A Methodological Comparison of Age-Period-Cohort Models: Fu's Intrinsic Estimator and Conventional Generalized Linear Models

Yang Yang, Duke University
Kenneth C. Land, Duke University
Wenjiang J. Fu, Michigan State University

Age-Period-Cohort (APC) accounting models have long been objects of attention in statistical studies of human populations. It is also well known that the identification problem created by the linear dependency of age, period and cohort (Period = Age + Cohort) presents a major methodological challenge to APC analysis and it still remains largely unsolved. This methodological problem has been widely addressed in demography, epidemiology, and statistics. This paper compares parameter estimates produced by the two solutions to the identifiability problem in age-period-cohort models, namely, Fu's intrinsic estimator method and conventional generalized log-linear models (Mason and Smith 1985), using population data on disease and mortality rates.