Administrative Costs of Federal Credit Programs

Summary

The federal government supports some private activities—such as homeownership, postsecondary education, and certain commercial ventures—through credit assistance offered to individuals and businesses. Some of that assistance is provided through direct loans, and some is provided through federal guarantees of loans made by private financial institutions.

In the process of providing such credit assistance, the government incurs administrative costs. Administrative costs include servicing and collection costs that are essential to the repayment of direct loans and the recovery of losses from loan guarantees. They also include costs for general administrative functions (such as credit-extension costs for underwriting and processing loans, as well as policy and oversight costs).

Those costs are measured separately in the federal budget and on a different basis from the credit subsidy—the main measure of costs for loan and loan guarantee programs under the Federal Credit Reform Act of 1990 (FCRA). Specifically, in the budget, the credit subsidy is measured as the present value of the lifetime cost of (or savings from) a loan at the time of disbursement or, for a loan guarantee, when the guarantee is made.1 (A present value is a single number that expresses the flow of current and future income or payments in terms of an equivalent lump sum received or paid at a specific time.) By contrast, administrative costs are measured on a cash basis—that is, they are recorded when the spending occurs. As a result, credit subsidies and administrative costs cannot be directly combined to produce an estimate of those programs’ total costs.2

To enhance transparency about the costs of certain federal credit programs, the Congressional Budget Office has developed a method for calculating the present value of their lifetime administrative costs using data collected from federal agencies by the Office of Management and Budget (OMB). In this report, those estimates are referred to as administrative cost subsidies. By combining the administrative cost subsidy with the credit subsidy, an estimate of the total cost of selected credit programs may thus be calculated.

Measuring Administrative Cost Subsidies

In its cost estimates and baseline budget projections, CBO generally follows the standard budgetary procedures for credit programs. In the budget, the credit subsidy is measured separately for each cohort of direct loans and loan guarantees. (A cohort refers to all loans or guarantees originated by a program in one fiscal year.) By contrast, administrative costs are not measured in the budget on a cohort basis. Instead, the administrative costs of loans and loan guarantees originated over many previous years and during the current fiscal year are combined into a single number on the basis of when the spending takes place.

CBO’s method for calculating administrative costs in this report differs from the approach used in the federal budget: The agency has estimated the administrative costs associated with particular cohorts. For the cohorts of direct loan programs analyzed in this report—which comprise loans originated in fiscal years 2019, 2020, and 2021—CBO estimated an average administrative cost subsidy rate of 2.0 percent (see Figure S-1).3 (That rate is equal to the administrative cost subsidy divided by the amount of credit obligations.) For loan guarantee programs, the average administrative cost subsidy rate was lower, 0.3 percent. CBO also estimated that larger programs (measured by the amount of credit obligations) tend to have lower administrative cost subsidy rates than smaller programs do.

In CBO’s view, an alternative method—the fair-value approach—would provide a more comprehensive estimate of the total cost of a credit program, particularly when measuring both the credit subsidy and the administrative cost subsidy on that basis. The advantage of such an approach is that it would incorporate market risk—the financial risk that remains even with a well-diversified portfolio and that depends solely on the performance of the economy. (For a discussion of fair-value estimates of administrative expenses, see Appendix A.)

Federal Direct Loan Programs Compared With Private-Sector Lending

This report also compares the administrative cost subsidies for federal direct loan programs with those of private financial institutions offering similar types of credit (such as housing and real estate loans and commercial loans). They may be different for a number of reasons. The federal government’s motivation for offering loan programs may differ from that of the private sector. For example, the government’s credit programs have goals that have been set by the Congress and the Administration, whereas private firms seek to maximize profits and may be subject to regulation that requires a certain amount of lending to low- and moderate-income borrowers. Additionally, the government does not always perform all administrative functions for the underlying loans. For example, it often outsources servicing and collection activities to the private sector. Those differences could partially explain why the federal government might spend more or less than private financial institutions do. For instance, CBO found that, on average, the administrative cost subsidy rates were lower for housing and real estate loans offered by the federal government than for those offered by private lenders but were higher for commercial loans.

Budgetary Measures and the Administrative Costs of Federal Credit Programs

Current budgetary measures for costs associated with federal credit programs were established by the Federal Credit Reform Act of 1990. The costs of credit programs fall into two categories, which are measured differently in the federal budget.

