Policy Lab

Run on 6 March 2026

Student Loans Reform in the US

A comprehensive restructuring of federal student lending: caps graduate borrowing at program earnings levels, holds institutions financially accountable for graduate outcomes, eliminates interest capitalization, reforms servicer contracts to outcome-based standards, restores bankruptcy discharge with a five-year waiting period, and clears the Borrower Defense backlog.

Full document →

The final result

Read the article →

This debate took roughly 75 minutes to run. We created AI agents for each major political constituency, surfaced their grievances, designed reform proposals, scored each other's reactions, and eventually produced a full legislative bill proposal.

If you just want the final result, the article explains what the bill would change and how it affects the different groups involved. Otherwise, keep reading below to follow the full deliberation.

Read the Article

or keep reading to discover the process behind it

Constituencies

Full map →

We began by identifying the main political constituencies involved in this debate.

Progressive DemocratsModerate DemocratsModerate RepublicansConservative Republicans

Grievances

Full reports →

We then mapped what each group wants, in their own voice.

Progressive Democrats

1Interest accrual turns manageable debt into an unpayable spiral
2Income-driven repayment plans fail the people they were designed to help
3Public Service Loan Forgiveness is broken by design and by practice

Moderate Democrats

1Income-driven repayment plans have become a bureaucratic trap
2Graduate and professional school debt has grown to levels that distort career choices
3The interest accrual system compounds debt even when borrowers are doing everything right

Moderate Republicans

1Taxpayers Are Being Asked to Pay for Degrees That Didn't Deliver
2The Income-Driven Repayment System Has Become a Hidden Subsidy with No Accountability
3Universities Face No Consequences for Graduates Who Cannot Repay

Conservative Republicans

1Debt Cancellation Punishes People Who Played by the Rules
2Federal Loan Guarantees Drive Tuition Inflation
3Income-Driven Repayment Plans Create Open-Ended Taxpayer Liability

Policy Areas

Full analysis →

Finally, we grouped these grievances into 8 negotiable policy areas.

01Institutional Risk-Sharing and Outcome Accountability
02Federal Credit Design and Graduate Borrowing Limits
03Loan Administration, Servicer Accountability, and IDR Integrity
04Loan Interest Mechanics and Amortization Reform
05Predatory Credentialing, Accreditation Oversight, and Borrower Fraud Relief
06Student Loan Bankruptcy Discharge
07Unassigned Grievances
08What This Debate Does Not Address

Alignment

Full analysis →

We then analyzed where the constituencies converge, where they clash, and what trade-offs might hold a coalition together.

For each of the 6 policy areas, we ran a reform loop: one agent proposed changes, another scored how each group would react, and the cycle repeated until the scores cleared the bar or no more gains were possible.

Since 1976, federal student loans have been treated as nearly non-dischargeable in bankruptcy. To eliminate a student loan balance through bankruptcy, a borrower must file a separate adversary proceeding and meet the "undue hardship" standard, which courts have historically interpreted to require near-total and permanent inability to repay. Most people in bankruptcy cannot afford the litigation this requires, and attorneys often decline these cases because the outcome is too uncertain to be worth the cost. No comparable restriction applies to other forms of unsecured consumer debt, such as credit cards or personal loans.

These were the proposed changes:

·The bankruptcy exit that doesn't exist for student debt
·Income thresholds that limit who qualifies
·The five-year clock and how it works
·The government must respond in 60 days or lose
·A fiscal reserve and a paired accountability bill
→For the full details, see the full debate.

By group

Progressive Democrats: The proposal reaches borrowers that the current legal system treats as assets on a ledger: people who have been in financial distress for years but cannot legally exit debt the way they could exit credit card debt. The core change, eliminating the adversary proceeding and replacing it with a motion, makes discharge practically accessible rather than theoretically available.

