dominance.
Building on earlier work published in the American Economic Review, Selten, Chmura et al. (2011) deepen the analysis of their data and Chmura, Goerg et al. (2012) study experimentally in which ways choices and thereby equilibria change over time in a set of repeated games where the Nash equilibrium of the stage game requires mixing. It turns out that two competing learning models explain different aspects of the data: impulse-matching learning, and self-tuned, experience-weighted attraction learning.
The game studied by Hortala-Vallve and Llorente-Saguer (2012) may be seen as an anti-coordination game. The game goes by a martial name: Colonel Blotto game. In the cover story, two colonels, fighting for different countries, decide how many troops to deploy in one of several battlefields. A battlefield is won by the colonel who has sent more troops. The war is won by the colonel who has won in more battlefields. The intuitive solution is an equilibrium in mixed strategies. The theory paper characterizes the conditions for equilibria in pure strategies and characterizes the set of games in which such equilibria exist (see also the ambiguous act equilibria developed by Bade 2011a).

g) Biases
While most of our experimental work is in the tradition of experimental economics, there are also a few contributions in the heuristics and biases tradition that originated in social and cognitive psychology. This programme focuses on potential defects of isolated individual choices, not on socially undesirable outcomes of interactive choices.
Towfigh and Glöckner (2011) address a hotly debated issue of German (legal) policy. In Germany, the länder hold a monopoly on betting. Critics assert that this is just a way for prime ministers to have a considerable amount of money at their personal disposal, rather than having to subject it to parliamentary control. Yet, the official justification of the monopoly is the purported danger of gaming for bettors. Critics say that the fact that sports betting is not part of the monopoly is just a historical accident. In official statements, however, the difference is justified with the smaller personal risk inherent in "games of skill", compared to "games of chance". In a field experiment, the authors show that the opposite holds true. In games of skill, over-optimism and the illusion of control make bettors seek even more risk.
Using a huge set of data from a Chinese online lottery, Ding (2011) demonstrates a whole panoply of biases: there is evidence for the gambler's fallacy, for hot-hand fallacy, and an inclination to choose culturally contingent prominent numbers.
Grechenig and Roberto (2011) discuss the importance of hindsight bias for the definition and application of negligent behavior in torts and contract.

h) General Behavioral Theory
We benefit from contributions to general behavior of theory made in other parts of the institute. Psy-chologists investigate in which way the personality trait of victim sensitivity influences the ability to detect cooperativeness, and show that both are negatively correlated (Gollwitzer, Rothmund et al.). Glöckner and Pachur (2012) demonstrate that cumulative prospect theory is able to predict risky choices over time, and that the predictive intertemporal power of the theory is increased if one fits parameters per individual rather than per sample. Fiedler and Glöckner (2012) use a whole set of measures, mostly from eye-tracking, to develop a process model for risky choice. Von Weizsäcker (2011) develops a behavioral model for past choices changing future preferences, and defines the conditions under which welfare statements can nonetheless be made.

i) Translation into Legal Doctrine
Translating the relevance of our findings on problem definition for legal doctrine has not been our focus. Yet, Engel (2013a) uses his dataset with all publicly available decisions of the German Constitutional Court promulgated in 2011 to show empirically that the court is substantially ahead of the academic literature when it comes to deciding which purposes justify intrusions into fundamental freedoms. While the court does not use technical language, concepts developed in the social sciences find their way into law through this channel.

2. Institutional Intervention
In economic parlance, the law shapes the opportunity structure for private action and interaction. Even if this is done with the intention to facilitate the exercise of free will and the decentral coordination among private parties, in a way the law defines the institutional framework for private activity, and it excludes courses of action that do not fit this framework. In that sense, any legal rule intervenes into freedom or property; much law is openly interventionist. This explains why the second strand of our work addresses the behavioral effects of intervention, be that minimal intervention (a), sanctions (b), post-hoc intervention into contractual relations (c), or formal legal intervention (d).

a) Triggering Social Preferences
Many of us have been interested in social preferences making minimal intervention practical. Engel and Kurschilgen (2013) show that it suffices to ask participants in a linear public good in every period whether they believe a social norm exists and, if so, which minimum contribution is required. These two questions are strong enough almost to stabilize cooperation. Framing these questions as statements about the emergence of a rule of customary law does not increase cooperation. Cooperation and normative expec-tations can be shown to co-evolve (see also Engel (2011d) on the relevance for the conceptualisation of customary law; further, see Petersen (2011) on the importance of uncertainty in the formation of customary law).
Engel, Beckenkamp et al. (2014) comparatively assess two recommendations derived from a prominent approach in criminal policy, the broken windows theory. The approach has been explicitly used in cities like New York and Los Angeles to curb crime, and concomitant with the intervention the crime rate has indeed been strongly reduced. In the policy debate, the approach has often been equated with "zero tolerance policies". If that was the core of the matter, the critical component would have to be early intervention. If society spots the first signs of even minor social disorder, it should immediately step in and go after the perpetrators. Using a rich set of data from public-goods experiments run all over the world, we demonstrate that there is indeed a small effect of early vigilance, but it is not lasting. By contrast, there is a much stronger effect from first impressions. What participants experience in the first round of interaction explains their behavior throughout the entire game. This suggests that criminal policy should focus on impression management, rather than deterrence.
In a follow-up paper, Engel, Kube et al. (2011) manipulate, rather than measure, first impressions. Exploit-ing the fact that there are pronounced differences between the general willingness to cooperate in a public good across locations, they give negative information to participants in a context where co-operation is generally high, and they give positive information to participants in a context where co-operation is generally low. Negative information has a strong effect, whereas the effect of positive information is much weaker. That asymmetry, too, is important news for criminal policy. It is much more difficult to put a community back on a good track once social order is on the decay. By contrast, if first impressions are poor, this is very likely even to affect individuals who, in principle, would have been willing to live up to more demanding social standards.
