This paper argues that gender-egalitarian work norms may translate into lower fertility when they conflict with institutions organized around a male-breadwinner model—a norm–institution mismatch. East Germany provides a setting in which these forces became sharply misaligned: schooling under socialism emphasized women’s continuous employment, while reunification weakened institutional support for that model. Exploiting the GDR’s strict school-entry cutoff, I find that one additional grade under the socialist school regime reduces fertility by about 0.07 children among cohorts forming families after reunification. Consistent with norm transmission, additional exposure increases the non-pecuniary value women attach to work and raises their employment, without corresponding changes in educational attainment, household resources, or stated fertility preferences. Consistent with the mismatch account, the fertility effect is substantially smaller in regions where post-reunification childcare availability remained higher. A within-person vignette experiment points to the same work–family margin: greater exposure increases responsiveness to childcare provision but not to broader economic conditions. Together, the evidence suggests that gender-egalitarian work norms reduce fertility primarily where institutional support for combining work and family is limited.
Journal of Development Economics, Vol. 182, June 2026, 103785
Summary: Parental mental illness shapes children’s economic preferences: in experiments with more than 4,500 children in Bangladesh, children of affected parents are less prosocial but more patient.
This paper provides the first evidence that children's economic preferences vary systematically with parental mental health. Using experimentally elicited measures of economic preferences from more than 4,500 children in Bangladesh, we document that children with parents who are affected by mental illness are less prosocial but more patient than their peers with mentally healthy parents. Attitudes toward risk remain unchanged. We discuss potential pathways through which parental mental health may influence the formation of children's preferences, documenting that children with mentally ill parents assume greater responsibilities within the family, experience less parental involvement, and are exposed to a more adverse home environment.
Summary: Adolescence is a sensitive period for time preferences: instability experienced between ages 13 and 17—but not earlier or later—leads to lasting reductions in patience across 76 countries.
This paper identifies adolescence as a formative period in the development of time preferences. Using data from 80,000 individuals in 76 countries, I exploit within-country, cross-cohort variation in exposure to aggregate instability—natural disasters, inflation crises, and sovereign defaults. Instability experienced between ages 13 and 17 reduces long-run patience, whereas comparable shocks experienced before or after this developmental stage show no detectable effects. Evidence from second-generation immigrants in Germany reveals similar age-specific effects when origin-country instability is transmitted only indirectly through family networks, suggesting that the mechanism operates through psychological rather than material channels.
Preferences as Memory States: Experience Effects at the Moment of Choice
Supported by research grants from Behavioral Finance e.V., the ifo Institute, and the Society for Experimental Economics (GfeW)
Summary: Risk valuations depend not only on the accumulation of experience, but also on which experiences are retrieved at the moment of choice: holding lotteries fixed, cueing different memories shifts valuations. Facial and gaze measures trace how these effects unfold.
Long-run risk preferences are shaped by experience, yet measured preferences also vary
across situations. This paper tests whether some of that variation reflects which past
experience is active at the moment of choice. In an account-history experiment, cueing a
−40% rather than a +40% episode lowers the certainty equivalent of a 50–50 lottery
paying $5 or $0 by 0.32 standard deviations when the history is tied to participants’
own earnings. The contrast is substantially attenuated in the payoff-irrelevant referenceaccount arm. Facial and gaze measures help trace this difference: in both arms, estimates
indicate more positive facial valence after gain than after loss cues and a greater tendency
to orient first toward the lottery’s $0 outcome after loss cues. The largest arm difference in
facial valence appears after participants orient toward $0, with loss cues reducing valence
more in the own-account arm. A two-wave panel extends the test to naturally occurring
memories: a cue to the 2008 financial crisis lowers valuations among respondents who
were young adults during the crisis, but changes little among those who were young
children. Within the older cohort, responses are stronger with greater local crisis severity
and self-rated exposure. The estimate one week later is close to zero.
Many Children, Few Paths: A Panel Study of the Development of Economic Preferences
(with Laura Breitkopf, Shyamal Chowdhury, Hannah Schildberg-Hörisch, and Matthias Sutter)
Summary: Children’s economic preferences develop along a small number of distinct trajectories, making the “average child” a misleading description of preference development—especially for risk-taking.
