Archived Information

Financing Postsecondary Education: The Federal Role - October 1995

Rethinking the Allocation of Pell Grants

David W. Breneman and Fred J. Galloway

Two facts motivate this analysis:

Our purpose is to explore alternative ways of allocating this $6 billion to regain some of the earlier promise of the program.

Before turning to the analyses, a few words about the underlying economics of the Pell Grant program are in order. Currently, nearly 4 million students receive Pell Grants, but most grants are not for the maximum award. The amount awarded to each student is based on both income and assets, and it is further reduced by the need to bring total eligibility into line with funds appropriated--Pell Grants are not true entitlements, but are based upon annual appropriations.

While the maximum award authorized for 1995-96 is $4,100, the maximum award funded is much lower--$2,340 this year. Given the large number of recipients, it costs roughly $320 million to increase the maximum award by $100. Thus, nearly $6 billion in additional spending would be required to fund the authorized maximum award, a sum not likely to be forthcoming. Were we living in a more perfect world, we would advocate full funding for the program, maintaining the promise of access. Given the world as we find it, however, we think it imperative to examine more effective allocations of the current $6 billion in order to ensure as much access as possible with that sum.

Baseline Projections and the Current Allocation Formula

The tables that follow are simulations derived from the Pell Grant cost estimation model at the American Council on Education. This microcomputer-based model was developed by the Department of Education and provides information on the cost of various programmatic changes to the Pell Grant allocation formula.1 In addition, the model also provides for the distribution of student awards by income level, dependency status, and institutional type. To generate cost estimates for the out years, the model uses the economic assumptions developed for the 1996 budget by the Office of Management and Budget.

Before proceeding to the various cost estimates, a brief discussion of the current allocation formula is in order. For each potential Pell Grant recipient, available income and discretionary net worth are first calculated based on the recipient's total income and total net worth, in a manner that sets aside certain sums for living expenses and other purposes.2 In the next step, assessment rates are applied to available income and discretionary net worth to produce contributions from both sources. These contributions are then summed to produce the "expected family contribution" or EFC. Once the EFC has been determined, it is then used in separate formulas to produce two potential award levels, with the smaller of the two award levels being the one the student actually receives.3

To calculate the annual cost of the Pell Grant program, the total value of all these awards is then added to an administrative allowance to produce the total annual program cost. In our analysis, we use this methodology to provide baseline cost estimates for the five-year period from 1995-96 through 1999-2000 under two distinct scenarios. The first scenario assumes that the current $2,340 maximum grant is unchanged over the five-year period, while the second assumes a $2,340 maximum for 1995-96, increasing to $2,620 for the subsequent four years, as presented in the President's FY 1996 budget proposal. The results presented in Table One show that, under the first scenario, budget outlays over the five-year period total $31.05 billion, while under the second scenario, outlays total $34.75 billion.

Table One: Baseline Cost Estimates for 1995-96 to 1999-2000 (Cost in Millions)
Award Year Scenario One
$2,340 Maximum
All Years
Scenario Two
$2,340 Maximum 95/96
$2,620 Maximum other years
1995/96 $5,990
$5,990
1996/97
6,085
6,982
1997/98 6,209 7,125
1998/99 6,322 7,257
1999/00 6,445 7,396
5 Year Total $31,051 $34,750

The Alternative Allocation Formulas

Seven alternative allocation formulas are considered in this paper:

  1. "Frontloading" the awards, by which we mean limiting Pell Grants to first- and second-year students only;

  2. Exclusion of all students attending proprietary institutions;

  3. The combination of (1) and (2), i.e., frontloading plus exclusion of proprietary school students;

  4. Targeting awards on lower-income students and families by raising the assessment rates on income;

  5. Exclusion of students enrolled in less than one-year programs;

  6. Exclusion of students enrolled in less than two-year programs; and,

  7. Eliminating all awards smaller than $600.
We believe that a reasonable educational case can be made for each of these options, and that none violates either equity or efficiency criteria. In the discussion section of the paper, we evaluate each of the alternative allocation formulas in terms of overall programmatic cost as well as the distributional consequences for institutions.

