Honig M, Hanoch G
Soc Secur Bull. 1980 Apr;43(4):29-39.
Identifying the separate effects of age, time period, and birth cohort is of obvious importance in studies of the older population and especially critical in analyzing the labor-market behavior of the elderly. Age and the aging process in particular are fundamentally associated with changes occurring in earnings and incomes of persons at or near retirement. This article estimates and analyzes the separate effects of age, cohort, and period for some key labor-market variables, utilizing panel data from the Retirement History Study of the Social Security Administration, matched with social security earnings records. The analysis applies a strategy recently developed by the authors for estimating these separate effects from period-cohort means. The approach provides a solution to the well-known identification problem that has long been an obstacle in these studies. Integrated with the age-period-cohort analysis is a behavioral simultaneous model of labor-force participation, annual hours and weeks of work, threshold labor-supply quantities and reservation wages, wage offers, and asset holdings of older married men and women who are heads of households.