For several years the senior author of this study has analyzed the relationship between state government spending in higher education and the rate of economic growth. Recently, working with colleagues Jonathan Leirer at the Center for College Affordability and Productivity and Tony Caporale at Ohio University, he has greatly expanded his investigation, using ever more elaborate models and econometric techniques. The results, however, remain the same: The statistical correlation between state and local governmental expenditures on higher education and the rate of economic growth (growth in real income per capita) is typically negative — higher spending for universities is associated with lower growth in a state, other things being equal. We certainly reject the hypotheses arising from conventional wisdom, namely that greater university appropriations will likely mean higher state growth rates.
For the interested reader, a sampling of the statistical results can be found in Table 3. We used a data set encompassing observations for all 50 states for each year over the 46 years from 1960 through 2005. Most of the statistical models run have far more than 1,000 observations. We incorporate lags to acknowledge the fact that money spent today may take years to have a pay-off — students for example, take four, five or even six years to get through school. Research monies similarly may have long term pay-offs. We also looked at economic growth over a short time horizon (five years), as well as longer periods (10 or 15 years). We incorporated a large number of non-higher education variables into our model to at least partially control for the considerable non-educational determinants of income growth over time.
The results are intriguing. In general, the model’s explanatory power is greatest for longer time lags. Consequently, we will talk mainly about equation three in the table, which looks at economic growth over a 15-year time horizon and relates it to state appropriations made 15 years previously, to allow generously for the lagged impact of appropriations.
There is much that can be said about the results. Most importantly, however, in all three equations (and dozens of other equations not presented here) we obtain a negative relationship between state and local higher education expenditures and economic growth. In two of the equations, the results are statistically significant at the 1 percent level.
Economists cannot say with precision what factors are most important for future economic growth in a state, but with these results in hand we can say this: The hypothesis that higher education spending promotes economic growth is rejected.
Some of the control variables introduced are interesting in their own right. The model shows a very strong negative correlation between the aggregate state and local tax burden and economic growth in a state. Since taxes are something controlled by state policymakers, the model suggests a better growth-enhancing factor than expanding higher education appropriations would be to lower the tax burden.
Also, we included non-higher education spending (K-12) as a variable in the model, and in the statistically most impressive equation three, we obtain a negative relationship between spending and growth — even after controlling for the tax burden. All of this suggests that government activities that crowd out private-sector activity (through higher taxes, for example) tend to lower growth, since on average the private sector utilizes resources more efficiently than the public sector.
There is one variable in the model that seems to suggest that universities have a positive growth effect. Economic growth is faster in states where the proportion of the population over the age of 25 possesses college degrees. Doesn’t this favor more state university funding — since more state subsidies will mean lower tuition charges, increasing access to colleges? There are several reasons why this argument is weak.
First, other empirical evidence we have gathered shows that a large majority of new state appropriations go to increase total university expenditures — not to lowering the rate of tuition increases. Second, there is only the very weakest of statistical correlations between state appropriations and the proportion of the population who are college graduates. It is worth remembering in this context that roughly half the students entering four-year degree programs fail to graduate from college within six years. More appropriations may merely lead to small increases in enrollments among marginally qualified students who then fail to graduate.
Most important, however, is the fact that college almost certainly acts as a "screening device." The persons who graduate from college derive higher earnings than non-college graduates and are clearly more productive. But is this because of what they learned in college? Or, is it because, on average, college graduates are brighter, more motivated, and more disciplined than non-college graduates? Even if they had not gone on to college, the college graduates would probably have fared far better than the typical non-college worker because of these other attributes.
The results presented here are merely representative of many models estimated. We have used alternative functional forms (ordinary least squares, generalized least squares), different independent (control) variables, different lags, allowed for individual state characteristics (so-called fixed effects models) and other permutations. The preponderance of evidence leads us to reject the hypothesis that university spending tends to increase the rate of economic growth. While the alternative hypothesis — higher university spending lowers growth — seems likely to be valid, it is not necessary to accept it in order to reject the validity of notions that greater fiscal efforts by the state will enhance growth.
Elsewhere, the senior author has shown that incremental resources of universities have often gone to fund things far removed from the core functions, including bloated administrative bureaucracies, elaborate student services, high salaries and reduced teaching loads for faculty. This is particularly true of the large research universities like the University of Michigan.