Policy 09: Bring back bankruptcy, including on student loans.

Researched: 2026-07-22. Status: draft.

The plank

Verbatim from README.md: "Policy 9. Bring back bankruptcy, including on student loans." — title only; the README gives no body text. app/templates/index.html matches (Policy 09 is also title-only), so there is no drift on this plank. (Unrelated to this plank: index.html's Policy 04 heading, "Separate healthcare and employment," differs from the README's "Disintermediate healthcare and employment" — that is a drift bug to fix in the Policy 4 doc's pass.)

Constructed meaning, from the title: restore a meaningful consumer bankruptcy fresh start. Two parts: (1) repeal the special nondischargeability of student loans — federal and private — so education debt is treated like other unsecured consumer debt, and (2) unwind the parts of the 2005 bankruptcy law (BAPCPA) that priced ordinary people out of the bankruptcy system itself. This doc treats the student-loan carve-out as the core and BAPCPA repair as the second front.

The problem

Bankruptcy is America's 230-year-old mechanism (Art. I, §8, cl. 4) for letting honest but overwhelmed debtors start over. Two legislative campaigns broke it for ordinary people:

1. Student loans were carved out of the fresh start, one ratchet at a time.

• Before 1976, student loans were dischargeable like any unsecured debt.
• 1976 Education Amendments: federally backed student loans nondischargeable for the first 5 years of repayment, absent "undue hardship" (now 11 U.S.C. §523(a)(8)).
• 1984: extended to loans funded or guaranteed by nonprofits.
• 1990 (Crime Control Act): waiting period stretched to 7 years.
• 1998 Higher Education Amendments: waiting period eliminated entirely — federal loans nondischargeable forever, absent undue hardship.
• 2005 BAPCPA: extended nondischargeability to private "qualified education loans" — loans with no income-driven repayment, no forgiveness programs, no federal safety valves at all.

"Undue hardship" is undefined in the statute. Most circuits apply the Brunner test (Brunner v. N.Y. State Higher Educ. Servs. Corp., 831 F.2d 395, 2d Cir. 1987): the debtor must show (1) she cannot maintain a minimal standard of living while repaying, (2) that state of affairs is likely to persist, and (3) she made good-faith repayment efforts. It requires a separate lawsuit (an adversary proceeding) inside the bankruptcy, which most debtors' flat-fee attorneys do not file.

The result, quantified: roughly 250,000 student-loan debtors file bankruptcy each year, and fewer than 300 even attempt to discharge their student loans (Iuliano, "The Student Loan Bankruptcy Gap," Duke Law Journal, 2020). In his earlier nationwide sample, only 0.1% of student-loan debtors in bankruptcy attempted discharge — but roughly 40% of those who tried won full or partial relief (Iuliano, "An Empirical Assessment of Student Loan Discharges and the Undue Hardship Standard," American Bankruptcy Law Journal, 2012). The wall works mostly by deterrence, not by adjudication.

Meanwhile the debt it traps has grown to systemic scale:

• $1.66 trillion in outstanding student loan balances (NY Fed Household Debt & Credit Report, 2026 Q1).
• Roughly 42–43 million Americans hold federal student loans (ED Federal Student Aid portfolio data, 2025; approximate).
• After the pandemic pause ended and delinquency reporting resumed in 2025: 10.3% of student loan balances were 90+ days delinquent in 2026 Q1 (up from 9.6% in 2025 Q4); roughly 1 million borrowers defaulted in 2025 Q4 and another 2.6 million in 2026 Q1; over 17% of borrowers have gone 90+ days past due at least once since reporting resumed in 2025 (NY Fed, Q1 2026 report and Liberty Street Economics, May 2026).

A defaulted federal borrower faces administrative wage garnishment, tax-refund seizure, and Social Security offset — with no statute of limitations and, uniquely among major consumer debts, no bankruptcy exit. That is how the status quo shows up in an ordinary life: a nurse's aide who dropped out with $20,000 of debt in 2009 can have her paycheck garnished in 2026 for a balance that has grown, not shrunk.

