The impact of changing union density on earnings inequality: Evidence from the private and public sectors

The degree of inequality in U.S. earnings has varied considerably over the past 20 years, including a dramatic, much documented rise since 1980. We examine empirically how changes in union density have contributed to these trends, using Current Population Survey data for 1977 and 1992. Inequality is measured as the mean logarithmic deviation of individual earnings from overall average earnings. A decomposition of the change in the inequality index reveals that decreases in private-sector union density have accounted for about 25 percent of the overall rise in earnings inequality during the past 15 years. Decompositions based on public-sector earnings indicate that increases in union density have produced inequality that is 29 percent below what it otherwise would have been. The analysis demonstrates that, among private sector workers, the results are sensitive to the population being studied: Changing union density accounts for 13 percent of the rise among prime aged males (a noticeably smaller fraction than found in existing studies) and only 4 percent among females and non-prime-aged males. The analysis also demonstrates that covariances between the subsamples explain why the union effect is larger in percentage terms for the whole sample than it is in either subsample.