We investigate the relationships between participation in the School Breakfast Program (SBP), National School Lunch Program (NSLP), Supplemental Nutrition Assistance Program (SNAP) and childhood obesity. Our focus is on the effects of simultaneous participation in multiple programs while attempting to address non-random selection into each program. Using Early Childhood Longitudinal Study, Kindergarten Class of 1998-99 data we find: (1) there exists evidence of selection into the SBP and NSLP on the basis of unobservable attributes; (2) among low-income students this non-random selection disappears once we condition on school fixed effects; (3) participation in the NSLP alone contributes to childhood obesity in low-income households; and, (4) simultaneous participation in all three programs (relative to participation in no programs) is unrelated to childhood obesity in low-income households. That said, we show that identification of the causal effects of participation in different combinations of these programs is quite difficult since participation decisions in the three programs are highly correlated.