Moderate Democrats: The implementation plan in this proposal covers both staffing and the case management technology platform needed to receive and track discharge motions from all 94 federal bankruptcy courts. Both are published in the Federal Register, funded in the same appropriations cycle, and must be confirmed operational before the first discharge motion can be filed. That matters because a staffing floor without a functioning intake system just recreates the backlog through a different bottleneck.

Moderate Republicans: The companion bill requirement is the one concrete concession this proposal offers: a companion institutional risk-sharing bill, holding universities accountable for graduate default rates, must be introduced and referred to a congressional committee within 180 days of enactment. That is a legislative fact, not a signal. A bill number and a committee referral exist on the public record.

Conservative Republicans: Conservative Republicans hold that debt freely chosen must be repaid. This is a moral premise, and no mechanism in this proposal addresses it. The five-year waiting period, the income thresholds, the tiered structure, the companion bill: all of these are design features that modify the scope of the discharge mechanism. None of them change whether discharge should exist.

The federal government extends student loans at the graduate level through the Grad PLUS program, which has no aggregate borrowing cap: a student can borrow the full cost of attendance at any accredited institution regardless of the program's earnings profile or the borrower's credit history. Undergraduate Direct Loans carry statutory annual and aggregate limits, but no equivalent constraint applies at the graduate level. The interest rate on federal graduate loans is set annually based on Treasury yields plus a statutory add-on, and has in recent years been above 8 percent for Grad PLUS borrowers [Federal Student Aid, 2024]. Schools have faced no regulatory requirement to publish median program-level debt or earnings outcomes as a condition of certifying loan amounts, though some disclosure requirements have been proposed and litigated across administrations.

These were the proposed changes:

·Grad PLUS borrowing caps tied to earnings
·Public-interest carve-out for teachers and nonprofit workers
·Mandatory earnings disclosure before loan certification
·Automatic cap reductions at persistently poor-outcome programs
·Data infrastructure as a gating condition
→For the full details, see the full debate.

By group

Progressive Democrats: Graduate students preparing for careers as Title I teachers and direct-service nonprofit workers gain access to a higher borrowing cap tier under the carve-out, which previously excluded them. For the students who fill these roles, it means the difference between being able to complete a graduate degree and being priced out by an earnings-indexed cap.

Moderate Democrats: The proposal delivers on a structural fix that many in this constituency have pushed for: tying federal credit to what graduates actually earn rather than to whatever schools charge. Future borrowers in fields where debt has dwarfed earnings for years will face a different cost environment.

Moderate Republicans: The outcome-triggered reduction mechanism is the component this constituency finds most compelling. Programs where graduates are buried in debt relative to what they earn now face a direct federal consequence, not just a reputational one. The appeal-stay cap prevents schools from running out the clock indefinitely.

Conservative Republicans: The distributional analysis requirement, which must be published before the carve-out takes effect and must be disaggregated by race, income at entry, and institution type, directly addresses this constituency's concern about who the carve-out actually benefits. They can see the data before the subsidy tier is activated.

Under current federal law, institutions that participate in Title IV federal student loan programs collect tuition revenue from students who borrow federally, but bear no direct financial consequence when those graduates cannot repay. The 1992 Higher Education Act reauthorization created cohort default rate thresholds that can trigger loss of Title IV eligibility, but these thresholds are set high enough that few schools have lost access on those grounds, and the sanctions apply only after sustained, severe default rates rather than as a proportional cost that rises with institutional risk. Schools set their own tuition, access federal credit regardless of program-level earnings outcomes, and retain the revenue whether graduates thrive or default.

These were the proposed changes:

·Institutions pay when their programs fail borrowers
·Graduate borrowing limits tied to what programs actually pay
·Program earnings data shown before every loan is certified
·A federal accountability threshold independent of accreditation
·Assessment calculations adjusted to avoid penalizing schools that serve low-income students
·The earnings data pipeline protected against legal and administrative disruption
→For the full details, see the full debate.

By group

Progressive Democrats: The equity-adjusted benchmark directly protects programs at historically Black colleges, community colleges, and other schools where high Pell enrollment correlates with lower early earnings. Without the adjustment, those schools would be most exposed to assessments despite serving the students the system is supposed to help. The adjustment is in statute, not subject to administrative revision.

