Implications of Student Loan COVID-19 Pandemic Relief Measures for Families with Children

Finance and Economics Discussion Series Federal Reserve Board, Washington, D.C. ISSN 1936-2854 (Print) ISSN 2767-3898 (Online) Implications of Student Loan COVID-19 Pandemic Relief Measures for Families with Children Sarena Goodman, Simona Hannon, Adam Isen, Alvaro Mezza 2023-025 Please cite this paper as: Goodman, Sarena, Simona Hannon, Adam Isen, and Alvaro Mezza (2023). “Implications ofStudentLoanCOVID-19PandemicReliefMeasuresforFamilieswithChildren,” Finance and Economics Discussion Series 2023-025. Washington: Board of Governors of the Federal Reserve System, https://doi.org/10.17016/FEDS.2023.025. NOTE: Staff working papers in the Finance and Economics Discussion Series (FEDS) are preliminary materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth are those of the authors and do not indicate concurrence by other members of the research staff or the Board of Governors. References in publications to the Finance and Economics Discussion Series (other than acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers.

Implications of Student Loan COVID-19 Pandemic Relief Measures for Families with Children* Sarena Goodman Federal Reserve Board Simona Hannon Federal Reserve Board Adam Isen Office of Tax Analysis, U.S. Department of the Treasury Alvaro Mezza Federal Reserve Board Abstract The initial years of the COVID-19 pandemic and the resulting economic fallout likely posed particular financial strain on U.S. households with children, who faced income disruptions from widespread jobs and hours cuts in addition to new childcare and instruction demands. One common expense for many such households is their student loan payment. The Coronavirus Aid, Relief, and Economic Security (CARES) Act included provisions to curb the impacts of these payments, which have been extended several times. These measures were not targeted and thus applied independent of need. This chapter analyzes two nationally representative datasets and finds: 1) families with children were more likely to benefit from pandemic student loan relief than those without children, but this relief was concentrated among higher-income and White families and 2) there were larger improvements in overall credit health and an increased use of other credit among families with student loan debt that was eligible for relief relative to those with student loan debt that was not. * Email: sarena.f.goodman@frb.gov, simona.m.hannon@frb.gov, adam.isen@gmail.com, and alvaro.a.mezza@frb.gov. Analysis prepared in Spring 2022 for a National Academies of Sciences, Engineering, and Medicine’s 2023 Report entitled Addressing the Long-Term Effects of the COVID-19 Pandemic on Children and Families. Hannah Case provided excellent research assistance. This analysis does not necessarily reflect the views or opinions of the Federal Reserve Board, the U.S. Treasury, or the National Academies of Sciences.

2 I. Introduction The COVID-19 pandemic and the resulting economic fallout likely posed particular financial strain on American households with children, as they potentially faced income disruptions from widespread jobs and hours cuts in addition to new childcare and instruction demands. Indeed, the Census Pulse Survey indicates that during the pandemic, households with children experienced heightened food insecurity and trouble meeting their expenses relative to households without.2 One common expense for many households is their student loan payment. According to the 2019 Survey of Consumer Finances (SCF), nearly one-third of households with children have a student loan. The Coronavirus Aid, Relief, and Economic Security (CARES) Act included provisions to curb the impacts of these payments, which were subsequently extended and expanded several times through the pandemic. This chapter reviews the implications of these provisions for families with children. II. Student Loans Borrowing for education has become increasingly common in recent decades, with about onehalf of both US undergraduate and graduate students relying on loans to help finance their education.3 According to official Federal Reserve statistics, there is currently over $1.7 trillion dollars in student loan debt. This figure exceeds the cumulative amount outstanding from any other source of non-mortgage household debt. Nearly all of this student loan debt is held or backed by the federal government, with a large majority originated under one of two lending programs established under Title IV of the Higher Education Act of 1965—the Federal Direct Loan (DL) Program and the (now-defunct) Federal Family Education Loan Program (FFELP).4 Undergraduate Stafford Loans, the main loan type offered through the DL program and FFELP, are subject to statutory borrowing limits and feature a congressionally set, often below-market interest rate and standardized terms. The standard repayment plan has fixed monthly payments paid over 10 years, though other repayment plans are available—most notably, low-income borrowers may be eligible for an income-driven repayment (IDR) plan with a longer term, reduced payments, and potential loan forgiveness for the remaining balance at the end of the term.5 According to standard economic models, this structure should encourage human capital investment by helping remove credit 2 See, for example: https://www.census.gov/library/stories/2021/08/economic-hardship-declined-in-householdswith-children-as-child-tax-credit-payments-arrived.html. 3 See https://nces.ed.gov/programs/digest/d19/tables/dt19_331.20.asp and https://nces.ed.gov/programs/digest/d19/tables/dt19_332.30.asp. 4 Under FFELP, loans were issued by private lenders but federally guaranteed, meaning that in the event that a borrower defaults, the federal government pays the bank and takes over the loan. The federal government pays approximately 97 percent of the principal balance to the lender. At that point, the federal government owns the loan and the right to collect payments on the loan. 5 While conceptually useful for borrowers struggling to make timely student loan payments, administrative hurdles have generally made it difficult for borrowers to enroll and/or remain in IDR programs. See: https://www.brookings.edu/wp-content/uploads/2022/03/IDR-student-loan-report.pdf.

