Much of the current labor market research on black-white inequality focuses on explaining differences between young African Americans and whites. However, since 1970 a major decline in the labor force attachment of older men occurred, with the participation rates of older African American men falling fastest. A variety of hypotheses have been offered to explain the general decline. One hypothesis is that the increase in social security and pension coverage rates and their benefit levels raised the incentive to leave the labor force at earlier ages. Another hypothesis is that the real wages of older men fell, lessening the incentive to remain in the labor force. Current research supports the social security and pension hypothesis. To my knowledge little work has been done to explain why the African American participation rate fell faster. This project proposes to answer the following questions: 1) How much faster did the labor force participation rate of older African American men fall? 2) Do changes in racial differences in social security and pension benefit levels and coverage rates explain the faster decline? Was the faster decline due to a change in the relative sensitivity towards benefit levels and coverage rates? 3) Did the real wages of older African American men fall faster than the real wages of older white men? If so, does this change in relative wages explain the faster decline? 4) What role do other factors such as changes in disability status, and the economy's shift away from manufacturing jobs toward service jobs play in explaining the faster departure of older African American men from the labor force?
The project first provides a detailed comparison of older African American and white men's labor force participation. The project then uses endogenous breakpoint methods to identify the rates at which the participation of African American and white men declined. Finally, the project uses decomposition methods to examine a variety of potential explanations. The analysis uses the Current Population Survey (CPS) March Annual Demographic files (1964 to 1997) to construct labor force participation rate time series. The files contain information about wages, educational attainment, industry and occupation of employment, and a variety of geographic data. The files also contain information on social security and pension coverage, social security and pension income, and disability status. Each yearly sample contains on average 20,000 men that are at least 45 years of age. Approximately, 8 percent are African American men. The CPS files have several drawbacks. They do not contain earnings histories, which are needed to construct social security coverage and benefit levels. Nor do they contain wages for respondents that are out of the labor force. To partially circumvent these limitations, the project uses two longitudinal data files that focus exclusively on the elderly: the Retirement Health Survey (1969-1979), and the Health Retirement Survey (1992, 1994, and 1996). These data contain a richer set of information on wages, employment' retirement decisions, and income. The RHS starts with 8,131 men in 1969. The HRS starts with 5,866 men in 1992. African American men comprise approximately 16 percent of the sample.
