The aim of our study is to test the theories of compression or expansion of morbidity on the basis of data on the elderly population of Austria. Our data come from four microcensus surveys for the years 1978, 1983, 1991, and 1998. We use self-perceived health ratings to calculate healthy-life expectancy for the elderly population aged 60-89. Because our data are based on four cross-sectional surveys, we devote the first part of the paper to the consequences of possible sampling and non-sampling errors in our analysis of time trends. We come to the conclusion that, although the absolute number of years lived in good health may be overestimated, the time trend in healthy-life expectancy over the 20 years most probably is unbiased. The second part of the paper describes trends in healthy-life expectancy for the Austrian population. Our results suggest that both healthy-life expectancy and the ratio of healthy years to life expectancy increased between 1978 and 1998. Thus, in Austria ill health seems to be more and more compressed into the later years of life. Contrary to Fries's hypothesis, however, life expectancy does not seem to be approaching a maximum average life span in Austria, as mortality rates at older ages have been continuously decreasing over the last 20 years.