PRRAC Poverty & Race Research Action Council
Home About PRRAC Current Projects Publications Newsletters Resources Contact Us Support PRRAC Join Our Email List

"Why Are African Americans and Latinos Under-represented Among Recipients of Unemployment Insurance and What Should We Do About It?,"

by Andrew Grant-Thomas May/June 2011 issue of Poverty & Race

While the Great Recession left no group or community unmarked, it marked some more than others. African Americans and Latinos, for example, have endured much higher rates of unemployment than have non-Hispanic Whites—16%, 12% and 8%, respectively, in January 2011. African Americans are also especially prominent among the long-term unemployed. According to the Current Population Survey, Black workers were 11% of the U.S. labor force but 22% of workers unemployed for 27 weeks or longer in 2009. Latinos fared better, but not well: 15% of the work force, they comprised 17% of the long-term unemployed.

How much of a safety net does unemployment insurance (UI) provide for jobless workers of color and their families? Data on the racial and ethnic identities of workers who apply for and get unemployment benefits are not reported fully or consistently by the states, each of which runs its own unemployment program under a broad set of federal guidelines. Using data from the Department of Labor Employment and Training Administration (“Characteristics of the Insured Unemployed”), we found 15 states for which the sum of the percentages of unknown racial identity (Black, White, etc.) and ethnic identity data (Hispanic, non-Hispanic) of UI recipients did not exceed 10% in 2009. These states are Alaska, Arkansas, Delaware, Georgia, Illinois, Louisiana, Maryland, Massachusetts, Mississippi, Montana, North Carolina, North Dakota, Ohio, Pennsylvania and Tennessee. Combined, those designated as unemployed in these states included 2.9 million White workers, 1.1 million African-American workers and 360,000 Latino workers. Although this sample is not as expansive as we might like, the findings it yields are suggestive.

A comparison between the racial and ethnic compositions of unemployed workers and UI recipients across all 15 states reveals that 43% of unemployed Whites, 39% of unemployed Blacks and only 32% of unemployed Latinos received UI benefits. This result is consistent with a National Urban League study finding that, in 2008, “17.5% of unemployed black workers received UI benefits as compared to 25.3% of white unemployed workers.” Allowing for the limitations of the data, it appears that African Americans are underrepresented, and Latinos very underrepresented, among UI recipients relative to Whites. In addition, given that solid majorities of workers in all three groups do not collect UI benefits, it is likely that many White, Black and Latino workers are foregoing benefits for which they are eligible.

The implications of these unemployment and UI recipiency figures for the well-being of people of color are grim in light of what we know about the financial resources available to Black and Latino families. In a recent presentation entitled “Social Security at 75: Building Economic Security, Closing the Racial Wealth Gap,” the Insight Center for Community and Economic Development provided data on racial differences in wealth that help put into relief the challenge many families of color face, especially when unemployment strikes.

Among married and cohabiting householders, the typical non-Hispanic White family ($193,400) had more than four times the wealth of its African-American counterpart ($46,900) and almost five times the wealth of its Latino counterpart ($39,100) in 2007. Among households led by single men and women, the disproportions were equally or more dramatic, but for our purposes what is perhaps more telling is the sheer wealth poverty of people of color. For example, whereas the median wealth of households led by single White women was $49,180 ($41,500 if we exclude the value of any vehicles owned), for single Black and Latino female-headed households those values were $5,000 and $2,680, respectively.

The underrepresentation of African Americans and Latinos among UI recipients is cause for alarm. This is so not simply because racial disparities in the distribution of important social benefits (and burdens) are in themselves suspect, but also because many African Americans and Latinos have little or no financial reserves with which to weather the storm of unemployment. No surprise, then, that local welfare offices and homeless shelters have seen a surge in people seeking help. Or that more and more people are turning for shelter and aid to families and friends, many of whom are struggling themselves. The costs of prolonged unemployment include health and mortality declines, marital conflict, and a rise in violent crime, poverty and debt. Extending UI coverage to more people, especially to more people of color, can help mitigate some of these costs.

Are Racial and Ethnic Disparities in UI Receipt Due to “Pre-existing” Conditions?

How can we account for the apparent underrepresentation of African Americans and Latinos among unemployment insurance beneficiaries? There are at least four possible answers, none of them exclusive of the others. First, it may be that Blacks and Latinos are less likely than Whites to apply for unemployment insurance. I am not aware of any good data that would allow us to evaluate that possibility effectively.

Second, it may be that Blacks and Latinos, but not Whites, happen to be most numerous in those states with the most restrictive UI eligibility criteria. This explanation has merit. I divided the 50 states and the District of Columbia into three tiers of 17 territories each according to the proportions of their unemployed workers receiving UI at the end of 2009. I then compared the distributions of the White, Black and Latino populations across these territories and found that Blacks and Latinos are unfavorably distributed with respect to state recipiency rates when compared to Whites. Whereas 35% of Whites lived in territories that fell into the least-generous third of the coverage distribution, 36% of African Americans and a sizable 53% of Latinos lived in such states. At the high end, 33% of Whites but only 27% of African Americans and 27% of Latinos lived in the most-generous third of states.