2. BAPCPA (2005) made bankruptcy itself costlier and scarcer for everyone.

• It imposed a Chapter 7 means test, mandatory credit counseling, and heavier attorney paperwork/liability. GAO found the average Chapter 7 attorney fee rose from $712 (Feb–Mar 2005) to $1,078 (Feb–Mar 2007), and filing fees rose from $209 to $299 for Chapter 7 (GAO-08-697, 2008). Implementing BAPCPA cost the U.S. Trustee Program ~$72.4M and the federal judiciary ~$48M in its first two years (same report).
• Filings collapsed from a record 2,078,415 in 2005 to 617,660 in 2006 (Administrative Office of the U.S. Courts / ABI, 2006–07). Some of that was the pre-deadline rush, but filings stayed structurally lower — while financial distress did not disappear, it just stopped being resolved.
• The perverse outcome: the people most likely to be screened out by fees and paperwork are the poorest debtors — the ones bankruptcy exists for.

Who profits from the status quo

• The credit-card industry (BAPCPA's authors). Over roughly eight years beginning in the late 1990s, the industry spent more than $100 million lobbying for what became BAPCPA — at the time the most expensive financial-industry lobbying campaign on record (widely documented; see Missouri Law Review BAPCPA symposium, 2006, and contemporaneous reporting; exact total is an advocacy-side tally, treat as approximate). What they bought: Simkovic ("The Effect of BAPCPA on Credit Card Industry Profits and Prices," American Bankruptcy Law Journal, 2009) found that after 2005, credit-card charge-offs fell and industry profits hit records, but the savings were not passed through to consumers in lower rates or fees.
• Sallie Mae / SLM and successor Navient. Sallie Mae was a leading lobbying force for BAPCPA's private-loan nondischargeability provision (reported contemporaneously; specific dollar figure for that provision: unverified). Nondischargeable private paper is more valuable paper: it can be priced like risky unsecured credit while carrying near-immunity from the debtor's main legal escape. Recent lobbying spend: Navient ~$1.7M (2020) and ~$970K (2021); Sallie Mae ~$1.4M (2020) and ~$960K (2021) (OpenSecrets, 2021). Navient's conduct record: sued by CFPB in 2017 for servicing failures; a 2022 39-state settlement cancelled $1.7 billion of subprime private loans; a 2023 settlement reportedly wiped ~$182M of loans that should have been dischargeable (reported; verify figure); a 2024 CFPB order imposed a further ~$120M in redress and penalties (reported; verify).
• Federal loan servicers. Five companies hold ED's Unified Servicing contract, announced 2023, worth up to ~$16 billion over 10 years: Maximus (Aidvantage, servicing ~9 million borrowers), Nelnet, MOHELA, EdFinancial, Central Research (ED announcement, 2023; Government Executive, 2024). Nelnet alone reported $277.7M of ED servicing revenue in the first nine months of 2024 (Nelnet 10-Q, Q3 2024). Servicers are paid per borrower per month; a system where debt can never be extinguished, only endlessly serviced, rehabilitated, and re-defaulted, is their revenue base.
• Debt buyers and collectors of both private student loans and post-BAPCPA credit card debt, whose collateral is precisely the debtor's inability to discharge.

This section survives a defamation review because it claims only documented lobbying expenditures, public contract values, SEC-filed revenues, and settled/adjudicated enforcement actions — not intent.