• Credit subsidy. FCRA requires the use of accrual accounting to measure the lifetime cost to the government of making direct loans or guaranteeing loans made by private-sector lenders. (Under accrual accounting, unlike cash accounting, the estimated present value of credit programs’ expenses and related receipts is recorded when the legal obligation is first made rather than when subsequent cash transactions occur.) As required by FCRA, the credit subsidy, in the case of direct loans, equals the lifetime cash flows between the government and the borrower recorded up front on a present-value basis. In the case of a loan guarantee, the credit subsidy is the present value of future cash flows between the government and the private institution making the loan.
• Administrative costs. Administrative costs—which include the cost of originating direct loans and loan guarantees, performing oversight functions, servicing loans, and collecting on defaulted loans—are treated differently in the federal budget. As required by FCRA, those costs are excluded from the credit subsidy and instead are measured on a cash basis—that is, the government’s expenditures are recorded as outlays in the year in which they occur and as the Congress appropriates funds for those expenditures. (For a discussion of CBO’s earlier report on the budgetary treatment of the administrative costs of federal credit programs and how those costs compare with the administrative costs of grant programs, see Box 1.)

  Box 1.

  Findings From CBO’s 1992 Report on Budgeting for Administrative Costs

  In January 1992, the Congressional Budget Office published Budgeting for Administrative Costs Under Credit Reform, which discussed the effects that the enactment of the Federal Credit Reform Act of 1990 (FCRA) had on credit program accounting and the budgetary treatment of administrative costs.1 For that report, CBO also estimated the long-term administrative costs of federal credit programs and showed how they compared with the administrative costs of private-sector credit programs.

  CBO concluded that the implementation of FCRA removed the bias—in terms of budgetary costs—in favor of loan guarantee programs and against direct loans and grants. Cash accounting, the method used in the budget before the enactment of FCRA, overstated the costs of direct loans when they were made by ignoring future repayments but treated loan guarantees as relatively costless when credit was extended because cash flows did not occur until a future date. FCRA did not include administrative costs in the credit subsidy, and thus administrative costs continued to be funded separately on a cash basis—even though those costs could be significant and were directly related to extending or guaranteeing credit.

  CBO estimated that the administrative cost subsidy rate was 7.7 percent of the total amount of direct loans issued in 1991 and 1.0 percent of the total amount of loan guarantees and grants made in 1991. Based on the limited program data available at the time, those results were similar to the administrative costs of private-sector loans. The data used to construct the estimates in this report are more comprehensive than the data that were available for the 1992 report.

  In its 1992 report, CBO found that continuing to budget for administrative costs on a cash basis would have advantages. Specifically, the agency concluded that a change in the budgetary treatment of administrative costs would have the following effects:

  • Increase the budgeting and accounting workload of federal agencies;
  • Reduce Congressional control over agency appropriations;
  • Increase the deficit in the short run because of the outlays required by a present-value method of budgeting;2 and
  • Potentially make credit programs less comparable to noncredit programs.

  This report builds on CBO’s prior work on the administrative costs of federal credit programs in three ways. First, the estimates in this report reflect changes in business practices and technology since 1992 that may have affected administrative costs. Second, CBO now has a wider range of data available on federal credit programs. For instance, for this report, CBO used loan performance data by cohort, a larger sample of federal credit programs, and supplemental information on federal credit program contracts, none of which were available to the agency in 1992. Third, this report includes an enhanced comparison between the administrative costs of credit programs administered by the federal government and by the private sector. CBO obtained more detailed data on private-sector lending activities than was previously available and developed a method that estimates the present value of long-term administrative costs for private-sector credit programs.

  ---

  1. Section 503 of the Federal Credit Reform Act of 1990, 2 U.S.C. § 661(b) (2006) required CBO to study and issue a recommendation on the budgetary treatment of administrative costs. See Congressional Budget Office, Budgeting for Administrative Costs Under Credit Reform (January 1992), www.cbo.gov/publication/20562.

  2. A present value is a single number that expresses the flow of current and future income or payments in terms of an equivalent lump sum received or paid at a specific time.

The different budgetary treatments—each with its own advantages and disadvantages—make it difficult to combine the two costs and then compare the overall costs of federal credit programs.

How Credit Subsidies Are Measured in the Federal Budget

Under FCRA, the present value of all projected cash inflows and outflows associated with a direct loan or loan guarantee is calculated by discounting those flows at Treasury rates with corresponding terms to maturity and adding them together to produce a single value.4 The budget records a separate credit subsidy cost for each set, or cohort, of loans or loan guarantees originated by a program in a fiscal year. For example, the 2021 cohort of student loans is the set of student loans originated during fiscal year 2021.

The expected cash flows incorporated in the credit subsidy of a direct loan include the disbursement to the borrower, the scheduled principal and interest payments net of defaults and recoveries, and the fees that the government expects to receive.5 For loan guarantees, the cash flows include payments that the government is expected to make to the private-sector lender when a borrower defaults on the loan, fees that the government receives, and recoveries. Those costs are updated each year to reflect changes in the credit subsidy on the basis of loan performance and interest rate changes.

