Demographic and Educational Effects on Long-Term Earnings

University Of Kansas Lawrence, Lawrence KS

Investigators

Linked publications, trials & patents

Abstract

DESCRIPTION (provided by applicant): This project seeks to evaluate the long-term effects of education and demographic variables on cumulative earnings over different stages of the adult life course and across successive birth cohorts. During the 20th century, the division between less-educated and more-educated Americans grew to become one of the most important determinants of life chances. While this conclusion is important and indisputable, it is nonetheless largely based on cross-sectional measures which may actually understate the rise in long-term earnings inequality. To date, due to a lack of appropriate data, the vast majority of research has been based on the measurement of earnings in a single year or even some shorter time period. However, workers experience earnings volatility over their careers and the pattern of this volatility is likely to vary by education and birth cohort. If low income earners ae more likely to retire early or to be unemployed for a longer period of time than high income earners, then long-term earnings differentials will be understated by estimates based on annual incomes. In addition, the effect of education on long-term earnings has considerable implications for retirement and aging including issues relating to Social Security benefit levels, work-related pensions, private retirement savings, and health and mortality outcomes. A few studies on the returns to education on lifetime earnings do exist, but they have been limited to synthetic cohort methods using cross-sectional data. In contrast to previous research, we propose to make use of a restricted data set that matches respondents in wave 2 of the Survey of Income and Program Participation (2004 panels) to their longitudinal earnings records in the Social Security Administration. Using this matched data, our research aims include to evaluate and contrast the actual 20 years cumulative earnings (from 1982 to 2008) for four different birth cohorts by educational subgroups; to apply a semi-synthetic cohort method to compute lifetime earnings (i.e., 50 year cumulative earnings from age 20 to 69) by educational groups, field of study, and race/ethnicity; and to estimate the causal effect of educational attainment on long-term earnings using propensity score matching methods.