The proposal, concretely

1. Amend 11 U.S.C. §523(a)(8) to strike student loans from the exceptions to discharge — restoring pre-1976 treatment. Discharge would still require a bankruptcy: means test, trustee scrutiny, asset surrender in Chapter 7 or a 3–5 year plan in Chapter 13, and a 7–10 year credit-report scar. Bankruptcy is not free money; it is a court-supervised loss-recognition process.
2. Fallback / phased version: the FRESH START Through Bankruptcy Act (Durbin (D-IL)–Cornyn (R-TX), S.2598, 117th Congress; reintroduced since, never brought to a floor vote). Federal student loans become dischargeable in ordinary bankruptcy 10 years after the first payment came due (mirroring pre-1998 law); undue hardship remains available before then; colleges with high default rates and heavy loan dependence must partially reimburse ED for discharged loans — a clawback that puts skin in the game on the school, where the pricing power sits.
3. Repeal BAPCPA's private-loan provision immediately (§523(a)(8)(B)). Private loans have no IDR, no forgiveness, no disability discharge guarantee; the case for special protection is weakest there, and this is the piece bought most directly by lobbying.
4. Codify the November 2022 DOJ/ED attestation process (currently only agency guidance, revocable by any administration): a standardized form keyed to IRS Collection Financial Standards, with a presumption of undue hardship for defined categories (age, disability, long-term unemployment, prolonged nonpayment).
5. BAPCPA repair for everyone: simplify or repeal the means test for below-median-income filers, cut filing fees for indigent debtors, and drop the credit-counseling ritual (completion rates near 100% show it screens nothing).

Where it's "already in use": the United States itself, pre-1976 (and, for the 10-year rule, pre-1998) — the mechanism is not exotic, it is the repeal of an exception.

Evidence

• The original justification was empirically false. GAO's 1977 study (HRD-77-83, sample of 606 bankruptcy claims paid FY1976) found less than 1% of federally insured/guaranteed student loans were discharged in bankruptcy, versus an 18% default rate — and that discharged debtors were generally bankrupt for other reasons, not gaming their student loans. Congress legislated against an abuse anecdote, not a measured problem. Strength: strong, primary.
• The wall deters, it doesn't adjudicate. Iuliano (2012): 0.1% of bankrupt student-loan debtors attempt discharge; ~40% of attempters succeed; winners are disproportionately unemployed, medically burdened, low-income — i.e., the standard finds real hardship when asked. Iuliano (2020): ~250,000 filers/year, <300 attempts. Strength: strong for the deterrence claim; based on original nationwide case datasets.
• When the government stopped fighting, the sky did not fall. Under the Nov. 2022 DOJ/ED guidance: 632 adversary cases in the first 10 months; 1,220 cases in the first 17 months; in 98–99% of decided cases the government recommended and courts granted full or partial discharge (DOJ status reports 2023; NCLC; Iuliano, "Bridging the Student Loan Bankruptcy Gap," ABLJ 2025–26, finds an 87% debtor success rate post-reform but filings still far below need). Cases in the low thousands, against 250,000 annual bankrupt student debtors, is the strongest available real-world test of the moral-hazard fear — and it shows a trickle, not a flood. Strength: moderate-strong; short window, selection effects possible.
• A fresh start has measured economic value. Dobbie & Song ("Debt Relief and Debtor Outcomes," American Economic Review, 2015; 500,000 filings, random judge assignment, linked tax records): winning Chapter 13 protection raises annual earnings by $5,562, cuts 5-year mortality by 1.2 percentage points, and cuts 5-year foreclosure by 19.1 percentage points versus dismissed filers. Caveat, conceded: this measures bankruptcy protection generally, not student-loan discharge specifically, and the effect is largely dismissed filers doing badly. Strength: strong internally; moderate as a mapping to this plank.
• Strategic default is real but small. Yannelis ("Strategic Default on Student Loans," working paper, 2016; and Darolia & Yannelis, "Strategic Default Among Private Student Loan Debtors: Evidence from Bankruptcy Reform," Education Finance and Policy, 2020, from Philadelphia Fed WP 17-38): around the 1998 reform, borrowers who retained discharge eligibility were about 0.25 percentage points (≈18% relative) more likely to default; the private-loan evidence around BAPCPA shows limited strategic response. So dischargeability does change behavior at the margin — by a quarter of a point, in a system now producing 2.6 million defaults a quarter. Strength: moderate; the best evidence the opposition has, and we cite it against ourselves.

