This work uses multilevel models to estimate the effects of individual characteristics, local economies, and state policies on Temporary Assistance for Needy Families (TANF) participation among unmarried mothers with 1-year-old children. It also assesses the contributions of the different sets of factors to city variation in rates of TANF participation. Results suggest that the many predictive individual-level factors account for one-quarter of the between-city variance in TANF participation. Most local economic and state policy measures are not associated with TANF participation, but those that are, together, explain another quarter of the between-city variance. The results will help to inform policy decisions at the state level, to design programs and policies in urban areas, and to understand the factors that promote economic self sufficiency.