Comments

Robert P. Inman: Thomas Kane, Peter Orszag, and Emil Apostolov address the declining trends in state appropriations for public higher education as a share of income and in relation to spending for students at private universities. In levels, real state spending per public university student has been just about constant over the past twenty-five years, but both average income and real spending per student at private universities have increased. While state government support for higher education is cyclical (see discussion below), the real concern for the authors is stagnant state support for state universities. The authors identify the culprit as the growth in states' Medicaid spending, acting to crowd out states' potential fiscal support for higher education. Kane, Orszag, and Apostolov see these trends as "grounds for significant concern that the relative quality of public institutions has declined in recent years." It is useful to ask what arguments and evidence might lead one to share their concerns.

The authors identify two key determinants of state higher education spending over the past twenty-five years, the business cycle and state Medicaid spending. They find higher education spending is cyclically sensitive, declining one year after an increase in the state unemployment by about $4 a person (a 2.6 percent decline) for each 1 percent increase in state unemployment (see table 3). These cutbacks in higher education spending during the downturn are allocated to help defray the higher welfare and Medicaid payments caused by the recession. As the economy recovers and welfare and Medicaid obligations decline, however, spending for higher education is increased by amounts sufficient to hold constant the level of higher education spending over the business cycle. During downturns, spending falls by $400 a student from the mean of $6,100 a student, but then rises by $400 a student above the mean during the upswings (see figure 2). Such within-budget reallocations over the business cycle allow the state to hold the aggregate level of state government spending as a share of state income constant over the business cycle (see table 3). A stable income share for aggregate state spending allows states to hold tax rates fixed over the business cycle. A stable tax rate over the business cycle, known as tax smoothing, is exactly what is recommended for efficient public finance. Precluded from borrowing on their own to cover increased poverty obligations during a downturn by state balanced budget rules, states turn to their middle-class residents with children at the state universities for their loans. During the downturn, states borrow from these families (on average about $400 a student), by reducing state support for a university education. The students' families can then borrow privately, perhaps even using favorable federal loans, to cover reduced state funding. The states finally repay the loans as the economy improves by increasing state support for higher education by $400 a student more than average. There is nothing particularly troubling about such fiscal behaviors. Balanced budget rules exist for good reasons, and tax smoothing is the efficient fiscal policy.

This pattern of spending during the business cycle offers interesting evidence in favor of tax smoothing, but the results are not at all worrisome as a matter of public policy. Instead, Kane, Orszag, and Apostolov are right to look for problems, if they exist, in the secular trends in higher education spending. Figure 1 shows the trend in state general support for higher education as a share of state income. That share is falling. What might explain the decline? A good place to begin the search is to examine the relevant economic trends over their sample period, four of which seem particularly relevant: 1) fewer college-age students; 2) rising incomes; 3) rising costs; and 4) increased competition from private universities.

College-age students (ages 18–24) as a share of national population have fallen from 13.25 percent in 1980 to 10 percent in 2003.¹ With fewer students, spending per student can be maintained at the historically stable mean of $6,100 a student using a smaller...