Privatization and cost inefficiency at U.S. Public research universities

ABSTRACT

This study examines the relationship between cost efficiency and privatization at 163 public research institutions in the United States between 2005 and 2015. We employ a spatial autoregressive (SAR) random-effects model and stochastic frontier analysis (SFA) to estimate the relationship between costs and four privatization variables: auxiliary enterprises as a percentage of total revenue, tuition and fees as a percentage of total revenue, private grants/contracts as a percentage of total revenue, and out-of-state first year enrollment. Results showed cost inefficiency at public research universities increased between 2005 and 2015, even as reliance on private sources of revenue increased. Public research universities exhibit 28.5% overall cost inefficiency over the time period studied, 85.6% of which is short-run cost inefficiency. This suggests that most of the cost inefficiency varies across years and may be the result of challenges that institutional leaders face adapting to short-term fluctuations in market-oriented sources of revenue. The results also show a nonlinear relationship between cost inefficiency and three of the privatization variables. Given the expectation of little to no increase in state support for public research universities, this study has implications for policy, institutional management, and future research.

KEYWORDS:

• Educational finance
• public research universities
• privatization
• cost efficiency
• stochastic frontier analysis
• spatial analysis

JEL Classification:

• I22
• I23
• I29
• H52

Disclosure statement

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

Notes

1 The authors looked for reliable and complete data on part-time faculty salaries. We investigated The Chronicle of Higher Education’s crowd-sourced data on salaries for instructors and adjuncts, but few institutions were present in the database for multiple years and our study spans 15 years.

2 See Jaquette and Parra (2014) regarding why the OPEID is preferred to the IPEDS UNITID when establishing the institution unit of analysis.

3 More specifically, the Stata user-written program sfmodel by Kumbhakar and Lovell (2000) is used to regress estimates of short-term (residual) and long-term (persistent) cost inefficiency on measures on privatization.