Using four waves of panel data from incentivized experiments with more than 5,600 children aged 7 to 18 years, we provide the first within-individual evidence on how economic preferences develop in childhood and adolescence. On average, risky choices converge toward risk neutrality, patience decreases, and prosociality increases. Yet, these aggregate patterns conceal a small number of distinct developmental types, which we identify by applying group-based trajectory models. We show that the average child can be a misleading representation of preference development: for risk-taking, most children do not follow the average trajectory, which lies between a decreasing majority and a hump-shaped minority. For patience and prosociality, by contrast, the dominant type follows the average trajectory. Types are largely preference-specific: a child’s path in one preference says little about her path in another. The parent’s own preference is a child’s strongest family correlate, and cognitive ability matters most for the risk-taking paths.
(with Wolfgang Dauth, Jens Südekum, and Ludger Woessmann)
IZA Discussion Paper No. 18247
Summary: Trade and technology shocks reshape how workers adjust: trade exposure reduces training, robot adoption increases it, and both accelerate early retirement, shifting adjustment costs onto public pension systems.
How do firms and workers adjust to trade and technology shocks? We analyze two mechanisms that have received little attention: training that upgrades skills and early retirement that shifts adjustment costs to public pension systems. We combine novel data on training participation and early retirement in German local labor markets with established measures of exposure to trade competition and robot adoption. Results indicate that negative trade shocks reduce training — particularly in manufacturing — while robot exposure increases training — particularly in indirectly affected services. Both shocks raise early retirement among manufacturing workers. Structural change thus induces both productivity-enhancing and productivity-reducing responses, challenging simple narratives of labor market adaptation and highlighting the scope for policy to promote adjustment mechanisms conducive to aggregate productivity.
(with Daniel Kamhöfer and Hannah Schildberg-Hörisch)
IZA Discussion Paper No. 16454
Summary: Education raises life satisfaction among the employed but lowers it among the unemployed, a pattern consistent with an aspiration-augmented utility function in which education raises aspirations about labor-market outcomes.
This paper investigates the causal effect of education on life satisfaction, exploring effect heterogeneity along employment status. We use exogenous variation in compulsory schooling requirements and the build-up of new, academically more demanding schools, shifting educational attainment along the entire distribution of schooling. Leveraging plant closures and longitudinal information, we also address the endogeneity of employment status. We find a positive effect of education on life satisfaction for employed individuals, but a negative one for those without a job. We propose an aspiration-augmented utility function as a unifying explanation for the asymmetric effect of education on life satisfaction.
Adverse Shocks Disrupt the Growth of Patience in Childhood and Adolescence
(with Laura Breitkopf, Shyamal Chowdhury, Hannah Schildberg-Hörisch, and Matthias Sutter)
Summary: Adverse shocks can disrupt the typical rise in patience during childhood and adolescence, flattening—or even reversing—its development across two distinct settings.
Higher levels of patience are, in general, associated with better education, higher wealth, and better health across the life course. In stable, high income societies, patience rises steadily through childhood and adolescence before stabilizing in adulthood, a trajectory widely treated as a universal feature of human development. Here we show that it is not. Using novel, longitudinal data tracking more than 5,600 children aged 7–18 from Bangladesh, a lower-middle-income country frequently struck by natural disasters, we find that average patience declines with age rather than rising. We reconcile these seemingly contradictory patterns by showing that exposure to adverse shocks during key developmental stages significantly reduces patience, driving the overall non-increasing age trend observed in Bangladesh. To assess the generalizability of this finding, we analyze variation in shock exposure among young adults in Germany, following a major flood in 2013. In both contexts, our results reveal that adverse shocks can flatten—or even reverse—the typical age trajectory of patience found in many high-income countries. Our findings underscore the critical role of adverse shocks in shaping the formation of time preferences and challenge the stylized fact that patience universally increases before adulthood.
Work in Progress
Building Mental Health Resilience to Life Shocks: Evidence from a Randomized Socio-Emotional Learning Program
(with Laura Breitkopf, Shyamal Chowdhury, Hannah Schildberg-Hörisch, and Matthias Sutter)
Summary: Socio-emotional skills may provide a buffer against later adversity: evidence from a randomized socio-emotional learning program suggests that they can mitigate the mental-health effects of subsequent life shocks.
