2nd Conference of the Swiss Society of Health Economics

Objectives:
In many countries, policy makers promote gradual retirement to incite older workers to prolong their working life. Yet little is known about how gradual retirement is associated with older workers’ health. We therefore analyze employment and health trajectories of workers aged 50 to 70 in four countries with gradual retirement policies: The United States (U.S.), Germany, Sweden and Italy. Moreover, we examine how these trajectories vary by type of welfare state and by age group.

Methods:
We use data from two large representative panel surveys, the Health and Retirement Study (HRS) for the U.S. and the Survey of Health, Ageing and Retirement in Europe (SHARE) for Germany, Sweden and Italy. We apply sequence and cluster analysis to identify workers’ most typical employment and health trajectories by type of welfare state and by age group.

Results:
We find that, compared with other welfare regimes, gradual retirement trajectories are associated with trajectories of better health in countries with a social-democratic and a corporatist welfare state such as Sweden or Germany. Moreover, the trajectories are substantially more heterogeneous in the U.S., a country characterized by a liberal welfare state. Similarly, we find more heterogeneity among the oldest age group of our sample.

Discussion:
Our findings suggest that adopting policies that allow gradual retirement and more generous welfare states may have a positive impact on health in old age.

Adolescent mental health is key for later well-being. Yet, causal evidence on environmental drivers of adolescent mental health is scant. We study how an important classroom feature---the gender composition in compulsory-school---affects mental health. We use Swedish administrative data (N=576,285) to link variation in gender composition across classrooms within cohorts to mental health. We find that a higher share of female peers in one's classroom increases the incidence of mental health diagnoses, particularly among boys. The effect persists after students transition to a different high-school classroom. Peer composition is thus an important and persistent driver of mental health.

Objectives:
Early pregnancy losses are the most frequent complications in pregnancy, they occur in about 10-20 percent of recognised pregnancies. An early pregnancy loss often leads to grief, stress, or guilt, and may result in mental health declines for both the woman and her partner.

Methods:
Linking Dutch administrative data on hospital admissions, health insurance expenditure, prescription drugs, and family links from 2009 to 2017, we estimate the effect of an early pregnancy loss on mental health care use. Conditional on being pregnant, early pregnancy losses are random to a large extent. Therefore, we use an event study model that accounts for treatment effect heterogeneity across treatment cohorts.

Results:
We find that early pregnancy losses increase both the probability of using mental health care and psychiatric drugs by 14 percent with respect to the baseline mean; and that mental health care spending increases modestly. Partners are also affected, but to a lesser extent. Preliminary results indicate worsened labour market outcomes and a higher likelihood of divorce for women who remain childless after an early pregnancy loss.

Discussion:
Integrating a systematic mental health care follow-up to the treatment of an early pregnancy loss may give affected women and their partners additional support to deal with the loss.

 Concerns about left-behind children and dangerous working conditions abroad have encouraged some governments to restrict women’s labor migration. I examine how women’s fertility responds to such restrictions in Sri Lanka. The Sri Lankan government introduced a policy in 2013 prohibiting women from migrating for work based on their age and the age of their youngest child. These restrictions could alter fertility decisions, with women simultaneously choosing between future employment and childbearing. Using a panel dataset created from the Demographic and Health Survey in a regression discontinuity in time framework, I find that women from poor households, who are most likely to migrate, change their fertility behavior. Young women, who are already restricted from migrating based on their own age increase their fertility. Older women, who are restricted from migrating only if they have young children, reduce their fertility. As a result, new mothers are less-educated and younger, which may have an impact on child outcomes. My findings contribute to the literature on migration policies in developing countries and trade-offs between women’s employment and fertility decisions.

In low-resource settings, where asymmetric information is widespread, visible status goods may
be exploited as cheap wealth screening devices. Obesity is visible, and in most low- income countries
appears to have the characteristics of a status good. My setting is Kampala, Uganda. I estimate
the causal effect of obesity by randomly assigning body mass using photo-morphed pictures. I
begin by providing quantitative evidence that obesity signals wealth in this context. Then, focusing
on credit markets, I use a field experiment with 124 financial institutions to show that obesity
has monetary benefits because it is exploited as a signal of wealth. Going from normal weight to
obese improves access to credit as much as increasing one’s income by 60 percent. I pinpoint the
mechanism by showing that reducing asymmetric information reduces obesity benefits. In terms
of welfare consequences, a model of credit discrimination suggests that loan officers over-respond
to obesity as compared to an unbiased discrimination benchmark. Further experimental evidence
shows that lay people anticipate and overestimate obesity benefits, leading to demand distortions.
I discuss policy implications including anti-discrimination and obesity prevention policies.