3 constraints from the decision to attend college, subsidizing its cost, and making resources available to students at a point in the lifecycle when credit is generally scarce. Companion lending programs also exist for graduate and parent borrowers, with the principal differences being that so-called PLUS Loans are neither subsidized nor subject to the statutory borrowing limits. These programs have grown in importance over time, with each accounting for around 5 percent of 2021 unconsolidated federal student debt.6 There is also a small private student loan market, representing less than 10 percent of outstanding student debt. Private student loans do not entail standardized terms and rates and often require an established credit record or a cosigner. The vast majority of students that finance their education with private loans also use federal loans. Compared with other forms of credit, including student loans made through the private sector, federal student loans are typically available to anyone who meets the basic eligibility criteria for federal financial aid programs, and there are no collateral requirements. Undergraduate borrowers can have relatively thin or adverse credit histories, while eligibility for loans to parents and graduate students is a bit more stringent, requiring the borrower is not delinquent or in collections on a large sum of debt and has not recently had a major negative credit event (e.g., bankruptcy). But, due to this weak eligibility criteria, the penalties for nonpayment are severe and include garnishment of wages, federal benefits, and tax refunds, with the loans extremely difficult to discharge in bankruptcy. III. Student Loan Pandemic Relief Measures The CARES Act, signed into law on March 27, 2020, included several pandemic relief measures to help federal student loan borrowers.7 These measures initially applied to “ED-owned loans”— that is, loans made under the Direct Loan program and the subset of loans made under FFELP that were held by the Department of Education—which together compose about 80 percent of the student loan market. For these loans, the Act mandated (1) a suspension of loan payments (i.e., administrative forbearance);8 (2) a zero percent interest rate on outstanding balances, and (3) a suspension of collection activities on defaulted loans. These measures technically went into effect March 20, 2020—ahead of, for example, the cash stimulus families received—and initially were in effect through September 30, 2020. These measures have subsequently been extended 8 times: on August 8, 2020 through December 31, 2020; on December 4, 2020 through January 31, 2021; on January 20, 2021 through September 30, 2021; on August 6, 2021 through January 31, 2022; on December 22, 2022 through May 1, 2022; on April 6, 2022 through August 31, 2022; on August 24, 2022 through December 31, 2022; and on November 22, 2022 through 60 days after either the Administration’s debt relief plan has been implemented (or the litigation has been 6 Factoring in the loan type at origination for consolidated debt would produce larger shares, particularly for graduate PLUS loans. 7 For more information, see https://studentaid.gov/announcements-events/covid-19. 8 According to Department of Education data, 20 million federal student loan borrowers were in repayment before the pandemic and were able to remain current even if they stopped making payments. This figure increases to 26 million if borrowers in forbearance or deferment due to economic hardship are included. See: https://studentaid.gov/sites/default/files/fsawg/datacenter/library/PortfoliobyLoanStatus.xls.

