UMBC logo

Department of Economics

Project Abstract

"Why Has U.S. Income Inequality Stopped Increasing Since the Mid-1990s?"

Marie-Claire A. Guillard

Income inequality increased in the United States from the 1970s through the 1980s and into the early 1990s. Since the U.S. has had one of the fastest rates of growth of income inequality among industrialized countries, it now has one of the highest levels of inequality as well. However, this trend of increasing inequality has leveled off since the mid-1990s. This paper examines the changes in the 1990s that contributed to the leveling off of income inequality. This is accomplished by using a technique recently developed in Fields (1998). The Fields' decomposition is based on earnings regressions and allows for the calculation of the proportion to inequality contributed by the changes in each independent variable.

The consensus in the literature is that the increasing inequality from the 1970s to the 1990s was attributed to increasing returns to skill, in particular education and experience, and increasing inequality within education and experience groups. Comparing earnings from 1971 to 1993, Fields' decomposition of earnings inequality during this period yields results that are consistent with the findings of the literature. For example, the significant positive change in the Gini index was 0.027. The change caused by the education variable alone was 0.0165, 0.0027 for the experience variable, and the increase in within-group inequality was 0.0327 (there were obviously some equalizing factors as well). Using Fields decomposition to compare 1993 to 1996 shows that the trnds from 1971 to 1993 continued into the early 1990s. That is, the significant positive change in the Gini index of 0.051 was primarily caused by education, which contributed to a change of 0.0102, and by within-group inequality, which increased by 0.0403.

The trends of earnings inequality from 1996 to 1999 are different from the 1993-1996 period. The top three determinants affecting earnings inequality during this period were education, gender, and within-group inequality. Unionization and race did not have significant impacts on inequality. Gender and experience were equalizing factors whereas education and within-group variation were still disequalizing. However, the contribution of the latter two factors was significantly less than in previous years. For example, the small insignificant change in the Gini index (a change o 0.005) contributed by the education variable was 0.0048, less than half as much as the 1993-1996 period, and experience decreased in equality by 0.0010. This demonstrates that skill was less disequalizing in the late 1990s than previously. Similarly, the contribution of within-group inequality decreased to 0.0022 during the 1996-1999 period. Thus the conclusion is that inequality has stopped rising in the late 1990s because returns to skill, in particular education and experience, stopped rising.

DEPARTMENT OF ECONOMICS • University of Maryland, Baltimore County
3rd floor Public Policy Building, Room 338
1000 Hilltop Circle • Baltimore Maryland 21250
phone: 410-455-2160 • email: Mrs. Kelly Hodges <>