Paid Work and Unpaid Caregiving Under Guaranteed Income: Evidence from Georgia's Pilot

Background and Purpose: Policy conversations around guaranteed income (GI) are increasingly active in the United States. Since 2018, over 150 local GI pilots have launched across more than 35 states, varying in benefit amount, duration, and eligibility—spurred in part by the Mayors for a Guaranteed Income initiative in 2021 (Stanford Basic Income Lab, 2025). One such effort, In Her Hands (IHH)—Georgia’s first GI pilot—began in summer 2022, providing monthly unconditional cash transfers ($850 on average) to 650 low-income Black women in the Atlanta Metro Area (Brugger et al., 2024). While much debate around GI has focused on labor effects (Widerquist, 2005), little attention has been paid to unpaid labor, such as caregiving. This study examines changes in paid labor and unpaid caregiving patterns among IHH participants.

Methods: The IHH pilot employed a randomized controlled trial with assigned treatment and control groups. Participants completed demographic questionnaires at the application stage. The treatment group was surveyed at five time points: before receiving GI (Wave 0), and at 6, 12, 18, and 24 months (Waves 1-4). The control group was surveyed at Waves 2 and 4 only. Our analytic sample included Black women with children and household incomes below 200% of the federal poverty line at the application stage (N=599). We conducted three sets of analyses. First, we explored within-group trends (Analysis 1) by conducting paired t-tests that compared treatment group averages for paid labor and unpaid caregiving hours at Waves 1-4 against baseline (Wave 0). Second, we conducted t-tests (Analysis 2) to compare treatment and control groups’ average labor and caregiving hours at Waves 2 and 4. Finally, we conducted a factorial ANOVA to examine whether changes in paid work hours were associated with changes in unpaid caregiving hours across groups between Waves 2 and 4 (Analysis 3). Inverse probability weighting, based on application-stage demographics, adjusted for baseline differences in Analyses 2 and 3.

Results: Analysis 1 found that participants in the treatment group significantly increased paid work hours at Wave 3 (t = 2.37, p < 0.05) and Wave 4 (t = 3.13, p < 0.01), while caregiving hours decreased across all waves. In Analysis 2, the treatment group reported fewer paid work hours than the control group at Wave 2 (t = -2.45, p < 0.05), but this difference disappeared by Wave 4. Analysis 3 showed no significant differences between treatment and control groups in caregiving hour changes based on paid labor hour shifts.

Conclusions and Implications: These findings in part challenge the dominant narrative that unconditional cash programs discourage paid labor. Among urban low-income Black women with children, GI did not reduce paid labor, particularly at later Waves. At the same time, reductions in unpaid caregiving were not directly tied to changes in paid work. Further research is needed to assess longer-term impacts of GI on labor dynamics, particularly across larger samples and diverse geographic settings.