Consequences of serious parental health events on child mental health and educational outcomes

Abstract

I show that serious, yet common, parental health events in childhood have immediate and lasting effects on mental health and educational outcomes for children. Following a parental health event, the children are more likely to receive therapy and consume anti-depressant (AD) medication. More so, the children achieve lower test scores and have lower school enrollment rates. The effect immediately occurs following the event and persists at least into early adulthood. I find that the effect on test scores doesn't differ significantly across family income, but that children from low-income families are more likely to be prescribed ADs following the event, while children from high-income families are more likely to receive therapy. Exploiting differences in general practitioners' behavior in prescribing AD and referring children to therapy, I find suggestive evidence that children who are more exposed to medical treatment of mental health issues have lower educational attainments in early adulthood.

Dark Passage: Mental Health Consequences of Parental Death

“…We contribute to the literature by analyzing the causal relationship between parental death and multiple measures of the child's mental health. We differ from Kristiansen's (2021) recent analysis of parental death in Denmark that studies therapy and anti-depressant medication as outcomes in four ways. First, we focus on mental health-related hospitalization and estimate separate effects by parent's gender.…”

“…Investigating the relationship between parental death and children's occupational status using historical data (1850 to 1952) from the Netherlands, Rosenbaum-Feldbrügge (2019) finds, for both men and women, larger adverse effects due to maternal loss compared to paternal loss in OLS models. 1 Our paper is most closely related to the recent work by Kristiansen (2021), 2 who examines the effect of parental health shocks, including deaths, on the likelihood of two mental health-related outcomes, therapy and anti-depressant medication. She studies health shocks for children aged 14 to 18 at the time of the shock, using Danish registry data.…”

“…Fourth, we test more formally the underlying assumption of parallel trends, using the recent techniques of Borusyak et al (2021). The event study analysis used here and in Kristiansen (2021) explicitly focuses on how parental death affects mental health over time, in contrast to previous works where the outcome is typically defined at a specific point in time, such as a college entrance exam (Chen, Chen, and Liu, 2009;Gould, Simhon, and Weinberg, 2019).…”

“…Our preferred specification is an event study framework akin to the model used by Kleven, Landais, and Egholt Søgaard (2019) and Kristiansen (2021), in which we follow individuals before and after a parental death at ages 10-20. An extension to the event study model uses children who did not experience a parental death as a comparison group.…”

“…In contrast, we generally cannot reject the possibility of no change in hospitalization rates for males experiencing maternal death or females experiencing paternal death.These main findings are robust across four different model specifications. To understand the impacts, our preferred specification is an event study model based onKleven et al (2019) andKristiansen (2021), where we explicitly model the dynamic effect of parental death over time. As the event study model includes only individuals experiencing a parental death, we also specify a difference-in-differences event study model using individuals without a parental death before age 30 as a comparison group.…”