Moderate Democrats: The combination of the risk-sharing fee (supply-side accountability) and the Grad PLUS cap (demand-side constraint) together address the tuition inflation loop they have named as a structural grievance. The fee changes institutional behavior around program pricing; the cap limits how much students can be certified to borrow at programs where the fee already signals weak outcomes.

Moderate Republicans: This proposal delivers the Grad PLUS borrowing cap they requested in each round of deliberation. The cap is tied to five-year post-graduation earnings for the specific program, applied at loan certification, and recalibrated annually. It is additive with the risk-sharing fee, not a substitute. A program subject to the fee also faces the cap.

Conservative Republicans: The Grad PLUS cap reduces federal credit exposure at programs where graduates cannot realistically repay. CBO would score this as a net fiscal benefit because it reduces lending at programs where a significant share of originations would otherwise end in income-driven repayment forgiveness or default. Fiscal hawks can note this without endorsing the broader mechanism.

Federal student loans accrue interest during deferment, forbearance, and any period when a borrower's income-driven repayment (IDR) payment falls below the monthly interest charge. At the end of those periods, the unpaid interest capitalizes, meaning it is added to the principal balance. A borrower who makes every required payment under an IDR plan for years can finish that period with a larger balance than they carried at the start of repayment. Interest rates on federal loans are set each spring by Congress based on Treasury yields plus a statutory add-on, with a statutory cap that has rarely been a binding constraint. This design is not shared by most consumer lending products, where scheduled payments are structured to retire principal over time.

These were the proposed changes:

·Interest capitalization eliminated
·24-month retroactive correction
·Interest subsidy for lower-income borrowers on income-driven plans
·Self-certification pathway for borrowers who do not file tax returns
·Payment floor on future income-driven plans
·Forgiveness cap at the original amount borrowed
→For the full details, see the full debate.

By group

Progressive Democrats: The subsidy threshold increase from the first-round level of 225 percent of poverty to 250 percent of poverty was the specific change this group pushed for. A public school teacher earning $42,000 was above the first-round line and received nothing while watching interest accumulate. Round 2 covers them.

Moderate Democrats: The non-filer self-certification provision was the specific gap this group identified after the first round. The proposal answers with operational specificity: a DOE form within 90 days, servicer outreach within six months, and an annual re-certification process tied to the existing income-driven repayment recertification. That is a workable pipeline, not a promise.

Moderate Republicans: The proposal includes a binding CBO score requirement before the subsidy can take effect. If the companion institutional risk-sharing policy area does not produce enough revenue to render the subsidy budget-neutral, the Department of Education must bring a secondary offset proposal to the relevant congressional committees within 90 days. This converts a fiscal assumption into a fiscal constraint with a defined process if it falls short.

Conservative Republicans: The one structural change this group gained is that collapsing the "interest-mechanics relief" and "loan forgiveness" reporting categories now requires a floor vote in both chambers. In the first round, a future appropriations rider or OMB scoring decision could have done it administratively. Round 2 makes that collapse a legislative act with floor accountability.

Federal student loan servicing is contracted to private companies paid a flat fee per account, giving them a financial incentive to process accounts at minimum cost rather than help borrowers navigate complex repayment options. The Consumer Financial Protection Bureau, the Government Accountability Office, and the Department of Education's Inspector General have each documented servicer failures including steering borrowers toward forbearance instead of IDR, miscounting qualifying payments, and losing payment history across servicing transfers. The Public Service Loan Forgiveness program, created in 2007, requires 120 qualifying payments under qualifying plans with qualifying employers, and has a documented history of rejections based on misapplied plan types and miscounted payments, often discovered only after a decade of repayment.

These were the proposed changes:

·Servicers get paid for outcomes, not volume
·A mandatory audit corrects years of miscounted payments
·IDR enrollment becomes automatic
·All IDR plans consolidate into one
·PSLF terms become law, not guidance
·Borrowers get a real appeals process
→For the full details, see the full debate.

