Tax season ended last week. Taxpayers have filed for over thirty billion dollars in credits and deductions for college expenses they paid in 2017.
Evidence now clearly shows that these credits have zero effect on college attendance. The tax credits surely make those who get them better off, but they do nothing to increase education. If their intent is to increase schooling, they are a failure.
The tax subsidies for education are now extensive, complicated, and expensive. The tax system subsidizes the families of college students through tax-advantaged savings plans, credits, a deduction for tuition costs and loan interest, an exclusion of scholarships, grants and tuition reductions from taxable income, and a dependent exemption for students aged 19 to 23. Finally, the tax code subsidizes college with a deduction for interest paid on student loans.
Economists George Bulman and Caroline Hoxby scoured hundreds of millions of tax returns searching for an effect of the tax credits and tuition deduction on educational outcomes.  They inspected anonymized, detailed, individual-level administrative data from the IRS on the population of potential tax filers. The IRS has developed secure mechanisms that allow these data to be analyzed without compromising the privacy of taxpayers.
Bulman and Hoxby exploit the complicated rules that define eligibility for tax credits to estimate their effects. The tax credits “phase-out” within certain income ranges, with their value dropping with each additional dollar of income. The American Opportunity Tax Credit is worth up to $2,500 per student, but above an adjusted gross income of $160,000 (for a married couple) the credit drops steadily, until it reaches zero at $180,000.
Tax credits are supposed to increase attendance by offsetting some of the costs that students and their families incur by going to college. If tax credits work as intended, the association between level of family income and college attendance should weaken where the tax credits phase out. In the phase-out region, rising income is offset by the decreasing credit, so college attendance should be negatively affected to the extent that it is positively affected by the tax credit in the first place. Statistical analysts call this method a “regression-kink design”. “Kink” refers to the change in the slope of the curve representing the correlation between two variables (in this case, family income and college attendance) that one should see in the portion of curve where the hypothesized causal intervention (in this case, the phase out of the tax credit) is operative.
In many applications, small data sets make it challenging to precisely identify such a kink in the data. But with millions of observations at their disposal, the authors are able to conclusively reject any effect of the tax credits. In other words, there were no kinks in enrollment at the point that tax credits fade-out.
In another paper, Bulman and Hoxby use a regression-discontinuity design to estimate the effects of the tuition tax deduction for families around the maximum income cutoff for eligibility.  Again, they find no evidence that the deduction increases college enrollment. They also find no effect of the deduction on enrollment intensity, college choice, tuition paid, or student debt.
Why no effect? Tax credits and deductions primarily go to middle- and upper-income families, whose decision on whether to send their kids to college is unlikely to be affected by a tax benefit that is relatively small in relation to their income or the costs of college attendance.
Another possible explanation is that the credits are delivered too late to affect enrollment. Families get their tax refunds well after tuition is due; a family who pays tuition in September won’t get a tax credit until at least the following January. At that point the credit is a nice windfall but has arrived too late to help with paying the tuition bills.
The complexity of the tax benefits also likely undermines their effect. We discuss this in greater detail in a paper that provides a comprehensive overview of the tax benefits for education. 
An alternative way of looking at tax benefits for education is that they are a transfer program for middle-income families, putting more money in their pockets for all manner of expenditures, not just the costs of college. But they are ill-designed for that purpose, since they impose extensive administrative burdens on households, colleges and government. Reducing the tax rates applied to these families would be a more transparent and less expensive approach to achieving this goal.
If the billions spent on the tax benefits are to have any effect on college attendance, they should be delivered when tuition bills are due. One proposal, suggested by Hoxby and Bulman, is to compute eligibility for the credits automatically, using income tax information when a dependent approaches college age. Families could then be proactively notified of their eligibility. The authors also suggest that colleges file to receive the benefits directly from the IRS, so that a student need only present evidence of eligibility in order to have their account credited immediately.
An even more comprehensive approach would be to consolidate the tax credits with the Pell Grant, creating a single grant program that pays college costs at the time of enrollment. Eligibility could automatically be calculated using tax data, with funds delivered by the Department of Education. Families could apply by checking off a box on their tax forms. This approach would cut back substantially on paperwork, a relief for the millions of students who complete both the 1040 and the Free Application for Federal Student Aid in order to get federal grants, loans and tax credits for college.
Streamlining the tax benefits for education could potentially enhance their efficiency. At a minimum, a simpler system of education tax benefits would decrease the administrative and time costs of transferring funds to households with postsecondary expenses. At best, simplification would clarify incentives and increase investments in human capital.
— Susan Dynarski and Judith Scott-Clayton
Susan Dynarski is a professor of public policy, education and economics at the University of Michigan. Judith Scott-Clayton is Associate Professor of Economics and Education at Teachers College, Columbia University.
This post originally appeared as part of Evidence Speaks, a weekly series of reports and notes by a standing panel of researchers under the editorship of Russ Whitehurst.
The author(s) were not paid by any entity outside of Brookings to write this particular article and did not receive financial support from or serve in a leadership position with any entity whose political or financial interests could be affected by this article.
1. Hoxby, Caroline and George Bulman. 2015. “The Returns to the Federal Tax Credits for Higher Education,” Tax Policy and the Economy, Volume 29.
2. Caroline M. Hoxby and George B. Bulman, 2015. “The Effects of the Tax Deduction for Postsecondary Tuition: Implications for Structuring Tax-Based Aid,” Economics of Education Review.
3. Dynarski, Susan and Judith Scott-Clayton. “Tax Benefits for College Attendance,” in Alan Auerbach and Kent Smetters, eds., The Economics of Tax Policy. 2017. Oxford University Press.