Skip Navigation

This memorandum presents information on federal spending in the major human resources programs (see Table 1 for a list of the programs considered here, which account for 93 percent of all human resources spending). It is divided into two sections. The first gives program-by-program estimates of spending if current policies were continued (the baseline), as well as that projected under the policies of the First Concurrent Resolution on the Budget--Fiscal Year 1984. The second presents estimates of how spending for a subset of these programs is distributed among households in different income categories. The latter analysis is subject to a number of important caveats, however, which are discussed below.
 

PROJECTED SPENDING BY PROGRAM

Three different projections are presented in Tables 1, 2, and 3, respectively. Table 1 gives baseline spending--that is, spending that would result if current policies were continued, including increases in funding for appropriated programs to offset projected inflation. The CBO's February 1983 economic assumptions are used, but the projections have been updated to reflect legislative actions through July 31, 1983. Although the CBO has since released newer economic assumptions, the February 1983 ones are used here, as requested, so that this analysis is consistent with the memorandum prepared for the Speaker of the House of Representatives entitled "Major Legislative Changes in Human Resources Programs Since January 1981."

Table 2 is a projection of spending under policies specified in the First Resolution. This projection is based on the economic assumptions embodied in the Resolution, but excludes the reserve fund.

Table 3 is the same as Table 2, except that it includes the reserve fund outlays for 1984 and 1985. It excludes the reserve fund outlays for 1983, since these funds have not been authorized.

The tables indicate that, despite the significant cuts enacted in 1981 and 1982, total outlays for these programs are projected to grow. Under the 1983 baseline projections, outlays are projected to increase at an annual rate of 7.3 percent compared with a 4.7 percent per year increase in the GNP deflator. Under the policies of the First Resolution, the rate of increase would be 7.2 percent per year without the reserve fund, or 7.4 percent per year with the reserve fund.

An important factor in the growth in spending is rapid increases in medical care costs. Outlays for medical programs are projected to grow at a 13.3 percent annual rate under the 1983 baseline. The growth rate for the remaining programs is 5.7 percent per year--one percentage point above general inflation.

Social Security and Medicare dominate federal spending for human resources programs. Under the 1983 baseline, they will account for 62 percent of spending in 1985 in programs listed in Table 1. In contrast, means-tested programs play only a limited role in human resources spending--accounting for just under 19 percent of the programs examined here by 1985.

The policies specified in the First Resolution, while of importance to certain individual programs, would have only a small effect on these proportions. For example, with the reserve fund, means-tested programs would account for 19.2 percent of spending, compared to 18.6 percent under the baseline, by 1985.
 

DISTRIBUTION OF SPENDING TO HOUSEHOLDS IN DIFFERENT INCOME CATEGORIES

This section reports on estimates of how federal spending for human resources programs is divided among households in different income categories. Considering the caveats discussed below, the results should be used with caution.

The Analytic Approach

Federal outlays for selected human resources programs have been distributed over five categories of households. Households are assigned to categories on the basis of the amount of cash income they received from all sources--including, for example, earnings, AFDC, and Social Security benefits. To maintain a roughly constant composition of households in the categories over time, the brackets used have been defined in constant 1982 dollars (the deflator being per household personal income), although the outlays are expressed in current dollars. Consequently, per household distributions of outlays will appear high relative to incomes, especially for 1985. Although in-kind benefits have been excluded from the process of sorting households into income categories, the distribution of both cash and in-kind benefits is shown on the tables.

There are two major reasons for this approach. First, benefits provided in-kind by those other than the federal government--for example, employer-provided fringe benefits--cannot be included in the measure of income used to categorize households, and the inclusion of such benefits only when they are federally provided would be inconsistent. Second, there is no universally agreed upon way to compare the value of in-kind benefits to recipients with that of cash benefits.

The analysis allocates outlays, rather than benefits, to households in different income categories. This means that if spending goes to reimburse providers for services, such as in the Medicare program, the analysis accounts for what the federal government paid providers, rather than what the services might be worth to the beneficiaries. Consistent with this rule, administrative costs are included in the analysis.

