POVERTY AND MORTALITY AMONG THE ELDERLY - MEASUREMENT OF PERFORMANCE IN 33 COUNTRIES 1960-92

This paper analyses the effect of income and education on life expecta ncy and mortality rates among the elderly in 33 countries for the peri od 1960-92 and assesses how that relationship has changed over time as a result of technical progress. Our outcome variables are life expect ancy at age 60 and the probability of dying between age 60 and age so for both males and females. The data are from vital-registration based life tables published by national statistical offices for several yea rs during this period. We estimate regressions with determinants that include GDP per capita (adjusted for purchasing power), education and time (as a proxy for technical progress). As the available measure of education failed to account for variation in life expectancy or mortal ity at age 60, our reported analyses focus on a simplified model with only income and time as predictors. The results indicate that, control ling for income, mortality rates among the elderly have declined consi derably over the past three decades. We also find that poverty (as mea sured by low average income levels) explains some of the Variation in both life expectancy at age 60 and mortality rates among the elderly a cross the countries in the sample. The explained amount of variation i s more substantial for females than for males. While poverty does adve rsely affect mortality rates among the elderly (and the strength of th is effect is estimated to be increasing over lime), technical progress appears far more important in the period following Igbo. Predicted fe male life expectancy (at age 60) in 1960 at the mean income level in 1 960 was, for example 18.8 years; income growth to 1992 increased this by an estimated 0.7 years, whereas technical progress increased it by 2.0 years. We then use the estimated regression results to compare cou ntry performance on life expectancy of the elderly, controlling for le vels of poverty (or income), and to assess how performance has varied over time. High performing countries, on female life expectancy at age bo, for the period around 1990, included Chile (1.0 years longer life expectancy), China (1.7 years longer), France (2.0 years longer), Jap an (1.9 years longer), and Switzerland (1.3 years longer). Poorly perf orming countries included Denmark (1.1 years shorter life expectancy t han predicted from income), Hungary (1.4 years shorter), Iceland (1.2 years shorter), Malaysia (1.6 years shorter), and Trinidad and Tobago (3.9 years shorter). Chile and Switzerland registered major improvemen ts in relative performance over this period; Norway, Taiwan and the US A, in contrast showed major declines in performance between 1980 and t he early 1990s.