Income-Contingent Student Loan Repayment in Australia
Introduction
Income-contingent student loan repayment insures borrowers against low post-school income by linking required repayments to realized income, but it can also create moral hazard because borrowers reduce repayments when they earn less. de Silva (2025) studies Australia’s Higher Education Loan Programme (HELP), the first national income-contingent student loan program, and focuses on the 2005 repayment schedule that raised the repayment threshold for existing and new borrowers.
The empirical analysis uses linked Australian administrative tax and student-loan records from the Australian Taxation Office Longitudinal Information Files, supplemented with Census, MADIP, and HILDA data. The identifying variation comes from the 2004-2005 HELP schedule change: the threshold above which borrowers begin repayment rose from approximately AU$26,000 to AU$35,000, and the repayment schedule creates a notch at which a borrower just above the new threshold makes a repayment of approximately AU$1,400. Borrowers bunch below the repayment threshold after the reform, with larger responses in occupations with more hourly flexibility.
The MVPF is computed from an estimated life-cycle model of consumption, labor supply, student debt repayment, income risk, borrowing constraints, and adjustment frictions. The baseline MVPF estimate corresponds to moving borrowers from a 25-year fixed-repayment contract to the 2005 HELP income-contingent loan, holding education choices, borrowing choices, wages, interest rates, and the existing tax-transfer system fixed.
MVPF = 3.7
Net Cost
The net cost is the reduction in the government budget from moving borrowers from the benchmark 25-year fixed-repayment contract to the 2005 HELP income-contingent loan. In Table IV of de Silva (2025), this fiscal cost equals AU$1,635 per borrower in 2005 dollars. It is the sum of two components: repayments fall by AU$185, and taxes net of transfers fall by AU$1,450. The second component is quantitatively important because income-contingent repayment changes labor-supply incentives, so the government loses revenue not only through the repayment system but also through the broader tax-transfer system.
The calculation is performed in the estimated structural model rather than directly in the reduced-form bunching exercise. Borrowers face income risk, borrowing constraints, and fixed labor-supply adjustment costs; they choose consumption and labor supply over the life cycle, and the government collects income taxes, debt repayments, and pays transfers. The model is estimated to match the bunching response around the HELP repayment threshold and standard moments of the income process, then used to simulate the fiscal effect of alternative repayment contracts.
Willingness to Pay
Willingness to pay is measured using a structural life cycle model as the model-implied equivalent variation for borrowers who are “behind the veil of ignorance” with respect to their initial states. In Table IV of de Silva (2025), moving from fixed repayment to the 2005 HELP income-contingent loan has a WTP of AU$6,004 per borrower in 2005 dollars, or a consumption-equivalent gain of 1.68% of lifetime consumption. This value reflects the insurance provided by reducing repayments when borrowers have low income, which is valuable because borrowers face uninsurable income risk and borrowing constraints early in the life cycle.
The WTP calculation nets out the utility cost of the labor-supply responses induced by income-contingent repayment. The same estimated model shows that these responses are economically meaningful: borrowers reduce labor supply to lower repayments, and the estimated Frisch labor-supply elasticity is approximately 0.15. Nevertheless, the insurance value of income-contingent repayment is large enough that borrower WTP substantially exceeds the fiscal cost in the baseline HELP 2005 comparison.
MVPF Summary
The baseline MVPF equals WTP divided by the fiscal cost of the policy. Using the Table IV estimates for the 2005 HELP income-contingent loan, the MVPF is 6,004 / 1,635 = 3.67. This means that each dollar of net fiscal cost generates approximately AU$3.67 of borrower willingness to pay in the model.
The estimate should be interpreted as the value of restructuring repayment terms for existing borrowers, holding education and borrowing choices fixed. This restriction is deliberate: for existing student debt, ex ante choices are fixed by definition, so the comparison isolates the insurance value and moral-hazard cost of the repayment contract itself. The paper also solves for a constrained-optimal income-contingent loan that raises the same revenue as fixed repayment; that contract has no net fiscal cost and increases borrower welfare by the equivalent of 0.79% of lifetime consumption, which reinforces the conclusion that income-contingent repayment can improve welfare even when labor-supply responses are taken seriously.
The paper contains numerous other MVPF estimates for alternative policy adjustments.
References
de Silva, Tim (2025). “Insurance versus Moral Hazard in Income-Contingent Student Loan Repayment.” Quarterly Journal of Economics 140(4): 2851-2905. https://doi.org/10.1093/qje/qjaf036.
de Silva, Tim (2025). “Replication Data for: Insurance versus Moral Hazard in Income-Contingent Student Loan Repayment.” Harvard Dataverse. https://doi.org/10.7910/DVN/D2G7CC.
Hendren, Nathaniel, and Ben Sprung-Keyser. 2020. “A Unified Welfare Analysis of Government Policies.” Quarterly Journal of Economics 135(3): 1209-1318. https://doi.org/10.1093/qje/qjaa006.
Policy Features
- Category
- Education
- Sub-Category
- Subsidized Loans
- Beneficiary Type(s)
- Adults, College enrollees, Students
- Average Age
- 29.5
- Average Income
- 28586
- Country of Implementation
- Australia
- Year of Implementation
- 2005
- Empirical Method
- Event Study
- Research Type
- Primary
- Peer Reviewed
- Yes
- MVPF Publication Link
- doi.org/10.1093/qje/qjaf036