Disentangling the Age, Period, and Cohort Effects using a Modeling Approach

Disentangling age, period, and cohort effects in explaining health trends is crucial to assess future prevalences of health disorders. The identification problem -- age, period, and cohort effects are perfectly linearly related -- is tackled by modeling cohort and period effects using lifetime macro-indicators. This approach -- innovative in analyses on health trends -- handles theidentification problem and explains mechanisms underlying cohort and period effects. The modeling approach is compared with graphical and two-factors methods. The methods are applied on Dutch trends in functional limitations using data from the Longitudinal Aging Study Amsterdam. We argue that the modeling approach is a highly appropriate alternative. We find that theprevalence of functional limitations increases in the nineteen-nineties due to adverse cohort and period effects. Cohort effects are explained by hygienic and socio-economic conditions during childhood and period effects by restrictions in availability of health care services.