Association for Public Policy Analysis & Management

Starting from the late 1990s, the U.S. federal government has adopted a new form of student financial aid: education tax benefits. Although education tax benefits represent a relatively new source of federal aid, there have been sharp increases in both federal spending on these benefits and the total number of recipients. Between 1997 and 2013, total federal spending on education tax benefits has increased by a factor of 13.5 from $1.5 to $20.3 billion. During the same period, the number of recipients grew from 2 to 15 million, almost twice the number of the recipients of Pell Grants, the primary federal aid program (The College Board, 2013).

The vast expansion of education tax benefits has attracted little scholarly attention. To my knowledge, only three studies have examined the enrollment effects of tax benefits and reached different conclusions (LaLumia, 2012; Long, 2004; Turner, 2011). Except for (LaLumia, 2012), no study has investigated the effect of tax benefits on college completion. Concerns about college completion, among both policy makers and academics, have grown over the last few years. The substantial growth in enrollment rate over the last four decades has not been accompanied by a comparable increase in completion rate. According to OECD (2013), the U.S has one of the highest dropout rates among all industrialized countries. More than 40 percent of U.S. students in postsecondary education stop short of a degree, compared to less than 25 percent in France, Japan, Denmark, and Finland.

This paper examines the impact of education tax benefits on college completion using a restricted-use dataset from the 2004-09 Beginning Postsecondary Students (BPS) survey. The paper employs Propensity Score Matching (PSM) procedure to correct for the endogeneity of education tax benefits. While PSM doesn’t entirely eliminate the bias resulted from the unobserved differences between students who receive tax benefits and those who don’t, it has two main advantages over other non-experimental approaches. First, PSM doesn’t impose functional form restrictions on the relationship between the treatment and the outcome. Second, and most importantly, PSM controls more effectively for the differences between tax benefits’ recipients and non-recipients because it requires that there is a sufficient overlap in the distribution of covariates between the two groups.

Empirical results suggest that education tax benefits increase college completion, within 6 years of initial enrollment, by 8 percentage points. These findings, however, vary by students and institutional characteristics. In particular, the evidence suggests that education tax benefits increase the completion gap between students attending private four-year institutions and other less-affluent students.

The paper contributes to the prior literature in two main ways. First, this paper is one of the few papers that examine the effects of education tax benefits on college outcomes. Second, the paper employs a very rich data set that allows for better measurement of eligibility to education tax benefits compared to prior research. In particular, the BPS data includes detailed information on student/parent Adjusted Gross Income (AGI), education expenses, sources of financial aid, enrollments status, degree level, and program of study.