ABSTRACT
This chapter provides additional models and data designs that provide promising avenues for future development and substantive research with regard to age-period-cohort (APC) models. A smoothing cohort model utilizes the fact that the overlap of adjacent cohorts requires that each cohort effect be estimated with contributions from near neighbors. This model replaces the fixed cohort effects in the APC accounting model with smoothed cohort effects through a nonparametric spline smoothing function. It is well known that the conventional APC linear accounting model suffers from the identification problem methodologically. The multicohort, multiwave data design is especially important for aging and cohort analysis. The technique of the Intrinsic Estimator for modeling APC tabular data helps one to obtain statistically sound solutions to the identification problem that has long compromised previous APC analyses when linear models or generalized linear models are applied to aggregate tabular data or contingency tables.