4 resolved) or June 30, 2023.9 They were also expanded over the course of the pandemic, most notably, on March 30, 2021, to make federal student loans made through FFELP that were not ED-owned but in default eligible for the zero percent interest rate on outstanding balances and the stoppage of collections activities (retroactive to March 13, 2020), and on April 6, 2022, to reinstate paused defaulted and delinquent federal loans to good standing upon expiration of the pandemic relief measures. Outside of these measures, any relief for payments on student loans held by private entities over the pandemic—that is, FFELP loans that are not held by the Department of Education and private student loans—was subject to the lender’s discretion and, if available, only upon request and for much shorter durations. Market participants in these segments have indicated in their earnings calls that forbearance rates peaked during the second quarter of 2020, at rates just below 30 percent and 15 percent, respectively. IV. Related Work Our analysis will principally focus on the implications of the suspension in loan payments and collection activities for families with children.10 A few related short pieces, all leveraging consumer credit records, have been written on these provisions more generally. Case, Hannon, and Mezza (2022), which developed some of the technology from which our analysis will draw, identified student loan borrowers that were making payments ahead of the pandemic (a proxy for eligibility for automatic forbearance) and found improvements in their credit profiles over the pandemic. They also estimated over $80 billion in aggregate savings through Spring 2022 among borrowers in this group whose balances were unchanged over the pandemic. In a complementary analysis, Goss, Mangrum, and Scally (2022) estimated around $200 billion in aggregate payments were waived under automatic forbearance among all borrowers that were eligible (whether or not they were current on their debt entering the pandemic) and found evidence that the student loan borrowers that were required to make payments struggled. Finally, Conkling, Gibbs, and Jimenez-Read (2022) identified potential risk factors as borrowers emerge from the pandemic forbearance and found that over 5 million borrowers have at least two of five potential risk factors and that borrowers with multiple risk factors are more likely to live in low-income or high-minority census tracts.11 V. Data Our analysis derives from two nationally representative datasets, the Federal Reserve Board’s Survey of Consumer Finances (SCF) and the Federal Reserve Bank of New York Consumer Credit Panel/Equifax (FRBNY CCP/Equifax). The SCF is a triennial survey of US families with rich detail on their current balance sheet, financial characteristics, and demographics that was 9 This sentence has been modified to reflect two additional extensions since the preparation of our analysis. 10 While not our express focus in this chapter, at least one short piece has analyzed the zero interest rate provision as a coarse form of loan forgiveness (Center for a Responsible Federal Budget, 2022), a topic that has remained central to student loan policy discussions through the pandemic. 11 Risk factors considered are: (1) delinquency on student loans in February 2020, (2) receipt of payment assistance prior to the pandemic, (3) multiple student loan servicers, (4) delinquency on active non-student loan products from April 2020 to February 2022, and (5) new non-medical collections reported from April 2020 to February 2022.