How Administrative Costs Are Measured in the Federal Budget

Because the administrative costs of direct loans and loan guarantees are measured in the federal budget on a cash basis, those costs are recorded when the cash flows from the government occur. For example, the administrative costs recorded in 2021 for student loans included the costs of originating loans made that year and the costs of servicing loans made in previous years. In addition, the administrative costs for each loan or loan guarantee program are not always measured separately in the budget. Budget accounts for administrative costs may cover more than one credit program and include administrative costs associated with other types of programs. Although cash accounting is also used for most other government activities, it differs from the accrual treatment of credit subsidies.

This report divides administrative costs into four main categories (see Figure 1):

• Credit-extension costs are the costs required to originate a direct loan or loan guarantee. The salaries of loan officers who assess the creditworthiness of borrowers seeking direct loans are an example of credit-extension costs.
• Policy and oversight costs are the costs of enforcing policies that govern the direct loan or loan guarantee program and of conducting oversight activities related to the program. A program incurs policy and oversight costs while an outstanding direct loan or loan guarantee exists or for as long as the credit program is active. A few examples of policy and oversight costs are the costs to update computer systems to comply with data security regulations, the costs to revise borrower forms to comply with fair-lending laws, and the costs to investigate potential cases of loan fraud.
• Servicing costs are those associated with contacting the borrower and collecting on a loan (before any default). Those costs are incurred when loans are modified, statements are sent, payments are received or attempts are made to collect such payments from the borrower or lender, and efforts are made to collect on outstanding nondelinquent loans or loan guarantees. Some examples of servicing costs include the cost to mail a loan statement to a borrower and the cost to process a borrower’s monthly payment on a direct loan.
• Collection costs are incurred when the government or private lender attempts to collect on defaulted loans (including those the government assumes when a payment is made on a loan guarantee) and when collateral from the borrower or lender is seized and liquidated for recovery by the government.6 Examples of collection costs are those required to locate a borrower who has defaulted on a direct loan but cannot be reached with the contact information on file and those associated with collecting from a borrower (including seizing and selling collateral) who defaulted on a loan guaranteed by the government.

In this report, servicing and collection costs are often discussed together because those costs are essential to preserving the value of the government’s credit obligations. Estimates of the administrative cost subsidy rate are presented for each administrative cost function, as well as in total.

Administrative cost subsidy rates vary widely not only between direct loans and loan guarantees but also across individual loan programs. The administrative cost subsidy rate is higher for direct loan programs than for loan guarantee programs because the government is responsible for all administrative costs associated with direct loans but not for all of the costs of loan guarantees (private-sector lenders share some of those costs). In terms of efficiency, any differences should be small when the administrative costs for loan guarantees that are borne by the private sector are included. However, borrowers may pay a higher interest rate or larger fees for guaranteed loans than they would for a similar direct loan in order to compensate the lender for its administrative costs.

Challenges in Combining Credit Subsidies and Administrative Cost Subsidies

The Congress often provides appropriations for the subsidy cost of multiple loan programs (including both direct loans and loan guarantees) along with an appropriation for the administrative expenses associated with operating those loan programs in the same program account.

In many cases, such appropriations are provided without any clear indication of how the administrative costs relate to the various loan programs. For example, the “Agricultural Credit Insurance Fund Program” is the program account that receives an annual appropriation to operate and administer several programs, including emergency disaster loans and direct loans and loan guarantees for farm operating and ownership. Including administrative costs, particularly expenses that are essential to preserve the value of the government’s obligations, in measures of the subsidy costs would make subsidy estimates across programs more comparable.

Given the different budgetary treatment of credit subsidies and administrative costs, combining the two measures to get a more comprehensive estimate of the overall cost of a credit program is not straightforward. The credit subsidy for a loan program in any year measures the lifetime cost of the loans originated in that year—that is, it measures the credit subsidy cost for only that cohort of loans. By contrast, administrative costs are not recorded in the budget for individual cohorts and not in a way that isolates the cost of each program. The appropriation provided for administrative costs not only covers the credit-extension costs for the cohort of loans to be made in the same year, but also includes servicing, policy and oversight, and collection costs in that year for many past cohorts.

Effects of Accrual Accounting for Administrative Costs

The primary goal of accounting as prescribed by FCRA is to recognize budgetary costs when commitments (direct loans and loan guarantees) are made, rather than when those costs are paid. FCRA grants agencies permanent indefinite authority to cover credit costs that exceed expected costs. For example, when a loan guarantee program needs to make larger-than-expected claim payments on defaulted loans, permanent indefinite authority allows the agency to make those payments without any new funding from the Congress. If credit subsidy costs are lower than expected for the credit program, those savings accrue to the Treasury and not to the program. Loan guarantees are contractual obligations, and this system ensures that the government will not default on those obligations. The credit subsidy cost is updated each year on the basis of data for the cash flows received to date and revised assessments of future cash flows.