Before discussing our results, a few words are in order regarding the process used to select the policy options for our analysis. For a particular programmatic change to be included, we required it to be policy-relevant and required that it could be estimated with a relatively high degree of precision. For many of the programmatic changes we considered, but rejected, failure to meet the second condition was the reason.4 For example, reintroducing the net value of family residences into the need analysis was a potentially interesting programmatic change; however, since 1993-94, this information has not been collected on the federal student aid application. In this case, as well as others, the analysis would have required us to make an unrealistic set of distributional assumptions to estimate the cost of the proposed change.

In the case of certain options we did choose to pursue, it is important that we explicitly state the assumptions used in our analysis. For example, the savings estimates associated with the exclusion of either proprietary students, or students enrolled in less than one- and two-year programs, represent an upper bound on overall programmatic savings in that we have made no allowance for the possibility that these students might enroll in other types of institutions where they would remain eligible for Pell Grants. Analysts at both the Congressional Budget Office and the Office of Management and Budget suggest that up to 50 percent of these students might enroll elsewhere, cutting our estimates of program savings and increases in the maximum award possibly in half. However, since no direct empirical evidence exists to guide us, we present our savings estimates with this caveat.

Table Two presents the five-year cost savings associated with each of the seven options, together with the new maximum awards that would become possible if such savings were redirected to the remaining eligible students. In calculating the new maximum awards, we were constrained by both the annual outlays contained in our baseline projections, as well as the five-year cumulative outlays of $31.05 billion under the first scenario and $34.75 billion under the second scenario. As such, the entries in the table are expenditure neutral with respect to the current program and to the President's FY 96 budget proposal in that the savings generated by excluding some individuals from the program are used to increase the maximum awards for the remaining eligible population.5

Table Two:
Five-Year Cost Savings Associated with Frontloading, Exclusion of Proprietary School Students, Income Targeting, Exclusion of Students in Less than One-and Two-Year Programs, and $600 Minimum Award Levels (Cost Savings in Millions)

Scenario One Scenario Two
Programmatic Change 5 Year
Cost Savings
Maximum
Grant
5 Year
Cost Savings
Maximum
Grant
Frontloading $8,403 $3,030 $9,475 $3,410
Exclusion of proprietary
school students
5,134 2,710 5,698 3,030
Exclusion of proprietary
school students with
frontloading
13,327 3,720 14,940 4,180
Income targeting 3,016 2,540 3,338 2,840
Exclusion of students
in less than one year
programs
1,634 2,440 1,810 2,740
Exclusion of students
in less than two year
programs
4,383 2,650 4,848 2,960
$600 minimum award 606 2,370 545 2,650

Table Three then presents the distributional consequences of the various alternative allocation formulas, showing how the public, private, and proprietary sectors fare under the revised formulas. In this table, the relative gains and losses are shown for all three institutional sectors by comparing the percentage increase or decrease in that sector's share of total Pell Grant expenditures.6 Given our assumption of revenue neutrality, the gains and losses, by definition, sum to zero.

Table Three:
Changes in the Relative Shares of Total Pell Grant Dollars in the Public, Private. and Proprietary Sectors as a Result of the Various Programmatic Changes

Scenario One
Programmatic Change Public Private Proprietary
Frontloading -2.14% -2.73% 4.87%
Exclusion of proprietary school students 13.62 3.36 -16.98
Exclusion of proprietary school students with frontloading 15.55 1.43 -16.98
Income targeting -.26 -.21 .47
Exclusion of students in less than one year programs 3.46 .83 -4.29
Exclusion of students in less than two year programs 8.67 1.96 -10.63
$600 minimum award -.09 -.02 .11

