by Alice Henriques, Columbia University
This project proposes to document the evolution of poverty for elderly women and quantify the relative importance of the factors that cause older women to be poor. In particular, this study will focus on both labor history and current and past marital status. Women have always been worse off than men, and this has not changed over the past several decades. It is well-known that widows tend to experience poverty at much higher rates than married women. In light of the feminist movement, we would hope that more than marital status determines poverty of women in old age. In contrast, we might expect that increased participation of women in the labor force would lead to increased economic well-being in the future. Many studies have documented possible causes and correlated of poverty of elderly women, but none have taken an integrated approach to consider multiple causes in tandem. This study aims to fill this gap. I plan on using a newly available dataset, the Synthetic SIPP Beta files, to document poverty patterns controlling for both labor force history and marital history. The advantage of this dataset is that it merges individual level survey data from all of the SIPP panels from the 1990s with confidential data from the Social Security Administration and the Internal Revenue Service. This provides a full earnings history in addition to information about Social Security claiming. The SIPP data contains a full history of all marital events up to the dissolution of a third marriage. The first step of this project is to carefully document the trends in elderly poverty for different groups in the population. The second step of the analysis is to decompose the poverty rates into changes due to shifts in labor participation, marriage and divorce rates, and changes in poverty which have arisen within each group. The final step is to identify the magnitude of the effect females’ labor force history has on old age poverty and I will use an instrumental variables strategy to address omitted variables bias in the regression equation. The results of this project will inform policymakers with information about how to target and reduce poverty of elderly women, a problem much documented but not solved at this juncture.