By contrast, under cash budgeting for administrative costs, the Congress appropriates funds for those costs for all cohorts, past and present, on an annual basis. Those annual appropriations give the Congress more control over administrative costs. However, some administrative activities—particularly the servicing activities necessary for the repayment of a direct loan or for recoveries in the case of default—are unavoidable after a loan has been originated if the government is going to recover any of the money it provided.

Covering administrative expenses with permanent indefinite authority would ensure that sufficient resources were available to carry out those administrative activities but would reduce the amount of control the Congress has over those expenses after the origination of a loan as well as reduce a program’s incentive to operate efficiently. It might also reduce an agency’s incentive to cut spending on administrative activities in ways that would result in larger credit losses.7 Fewer resources for administrative activities could increase credit losses, either because the agency could not thoroughly screen borrowers at origination or because it would have to allocate fewer resources for servicing and collection.

CBO’s Estimates of Administrative Cost Subsidy Rates for Selected Credit Programs

For the programs included in this report, CBO estimates that the administrative cost subsidy rates for direct loan programs are, on average, higher than those for loan guarantee programs.8 To allow for direct comparison with the credit subsidy rate, CBO estimates an administrative cost subsidy rate that is equal to the present value of administrative costs divided by the amount of credit obligations. For the 2019, 2020, and 2021 cohorts, the weighted average estimate of the total administrative cost subsidy rate was 2.0 percent for direct loan programs and 0.3 percent for loan guarantee programs (see Table 1). Large programs (those with relatively larger amounts of credit obligations) tend to have lower administrative cost subsidy rates, on average, than small programs, suggesting that there are some economies of scale.9 (For estimates of those subsidy rates for individual programs and cohorts, see Appendix B.)

Direct Loan Programs

In its direct loan programs, the federal government bears all costs for credit extension, policy and oversight, loan servicing, and debt collection. The total administrative cost subsidy rate for those activities varied by the type of lending for direct loans issued between 2019 and 2021 (see Table 2).

Student Loans. The federal direct student loan program account had the largest amount of credit obligations ($358 billion) among the departments and agencies that provide direct loans. The program had a total administrative cost subsidy rate of 2.1 percent. Nearly all of the program’s administrative expenses were related to servicing and collection activities, which accounted for 1.9 percentage points of that subsidy rate.

Housing and Real Estate Loans. CBO estimated the administrative cost subsidy rate for two housing and real estate programs that offer direct loans, both of which are administered by the Department of Agriculture. The first program is the Rural Housing Service’s Multifamily Housing Revitalization Seconds program (part of the Multifamily Housing Revitalization program account), which offers second mortgages to finance the repair and rehabilitation of multifamily housing projects. That program had the second-largest amount of credit obligations among all direct loan programs ($76 billion) over the 2019–2021 period, which was significantly more than all other direct loan programs except for student loans. It had the lowest total administrative cost subsidy rate (less than 0.5 percent).

The second housing and real estate program is the Section 502 Single Family Housing Direct Loan program (part of the Rural Housing Insurance Fund, or RHIF, program account). That program offers payment assistance to low-income borrowers in rural areas. The RHIF program account had the highest total administrative cost subsidy rate (27.8 percent) of all direct loan programs. It also had the highest administrative cost subsidy rate (6.7 percent) for servicing and collection activities.

Commercial Loans. Most of the direct loan programs that CBO analyzed offer commercial loans, yet the amount of credit obligations for those programs ($39 billion) is less than the amounts for student loans and for housing and real estate loans. CBO estimated a total administrative cost subsidy rate of 2.8 percent for commercial loan programs; subsidies for credit extension and policy and oversight costs (1.5 percent) were similar to those for servicing and collection costs (1.3 percent).

Loan Guarantee Programs

In loan guarantee programs, the federal government bears all costs for conducting credit extension and policy and oversight activities, but most costs for loan servicing and debt collection are typically borne by private-sector lenders.10 CBO found small differences in the estimates of the administrative cost subsidy rate by lending category—in total and by administrative function—for loan guarantees issued between 2019 and 2021.

Housing and Real Estate Loans. CBO found that the total administrative cost subsidy rate (0.3 percent) for housing and real estate loan guarantees was the lowest among all lending categories, including both direct loans and loan guarantees. Those programs had the largest amount of obligations ($2.0 trillion) among all programs in CBO’s analysis and a low administrative cost subsidy rate for servicing and collection activities (0.2 percent). One reason for the low subsidy rate for those activities is that loan guarantee programs incur servicing costs only after they assume defaulted loans. When a borrower defaults, private lenders attempt to make the loan current (for example, through collections or loan modifications) before making a claim on the loan guarantee. In most cases, once the government makes a claim payment to the lender, it assumes the defaulted loan and incurs all future administrative costs for the loan, just as if that loan was a direct loan. Those future costs may include seizing and selling any collateral backing the loan.

