For journalists, researchers and policymakers interested in understanding how job earnings differ across demographic groups, keep this mind: Measures of racial wage gaps that don’t account for prisoners and people who have been out of work a long time are probably missing the full economic picture, suggests new research by University of Massachusetts Amherst economist Jeannette Wicks-Lim.
The paper, “Revising the Racial Wage Gap Among Men in the United States: The Role of Nonemployment, Underemployment, and Incarceration,” is based on an analysis of nearly three decades of wage data and survey panels, spanning 1981 to 2008, of prime working-age men, 25 to 54.
The racial wage gap usually refers to the difference in average reported employment earnings between demographic groups — between white and Black men, for example.
The new findings indicate that accounting for men who were incarcerated and men who were out of work for more than one year expands the racial wage gap by nearly one-third during the timeframe studied.
The data used in the current paper comes from nationally representative, longitudinal and confidential survey panels conducted by the U.S. Census Bureau. Panel studies follow the same group of people over a period of time, typically four years in this case.
The panels included varying numbers of participants, with an average of 21,000 men per panel, Wicks-Lim says. The Census Bureau matched survey responses with Social Security and Internal Revenue Service information, allowing for tracking of individual earnings over time.
The 2008 survey panel was the most recent the Census Bureau matched with the administrative records.
The racial wage gap and long-term unemployment
Across the period studied, Wicks-Lim finds white male wage earners made 46% more per year, on average, than their Black counterparts — when her analysis only included men who worked regularly or were unemployed for less than one year.
But when Wicks-Lim included incarcerated men and those who were “nonemployed,” the gap grew 28 percentage points, to 74%. In the paper, Wicks-Lim counts individuals as nonemployed if they had no earnings for at least one year during the four-year panel they took part in.
“We know that there are more and more people who are not participating in the labor market,” she says. “And we know that this has a racial dimension to it because of the very publicized phenomenon of mass incarceration — but also the labor force participation rate among men has been declining.”
From the 1980s through the 2000s, Black men were increasingly and disproportionately incarcerated compared with white men. While the racial incarceration gap has shrunk in recent years, stark differences in incarceration rates by race persist. Across state and federal penal institutions, Black people were imprisoned at a rate of 901 per 100,000 residents, compared with 181 per 100,000 residents for white people, according to a 2021 report from the U.S. Bureau of Justice Statistics, the most recent available.
The lowest period of nonemployment among Black men was still 3 percentage points higher than the highest period of nonemployment among white men, the new paper finds.
The highest period for white men was during the Great Recession of the late 2000s, when their nonemployment share rose to 23%, according to Wicks-Lim’s analysis.
In other words, during the 2008-to-2011 panel, 23% of white men had at least one year without earnings.
The lowest period for Black men was during the late-1980s economic expansion, when their nonemployment share fell to 26%. Black men also experienced their highest rate of nonemployment, 35%, during the Great Recession.
Analyses of racial wage gaps only among people with job earnings aim to understand how race affects what working people earn. “I don’t mean to criticize that at all,” Wicks-Lim says. “I think that’s a perfectly useful exercise.”
But, for a broader view of earnings outcomes across racial groups, she says, “it’s really important to look at the employment dimension. It’s really important to know who is in and out of the labor force, who is in and out of a job — because we know that those things are racialized.”
Crucially, being out of work for a long while makes it difficult to earn in the future. Those who work sporadically and experience long spells of being out of a job earn less than those continually employed — with Black men more affected by this trend than white men, Wicks-Lim finds.
Other research has drawn similar conclusions about the relationship between extended unemployment and future earnings, including recent work published in The Industrial and Labor Relations Review and in American Psychologist.
Foundational research
Forty-three years ago, Duke University economist William Darity noted that selection bias — meaning a subset of a population a researcher is studying is excluded from analysis — was leading academics and journalists to overestimate the economic progress of Black workers.
In the early 1980s, news stories and research papers proclaimed an astonishing, shrinking earnings gap between Black and white workers during the 1960s and 1970s.
Darity was among the first to interrogate the underlying Census data used in most racial wage gap analyses at the time — the datasets left out people without earned income.
“The central argument is this: Blacks who are receiving incomes may well be moving closer to whites who are also receiving incomes,” Darity wrote in his 1980 paper, “Illusions of Black Economic Progress,” also published in The Review of Black Political Economy. “Nevertheless, if the proportion of Blacks who receive no income is consistently larger than the proportion among whites, racial inequality measured by the ratio of per capita incomes may not reveal the same evidence of approaching parity.”
Looking at earnings data from 1967 to 1977, Darity found “a clear decline in the share of black men with earned income while the white share remained stable over the decade.”
Similar research published in 1980 by Darity and University of Minnesota economist Samuel Myers Jr. found “Black men display virtually no convergence toward white men,” in terms of average income when they included men without earnings.
More than four decades later, the new paper reinforces Darity’s and Myers Jr.’s observations and serves as an important reminder that what is left out of an analysis can be as important as what is in it.
Expert Commentary