This paper proposes a new way of modeling age, period, and cohort effects that improves substantively and methodologically on the conventional linear model. The linear model suffers from a well-known identification problem: If we assume an outcome of interest depends on the sum of an age effect,...