Commercial Loans. Loan guarantees for commercial loans, in aggregate, account for just $113 billion of the $2.1 trillion in loan guarantees issued by the programs that CBO analyzed. Those programs had the same low administrative cost subsidy rate for servicing and collection expenses (0.2 percent) as housing and real estate loans. That result is expected given that most of the programs have very low default rates; thus, the government rarely assumes defaulted loans.

CBO did not include the Small Business Administration’s (SBA’s) Paycheck Protection Program (PPP) in its estimates for commercial loan guarantees, because that program has characteristics that are more similar to a grant program than to a typical loan guarantee program. However, CBO did include estimates as an addendum in the tables that accompany this report. Credit obligations for the PPP totaled $795 billion in 2020 and 2021, and CBO estimated that the administrative cost subsidy rate was just 0.1 percent—split roughly equally between costs for credit extension and policy and oversight activities and costs for servicing and collection.

The PPP processed a large number of loans just a few weeks after the program was created in 2020 to assist small businesses that were adversely affected by the coronavirus pandemic. The SBA estimated that more than 80 percent of the administrative costs for the PPP were related to credit extension for the 2020 cohort; that percentage fell to less than 40 percent for the 2021 cohort of loans as the program became more streamlined. Conversely, just 10 percent of administrative costs was related to servicing activities for the 2020 cohort, but that percentage increased to half of all administrative costs for the 2021 cohort. Policy and oversight activities were initially limited and accounted for about 10 percent of total administrative costs for both cohorts. A significant amount of fraud occurred soon after the program was implemented because the initial application for forgiveness on PPP loans of $150,000 or less required no documentation.11 (For further discussion about the PPP, which ended on May 31, 2021, see Box 2.)

Data and Methods Used to Estimate Administrative Cost Subsidies

CBO received data for the administrative costs of most credit programs from the agencies that administer those programs, facilitated by a request from the Office of Management and Budget. CBO, in consultation with OMB, selected a subset of agencies to represent a broad range (in terms of size and sector) of direct loan and loan guarantee credit programs. The data are from the 2019, 2020, and 2021 cohorts and include the following: the total amount of administrative costs in each credit program account; an allocation of those costs across four categories of administrative functions; and general information about the credit portfolio, including data regarding the average maturity, the average number of years from origination to default, and the average number of years that a collection effort is active on defaulted loans. CBO estimated missing values using data from its credit subsidy models and from the appendixes and Federal Credit Supplements included in the federal budgets proposed for fiscal years 2021 through 2024.12

Each program account—which may include several individual credit programs—receives an annual appropriation for its administrative expenses. Agencies are not required to account for the administrative costs of individual programs or cohorts but did provide estimates to CBO for use in this report. Those estimates are inherently imprecise given the lack of detailed accounting for administrative functions within and across individual programs. All agencies provided an estimate of the allocation of their program accounts’ administrative costs to individual programs, with the exception of the Development Credit Authority (an agency within the Department of State).13 For that agency, CBO distributed the annual administrative costs for the program account weighted by the volume of outstanding loans and new credit obligations across all of its credit programs.

Agencies also estimated how each program’s administrative costs were distributed to the four administrative functions used in this report—credit extension, servicing, policy and oversight, and collection. For most programs, CBO used the data reported by the agencies. In some cases, however, agencies did not report any expenses for credit extension or for policy and oversight activities; in those instances, CBO applied a minimum allocation of 5 percent to each function to account for general program expenses. (The minimum allocation applied to only a few programs.) In most of those cases, CBO incorporated the remaining costs across the other administrative functions using the agencies’ initial allocation to those functions. For student loans, however, CBO estimated the allocation of administrative costs to each function using open government databases that reported agencies’ spending by program account.14

CBO estimated annual costs for each administrative function as follows:

• Credit-extension costs. CBO estimated that all the costs of extending credit to borrowers were realized in the first year after the loan was disbursed or the loan guarantee was made. Although programs that disburse loans over multiple years may incur additional credit-extension costs with each disbursement, CBO assigned all costs to single disbursements for simplicity and comparability with other cash flow estimates for credit subsidy costs.
• Policy and oversight costs. To estimate the annual costs for policy and oversight functions for each loan, CBO divided the share of annual program costs allocated to those functions by the number of loans outstanding. For each year until the average maturity of the loans, those amounts (increased by an estimated long-run average inflation rate of 2 percent each year) were then multiplied by the number of loans in the cohort to estimate policy and oversight costs.
• Servicing costs. To estimate the annual costs for servicing activities for each loan, CBO divided the share of annual program costs allocated to servicing by the number of loans outstanding. For each year until the average maturity of the loans, those amounts (increased by an estimated long-run average inflation rate of 2 percent each year) were then multiplied by the number of loans in the cohort to estimate servicing costs.
• Collection costs. To estimate the annual cost per loan for collection activities, CBO divided the share of annual program costs allocated to collection efforts by the number of loans in default. Beginning in the average year of default and continuing for each year that a collection is active, the annual cost per loan (increased by an estimated long-run average inflation rate of 2 percent each year) was multiplied by the expected number of defaulted loans to estimate the administrative cost for collection activities.

