Carruthers and Wanamaker seek to better understand changes in the black-white wage gap, hypothesizing that the period between 1940 and 1980 is especially important because these were the years in which first, school desegregation led to a reduction in the quality gap between black and white schools, and second, the labor market moved from one without explicit protections for racial minorities to one with codified protections in the Civil Rights Act of 1964. They propose to use newly digitized data on race-specific school quality at the county level, combined with a long-running panel of longitudinal earnings data, to answer three questions. First, for birth cohorts (1902-1925) educated in the segregated pre-World War II South (1910-1940), did labor market outcomes reflect the poorer school quality of black men and women and to what extent are school quality differences associated with the black-white wage gap over the lifecycle? Second, what effect did Civil Rights legislation have on this cohort of workers and their labor market outcomes and did these effects differ by age, experience level, and schooling-related human capital at labor market entry? And last, how did human capital gaps and civil rights legislation interact for this cohort?
The PIs will link data from several sources. They have already assembled county-level data on school quality for ten Southern states between 1910 and 1940, with measures of spending by race, teacher salaries, student-teacher ratios, number of certified teachers by race, and length of school year. School quality will be constructed as an index from each measure of school resources in the digitized 1910-1940 panel. These data will be merged with the restricted-access Continuous Work History Sample (CWHS) provided by the Social Security Administration (SSA). The CWHS contains earnings records on covered jobs from 1937 to 1977 for a 0.01 percent sample of all social security numbers (SSNs). A larger one percent sample contains income information between 1951 and the present. These longitudinal data contain the earnings, industry, and work histories of individuals, and critically, their place of birth. Birthplace information is essential in order to match individuals to their likely county of schooling. The data also include demographic information, such as race and sex, from the social security applications. They also plan to merge in county-level socioeconomic factors from the Censuses.