Small-Area Estimates of School-Age Children in Poverty: Evaluation of Current Methodology (2000)

This chapter summarizes the procedures used to allocate Title I funds, describing features that need to be considered when evaluating the reliability and appropriateness of the Census Bureau's SAIPE estimates of poor school-age children for use in the allocations. Following a summary description of the Title I formulas, the chapter describes the two-stage procedure that was used for county and school district allocations from the inception of the program in 1965 through the 1998-1999 school year and the direct allocation procedure that was first used for school district allocations for the 1999-2000 school year.¹

Title I allocations are based on estimates of formula-eligible children, which comprise four groups: related poor school-age children, as estimated by the Census Bureau; children in foster homes; children in families above the poverty level that receive Temporary Assistance to Needy Families (TANF);² and children in local institutions for neglected and delinquent children. The Census Bureau's estimates of poor school-age children account for about 95 percent of total formula-eligible children.

The statute contains four formulas for allocating Title I funds–basic grants, concentration grants, targeted grants, and the Education Finance Incentive Program–but Congress has to date appropriated funds only for the basic and concentration formulas. Basic grants have existed since the program began in 1965; concentration grants were added in 1978 to provide additional funds to school districts with high concentrations of school-age children in poverty (Moskowitz et al., 1993). The total amount of Title I funds allocated for the 1998-1999 school year was $7.3 billion–$6.2 billion for basic grants (85% of the total) and $1.1 billion for concentration grants; the total amount allocated for the 2000-2001 school year was $7.7 billion (U.S. Department of Education, 2000).

The basic grant formula has a low threshold for eligibility for school districts to receive funds: eligible districts must have at least 10 formula-eligible children and the number of eligible children must exceed 2 percent of the district's population aged 5-17. There were no minimum eligibility criteria for a county to receive a basic grant under the two-stage procedure.

In contrast, the concentration grant formula has a high eligibility threshold, allocating funds only to jurisdictions–counties and school districts for the two-stage procedure, school districts for the direct procedure–with high numbers or high percentages of poor school-age children. To be eligible to receive a concentration grant, a jurisdiction must have more than 6,500 formula-eligible children or more than 15 percent of the children in the jurisdiction must be formula eligible. Under the two-stage procedure, a school district could meet the eligibility threshold for a concentration grant but not receive funds because the county in which it is located was not eligible. States could reserve up to 2 percent of their concentration grants funds to allocate to such districts.

The two formulas take account not only of numbers of formula-eligible children in each jurisdiction, but also of each state's average per-pupil expenditure, a factor intended to compensate for state differences in the cost of education. A state minimum grant provision applies to each of the two formulas as well. For basic grants, the state minimum grant is equal to the lesser of (1) 0.25 percent of total funds available for Title I basic grants and (2) the average of 0.25 percent of total funds and 150 percent of the national average grant payment per formula-eligible child multiplied by the number of formula-eligible children in the state.³ There is some added complexity for the state minimum for concentration grants (Moskowitz et al., 1993).

Finally, the Title I formulas include hold-harmless provisions to cushion the impact of decreases in allocations. Historically, there was no hold-harmless provision for concentration grants, but for school year 1996-1997, the Title I

legislation specified a hold-harmless provision for both formulas at a rate of 100 percent: that is, a jurisdiction that met the eligibility threshold could not receive fewer funds than it had received in the previous school year. For school year 1997-1998 and beyond, the legislation specified for basic grants that the hold-harmless provision was to be applied at variable rates, with a higher rate for higher poverty jurisdictions: those with 30 percent or more poor school-age children were to be guaranteed at least 95 percent of the prior year's basic grant; the guarantee was to be 90 percent for those with 15-30 percent poor school-age children and 85 percent for those with fewer than 15 percent poor school-age children. For school year 1997-1998 and beyond, the legislation did not include a hold-harmless provision for concentration grants.

However, for allocations for school years 1998-1999, 1999-2000, and 2000-2001, Congress enacted a 100 percent hold-harmless provision for both basic and concentration grants. In addition, for school years 1999-2000 and 2000-2001, Congress stipulated that school districts that were no longer eligible for a concentration grant under the new direct allocation procedure using the Census Bureau's 1995 estimates of poor school-age children would nonetheless receive 100 percent of the previous year's concentration grant. Previously, a district had to meet the eligibility threshold (more than 6,500 or more than 15% formula-eligible children) to receive concentration grant funds. Districts continued to have to meet the much lower eligibility threshold to receive a basic grant.