Cornelißen, Himmler et al. (2013) complement the findings from the lab with evidence from the field. If individuals gain the impression that society is not applying fair standards, this can spill over to other, unrelated areas of life. Specifically if, in a country, the rich are perceived to pay too little in taxes, this engenders a considerably higher degree of absenteeism at the workplace.
A final experiment is related, although it really is only a preparatory step towards the legal research question that has triggered the endeavor. Arguably, many legal rules do not achieve their stated purpose because their addressees ad hoc compare the benefit from violating the rule with the expected cost from legal intervention. Rather, the legal rule shapes how addressees construe the issue, and it leads to the adoption of behavioral routines. While automization in principle enhances the governance effect of legal intervention, it potentially has a downside. Once addressees have established their routines, it may be very difficult for the law to change them, even if the old rule is no longer functional. Our experiment tests how difficult it is to overcome a routine, and in which ways the presence of significant others (arguably other individuals the law effectively alerts to the rule change) helps. Being paired with a role model indeed helps participants realize that the world has changed (Betsch, Lindow et al. 2011).

b) Punishment
The quintessential technique to curb socially undesired behavior is punishment. One of the surprises of the literature in experimental economics has been the willingness of experimental subjects to engage in costly punishment, although this means contributing to a second-order public good. Balafoutas and Nikiforakis (2012) test, in a field experiment, whether this is just an artifact of the experimental setting. They study violations of two well-known social norms by passengers in a large Athens subway station. This provides them with an almost completely anonymous, one-shot interaction. They find punishment, but it is confined to a minority.
Traxler and Winter (2012) use survey evidence to demonstrate that individuals' willingness to punish violations of legal obligations like tax evasion or fare-dodging by disapproval or social exclusion depends on the perceived prevalence of the violation; the more a violation is common, the less the willingness is to engage in costly social punishment.
Engel and Zhurakhovska (2013b) rely on punishment in a linear public good to cast new light on an old puzzle in the legal literature. All over the world, judges receive a fixed salary. There are indirect incentives like career concerns or the prospect of re-election. But incentives could at best partly explain why judges overall seem to do a reliable job. The authors isolate an alternative motive by adding an additional player to the game. This player does not benefit from the public good. She receives a flat fee plus an additional endowment she may use for punishing active players. Most of these additional players neither keep their endowments for themselves nor use them in a spiteful manner. Rather they discipline the groups they happen to be assigned to. Being in charge suffices to motivate them.
In the field, the threat of sanctions is frequently imperfect. Arguably, the likely sanction, multiplied by the probability of its enforcement, is often below the benefit from breaking the legal rule. If addressees were to maximize profit, the sanction should be devoid of effect. This is not the impression one gains in legal practice. Individuals seem to care, even if the sanction does not deter. A public-goods experiment allows us to isolate one explanation. Individuals might be weakly averse against exploiting others. The degree of this aversion would not be strong enough to constrain the selfish motives inherent in a social dilemma. Yet, a small sanction may tilt the balance in favor of cooperation. Classifying participants by their individual degree of social value orientation, this is indeed what Engel (2013c) finds.
Glöckner, Kube et al. (2011) give participants two options for punishment: punishment that becomes immediately notified to the addressee, and punishment of which feedback is withheld until the end of the game. This combination is a much more effective deterrent than either form of punishment in isolation (see also Hilbig, Zettler et al. (2012): on the moderating role played by dispositional honesty/humility).
In the literature on social dilemmas, punishment is usually presented as a solution. If only those exposed to the dilemma are given the opportunity to sanction each other, even if sanctions are costly, this gives them a technology to discipline free-riders. Yet, punishment potentially has a dark side (which, of course, is a prominent justification for the state's monopoly of force). Those who have been punished may retaliate, which may trigger a feud. Nikiforakis, Noussair et al. (2012) demonstrate that the risk of feuds is much more pronounced if there is room for normative conflict. Specifically, they create conflict by offering two equally convincing fairness norms, equality and desert.
While feuds have the potential to be disastrous, experimental participants are not blind to this risk. If the design of the experiment makes it possible for the feud to destroy more and more resources, participants are less rather than more likely to let this happen (Nikiforakis and Engelmann (2011)). Nicklisch and Wolf (2011) do not find a pronounced negative effect of counter-punishment either.
Fischer, Grechenig et al. (2013) find another limitation of punishment. The limitation does not depend on whether punishment is by other group members or by an authority. While punishment stabilizes contribu-tions if the punisher is perfectly informed about the recipient's choices, punishment loses its beneficial effect if the information about the recipient's contributions to the public good is noisy.
Legal orders do not exclusively rely on criminal law for disciplining those who violate legitimate normative expectations. Frequently the law either exclusively or additionally gives those who suffer from norm violations the right to claim damages. In the law and economics literature, damages are routinely given a forward-looking interpretation. They are equated with fines. If that interpretation was correct, the right to claim damages should have the same disciplining effect as punishment. Eisenberg and Engel (2012) test this hypothesis in a linear public good. They find that compensation indeed has a disciplining effect, but it is confined to the slope of contributions. If the expected value of compensation is sufficiently high, contributions (almost) stabilize. Giving participants the choice between punishment and compensation does not change results. This speaks against recipients caring about intentions. They are no less disci-plined if they know that the punisher has been selfish (see also Grechenig, Baumann et al. (2011): on the deterrent effect of incomplete strict liability).
Kube and Traxler (2011) study the interaction of central and decentral sanctions. In a linear public good, participants are given the possibility to punish each other. In the treatment, participants are additionally exposed to a small risk of an automatic fine for any contribution less than the complete endowment. When participants decide about punishment, they do not know whether the recipient will also be fined. This reduces decentral punishment. Nonetheless, the welfare balance from introducing the second sanction is positive.
In the criminology literature, it is taken for granted that the certainty of punishment is more important than its severity. One explanation of the effect goes back to Gary S. Becker. He has shown that, in equilibrium, only risk-seeking individuals commit crimes. Risk-seeking implies that an increase in certainty is a stronger deterrent. Engel and Nagin (2013) implement a stealing game to test the hypothesis. It is only supported if the expected value of stealing is negative, i.e., if the sanction would deter a profit-maximizing individual. If the expected value of stealing is positive or if it is zero, i.e., if a profit-maximizing agent is indifferent between stealing and refraining from it, even risk-seeking participants exhibit a stronger reaction to an increase in severity.
