The U.S. federal government funds several large food assistance programs, including the Supplemental Nutrition Assistance Program (SNAP, formerly the Food Stamp Program), the National School Lunch Program (NSLP), the School Breakfast Program (SBP), and the Supplemental (WIC). In FY 2011, the federal costs for these programs were $75.3 billion, $11.3 billion, $3.0 billion, and $7.2 billion, respectively. The programs all share a common goal of helping disadvantaged people get access to more nutritious food than they could otherwise afford, but the programs differ in how they try to achieve that goal. The SNAP increases food resources for entire families, while the other programs only benefit specific members. In particular, the NSLP and SBP are targeted at school children, and WIC is targeted at infants, very young children, and pregnant, postpartum, and breastfeeding women.
A natural question regarding the targeted programs (natural to an economist anyway) is "do they benefit other members of a household?" Consider the WIC program. WIC provides its beneficiaries with vouchers that can be used for specific foods--for example, juice, milk, regular cereal, eggs and legumes for beneficiaries who are young children or mothers. In a household that also includes older children, it's possible (and maybe even reasonable to expect) that those children might consume some of the food that was intended for their mother or younger siblings.
The SBP and NSLP programs are different in that they serve meals outside the home and directly to their beneficiaries, so other household members can't literally take the food out of the school children's mouths. However, families can redirect resources. If participating families would have otherwise spent money on school-day breakfasts or lunches for their children, they could redirect that money to other members' food consumption.
A UNCG Ph.D. student, Jonathan Woodward, and I received a small grant from the U.S. Department of Agriculture to look at these issues, using 2002-3 survey data from the Child Development Supplement (CDS) of the Panel Study of Income Dynamics (PSID). The CDS is really helpful in this regard because it asked 1,744 children aged 10 or older about the foods they usually ate for breakfast and the foods that they had eaten in all meals during the previous week. The CDS and the PSID also asked about households' participation in each of the food assistance programs described above and about other economic and demographic characteristics of the households.
Results from Jonathan's and my research were just published in the open-access journal, Food and Nutrition Sciences. We found evidence that was consistent with targeted food assistance being shared. Specifically, school children who had mothers or younger siblings who received WIC reported drinking more milk and eating more cereal (foods that are in the WIC basket) for breakfast than other school children but also reported eating less toast (a food that isn't in the WIC basket). School children also benefited from the programs that were supposed to help them, generally eating more if they got school meals or if their families received SNAP.
An implication from this research is that it is really hard to target benefits for specific family members. In the case of WIC, we might consider whether it's really worth operating a separate program to do this or whether we should just fold the program into the larger SNAP. Merging the programs would allow for administrative efficiencies for the government and for households. Part of the cost savings could be used to expand benefits and participation among the intended recipients of WIC (WIC isn't able to help all eligible mothers and children now). Part of the savings could also go to deficit reduction. Families would also probably appreciate this because it would give them more flexibility in purchasing food.