By group

Progressive Democrats: Public sector workers, teachers, nurses, and social workers who submitted employer certification forms and never heard back get automatic PSLF certification on day 181. A future administration cannot run out the clock on them anymore.

Moderate Democrats: Mid-career professionals with graduate debt who hold FFELP loans can now consolidate to access the new single IDR plan without losing the IDR payment history they accumulated before consolidation. Previously, consolidation meant restarting the forgiveness clock; that trap is closed.

Moderate Republicans: The private companies that caused the documented errors, the servicers and their technology subcontractors, pay the penalties. The Servicer Error Remediation Fund draws from those penalty assessments, not appropriations. Servicers cannot blame their data vendors and walk away; the proposal explicitly extends liability to subcontractors.

Conservative Republicans: The PSLF deemed-certification is explicitly limited to borrowers who filed complete forms under the original 2007 eligibility criteria. It does not use Biden-era expanded qualifying employment definitions or waiver-based eligibility.

The federal government conditions access to Title IV student aid on accreditation by agencies recognized by the Department of Education. In practice, these accrediting agencies are substantially self-governed by the institutions they oversee, which has allowed a subset of schools, concentrated in the for-profit sector, to maintain accreditation and access to federal loan dollars despite documented patterns of deceptive recruiting, inflated job placement claims, and poor student outcomes. When such schools have collapsed or been exposed, the statutory remedy for defrauded borrowers, Borrower Defense to Repayment, has been chronically underfunded and subject to reversal across administrations, leaving large numbers of approved claimants in limbo while their balances continued to accrue interest. A gainful employment rule, which would have conditioned aid eligibility on program-level debt-to-earnings thresholds, was issued, litigated, and reversed without ever reaching stable enforcement.

These were the proposed changes:

·Program-level outcome thresholds
·No more debt-metric gaming
·Locked-in data infrastructure
·Clearing the fraud claim backlog
·Full discharge, not partial credit
·A path to relief for closed-school cases
·Automatic discharge when fraud is proven institutionally
·Fraud investigations on a deadline
·Pre-funded protection for students when schools collapse
·Independent accreditors, not industry clubs
→For the full details, see the full debate.

By group

Progressive Democrats: Former students of collapsed for-profit schools who filed fraud claims years ago and got nothing now have a statutory timeline: 24 months to resolution, with an independent panel as backstop if the Department falls behind. That panel's decisions are binding, meaning the Department cannot ignore them.

Moderate Democrats: The 24-month BDR resolution mandate now has teeth for the cases that were most likely to be denied under the old approach: older applications from closed schools with missing records. The alternative evidentiary standard was not in round 1. It means borrowers who attended schools in the 1990s and 2000s that have since closed are not simply out of luck.

Moderate Republicans: The debt metric now covers what students actually owe, not just the federal portion. A school cannot approach the warning threshold and then substitute institutional loans and deferred-fee arrangements to pass the calculation while students carry the same total burden. The statute closes that substitution pathway explicitly, and underreporting non-federal debt costs a school its entire Title IV certification authority.

Conservative Republicans: The biggest structural problem with the round-1 design was that the entire accountability mechanism rested on an administrative memo between two federal agencies, which any administration could let lapse without passing a law. That pathway is closed. The DOE-SSA data-sharing authority is now in statute, with ranked fallback sources. Disabling the threshold system now requires an affirmative act of Congress, not a quiet non-renewal.

Executive Summary

Full debate →

Across all 6 policy areas, average constituency approval rose from 22% to 58% and satisfaction from 20% to 53%.

The full debate page has the per-policy-area executive summaries and the complete round-by-round record.

