Simon Columbus

Herrmann et al. study widespread antisocial punishment in public goods games across societies. Punishment behavior differed strongly across societies. Antisocial punishment might be motivated by revenge. Cooperation stabilized in all pools, but contributions differed across pools; either remaining stable or increasing. Average cooperation level was negatively associated with antisocial punishment; net earnings differed widely. Reactions to punishment differed; received punishment mostly led to increased contribution in t+1, but negatively associated with antisocial punishment. Without punishment, cooperation decreased across all pools. Punishment did not increase cooperation in all pools; the higher antisocial punishment was, the less punishment increased cooperation. Punishment might be related to social norms of cooperation. Prosocial punishment is positively related to norms of civic cooperation. Antisocial punishment occurs where civic cooperation norms and rule of law are weak. - Simon Columbus

Herrmann et al. study widespread antisocial punishment in public goods games across societies. Punishment behavior differed strongly across societies. Antisocial punishment might be motivated by revenge. Cooperation stabilized in all pools, but contributions differed across pools; either remaining stable or increasing. Average cooperation level was negatively associated with antisocial punishment; net earnings differed widely. Reactions to punishment differed; received punishment mostly led to increased contribution in t+1, but negatively associated with antisocial punishment. Without punishment, cooperation decreased across all pools. Punishment did not increase cooperation in all pools; the higher antisocial punishment was, the less punishment increased cooperation. Punishment might be related to social norms of cooperation. Prosocial punishment is positively related to norms of civic cooperation. Antisocial punishment occurs where civic cooperation norms and rule of law are weak. - Simon Columbus

Simon Gächter et al. investigate long-run social goods games with and without costly punishment in small, constant groups. They find that cooperation is higher with punishment option both in the long and in the short run. However, average net earnings are higher without punishment in the short run; this is reversed in the long run. The prospect of many iterations already increases cooperation in early periods. The authors conclude that "punishment not only increases cooperation, it also makes groups and individuals better off in the long run". This finding supports evolutionary models in which costly punishment sustains costly cooperation. - Simon Columbus

Urs Fischbacher and Simon Gächter find that cooperation declines in repeated public goods experiments. They argue that this is not due to learning the free-rider strategy. Rather, they argue that most people are imperfect conditional cooperators who match others' contributions only partly; this will cause a decline in contributions even if free riders are absent. In fact, most people are conditional cooperators, others are free riders or "triangular cooperators". A large share (10%) was found to be confused, i.e. not following a clear pattern. Actual contributions are determined by people's preferences and their beliefs about others. Beliefs are adjusted in the direction of previous contributions of others, i.e. "beliefs in a given period are a weighted average of what others contributed in the previous period and one's own belief in the previous period." Belief formation does not change over time. Preference heterogeneity is mostly irrelevant for the decline of cooperation. - Simon Columbus

Simon Gächter et al. investigate long-run social goods games with and without costly punishment in small, constant groups. They find that cooperation is higher with punishment option both in the long and in the short run. However, average net earnings are higher without punishment in the short run; this is reversed in the long run. The prospect of many iterations already increases cooperation in early periods. The authors conclude that "punishment not only increases cooperation, it also makes groups and individuals better off in the long run". This finding supports evolutionary models in which costly punishment sustains costly cooperation. - Simon Columbus

Urs Fischbacher and Simon Gächter find that cooperation declines in repeated public goods experiments. They argue that this is not due to learning the free-rider strategy. Rather, they argue that most people are imperfect conditional cooperators who match others' contributions only partly; this will cause a decline in contributions even if free riders are absent. In fact, most people are conditional cooperators, others are free riders or "triangular cooperators". A large share (10%) was found to be confused, i.e. not following a clear pattern. Actual contributions are determined by people's preferences and their beliefs about others. Beliefs are adjusted in the direction of previous contributions of others, i.e. "beliefs in a given period are a weighted average of what others contributed in the previous period and one's own belief in the previous period." Belief formation does not change over time. Preference heterogeneity is mostly irrelevant for the decline of cooperation. - Simon Columbus

Forsythe et al. study fairness in dictator and ultimatum games. They find that offers in both games differ significantly from the expected Nash equilibria (0 and 0.01 resp.). Offer distributions in dictator games with pay are bimodal and peak at the Nash equilibrium and at an equal share; distributions in ultimatum games with pay peak at an equal share. The study rejects the fairness hypothesis according to which offers can be explained by proposers' concerns with fairness, as distribution between both games differ. The authors propose instead that there are different types of players; some are pure gamesmen, whereas others are concerned (to varying degrees) with fairness. The study also tests a pay hypothesis according to which payment makes a difference; this is rejected for the dictator game, but there is some evidence in favor for the ultimatum game. - Simon Columbus

Forsythe et al. study fairness in dictator and ultimatum games. They find that offers in both games differ significantly from the expected Nash equilibria (0 and 0.01 resp.). Offer distributions in dictator games with pay are bimodal and peak at the Nash equilibrium and at an equal share; distributions in ultimatum games with pay peak at an equal share. The study rejects the fairness hypothesis according to which offers can be explained by proposers' concerns with fairness, as distribution between both games differ. The authors propose instead that there are different types of players; some are pure gamesmen, whereas others are concerned (to varying degrees) with fairness. The study also tests a pay hypothesis according to which payment makes a difference; this is rejected for the dictator game, but there is some evidence in favor for the ultimatum game. - Simon Columbus

Paul J. Zak and Ahlam Fakhar study whether the endocrinological basis for trust between humans scales up to the country level. They find that two classes of factors are related to trust. Consumption of plant-based estrogens, which approximates oxytocin levels, is positively correlated with generalized trust. Environmental conditions that include measures of estrogen-like molecules are negatively associated with trust; these presumable operate through measures of poverty not captured by income as well as pollution with estrogen antagonists. These findings hold after controlling for income. Hence, nutrition and pollution could have an impact on nation-wide interpersonal trust, either because oxytocin directly raises trust, or because it stimulates trust-building interactions. - Simon Columbus

Paul J. Zak and Ahlam Fakhar study whether the endocrinological basis for trust between humans scales up to the country level. They find that two classes of factors are related to trust. Consumption of plant-based estrogens, which approximates oxytocin levels, is positively correlated with generalized trust. Environmental conditions that include measures of estrogen-like molecules are negatively associated with trust; these presumable operate through measures of poverty not captured by income as well as pollution with estrogen antagonists. These findings hold after controlling for income. Hence, nutrition and pollution could have an impact on nation-wide interpersonal trust, either because oxytocin directly raises trust, or because it stimulates trust-building interactions. - Simon Columbus

Two games are used: an ultimatum game with punishment option, and a dictator game as control condition. Under the punishment condition, the fairness norm is not complied with; but not under the control condition. There is a negative correlation between Machiavelli test score and transfers in the control condition, and a positive correlation with behavioral changes caused by the social punishment threat; i.e. Machiavellian subjects earned the highest incomes. Brain activity under the punishment condition was higher in brain areas implicated in the evaluation of punishing stimuli and in processing information about positive and negative reinforcers. If player B was nonhuman, activation was lower. Higher activity in the activated areas was correlated with increased sharing under the social punishment threat. Machiavellism was positively correlated with activity in the left anterior OFC, as well as the right insula; perhaps indicating stronger negative affect from the punishment threat. - Simon Columbus