5 most recently conducted in 2019. The FRBNY CCP/Equifax is an anonymized administrative panel drawn from the universe of consumer credit records with rich detail on interactions with credit markets and borrowers’ age and geography, monthly from December 2019 through December 2021. Both datasets provide important detail for describing U.S. households with student debt, including balances and required monthly payments, and offer complementary insight into our question of interest. The SCF allows insight into the composition, demographics, and comprehensive financial positions of these households just ahead of the pandemic, as well as the ability to distinguish between federal and private student loans and among student loan borrowers who are not repaying, and those that are in school from those not repaying for other reasons. However, it does not provide a lens into changes after the onset of the pandemic and, because of its design, excludes some segments of the student loan market.12 The FRBNY CCP/Equifax provides that lens and generally has better coverage of the student loan market but is limited to characteristics available on credit reports and does not allow for direct identification of families or their demographics or for distinction between private and federal student loans or the rationale behind nonpayment. VI. SCF Analysis The first question is whether the relief measures were beneficial to families with children. In broad terms, using whether a family had a federal student loan in the 2019 SCF as the criterion for eligibility, the relief measures were more likely to reach families with children than families without: a greater share of eligible families included children, and a greater share of families with children was eligible relative to families without.13 Per the former, over half of families with student loan debt included at least one child under 18 years old (53 percent), and nearly all of these families had a federal student loan (86 percent); all in, 52 percent of eligible families were families with children.14 Per the latter, one-quarter of families with children had a federal student loan, compared to 15 percent of families without children.15 Another question is whether the families with children that were eligible for relief had relatively fragile financial positions, particularly in relation to other families with children (Table 1).16 The 12 The SCF sampling frame is household-based and within the household, principally collects data on the primary economic unit—that is, the economically dominant individual or family (if several individuals are financially interdependent) in a surveyed household. It excludes student loan debt held by those in institutional settings (e.g., dormitories) and by roommates or live-in relatives, including children, that are financially independent of this unit. As a result, our analysis of the SCF may miss some younger families with student loan debt and families with student loan debt that reside in multi-family households. 13 This method will count families whose only federal student loans were non-defaulted FFELP loans held by private lenders as eligible for relief when they were not. 14 A similar share of families with student loan debt that did not have children had a federal student loan. 15 About one-fifth of families reported that they were not making payments due to being in school or a payment grace period, and this fraction is relatively stable whether we restrict to families with federal student loan debt or not. Presumably, many of the families that were in school or grace in 2019 would have moved into repayment some time over the past two years, but assuming none have and taking this nonpayment into account, at minimum, 20 percent of families with children, and 12 percent of families without, benefitted from the relief measures. 16 The SCF analysis focuses on medians due to the potential influence of extreme values on averages for some of the groups with relatively small sample sizes.

6 typical family with children earned about $75,000 per year.17 Among them, families eligible for relief tended to out-earn ineligible families by about $11,000—$81,450 versus $70,250—and had greater monthly overhead from their debt (e.g., mortgages, credit cards, car loans, student loans). Families eligible for relief typically paid about $1,400 per month in total to service their debt, while ineligible families paid about $800 per month. Importantly, the median monthly debt service among eligible families includes those that were not making payments on their student loan debt, either because they were not required to or because they could not afford to, who thus would have a $0 student loan payment.18 Of those that were making student loan payments, the typical family with children paid $200 per month toward their loans, such that the automatic forbearance left them, all else equal, with as much as an extra 3 percent of monthly income at their disposal to save or spend.19 Further, the 10 percent of eligible families that could not afford their payment likely still benefited financially from the stoppage of collections activities, better terms on other credit applications (due to the reinstatement of previously delinquent student loans to current status and the likely corresponding improvement in credit health), and/or the zero interest provision.20 A last question is the extent to which the relief reached families with children by race/ethnicity and income (Table 2). Zooming in first on a family’s race/ethnicity, among families with children, White families were the most likely to be eligible for pandemic relief (representing 31 percent of all eligible families, both with and without children), followed by Black families, then Hispanic families.21 With respect to families with children of a given race/ethnicity, Black families had the greatest share eligible (33 percent versus 17 percent and 16 percent of White and Hispanic families, respectively). Among these eligible families, the median student loan debt was also highest for Black families ($30,000 versus $20,000 and $14,000 for White and Hispanic families, respectively), while their median income was the lowest of the 3 groups ($61,000 versus $93,000 and $76,000, respectively). With respect to savings on monthly payments from the automatic forbearance, eligible Black families typically saw less: the median student loan payment for an eligible Black family was $30 versus $150 and $80 for eligible White and Hispanic families, respectively, largely reflecting a greater share that were not making payments (53 percent versus 33 percent and 42 percent, respectively). Among those making payments, the median monthly student loan payment was $160 per month for eligible Black families, $200 per month for eligible White families, and $140 per month for eligible Hispanic families. About 11 percent of eligible Black families, 7 percent of eligible White families, and 14 percent of eligible Hispanic families could not afford payments and likely experienced some financial benefits from the other relief measures. 17 This section’s analysis measures income using “usual income”—that is, a family’s earnings in a “normal year” for them, net of any temporary deviations in the survey year. 18 Including these families, the typical student loan payment was $110. 19 These figures are upper bounds, as not necessarily all the student loan payments within an eligible family were eligible for the automatic forbearance. 20 The SCF classifies families that were making zero payments under IDR plans as not able to afford their payment. 21 A family’s race/ethnicity is classified based on the race/ethnicity of the respondent. This discussion excludes families in which the respondent’s race/ethnicity was classified as “Other.”

