Researchers have long observed that the simultaneous phaseout of benefits under multiple means-tested transfer and tax programs as household income increases can, in theory, result in shockingly high effective marginal tax rates for some low-income households (Ellwood & Liebman, 2000; Holt & Romich, 2007). However, those observations are typically based on static calculations or cross-sectional data and thus do not reflect the dynamics of actual income and benefit changes. Although tax benefits adjust immediately to household income changes, transfer benefits do not. Under program laws and regulations, households may not realize benefit reductions in response to an income increase for as long as 24 months under Medicaid; 12 months under the Children’s Health Insurance Program (CHIP), the National School Lunch Program (NSLP), and the Supplemental Nutrition Assistance Program (SNAP); and 6 months under the Special Supplemental Nutrition Program for Women, Infants and Children (WIC).

Although some have estimated the dynamics of income and benefit changes using simulations (Dickert, Houser & Scholz, 1994; Keane & Moffitt, 1998; Leguizamon, 2012; Maag et al., 2012) or small samples (Romich, 2006), large sample estimates of these dynamics are notably absent from the literature. We address this gap by using full panel data from the 2008 panel of the Survey of Income and Program Participation. Our main analysis sample includes 483,666 month observations from 22,795 persons in 8,012 low-income households without any elderly or disabled members.

We first examine whether benefits continue to be affected by prior incomes for as long as the program rules permit. We find that these benefit continuations in fact persist even longer under all of these programs. Households continue to receive significantly higher benefits after an income increase for at least 35 months under Medicaid, 33 months under CHIP, 31 months under NSLP, 27 months under SNAP and 9 months under WIC, after controlling for the applicable poverty line and for household income in each intervening month.

We then consider how these benefit continuations affect the high marginal tax rates often associated with benefit phaseouts under these programs. We concentrate this analysis on households that are typically thought to face the highest such rates: single households with children and annual household incomes between $15,000 and $35,000, for whom the theoretical rates average about 80%. We find that benefit continuations under SNAP and Medicaid, in particular, do significantly mitigate the high marginal tax rates for these households. Overall, for each dollar income increase, the households receive total additional benefits of $0.37 under these five programs over the subsequent 6 months, $0.68 over the subsequent 12 months, and $1.22 over the subsequent 24 months, after controlling for the applicable poverty line and for household income in each subsequent month. Thus, benefit continuations under these five transfer programs almost entirely eliminate the theoretically high marginal tax rates during the first six months after an income increase and reduce those rates by about 75% during the first two years after an income increase.