I appreciate the opportunity to respond to the thoughtful comments by Yang and Land (henceforth Y&L), O’Brien, Held and Riebler (henceforth H&R), and Fienberg. In this response, I will (1) elaborate on the limitations of the age-period-cohort (APC) accounting model, (2) clarify apparent misunderstandings about the intrinsic estimator (IE), and (3) outline a methodological research agenda that will facilitate improvements to substantive research.
My article focuses on IE, an estimation technique developed under the APC accounting model—an analysis of variance (ANOVA) model that includes age, period, and cohort as three independent variables predicting some outcome. The identification problem inherent in the APC accounting framework is not news. Researchers in many disciplines have participated in the century-old debate about the identification problem and possible solutions (see, e.g., Clayton and Schifflers 1987; Fienberg and Mason 1979, 1985; Glenn 1976, 2005; Held and Riebler 2012; O’Brien 2000...
