A model for predicting HEAL repayment patterns and its implications for medical student financing

R Cooter,J B Erdmann

Abstract:Background: The current level of Health Education Assistance Loan (HEAL) defaults has exceeded the original estimate, and as such is producing an unforeseen federal expenditure. Predicting repayment patterns of HEAL borrowers is an important step in assessing the impact that this unforeseen expenditure will have on HEAL and other financial aid programs. Method: Since prior research on educational lending has not identified a practical model for predicting repayment patterns, the authors turned in 1994 to an alternative discipline, research on consumer lending. Using the multivariate discriminant analysis credit scoring model framework, the authors incorporated operational definitions of the borrower's character, capacity, and capital. To identify factors that are significant in establishing repayment category membership, the framework was applied to a group of 233 HEAL borrowers who graduated from five medical schools in 1988. Because of the small numbers of delinquent borrowers and defaulters, the repayment categories were restricted to repayment, deferment, and forbearance. Results: The level of unsubsidized debt, in conjunction with financial resources (including the potential resource of parental support), may be significant in identifying those HEAL borrowers who may confront repayment difficulty. Conclusion: While the results are not surprising, they do lend documented support for a review of financial aid policies at both institutional and governmental levels, and a framework within which this review may be conducted.