Overall strength of evidence: moderate-to-strong. The deterrence and false-premise claims are strong; the projected behavior under full repeal rests on the 2022–25 natural experiment and pre-1998 history, both encouraging but not dispositive.

Who wins, who loses

Transition: discharge losses would front-load into the first 2–3 years (pent-up insolvency from the 2025–26 default wave). No compensation mechanism is proposed for lenders or servicers, and none is owed: they priced and lobbied for a legal privilege, not a property right. The FRESH START school-clawback is the only novel compensation flow, and it runs toward the Treasury.

Fiscal impact

No CBO score exists for §523(a)(8) repeal or for FRESH START (never reached a floor vote). Everything below is back-of-envelope, labeled as such.

• Base rates. Historical full-dischargeability experience: <1% of the federal portfolio discharged per year (GAO 1977). Applied naively to today's ~$1.6T federal portfolio (2025): an upper-band gross discharge of ≤$10–16B/year, and realistically far less, because bankruptcy filing itself gates access (only ~250K student debtors/year file at all; at an average federal balance of ~$38–39K (2025, approximate), even every filer discharging everything caps gross discharge near $9–10B/yr, and the 2022–25 attestation experience — ~1,220 cases in 17 months — suggests take-up in the low single-digit percentages of filers for years).
• Much of the "cost" is already lost. 10.3% of balances were 90+ days delinquent in 2026 Q1 (NY Fed); recoveries on deep-delinquent student debt come slowly via garnishment and offset, at high administrative cost, from people near insolvency. Discharge largely converts an unrecoverable receivable into an acknowledged loss — a recognition event, not a new outlay. (Precedent for the accounting gap: ED's own re-estimates have swung the direct-loan program's projected budget effect by hundreds of billions; GAO, 2022.)
• Offsets: reduced collection-contract and servicing outlays on defaulted paper; Dobbie & Song (2015)-type earnings recovery implies higher income-tax receipts and lower safety-net draw among discharged debtors (direction well-evidenced, magnitude unscored).
• Private-loan repeal costs the Treasury ~nothing (private losses fall on private lenders).
• Honest bottom line (estimate, weak evidence): plausible net federal cost in the low single-digit billions per year in the early years, declining as pent-up insolvency clears — small against a $1.6T portfolio and against the ~$100B+/yr the program already swings on IDR/forgiveness re-estimates. The platform should demand a CBO score of FRESH START as its first legislative ask; this section is the doc's weakest and says so.

