In order to estimate the places in which student debt represents a greater burden, we rely on the IRS’s Statistics of Income data, which includes the number and percentage tax returns that claimed the student loan interest deduction, earned income tax credit, and other federal tax programs at the ZIP code level. We then aggregate this data at the congressional district and state levels, assigning split ZIP codes based on the shares of their population in each district using the Department of Housing and Urban Development’s crosswalk. To account for the fact that the same amount of student debt will be less of a burden for high-income debt holders, we constructed a debt burden index that adjusts each state and district’s overall rate of claiming the student loan interest deduction by their median household income, racial composition, and rate of reporting income from the earned income tax credit. The result is an estimate of not only how broadly student debt affects the population of a given geographic region, but how much of a burden that debt is likely to represent for those who hold it.
To be clear, this method is inherently limited in certain respects. While the Statistics of Income data is the most comprehensive data on student debt at the most granular geographic level available, it is a measure of tax returns, not individuals. Some people file their taxes jointly, and not everyone who is responsible for paying off a student loan claims the deduction on their taxes. However, there is a strong correlation between the share of student loan interest deduction filings and the share of survey respondents who reporting being responsible for student debt payments at various geographic levels, so this does not seem to have systematically biased our results. With respect to the construction of our burden index, adjusting the rate of student loan deductions by other economic indicators imposes a strong ecological assumption that doesn’t hold in all cases -- there are certainly low-income student debt holders who live in relatively wealthy areas. However, this adjustment should, on average, move us closer to the accurate reflection of where student debt burden is most acutely felt that would be available in an ideal world where we had similar data at the individual level.