34
find a yearly reduction in mortality by around 26 percent, which implies a yearly
reduction by around 0.36 percentage points and a reduction by around 5.4 percentage
points over the 15 years study period.
30
This may seem to be a quite large effect when
comparing with the results from the only known earlier study using Swedish employees
(Hult et al. 2010). They could not find an effect of early retirement on mortality using
data on male builders born 1920 to 1932. However, in a recent study, also using a quasi-
experimental design, Bloemen et al. (2013) found point estimates of the same
magnitude as ours. They found no effects for females but a decrease in mortality due to
an early retirement offer for male civil servants by 42.3 percent over a 5-year period or
a reduced probability to die within five years by 2.5 percentage points.
Our results indicate a larger negative effect on inpatient care days, both for those
with low pre-retirement incomes and for those without college education. Those with
low income before retirement (or with low education level) most likely have worse
work environments on average and/or less good health, in comparison to others, which
means that the improved health is linked to less exposure to workplace hazards, which
would result from the immediate risk reduction of the retirement or from an indirect
effect via a poorer health at the time of the retirement offer.
Most OECD countries are undertaking measures to prolong the careers of older
workers. A natural question concerns the effect that postponing retirement may have on
individual well-being and, in particular, on health. Unfavorable (or favorable) effects
from retirement timing on post-retirement health not only influence individual health
but also directly affect health care costs among retirees. Taking cross-sectional studies
as evidence of negative effects of retirement on health, suggests a “win-win” situation
for extending retirement age in the population. Our findings suggest that increasing
mandatory retirement may not, unfortunately, be a “win-win” situation. Early retirement
is instead shown to have positive health effects. The positive income effect for the
government will most likely also create negative side-effects in terms of increasing
costs for health care.
Sweden has obligatory public sickness and disability insurance schemes that, in an
international context, are generous, both when it comes to the levels of income
30
The estimation is based on the upper left panel o
From this panel we can see that we have have a survival
rate of approxiamtly 0.80. This yields an aproximatly 1.5 percent yearly hazard. A 26 percent decrease imply an
yearly reduction by 0.36 (= 1.5*(1-0.76)) percentage points and hence a total of 5.4 percentage points over the 15
years period.