Corruption is a vexing evil of modern societies. While legal orders all over the world make it a crime, they differ in one respect: some legal orders target the briber and the official with equally severe sanctions; other legal orders treat the briber more mildly. Engel, Goerg et al. (2012) demonstrate that asymmetric punishment is a bad idea. Punishing sentiments give bribers a psychologically credible threat if the recipient takes the bribe, but does not return the favor she was asked for. The asymmetry thus gives the briber a technology to enforce the corrupt deal.

c) Post Hoc Intervention
Not so rarely, the law post hoc changes the outcomes of relations individuals have created voluntarily. We have experimentally studied two such instances. The German statute on copyright gives authors the right to ask for an additional license fee if the work turns out a blockbuster. We show that this provision not only assuages the hurt feelings of individuals who find out they have sold a jewel for a trifle. Actually the rule is even efficient. More goods trade at a lower price. The rule implements the functional equivalent of a success-contingent contract. It does so despite the fact that our experimental authorities were almost exclusively guided by ex post equality, and disregarded the fact that the original fee insured the author against the risk of project failure (Engel and Kurschilgen 2011).
The second experiment addresses straightforward redistribution. Arguably, effort and redistribution are negatively correlated. The more income is redistributed, the lower the incentive to engage in productive effort. And the higher the effort, the smaller the support for redistribution. Using data from the World Value Survey, we find the former relationship, but not the latter. In the lab, we again find that redistribution reduces effort, but now we even find a positive willingness of those engaging in high effort to support redistribution (Buch and Engel 2012).
d) Formal Legal Intervention
In the law and economics literature, legal intervention is normally modeled as a change in the opportunity structure. The law changes the game, be that by changing payoffs, or by changing the structure of the game. This is often an elucidating perspective. But this view neglects many subtleties of the law in action. One strand of our work explicitly addresses these additional features.
In the law and economics literature, the situation into which the law intervenes is usually modeled as a well-defined problem. In the courtroom, this is frequently not a plausible assumption. The very essence of the conflict the court is asked to decide upon is imperfect knowledge. All agree that the problem to be solved is not well defined. There tends to be room for improvement, but even at the end of the legal procedure, the uncertainty is normally only reduced, but not completely removed. Essentially, court cases are ambiguous.
Relying on the parallel constraint satisfaction model that is at the core of the work of the independent research group in psychology, we study two implications of this ambiguity. Intuition serves an important function. It enables individuals to make choices even if they know they have only partly understood the issue. This might hold a piece of troubling news for the law. Intuition might be so powerful that it overrides standard-of-proof instructions. Intuitive jury members would override the distinction between preponderance of the evidence and beyond-reasonable-doubt instructions. Happily, this hypothesis turns out to be wrong. Apparently, the intuitive process of story construction can be fine-tuned by explicit intervention (Glöckner and Engel (2013)).
In a second respect, the news for the law is less benign. The adversarial system critically relies on the expectation that ritualized conflict is the best way of approximating truth. A classic illustration is criminal procedure. Using the same, deliberately ambiguous case as in the previous study, we randomly assign the roles of prosecutor and defense counsel. We have participants sketch pleadings. The task, however, is not to convince the jury, but to predict the decision of a neutral court; in criminal procedure, this is particularly relevant when deciding about a plea bargain. Despite the fact that we give a strong financial incentive for accuracy, random assignment of the role creates pronounced bias (Engel and Glöckner (2013)).
Intuitively, most lawyers would think that training and experience matter. But can the effect actually be shown, and how strong is it? Dickert, Herbig et al. (2012) test experienced lay judges, law students, and controls on a series of difficult criminal law cases that all center around the question whether action was premeditated. Law students are significantly and strongly more likely than the other two groups to decide the same way as the German Supreme Court. They are also least likely to show an emotional reaction, in particular when the information they receive about the case is rich (see also the overview on legal expertise and intuition by Glöckner and Ebert (2011)).
In most procedures, courts and other legal authorities do not just decide. They also have to justify their decisions. Engel and Zhurakhovska (2013a) implement a linear public good with an additional player in the role of authority to test the power of justification. In the baseline, the reasons given only go to the experimenter. Active group members know that authorities have to justify themselves, but they do not learn what authorities say. In a first treatment, reasons are only communicated to the recipient of punish-ment. In a second treatment, all group members learn all reasons. Authorities reduce monetary punish-ment whenever reasons are communicated to addressees. Yet, reasons actually do only serve as substi-tutes if they are made public.
The right to be heard is a fundamental feature of the rule of law. Kleine, Langenbach et al. (2013) use a lab experiment to measure how much individuals care. In the first part of the experiment, two participants are given a real-effort task with asymmetric workload and piece rate. That way the authors induce a conflict between fairness norms. An impartial third participant decides upon the distribution of profit between the two active participants. In the treatment, active participants may send the authority a message about the distribution they deem fair. The actual measure is an unannounced dictator game where active participants may share a fraction of a new endowment with the authority. If they had voice, they would share almost twice as much, even if the decision of the authority were not in their favor.
In the old days, the roles of prosecutor and judge were not separated. It has been a major advancement in the rule of law to make judges impartial. The price was the openly one-sided role of prosecution. Now prosecutors are frequently suspected of using whatever ploy they can to get the defendant convicted. Engel and Pluta (2011) model the situation as a sequential game and test it in the lab. If, but only if they frame the situation as criminal procedure, the willingness to exploit an ambiguous rule to tilt the odds in favor of conviction is significantly and substantially reduced. At least in the lab, prosecutors are not the people's hired guns (do also see van Aaken, Feld et al. (2010): on prosecutor independence as a safeguard against corruption).
