All students deserve to attend programs and institutions that meet high quality standards. But different program goals, investments of time, and eligibility criteria make it difficult to set specific standards that will apply equally well to all types of programs and institutions. Though no criteria will perfectly measure institutional quality, adhering to basic policy design principles could lead to an effective system that reduces the number of students enrolling in programs and institutions that are unlikely to help them achieve their goals.

This report sets out principles for developing an accountability system for postsecondary institutions to improve outcomes for students, protecting both them and the investments taxpayers make in their education. We propose a system based on multiple metrics that requires institutions to pass three out of four thresholds, which diminishes the risk of institutions manipulating their outcomes and requires satisfactory performance in more than one area, while allowing flexibility for differing programs, missions, and circumstances. These proposed metrics would be based on the following:

• Share of students defaulting. Weighting the share of borrowers defaulting by the share of students who borrow gives an institution where most students have debt a lower score than an institution with the same share of borrowers defaulting but a lower share with debt, providing a better indication of the share of students with poor outcomes.
• Share of total debt repaid. Using a metric related to the share of balances repaid to complement the student-based default rate metric captures debts disproportionate to incomes even when borrowers are delinquent but not in default or are in good standing because of income-driven repayment, despite not retiring their debts.
• Completion rate. Completion is a critical outcome of higher education that applies to all students, not only those who rely on debt financing. Unlike default, repayment, and earnings measures, completion rates are available quickly. Thresholds should differ for four-year, two-year, and shorter programs.
• Postcollege earnings. The difficulty of adequately measuring rates of return makes earnings a reasonable metric. Thresholds should differ by type of credential, might be adjusted for geographic differences, and should account for cyclical variation.

The outcomes we document in this report indicate that even if we implement effective accountability standards that steer students away from the institutions with the weakest outcomes, much work will remain to ensure students achieve their educational goals. In addition to eliminating programs and institutions that fail the standards, setting specific standards for outcomes should provide incentives for improved performance. Highlighting outcomes on the proposed metrics should also inspire policymakers and educators to redouble efforts to provide institutions and students the resources and strategies required to support student success.