THE USE OF ECONOMIC DATA IN CHILD SUPPORT GUIDELINES:
THE UNITED STATES DEPARTMENT OF AGRICULTURE'S EXPENDITURES ON CHILDREN
Laura Wish Morgan
Child support guidelines build into the presumptive award certain economic assumptions about expenditures. In fact, the guidelines must be based on economic data concerning the cost of raising children:
As part of the review of a State's guidelines required under paragraph (e) of this section, a State must consider economic data on the cost of raising children and analyze case data, gathered through sampling or other methods, on the application of, and deviations from the guidelines. The analysis of the data must be used in the State's review of the guidelines to ensure that deviations from the guidelines are limited.
45 C.F.R. § 302.56(h). Thus, economic data form the basis of the economic elements of the guidelines: the percentages for Percentage of Income Model States; the schedules of support for Income Shares Model States; and the primary support levels and standard of living percentage for Melson Formula Model states. See Laura Wish Morgan, Child Support Guidelines: Interpretation and Application § 4.02 (Supp. 1998).
When the states began adopting their guidelines in 1988 through 1990, the most commonly used economic data underlying the state guidelines was the result of research conducted by Thomas J. Espenshade. See Thomas J. Espenshade, Investing in Children: New Estimates of Parental Expenditures (1984). Espenshade used 1972-73 Consumer Expenditure Survey data to estimate average expenditures for children in the home. This economic data was later adopted by Williams of the federal guidelines' advisory panel. Robert G. Williams, Guidelines for Setting Levels of Child Support Orders, 21 Fam. L. Q. 281, 289 (Fall 1987). The most important thing to note about this economic data at this time is that Espenshade himself does not advocate the use of his now-outdated figures. Barbara R. Bergmann & Sherry Wetchler, Child Support Awards: State Guidelines vs. Public Opinion, 29 Fam. L.Q. 483-486 (Fall 1995). See also David Betson, Alternative Estimates of the Cost of Children from the 1980-1986 Consumer Expenditure Survey at 57 (1990).
As the states began to go through their revisionary process, some states began to use economic data from other sources. For example, the Kansas Child Support Guidelines are derived from an economic model initially developed by William Terrell in 1987, updated in 1989 by Ann Coulson using more current data, and then adjusted for current economic data in 1993.
The annual reports put out by the United States Department of Agriculture (USDA), based on the Consumer Expenditure Survey of the Bureau of Labor Statistics, United States Department of Labor, may also be used to determine child support guidelines.
The United States Department of Agriculture recognizes that its data is used in the formation of child support guidelines. Upon release of the 1995 report on May 31, 1996, Secretary of Agriculture Dan Glickman stated, "This report has proven to an invaluable resources to states in determining child support guidelines and foster care payments." Mark C. Lino, Expenditures on Children by Families: 1995 Annual Report (United States Department of Agriculture, 1996). Further, a cover story in the March 30, 1998 edition of U.S. News and World Report, called the USDA study a starting point to estimate overall expenses on children.
Since 1960, USDA has provided annual estimates of expenditures on children by married-couple and single-parent families from birth through age 17. See Mark C. Lino, Expenditures on Children by Families: 1997 Annual Report (United States Department of Agriculture 1998). These expenditures on children are for the major budgetary components: housing, food, transportation, clothing, health care, child care and education, and miscellaneous goods and services (personal care items, entertainment, etc.). The latest child-rearing expenses are based on the 1990-92 Consumer Expenditure Survey (CE) updated to 1997 dollars using the Consumer Price Index (CPI). The CE is the only Federal survey of household expenditures collected nationwide. It collects information on socio-demographic characteristics, income, and expenditures of a nationally representative sample of households.
The methodology employed by USDA in determining child-rearing expenses specifically examines the intra-household distribution of expenditures using data for each budgetary component. The CE contains child-specific expenditure data for some budgetary components (clothing, child care, and education) and household level data for the other budgetary components. Multivariate analysis is used to estimate household and child-specific expenditures, controlling for income level, family size, and age of the child so expenses can be determined for families with these varying characteristics (regional expenses are also derived by controlling for region).
