33
8
Conclusions
This study contributes to the empirical literature by increasing the understanding of the
effects of retirement on health. The analysis makes use of detailed longitudinal data on
inpatient care, mortality and labor market outcomes. The estimation exploits a targeted
retirement offer during 1992-1994 to military employees 55 years of age or older.
Before the offer was made, the normal retirement age for military officers was 60 years
of age. We estimate the effect on individuals‟ health during ages 56-70, after being
offered retirement at age 55. The identification strategy is based on cohort variation in
the timing of the offer, and by using other civil servants not affected by the early
retirement offer, to control for secular trends in schooling, nutrition (i.e., early
childhood difference at the cohort level), health care technology, and general period
effects.
We show that the reform increased early retirement and decreased labor market work
in ages 55-59. Moreover, the reform had only short-run effects on disposable income,
which is very different from the literature on the effects of job loss or unemployment
(e.g., Eliason & Storrie, 2009a, 2009b, 2011; Browning & Heinesen, 2012; Black et al.,
2013).
29
More importantly, we find that the opportunity to retire early had positive
effects on health; it decreased the number of days in inpatient care and also reduced
mortality. Our result is robust to different specifications and alternative inpatient care
outcome measures. The effect on mortality is larger when we censor our data at the age
of 66 instead of at age 71, and we also find that the risk of dying in acute myocardial
infarction is reduced before age 66. This suggests that some of the effect of early
retirement on health is from reduced workplace stress.
When we estimate the effect of retiring for the compliers, we find that the number of
days in inpatient care is reduced by around 8 days over the 14 years under study (i.e.,
from 56 to 70 years of age).
The question is how the results compare with earlier results? There are no other
comparable studies using inpatient care data why there is no possibility to compare the
magnitude of these effects. Based on our point estimates from the Cox regression we
29
This literature makes use of displaced worker and not unemployment per se. The results from US studies find that
the earnings losses of displaced workers are large: up to 25 percent five to six years after the displacement (see
Kletzer 1998). Using Norwegian data Huttunen et al (2011 finds that a 31 percent increase in the exits from the labor
market for the displaced. The long run effect for those remaining in the labor force is modest however. Seven years
after the displacement the income loss is estimated to be 3 percent.
