January 2019 NBER Digest

Early Retirement and Mortality Rates of Blue-Collar Men


For blue-collar male workers in Austria, an extra year of early retirement, induced by a policy change, was associated with an increase in the probability of death before age 73 of 1.85 percentage points.


Many workers dream of retiring as early as possible to pursue travel, leisure, sport, and other pursuits. But the research findings in Fatal Attraction? Extended Unemployment Benefits, Labor Force Exits, and Mortality (NBER Working Paper No. 25124), a study by Andreas Kuhn, Stefan Staubli, Jean-Philippe Wuellrich, and Josef Zweimüller, suggest that some individuals, particularly men, might want to postpone retirement if possible. Studying a temporary change in unemployment insurance rules in Austria which allowed workers to retire early, they find that men who retired before the normal retirement age experienced an increased risk of premature death. They do not find any statistically significant effect of early retirement on women.

Many nations are grappling with aging populations and the strains that pension and medical-care obligations place on government budgets. Some are considering changes to retirement programs, such as raising the age of eligibility or reducing benefits. The effect of such changes on the health and well-being of the elderly is a subject of ongoing debate.

To shed light on this question, the researchers analyze a unique public program in Austria in the late 1980s and early 1990s that was adopted when that nation's steel sector underwent dramatic downsizing. The shock affected tens of thousands of workers and their family members, and to cushion the economic blow to older workers, the Austrian government implemented the Regional Extended Benefits Program (REBP). This program effectively allowed workers to take early retirement via disability insurance or old-age pension programs. The program was only available in some regions of the country.

Using information from the Austrian Social Security Database, the researchers were able to compare the employment histories, incomes, gender, age, retirement dates, and age at death of those who took early retirement and those who were eligible but did not. They ultimately compiled information on 310,440 men and 144,532 women — excluding those from the steel sector — and compared data from REBP — eligible regions and nearby non-REBP regions.

The program induced a significant increase in early retirement. When the researchers examined the mortality rates of those who took early retirement, they found that an additional year in early retirement increased a man's probability of death before age 73 by 1.85 percentage points — equivalent to a relative increase of 6.8 percent — and reduced the age at death by 0.2 years. For women, early retirement was not associated with elevated mortality, a finding that is in line with previous research by others.

Men in blue-collar occupations, men with low work experience, and men who had some pre-existing health impairment displayed higher mortality effects than men in white-collar occupations. An additional year in early retirement increased the probability of death before age 73 by 1.91 percentage points for blue-collar men, 3.45 percentage points among men who have spent some time on sick leave, and by 2.42 percentage points among men with low work experience.

To check the robustness of their findings, the researchers analyzed data from before and after the early retirement program and found no differences in mortality and early retirement trends between those two periods. They also found that the changes in lifetime income associated with early retirement were negligible, particularly when generous government old-age benefits were counted, and that they could not explain the increased mortality among certain groups of the population. The researchers suggest that lifestyle changes may explain the study's mortality findings.

— Jay Fitzgerald

The Digest is not copyrighted and may be reproduced freely
with appropriate attribution of source.