In 1996, the United States fundamentally changed the rules of welfare for the very poor. Instead of giving cash benefits to poor mothers, the new Temporary Assistance for Needy Families program emphasized moving recipients quickly into the paid labor force, no matter how low the wages. The federal government gives states funds to support public assistance, but in return for temporary assistance and very limited services, clients must prove they are completing work activities and looking for employment. The new approach assumes that poor people mismanage their lives and need to develop greater self-discipline to become self-sufficient. Front-line administrators set clear behavioral expectations, monitor compliance, and use incentives to encourage compliance. When rules and incentives do not work, case-workers must apply penalties – up to and including the termination of aid altogether.

Today as in the past a high proportion of welfare clients come from racial minorities. In the tough new U.S. system, how do front-line welfare administrators decide which clients should be sanctioned for misbehavior? My colleagues and I have done research to nail down the answers.

Race, Rules, and Penalties

After 1996, each U.S. state participating in Temporary Assistance for Needy Families had to make decisions about such matters as time-limits for assistance, definitions of work and searching for work, and what to do when clients break rules. When states are classified according to the stringency of their rules, research shows that African Americans face tougher welfare systems. By 2001, just five years after passage of the new law, more than three-fifths of welfare families in the least stringent state programs were classified as white, while more than three-fifths of those in the most restrictive programs were identified as black. In 2001, more than three-fifths of welfare families in the least stringent programs were white; three-fifths of those in the most restrictive programs were identified as black.

Racial biases may also influence the local implementation of sanctions for noncompliant behavior. I have developed a novel theoretical framework for understanding when, exactly, racial stereotypes are likely to influence social policy design and implementation. I hypothesize that racial stereotypes operating below the level of conscious reflection are more likely to influence people’s decisions and actions when cues in the situation trigger ingrained racial assumptions.

For example, African Americans have long been stereotyped as lazy, not willing to work, and inclined to stay on cash assistance indefinitely. Hence, my perspective suggests that administrators may be especially likely to let unconscious racial biases creep in when they are dealing with rule-breaking welfare clients who have been on assistance for longer periods. Overall, the evidence confirms my hypothesis: blacks are far more likely than whites to get sanctioned if they have already received aid for relatively long periods.

Stronger Evidence from Welfare Administration in Florida

But maybe the blacks and whites were sanctioned differently in the findings just summarized because many were subtly different in other relevant ways. My colleagues and I wanted to find a research strategy that would offer a more precise test of racial bias. Turning to Florida’s version of get-tough welfare, the Welfare Transitions program, we used two research approaches to get beyond earlier limitations:

• First, we did a survey of front-line case workers that asked them to decide about sanctions for hypothetical welfare clients who broke various types of rules. We were able to make the vignettes about the rule-breakers realistic and identical except for the race of the client.
• Second, we corroborated our experimental results by analyzing large amounts of administrative data about how Florida caseworkers in various counties have actually made decisions about sanctions for rule-breakers with various characteristics.

As in all U.S. states, the guidelines in Florida’s program are race-neutral. In principle as well as law, sanction decisions must respond only to a client’s current behavior – not to her personal characteristics or prior experiences. Two-thirds of the Florida case managers were themselves women of color, and we learned from interviews as well as surveys that the vast majority were committed to norms of racial equality and unbiased treatment.

In our survey, we asked case managers to decide whether to sanction a fictional client. The facts of the case were always the same except for two elements that, according to the guidelines, should be irrelevant: the race of the client (black versus white) and whether the record indicated that the individual had received a sanction in the past.

Our results are telling: Case managers’ decisions about whether to sanction white clients did not change even if prior sanctions were mentioned in the vignette, and black and white clients with no prior bad marks in their vignettes elicited the same decisions. But when a prior sanction was indicated in a vignette about a black client, case managers became significantly more likely to impose a new sanction. In fact, they imposed the new sanction in overwhelming numbers and did so regardless of their own racial identities. Overall, the black client with a prior sanction was 29 percent more likely to be punished than a white client who had been previously sanctioned.Listen to an Office Hours interview with the author about “poverty governance”.

Random-assignment experiments like this are considered the “gold standard” in empirical research. But still, we wondered if real-world Florida welfare decisions would show the same patterns of racial bias. Our analysis of Florida-wide data revealed that African-American clients with no prior sanction were twice as likely to receive a sanction as white clients who shared their characteristics (28 percent versus 14 percent, respectively). The field results were statistically significant, and in the same direction as the experiments using fictional vignettes.

Various research results, in short, underscore that race-neutral rules, racially diverse personnel, and commitments to equal treatment are not enough to prevent racial biases in poverty programs. As welfare becomes ever tougher in today’s United States, we all have reason to worry that, once again as so often in America’s past, the minority poor are treated more severely than others.

—–

Read more in Joe Soss, “Deciding to Discipline: Race, Choice, and Punishment at the Frontlines of Welfare Reform” (with Sanford Schram, Richard Fording, and Linda Houser). American Sociological Review 74, no. 3 (2009): 398-422, and Disciplining the Poor: Neoliberal Paternalism and the Persistent Power of Race (with Richard C. Fording and Sanford F. Schram) (University of Chicago Press, 2011).