Bill

Full bill →

An independent legislator agent implemented the outcomes of the debate in a 7-title, 52-section bill proposal, with some highlights:

·Interest capitalization eliminated: Right now, unpaid interest is added to principal when a borrower exits deferment or forbearance, and interest then accrues on the inflated balance. This bill bans that practice for all Direct Loans and reverses capitalization events from the prior 24 months. For borrowers on income-driven plans earning below 250 percent of the federal poverty level, the federal government covers the monthly interest gap so the balance does not grow while payments are being made.
·Institutions pay when graduates cannot: Universities currently have no financial stake in whether graduates repay their loans. This bill imposes a program-level risk-sharing fee of 5 to 15 percent of loans certified when a program's graduates carry debt exceeding 12 percent of annual earnings for two consecutive years, and caps Grad PLUS borrowing at 150 percent of the program's own five-year median post-graduation earnings. Institutions that set tuition knowing their graduates will not repay now face a direct cost.
·Servicers held to outcomes, not activity: Servicer contracts currently pay a flat fee per account, creating an incentive to minimize borrower contact. This bill converts contracts to outcome-based performance metrics covering IDR enrollment rates, payment-count accuracy, and dispute resolution. A mandatory 24-month audit reconstructs qualifying payment counts for all active IDR borrowers, with the burden of proof on the Department rather than the borrower.
·and more in the full bill

Constituency Reviews

Full reviews →

Each constituency reviewed the final bill and assessed how it compares to the status quo.

Progressive Democrats: view the interest capitalization ban and the income-graduated subsidy as real wins, particularly the threshold of 250 percent of poverty, which covers a public school teacher earning $42,000 who would have been above the original proposal's line. The payment-count audit with the burden of proof inverted is the provision they describe as overdue by a decade. Codifying the Public Service Loan Forgiveness program into statute removes the administrative reversibility that cost many of their constituents years of credit toward forgiveness. The bankruptcy discharge, even with the five-year wait, opens an exit that did not exist.

Moderate Democrats: emerged from this process closest to getting what they asked for. Their 1.2 grievances mapped almost exactly to the provisions included: the interest capitalization fix, the servicer accountability conversion, the Grad PLUS caps, the Borrower Defense backlog, the bankruptcy discharge, and the reliance-harm remedy for borrowers who reorganized their lives around programs that were later reversed. The coalition they positioned themselves to anchor, structural reform without broad cancellation, held together.

Moderate Republicans: secured their two most consistent asks: a Grad PLUS borrowing cap tied to program-level earnings (requested in every round of the reform process) and institutional risk-sharing fees that impose a direct financial cost on schools whose graduates cannot repay. The Congressional Budget Office offset requirement before the interest subsidy takes effect, the congressional authorization requirement for future forgiveness programs exceeding $10 billion, and the means-testing phase-out above $125,000 individual income address their fiscal accountability and distributional fairness objections concretely.

Conservative Republicans: got a version of the fiscal discipline provisions they pushed for: Grad PLUS caps that reduce federal credit exposure at low-value programs, risk-sharing fees that make institutions pay when graduates default, a forgiveness cap at original principal that limits how much income-driven repayment forgiveness can exceed what was actually borrowed, and a congressional authorization requirement for large new forgiveness programs that addresses their separation-of-powers objection. These are not minor concessions from the rest of the coalition.

Process Audit

Full audit →

Finally, an independent auditor agent reviewed the full deliberation process, flagged structural risks, and assessed whether the outcomes hold up to scrutiny.

Article

Read the article →

A plain-language explanation of the final bill, written for a general audience with no knowledge of the policy area or deliberation process.

Student Loans Reform in the US

Phase 1: Preparation

Phase 2: Iteration

Bankruptcy Discharge

Federal Credit Design

Interest Mechanics

Loan Administration

Predatory Credentialing

Phase 3: Final

Phase 4: Publication

Student Loans Reform in the US

Phase 1: Preparation

Phase 2: Iteration

Bankruptcy Discharge

Federal Credit Design

Institutional Risk Sharing

Interest Mechanics

Loan Administration

Predatory Credentialing

Phase 3: Final

Phase 4: Publication