Association for Public Policy Analysis & Management

This paper investigates how changes in the accessibility and generosity of TANF and SNAP benefits from 1997 to 2015 have affected child poverty rates. Specifically, we measure whether increases in SNAP participation and benefit levels have ‘offset’ the income losses due to the decline of TANF throughout the past two decades. Our indicators of accessibility measure the participation rate of TANF (and SNAP) among all income-eligible households in each state-year. We utilize the TRIM3 micro-simulations to estimate whether each household in the CPS ASEC is eligible for TANF (or SNAP) benefits and the likelihood that each household receives the benefit. Our indicators of generosity capture state-year policy changes to the level of TANF (or SNAP) benefits. Results from a longitudinal two-way fixed effects model suggest that the steady decline of TANF participation among eligible households after the 1996 ‘welfare reform’ has contributed to a higher likelihood that a child lives in poverty. However, increases in SNAP benefits and take-up rates during the same period more than offset the poverty-reduction effect of TANF’s decline. In a counterfactual analysis, we show that if TANF accessibility and generosity in 2015 were the same as its 1997 levels, the estimated share of children living below 50 percent of median income in 2015 would fall by 6.4 percent (16.9 to 15.8 percent), though the share of children living below 10 percent of median would fall by 58 percent (0.43 to 0.18 percent). The results suggest that SNAP benefits have largely filled the social assistance gaps after TANF’s decline, but child poverty rates could nonetheless be far lower if TANF were accessible today as it was immediately following welfare reform.