Determinants of Health

Einleitung / Introduction

The large inflow of refugees in 2015 was associated with a marked increase in hate crimes directed against this group. These hate crimes potentially impair the refugees in the long-term. The reason is that impaired mental health is associated with worse decision making. Therefore, this can affect the refugees’ accumulation of country-specific human capital. Against this background, we estimate the effect of hate crime on refugees’ mental health in Germany.

Methode / Method

For this purpose, we combine two innovative datasets: administrative records on xenophobic crime against refugee shelters by the Federal Criminal Office and the IAB-BAMF-SOEP Survey of Refugees. The data on hate crimes against refugee shelters stem from small inquiries of the party “Die Linke” in the German parliament, which is published on a quarterly basis. We digitized and geo-referenced this data, allowing us to assign each hate crime to the county in which it happened. The IAB-BAMF-SOEP Survey of Refugees is a unique representative and longitudinal survey of refugees that entered Germany in 2015. It comprises rich (predetermined) information about the refugees and their mental health. Our main mental health outcomes are the Mental Component Summary (MCS) and Patient-Health-Questionnaire 4 (PHQ 4) score. Clearly, a simple regression of the mental health outcome on an indicator for the occurrence of a hate crime would be misleading. The reason is that unobserved confounders, such as differential characteristics of refugees in regions where a hate crime occurs and where none occurs, may bias the coefficient of interest. Therefore, we apply a regression discontinuity in time design to consistently estimate the effect of interest. That is, we compare refugees just before and after a hate crime in their respective county of residence happened.

Ergebnisse / Results

Our results indicate that hate crime has a substantial negative effect on several mental health indicators. The effect sizes correspond to about one third of a standard deviation. The effects are stronger for refugees with closer geographic proximity to the focal hate crime and refugees with low country-specific human capital. While the estimated effect is only transitory, we argue that negative mental health shocks during the critical period after arrival have important long-term consequences.

AutorInnen

Daniel Graeber, DIW Berlin

Felicitas Schikora, Konrad Adenauer Stiftung e.V.

Einleitung / Introduction

J1 is a preventative examination recommended for children at the age of 13. In contrast to the well-established U1-U9 examinations for younger children, with participation rates over 90%, the attendance rate in J1 is about 40%. One of the most frequent reasons for not attending J1 is that parents and their teenage children are not aware of this examination. Several German federal states introduced regional policies motivating adolescents to participate in J1. One of such policies, “Ticket to J1”, was introduced in Bavaria in mid-2017. It is an information leaflet about the J1 examination in an appearance that appeals to teenagers and aims to reduce potential hesitations against the examination. The aims of the present analysis are to: a) investigate if the regional initiative was effective at increasing attendance in J1 and if the effects vary by family socioeconomic status (SES); and b) explore which meso-level characteristics of the healthcare system correlate to attendance rates in J1.

Methode / Method

We use anonymised routine data of Techniker Krankenkasse in the timeframe of 2016-2018. To investigate the effect of the policy, a difference-in-difference design on an individual level is used. Assuming a parallel trend on the level of federal states, the likelihood of attendance in J1 of 13-year olds is compared between Bavaria and other federal states pre and post policy introduction. We take into account regional historical differences in attendance rates in J1 and other federal states where similar policies were introduced in the past. To investigate differences in policy effects by SES, the analysis is additionally conducted for different SES strata

Ergebnisse / Results

Preliminary findings show that “Ticket to J1” had a positive significant effect on the participation in J1 for 13-year olds. Furthermore, the effect was stronger for children from families with lower SES. Meso-level characteristics of the healthcare system seem to have a significant effect on participation rates in J1. The number of pediatricians at the community level was positively correlated with the participation rates in J1.

Zusammenfassung / Conclusion

Preliminary results show that health policy has a significant impact on the utilization of J1. Further efforts should be taken to inform adolescents about J1, as it seems to be the unawareness that prevents adolescents to attend it. Furthermore, the policy has a potential to reduce socioeconomic inequality in health care utilization, as the effect of the policy was stronger for children from families with lower SES. More research should be done to investigate if the way such policies are introduced affects its effectiveness, such as using multiple channels of communication of a new policy, e.g. school/pediatrician, as well as multiple targets of the policy message, e.g. children and parents.

