Document Type

Working Paper

Publication Date



Relationships among socio-demographic characteristics, general assessments of health, and old-age mortality are well established in developed countries. There is also an increasing focus on the connection between early-life experiences and latelife health. This paper tests these and other associations using representative survey data from Taiwan on the population aged 60 and older in 1989, 1993, and 1996 that have been linked to data on deaths between 1989 and 1999 from a national death registry. The study also explores the possible influence of Taiwan’s Universal Health Insurance Program, instituted in 1995, and whether or not the survival of some groups of older people may have been differentially enhanced. Mortality is modeled using Gompertz regression. Multiple survey waves are employed to construct time-varying covariates. Some results verify findings of past studies; others are new. Effects of education are attenuated after the introduction of some health indicators. Functional and global assessments of health have stronger associations with mortality than do self-reports of health behaviors or particular chronic conditions such as diabetes. Mainlanders have higher survival than others. The survival of older adults with the greatest number and severity of functional limitations improved over the 1990s, suggesting a possible beneficial influence of the insurance program.