7 Classifying families with children next according to their income, families in the top-middle income quartile among families with children were the most likely to be eligible for pandemic relief (representing one-fifth of all eligible families, both with and without children), followed by the bottom-middle quartile, the top quartile, and the bottom quartile. Eligibility for relief within an income group takes on an inverse U shape over the distribution: we estimate 17 percent of families in the bottom income quartile, 28 percent of families in the bottom-middle quartile, 33 percent of families in the top-middle quartile, and 22 percent of families in the top quartile were eligible for relief. Among these eligible families, the median student loan debt is highest among the top quartile, relatively stable between the two middle quartiles and lowest among the bottom quartile. Unsurprisingly, median student loan payments rise with income, with eligible families in the bottom quartile typically not making payments on their student loan debt pre-pandemic. Somewhat more surprisingly, this relationship holds even conditioning on families that were making payments, despite the likelihood of making payments also increasing with income; that said, the maximum share of monthly income leftover to save or spend due to automatic forbearance decreases over the distribution, once pre-pandemic payment activity is taken into account. VII. FRBNY CCP/Equifax Analysis The FRBNY CCP/Equifax analysis evaluates how student loan borrowers eligible for pandemic relief fared over the pandemic, drawing comparisons between 3 mutually exclusive subsamples of student loan borrowers that faced different intensities of relief: 1) eligible student loan borrowers that were current (or delinquent) on their student loan debt just before the pandemic, who therefore would be likely to have received automatic forbearance;22 2) eligible student loan borrowers that were in default just before the pandemic, who therefore would be likely to have had their collections activities stopped; 3) and ineligible student loan borrowers.23 Within each of these subsamples, a borrower is associated with a likelihood that they are in a low-income family 22 This subsample contains about 24.2 million borrowers, in line with the published statistics in footnote 6, with average and median required monthly payments in 2019:12 of $197 and $99, respectively, and it includes borrowers making payments as well as borrowers not making payments on their student loan debt prior to the pandemic. The first group is, all else equal, more likely able to save or spend more of their income due to the automatic forbearance provision. Potential reasons for zero payment include enrollment in deferment due to economic hardship, forbearance, or an IDR program with low enough income such that payments were not required on the loan, and the circumstances governing these different statuses might have extended into or through the pandemic. 23 These groups must be inferred from the detail in credit records, and to belong to our sample, a borrower must have student loan debt. For simplicity, the analysis removes borrowers whose most recent loan was originated within the 365 days before December 2019 and who are less likely to be in repayment during the pandemic. Subsample 1 excludes borrowers with only defaulted loans, cosigned loans (which are likely to be issued by private lenders), or loans in which payments were due over the automatic forbearance period, defined as the period between August 2020 and December 2021 (which are likely to either be non-cosigned private student loans or FFELP loans that were ineligible for relief). Subsample 2 imposes most of the same restrictions, but the borrower must not belong to subsample 1 and must have a non-cosigned loan that was in default prior to the pandemic. Subsample 3 contains the remaining student loan borrowers, who likely exclusively held FFELP and/or private student loans. Most private student loans are cosigned, especially those originated after the Great Recession. For example, according to MeasureOne, about 91 percent of private undergraduate loans and 63 percent of private graduate loans were cosigned in the 2020–21 academic year. See https://fs.hubspotusercontent00.net/hubfs/6171800/assets/downloads/MeasureOne%20Private%20Student%20Loan %20Report%20Q3%202021%20FINAL%20VERSION.pdf. Borrowers with at least one federal loan in default are usually in default on all their federal loans.

