Herwig Birg (2000)
An Approach for Forecasting Life Expectancy and its Application in Germany
In: Zeitschrift für Bevölkerungswissenschaft, Vol. 25, 1/2000, p. 175-198, Opladen: Verlag Leske + Budrich, ISSN: 0340-2398
In most mortality studies, the dynamism of changes in life expectancy is based on period analysis using cross-sectional data referring to a specific period. The cohort approach is methodologically more satisfactory, but it has the disadvantage that, by the time the mortality figures for a particular cohort are complete, they reach back up to 100 years in time, which means the approach has little value as a basis for forecasting.
This paper presents a new approach which combines period and cohort analysis. It therefore offers the advantage of period analysis (the use of more up-to-date statistical data) coupled with the methodological benefits of the cohort approach. The analysis is based on German life tables going back as far as 1871. On this empirical basis, the approach is used to issue a forecast of life expectancies up to the year 2080.
The middle variant forecasts a growth in male life expectancy from 73.3 years in 1994/96 to 81 years in 2080, and in female life expectancy from 79.7 to 87 years. The gain in life expectancy (i.e., in further life expectancy) is considerably greater for the older age-groups than it is at ages below 50 years. The results are presented in the form of survival functions from which age-specific death probabilities for both sexes and further life expectancies can be derived for each year during the forecast period.