Scenario Two

Public
Private Proprietary
Frontloading -2.22% -2.67% 4.89%
Exclusion of proprietary school students 13.46 3.34 -16.80
Exclusion of proprietary school students with frontloading 15.26 1.54 -16.80
Income targeting -.32 -.18 .50
Exclusion of students in less than one year programs 3.44 .97 4.41
Exclusion of students in less than two year programs 8.49 2.00 -10.49
$600 minimum award -.07 -.01 .08

Notes:

a. For the 1999-2000 academic year, under Scenario One, every 1 percent change in the dollar distribution among the sectors amounts to a shift of about $65 million in Pell Grant funds.

b. For the 1999-2000 academic year, under Scenario Two, every 1 percent change in the dollar distribution among the sectors amounts to a shift of about $75 million in Pell Grant funds.

Discussion

"Frontloading"' the grants to first- and second-year students provides the largest savings of any single option and, if implemented, would allow maximum grants of $3,030 or $3,410 under the two budgetary scenarios. The rationale for frontloading is that it would provide larger grant support for entering and second-year students, for whom higher education represents substantial risk and uncertainty. For those students who succeed in their first two years, much of that risk and uncertainty is diminished, and it is reasonable to expect them to borrow more heavily for the final two years. In essence, this option would reduce borrowing for first-and second-year students and increase it for third- and fourth-year students, while providing larger grant support in the first two years. Such a policy change would encourage more students to try higher education, while reducing loan defaults by students who start but do not complete a four-year degree.

From a distributional standpoint, frontloading would reallocate Pell dollars away from public and private institutions to those in the proprietary sector. In the 1995-96 award year, for example, the percentage of Pell dollars going to public colleges and universities would decline from 67 percent to almost 65 percent, while the percentage going to private colleges and universities would decline from about 17 percent to about 14 percent. In terms of actual dollars, the 5 percent increase for the proprietary sector translates into an additional $300 million Pell dollars.7

Exclusion of proprietary school students would allow an increase in the maximum grant to $2,710 under scenario one, or to $3,030 under scenario two, while the combination of frontloading with exclusion of proprietary school students would allow the maximum grant to rise to $3,720 or $4,180 under the two scenarios respectively. The rationale for excluding proprietary school students from the Pell Grant program is that many of these short course vocational programs might be more effectively and efficiently supported by the Labor Department through direct contracts with specific proprietary schools. To preserve our assumption of expenditure neutrality, it would be necessary to estimate the cost of these Labor Department programs, which would presumably be less than what is currently spent on proprietary school training through Pell Grants. Were that done, the effect would be to reduce the maximum Pell Grant made possible under this option.

As expected, excluding proprietary school students from Pell Grants benefits both public and private sector institutions. For example, in the 1995-96 award year, the percentage of Pell dollars going to public colleges and universities would increase from 67 percent to almost 80 percent, resulting in an additional $770 million Pell dollars. For private sector institutions, the 3 percent gain is more modest, reflecting a $200 million increase in funding.

However, when frontloading is combined with excluding proprietary school students, the nation's public colleges and universities are even better off, reflecting the disproportionately large number of students at two-year public, rather than two-year private institutions. Under this scenario, public sector institutions would see their share of Pell dollars increase in 1995-96 from 67 percent to almost 82 percent, while private sector colleges and universities would see their share rise by slightly less than 2 percent. In terms of actual dollars, public sector institutions would gain almost $900 million, while private sector institutions would gain less than $100 million.

Income targeting in the version considered for this paper assumes a 20 percent increase in the assessment rates on student and family income, and produces relatively modest savings, allowing maximum grants of either $2,540 or $2,840 under the two scenarios. The rationale for income targeting is that it concentrates grant funds on the most needy students; a number of approaches could have been chosen to achieve this objective. Under this particular option, the distributional consequences for all sectors are relatively minor, with the proprietary sector gaining less than half of a percent in overall program revenues. In terms of actual dollars, students at public sector colleges and universities would lose about $28 million in 1995-96, while those at private sector institutions would lose over $17 million.