The opposition's best case

1. "Dischargeability invites strategic default." Strongest proponent: Constantine Yannelis (Chicago Booth), whose 1998-reform evidence shows discharge availability raised defaults ~0.25pp (2016 wp; Darolia & Yannelis 2020). Answer: we accept the finding — the effect is real, and it is small: a quarter of a percentage point, versus a status quo now producing 2.6M defaults in one quarter (2026 Q1) with no discharge available at all. Bankruptcy's built-in costs (court scrutiny, asset surrender, 7–10 year credit scar) are the deterrent; FRESH START's 10-year wait removes the fresh-graduate strategic case entirely. Concession in writing: full repeal with no waiting period would raise defaults somewhat; the 10-year version is the defensible core.
2. "Lenders will ration credit and raise prices." Standard credit-supply economics (see Gross–Notowidigdo-style bankruptcy-protection literature). Answer: for federal loans, rates and access are set by statute — Congress, not risk models, prices them, so student access to federal aid is unchanged. For private loans, yes: rates rise and marginal underwriting tightens. That is risk pricing returning to a market that had legislated it away — and tighter private credit for low-value programs is partly the point. Concession: some borrowers who value private credit access will lose it or pay more.
3. "Easy credit fuels tuition inflation — don't make lending consequence-free for schools." Strongest version: Lucca, Nadauld & Shen (Review of Financial Studies, 2019): a $1 increase in subsidized federal loan caps passed ~60¢ into tuition (≈20¢ for unsubsidized). Answer: this cuts against expanding credit, not against dischargeability — if anything, dischargeability contracts risky credit. And FRESH START's clawback makes high-default schools repay discharged loans, aiming the incentive at the institution with the pricing power. Concession: dischargeability alone does nothing about tuition; it needs the clawback or accountability rules to touch the root cause.
4. "Taxpayers eat discharged federal loans." Proponents: BAPCPA-era ED and Treasury officials; fiscal hawks today. Answer: true, and unscored — see fiscal section. But taxpayers are already eating these losses slowly and expensively through default, garnishment machinery, and IDR write-downs; bankruptcy is the honest, judicially supervised way to book them, and GAO showed in 1977 that actual discharge rates under full dischargeability were under 1%. Concession: near-term deficit impact is real and we should get it scored before claiming precision.
5. "IDR already protects distressed borrowers; bankruptcy is redundant and blunter." Proponents: defenders of the current safety-valve architecture (e.g., Beth Akers, AEI, and similar). Answer: the 2025–26 restart is the counter-evidence — 3.6 million borrowers defaulted in two quarters (NY Fed 2026) with IDR on the books, because IDR is administratively fragile, litigation-whipsawed, and requires continuous paperwork from the least-resourced borrowers. And IDR does not exist at all for the private loans BAPCPA protected. Bankruptcy is the backstop for when the safety valves jam — which is now. Concession: for most merely-struggling federal borrowers, a functioning IDR is the better first resort; bankruptcy is for insolvency, not inconvenience.

Talking points

• One line: Every other honest debt in America can be discharged in bankruptcy — a casino loss can, a yacht loan can — but a nurse's student loan can't; bring back bankruptcy.
• Thirty seconds: Congress walled student loans off from bankruptcy in stages from 1976 to 2005, based on a fraud scare the government's own auditors had already debunked — GAO found under 1% of federal student loans were ever discharged. Today $1.66 trillion is outstanding, 3.6 million borrowers defaulted in just two quarters after payments restarted, and a quarter-million student debtors a year go through bankruptcy without relief. Restore discharge — at minimum the bipartisan Durbin–Cornyn 10-year rule — and let courts do what they did for 200 years.
• Two minutes: adds the villain and the honest cost. The 2005 bankruptcy bill was bought: the credit-card industry spent over $100 million lobbying for it, and research shows their losses fell and profits hit records while consumers' rates never dropped; Sallie Mae got private student loans — loans with no income-based repayment at all — made nondischargeable in the same bill. Servicers like Maximus and Nelnet sit on a $16 billion federal contract that pays them to administer debt that can never die. The fears are testable and tested: when DOJ stopped fighting hardship cases in 2022, about a thousand borrowers used the process in the first year and a half — 98% of decided cases won relief — a trickle, not a flood; and the best academic estimate of "strategic default" under dischargeability is a quarter of a percentage point. The honest cost: taxpayers book losses on discharged federal loans sooner — losses largely already baked in — and private lenders tighten risky lending; there is no CBO score yet and we should demand one. Economists Dobbie and Song showed a bankruptcy fresh start raises a debtor's earnings by about $5,500 a year and measurably cuts mortality. Debt relief for the insolvent isn't charity; it's how a market economy recycles people back into productivity.

Every figure above traces to the sourced claims in earlier sections.