Sometimes the institute is the locus of friendly, but lively academic debate. The law of evidence is an instance. Mark Schweizer has joined the institute with a stipend from the Swiss National Science Foundation to finish his habilitation on the law of evidence. In his book, he forcefully argues in favor of a Bayesian approach to the assessment of disputed and incomplete evidence. Engel (2012b) invokes a series of largely undisputed solutions to standard legal cases to argue that, quite often, the legal order deliberately neglects the base rate. Schweizer (2012) counters that, if he forces triers to assess the probative value of each individual piece of evidence separately, and if he uses Bayesian network analysis to aggregate these assessments into an overall assessment of guilt, the bias in the assessment of the evidence resulting from being coherent with one's own decision largely disappears. (This is, however, not how trials are organized, and as a practical matter it seems hard to imagine how they could be reorganized along those lines). The debate between the two approaches is likely to remain lively.

3. Institutional Design
Over the last years, we have become more active in the area of institutional design than before. Casella, Llorente-Saguer et al. (2012) model a proposal that keeps being made. If a committee decides by vote, votes tend to be non-transferable. In particular, it tends to be forbidden to trade votes against other goods, be that money or favors. From a normative perspective, the ban on trading votes is not obvious. If the valuations for the outcome of the vote are heterogeneous, arguably trade improves welfare, in that those with low valuation for the issue may get something in recompense for letting those with high valuation have their will. In theory, this intuition only tends to get it right under special circumstances. If agents hold standard preferences, in equilibrium a single agent always becomes a dictator. If the commit-tee is not small, this reduces welfare, since the probability of being pivotal is larger than the probability of being dictator. The authors test the predictions of the model in the lab. There is less dictatorship than predicted. Choices and thereby outcomes are influenced by the fact that the majority of participants is risk-averse. Yet, with these qualifications, the model is supported (see also Bade (2011b): on electoral competition if parties are averse against uncertainty).
Bouton, Castanheira et al. (2012) study both theoretically and experimentally the properties of plurality and approval voting. They consider a case in which the majority is divided between two alternatives as a result of information imperfections, and the minority backs a third alternative which the majority views as strictly inferior. The majority thus faces two problems: aggregating information and coordinating to defeat the minority candidate. They show that, under plurality, voters may have to concentrate all their strength in fighting one of these problems. With approval voting, voters can address both problems simultaneously.
Statements made during election campaigns are often accused of merely being cheap talk. Corazzini, Kube et al. (forthcoming) use a lab experiment to cast doubt on this skeptical position. They let two out of seven participants run for office and give them the possibility to make non-binding commitments in a campaign stage. Commitments are predominantly in the interest of voters, and the one candidate winning the election predominantly keeps her word.
There is an old debate over representative versus direct democracy. This debate usually discusses the issue in normative terms. Towfigh, Glöckner et al. (2013) add an experimental perspective. In an online vignette study, they present voters in the wake of regional elections with plausible decisions taken on three hotly debated issues, and elicit acceptance scores. The subjective importance of the issue turns out to be a critical moderator. Acceptance is higher if the decision is taken by referendum only if the individual classifies the issue as one of highest importance for herself.
In a theory paper, Grechenig and Kolmar (forthcoming) add one argument to the justification of internal sovereignty. It provides individuals with a commitment device. They need not engage in a war of ever more powerful, but also more costly, devices for the protection of their property.
In the ordo-liberal tradition, antitrust is not only propagated as a means for preventing customers from being exploited. Ordo-liberals also expect economic power to corrupt democracy. Using the fact that many countries have adopted antitrust over the past decade, and exploiting deeper lags in the develop-ment of a well-established democracy index for the country in question as an instrument, Petersen (2013a) demonstrates that the introduction of antitrust does not have a significant effect on the level of democracy (while it has a beneficial effect on economic growth).
In a doctrinal piece, Petersen (2013c) shows that the German Constitutional Court uses legislative incon-sistency as a signal that the justification brought forward for a legislative intervention is not sincere and hides illegitimate motives.
In a series of papers, political scientist Philip Leifeld demonstrates the power of his methodology for the analysis of political discourse, using concepts from graph theory. Leifeld and Schneider (2012) introduce the methodology and demonstrate in which ways it reflects the theory of political discourse. Leifeld applies his methodology to American climate politics (Fisher, Waggle et al. 2013a; Fisher, Waggle et al. 2013b), to the European conflict over software patents (Leifeld and Haunss 2012), and to the German debate over pension reform (Leifeld 2013).

4. Meta Questions
a) Social Sciences and the Law
The lawyers among us openly adopt methods from the social sciences. While this approach would seem quite natural in a good US law school, it still bears explanation, if not justification, in Germany (and most other European countries, for that matter). We therefore cannot avoid engaging ourselves in the legal debate over methodology. Petersen (2013b) advocates empirical methods from the social sciences as a safeguard against the constitutional courts that fall prey to widespread, but erroneous beliefs. In his contribution to the upcoming Oxford Handbook of Behavioral Law and Economics, Engel (2013b) discuss-es the pros and cons of field data, survey data, vignette studies, and incentivizing lab experiments. Engel (2013d) is a systematic account of the most prominent objections against lab experiments on legal issues, and demonstrates in which ways they can be addressed by the design of the experiment and of experi-mental papers.
Further contributions demonstrate the fruitfulness of economic theory for understanding and interpreting fundamental freedoms (Engel 2012c) and economic law (Lüdemann 2013).
Finally, two empirical papers turn legal academia into an object of study. German legal scholars tend to think that the discipline specific type of training is effective. Using the complete data from all mock exams written in a large German law faculty over an entire year, Glöckner, Towfigh et al. (2013) find an interesting non-linearity. Over the course of the year, candidates' performance does not improve monotonically. Rather, they undergo characteristic dips, which are likely caused by drops in motivation.
Engel and Hamann (2012) investigate another non-linearity. Observers have long suspected that the supply of law professors is irregular in a rather specific sense: if there have been too many good candidates, some of whom had a hard time becoming professors, some years later, low and behold, there is a dearth of candidates. Likewise, if universities have had a hard time for a while finding convincing candidates, a few years later there are too many of them. It seems that the supply of law professors has the characteristics of a hog cycle. Using data about individuals passing the entry barrier of habilitation, the authors demonstrate that the suspicion actually holds true. The supply of law professors is negatively correlated to the supply of eight years ago.

