Welcome to the International Max Planck Research School on Behaviorally Smart Institutions!

Mission

Institutions are ubiquitous. All human behavior is embedded in and guided by formal and informal institutions. The discipline of economics models individuals as having well-defined preferences, and as maximizing their personal well-being. In this perspective, institutions are constraints. The economic model predicts that different institutions give rise to differences in human behavior. This prediction resonates with observation:  if one is by default an organ donor, many more organs are donated; if a reputation system reliably tracks the trustworthiness of sellers in an online platform, there is less fraudulent behavior; if the usage of roads is priced contingent on traffic conditions, there is less congestion; if freeriders can be sanctioned, all group members cooperate more. All of these examples illustrate how human behavioral dispositions and institutions interact.

Most formal institutions are legal, and many non-legal institutions are regulated by law. The disciplines of both law and economics are therefore complementary for analyzing the interplay of human behavior and institutions, and to develop institutions that promise outcomes that are individually or socially more desirable. Yet (neoclassical) law and economics starts from a very narrow definition of human behavior. This is worrisome both from an analytic and a normative perspective. While a particular institution might be optimal in case of a rational, well-informed, and forward-looking individual, this may not at all be the case when human behavior deviates from these assumptions. Addressing the key limitations of human behavior and deriving implications for institutional design from these limitations is the task of the proposed graduate school. In the following, we sketch some of the most troublesome, but also intellectually most interesting limitations. Informing institutional analysis and design about these limitations are the focus of the school.

Real institutions do not address the agents of economic textbooks; they address real people. Real people care about more goods than money. This can be captured by shifting from profit to utility space. But a richer utility function is still exclusively motivational. Arguably the reason why real people behave in normatively undesirable ways is also cognitive. They may not understand what would be in their individual best interest. They may misinterpret the situation. They may overestimate a risk and underestimate an opportunity, or they may be overly optimistic. The way how they read the available incomplete evidence may be tilted towards their personal interest. Even if an individual has the best of all intentions, she may lack the ability to act upon them. Socially undesirable behavior may be a habit that she cannot overcome on short notice. She may lack the necessary self-control, or the ability to plan her life on a sufficiently long time horizon. For these and many related reasons it is paramount for institutional designers to understand the mental mechanisms and motivations that determine choices.

It is by no means the case, however, that all humans are the same. For the analyst and designer, heterogeneity in human behavior is a challenge. First-generation behavioral analysis tends to bracket heterogeneity. It implicitly assumes that variation in behavioral determinants is distributed in some well-behaved way around some central tendency. Technically, the variance is put into a noise term. Analysis focuses on the average population effect. Now again reality can be quite different. A substantial literature has, for instance, shown that most populations consist of a sizeable minority of selfish individuals, a small minority of unconditional altruists, and a majority of conditional cooperators. This majority is good-natured only if it knows or expects a sufficient fraction of their interaction partners to be good-natured as well. Such patterned heterogeneity is not only much harder to identify and measure. It also exposes institutional designers to a much harder problem, because institutions should provide a useful framework not just for one particular type of human behavior, but for a possibly large set of different types. Ideally institutions should engender desirable behavior even when humans are characterized by multiple heterogeneous behavioral limitations.

Sometimes institutional designers can adopt a two steps approach. In the first step they reduce behavioral complexity to a degree that makes it possible to ignore occasional deviations. Markets often have this effect, as suppliers who ignore market pressure are forced to leave the market. If behavior is embedded in formal or informal institutional structures, it becomes much more predictable what agents are likely to do. But quite often institutional designers have to take their addressees with the behavioral patterns and limitations they happen to have. Therefore, in a second step it is necessary to study these patterns empirically. Running experiments will be the prime method to do so in this graduate school, since by random assignment of subjects to different treatment conditions it is possible to causally identify which institutional features generate which type of behavior, conditional on behavioral patterns and limitations. As such, experiments in the laboratory or randomized control trials in the field allow for testing how institutional design affects human behavior.

Behaviorally informed institutional design need, however, not be more challenging than designing institutions for agents one expects to maximize profit. In many contexts many individuals are good-natured. Or to use the language introduced by Aristotle: man is a social being. When taking action, humans tend to reflect the effects on others. Many care about being good members of the social groups to which they belong. This opens up an avenue for less intrusive intervention. It may suffice to overcome temptation, or the unsubstantiated fear of being harmed. It may even be enough to make the individual see the situation in the appropriate light. This approach has been prominently labeled nudging. Compared with the inhabitants of the animal kingdom, the behavior of humans is extraordinarily plastic. One might even say: the human species specializes in reinventing itself with every new generation. It is not pigeonholed in evolutionary niches, and has the ability to conquer whatever new environment it faces. Humans are the more plastic the younger they are. This makes it important to understand in which ways the behavior of the next generation is shaped in their youth, and to which degree this process is open to purposeful institutional intervention.

An institutional designer is a social engineer. She wants to change the behavior of some discernible fraction of the population. If unconstrained, she goes for the intervention that promises to be most effective, or the monetary or political cost of which she deems affordable. Yet humans do not want to be treated like pawns on a chessboard. They have dignity and care about being respected by the state and others under whose sway they are. Not everything that can be done should be done. A normative discourse is required. But for this discourse to be meaningful, one must understand the object of intervention, i.e. human behavior. Handing the issue over to the inhabitants of the philosophical ivory tower is not enough. The normative discussion must go hand-in-hand with a growing understanding of mental and social mechanisms.

In our research program, the following questions will guide us: How can one help institutional analysts and designers make more adequate definitions of the problems that call for intervention? And how can one help them design interventions that are more effective, and ideally also less intrusive? How can this program of making institutional design smarter come to life? The answers will not be found in the silo of one discipline. One needs behavioral research to deeply understand the determinants of human behavior. One needs comparative research to assess the variability of human behavior, and its plasticity. One needs the analysis of existing institutions: in which contexts have the interventions delivered on their promises? In which contexts have they been counter-productive? And when have they been robust to which changes in the environment, or the political landscape? And one needs input from those who understand the existing arsenal of interventions: what has been used where, and for which purposes? Which are the framework conditions that must be respected?

Go to Editor View