8 with children based on their neighborhood of residence, and those most likely to be in such a family are separately analyzed.24 Table 3 describes the pre-pandemic student debt held by each of the borrower subsamples, further partitioning them by the likelihood they belong to a low-income family with children. Expectedly, a large majority of outstanding student debt at the end of 2019 was held by borrowers eligible for automatic forbearance. Average debt is similarly much higher for borrowers in this group than the others. Across those groups, the amount of student debt held by borrowers eligible for their collections activities to be stopped was just below that held by ineligible borrowers, but their average debt was almost $7,000 higher (an artifact of relatively few borrowers in subsample 2 in comparison to subsample 3). Notably, with limited exception, a family’s average amount of student loan debt decreases in the likelihood they are a low-income family with children. This pattern is likely a byproduct of lower-income families generally having less postsecondary education. Turning to how borrowers in the different subsamples fared over the pandemic, first, credit metrics for borrowers eligible for automatic forbearance broadly improved, with evolutions in credit among low-income families with children largely tracking the broader group of eligible families in subsample 1. First, interaction with other credit markets increased (Figures 1-4), on balance, with a notable increased in the likelihood of having a mortgage, a proxy for homeownership. In 2019:12, 26 percent of families eligible for automatic forbearance, and 20 percent of eligible low-income families with children, had a mortgage, versus 31 percent and 24 percent, respectively, in 2021:12. Interestingly, the likelihood of having a credit card also increased several percentage points for both groups, but most of the increase took place over 2021. There was little change, on balance, within either group in aggregate credit card debt or the likelihood of having an auto loan.25 Turning to nonpayment of other types of debt, first, it is worth noting that credit card, auto loan, and mortgage debt held by low-income families with children is generally more likely to be delinquent (more than 30 days past due) than debt held by the broader group of eligible families in subsample 1 (Figure 5). Still, patterns in delinquency are similar. For auto loans and mortgages, nonrepayment declined sharply in the first half of the pandemic, coincident with participation in forbearance programs (Figure 6), but has since remained unchanged, even as borrowers exited forbearance. For credit cards, delinquency rates are, on balance, below prepandemic levels, but they rose toward the end of each calendar year, and more so for low-income families with children. Finally likely reflecting the increases in repayment across the board, overall credit health—as measured by the Equifax Risk Score 3.0—improved both overall for 24 Family composition and income are not observable in credit records. Using an external dataset derived from tax records for the 2019 tax year, zip codes are grouped into quartiles based on the proportion of households in that zip code that are low-income families with children, where low income is defined as having an adjusted gross income below $75,000 and the presence of children is measured based on the election of the child and other dependent credit. Student loan borrowers that resided in the top quartile zip codes according to their 2019:12 address in the credit data are classified as “low-income families with children” for separate analysis. 25 Dollar amounts are expressed in real 2021:12 dollars, which may exaggerate changes in balances due to the inflationary environment.