AutorInnen

Iryna Iashchenko, Technical University Munich

Leonie Sundmacher, Technical University Munich

Einleitung / Introduction

Loneliness is a subjective and complex feeling that includes perceived social, and - simultaneously or independently - a feeling of psychological or emotional isolation (Weiss 1973). It is strongly correlated with a higher risk of developing mental conditions (depression, dementia, etc.) and a deterioration in physical health - less active lifestyles, diabetes, stroke, coronary heart disease, etc. (Fees, Martin et al. 1999; Hawkley, Burleson et al. 2003; Cantarero-Prieto, Pascual-Sáez et al. 2018; Christiansen, Larsen et al. 2016). Loneliness and social isolation are also linked to all causes of mortality (Luo, Hawkley et al. 2012; Steptoe, Shankar et al. 2013; Holt-Lunstad, Smith et al. 2015; Rico-Uribe, Caballero et al. 2018) and have an impact on health care utilisation (Gerst-Emerson and Jayawardhana 2015). As a result, loneliness can be considered a major public health problem on its own and is of growing interest to policy makers. This is particularly relevant for European countries, where, with an ageing society, many older adults face a higher risk of loneliness. While a correlation of loneliness and retirement is well established, much less is known about the causal effect. The objective of this paper is therefore to study the impact of retirement on loneliness, and to differentiate between short and longer-run causal effects.

Methode / Method

We investigate the short and long-term causal impact of retirement on loneliness among Europeans using a fuzzy regression discontinuity design and data from multiple waves of the SHARE database. To resolve endogenous retirement behaviour, we use the differences in statutory pension age between countries and identify a causal effect of retirement. The panel structure of the dataset with information from at least three waves of SHARE allows us to differentiate between short and long-term effects.

Ergebnisse / Results

Our preliminary results indicate that, in the short term, retirement leads to an increase in loneliness. However, over the long run, the results reverse, and we find retirement to decrease loneliness. Intuitively, these results may be explained by a sudden absence of work-related contacts right after retirement, and an overall different lifestyle people need time to adapt to. Once they do, these negative effects not only subside, but reverse. The short-term effect seems to be particularly large for women and two-person households, while there seem to be no differences in education or between rural and urban households.

Zusammenfassung / Conclusion

To get a clearer understanding of what drives these effects, more thorough heterogeneity analyses (e.g. concerning personality types, occupation, marriage status, grandchildren, etc.) will be conducted alongside an investigation of how retirement affects related outcomes such as social networks and social participation of elderly Europeans.

AutorInnen

Sophie Guthmuller, Wirtschaftsuniversität Wien

Johannes Hollenbach, RWI – Leibniz-Institut für Wirtschaftsforschung

Anna Werbeck, RWI – Leibniz-Institut für Wirtschaftsforschung

Dörte Heger, RWI – Leibniz-Institut für Wirtschaftsforschung

Einleitung / Introduction

In past research, we have documented surprising regional differences in spousal lung function, a continuous measure of health, across European regions. In this project, we take a broader look at whether and why associations in spousal health across different regions in Europe are different. In particular, we want to gain a better understanding of the role of assortative matching on pre-marriage characteristics, i.e., childhood health, childhood socio-economic status, and education.

Methode / Method

For our empirical analysis, we make use of the Survey of Health and Retirement in Europe (SHARE). We concentrate our analysis on a random sample of older couples from 11 European countries. The first step in our analysis is documenting patterns of associations in different health measures across regions in Europe. As outcomes, we choose two measures of physical health that are frequently used in the literature (major and minor conditions) as well as one measure of mental health (depression). We then investigate assortative mating based on early-life factors and their role for health associations later in life. We estimate a static matching model with transferable utility and multiple continuous matching attributes. This matching model provides us with structural estimates of matching preferences, thus allowing us to investigate which and to what extent pre-marital factors matter for assortative matching in the three European regions. With these structural estimates at hand, we moreover conduct a counterfactual exercise to predict how similar spousal adult health would have been in the South when sorting patterns had been those of the North and vice versa.

Ergebnisse / Results

We document that correlations between couples’ health is strong and exhibits strong regional differences. Health associations among couples are strongest in the South, followed by Central Europe, and they are lowest in Northern Europe. There is strong assortative mating on early-life factors, and it follows a similar regional pattern as associations in spousal health. By linking our matching estimates to spousal health correlations, we find that matching on early-life factors explains only little of the variation in late-life spousal health associations.

Zusammenfassung / Conclusion

This is in line with research showing that matching on early-life factors matters more for health correlations in the early stage of marriage. We conduct a counterfactual analysis by imposing matching preferences from one region on couples in another region, showing that regional differences in spousal associations are indeed at least in part caused by differences in matching on early-life factors. Our findings open up new possibilities for future research combining insights from health economics with the matching literature.

AutorInnen

Bettina Siflinger