The administrative cost subsidy for each cohort of loans is the present value of the sum of the annual costs for each administrative function: credit extension, policy and oversight, servicing, and collection. The discount rates used to compute the present value are the same as those used to estimate the credit subsidy for each cohort. That construction allows the administrative cost subsidy to be compared directly with the credit subsidy.

Administrative Cost Subsidy Rates for Federal Direct Loan Programs and for Private-Sector Credit Activities

Private financial institutions provide a useful benchmark for the administrative costs of federal credit programs because they engage in similar lending and financial activities. Private institutions may maximize their profit from lending activities either by spending less on total administrative costs or by spending more on certain administrative activities that directly reduce credit losses. By contrast, government agencies have permanent indefinite authority to cover higher-than-expected credit losses and separately receive an appropriation for administrative costs.

For private lenders, CBO estimated the administrative cost subsidy rate differently than for federal programs. Specifically, CBO defined the subsidy rate as the noninterest expense associated with making a set of loans divided by the dollar amount of those loans. The agency found that the rate varied across four private lending categories: student loans, housing and real estate loans, commercial loans, and consumer loans (see Table 3). For loans originated between 2019 and 2021, CBO estimated that the administrative cost subsidy rate for private lenders was 2.5 percent for student loans, 2.9 percent for housing and real estate loans, 1.7 percent for commercial loans, and 2.9 percent for consumer loans.

CBO did not have administrative cost data by administrative function for private lenders. Because consumer loans have higher default rates than housing and real estate loans and commercial loans, they are likely to have greater administrative costs for servicing and collection activities.15 Furthermore, in terms of loan policies, underwriting practices, and loan modification and restructuring procedures, commercial loans vary more than housing and real estate loans and are therefore likely to have greater expenses for credit extension and policy and oversight activities.16

Overall, the federal direct loan programs in CBO’s analysis had lower administrative cost subsidy rates, on average, than similar private-sector credit programs for housing and real estate loans but a higher administrative cost subsidy rate for commercial loans. The results in each lending category varied because of differences in volume, as well as differences in the credit risk profiles of the programs, which generated different default rates and collection efforts.

Student Loans

The federal government is responsible for nearly all lending to students. Consequently, it is not easy to compare private-sector lenders’ administrative cost subsidy rates with those of the government. Sallie Mae is a private-sector company that originates and services student loans. Sallie Mae’s administrative cost subsidy rate was 2.5 percent, on average, over the 2019–2021 period compared with a subsidy rate of 2.1 percent for federal direct student loans. The estimate for federal direct student loans may not be representative of typical administrative costs, however, because the Department of Education paused repayments and did not attempt to collect on delinquent debt during the coronavirus pandemic.

Private-sector student loans typically have lower default rates—and thus potentially lower collection costs—than federal student loans because private lenders have stricter underwriting requirements. By contrast, federal student loans are available to all borrowers regardless of their creditworthiness.17 Although the amount of outstanding private-sector loans administered by Sallie Mae between 2019 and 2021 totaled $67 billion—much less than the $358 billion in federal obligations over the same period—its administrative costs for student loans may have been larger than those for federal loans because of underwriting practices. For example, private-sector loans are often larger than federal student loans (which are limited in size by statute), and Sallie Mae’s underwriting efforts are probably more comprehensive, including income verification and other criteria to establish creditworthiness.

Housing and Real Estate Loans

The federal government operates several large mortgage guarantee programs but just a few direct loan programs for housing and real estate loans. CBO’s data set included two programs offered through the Department of Agriculture’s Rural Housing Service: the Multifamily Housing Revitalization Seconds program and the Section 502 Single Family Housing Direct Loan program. The multifamily housing program had the largest amount of credit obligations for direct loans ($76 billion), and administrative costs were minimal across all administrative functions (less than 1 percent). By contrast, the single-family housing program had the least amount of credit obligations ($3 billion) but a high administrative cost subsidy rate for policy and oversight costs (21 percent), as well as a high subsidy rate for servicing and collection costs (6.7 percent). Servicing and collection costs in that program were high because of high default rates, and credit-extension costs were also high because borrowers in rural areas are frequently hard to reach throughout the loan origination and underwriting process.