The next two options focus on limiting eligibility to students who are enrolled in programs of at least two years duration, and which award a degree rather than a certificate of completion. These options would exclude students enrolled in less than two year programs in both the traditional and proprietary sectors but would include proprietary programs of at least two years duration. Savings are relatively small for the less than one-year exclusion, but the two-year limit would allow maximum grants of $2,650 or $2,960 under the two scenarios.

The distributional consequences of excluding students in less than one- and two-year programs, however, are more severe than those under income targeting. For the 1995-96 award year, excluding students enrolled in less than one year programs would produce a 3 percent gain for the public sector (from 67 percent to 70 percent) and a 1 percent gain for the private sector (from 17 percent to 18 percent). Under this scenario, students at proprietary schools would lose almost $250 million.

When students enrolled in less than two year programs are excluded from the eligibility formulas, public and private sector institutions continue to gain at the expense of the proprietary sector. If this programmatic change were to take effect in 1995-96, the proprietary sector would lose almost $610 million, as their share of total Pell dollars declined from 16 percent to 6 percent. The biggest winners would be students at public sector institutions, whose share would increase by 9 percent (from 67 percent to 75 percent). Students enrolled at private colleges and universities would also benefit as their share increased from 17 percent to 19 percent.

The final option, eliminating awards smaller than $600, produces modest savings and would allow maximum grants to increase only to $2,370 or $2,650 under the two scenarios. In addition, it has barely any distributional impact. For example, in the 1995-96 award year, the private sector's share of total Pell dollars remains essentially unchanged, while the proprietary school students gain one-tenth of a percent at the expense of students in the public sector. In total, the gain to students in the proprietary sector is less than $3 million.

Conclusions

Although we believe a case can be made for all seven of these options on grounds of efficiency and equity, we recognize that none would necessarily have a clear path in the policy arena. Hard choices must be made, however, and we would argue that these simulations demonstrate that, within current budgetary limits, greater access can be achieved than the current allocation formula permits.


1 The simulations in the paper were generated by Version 1.3, Update 96B Grant Cost Estimation Model. In this version, 240,000 sample records, stratified by dependency status and income level, were used together with 1993-94 program participation rates and 1994-95 application flow rates to produce the various cost estimates.

2 For dependent students, the calculations are also made for parental income and assets.

3 The two formulas are the entitlement rule, which subtracts the student's EFC from the maximum award level, and the traditional need-based rule, which subtracts students' EFC from their cost of education. When the maximum award level exceeds $2,400, the percent of cost rule, introduced in the 1992 Higher Education Amendments, is also used so that three potential award levels are calculated, with the lowest award being the one the student receives.

4 In addition to inclusion of the net value of the family home into the need calculations, other programmatic changes considered were limiting awards to those with a high school diploma or GED, eliminating the financial aid administrator's authority to use professional judgment, introducing tuition sensitivity into the award formula, as well as numerous variations on the frontloading theme.

5 The maximum awards presented in the table are for the 1999-2000 academic year, although little variation exists across award years for a given programmatic change.

6 As in the previous table, the percentage gains and losses presented are for the 1999-2000 academic year, although the gains and losses for the other years are very similar.

7 In calculating the distributional consequences of various programmatic changes, the baseline distribution of awards across sectors was compared to the distribution of awards that resulted from the new higher maximum award level. In this manner, the revenue neutrality of total programmatic cost was preserved.

David W. Breneman is Professor and Dean of the Curry School of Education at the University of Virginia. Fred J. Galloway is Director of Federal Policy Analysis at the American Council on Education.
-###-


[The Federal Role in Financing Higher Education: An Economic Perspective] [Table of Contents] [Implications of Demographic Trends in Higher Education on Student Financial Aid Over the Next Ten Years]