9 this group and for the subset of low-income families with children (Figure 7), as did both groups’ success rates in obtaining new credit (Figure 8).26 Credit metrics for borrowers eligible for the stoppage of collections activities also saw improvements, dramatically so for their overall credit health and in their success in getting new credit (Figures 15-16). That said, as would be expected, their credit metrics were generally worse than those eligible for automatic forbearance, before and throughout the pandemic. And, as was the case for those eligible for automatic forbearance, metrics for low-income families with children were worse than those in the broader sample, but the two groups also evolved similarly over the pandemic. Notably, increases in the likelihood of having a credit card and in having an auto loan were larger for the groups in subsample 2 than those in subsample 1, though increases in the likelihood of having a mortgage were near-negligible (Figures 9-12). Delinquency rates were also less differentiated between subsample 2 and the subset of subsample 2 composed of low-income families with children (Figure 13), probably because all of the families in subsample 2 were likely in financial distress to some extent (given they were in default on their student loan debt). With respect to changes in nonpayment over the pandemic, mortgage delinquency rates dropped precipitously over 2020, likely reflecting movement into forbearance (Figure 14). In contrast to subsample 1, forbearance rates on mortgages for these groups were much more persistent. Decreases in auto loan delinquencies were less stark than in subsample 1, but enrollment in forbearance was similarly short-lived. Other pandemic relief measures (e.g., federal mortgage forbearance, the stimulus checks, more generous unemployment insurance, the expanded child tax credit) may have driven some of the changes in the credit metrics of student loan borrowers over this period. One way to try to parse out the role of the federal student loan relief is to examine the credit market interactions of student loan borrowers who were ineligible for relief, noting, of course, that this group is composed of student loan borrowers with, on average, older FFELP loans and/or private student loans and, as such, likely differ in important ways from borrowers eligible for relief. For instance, the ineligible groups clearly had better overall credit health pre-pandemic than those eligible for automatic forbearance (or those eligible for the stoppage of collections). Relative to those eligible for automatic forbearance, the ineligible group did not appear to see the same improvements in their credit metrics during the pandemic. The use of other forms of credit was either unchanged or decreased, on balance, for both the full set of ineligible families as well as those that were low-income families with children (Figures 17-20). Nonpayment was somewhat improved in each credit category (Figure 21), but less so than for families eligible for automatic forbearance, while forbearance on auto loans and mortgages was similarly persistent (Figure 22). Reflecting the more-muted changes among ineligible families, overall credit health improved to a lesser extent (Figure 23), both for this group overall and for the subset of low-income families with children, as did their success rates in obtaining new credit (Figure 24). All in all, these dynamics suggest that the improvements in credit health experienced by eligible borrower groups over the pandemic can be at least partially attributed to the federal student loan relief measures.27 26 The success rate measures the fraction of individuals opening a new account in the past 12 months, given at least one inquiry in the past 12 months. 27 There is no adjustment for differences in baseline characteristics in this assessment.

10 VIII. Discussion The majority of pre-pandemic outstanding student loan debt was eligible for federal student loan pandemic relief, which provided automatic forbearance, stoppage of collection activities on defaulted loans, and zero interest accumulation. These measures were not targeted and thus applied independent of need. That said, given the distribution of student loan debt, families with children were more likely to benefit: more than half of families with eligible student loans had children, and the relief reached a greater share of families with children than those without. Still, based on their income, families with children that were eligible for student loan relief were more well off than those with children that were not, likely at least partly driven by the fact that families with student loan debt tend to have more education than those without. Other findings echo this pattern. First, both analyses underscored that families with higher incomes tend to have more student loan debt. Second, families toward the bottom of the income distribution were less likely to be eligible for relief (i.e., have student federal loan debt).28 Third, families with higher incomes likely saw more savings from the automatic forbearance provision, as they tend to have higher monthly payments, due partly to having more debt and partly to reduced eligibility for IDR programs. Among families with children, White families were more likely to be eligible for relief, followed by Black, and then Hispanic families, presumably driven by the composition of the U.S. population and, in particular, those who have pursued a post-secondary education. A much larger share of Black families with children were eligible for the relief than Hispanic or White families with children, and such families typically had the most student loan debt and the lowest income. That said, they also typically had the lowest monthly payments, implying less savings from the automatic forbearance provision in particular. Based on credit records, student loan borrowers eligible for federal student loan pandemic relief, including those in higher risk groups, were able to improve their credit profiles. Some of this improvement likely owed to other fiscal stimulus and relief measures, but the analysis indicates outsized improvement for those eligible for federal student loan pandemic relief relative to those that were not, indicating some of the improvement in overall credit health and interactions with other credit markets may be attributable to the cessation of student loan payments and collections activities. 28 Families with children in the top quartile are moderately less likely to be eligible than those in the middle segments, as, while such families are presumably highly educated, they are also less reliant on student loans for financing and/or more likely to have repaid the loans they had.

11 Work Cited Case, H., Hannon, S., and A. Mezza (2022). “A Note on the Expected Expiration of Federal Student Loan Forbearance.” FEDS Notes. Washington: Board of Governors of the Federal Reserve System, May 27, 2022. Committee for a Responsible Federal Budget (2022). “How Much Student Debt Has Already Been Cancelled?” [Blog Post]. Retrieved from https://www.crfb.org/blogs/how-much-studentdebt-has-already-been-cancelled. Conkling, T., Gibbs, C., and V. Jimenez-Read (2022). “Student Loan Borrowers Potentially At- Risk When Payment Suspension Ends.” Consumer Financial Protection Bureau, April 2022. Goss, J., D. Mangrum, and J. Scally (2022). “Student Loan Repayment During the Pandemic Forbearance.” Federal Reserve Bank of New York Liberty Street Economics, March 22, 2022.