Third, it is also possible that African Americans and Latinos are more likely than Whites to fall into worker status categories that make them less likely to meet state UI eligibility criteria. This possibility also seems to bear some explanatory weight. Among the unemployed, African Americans were less likely than Whites to be “job losers” in fourth quarter, 2010: 58% of Blacks and 64% of Whites were “job losers.” We know from previous research that job losers are more likely to meet UI eligibility criteria than other categories of unemployed, including “new entrants” and “reentrants” to the job market.

Blacks and Latinos are also disproportionately low-income workers. The Economic Policy Institute estimated, for example, that in 2009 Blacks were 11% of the workforce, but 18% of workers whose incomes were low enough to be affected by the minimum wage increase to $7.25/hour in that year. Latinos were 14% of the workforce and 19% of workers affected. New entrants to the labor force, re-entrants, and low-income workers are all less likely to meet state monetary eligibility criteria for unemployment benefits.

Are Racial and Ethnic Disparities in UI Receipt Also Due to Direct Racial Bias in the UI System?

A final possible reason for racial disparities in UI receipt is that the unemployment insurance program itself is racialized in ways that merit attention and redress. Of course, race and racial bias may actually play causal roles in the explanations outlined above as well.

It is possible that the original design of UI criteria was racialized and continues to do its racial work today. Note that the state-by-state distribution of the Black population has changed only moderately in the 75 years since passage of the Social Security Act that established the UI program. The same band of states and territories that had the largest shares of Black Americans and included many of the most restrictive UI eligibility rules back in the 1930s—from the eastern half of Texas through the northern tip of Florida and up to the District of Columbia—still remain home to the country’s largest shares of African Americans and many of its most restrictive UI criteria. Race is hardly the only plausible reason for this convergence of geography and rule-making. That said, from the exclusion of agricultural and domestic workers from the original provisions of the Social Security Act itself to the mid-19th Century establishment of felon disenfranchisement laws in some states that attached the sharpest penalties to crimes thought to be committed most often by Blacks, many scholars have argued that brute racial bias accounts for much of the contemporary racialized impact of some of our most important social policies.

On a second front, a large and growing research literature on implicit cognitive bias provides a strong prima facie case for the role of racial bias in the distribution of UI benefits. Implicit or hidden biases refer to the automatic beliefs, attitudes and stereotypes we hold about categories of people or things. As Jerry Kang notes in his “primer” on implicit bias for the National Center for State Courts, researchers have found that implicit bias predicts how much more readily players will “shoot” African Americans compared to Whites in a videogame simulation when given a split-second to decide whether the target represents a danger to oneself or others. Greater levels of implicit bias have been linked to more negative assessments of ambiguous actions taken by African Americans. Researchers have documented a positive association between medical doctors’ implicit racial attitudes and their unequal treatment for Latinos and Black patients compared to White patients. More or less implicit bias corresponds to comfort level and body language in interracial interactions. And so on.

Especially at a time when the number of contested UI benefit claims have soared—56,000 in the year ending June 2010 in Maryland alone, according to the Baltimore Sun—the workings of the UI system rely substantially on human judgment. Does the evidence support the claimant’s argument that she was fired rather than the employer’s contention that she quit? Is ambiguity in this or that regulation best resolved in the claimant’s favor or not? On the strength of the rapidly mounting evidence of the real-world effects of implicit bias, it is not cynical to suppose that bias, implicit and otherwise, might play a systematic role in how UI bureaucrats (i.e., human beings) answer these questions. Arguably, it would be naďve to suppose otherwise.

If racial and ethnic biases help produce racial and ethnic disparities in the receipt of unemployment benefits, one might expect, for example, a negative association between state recipiency rates and the proportion of the state population that is African- American or Latino. On the assumption that racial biases will most often operate against Blacks and Latinos, the logic here is that states with larger proportions—not numbers, but proportions—of Blacks or Latinos will also have lower recipiency rates. That is what we see: a negative association for African Americans (correlation coefficient of -.40) and Latinos (correlation coefficient of -.16), but a positive association for Whites (correlation coefficient of .22). This I take to be fairly compelling evidence for the bias claim.

By a similar logic, if bias were systematically skewing outcomes against Blacks and Latinos, we would expect to see an association between the proportion of cases in which UI benefits are denied in error and the proportion of claimants who are Black or Latino. Using data from the Department of Labor Employment and Training Administration’s Benefit Accuracy Measurement Report for 2009, we found that states’ shares of African Americans and Latinos were indeed positively associated with the rates of one kind of improper denial (improper monetary denial rates, correlation coefficient of .27), but not with either of the other two kinds (improper separation and non-separation error rates).

