|
Abstract:
|
Economic and social disparities have received growing attention in studies of economic mobility and intergenerational mobility. Although these measurements have been well studied, statistical methods have been developed for accurate estimation. For example, the intergenerational elasticity has been found to be biased due to an identifiability problem (IP) in age-period-cohort model in studying income across generations (Lee and Solon 2009), and less efficient (Mitnik 2020, Mitnik and Grusky 2020). In life cycle studies of income, investment, and savings, it is crucial to dissect separate effects of age, period, and cohort yet the three factors are well-known confounded. The period effect attribution approach (Deaton and Paxson 1994) to the IP has been challenged by a cohort effect attribution (Chamon and Prasad 2010), raising further questions: Which one to select? How and Why? These fundamental questions need to be answered, not only to address specific questions, but also to develop methods that would work for all. In this paper, I will present novel methods using an intrinsic estimator to address these issues and illustrate with the SCF, CPS, and PSID data on income and savings.
|
