ABSTRACT
In this chapter, the author addresses the identification problem. He provides a complete structure of the multiple estimators of the age-period-cohort (APC) model. This structure of estimators leads to the discovery of the estimable function that determines completely the trend in the age, period, and cohort. The author discusses the properties of the intrinsic estimator defined based on the estimable function. He then provides an outline of the justification, and a heuristic explanation of why this method works while others may not. The unbiased estimator of the estimable function is called the intrinsic estimator, which was first discovered in Fu as the limit of the ridge estimator when the penalty tuning parameter diminishes to 0. The author provides details of a robust parameter estimation study through a sensitivity analysis and summarizes the asymptotic properties of the intrinsic estimator and the multiple estimators based on the large sample theory.
