Key Points
The Department of Education recently announced final regulations on the proposed Income-Driven Repayment Plan called “SAVE.” These final regulations follow the Supreme Court’s recent decision to invalidate President Biden’s previous Student Loan Forgiveness Plan. We estimate SAVE will incur a net cost of $475 billion over the 10-year budget window.
About $200 billion of that cost will come from payment reduction for the $1.64 trillion in loans already outstanding in 2023. We estimate that about 53 percent of the current loan volume will move to SAVE after it goes active in July 2024, implying that about $869 billion will be subject to enhanced subsidies under SAVE.
The remainder of the budget cost, or about $275 billion, comes from reduced payments for about $1.03 trillion in new loans that we estimate will be extended over the next 10 years. We estimate a take-up rate for future loans of 70 percent,1 implying that about $645 billion in future loans will be subsidized. About 6.57 percent of future borrowers (or 4.98 percent of total predicted loan volume) will never have to make any payments under SAVE.
Biden’s New Income-Driven Repayment (“SAVE”) Plan: Budgetary Cost Estimate Update
On July 10, 2023, the Department of Education issued final regulations for President Biden’s new Income-Driven Repayment (IDR) Plan (also known as “Saving on a Valuable Education” or “SAVE”). This new IDR plan will reduce federal student loan repayment for many borrowers, including borrowers in existing IDR plans as well as borrowers using the standard repayment option.
Our previous estimate of the new IDR plan was based on the August 2022 rule language that was updated on July 10, 2023. We also updated our methodology to include two more behavioral responses to the enhanced IDR program discussed below.2 Most importantly, our previous estimate was made after the Student Loan Forgiveness Plan was proposed but before the Supreme Court’s recent decision to invalidate the Student Loan Forgiveness Plan. This ruling increases the amount of total borrowing that can qualify for the new IDR plan, thereby increasing its cost.
Table 1 shows our estimates of the proposed SAVE plan. The first row shows our estimate of SAVE which includes changes in IDR take-up rates that we considered in earlier work but before any new behavioral effects considered in the current brief. We also updated our previous work to include the recent SCOTUS decision to prohibit student loan forgiveness. We also updated various other assumptions and now focus on a single student take-up optimization rule based on our recent analysis. (See the Technical Appendix to this brief for more information.) Under these assumptions, we estimate a total cost of $342 billion over the 10-year budget window.
| Scenario | Year | Cohorts 2024-2033 | Budget Window (2023-2033) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2023 | 2024 | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | 2031 | 2032 | 2033 | |||
| Updated budgetary cost estimate without considering changes in community college or student borrowing decisions. | 200.2 | 10.8 | 10.4 | 10.9 | 11.8 | 13.9 | 14.5 | 15.3 | 16.5 | 18.0 | 19.6 | 141.7 | 341.9 |
| Adding new community college borrowers qualifying for Title IV student loan program | 200.2 | 13.2 | 12.9 | 13.4 | 14.4 | 16.6 | 17.2 | 18.1 | 19.4 | 20.9 | 22.5 | 168.4 | 368.6 |
| Adding increase in student borrowing | |||||||||||||
| Conservative Estimate | 200.2 | 14.9 | 14.5 | 15.1 | 16.2 | 18.8 | 19.5 | 20.5 | 22.0 | 23.7 | 25.6 | 190.7 | 390.9 |
| Maximum Estimate | 200.2 | 29.3 | 28.0 | 28.9 | 30.8 | 35.6 | 36.6 | 38.2 | 40.7 | 43.7 | 46.9 | 358.6 | 558.8 |
| Medium Estimate | 200.2 | 22.1 | 21.3 | 22.0 | 23.5 | 27.2 | 28.0 | 29.4 | 31.3 | 33.7 | 36.3 | 274.7 | 474.9 |
Under standard scoring rules, the budget cost for existing loans appears in the current budget year (2023). This budget cost is the full amount, not a 1/2 fraction sometimes used prior to the posting of final regulations.
Table 1 also presents our estimate after allowing more community colleges to join the Title IV federal student loan program due to a reduced “Cohort Default Rate.” Under current law (that is, prior to the new SAVE program), default rates for students at many community colleges exceed the limit required for IDR program qualification. The SAVE plan will reduce the default rate, thereby allowing more community colleges to participate, and increasing the program cost. (See the Technical Appendix for more details on our estimation methodology.) Including this effect increases the 10-year budget cost to $369 billion.
Table 1 also presents new budget estimates that include alternative assumptions about how much students might increase their federal eligible borrowing in response to the SAVE plan. Under the law, students cannot borrow more than the “cost of attendance” at their institutions, but many students currently (before SAVE) borrow less than that amount. The “conservative” estimate shown in Table 1 includes an increase in student borrowing only to the extent that it is essentially “free money” to those students, as future repayment will not increase due to the lower payment caps relative to the new definition of discretionary income under SAVE.3 Under this assumption, the 10-year SAVE program cost increases to $391 billion. The “maximum estimate” includes borrowing that increases to equal the minimum of the statutory loan limit or the cost of attendance at the chosen institution. Under this alternative assumption, the 10-year SAVE program cost increases to $559 billion. Our medium estimate, which essentially falls in between these two extremes, produces a 10-year budget cost of $475 billion.
