The federal government's Direct Loan program dominates the student-loan market today, issuing 90 percent of all loans made across the country each year. Students pursuing everything from short-term certificates to master's degrees qualify for nearly $100 billion in loans every year at terms more generous than most private lenders would offer. Given the size and scope of the loan program, it is important to understand that the loan program imposes costs on taxpayers. Such costs speak directly to the need for policies that guard against fraud, waste, and abuse along with policies that provide information about loan performance. Borrowers who attend poor quality or overpriced programs will struggle to repay their debt and in turn impose losses to taxpayers. In the early 1990s, Congress enacted its first loan-based accountability policy: the cohort default rate. The cohort default rate measures the share of an institution's former students who borrow federal loans and default within three years of entering repayment. Institutions with high default rates lose eligibility for federal student aid programs because lawmakers saw high default rates as a proxy for low-quality institutions of higher education. The Obama administration's "gainful employment" regulations again sought to use loans as a proxy for value and quality, but in a different way. The initially proposed rule included a measure of whether borrowers who completed a particular program paid down principal on their student loans. The final rule does not include that measure but instead uses the amount of debt a student takes on (relative to his earnings) to gauge eligibility for federal aid by program. There are also proposals for a third loan-based accountability measure: risk sharing. These proposals -- advanced by think tanks, researchers, advocates, and some lawmakers -- would require institutions that pass the other measures of accountability to pay penalties to the federal government commensurate with the amount of loans that perform poorly. Despite the sound rationale for loan-based accountability policies, these measures still have limitations. The testimony presented here by Jason D. Delisle examines the loan program by looking at four categories of costs: loan defaults; Income-Based Repayment and loan forgiveness programs; loan discharges for fraud and closed schools; and lastly, comprehensive budget cost estimates for the entire loan program. These categories are not mutually exclusive, but they provide a useful framework for evaluating the major costs within the loan program. In discussing costs in these categories, Delisle disputes the view that the government profits when borrowers default on their loans and that it profits on the overall loan program. In concluding remarks, Delisle offers some general principles that he believes should guide any reform to accountability policies for federal student aid. Guiding principles for federal student aid accountability policies are suggested.