b) Experimental Methodology
Since the group predominantly uses the experimental method, it also seemed natural to make a few contributions to experimental methodology. Cleave, Nikiforakis et al. (2013) test whether participants with particularly pronounced social preferences self-select into laboratory experiments. Overall, they do not find signs of selection bias. But participants who are less likely to send a substantial amount in the trust game are significantly more likely to participate in later experiments.
Jekel, Fiedler et al. (2011) introduce a technology for classifying judgment and decision-making by choices in diagnostic tasks, and demonstrate how the resulting data can be analyzed.

c) Data Analysis
We finally make contributions to the techniques of data analysis. The data generated by our experiments are frequently censored. The standard statistical approach to analyze such data is the Tobit model. In the frequent case of censoring at zero, the Tobit model assumes that a fraction of the observed zeros would actually have been negative, had the design of the experiment not made this impossible. In the case of punishment, one must therefore assume that participants who have not punished would actually have wanted to reward other participants. Depending on the nature of the game, this can be a fairly strong assumption. It may be more plausible to assume that participants who do not punish just do not want to contribute to the second-order public good. If that assumption is correct, a two-part of model is appro-priate. One first explains why participants do not want to punish at all. Conditional on their decision to engage in punishment, one separately explains severity. If this is the appropriate statistical model, one can estimate a logit or probit for the first part, and can use OLS for the second part. The only adjustment is for the truncated distribution of the error terms in the second part of the model. Yet, this still requires a fairly strong assumption. If a participant does not punish, this implies that said participant would have punished under no circumstances. Yet, why should a participant punish another participant who seems to have behaved well? If one wants to make room for that, one needs a statistical model that allows for two types of zeros: those resulting from the unwillingness to punish in the first place and those resulting from local conditions, and in particular the amount contributed to the public good by the would-be recipient. Engel and Moffat (2012) work out the theory of such a double-hurdle model, program separate versions for cross section and panel data in Stata, and demonstrate the power of the approach with a long-standing bone of contention among experimentalists: the house money effect (see also Leifeld: on a software tool to export tables in R into LaTeX or HTML).

C.II.1.3 References

Aretz, Bodo and Sebastian Kube (2013). "Choosing Your Object of Benevolence: A Field Experiment on Donation Options." The Scandinavian Journal of Economics 115(1): 62-73.
Bade, Sophie (2011a). "Ambiguous Act Equilibria." Games and Economic Behavior 71: 246-260.
Bade, Sophie (2011b). "Electoral Competition with Uncertainty Averse Parties." Games and Economic Behavior 72: 12-29.
Balafoutas, Loukas and Nikos Nikiforakis (2012). "Norm Enforcement in the City: A Natural Field Experiment." European Economic Review 56: 1773-1785.
Betsch, Tilmann, Stefanie Lindow, Christoph Engel, Corina Ulshöfer and Janet Kleber (2011). Has The World Changed? My Neighbor Might Know Effects of Social Context on Routine Deviation. Bonn, Max Planck Institute for Research on Collective Goods
Bouton, Laurent, Micael Castanheira and Aniol Llorente-Saguer (2012). Divided Majority and Information Aggregation: Theory and Experiment Bonn, Max Planck Institute for Research on Collective Goods
Buch, Claudia M. and Christoph Engel (2012). Effort and Redistribution: Better Cousins Than One Might Have Thought. Bonn, Max Planck Institute for Research on Collective Goods.
Bünstorf, Guido, Christoph Engel, Sven Fischer and Werner Güth (2013). Win Shift Lose Stay. An Experimental Test of Non-Compete Clauses.
Casella, Alessandra, Aniol Llorente-Saguer and Thomas R. Palfrey (2012). "Competitive Equilibrium in Markets for Votes." Journal of Political Economy 120: 593-658.
Chmura, Thorsten, Christoph Engel and Markus Englerth (2013). Selfishness As a Potential Cause of Crime. A Prison Experiment. Bonn, Max Planck Institute for Research on Collective Goods

Chmura, Thorsten, Sebastian J. Goerg and Reinhard Selten (2012). Learning in experimental 2 x 2 games. Games and Economic Behavior 76:44-73.
Cleave, Blair L., Nikos Nikiforakis and Robert Slonim (2013) "Is There Selection Bias in Laboratory Experiments? The Case of Social and Risk Preferences." Experimental Economics 16(3): 372-382.
Corazzini, Luca, Sebastian Kube, Michel Maréchal and Antonio Nicolo (forthcoming). "Elections and deceptions: an experimental study on the behavioral effects of democracy." American Journal of Political Science.
Cornelißen, Thomas, Oliver Himmler and Tobias Koenig (2013). "Fairness Spillovers – The Case of Taxation." Journal of Economic Behavior & Organization 90: 164-180.
Deck, Cary and Nikos Nikiforakis (2012). "Perfect and Imperfect Real-Time Monitoring in a Minimum-Effort Game." Experimental Economics 15: 71-88.
Dickert, Stephan, Britta Herbig, Andreas Glöckner, Christina Gansen and Roman Portack (2012). "The More the Better? Effects of Training and Information Amount in Legal Judgments." Applied Cognitive Psychology 26: 223-233.
Dickert, Stephan, Janet Kleber, E. Peters and P. Slovic (2011). "Numeracy as a precursor to pro-social behavior: The impact of presentation format on the cognitive mechanisms underlying donation decisions." Judgment and Decision Making 6: 638-650.
Dickert, Stephan, N. Sagara and P. Slovic (2011). "Affective motivations to help others: A two-stage model of donation decisions." Journal of Behavioral Decision Making 24: 361-376.