Open questions

1. Is the November 2022 DOJ/ED guidance still in force as of mid-2026? The attestation form was updated May 2025, but the guidance is revocable policy, not law; verify current administration posture before using the 98% figure in public.
2. Get a CBO score. The fiscal section is a back-of-envelope; the plank's fiscal claim is not yet quantified to platform standard. An economist reviewer should check the discharge-rate and average-balance assumptions first.
3. Exact Sallie Mae lobbying spend on the BAPCPA private-loan provision — widely asserted, never pinned to a verified dollar figure here. Find FEC/lobbying-disclosure records or drop the specific attribution.
4. Verify the Navient 2023 ($182M) and 2024 (~$120M) settlement figures against primary CFPB/court documents.
5. Does the 87–98% success rate persist as volumes grow, or does it reflect early self-selection by the clearest hardship cases? Iuliano's post-2026 data will tell.
6. Private student loan pricing response: find or commission an estimate of the spread effect of restoring dischargeability (the 2005–07 natural experiment ran the reverse direction and, per Darolia & Yannelis, lenders' windfall did not visibly reach borrowers).
7. What we'd change our mind on: credible evidence that a no-waiting-period repeal produces discharge rates an order of magnitude above the pre-1976/GAO base rate would push us to the FRESH START 10-year version as the ceiling, not the floor.

