Citation
Blanchflower, D.G., & Wainwright, J. (2005). An analysis of the impact of affirmative action programs on self-employment in the construction industry. Working paper no. 11793. Cambridge, MA: National Bureau of Economic Research.
Highlights
- The study’s objective was to examine the impact of City of Richmond v. J.A. Croson Co. (referred to as the Croson decision) in 1989, which weakened affirmative action policies, on self-employment, with a focus on the construction industry, by race, ethnicity, and gender.
- The study used a nonexperimental analyses to compare self-employment before and after the Croson decision. The authors used national data from the Current Population Survey (CPS) and the 2000 decennial U.S. Census.
- The study found that gender disparities in self-employment rates have declined since the Croson decision in all industries, but have declined most in construction. In contrast, racial and ethnic disparities in self-employment have declined since the Croson decision in all industries except construction.
- The quality of causal evidence presented in this study is low because the authors did not control for trends in self-employment or earnings before the Croson decision. This means we are not confident that the estimated effects are attributable to the Croson decision. Other factors are likely to have contributed.
Intervention Examined
City of Richmond v. J.A. Croson Co.
Features of the Intervention
In 1989, the U.S. Supreme Court ruled in the case of City of Richmond v. J.A. Croson Co. This decision instituted more stringent tests for the legitimacy of race- and gender-based affirmative action programs in public contracting. The Croson decision required a “strict scrutiny test” that there was actual discrimination to warrant race-based policies; otherwise, such race-based policies were deemed unconstitutional. Although the Croson decision was specific to Richmond, Virginia, other cities’ race-based affirmative action policies were challenged in court following the Croson decision, with many deemed unconstitutional, including in Philadelphia, Pennsylvania; Columbus, Ohio; Fulton County, Georgia (Atlanta); Dade County, Florida (Miami); Cook County, Illinois (Chicago); the District of Columbia; and the states of Florida and Ohio. The authors argue that such tests dampen the effects of affirmative action programs on promoting business formation by minorities, proxied in this study by self-employment rates.
Features of the Study
This study used a nonexperimental design to examine self-employment rates and earnings in the construction industry and on average across all industries by race, ethnicity, and gender. The analysis included noninstitutionalized men and women ages 16 and older in all states and industries who were either self-employed or employed in the private sector. The authors estimated regression models comparing the outcomes for different race, ethnicity, and gender groups before (1979 to 1991) and after (1992 to 2004) the Croson decision, controlling for age, education, industry, state, and year. The authors used data from the CPS March Annual Demographic files to estimate statistical models of earnings, with additional controls included for occupation. Data from the CPS Merged Outgoing Rotation Group files and the 2000 decennial U.S. Census were used to estimate models of self-employment.
Findings
- Across all industries, disparities in self-employment rates by race and ethnicity declined somewhat following the Croson decision. However, these disparities changed little in the construction industry following the Croson decision. The authors did not report whether the change before and after the Croson decision was statistically significant.
- Disparities in self-employment rates by gender declined following the Croson decision, especially in the construction industry. The authors did not report whether the change before and after the Croson decision was statistically significant.
- Disparities in self-employment earnings by race and ethnicity fell in the construction industry after the Croson decision, but increased on average across all industries. In contrast, disparities in self-employment earnings by gender increased in the construction industry, but fell on average across all industries. The authors did not report whether the change before and after the Croson decision was statistically significant.
Considerations for Interpreting the Findings
The authors defined the period before the Croson decision as 1979 to 1991, but the Croson decision occurred in 1989. Thus, some of the effect of the decision might be captured in the last two years of the comparison period as defined by the authors. In addition, the authors did not adequately control for pre-decision trends in self-employment and earnings for the groups being examined; therefore, the estimated effects might reflect pre-existing differences in employment and earnings and not the effect of the court ruling.
Causal Evidence Rating
The quality of causal evidence presented in this report is low because the authors did not control for pre-intervention trends in self-employment or earnings, which means the comparability of the treatment and comparison groups before the intervention was not established. This means we are not confident that the estimated effects are attributable to the Croson decision. Other factors are likely to have contributed.