12 Tables and Figures Table 1. Pre-Pandemic Financial Positions of Families with Children, By Eligibility for Pandemic Relief Eligible Ineligible Median student loan debt $20,000 $0 Annual Income $81,450 $70,250 Median monthly debt service $1,430 $820 Median monthly debt service, ex student loan payments $1,230 $810 Median monthly student loan payment $110 $0 Median share of annual income 2% -- Median monthly student loan payment, cond. on making payments $200 -- Median share of annual income, cond. on making payments 3% -- Share not making student loan payments 40% -- Share not making student loan payments because could not afford to 10% -- Note: Eligibility for pandemic relief inferred based on the presence of a federal student loan. Student loan payment includes all student loans, whether or not they would be eligible for automatic forbearance. Source: Survey of Consumer Finances, 2019.

13 Table 2. Pre-Pandemic Financial Positions of Families with Children that were Eligible for Pandemic Relief, By Risk Group Race/Ethnicity Income Quartile White Non-Black Non- Bottom Top Hispanic Hispanic Hispanic Bottom Middle Middle Top Share of all eligible families in the given category 31% 10% 5% 8% 15% 17% 12% Share in the given category that were eligible 26% 33% 16% 17% 28% 33% 22% Median student loan debt $20,000 $30,000 $14,000 $16,000 $20,000 $19,000 $22,000 Annual Income $92,660 $61,090 $76,360 $28,470 $54,570 $99,250 $178,170 Median monthly debt service $1,620 $640 $1,210 $340 $780 $1,690 $2,940 Median monthly debt service, ex student loan payments $1,420 $540 $1,210 $270 $670 $1,530 $2,620 Median monthly student loan payment $150 $30 $80 $0 $60 $160 $230 Median share of annual income 2% 0% 1% 0% 2% 2% 1% Median monthly student loan payment, cond. on making payments $200 $160 $140 $110 $220 $270 $420 Median share of annual income, cond. on making payments 3% 3% 2% 6% 3% 3% 2% Share not making student loan payments 33% 53% 42% 58% 45% 32% 30% Share not making student loan payments because could not afford to 7% 11% 14% 23% 13% 7% 1% Note: Eligibility for pandemic relief inferred based on the presence of a federal student loan. Families whose race/ethnicity were classified as "Other" excluded from table. Income quartiles derived from distribution of families with children. The top row includes eligible families without children in the denominator. Student loan payment includes all student loans, whether or not they would be eligible for automatic forbearance. Source: Survey of Consumer Finances, 2019.

14 Table 3. Pre-Pandemic Student Loan Debt, By Eligibility for Pandemic Relief Total (Billions) Average Student loan borrowers eligible for automatic forbearance (subsample 1) $930 $38,362 Likelihood of being a low-income family with children based on zip code Highest $192 $34,926 Middle-High $202 $36,246 Middle-Low $210 $37,041 Lowest $310 $43,783 Student loan borrowers eligible for their collections activities to be stopped (subsample 2) $122 $28,184 Likelihood of being a low-income family with children based on zip code Highest $33 $23,434 Middle-High $32 $28,604 Middle-Low $25 $28,225 Lowest $28 $34,640 Ineligible student loan borrowers (subsample 3) $136 $21,459 Likelihood of being a low-income family with children based on zip code Highest $20 $20,991 Middle-High $25 $20,454 Middle-Low $30 $19,196 Lowest $58 $23,009 Note: Dollar values expressed in 2021:12 dollars. Not all student loan debt held by borrowers in subsamples 1 or 2 was necessarily eligible for the federal student loan pandemic relief. Groups defined as described in the main text. Source: FRBNY CCP/Equifax, 2019:12; Internal Revenue Service, Statistics of Income, Individual Income Tax Statistics by Zip Code, 2019.

15

16

17

18