We present a model where individuals have diverging policy expectations, which has two important effects on trust. First, it drives a wedge between what citizens expect the government to deliver and the rules that are eventually put in place, which fuels institutional distrust. Second, it drives a wedge between what citizens deem good behaviour and what the rest of society actually does, which feeds social distrust. Both effects are determined by citizens' perceptions of the negative externalities caused by their and other people's actions. The more they are concerned about these issues, the more they will distrust others that are less supportive of regulation limiting the negative effects. The less they care, the more they will distrust the government for setting rules they believe are intrusive. Our empirical analysis suggests that these trust dynamics indeed come to the surface during an externality crisis. In an online survey experiment conducted in four European countries we find that the first wave of the Covid-19 pandemic eroded institutional trust among citizens who were not worried about the virus, whereas social trust declined (more) for concerned individuals. We lastly find that support for the welfare state erodes alongside sliding trust levels.

Experience of adverse events in the past is increasingly understood to matter for future behavior in ways that go beyond informational updating. This paper shows that such experience effects differ depending on whether the adverse event resulted from social or natural causes. We develop a novel experimental design that allows the experimenter to manipulate exclusively the source of risk that creates a subject’s experiential biography and to observe repeat choices at the individual level, all within the same stochastic environment. In this setting, we investigate whether the source, natural or social, that caused an adverse event in a past interaction makes a difference to how decisions about a future interaction are made. We generate our evidence by drawing upon a large sample of the US population (n=4,990) who participate in an online experiment. The evidence shows that experiencing an adverse outcome caused by another human (social risk) reduces future risk-taking, but experiencing the same outcome caused by nature (natural risk) does not, provided exposure to the social risk was intentional. The evidence is consistent with the hypothesis that adverse experiences with a social cause trigger amplified regret, driving subjects towards avoiding social risk in future decisions.

We study the developmental roots of cooperation in 929 young children, aged 3 to

6. In a unified experimental framework, we examine pre-registered hypotheses about which of

three fundamental pillars of human cooperation – direct and indirect reciprocity, and third-party

punishment – emerges earliest as a means to increase cooperation in a repeated prisoner’s

dilemma game. We find that already children as young as age 3 act in a conditionally

cooperative way, thus reciprocating others’ behavior. Yet, in the aggregate, direct and indirect

reciprocity do not increase overall cooperation rates beyond the levels in a control condition.

Compared to the latter, third-party punishment more than doubles cooperation rates, making it

the most effective and earliest developmental mechanism to promote cooperation. We also find

that children’s cognitive skills and parents’ socioeconomic background influence cooperation

levels, yet there are important interaction effects with the three pillars of human cooperation.

This study investigates the causal effect of financial literacy on mortgage payment delinquency. Employing the National Financial Capability Study (NFCS) survey data and the Instrumental-Variable (IV) approach, we find that increased financial literacy significantly reduces mortgage delinquency likelihood. This effect cannot be explained by formal education, income, and many other individual characteristics. Besides, our study also examines the heterogeneity of the impact across various demographic groups. We find that the effect of financial literacy on delinquency likelihood is negative and significantly different from zero for any age, gender, income, or education level. Yet, the magnitude of the effect decreases with age and is higher in states where financial literacy of the population is low, as compared to high-literate states.