Ding, Jieyao (2011). What Numbers to Choose for My Lottery Ticket? Behavior Anomalies in the Chinese Online Lottery Market. Bonn, Max Planck Institute for Research on Collective Goods
Ding, Jieyao (2012). A Portfolio of Dilemmas: Experimental Evidence on Choice Bracketing in a Mini-Trust Game. Bonn, Max Planck Institute for Research on Collective Goods.
Eisenberg, Theodore and Christoph Engel (2012). Assuring Adequate Deterrence in Tort: A Public Good Experiment. Bonn, Max Planck Institute for Research on Collective Goods
Engel, Christoph (2011a). Competition as a Socially Desirable Dilemma. Theory vs. Experimental Evidence. Competition Policy and the Economics Approach. Josef Drexl, Wolfgang Kerber und Rupprecht Podszun. Cheltenham: 245-269.
Engel, Christoph (2011b). "Dictator Games: A Meta Study." Experimental Economics 14: 583-610.
Engel, Christoph (2011c). "Die Bedeutung der Verhaltensökonomie für das Kartellrecht." Beitrag der Verhaltensökonomie (Behavioral Economics) zum Handels- und Wirtschaftsrecht (Beiheft der Zeitschrift für das gesamte Handelsrecht und Wirtschaftsrecht: 100-121.
Engel, Christoph (2011d). "The Emergence of a New Rule of Customary Law: An Experimental Contribut-ion." Review of Law and Economics 7: 767-789.
Engel, Christoph (2011e). An Experimental Contribution to the Revision of the Guidelines on Research and Development Agreements. FS Möschel. Baden-Baden, Nomos: 227-240.
Engel, Christoph (2011f). "When is Intellectual Property Needed as a Carrot for Innovators?" Journal of Competition Law and Economics 7: 277-299.
Engel, Christoph (2012a). Low Self-Control As a Source of Crime. A Meta-Study. Bonn, Max Planck Institute for Research on Collective Goods
Engel, Christoph (2012b). Neglect the Base Rate: It's the Law! Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph (2012c). Öffentliches Wirtschaftsrecht aus der Sicht der ökonomischen Theorie. Besonders Verwaltungsrecht I. Dirk Ehlers, Michael Fehling und Hermann Pünder. Heidelberg. I: 40-62.
Engel, Christoph (2013a). Aufgaben. Leitgedanken des Rechts. FS Paul Kirchhof. Hanno Kube und et al. Heidelberg: 57-69.
Engel, Christoph (2013b). Behavioral Law and Economics: Empirical Methods. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph (2013c). Deterrence by Imperfect Sanctions – A Public Good Experiment. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph (2013d). Legal Experiments: Mission Impossible? Eleven. Den Haag.
Engel, Christoph, Martin Beckenkamp, Andreas Glöckner, Bernd Irlenbusch, Heike Hennig-Schmidt, Sebastian Kube, Michael Kurschilgen, Alexander Morell, Andreas Nicklisch, Hans-Theo Normann and Emanuel Vahid Towfigh (2014). "First Impressions are More Important than Early Intervention. Qualifying Broken Windows Theory in the Lab." International Review of Law and Economics 37:126-136.
Engel, Christoph and Joshua Fairfield (2013). Big Brother is Watching You – Because Little Brother Has Opened the Door. An Experiment on Information Sharing in Social Networks.
Engel, Christoph and Andreas Glöckner (2013) "Role Induced Bias in Court: An Experimental Analysis." Journal of Behavioral Decision Making 26: 272-284.
Engel, Christoph, Sebastian J. Goerg and Gaoneng Yu (2012). Symmetric vs. Asymmetric Punishment Regimes for Bribery. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Hanjo Hamann (2012). The Hog-Cycle of Law Professors. Bonn, Max Planck Institute for Research on Collective Goods
Engel, Christoph and Marco Kleine (2013). Who is Afraid of Pirates? An Experiment on the Deterrence of Innovation by Imitation. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph, Sebastian Kube and Michael Kurschilgen (2011). Can we manage first impressions in cooperation problems? An experimental study on "Broken (and Fixed) Windows". Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Michael Kurschilgen (2011). "Fairness Ex Ante & Ex Post – An Experimentally Testing Ex Post Judicial Intervention into Blockbuster Deals." Journal of Empirical Legal Studies 8: 682–708.
Engel, Christoph and Michael Kurschilgen (2013). "The Coevolution of Behavior and Normative Expectations. Customary Law in the Lab." American Law and Economics Review 15(2): 578-609.
Engel, Christoph and Peter G. Moffat (2012). Estimation of the House Money Effect Using Hurdle Models
Engel, Christoph and Daniel Nagin (2013). Who is Afraid of the Stick? Experimentally Testing the Deterrent Effect of Sanction Certainty. Working Paper.
Engel, Christoph and Axel Ockenfels (2013). Maverick. Testing a Conjecture of Antitrust Policy in the Lab. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Alicja Pluta (2011). The People's Hired Guns? Experimentally Testing the Inclination of Prosecutors to Abuse the Vague Definition of Crimes. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Lilia Zhurakhovska (2011). Oligopoly as a Socially Embedded Dilemma. An Experiment. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Lilia Zhurakhovska (2012a). Harm on an Innocent Outsider as a Lubricant of Cooperation – An Experiment. Bonn, Max Planck Institute for Research on Collective Goods
Engel, Christoph and Lilia Zhurakhovska (2012b). When is the Risk of Cooperation Worth Taking? The Prisoner's Dilemma as a Game of Multiple Motives. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Lilia Zhurakhovska (2013a). Words Substitute Fists. Justifying Punishment in a Public Good Experiment. Bonn, Max Planck Institute for Research on Collective Goods.
Engel, Christoph and Lilia Zhurakhovska (2013b). You are in Charge! Experimentally Testing the Motivating Power of Holding a (Judical) Office. Working paper.
Englerth, Markus and Yoan Hermstrüwer (forthcoming). "Die Datenkrake als Nutztier der Strafverfolger – Zum strafprozessualen Zugriff auf Facebook-Profile." Rechtswissenschaft.
