State divestment and tuition at public institutions (2017)

This study examines the pass-through rate of changes in public funding to tuition and fees paid by students. Using an instrumental variable-fixed effects identification strategy, I estimate that a $1000 per student decrease in funding leads to the typical student paying $257 more each year in costs, with. However, both the pass-through rate and the proportion of tuition increases which can be explained by state divestment have increased over time. The pass-through rate increased from 10.3% prior to the year 2000 to 31.8% post-2000. I outline several avenues of future research which should be pursued in order to more fully understand which students shoulder the burden of reductions in public support.

Do expenditures other than instructional expenditures affect graduation and persistence rates in American higher education? (2010)

During the last two decades, median instructional spending per full-time equivalent (FTE) student at American 4-year colleges and universities has grown at a slower rate than median spending per FTE student in a number of other expenditure categories, including academic support, student services and research. Our paper uses institutional level panel data and a variety of econometric approaches, including unconditional quantile regression methods, to analyze whether these non-instructional expenditure categories influence graduation and first-year persistence rates of undergraduate students.

Our most important finding is that student service expenditures influence graduation and persistence rates and their marginal effects are higher for students at institutions with lower entrance test scores and higher Pell Grant expenditures per student. Put another way, their effects are largest at institutions that have lower current graduation and first-year persistence rates. Simulations suggest that reallocating some funding from instruction to student services may enhance persistence and graduation rates at those institutions whose rates are currently below the medians in the sample.