Spending Inequalities for Children's Programs in Texas

Abstract

Although researchers and reformers have attacked inequalities in school expenditures throughout this century, funding patterns in noneducational programs for children and youth have been generally ignored. This article presents measures of expenditure inequalities among the 254 counties in Texas for a variety of income support, nutrition, child care, and juvenile justice programs and for several social and psychological services. In every case nonschool spending differences are greater than variations in school spending. The intrastate variations result from a variety of mechanisms, some of which are difficult to ascertain and others difficult to correct, but the magnitude of the inequalities suggests that the problem is too serious to ignore.

Get full access to this article

View all access options for this article.

References

1. See, for example, Stephen Carroll with Millicent Cox and William Lisowski , The Search for Equity in School Finance: Results from Five States, Rand Report R-2348-NIE, Mar. 1979. For a brief history of school finance reform, see W. Norton Grubb and Stephen Michelson, States and Schools: The Political Economy of Public School Finance (Lexington, MA: Lexington Books, 1974). ch. 2.

2. Michael Kirst , Walter Garms , and Theo Opperman , “State Services for Children: An Exploration of Who Benefits, Who Governs,” Public Policy, 28:185-206 (Spring 1980).

3. See the results from the Gary income maintenance experiments published in the Journal of Human Resources, 14 (Fall 1979).

4. Community Services Administration, Geographic Distribution of Federal Funds in Texas, 1978 (Washington, DC: U.S. Government Printing Office, 1979). Some of the figures are prorated across counties by population or some other proxy, and so give no real information on county spending patterns.

5. In general, the variance of AFDC funds per eligible person will be a function of both the variances of the two components and their covariances. If we represent equation 2 as f= b.p, or a funding rate as the product of a benefit rate and a participation rate, then an approximation is var f = b̄ var p + p̄ var b + 2 b̄ p̄ cov (b,p) Leo Goodman, “On the Exact Variance of Products,” Journal of the American Statistical Association, 55:708-713 (Dec. 1960). In the Texas sample, benefit rates and participation rates tend to be positively correlated: the population-weighted correlation coefficients are .249 for AFDC individuals, .379 for AFDC families, and .131 for food stamp recipients.

6. For tests of different measures of variation, see Robert Berne , “Alternative Equity and Equality Measures: Does the Measure Make a Difference?” in Selected Papers in School Finance, 1978 (Washington, DC: U.S. Office of Education, 1978), pp. 1-56.

7. The weight used is the number of children under 20. Since all measures of population size are so highly correlated across counties, it makes no real difference to the results whether the weight is the number of children or the population.

8. See the Texas State Plans for the WIC program, 1978, 1979, and 1980.

9. That is, we plan to use the residuals, from regressions of service measures on indicators of need, as measures of inequity. Some indicators of need will also include composite variables from a factor analysis of a variety of social indicators. This method has been suggested in Leonard Kogan and Shirley Jenkins , Indicators of Child Health and Welfare: Development of the DIPOV Index (New York: Center for Social Research, CUNY, 1974).