Health problems as determinants of retirement: Are self-rated measures endogenous?
Introduction
Poor health can affect preferences for work, wage opportunities, and the time horizon workers face, and may therefore play an important role in the decision to retire. Previous research on this topic suggests that health problems do influence older workers' labor force participation, but the magnitude of their effect has yet to be precisely determined. The few measures available in most prior datasets tend to be subjective proxies measured with error and endogenous to labor supply choices. This paper uses the Health and Retirement Study (HRS), a new nationally representative survey of people age 51–61 in 1992 and their spouses. The data set offers a rich variety of economic and personal information including two distinct types of health measures which are the central focus of the analysis. We expect to minimize measurement error in health by using good objective measures of work capacity and exogenous health determinants as instruments.
In what follows we first provide background and offer a conceptual model, and then describe our empirical specification. The data set used and findings for measurement error models are reported, followed by an examination of the effects of specific conditions. Section 5 concludes.
Section snippets
The framework
Identifying the effects of health and economic variables on retirement is complicated because people's health status is not directly observable. Surveys tend to record respondents' subjective self-assessments of their physical capacity for work, and researchers have used these to capture the effect of health capacity on work potential. However, estimated health effects using these subjective measures may be mis-estimated if individuals use health as a justification for leaving the labor force
Alternative empirical formulations
In order to ascertain precisely how health problems influence retirement, an empirical model must be formulated. Econometric issues arise because a worker's `true' health status is not known precisely; that is, the health measures observed in surveys may be incomplete or even biased indicators. For instance, if the justification hypothesis is accurate, people wanting to retire may exaggerate their bad health conditions. As a consequence, health effects will be overstated, and self-rated poor
Data and empirical results
The empirical analysis uses the first wave of responses to the Health and Retirement Study (HRS), focusing on men age 51 to 61 in 1992. Because most of the HRS sample had not yet retired by 1992, the dependent variable we used is the worker's expected age of retirement.
Discussion
We have used the Health and Retirement Study to evaluate the impact of health problems on men's retirement plans. Though estimated health effects vary with measures used, health problems always have a potent influence on encouraging earlier retirement. Specifically, people self-reporting poor health, or indicating they suffer functional limits, planned retirement between one and two years earlier than average. Further, we find that self-rated health measures are not endogenously determined with
Acknowledgements
The paper was written while Dwyer was a Research Associate at the Center for Policy Research at Syracuse University. Mitchell is the International Foundation of Employee Benefits Plans Professor of Insurance and Risk Management at the Wharton School. Research support was provided by the National Institute on Aging (grant # 1 R03 Ag14220-01) and Cornell University, the Wharton School, and Syracuse University's Maxwell School. The authors are grateful for comments from John Abowd, Larry Blume,
References (28)
- et al.
Economic determinants and consequences of self-reported work disability
J. Health Econ.
(1984) - et al.
Market work, wages, and men's health
J. Health Econ.
(1994) - et al.
Hierarchical relationship between activities of daily living and instrumental activities
J. Chronic Disabil.
(1987) - et al.
The retirement-health nexus: a new measure of an old puzzle
J. Hum. Resour.
(1985) The early retirement decision: new empirical evidence on the influence of health
J. Hum. Resour.
(1985)- Bernheim, B.D., 1987, The timing of retirement: a comparison of expectations and realizations, in: Wise, D. (Ed.), The...
Self-reported versus objective measures of health in retirement models
J. Hum. Resour.
(1991)The pension acceptance decision of older workers
J. Hum. Resour.
(1979)- Dominitz, J., 1996. A comparison of subjective expectations elicitation methods in the Health and Retirement Study, the...
Self-ratings of health among the old and old–old
J. Health Soc. Behav.
(1980)
The economics of retirement behavior
J. Labor Econ.
On the concept of health capital and the demand for health
J. Political Econ.
Specification tests in econometrics
Econometrica
Cited by (360)
Stress and retirement
2024, Economic ModellingAnalytic approach for models of optimal retirement with disability risk
2023, Mathematical Social SciencesNo “honeymoon phase”: whose health benefits from retirement and when
2022, Economics and Human BiologyRoutes to retirement, working conditions and quality of life in comparative perspective
2022, Health and PlaceThe Impact of (In)Stability of Pension System on Retirement Timing: Macro-Level Analysis Based on "Certainty Effect"
2023, Journal of Social Policy
- 1
Tel.: +1-215-898-0424; e-mail: [email protected].