Table 2 presents the repayment ratio distribution for future borrowers over the next 10 years under our medium estimate discussed above. The repayment ratio measures how much of future projected loans are expected to be repaid either as payments under the non-IDR fixed-length plans or under IDR. More formally, the ratio is defined as the net present value of all future payments divided by the original loan amount borrowed. This ratio can exceed 100 percent because, by convention, the interest rate used for “budget discounting”4 is effectively not risk-adjusted and this rate is, therefore, lower than the federal student loan interest rate. A repayment ratio of 0 percent means those borrowers are projected to not repay any of their loans under IDR.
| Repayment Ratio | Fraction of Borrowers | Fraction of all Borrowing |
|---|---|---|
| 0% | 6.57% | 4.98% |
| Greater than 0%, smaller or equal to 20% | 13.76% | 10.66% |
| Greater than 20%, smaller or equal to 40% | 15.33% | 10.70% |
| Greater than 40%, smaller or equal to 60% | 14.53% | 12.10% |
| Greater than 60%, smaller or equal to 80% | 13.31% | 12.26% |
| Greater than 80%, smaller or equal to 100% | 11.90% | 11.70% |
| Greater than 100%, smaller or equal to 120% | 24.60% | 37.60% |
Note: PWBM calculation based on merged NPSAS18 data and PWBM Microsim.
As shown in Table 2, about 6.6 percent of future borrowers will never make a payment under SAVE, or about 5.0 percent of the total predicted loan volume disbursed. Among all federal student loan borrowers, 24.6 percent of them fully repay their loans, which accounts for 37.6 percent of the predicted loan disbursement. Therefore, about 62.4 percent of the predicted loan amount disbursed is projected to not be fully repaid.
The take-up rate of IDR – that is, the number of borrowers and the concomitant loan volume electing the IDR program – is central to any analysis of IDR costs. Previous literature includes a range of different possible “optimization” rules and heuristics that borrowers might follow when electing a loan repayment option. A companion piece to this brief examines these different options to see how well they explain the choice between existing IDR and standard repayment options that are currently in place, that is, prior to SAVE. In fact, the most obvious and most financially beneficial rule – simply minimizing the present value of future repayments – appears to best explain most borrower behavior with a surprisingly high level of accuracy.
According to IPEDS data, in 2021 approximately 1 million students were enrolled in community colleges that did not provide access to the Title IV federal student loan program. Most of these community colleges voluntarily do not participate in the Title IV student loan program because a high cohort default rate (CDR) could jeopardize their access to other Title IV funds, such as the Pell Grant.
However, because the proposed IDR plan introduces auto-enrollment with borrowers who are 75 days delinquent or more, the CDR merit of these community colleges would dramatically drop, allowing them to be able to participate in the federal student loan program under the new circumstance.
We incorporate this borrower base increase of newly eligible community college students into our IDR cost projection model using enrollment, cost of attendance, and transfer-out rate data from IPEDS. After merging with PWBM Microsim, we calculated the estimated budgetary cost in addition to the original projection. The analysis suggests that the borrower base increase would add approximately $2.7 billion per year for future years.
Due to the increased generosity of the newly proposed IDR plan, future student borrowers have the incentive to increase their federal student loan borrowing, shifting the current college financing pattern towards more loan borrowing instead of paying out-of-pocket. We investigate the potential borrowing increase scenarios due to different reasons as follows:
Under this scenario, we consider the situation where future students react to the increased generosity by maxing out the federal direct loan amount available to them. Using the NPSAS18 data, we construct the variable for each student borrower’s maximum applicable loan amount 
:
where COA is the cost of attendance, which includes tuition, room and board, books, supplies, transportation, loan fees, and miscellaneous expenses. is the maximum loan amount one student could borrow, according to the Federal Direct Loan rules.
For undergraduate students, the maximum amount (Loan limit) each year in Direct Subsidized Loans and Direct Unsubsidized Loans ranges from $5,500 to $12,500 per year, depending on the student’s year in school and dependency status. Graduate or professional students can borrow up to $20,500 each year in Direct Unsubsidized Loans. Direct PLUS Loans can also be used for the remainder of the college costs, as determined by the COA.
At the same time, we observe 
for each student. According to the data, the aggregate fraction of 
is about 52 percent.
Under this scenario, we assume every borrower would only increase their federal student loan borrowing amount to , if their present value of all projected future payments did not increase. This means only a small fraction of low-income borrowers (during their entire loan lifespan) would increase their federal borrowing. Specifically, borrowers predicted to be enrolled in the fixed-length repayment plans would not increase their borrowing. At the same time, most borrowers predicted to enroll in the new IDR plan would not increase their borrowing because their lifetime income is not low enough to guarantee no increase in the present value of future payments.
In this scenario, the aggregate fraction of 
is approximately 90 percent.
Written by Junlei Chen under the direction of Kent Smetters. Prepared for the website by Mariko Paulson.
-
The 70 percent includes only loan volume enrolled in SAVE. The aggregate IDR take-up rate is about 75 percent, including SAVE and Income-Based Repayment (IBR) plan. ↩
-
Our updated cost methodology now incorporates college students changing college financing patterns and more community colleges joining Title IV student loan program due to reduced Cohort Default Rate. See the Technical Appendix for more details. ↩
-
See our previous brief for a discussion on how SAVE increases the income no longer counted as discretionary income while also lowering the caps on repayment relative to discretionary income. ↩
-
Here we used interest rates on Treasury securities from the year of the loan’s disbursement with maturities that match the timing of the cash flows, following CBO 2020. ↩