Fannie Mae and Freddie Mac provide a private-sector comparison for the administrative costs of large single-family mortgage guarantee programs, such as the Federal Housing Administration’s (FHA’s) Mutual Mortgage Insurance Program.18 For loans issued from 2019 to 2021, on average, Fannie Mae and Freddie Mac spent 0.2 percent of outstanding loan balances on administrative expenses as compared with an administrative cost subsidy rate of 0.1 percent for the FHA. Fannie Mae’s and Freddie Mac’s expenses may have been higher because their activities are broader than the FHA’s. In addition to underwriting and seeking recovery in the case of losses, they also package mortgages into securities and sell them to investors (with guarantees for timely payment of principal and interest), and they carry out other activities such as the financing of multifamily homes.

Commercial Loans

The administrative cost subsidy rate was higher for federal direct loan programs for commercial lending (2.8 percent) than for similar private-sector loans (1.7 percent). Direct loan programs for commercial loans incur higher administrative costs for credit extension as well as for policy and oversight activities than other direct loan programs. There are more federal programs for commercial lending than for any other lending category, but many of those programs are small (in terms of credit obligations) or operate for just a few years.19

Consumer Loans

CBO’s data set included just one federal direct loan program that it classified as consumer lending: repatriation loans offered by the State Department. Repatriation loans are provided to Americans abroad who are unable to finance their return to the United States. That program is very small, with just $11 million in credit obligations between 2019 and 2021. However, it is relatively expensive to administer. CBO estimated that the administrative cost subsidy rate was 27.5 percent over that period, nearly all of which was attributable to the costs of credit extension. By contrast, because the private sector has underwriting standards and collateral requirements for consumer loans—as it does for all types of lending—its administrative cost subsidy rate for consumer loans (2.9 percent) was similar to that for housing and real estate loans.

Method Used to Estimate the Administrative Cost Subsidy Rate for Private Lending

CBO estimated the administrative cost subsidy rate for private financial institutions using data from S&P Capital IQ. The sample included data from the financial statements of 20 banks and a total of 26 “business segments” representing distinct activities within the organization from 2019 to 2021. Institutions may have different nomenclatures for their business segments and may report different types of data for each activity. For example, some institutions may divide their activities into “corporate banking,” “consumer banking,” and “mortgage banking” business segments, and they may provide balance sheets and income statements for each category. Others may use different names, such as “retail banking,” to refer to any or all of those same categories.

CBO identified business segments that aligned with at least one of three lending categories—housing and real estate loans, commercial loans, and consumer loans—and obtained data for noninterest expenses. To account for multiple lending categories within a single business segment, CBO proportionally allocated a segment’s noninterest expenses to each lending category using the reported amount of loans for the category. The administrative cost subsidy rate for private lenders was estimated as the ratio of noninterest expenses to assets for each business segment.

Noninterest expenses for the selected business segments in the private sector include administrative expenses that are similar to those for the government programs discussed in this report. Noninterest expenses measure administrative costs for private lenders in a way that is more comparable to the cash measures used in the budget than the cohort-based subsidy measures used for this report. However, CBO lacked the data to do a comparison using the same measure.

In its analysis, CBO considered the fact that some business segments encompass both an institution’s lending activities (which are reported as assets) and its liabilities. For example, the noninterest expenses associated with retail banking might include the costs of managing customers’ deposits in checking and savings accounts (which are reported as liabilities), as well as the costs of making automobile and personal loans (which are reported as assets) to those same customers. Because the report focuses on lending activities, the proportion of noninterest expenses attributable to liabilities was removed from CBO’s analysis.

CBO’s estimates of the administrative cost subsidy rate for private lenders are highly uncertain primarily because the financial statement data that the agency used were not standardized across business segments for each bank, making it difficult to clearly identify and assign each business segment to a lending category.20 As a result, for its comparison with federal programs, CBO used data for private lenders only when it was possible to clearly identify lending categories.

---

Appendix AFair-Value Estimates of Administrative Cost Subsidies

In its analyses of the costs of federal credit programs, the Congressional Budget Office measures credit subsidies using the method prescribed by the Federal Credit Reform Act of 1990 (FCRA). That method determines the cost of direct loans and loan guarantees on the basis of expected cash flows—primarily principal and interest payments, defaults, and recoveries—that are discounted at Treasury rates with corresponding terms to maturity.

An alternative approach, called the fair-value approach, accounts for a greater degree of financial risk than is reflected in the average budgetary effects. Specifically, the fair-value method includes market risk, which is the financial risk that remains even with a well-diversified portfolio and that depends solely on the performance of the economy. Government stakeholders are exposed to that risk when the government provides credit assistance or invests in a financial asset, such as an ownership stake in a private business. The fair-value approach provides information to policymakers about the cost of such risk, whereas the FCRA approach does not.