Household and child-specific expenditures are allocated among family members (i.e., in a married-couple, two-child family: the husband, wife, older child, and younger child). Since the expenditures for clothing, child care, and education apply only to children, allocations of these expenses are made by dividing them equally among the children. Because the CE does not collect expenditures on food and health care by family member, data from other Federal studies are used to apportion these budgetary components to a child by age. USDA food plans are used to allocate food expenses among family members. These plans, derived from a national food consumption survey, show the share of food expenses attributable to individual family members by age and household income level. These member food budget shares are applied to household food expenditures to determine food expenses on a child. Health care expenses are allocated to each family member based on data from the National Medical Expenditure Survey. This survey contains data on the proportion of health care expenses attributable to individual family members. These member budget shares for health care are applied to household health care expenditures to determine expenses on a child.
Unlike food and health care, no authoritative base exists for allocating household expenditures on housing, transportation, and other miscellaneous goods and services among family members. Two common approaches used in allocating these expenses are the per capita method and the marginal cost method. The marginal cost method measures expenditures on children as the difference in expenses between couples with children and equivalent childless couples. The method depends on development of an equivalency measure; however, there is no standard accepted measure. Various measures have been proposed, each yielding different estimates of expenditures on children. Also, the marginal cost approach assumes that the difference in total expenditures between couples with and without children can be attributed solely to the children in a family. This assumption is questionable, however. In addition, couples without children often buy homes larger than they need at the time of purchase in anticipation of children. Comparing the expenditures of these couples to similar couples with children could lead to underestimates of expenditures on children.
For these reasons, USDA uses the per capita method to allocate housing, transportation, and miscellaneous goods and services among household members. This method allocates expenses among household members in equal proportions. Although the per capita method has its limitations, these limitations were considered less severe than those of the marginal cost approach. It should be noted that for homeowners, housing expenses do not include mortgage principal payments; such payments are considered in the CE to be part of savings. Also, because transportation expenses resulting from work activities are not related to expenses on a child, these costs are excluded when determining children's transportation expenses.
There are other estimators to determine child-rearing expenses. Two of the most commonly used are the Engel and Rothbarth estimators. Both of these estimators are marginal cost approaches that are applied to overall expenses: expenses on children are measured as the difference between overall expenses of couples with children and equivalent childless couples. This difference is thought to represent additional or marginal expenditures couples make on a child. The two estimators differ in the equivalency scale they use to compare couples with and without children.
Almost all studies applying the marginal cost method apply it to overall expenses, for to apply it to individual expenses may produce illogical results. For example, couples with children often have lower food expenses than couples without children as the former group eats out less often, uses coupons, and purchases larger sizes that cost less per unit. In such a case, the marginal cost method, which would measure food expenses on children as the difference in food expenses between couples with children and those without children, would yield the improbable result that children have negative food expenses.
The Engel estimator assumes that if two families spend an equal percentage of their total expenditures on food they are equally well-off. The Rothbarth estimator uses the level of excess income available to people after necessary expenditures on family members are made as the equivalency measure. Rothbarth defined excess income to include luxuries (alcohol, tobacco, entertainment, and sweets) and savings.
Both estimators have their limitations, as previously explained. Each assumes a "true" equivalency measure. Neither of the equivalency measures, however, have been validated as the "true" measure in the economics literature. More importantly, the marginal cost estimators do not provide direct estimates of how much is spent on a child. They estimate how much money families with children must be compensated to bring the parents to the same utility level (as gauged by an equivalence scale) of couples without children. This is a different question than "how much do parents spend on children?"
According to Burt S. Barnow, an economist who has studied the issue of estimating expenditures on children, "While they (the Engel and Rothbarth estimators) undoubtedly yield biased estimates of the true level of expenditures made on behalf of children, the direction of the bias is believed to be known." Burt S. Barnow, "Economic Studies of Expenditures on Children and Their Relationship to Child Support Guidelines," in Child Support Guidelines: the Next Generation at 22-3 (Margaret Campbell Haynes, ed., United States Department of Health and Human Services 1994). He makes the argument that "the Rothbarth estimator is likely to provide a lower bound estimate of actual expenditures on children, while the Engel estimator is likely to provide an upper bound." Id. The precise magnitude of the overestimate in the case of the Engel estimator or the underestimate in the case of the Rothbarth estimator is unknown. Barnow states the Engel estimator yields results too high to be believed and so recommends the Rothbarth estimator be slightly increased to determine child-rearing expenditures.
