Skip to main content

Research Note: 1990s Cash Assistance Policy Changes Associated With Negative Social Behaviors for Adolescents, Particularly Boys

Federal and state policy changes to the nation’s primary cash assistance program for low-income mothers and their children in the 1990s, including time limits and loss of the entire family’s assistance if a parent didn’t meet a work requirement, “have come at a cost to the next generation, particularly to boys,” according to a new study.[1] Adolescent boys most at risk of exposure to these policies were likelier to skip school, damage school property, or get in a serious fight; adolescent boys and girls most at risk of exposure were likelier to use marijuana and other substances. These behavioral effects may have hindered the adolescents’ well-being, health outcomes, and socioeconomic trajectories as they transitioned into adulthood, the researchers explain. The study’s findings “do not support culture of poverty arguments that requiring poor mothers to work would make the next generation more responsible,” they add.[2]

States enacted experiments and policy changes to their cash assistance programs at various points during the 1990s, which enabled the researchers to use statistical methods exploiting variations in the timing of those changes to evaluate adolescents’ behavioral outcomes.[3] In the early 1990s, when cash assistance was still an entitlement program known as Aid to Families with Dependent Children (AFDC), some states began testing policy changes through waivers, including conditioning receipt of assistance on meeting work requirements.[4] Before some of those experiments concluded, federal policymakers replaced AFDC with Temporary Assistance for Needy Families (TANF), which put time limits on receipt of TANF and allowed states to impose harsher work penalties. Also, under TANF, states no longer had to apply for waivers to impose additional sanctions and work requirements on recipients, which meant these policies were not subject to the rigorous evaluations required before 1996.

In states where AFDC experiments or TANF implementation had been in place for at least 12 months, the researchers evaluated the behavioral outcomes for 10th- and 12th-grade youth of unmarried single mothers with at most a high school degree — that is, families most at risk of being affected by the policy changes in their state’s cash assistance program. Researchers compared outcomes overall and by gender for these “exposed youth” (the target group) to those of “unexposed youth,” or their youth counterparts with married mothers who had otherwise similar backgrounds. The researchers disaggregated their findings by gender but by not race or ethnicity.

The researchers used an annual, nationally representative survey called Monitoring the Future, which asks respondents about positive and negative social behaviors, to evaluate the adolescents’ behavioral outcomes. The positive behaviors measured include participation in volunteering, school clubs, and school athletics. The negative behaviors measured include skipping school, damaging school property, fighting, stealing, hurting someone, and using substances such as marijuana, alcohol, cigarettes, and other illicit drugs.

While lacking the rigor of a randomized experimental evaluation, the study observed associations between exposure to the policy changes and the measured behaviors.[5]

The researchers found:

  • A sizeable increase in negative social behaviors for adolescent boys at risk of exposure to the policy changes. Compared to youth not at risk of exposure, exposed youth were 7.2 percent likelier to skip school, 18.1 percent likelier to damage school property, and 20.8 percent likelier to get involved in a serious fight in school or at work (see Figure 1). Among adolescent girls, the researchers found one negative outcome of those they evaluated: exposed youth were 6.5 percent likelier to skip school than unexposed youth.
  • No evidence that exposure to the policy changes boosted positive behaviors among adolescent girls or boys. In fact, exposed adolescent girls were 6.1 percent less likely to volunteer than unexposed adolescent girls.
  • Increased substance use among exposed adolescent boys and girls. Exposed adolescent boys were 5.8 percent likelier than unexposed boys to use any substance. [6] They also were 8.4 percent likelier to smoke cigarettes, 12.8 percent likelier to use marijuana, and 16 percent likelier to use other illicit drugs. Exposed girls also had an increased likelihood of using any substance (4.2 percent), along with smoking cigarettes (10.6 percent) and marijuana (13.2 percent), relative to girls in the unexposed group. The rates of marijuana use were especially high for both genders in the exposed group relative to the unexposed group.

End Notes

[1] Dhaval Dave et al., “Effects of Maternal Work Incentives on Adolescent Social Behaviors,” NBER Working Paper No. 25527, February 2019, p. 39,

[2] Ibid.

[3] The researchers use the phrase “welfare reform” throughout their report to describe changes in cash assistance through AFDC waivers and TANF implementation, such as harsh work requirement penalties and federal time limits.

[4] The researchers classified an AFDC waiver as significant if it substantially altered the nature of the time limits, work requirements, earning disregards, sanctions, and/or family caps in the state AFDC program.

[5] Data limitations restrict researchers’ ability to study the effects of AFDC waivers and TANF implementation on the behaviors in this study. Accordingly, the study lacks the rigor of randomization and focuses instead on the timing of states’ AFDC waivers and TANF implementation. Consistent with several previous studies of the AFDC and TANF programs, the authors used the year-by-year variation in when AFDC waivers and the TANF program took effect in each state over a span of years as a rough indicator of policies such as work requirements and time limits on receiving cash assistance. Rebecca M. Blank, “Evaluating Welfare Reform in the United States,” NBER Working Paper No. 8983, June 2002,

[6] The Monitoring the Future survey includes the following substances: marijuana, alcohol, cigarettes, and other illicit drugs. The survey asked respondents how frequently they had used the substance in the past 30 days.