joint with Leah Boustan*, Mathias Fjællegaard Jensen*, Ran Abramitzky*, Elisa Jácome*, Alan Manning*, Santiago Pérez*, Analysia Watley*, Adrian Adermon, Jaime Arellano-Bover, Olof Åslund, Marie Connolly, Nathan Deutscher, Anne C. Gielen, Yvonne Giesing, Yajna Govind, Dominik Hangartner, Yuyan Jiang, Cecilia Karmel, Fanny Landaud, Lindsey Macmillan, Isabel Z. Martínez, Alberto Polo, Panu Poutvaara, Hillel Rapoport, Sara Roman, Kjell G. Salvanes, Shmuel San, Michael Siegenthaler, Louis Sirugue, Javier Soria Espín, Jan Stuhler, Gianluca L. Violante, Dinand Webbink, Andrea Weber, Jonathan Zhang, Angela Zheng, and Tom Zohar. (* indicate lead authors)
We estimate intergenerational mobility of immigrants and their children in fifteen receiving countries. We document large income gaps for first-generation immigrants that diminish in the second generation. Around half of the second-generation gap can be explained by differences in parental income, with the remainder due to differential rates of absolute mobility. The daughters of immigrants enjoy higher absolute mobility than daughters of locals in most destinations, while immigrant sons primarily enjoy this advantage in countries with long histories of immigration. Cross-country differences in absolute mobility are not driven by parental country-of-origin, but instead by destination labor markets and immigration policy.
Paper download: link [IZA Discussion Paper][NBER WP]
- Status: R&R at AER
- Media: DiePresse
- BlueSky: link
Suicide is a leading cause of death worldwide and a critical public health concern. We examine the hypothesis of suicide contagion within in the workplace, investigating whether exposure to a coworker's suicide increases an individual's suicide risk. Using high-quality administrative data from Austria and an event study approach, we compare approximately 150,000 workers exposed to a coworker's suicide with a matched group exposed to a "placebo suicide". We find a significant increase in suicide risk for exposed individuals, with a cumulative treatment effect of 0.04 percentage points (33.3 percent) over a 20-year post-event period. Exposed individuals who also die by suicide are more likely to use the same method as their deceased coworker, strongly suggesting a causal link. Two placebo tests bolster this interpretation: workers who left the firm before the suicide and those exposed to a coworker's fatal car accident do not show an elevated suicide risk.
Paper download: latest version: link (last update August 2025) [IZA Discussion Paper] [slides]
- Status: submitted to a journal
- Media:
- X/Twitter: link
How does military occupation affect long-term economic development? We use the post-World War II occupation of Austria as a laboratory setting. Austria was divided into different occupation zones for ten years. The Soviet occupation was exploitative, while the Western Allied occupation was more supportive. After ten years of different occupation regimes, the regions returned to a single nation-state. We estimate the effect of different occupation regimes on long-term economic development. Methodologically, we combine a spatial regression discontinuity design with a difference-in-differences approach. We find that areas in the former Soviet zone are still less economically developed today. These areas are less populated, host fewer and lower paying jobs, and their residents are more likely to commute outside the former Soviet zone. The most plausible mechanism for these long-lasting effects are agglomeration effects triggered by a large migration shock from East to West as the population fled the advancing Soviet army.
Paper download: link (last update August 2025).
Note, this is a completely revised version of the IZA Discussion Paper No. 10095 (link) published originally in July 2016.
- Status: Submitted to a journal
- Media: [OÖN]
Using Austrian and Danish administrative data, we examine the impacts of parenthood on mental health. Parenthood imposes a greater mental health burden on mothers than on fathers. It creates a long-run gender gap in antidepressant prescriptions of about 93.2% (Austria) and 63.2% (Denmark). These parenthood penalties in mental health are unlikely to reflect differential help-seeking or postpartum depression. Instead, they are related to mothers' higher investments in childcare: Mothers who take extended maternity leave in quasi-experimental settings are more likely to face mental health problems.
Paper download: link , IZA-Version: link
- Status: R&R at JEEA
- Media: [DiePresse][DerStandard][ORF.at][Weekendavisen] [Schrödingers Katze] [IZA World of Labor, Opinions] [religion.orf.at] [Ö1-radio]
We examine how the number of beds available in a maternity ward affects the likelihood of cesarean delivery and maternal health. Our analysis is based on administrative data from Austria. We exploit idiosyncratic daily variation in the occupancy of maternity hospital beds. We find that empty beds increase the probability of cesarean delivery, hospitalization, and readmission. A one standard deviation decrease in maternity bed occupancy increases the probability of cesarean delivery by 4.0% and subsequent hospitalization by 5.8%. Expectant mothers may benefit from a crowded hospital, even at unfavorable patient-staff ratios, because it may lead to less harmful overtreatment..