Fehr, Dietmar, Frank Heinemann and Aniol Llorente-Saguer (2011). The Power of Sunspots: An Experimental Analysis. Bonn, Max Planck Institute for Research on Collective Goods.
Fiedler, Susann and Andreas Glöckner (2012). "The dynamics of decision making in risky choice: An Eye-tracking Analysis." Frontiers in Psychology 3.335.
Fiedler, Susann, Andreas Glöckner, Stephan Dickert and Andreas Nicklisch (2013) "Social Value Orientation and information search in social dilemmas: An eye-tracking analysis." Organizational Behavior and Human Decision Processes 120: 272-284.
Fiedler, Susann, Andreas Glöckner and Andreas Nicklisch (2012). The Influence of Social Value Orientation on Information Processing in repeated Voluntary Contribution Mechanism (VCM) Games: An Eye-tracking Analysis. Neuroscience and the Economics of Decision-Making. A. Sirigu und A. Innocenti: 21-53.
Fischer, Ilan, Alex Frid, Sebastian J. Goerg, Simon A. Levin, Daniel I. Rubenstein and Reinhard Selten (2013) "Fusing enacted and expected mimicry generates a winning strategy that promotes the evolution of cooperation." Proceedings of the National Academy of Sciences 110(25): 10229-10233.
Fischer, Sven, Sebastian J. Goerg and Hanjo Hamann (2013). Cui Bono, Benefit Corporation? An Experiment Inspired by Social Enterprise Legislation in Germany and the US. Bonn, Max Planck Institute for Research on Collective Goods.
Fischer, Sven, Kristoffel Grechenig and Nicolas Meier (2013). Cooperation under punishment: Imperfect information destroys it and centralizing punishment does not help. Bonn, Max Planck Institute for Research on Collective Goods.
Fischer, Sven and Werner Güth (2012). "Effects of exclusion on acceptance in ultimatum games." Journal of Economic Psychology 33: 1100-1114.
Fisher, Dana R., Joseph Waggle and Philip Leifeld (2013a). "Mapping the Ideological Networks of American Climate Politics." Climatic Change 116: 523-545.
Fisher, Dana R., Joseph Waggle and Philip Leifeld (2013b). "Where does Political Polarization Come From? Locating Polarization Within the U.S. Climate Change Debate." American Behavioral Scientist 57: 70-92.
Glöckner, Andreas and I. D. Ebert (2011). Legal intuition and expertise. Handbook of Intuition Research. M. Sinclair. Northampton, MA, Edward Elgar: 157-167.
Glöckner, Andreas and Christoph Engel (2013). "Can We Trust Intuitive Jurors? Standards of Proof and the Probative Value of Evidence in Coherence Based Reasoning." Journal of Empirical Legal Studies 10: 230-252.
Glöckner, Andreas and B. E. Hilbig (2012). "Risk is relative: Risk aversion yields cooperation rather than defection in cooperation-friendly environments." Psychonomic Bulletin & Review 19: 546-553.
Glöckner, Andreas, Bernd Irlenbusch, Sebastian Kube, Andreas Nicklisch and Hans-Theo Normann (2011). "Leading with(out) Sacrifice? A Public-Goods Experiment with a Super Privileged Player." Economic Inquiry 49: 591-597.
Glöckner, Andreas, Sebastian Kube and Andreas Nicklisch (2011). The Joint Benefits of Observed and Unobserved Punishment: Comment to Unobserved Punishment Supports Cooperation. Bonn, Max Planck Institute for Research on Collective Goods.
Glöckner, Andreas and Thorsten Pachur (2012). "Cognitive models of risky choice: Parameter stability and predictive accuracy of Prospect Theory." Cognition 123: 21-32.
Glöckner, Andreas, Emanuel Towfigh and Christian Traxler (2013). "Development of legal expertise." Instructional Science 41:989-1007.
Goerg, Sebastian J., Johannes Abeler, Steffen Altmann, Sebastian Kube and Matthias Wibral (2011). "Equity and Efficiency in Multi-Worker Firms: Insights from Experimental Economics." Analyse & Kritik 33: 325-347.
Goerg, Sebastian J. and Sebastian Kube (2012). Goals (Th)at Work. Goals, Monetary Incentives, and Workers Performance. Bonn, Max Planck Institute for Research on Collective Goods.
Gollwitzer, M., T. Rothmund, B. Alt and Marc Jekel "Victim sensitivity and the accuracy of social judgments." Personality and Social Psychology Bulletin 38: 975-984.
Grechenig, Kristoffel, Florian Baumann and Tim Friehe (2011). "A note on the optimality of (even more) incomplete strict liability." International Review of Law and Economics (IRLE) 31: 77-82.
Grechenig, Kristoffel and Martin Kolmar (forthcoming). "The State's Enforcement Monopoly and the Private Protection of Property." Journal of Institutional and Theoretical Economics.
Grechenig, Kristoffel and V. Roberto (2011). "Rückschaufehler («Hindsight Bias») bei Sorgfaltspflicht-verletzungen." Zeitschrift für Schweizerisches Recht (ZSR) 130: 5-26.
Hamann, Hanjo (2012). "Reflektierte Optimierung oder bloße Intuition?" ZGR, Zeitschrift für Unternehmens- und Gesellschaftsrecht 41: 817-834.
Han, G. and Marc Jekel (2011). "The mediating role of job satisfaction between leader-member exchange and turnover intentions." Journal of Nursing Management 19: 41-49.
Hilbig, B. E., I. Zettler and T. Heydasch (2012). "Personality, punishment, and public-goods: Strategic shifts towards cooperation as a matter of dispositional Honesty-Humility." European Journal of Personality 26: 245-254.
Himmler, Oliver, Thomas Cornelissen and Tobias Koenig (2011). "Perceived Unfairness in CEO Compensation and Work Morale." Economics Letters 110: 45-48.
Hortala-Vallve, Rafael and Aniol Llorente-Saguer (2012). "Pure Strategy Equilibria in Non-zero Sum Colonel Blotto Games." International Journal of Game Theory 41: 331-343.