How do child-rearing expenses derived from the Engel and Rothbarth estimators compare with USDA estimates? This comparison is shown in Table I by number of children and total household expenditures. The results for the Engel and Rothbarth estimators are from a 1990 study sponsored by the United States Department of Health and Human Services that estimated husband-wife child-rearing expenses based on the 1980-87 CE; this study contains the most recent child-rearing expenses using the Engel and Rothbarth approaches. USDA expenses are based on the 1995 study. In this study, average expenditures in families in each income level were used to make comparison. Percentages by number of children are based on average expenditures in middle-income families. Household expenditure level figures are for a family with two children. The comparison is based on child-rearing expenses as a percentage of total family expenditures so the figures did not have to be converted into real dollars.
The Engel and Rothbarth techniques yield varying child-rearing expenses, which differ as much as 20 percentage points for a family with three children. Thus, when using the marginal cost method in estimating expenditures on children, the choice of an equivalency measure is obviously critical since different measures yield different results. If Barnow is correct in that the Rothbarth technique is a lower-bound estimator of child-rearing expenses and the Engel technique is an upper-bound estimator, this gives credence to USDA child-rearing expenses because they are between those produced by the Engel and Rothbarth techniques. For families with one child and for families with a high expenditure level, USDA child-rearing expenses are closer to the Rothbarth estimates, whereas for families with a low expenditure level, USDA child-rearing expenses are closer to the Engel estimates. For families with two or more children and for families with an average household expenditure level, USDA child-rearing expenses are about in the middle of the Rothbarth and Engel estimates.
It is sometimes argued that the USDA method overestimates child-rearing expenses since the per capita method is used to allocate housing, transportation, and miscellaneous expenses among household members. These three budgetary components account for about 60 percent of USDA child-rearing costs. One study argues that child-related housing expenses should be measured as the difference in rent between one- and two-bedroom apartments. Donald J. Bieniewicz, "Child Support Guideline Developed by Children's Rights Council," in Child Support Guidelines: The Next Generation at 108 (Margaret Campbell Haynes, ed., United States Department of Health and Human Services 1994). This argument assumes all children will reside in rental property. Housing expenses on an only child in a lower-income and middle-income family for the overall United States are estimated by USDA to be about $205 and $285 per month, respectively, in 1996. This includes the cost of shelter as well as utilities, furnishings, home insurance, and appliances. According to the Census Bureau, the difference in median rental price between an efficiency/one-bedroom housing unit and a two-bedroom housing unit was about $100 per month in 1996 dollars. Mark C. Lino, Do Child Support Awards Cover the Cost of Raising Children?, 11 Family Economics and Nutrition Review No. 1/2 at p. 29 (1998). This does not include cost for utilities for many units, furnishings, insurance, or appliances. Also, the USDA child-rearing housing expense includes homeowners and renters; housing costs for homeowners are typically higher than the costs for renters, as owned housing usually has more space than rental housing. The alternative, using a marginal cost procedure, could lead to severe underestimates of housing expenses on a child, for as previously explained, many couples without children purchase a home in anticipation of having children. Hence, these couples without children would have housing expenses similar to couples with children.
As for transportation expenses, USDA child-rearing expenses do not include the 40 percent of total transportation expenses deemed to be work-related. Miscellaneous expenses include expenditures on personal care (toothpaste, haircuts, etc,), entertainment (video cassettes, toys, etc.), and reading material (books, magazines, etc.). Many of the goods and services in this category are child-oriented, so a per capita approach seems reasonable in allocating these expenses. Based on some of the goods and services that are included in this category, it could be argued that children consume more than a per capita share of these expenses. Therefore, it is not likely that USDA child-rearing expenses provide gross overestimates of expenditures on children for housing, transportation, and miscellaneous goods and services.
If child support guidelines can adequately reflect the cost of raising children by incorporating this data, then at least one of the goals of the guidelines will be met: adequacy of awards.
Looking for previous articles? Check the archive.