Financing and Enrollments in Public Universities

ABSTRACT

State budget cuts have raised concerns about the disproportionate enrolment of out-of-state students. These students pay higher tuition fees at public universities. We investigate the decision-making of public universities about in-state and out-of-state enrolments by considering tuition rates as a financial incentive. We find that out-of-state enrolment is elastic, while in-state enrolment is relatively less elastic. Universities participating in reciprocity programs show that local student share is relatively insensitive to tuition changes; however, non-local student enrolments are more sensitive to these tuition changes. Overall, there is no evidence of the crowding-out effect on local student enrolment.

KEYWORDS:

• Higher education
• enrolment decisions
• tuition rates
• reciprocity programs
• public universities

JEL CLASSIFICATION:

• H52
• I22
• R23

Acknowledgement

The authors thank Mark D. Partridge, Ian Sheldon, Abdoul Sam, Wuyang Hu, John V. Winters, and Amada Ross for their helpful comments. Any remaining errors are our own.

Data availability statement

The data used in this article and the program file necessary to replicate the results are available online at https://doi.org/10.7910/DVN/4Z93LC.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ We use the terms ‘university’, ‘institution’, and ‘college’ interchangeably.

² We use the terms ‘reciprocity programs’ and ‘exchange programs’ interchangeably.

³ For more information about the particular universities participating in the various cross-state tuition arrangements, see: https://www.nasfaa.org/State_Regional_Tuition_Exchanges.

⁴ Using enrolment shares instead of pure headcounts can reduce bias caused by university capacity changes. For example, a university could have 100 in-state and 50 out-of-state students enrolled in the previous year. So, the previous-year total enrolment was 150 students. In the current year, the in-state and out-of-state enrolments are 105 and 70 students, which leads to 175 students in total. From headcounts, both enrolments have increased. However, the in-state student share has decreased from 66% to 60% of total enrolment. Our study used a share of enrolment controls for total seats that a university can offer to enrollees. Machin and Murphy (2017) also explain that a change in university capacity can cause endogenous sorting. For instance, a university can introduce a new department, which may generate an enrolment expansion.

⁵ Most exogenous institution controls rarely change over time (Deming and Walters 2017). Consequently, adding a time-fixed effect can control for invariant control variables. We also add the 25^th and 75^th percentiles of the ACT composite score to control for the different admission standards for in-state and out-of-state enrolment (Groen and White 2004). Admission scores can also mitigate the sorting bias in our estimations, since the students who perform less well than other student groups from secondary school are disadvantaged in competing for access to the most prestigious institutions (Astin and Oseguera 2004). The results are not reported here, since there is a great deal of missing data for admissions scores, which largely reduces our observations. Overall, however, our results are unchanged.

⁶ For example, Deming and Walters (2017) find that four-year, less-selective universities are more reliant on state appropriations because of their financial constraints from other revenue sources. Dependence on state appropriations is positively correlated with dependence on tuition revenue, which can impact tuition-setting.

⁷ Distance d (in km) between the centroid of state s to state s’ is used as a weight. A centre-point for each state is obtained from the 2000 Census of Centers of Population, which provides the population-weighted centroids.

⁸ There may be a concern that our instruments could suffer from spatial correlations, especially when we consider states participating in regional reciprocity programs. It is important to note that we include institution-fixed effects in all regressions, which they could alleviate this concern. Additionally, we also conduct a sensitivity analysis by adding region-fixed effects, such as Midwestern, Southern, and Western. The results of this sensitivity analysis reported in the Appendix (Table A3) resemble those of our baseline results.

⁹ We also constructed instruments that capture only the average weighted tuition growth rates in neighbouring states that are located within the region. For example, considering The Ohio State University, the average weighted tuition growth rates were included for Illinois, Indiana, Michigan, and Wisconsin. The results are not reported here, but overall, different instruments did not affect our baseline results.

¹⁰ Moreira, Sharifvaghefi, and Ridder (2021) employ model symmetries in the instrumental variable regression to obtain the conditional integrated test for the causality parameter in the over-identified model. This test is inappropriate for our study because our regressions are a just-identified model.

¹¹ Tuition rates do not significantly affect enrolment decisions for universities that are not participating in reciprocity programs. We believe that this is because universities not participating in the programs have different tuition prices for in-state and out-of-state students. Consequently, tuition prices may not be an important factor for universities deciding on their enrolments. By contrast, for universities that are participating in the programs, exchange students are not counted as residents but pay the same tuition rate as residents. Tuition prices, therefore, become a factor in deciding on the enrolment of these two groups. The reported results are in the Appendix (Table A2).

¹² The total expenditure includes instruction, research, public service, academic support, student service, institutional support, and transfers. We follow Leslie et al. (2012) to exclude the expenditure related to hospital services since some universities do not have a hospital.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.