Under the two-stage procedure used for Title I allocations through the 1998-1999 school year, the U.S. Department of Education determined the allocation amounts for each county, and the states then suballocated these amounts to the school districts in their state. The Department of Education calculated basic grant allocations in an iterative process. The number of formula-eligible children in each county was multiplied by 40 percent of the state's per-pupil expenditure;⁴ the resulting allocations were then proportionally reduced so that the total matched the total appropriation for basic grants. Allocations were then adjusted to meet hold-harmless and state minimum grant provisions (see above). The department calculated concentration grant allocations on the basis of numbers and proportions of formula-eligible children, the state's per-pupil expenditure, a state minimum provision, and a hold-harmless provision.

States used a variety of data sources for determining suballocations of Title I funds to school districts, as listed below for school year 1997-1998 from a chart provided by the department. The Department of Education approved each state's

allocation plan but not the specific estimates used by a state or the allocation amounts.

• Seven states and the District of Columbia made no suballocations to districts because their school districts are coterminous with counties (three of these states made suballocations to a few districts in their states that are not coterminous with counties, such as a city that is a separate district from the remainder of the county).

• Eight states used 1990 census data alone.

• Ten states used 1990 census data and estimates of the other categories of formula-eligible children, such as foster children.

• Nine states used a combination of 1990 census data together with counts of children approved to receive free meals or free or reduced-price meals under the National School Lunch Program, or counts of children in families receiving AFDC, or a composite of AFDC, food stamps, and Medicaid data.

• Eight states used free lunch data only.

• Three states used free and reduced-price lunch data.

• One state used free lunch and state tax information.

• Three states used AFDC data only or in combination with foster child data.

• One state used food stamp data.

Over time, it appears that a growing number of states elected to use free (or free and reduced-price) school lunch data to distribute the county allocations to school districts. Interviews conducted by panel staff in early 1999 determined that this trend was continuing: several states that had used AFDC or food stamp data for suballocations were found to have switched to using school lunch counts, although such data include children with family incomes above the poverty threshold. Children with family incomes up to 130 percent of the poverty threshold may receive a free school lunch; children with family incomes between 130 percent and 185 percent of the poverty threshold may receive a reduced-price lunch.

Under the two-stage procedure, most states were constrained to suballocate county amounts to the school districts (or parts of school districts) within each county. However, the Department of Education permitted nine states to make direct allocations of basic grants–but not concentration grants–to school districts without regard to the county allocation amounts because so many of their school districts crossed county boundaries. Of these nine states, one used 1990 census data to make direct allocations of basic grants; five used 1990 census data and estimates of the other categories of formula-eligible children; one used a combination of 1990 census and free and reduced-price lunch data; one used free lunch data; and one used free and reduced-price lunch data.

The Department of Education made direct allocations of Title I funds to school districts for the first time in spring 1999 for the 1999-2000 school year. As directed by the “Improving America's Schools Act” of 1994, the Department based these allocations on the Census Bureau's updated estimates of the numbers of poor school-age children at the school district level (which the panel determined were not unreliable or inappropriate for this purpose when compared with other estimates–see National Research Council, 1999) together with estimates of the other groups of formula-eligible children for school districts that it obtained from the U.S. Department of Health and Human Services and the states. In some cases, the department had to prorate county totals for these other groups to districts when no district-level information was available.

The Census Bureau provided three sets of school district estimates: (1) estimates of school-age children (aged 5-17) in 1996 who were living in and related to a family in poverty in 1995;⁵ (2) estimates of all school-age children; and (3) estimates of the total population of the district. The first two sets of estimates were needed to implement the allocation formulas for both basic and concentration grants. The third set of estimates was needed to identify school districts with fewer than 20,000 people. These school districts had to be identified because the 1994 act provided that states, at their discretion, could aggregate the fund allocations for districts with less than 20,000 population and redistribute the funds by using another method approved by the Department of Education.

Nine states applied and were granted approval to reallocate the department 's allocations for school districts with less than 20,000 population. All nine states had previously used data sources other than the decennial census, or in combination with the census, to suballocate the department 's county allocations under the old two-stage procedure. The data sources the nine states used for reallocating funds for small school districts included:

• Two states used the 1995 SAIPE estimates (weighted at one-half) and another source (weighted at one-half), either counts of children in families receiving public assistance in the state or free lunch counts;

• One state used the 1995 SAIPE estimates (weighted at three-fourths) and free and reduced-price lunch counts (weighted at one-fourth);

• One state used free lunch counts (weighted at one-half) and poor children estimated from state income tax data (weighted at one-half);

• One state used free lunch counts;

• One state used free lunch and free milk counts from public and nonpublic schools;

• Two states used free and reduced-price lunch counts;

• One state used the 1995 SAIPE estimates (weighted at 0.155), foster children counts (weighted at 0.155), free lunch counts (weighted at 0.46), and reduced-price lunch counts (weighted at 0.23).

All nine states used counts of other categories of formula-eligible children, such as children in locally operated institutions for neglected children, in addition to the sources listed above.