Comparative Policy Analysis

Designing a student-finance system from first principles

A cross-national look at how Australia, England, New Zealand, and Germany finance higher education — and a fully specified 2027 U.S. blueprint that puts colleges, not just students and taxpayers, on the hook for outcomes.

Institutional risk-sharing Income-contingent repayment Sept. 2026 data

$1.86T

Total U.S. student debt, 2026 Q2

10.6%

Balances 90+ days delinquent

A$0

Real interest under Australia's HECS-HELP

20–50%

Proposed institutional co-insurance share

How other countries do it

Australia, England, and New Zealand collect repayment automatically through the tax system as a share of income — not as a fixed installment owed to a servicer. Germany avoids the loan problem largely by keeping tuition free.

System	Tuition & subsidy	Repayment mechanism	Threshold & rate	Interest treatment	Write-off
Australia (HECS-HELP)	Publicly subsidized; capped student contribution	Automatic via ATO/employer withholding	No pay below A$69,528; marginal 1–10%	No real interest; CPI/WPI-indexed (2.8% in 2026)	None — follows borrower indefinitely
England	High tuition (~£9,535/yr), loan-financed	Automatic via HMRC/PAYE	No pay below ~£25k–£28k; 9% (6% postgrad) above	RPI to RPI+3%	25–40 years depending on plan
New Zealand	Moderate tuition, subsidized	Automatic via Inland Revenue	No pay below ~NZ$19–24k; 10–12 cents/$	Interest-free for residents	Hardship interest write-off available
Germany (BAföG)	Tuition free; nominal semester fee	Fixed quarterly installments, begin 5 yrs post-study	Fixed ~€390/quarter	0% on loan portion	Capped at €10,010 total; forgiven after 30 yrs low-income
United States (2026 reform)	High, rising; loans finance tuition + living	Servicer-based; Tiered Standard or RAP	RAP: as low as $10/mo; income-tiered	6.5–9.1% statutory, capitalizes	RAP: 30 years; taxable pre-2026 issue

What the evidence actually shows

Sorted by confidence level — strong causal evidence, mixed findings, and genuinely unresolved questions.

Strong evidence

Expanded loan availability passes through into tuition (Bennett hypothesis), concentrated in for-profit and vocational programs
Automatic, tax-collected income-contingent repayment lowers default and admin cost vs. servicer-based systems
Racial gaps in repayment are large and persistent — Black borrowers retain ~95% of balance after 20 years vs. 5% for white borrowers

Mixed evidence

Whether broad cancellation reduces racial wealth gaps — regressive by income, progressive by net worth
Causal size of debt's effect on homeownership is much smaller than popular narratives suggest
Entrepreneurship and fertility effects are largely correlational, confounded by family background

Open questions

No consensus empirical threshold for "unaffordable" debt-to-income ratio
Long-run effects of 2026 U.S. reforms (RAP, earnings-test rule) not yet observable
How self-employment/gig income should be captured in payroll-based collection

Repayment calculator: four systems compared

Illustrative model — not a substitute for an official loan estimate. Adjust income and balance to compare annual repayment burden across systems.

Annual income: $55,000

Loan balance: $30,000

Model simplifications: flat marginal bands per published 2026 thresholds; U.S. "current" reflects Tiered Standard 10-yr amortization at 7% interest; U.S. RAP reflects the 2026 income-tiered schedule.

The proposed 2027 U.S. model

Grant-first tuition financing, automatic payroll-collected repayment, and institutional risk-sharing — the three pillars that international evidence supports most strongly.

Grant structure

Universal minimum grant covers ~50% of public tuition, plus income-scaled Pell-style top-up

Max annual borrowing

$6,000 tuition loan + $4,000 living-cost loan per year

Lifetime cap

$24,000 tuition-loan lifetime cap (undergraduate)

Net tuition exposure

Target $0–$4,000/year at public institutions after grants

Graduate/professional

Uncapped but risk-shared track; no grant subsidy extends to graduate tuition

Private loans

Must offer income-contingent option; bankruptcy parity required

Interest

0% real interest; balance indexed to lower of CPI or wage growth

Income threshold

No repayment below 150% of federal poverty line

Repayment rate

Marginal bands, 3%–8% of income above threshold

Collection

Automatic IRS payroll withholding — no third-party servicers

Max repayment period

20 years, then automatic forgiveness, no application required

Unemployment/disability

Automatic pause via UI/SSDI data match, zero accrual

Co-insurance share

Institutions liable for 20–50% of principal unpaid after 10 years, scaled to program-level non-repayment rate

Accountability floor

2026 Earnings Accountability rule retained as eligibility gate beneath the co-insurance layer

Program-level scope

Risk-shared at the program level, not individual student level, to avoid admissions discrimination

Professional programs

Higher co-insurance share for programs with weak completion-to-earnings ratios

Bankruptcy

Full parity with unsecured consumer debt after defined post-obligation period

Public service

Qualifying nonprofit/government employment counts double toward 20-year clock

Existing borrowers

Voluntary conversion to new schedule; legacy borrower-defense and PSLF relief continue in parallel

Credit reporting

No adverse credit reporting tied to payroll-collected obligation

Strongest objections, considered fairly

Objection — Admissions discrimination

Institutional risk-sharing could cause colleges to deny admission to students perceived as higher-risk of non-repayment — precisely the students grant aid is meant to reach.

Mitigation: apply co-insurance at the program level rather than per-student, and risk-adjust the formula for entering cohort characteristics, similar to Pell-share adjustments in existing cohort default rate models.

Objection — Fiscal cost of zero real interest

Removing interest revenue increases the static federal cost of the loan program; CBO scoring shows swings of tens of billions of dollars depending on interest and forgiveness parameters alone.

Response: weigh against reduced default-collection costs and modest but positive completion/earnings effects documented in NBER loan-limit research — genuine tradeoffs, not free.

Objection — Income-manipulation loophole

A 20-year automatic forgiveness horizon could reward high earners who structure income to appear low, particularly self-employed and gig workers.

Payroll withholding closes this for salaried workers; New Zealand's repeated amendments to its income definition show this remains a genuine, unresolved edge case.

Objection — Bankruptcy parity raises private lending costs

Removing bankruptcy exceptionalism could tighten private-loan underwriting or raise borrowing costs.

Private loans are a small share (~7–9%) of total outstanding debt, and federal loans dominate the market this reform targets — an acceptable tradeoff given that scope.