Paper download: link, IZA-version: link
- Status: revising
- Slides: link
Superstition is a widespread phenomenon. We empirically examine its impact on health-related behavior and health outcomes. We study the case of the Taiwanese Ghost Month. During this period, which is believed to increase the likelihood of bad outcomes, we observe substantial adaptions in health-related behavior. Our identification exploits idiosyncratic variation in the timing of the Ghost Month across Gregorian calendar years. Using high-quality administrative data, we document for the period of the Ghost Months reductions in mortality, hospital admissions, and births. While the effect on mortality is a quantum effect, the latter two effects reflect changes in the timing of events. These findings suggest potential benefits of including emotional and cultural factors in public health policy.
Paper download: link (updated in October 2021)
- Slides: link
The East-West gap in the German population is believed to originate from migrants escaping the socialist regime in the German Democratic Republic (GDR). We use newly collected regional data and the combination of a regression discontinuity design in space with a difference-in-differences approach to document that the largest part of this gap is due to a massive internal migration wave 3 years prior to the establishment of the GDR. The timing and spatial pattern of this migration movement suggest that the dominant motive was escaping physical assault by the Soviet army and not avoiding the socialist regime. The skill composition of these migrants shows a strong positive selection. The gap in population has remained remarkably sharp in space and is growing. These patterns are equally evident when including all counties in the estimation sample.
Paper download: link (updated in May 2025)
- Status: R&R at JOPE
To test for ethnic discrimination in access to outpatient health care services, we carry out an email-correspondence study in Germany. We approach 3,224 physician offices in the 79 largest cities in Germany with fictitious appointment requests and randomized patients' characteristics. We find that patients' ethnicity, as signaled by distinct Turkish versus German names, does not affect whether they receive an appointment or wait time. In contrast, patients with private insurance are 31 percent more likely to receive an appointment. Holding a private insurance also increases the likelihood of receiving a response and reduces the wait time. This suggests that physicians use leeway to prioritize privately insured patients to enhance their earnings, but they do not discriminate persons of Turkish origin based on taste. Still, their behavior creates means-based barriers for economically disadvantaged groups.
Paper download: link (February 2022)
- Status: submitted to a journal
- Non-technical summary (in German): link
- Media: aerzteblatt.de
- Tweet: link
Social insurance programs typically comprise sick-leave insurance. An important policy parameter is how the costs of lost productivity due to sick leave are shared between workers, firms, and the social security system. We show that this sharing rule affects not only absence behavior but also workers' subsequent health. To inform our empirical analysis, we propose a model in which workers' absence decisions are conditional on the sharing rule, health, and a dismissal probability. Our empirical analysis is based on high-quality administrative data sources from Austria. Identification is based on idiosyncratic variation in the sharing rule caused by different policy reforms and sharp discontinuities at certain job tenure levels and firm sizes. An increase in either the workers' or the firms' cost share, both at public expense, decreases the number of sick-leave days. Policy-induced variation in sick leave has a significant effect on subsequent healthcare costs. The average worker in our sample is in the domain of presenteeism, that is, an increase in sick leave due to reductions in workers' or firms' cost share would reduce healthcare costs and the incidence of workplace accidents.
Paper download: link (updated in January 2017)
Little is known about the response behavior of parents whose children are exposed to an early-life shock. We interpret the prenatal exposure of the Austrian 1986 cohort to radioactive fallout from the Chernobyl accident as a negative human capital shock. We can rely on exogenous variation in the exposure to radioactive fallout (over time and) between communities due to differences in precipitation at the time of the accident. Our approach provides robust empirical evidence for compensating investment behavior. Families with low socioeconomic status reduce their family size, while families with higher status respond with reduced maternal labor force participation.
Downloads: paper, web-appendix (updated in November 2014)
media: IZA Newsroom
We examine the effect of the introduction of joint custody after divorce on marriage, divorce, fertility, and female employment. Our analysis is based on Austrian individual-level administrative data, which cover the entire population. We em-ploy a difference-in-differences estimation approach and use older couples with out children, and couples from neighboring Germany, as a control group. Our estima-tion results suggest that the reform has significantly increased marriage and marital births rates and has reduced divorce rates and the likelihood of female employment.Our findings are robust to several sensitivity checks and confirm most of the results previously found for the US.
Paper download: in February 2016: [paper]
(A previous version of this paper was circulated under the title, "Does Custody Law Affect Family Behavior In and Out of Marriage?": old paper version)Divorce and the Excess Burden of Lawyers, July 2007:
[paper] lost in a R&R
media: The Financial Times Undercover Economist
Bargaining at Divorce: The Allocation of Custody with Christine Hölzl December 2007:
[paper] lost in a R&R