Hortala-Vallve, Rafael, Aniol Llorente-Saguer and Rosemarie Nagel (2013). "The Role of Information in Different Bargaining Protocols." Experimental Economics 16: 88-113.
Jekel, Marc, Susann Fiedler and Andreas Glöckner (2011). "Diagnostic task selection for strategy classification in judgment and decision making." Judgment and Decision Making 6: 782–799.
Kleine, Marco, Pascal Langenbach and Lilia Zhurakhovska (2013). Voice Effects on Attitudes towards an Impartial Decision Maker. Experimental Evidence.
Kube, Sebastian, Clemens Puppe and Michel André Maréchal (2013) "Do Wage Cuts Damage Work Morale? Evidence from a Natural Field Experiment." Journal of the European Economic Association 11(4): 853-870.
Kube, Sebastian, Clemens Puppe and Michel André Maréchal (2012). "The Currency of Reciprocity – Gift Exchange in the Workplace." American Economic Review 102: 1644-1662.
Kube, Sebastian and Christian Traxler (2011). "The interaction of legal and social norm enforcement." Journal of Public Economic Theory 13: 639-660.
Leifeld, Philip (forthcoming) "texreg: Conversion of Statistical Model Output in R to LaTeX and HTML Tables." Journal of Statistical Software.
Leifeld, Philip (2013). "Reconceptualizing Major Policy Change in the Advocacy Coalition Framework. A Discourse Network Analysis of German Pension Politics." The Policy Studies Journal 41: 169-198.
Leifeld, Philip and Sebastian Haunss (2012). "Political Discourse Networks and the Conflict over Software Patents in Europe." European Journal of Political Research 51: 382-409.
Leifeld, Philip and Volker Schneider (2012). "Information Exchange in Policy Networks." American Journal of Political Science 53: 731-744.
Lüdemann, Jörn (2013). Öffentliches Wirtschaftsrecht und ökonomisches Wissen: Extrajuridisches Wissen im Verwaltungsrecht. I. Augsberg: 121-149.
McDonald, Ian, Nikos Nikiforakis, Nilss Olekalns and Hugh Sibly (2013) "Social Comparisons and Reference Group Formation: Some Experimental Evidence." Games and Economic Behavior 79: 75-89.
Morell, Alexander (2011). (Behavioral) Law and Economics im europäischen Wettbewerbsrecht. Wirtschaftsrecht und Wirtschaftspolitik. Baden-Baden, Nomos 284: 1-284.
Nicklisch, Andreas and Irenaeus Wolf (2011). "The cooperation norms in multiple-stage punishment." Journal of Public Economic Theory 13: 791-827.
Nicklisch, Andreas and Irenaeus Wolff (2012). "On the nature of reciprocity: Evidence from the ultimatum reciprocity measure." Journal of Economic Behavior and Organization 84: 892-905.
Nikiforakis, Nikos and Dirk Engelmann (2011). "Altruistic Punishment and the Threat of Feuds." Journal of Economic Behavior and Organization 78: 319-332.
Nikiforakis, Nikos, Charles Noussair and Tom Wilkening (2012). "Normative Conflict and Feuds: The Limits of Self-Enforcement." Journal of Public Economics 96: 797-807.
Nisvan, Erkal, Lata Gangadharan and Nikos Nikiforakis (2011) "Relative Earnings and Giving in a Real Effort Experiment." American Economic Review 101(7): 3330-3348.
Petersen, Niels (2013a) "Antitrust Law and the Promotion of Democracy and Economic Growth." Journal of Competition Law & Economics 9: 593-636.
Petersen, Niels (2013b) "Avoiding the Common Wisdom Fallacy: The Role of Social Sciences in Constitutional Adjudication." International Journal of Constitutional Law 11(2): 294-318.
Petersen, Niels (2013c) "Gesetzgeberische Inkonsistenz als Beweiszweichen." Archiv des öffentlichen Rechts 138(1): 108-134.
Petersen, Niels (forthcoming). The Role of Consent and Uncertainty in the Formation of Customary International Law. Reexamining Customary International Law. Brian D. Lepard. Cambridge, Cambridge University Press.
Schweizer, Mark (2012). Comparing Holistic and Atomistic Evaluation of Evidence. Bonn, Max Planck Institute for Research on Collective Goods.
Selten, Reinhard, Thorsten Chmura and Sebastian J. Goerg (2011). "Stationary Concepts for Experimental 2x2 Games: A Reply." The American Economic Review 101: 1041-1044.
Towfigh, Emanuel Vahid and Andreas Glöckner (2011). "Game over: Empirical support for soccer bets regulation." Psychology, Public Policy, and Law 17: 475-506.
Towfigh, Emanuel Vahid, Andreas Glöckner, Sebastian J. Goerg, Philip Leifeld, Carlos Kurschilgen, Aniol Llorente-Saguer and Sophie Bade (2013). Does Political Representation through Parties Decrease Voters' Acceptance of Decisions? Bonn, Max Planck Institute for Research on Collective Goods
Traxler, Christian and Mathias Spichtig (2011). "Social Norms and the Indirect Evolution of Conditional Cooperation." Journal of Economics 102: 237-262.
Traxler, Christian and Joachim Winter (2012). "Survey Evidence on Conditional Norm Enforcement." European Journal of Political Economy 28: 390-398.
van Aaken, Anne, Lars Feld and Stefan Voigt (2010) "Do Independent Prosecutors Deter Political Corruption? An Empirical Evaluation across 78 Countries." American Law and Economics Review 12(1), 204-244.
von Weizsäcker, Carl Christian (2011). Homo Oeconomicus Adaptivus – Die Logik des Handelns bei veränderlichen Präferenzen. Wohin steuert die ökonomische Wissenschaft? Ein Methodenstreit in der Volkswirtschaftslehre. V. Caspari und B. Schefold. Frankfurt/New York. 3: 221-255.
Weinschenk, Philipp (2011). Procrastination in Teams, Contract Design and Discrimination. Bonn, Max Planck Institute for Research on Collective Goods.