Sources

1. GAO, Guaranteed Student Loan Program Bankruptcies, HRD-77-83 (1977). https://www.gao.gov/products/hrd-77-83 — primary.
2. 11 U.S.C. §523(a)(8) (current text; exceptions to discharge) — primary.
3. Brunner v. New York State Higher Education Services Corp., 831 F.2d 395 (2d Cir. 1987) — primary.
4. Congressional Research Service, Bankruptcy and Student Loans, R45113 (2018–19). https://www.congress.gov/crs-product/R45113 — primary (legislative history: 1976, 1984, 1990, 1998, 2005).
5. Jason Iuliano, "An Empirical Assessment of Student Loan Discharges and the Undue Hardship Standard," 86 Am. Bankr. L.J. 495 (2012). https://papers.ssrn.com/abstract=1894445 — primary (academic).
6. Jason Iuliano, "The Student Loan Bankruptcy Gap," 70 Duke L.J. 497 (2020). https://scholarship.law.duke.edu/dlj/vol70/iss3/1/ — primary (academic).
7. Jason Iuliano, "Bridging the Student Loan Bankruptcy Gap," 99 Am. Bankr. L.J. (2025–26). https://www.ablj.org/bridging-the-student-loan-bankruptcy-gap-vol-99-issue-3-html/ — primary (academic; 87% post-reform success rate, low filing volume).
8. DOJ, Guidance for Department Attorneys Regarding Student Loan Bankruptcy Litigation (Nov. 17, 2022). https://www.justice.gov/d9/pages/attachments/2022/11/17/student_loan_discharge_guidance_-_guidance_text_0.pdf — primary.
9. NCLC, "New Process to Discharge Student Loans in Bankruptcy" (updated 2023–25; 632 cases/10 months; 99% of decided cases discharged; 1,220 cases/17 months). https://library.nclc.org/article/new-process-discharge-student-loans-bankruptcy — secondary.
10. GAO, Bankruptcy Reform: Dollar Costs Associated with BAPCPA, GAO-08-697 (2008; attorney fees $712→$1,078; filing fees $209→$299; agency implementation costs). https://www.gao.gov/assets/a277577.html — primary.
11. Administrative Office of the U.S. Courts / American Bankruptcy Institute, filings data (2,078,415 in 2005; 617,660 in 2006). https://www.uscourts.gov/data-news/judiciary-news/2022/08/09/just-facts-consumer-bankruptcy-trends-2005-2021 ; https://www.abi.org/newsroom/press-releases/2006-bankruptcies-fall-to-lowest-levels-since-1980s — primary.
12. Michael Simkovic, "The Effect of BAPCPA on Credit Card Industry Profits and Prices," 83 Am. Bankr. L.J. 1 (2009). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1157158 — primary (academic).
13. Credit-card industry ">$100 million" BAPCPA lobbying total — Missouri Law Review BAPCPA symposium (2006) and contemporaneous reporting, https://scholarship.law.missouri.edu/mlr/vol71/iss4/13 — secondary (approximate advocacy-side tally; flagged).
14. Federal Reserve Bank of New York, Quarterly Report on Household Debt and Credit, 2026 Q1 (May 2026; $1.66T balances; 10.3% of balances 90+ days delinquent; total household debt context). https://www.newyorkfed.org/newsevents/news/research/2026/20260512 — primary.
15. NY Fed Liberty Street Economics, "Federal Student Loan Defaults Return After Pandemic Pause" (May 2026; ~1M defaults 2025 Q4, 2.6M in 2026 Q1; 17% of borrowers 90+ days past due at least once since 2025 reporting restart). https://libertystreeteconomics.newyorkfed.org/2026/05/federal-student-loan-defaults-return-after-pandemic-pause/ — primary.
16. ED Federal Student Aid Data Center, portfolio reports (borrower count ~42–43M; average balance ~$38–39K, 2025). https://studentaid.gov/data-center/student/portfolio — primary (figures approximate; verify current quarter).
17. S.2598, FRESH START Through Bankruptcy Act, 117th Congress (Durbin–Cornyn, 2021; reintroduced later Congresses). https://www.congress.gov/bill/117th-congress/senate-bill/2598/text — primary.
18. Will Dobbie & Jae Song, "Debt Relief and Debtor Outcomes: Measuring the Effects of Consumer Bankruptcy Protection," 105 Am. Econ. Rev. 1272 (2015). https://www.aeaweb.org/articles?id=10.1257%2Faer.20130612 — primary (academic).
19. Constantine Yannelis, "Strategic Default on Student Loans" (working paper, 2016). https://fmaconferences.org/Napa/2017/Strategic_Default.pdf — primary (academic, unpublished wp; flagged).
20. Rajeev Darolia & Constantine Yannelis, "Strategic Default Among Private Student Loan Debtors: Evidence from Bankruptcy Reform," 15 Educ. Fin. & Pol'y 487 (2020); Philadelphia Fed WP 17-38. https://www.philadelphiafed.org/-/media/frbp/assets/working-papers/2017/wp17-38.pdf — primary (academic).
21. David Lucca, Taylor Nadauld & Karen Shen, "Credit Supply and the Rise in College Tuition," 32 Rev. Fin. Stud. 423 (2019). https://academic.oup.com/rfs/article-abstract/32/2/423/5042299 — primary (academic).
22. OpenSecrets, "Student loan companies spend millions on lobbying amid extended loan moratorium" (2021; Navient/Sallie Mae/Nelnet figures). https://www.opensecrets.org/news/2021/08/student-loan-companies-spend-millions-lobbying-amid-extended-moratorium/ — secondary (built on primary lobbying disclosures).
23. Government Executive, "Education Department withholds payments from student loan servicers" (Jan. 2024; five servicers, ~$16B/10-year contract; Aidvantage ~9M borrowers). https://www.govexec.com/management/2024/01/education-withholds-payments-student-loan-servicers/393153/ — secondary.
24. Nelnet Inc., Form 10-Q (Q3 2024; $277.7M ED servicing revenue, nine months ended Sept. 30, 2024). https://www.sec.gov/Archives/edgar/data/1258602/000125860224000058/nni-20240930.htm — primary.
25. Navient enforcement: CFPB v. Navient (2017 complaint); 39-state AG settlement, $1.7B private-loan cancellation (2022); 2023 bankruptcy-related settlement (~$182M) and 2024 CFPB order (~$120M) — unverified as to the last two figures; check primary CFPB/court documents before public use.
26. GAO, Student Loans: Education Has Increased Federal Cost Estimates of Direct Loans (2022; program re-estimate swings). https://www.gao.gov/ — secondary reference for re-estimate context (report number to confirm).