TY - JOUR
T1 - The causal effect of early retirement on medication use across sex and occupation
T2 - evidence from Danish administrative data
AU - Cremers, Jolien
AU - Nielsen, Torben Heien
AU - Ekstrøm, Claus Thorn
N1 - © 2024. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.
PY - 2024
Y1 - 2024
N2 - We examine the causal effect of early retirement on medication use using Danish registry data. A reform in early retirement schemes in 2006 gradually increased eligibility ages from 60 to 64 differentially across birth cohorts. This enables an instrumental variable design that was applied using novel g-estimation methods that alleviate bias in binary outcome IV models. Our data allow studying patterns in the short run (ages 59½-60½) and in the long run (ages 57-63). For those who were eligible already at age 60, retirement did not change overall medication use. However, when investigating medication and population subgroups, we see that painkiller use decreases and hypertension medication as well as mental health medication use increase after retirement in almost all population subgroups. Moreover, males as well as the blue-collar occupation subgroups do show decreases in overall medication use after early retirement. In conclusion, our analyses reveal that retirement can have important heterogeneous health effects across population groups and are potentially informative about the welfare benefits of social insurance more broadly.
AB - We examine the causal effect of early retirement on medication use using Danish registry data. A reform in early retirement schemes in 2006 gradually increased eligibility ages from 60 to 64 differentially across birth cohorts. This enables an instrumental variable design that was applied using novel g-estimation methods that alleviate bias in binary outcome IV models. Our data allow studying patterns in the short run (ages 59½-60½) and in the long run (ages 57-63). For those who were eligible already at age 60, retirement did not change overall medication use. However, when investigating medication and population subgroups, we see that painkiller use decreases and hypertension medication as well as mental health medication use increase after retirement in almost all population subgroups. Moreover, males as well as the blue-collar occupation subgroups do show decreases in overall medication use after early retirement. In conclusion, our analyses reveal that retirement can have important heterogeneous health effects across population groups and are potentially informative about the welfare benefits of social insurance more broadly.
U2 - 10.1007/s10198-023-01660-0
DO - 10.1007/s10198-023-01660-0
M3 - Journal article
C2 - 38472724
VL - 25
SP - 1517
EP - 1527
JO - European Journal of Health Economics
JF - European Journal of Health Economics
SN - 1618-7598
ER -