Is It Time to Fully Federalize the Unemployment Insurance System?

There are several compelling reasons to federalize unemployment insurance. Perhaps most compelling is the matter of funding, the “elephant in the room” with respect to UI. The huge number of unemployed workers receiving regular and extended federally-funded benefits makes the program de facto a federal one already. As noted in a February 9, 2011 New York Times editorial (“Relief for States and Businesses”), UI trust funds in 32 states are currently insolvent; those states are $42 billion in debt to the federal government, with more states likely to follow suit in 2011. In the 1980s, a similar crisis led 44 states to reduce the extent and amount of their benefit coverage.

Given the limits of the states’ ability or willingness to manage their trust funds effectively, it may make sense now to hand over the program entirely to the federal government. If the “modernization” reforms recently adopted by a majority of states are to be made permanent, federal control makes even more sense.

As of December 2010, 18 states had left unclaimed some or all of the federal funds available to them if they modernize their UI systems. The main thrust of the modernization reforms has been to extend coverage to previously excluded workers, including part-time and low-wage workers. As we have seen, people of color feature prominently among those too often left out before the recent wave of modernization. Governors and some legislators in the holdout states express concern about the conditions attached to the federal dollars. Some, like Texas Governor Rick Perry, worry that the demand for a more inclusive UI system will remain in place long after the federal funding to support it has dried up.

In light of such resistance, those concerned to see more vulnerable workers of color covered now and in the future, as needed, should hope to see the program federalized and modernization reforms extended to more states and made permanent. Our analyses show that for both Blacks and Latinos, modernization measures enacted in 2009 and 2010 mitigated the relationships between racial proportionality and recipiency rate at the state level. As recent work by Wayne Vroman and Jacob Benus demonstrates, because unemployment insurance dollars are typically spent, rather than saved, states with higher UI recipiency rates also provide greater boosts to their economies than states with lower recipiency rates. More extensive modernization efforts, then, would likely be a boon for the economies in affected states, as well as for unemployed workers of color in those states.

The key caveat to the claim that UI is effectively a federal program is this: The federal government can only “incentivize” certain state-level choices, as it did in the case of modernization. It lacks the authority to make efficient programmatic decisions. The federal government cannot easily make mandatory the collection of racial and ethnic identity data for UI claimants and recipients, though this would greatly improve our ability to assess the program’s fairness. The federal government can conduct audit tests of bias in claims processing, and should do so, but how readily it could provide de-biasing training in those places showing evidence of bias is unclear.

Finally, placing the unemployment insurance system under federal control would make it much more possible to meet the challenge of creating greater equity across states, as well as within them. As discussed above, both African Americans and Latinos are substantially underrepresented, relative to White workers, in the least restrictive UI states. Relative to White workers, Latinos are also greatly overrepresented in the most restrictive states. These “horizontal” inequities extend to other important dimensions of the system, such as large state variation in the portion of the weekly paycheck replaced by the UI benefit and the degree to which unemployed workers have exhausted their benefits.


It was only a short while ago that unemployment insurance was national news. The occasion was President Obama’s agreement with his political opponents to extend “tax cuts for the wealthy” as well as unemployment benefits. Little attention was paid then, before or since to questions of racial equity in the disbursement of UI benefits. The suggestive findings reported here regarding the underrepresentation of African Americans and Latinos receiving UI benefits, coupled with the undeniable social and economic vulnerability of many members of those two groups, make it imperative that much more attention be paid to the matter of UI equity.

Given the scarcity and other limitations of the data, some energy should be devoted to data-gathering and fact-finding. Not least, we need rigorous collection of racial data on who applies for and receives UI across all 50 states, the District of Columbia and Puerto Rico. We also need to understand why so many potential claimants do not file for UI benefits in the first place. Beyond this, the UI system needs higher levels of sustainable funding; greater “vertical” equity across worker status categories within states; and greater “horizontal” equity to bridge disparities in payment amounts, exhaustion rates and worker status treatment across states. In sum, from funding to operations, the unemployment system needs the resources and coordination capacity that only the federal government can provide.


National Employment Law Project,

National Urban League, “Achieving Fairness and Efficiency in Unemployment Insurance,” _June2010.pdf

American Values Institute,

Jerry Kang, "2009 Implicit Bias Primer for Courts,"

Wayne Vroman & Jacob Benus, "The Role of Unemployment Insurance as an Automatic Stabilizer During a Recession,"

Insight Center for Community and Economic Development, "Social Security at 75: Building Economic Security, Narrowing the Racial Wealth Divide,"

Andrew Grant-Thomas is Co-Founder and Co-Developer of EmbraceRace and works as an independent consultant. Formerly Senior Researcher at the Harvard Civil Rights Project and Deputy Director at the Kirwan Institute for the Study of Race & Ethnicity, he has written, spoken, and worked on a wide range of race-related issues. agrantth

Join Our Email List
Search for:             
Join Our Email List