Jörgen Weibull

We analyze a principal-agent model with moral hazard where the principal is risk neutral and the agent is risk averse or risk neutral. The agents' actions determine the probability distribution over outcomes, where some probability distributions require more effort than others. We analyze a family of effort disutility functions with arguably desirable properties. For a canonical class of such functions and arbitrary outcome spaces, the model is explicitly solvable. Optimal contracts combine debt with a monotonic sharing rule for the surplus created. When the agent is risk neutral, the contract is a pure debt contract. For agents with unit degree of relative risk aversion, the surplus is divided in fixed shares

Theorem 1 in Alger and Weibull (Games and Economic Behavior, 2016) consists of two statements. The first establishes that Homo moralis with the right degree of morality is evolutionarily stable. The second statement is a claim about sufficient conditions for other goal functions to be evolutionarily unstable. However, the proof given for that claim presumes that all relevant sets are non-empty, while the hypothesis of the theorem does not guarantee that. We here prove instability under a stronger hypothesis that guarantees existence, and we also establish a new and closely related result. As a by-product, we also obtain an extension of Theorem 1 in Alger and Weibull (Econometrica, 2013).

In recent years many parliaments have become fragmented and polarized, with difficulties to form government. Traditional party alliances have changed, new parties have been established, and some of these are shunned by the traditional parties. Fragile minority governments are formed and break up. We propose a game-theoretic framework for analysis of political power in parliaments for situations when party affilations change and new coalitions are formed. More speciffically, we use and modify tools from cooperative game theory, in particular Myerson (1977) and Alonso-Meijide and Carreras (2011). The analytical framework is applied to the current Swedish parliament

A rationally inattentive screener evaluates a heterogeneous pool of candidates, aiming to let only high-productivity candidates pass. Candidates belong to distinct and observable social categories. There is heterogenity across categories along multiple dimensions, including the costs of investment in skills, the costs of being screened, and the degree of bias faced in the screening process. Candidates choose how much effort to invest in skill development before being screened, with a payoff in a post-screening market that depends on the screening outcome. We characterize equilibrium behavior in this model, and use it to unify and extend several strands in the literature on categorical inequality, including statistical discrimination, prejudice, and social capital.

We consider a rationally inattentive screener who evaluates a pool of candidates composed of distinct and observable social categories. There is heterogeneity across categories in the costs of being screened, the degree of bias faced in the screening process, and the costs of investment in skills. Candidates choose how much effort to invest before being screened, with a payoff in a post-screening market that depends on the screening outcome. We characterize equilibrium in this model, and use it to unify and extend several strands in the literature on categorical inequality, including statistical discrimination, prejudice, and social capital.

Recent theoretical work suggests that a form of Kantian morality has evolutionary foundations. To investigate the relative importance of Kantian morality and social preferences, we run laboratory experiments on strategic interaction in social dilemmas. Using a structural model, we estimate social preferences and morality concerns both at the individual level and the aggregate level. We observe considerable heterogeneity in social preferences and Kantian morality. A finite mixture analysis shows that the subject pool is well described as consisting of two types. One exhibits a combination of inequity aversion and Kantian morality, while the other combines spite and Kantian morality.

The literature on the evolution of preferences of individuals in strategic interactions is vast and diverse. We organize the discussion around the following question: Supposing that material outcomes drive evolutionary success, under what circumstances does evolution promote Homo oeconomicus, defined as material self-interest, and when does it instead lead to other preferences? The literature suggests that Homo oeconomicus is favored by evolution only when individuals' preferences are their private information and the population is large and well-mixed so that individuals with rare mutant preferences almost never get to interact with each other. If rare mutants instead interact more often (say, due to local dispersion), evolution instead favors a certain generalization of Homo oeconomicus including a Kantian concern. If individuals interact under complete information about preferences, evolution destabilizes Homo oeconomicus in virtually all games.

We experimentally investigate behavior and beliefs in a sequential prisoner’s dilemma. Each subject had to choose an action as first-mover and a conditional action as secondmover. All subjects also had to state their beliefs about others’ second-mover choices. Using the elicited beliefs, we use a novel and intuitively appealing hit-rate appraoch to compare the explanatory power of a few current models of social and moral preferences. The data show clear differences in explanatory power between the preference models, both without and with control for the number of free parameters. The best-performing models explain about 80% of observed behavior. We also compare the results with a conventional maximum-likelihood method, and conclude that results by and large agree. We finally use the estimated preference parameters to identify biases in subjects’ expectations. We find a consensus bias (whereby subjects believe others behave like themselves) and a slight optimism (whereby subjects overestimate probabilities for favorable outcomes).

We analyze investment decisions when information is costly, with and without delegation to an agent. We use a rational-inattention model and compare it with a canonical signal-extraction model. We identify three "investment conditions". In "sour" conditions, no information is acquired and no investment made. In "sweet" conditions, investment is made "blindly", i.e. without acquiring costly information. In intermediate, "normal" conditions, the decision-maker acquires information and conditions the investment decision upon the information obtained. We investigate if the investor can benefit from employing an agent when the agent’s effort and information is private. Not even in the case of a risk neutral agent will the principal perfectly align the agent’s incentives with her own at the moment of investment (had the principal known the agent’s private information). Optimal contracts for risk neutral agents not only reward good investments but also punishes bad investments. Such contracts include three components: a fixed salary, stocks and options.

We experimentally investigate behavior and beliefs in a sequential prisoner’s dilemma. Eachsubject had to choose an action as first-mover and a conditional action as second-mover. Allsubjects also had to state their beliefs about others’ second-mover choices. We find thatsubjects’ beliefs about others’ choices are fairly accurate on average. Using the elicited beliefs,we compare the explanatory power of a few current models of social and moral preferences. Thedata show clear differences in explanatory power between the preference models, both withoutand with control for the number of free parameters. The best-performing models explain about80% of observed behavior. We use the estimated preference parameters to identify biases insubjects’ expectations. We find a consensus bias (whereby subjects believe others behave likethemselves) and a certain optimism (whereby subjects overestimate probabilities for favorableoutcomes), the former being about twice as strong as the second.