For several reasons, some outlays--accounting for 11 percent of those for 1983 in Table 1--had to be excluded from this part of the analysis. First, several programs were excluded because of lack of data on the incomes of households that benefit from them. In all cases, these programs do not provide benefits directly to individuals, but make grants to states or other organizations, which provide services.¹

Second, outlays for persons who are institutionalized and who are not part of a household with a noninstitutionalized person have been excluded. Since such persons are not covered by the Current Population Survey, the primary data base used for this analysis, including them would have delayed completion substantially, without changing the conclusions in a significant manner. In the estimates provided on a per-househoid basis, these units have also been excluded from the denominator (the total number of households). Third, in the Medicare program, voluntary premiums paid by beneficiaries for Supplementary Medical Insurance (Part B) have been subtracted from outlays for this analysis. In other words, these premiums are considered to offset an equal amount of benefits paid for by Medicare. Finally, outlays going to individuals residing in Puerto Rico and U.S. territories were excluded because the CPS does not survey households there. Similarly, aliens other than permanent residents were excluded.

For allocation among households in different income groups, outlays have been divided into three categories--outlays from programs providing cash benefits, outlays from medical programs, and outlays from programs providing other in-kind benefits. The separation between cash and in-kind benefits reflects the difficulty of valuing the latter from the perspective of the recipient. In-kind benefits are often valued more or less highly by the recipient than an equivalent cash benefit. For example, housing assistance provided through the public housing program may be worth less to recipients than cash benefits, because to receive the aid the households must live in designated projects.

Programs providing medical care have been separated from other programs providing in-kind benefits to avoid a problem of inconsistency with the allocation of program reductions in the previous memorandum. In that study, reductions in reimbursements to hospitals were excluded from the allocations because they did not affect individuals directly. However, virtually all outlays in Medicare and Medicaid are reimbursements to providers. Consequently any comparison of the distribution of reductions in spending for medical programs with total spending in those programs would be misleading. With minor exceptions, this problem is limited to medical programs, however.

Limitations of the Analysis

A number of important caveats about this part of the analysis need to be discussed, some of which are covered in more detail in earlier CBO studies of the distributional impacts of tax and benefit changes.²

First, a breakdown of the distribution of outlays by income category may be misleading in some respects. For example, average household size varies somewhat by income group; thus, a distribution of benefits by income category will not take into account differences in the relative needs of the different sized households.

Second, although the overall distribution of households stays fairly constant over time, individual households may move among groups relatively often. Such movement may be particularly likely for households in the bottom category, which contains a comparatively large share of single persons attending school. Similarly, events such as marriage, divorce, sustained unemployment, and retirement may also cause households to move among economic groups over time. Thus, the allocation of federal outlays for human resources programs to an income category may correspond with benefits received by a changing group of persons over time.

Third, the estimates presented here represent average outlays for benefits received by households in a given income category. Individual households are likely to receive benefits differing greatly from the average. For example, only a minority of households in the lowest income category receive Aid to Families with Dependent Children or Medicaid. Only households with college students are potentially eligible for higher education assistance programs.

Finally, a general caution in interpreting the per-household allocations is that they are not a good measure of the "average" increase in well-being. As indicated above, some individual households within a category receive no benefits from any program, while some recipients gain from several different programs. Also, some of the in-kind benefits, while very valuable to the recipient, do not redress problems of poverty. For example, an expensive episode of medical care treatment paid for by Medicaid may save a life, but nevertheless leave the family with insufficient resources to purchase other goods and services. Similarly, a low-income household may benefit enormously from a Pell grant that makes the difference between a child attending college and not, but the household members not attending will have the same difficulties in obtaining adequate food and housing.

The Distribution of Spending

The distribution of spending by household category differs among programs providing cash benefits, medical benefits, or other in-kind benefits. In fiscal year 1985, 70 percent of the outlays included in this part of the analysis are for cash benefit programs, 21 percent for medical programs, and 8 percent for other programs delivering in-kind benefits (see Table 4 for a list of the programs).³

For baseline spending under cash benefit programs in 1985, 33 percent of outlays go to households in the under $10,000 income category, and 30 percent and 27 percent go to households in the $10,000-20,000 and $20,000-40,000 categories, respectively (see Table 5).⁴ The remaining 10 percent of outlays go to households with incomes above $40,000.