For this report, CBO estimated the combined cost of credit subsidies and administrative activities using FCRA procedures, thus including administrative costs but not the cost of market risk. Using fair-value measures for those combined cost estimates would incorporate both market risk and the costs associated with administering loan and loan guarantee programs. Such an approach would thus represent a more comprehensive cost measure for credit programs, in CBO’s estimation.

As an example, the Administration’s estimate of the FCRA credit subsidy rate for student loans originated in 2019 is 45.3 percent, and the administrative cost subsidy rate associated with those loans is 2.2 percent, resulting in a total subsidy rate of 47.5 percent on a FCRA basis.1 CBO estimated that same credit subsidy to be 62.1 percent on a fair-value basis. Incorporating the fair-value estimate of the administrative cost subsidy rate would yield a combined fair-value subsidy rate estimate of 64.4 percent (see Table A-1).

Ideally, CBO would have presented fair-value estimates of administrative costs in this report along with credit subsidies estimated on that basis. But estimating administrative costs on a fair-value basis would require data that were not available when this analysis was undertaken and therefore is beyond the scope of this report.

A fair-value estimate of the administrative costs associated with a cohort would differ from a FCRA estimate because administrative costs are uncertain. (A cohort refers to all the loans originated by a program in one fiscal year.) The cash flows discounted to yield the credit subsidy are also uncertain. Such costs might be higher than under the FCRA method because a fair-value approach places more weight on costs during an economic downturn when the cost of servicing delinquent loans might increase. However, fair-value estimates might be lower during an economic downturn because more loans are written off and thus servicing and collection costs could decrease.

---

Appendix BEstimates of the Administrative Cost Subsidy for Individual Programs and Cohorts

Using data collected from federal agencies by the Office of Management and Budget (OMB), the Congressional Budget Office has developed a method for calculating the present value of the lifetime administrative costs of credit programs—referred to as the administrative cost subsidy.1 This appendix presents estimates of the administrative cost subsidy rate for both individual programs and cohorts of loans. (A cohort refers to all the loans originated by a program in one fiscal year.) By contrast, the main text provides weighted-average estimates across the 2019, 2020, and 2021 cohorts for direct loans and loan guarantees by program account, which may consist of several individual programs. Significant variability exists across programs and within program accounts, as well as across cohort years for individual programs. (For the effects of such variability on direct loans, see Table B-1. For the effects on loan guarantees, see Table B-2. Those tables also present estimates of the subsidy rates calculated using procedures prescribed by the Federal Credit Reform Act of 1990. Those estimates include both the credit subsidy and the administrative cost subsidy.)

Variation in Estimates of Administrative Cost Subsidy Rates From Year to Year

Estimates for an individual program may vary across years for a several reasons, including changes in economic conditions or the scope of the program. For example, the Department of Education reported no outstanding defaulted loans during the coronavirus pandemic because the Administration paused student loan repayments. As a result, there were minimal administrative expenses attributable to collection activities for those student loans.

Variation in Estimates of Administrative Cost Subsidy Rates Across Programs

The administrative cost subsidy rate for programs with a larger amount of credit obligations tends to be lower than the subsidy rate for programs with smaller obligations, suggesting that administrative functions in large credit programs may operate more efficiently than those in smaller programs. The federal government’s role in the operations of credit programs also partially explains differences across programs. For example, private lenders and firms originate, service, and manage defaults on loans guaranteed by the Federal Housing Administration’s single-family mortgage program. By contrast, the Department of Homeland Security’s disaster assistance loan program requires specialized natural disaster assessors to be involved in making credit decisions, as well as additional resources to make and service loans to itinerant borrowers. Those costs are borne by the federal government.

Variation in Estimates of Administrative Cost Subsidy Rates Across Cohorts

The variation in estimates across cohorts for each program is also affected by differences in the costs for each administrative function, as estimated by individual agencies at the request of CBO and OMB. Those agencies may have estimated that a larger portion of total administrative costs was spent on credit-extension functions in years in which a larger number of loans were issued or that a larger portion of total costs was spent on collection activities in years in which economic conditions declined and defaults increased. For example, the costs of originating a loan are generally less than the costs of other administrative functions because origination costs generally occur in the year the loan is issued rather than over the lifetime of the loan. Therefore, a larger portion of total administrative costs spent on credit-extension functions—and a smaller portion spent on other administrative functions—will generally result in a smaller administrative cost subsidy. Finally, because servicing costs tend to be expensive and occur throughout the loan’s lifetime, a larger portion of total administrative costs spent on loan servicing functions for one cohort versus another will generally increase that cohort’s administrative cost subsidy.

---