Collecting Expenditure Data
Once we have defined which programs to include, we col- lect spending data for each program, drawing primarily from the federal budget of the United States government (fiscal year 2011 and past years), its appendices, and special analyses for historical data and projections. Our database includes estimates of federal expenditures on children for five-year intervals from 1960 to 1995 and annually from 1996 to 2009. For most programs, we start with out- lay estimates from the Appendix to the Budget of the U.S. Government or, in the case of tax expenditures, from the Analytical Perspectives volume of the budget. The 2009 ex- penditure estimates released in this year’s report are based on detailed spending information released in February 2010 in the Appendix to the Budget of the U.S. Government for fiscal year 2011. All budget numbers presented in this report represent fiscal years and are expressed in 2009 dol- lars, unless otherwise indicated.
We also have estimates of state and local spending annually from 1998 to 2007, drawn from the Rockefeller Institute State Funding Database. Consultations between the authors of this report and researchers at the Rock- efeller Institute have increased consistency between the two sets of estimates, but some differences remain.4
Calculating Share Spent on Children
For programs that serve children only, we assign 100 percent of program expenditures (benefits and associated adminis- trative costs) to children, whether the expenditure is a direct service to children (e.g., education) or a child benefit paid through parents or guardians (e.g., SSI disabled children benefits). Where a program provides direct services to both children and adults (e.g., Medicaid), we calculate the percentage of program expenditures that goes to children. In the more difficult case where benefits are provided to families without any delineation of parents’ and children’s shares, we generally estimate a children’s share based on the number of children and adults in the family and assum- ing equal benefits per capita. For example, in a two-child, one-adult family, two-thirds of any public assistance to the family would go to the children and one-third to the adult.
We put significant effort into estimating the portions of large programs, such as SNAP/Food Stamps, Medicaid, or Supplemental Security Income, that go just to children. For these calculations, the most frequently used data sources
are the House Ways and Means Committee’s Green Book (various years), the Annual Statistical Supplement to the Social Security Bulletin (various years), reports from the agencies that administer the programs, and discussions with agency staff. We also rely on unpublished tabulations of administrative or survey data generated by ourselves or other researchers.
Our approach to defining spending on children and es- timating the children’s share of spending is summarized in figure 1, which also provides examples of specific programs included in our analysis. (For a full list of spending and tax programs, see table 1, which lists three dozen major pro- grams directly in the table and dozens of smaller programs in the notes.)
While we attempt to apply general rules consistently across all programs, many programs present specific chal- lenges, which are discussed in more detail in the Data Appendix to Kids Share 2010: Report on Federal Expen- ditures on Children through 2009, a separate publication that provides data sources and methodology used in estimating current and projected expenditures for each program included in our analysis.5 As an example, while we have a general rule of defining children as residents of the United States under age 19, we make exception to this rule for programs that collect programmatic data under different definitions (e.g., Social Security and Sup- plemental Security Income define children as under 18).
Changes in Methods in This Year’s Report
We made minor changes to the list of programs included in our analysis. Our largest addition was to expand the tax provisions benefitting children to include the new Quali- fied Bonds for School Construction, established under ARRA, as well as the pre-existing Qualified Zone Academy Bonds.When new programs are enacted or older programs are brought to our attention, we review the program eligi- bility rules to see whether the program’s benefits fit with our definition of spending on children, and we roughly estimate spending on children to determine whether the program has sufficient spending on children (at least $50 million in real dollars in any year) to justify the time need- ed to add to our analysis, including the historical database. This year, we dropped a few smaller programs to remain consistent with this general rule; again, these and other minor changes are detailed in the data appendix.
AN ANALYSIS OF FEDERAL EXPENDITURES ON CHILDREN