Outlays for medical programs are more concentrated in the lower-income categories of households. The under $10,000 category accounts for 47 percent of 1985 outlays, for example. This concentration results from two factors. First, means-tested programs providing medical benefits comprise a higher proportion of outlays for medical programs than means-tested cash benefit programs do for cash benefit programs. Second, benefits from the non-means-tested cash benefit programs--in particular, retirement programs--tend to increase with income, whereas average outlays for Medicare are relatively level, and actually decline slightly with household income.

Outlays for other in-kind benefit programs are the most concentrated in the lowest-income categories. Households in the lowest-income category account for 73 percent of outlays for other in-kind benefits, and in the lowest two categories, for 89 percent. This concentration reflects the fact that most other in-kind benefit programs included in this analysis are means-tested.

Table 6 represents the distribution of baseline outlays on a per-household basis. As noted earlier, a general caution in interpreting the per-household allocations is that they are not a good measure of the "typical" household's increased well-being, especially for the in-kind benefits. Like the data in Table 5, the degree of concentration of outlays in the lower-income categories of households varies markedly by type of benefit.

Cash benefits are less concentrated among the lower-income categories than medical or other in-kind benefits. The average household in the under-$l0,000 category gets the highest amount of outlays for cash benefit programs--$4,300 in 1985. This compares to $3,550 in the $10,000-$20,000 category, $2,300 in the $20,000-$40,000 category, and $1,550 in the $40,000-$80,000 category.

Average benefits in the above-$80,000 category--at $3,000 in 1985--are, however, twice the amount for those in the $40,000-$80,000 category. This increase reflects three factors. First a large proportion of cash benefits go to persons aged 65 and over, and households in this age group comprise a smaller proportion of those in the $40,000-$80,000 category than in the $80,000 and over category. Second, both the Social Security and Civil Service Retirement systems base benefits on past earnings. Since high-income households receiving benefits from those programs tend to have had high earnings before retirement, their benefits from these programs tend to be relatively high. Third, the number of households in the above-$80,000 category is very small (about 1 percent of the total), so the accuracy of estimates for this group is relatively low.

As in the distribution of outlays to household categories, the per-household outlays for medical care and other in-kind benefits are more concentrated in the lower-income categories than are outlays for cash benefit programs. Under the baseline, the average household in the under $10,000 category would receive medical services that cost the federal government $1,900 in 1985, compared with an average of $900 for all households. For other in-kind benefits, households in the lowest category would receive services costing $1,150, compared with $350 for all households.

The distribution of spending under the First Resolution is very similar to that under the baseline (see Tables 7 through 12). In most cases, outlays projected to go to households in the lowest income category in 1985 round to the same number under all three sets of projections. In essence, the differences between the First Resolution and the baseline in spending or in spending per household across the five income categories are small because the differences in total outlays are small when measured in percentage terms (see Table 4).

This document is available in its entirety in PDF.

1. Programs listed in Table 1 that were excluded from the analysis of the distribution of spending are WIC, other health services, Compensatory Education, Head Start, Vocational Education, Community Service Block Grants, Social Services Block Grants, Veterans' Readjustment Benefits, General Employment and Training, Job Corps, Public Service Employment, and the Work Incentive Program. This is the same list of programs that were excluded from the distributional analysis summarized in the earlier memorandum.

2. The most detail is in Congressional Budget Office, "Effects of Tax and Benefit Reductions Enacted in 1981 for Households in Different Income Categories," Special Study (February 1982).

3. These percentages do not sum to 100 percent because of rounding.

4. Note that, for all years examined here, the income categories are described in terms of the dollar divisions in 1982. The divisions for future years expressed in dollars of those years would be higher. For example, the divisions between categories for 1985 expressed in 1985 dollars would reflect three years of growth in personal income.