Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Objective
To assess the selective and protective impact of early retirement on life.
Design
Long term prospective cohort study.
Subjects
Results are based on 88,941 men and 41,762 women, all members of Gmünder Ersatzkasse (GEK), a German compulsory health insurance with approximately 1.5 million insured persons, who retired at the age 50-65 between January 1990 and December 2004.
Main outcome measures
Hazard ratio for death adjusted for age, age at retirement, year of observation, sex, socioeconomic status, and form of retirement scheme.
Results
Selection: We find significantly higher mortality risks among pensioners with reduced earning capacities than among old-age pensioners who either exited from the labour market at 56-60 or 61-65 for both sexes (P < 0.001, log rank test). After 7.2 years 2 out of 10 male pensioners with reduced earning capacities who retired between 56 and 60 had died while men of the same age group who receive an old-age pension lived on average 5.8 years longer. This difference is 4 years for pensioners who retire at 61-65. There is also a survival gap between women with different retirement schemes despite their higher general life expectancy. Strikingly, the youngest male and female pensioners who leave the labour market at the age of 51-55 because of their reduced earning capacity face the highest mortality risks. Protection: People who retire early and have no health problems do not experience lower long term survival than people who retire late. For women, early retirement even lowers their mortality risks significantly by 25% (0.75, 59 to 95).
Conclusion
With respect to mortality, early retirement triggers both, selective and protective processes. On the one hand, people in ill-health with lower survival chances are selected out of the labour market. On the other hand, healthy pensioners are protected by a less demanding and less stressful life style during retirement. For the former, early retirement is essential, for the latter it is an asset. Pension reformer should take this difference into consideration when cutting back pension programs and increasing retirement age.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: As the workforce of the United States is aging, some employees are able to modify full-time work status in order to “phase down” their career employment as they approach full retirement, and phased retirement may become increasingly common and important as the workforce continues to age. Our paper explores whether the ability to phase down one’s employment is associated with unequal endowments of human and financial capital. Using a unique definition of phased retirement based on data from a large, longitudinal dataset over a 10-year time period, this paper uses multinomial logistic regression and growth curve analyses to test the association between phased retirement and a variety of human and financial capital characteristics. We find that there are significant personal, household, and occupational differences between phased retirees and those who do not engage in phased retirement. Phased retirees appear to be better educated, have more resources, and occupy superior occupational positions. These results no doubt reflect that ad hoc nature of phased retirement, in which advantaged employees are able to discern and utilize opportunities as they arise.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: There are over 14 million older workers in the United States. There is a pressing need for a better understanding of their employment and the nature and magnitudes of the health risks they face. In this paper we address three research questions: [1] what are the concurrent associations between health and both hours worked and occupational substantive complexity among older workers; [2] are there reciprocal effects over time; and [3] are the associations between work and health different by gender or retirement status? We examined these research questions using data on 3,027 employed older Americans (ages 50 to 70), from the 1996 and 1998 waves of the Health and Retirement Study, a multi-wave, national longitudinal study of older Americans, supported by the National Institute on Aging. We found that this cohort of older workers is heavily involved in the labor force, with 80 percent- of women and 90 percent- of men employed 30 or more hours per week. Overall, these older workers are in good to excellent health, with only 10 percent- rating their health as fair or poor. Yet over half of the women and a third of the men reported some health-related limitations to their activities of daily living. We found few differences between older workers employed after retirement and those who were not yet retired on either substantive complexity or health. However, retired older workers were more likely to be employed part-time than were those who were not retired. We also found that, among older workers, occupations that require complex problem-solving help to maintain cognitive functioning. However, longer hours of employment may have negative health consequences for adults who continue to work post-retirement, and older women's morbidity may have negative consequences for their job choices and related well-being.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: We consider the redistributive effects of social security old age pensions as compared to the tendency for private pensions to exacerbate gender inequality while asset earnings compound racial inequality—effects that we documented in research on the retirement cohort of 1980-1981 (Hogan and Perrucci 1998). Then we suggest how changes in the past two decades might have increased or decreased gender and racial gaps in retirement earnings for younger cohorts. Our analysis of retirees extracted from AHEAD, CODA, HRS, and War Babies birth cohorts indicates some evidence of progress for white women and black men but little change in the general pattern of reproducing inequality in retirement. We find that cohort effects (left and right censoring) are generally more significant than what might at first appear to be temporal effects. Gender progress may be exaggerated by these effects, while racial progress may be under-estimated. In either case, however, there is no evidence of progress for black women--the truly disadvantaged. Furthermore, we find no evidence to suggest that we should reconsider our assertion (Hogan, Perrucci and Wilmoth 2000) that the privatization schemes currently being offered as alternatives to social security old age pensions will increase racial and gender inequality.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: In this study, we document the ways in which the Social Security system creates major disparities within ethnic minority groups. Immigrants and individuals who never participated in the Social Security program or those who immigrated to the United States in middle or late adulthood face particularly serious problems. Our purpose is to focus closely on the interplay of marital history, immigration, and retirement income supports. We address the following questions: (1) Does race/ethnicity and nativity affect changes in the sources of income for elderly women in the years prior to retirement and in the years after retirement? (2) What are the impacts of nativity and immigration on the economic security of pre-retirement age Latinos, in particular? (3) Do sociocultural factors such as age at migration and family structure make a significant difference in predicting Hispanics’ economic well-being defined in terms of personal income and assets, in late life? These analyses extend previous research on Latino economic security by employing several waves of the Health and Retirement Survey (HRS). The HRS was initiated in 1992, and is an excellent data set to examine these research questions since it includes in-depth economic, social, and program participation information on a nationally representative sample of individuals between ages 50 and 61. The large cohort, in addition to purposive over sampling of Latinos, provides ample opportunity to consider the importance of nativity, generally, and the specific effect of age at migration on Latino economic profiles and their implications for social security reform.