Cohort Based Analysis of Income Shocks Over the Life Cycle

TitleCohort Based Analysis of Income Shocks Over the Life Cycle
Publication TypeThesis
Year of Publication2011
AuthorsMoulton, JG
Number of Pages141
UniversityUniversity of California, Davis
CityDavis, CA
KeywordsAdult children, Employment and Labor Force, Methodology, Retirement Planning and Satisfaction, Social Security

The following three chapters investigate the first job, retirement decisions, and bequests of birth cohorts born between 1908 and 1926. I employ cohort based natural experiments using the Great Depression and Social Security Amendments of 1972 and 1977 to estimate the effect of entering the labor market during a recession, choice of retirement age in response to a retirement benefit shock, and financial transfers with children when facing a late life income shock. The first chapter uses a regression discontinuity framework to show that less educated workers entering the labor market during the Great Depression experienced 7 to 10% lower earnings roughly ten years after the onset of the Great Depression, in 1940. These effects fade substantially ten more years later in the 1950 Census. More educated workers entering the labor market experienced no loss to earnings in 1940 and positive earnings in the 1950 census, but these results may be driven by the effects of World War II or a more favorable job market for educated laborers during the Great Depression. Using the same regression discontinuity framework, I also show that young workers did not significantly delay their entry into the labor market during the Great Depression by prolonging their stay in high school. The second chapter reinvestigates and helps explain the findings of Krueger and Pischke (1992), who find no significant response of retirees in response to the Social Security Notch (Amendments of 1972 and 1977). Using a hazard model and data from the AHEAD cohort of the Health and Retirement Study, we find that using a long-term incentive measures, of the form found in Coile and Gruber (2000), provides a better estimate of retirement decision making than single year accrual measures. We find that individuals respond to both the long-term incentive measure (decreases the hazard of retiring) and the present value of their Social Security wealth (increases hazard). However, using the exogenous variation caused by the Social Security Notch results in opposite signed effects for the incentive measure. We show that this odd result and the reason that Krueger and Pischke do not find a significant change in retirement is that the Social Security Notch is one experiment, but has two offsetting, but correlated treatments. The Social Security amendments increased both the incentive measure and Social Security wealth for the cohorts subject to the 1972 amendments (Pre-Notch cohort) and the opposite for those subject to the 1977 amendments (Notch cohort). We show that the lack of retirement response is not due to no response, but to large offsetting effects. The last chapter investigates intergenerational wealth transmission between fathers and children using the 1972 and 1977 Amendments to the Social Security Act. As shown in chapter 2, individuals did not change their retirement age in response to these amendments, so those affected by the 1972 amendments experienced increased growth in Social Security benefits while those subject to the 1977 amendments experienced a negative shock (the "Notch Generation"). Using the Health and Retirement Study, I find that sons with fathers born in the Notch Generation have significantly lower wealth levels than those with fathers born in surrounding cohorts, most notably the cohort born just prior. Further evidence suggests that the wealth difference is likely due to a higher probability of providing financial support to parents and a lower probability of receiving an inheritance and financial assistance from parents.

Endnote Keywords

labor market

Endnote ID


Citation Key6020