Abstract
This paper examines how personal, institutional, and legal factors affect where college presidents are placed and how much they earn given their placements. We find that controlling for selection into institutional type is important, suggesting that presidents nonrandomly sort into institutions based on unobserved characteristics that also relate to wages. We also find evidence that state “sunshine” laws governing whether applicants’ names must be disclosed in searches are related to placements and wages. Presidents hired in states that exempt the names of all but finalists from disclosure are more likely to be placed in public research universities and less likely to be placed in private institutions. There is also evidence that presidents hired in open records states earn compensating differentials, but we are ultimately unable to distinguish this from a state-specific effect. We also find wage discounts for presidents hired at times with larger numbers of states with open records and with exemptions to disclosure for non-finalists. Thus, presidents and institutions appear to respond to market-wide incentives created by sunshine laws.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
The sample was initially drawn from 572 Carnegie Tier I research institutions and all master’s-granting institutions; this initial sample was expanded to include all institutions in the same athletic conferences as the original institutions and the liberal arts institutions in the Consortium on Financing of Higher Education (COFHE).
We have a relative lack of compensation information for presidents at public schools in the 2001 and 2005 data. We thus control for both institutional categories and survey years in the wage regressions described below.
Although alternative measures of compensation are available in the Chronicle survey, we use total compensation from all sources as it provides the most comprehensive measure of presidential remuneration. It also includes 20% more observations than are available for base salary.
There are (4 choose 2) = 6 possible pairwise comparisons among the four categories, and for each of these there are 2 irrelevant alternatives that can be included or excluded, for a total of 12 possible unique comparisons. Note that there are only 12 unique values in Table 2 since each value recurs when the base categories are interchanged.
Indeed, when we include this variable in the final wage regression, its coefficient is insignificant (p > 0.4) with a small magnitude (β = 0.003).
Including predicted values of the institutional categories in the wage regression leads to the “generated regressor problem” in which the usual OLS standard error formulae are invalid. Thus, we report bootstrapped standard errors.
The effects of sunshine laws on wages that we are able to estimate are likely to be somewhat diffuse since in many cases we are observing the wage well after the law would have had an effect. We only see the current wage, not the wage when the person was hired; however, current wages are at least to some extent a reflection of how wages were initially set at the time of hiring.
The only notable changes from the model without sunshine controls are: (1) appointment year is insignificant for public non-research institutions but significant and negative for private non-research schools; and (2) having a prior position as a senior administrator is insignificant for private non-research institutions.
The estimates in the last row of the table show that when the sunshine controls are added, the number of in-state institutions still significantly affects placement into at least one of the institutional types. When included in the wage regressions that include the full set of controls, the number of institutions again has an insignificant coefficient.
The main changes are (1) the coefficient on tenure is generally less significant now until we control for selection in models 2 and 3, where it becomes negative and significant, suggesting wages for new hires rise faster than those of incumbents (consistent with Pfeffer and Ross, 1988; Langbert and Fox 2013); (2) the coefficient on nonwhite is now insignificant in the selection version of model 2 although it remains insignificant in the selection version of 3 as before; and (3) the coefficient on year 2005 is now significantly negative in all models, confirming that wages have increased across our successive survey years.
We also estimated specifications with the age of the oldest college in the state as an alternative instrument. Like the number of in-state institutions, this variable is significant in the placement model but insignificant in the wage model with the full set of controls. With this alternative instrument, the results for the sunshine variables in the final model remain similar, though exemptions for non-finalists have a significant positive effect and the number of states with open records just misses 10% significance (p = .116). But being an open records state still has a significant positive effect and the number of states with exemptions for non-finalists still has a significant negative effect.
Another option, rather than using instruments, is to also include the number of in-state institutions in the wage regression and identify the effect of placement from the nonlinearity of the multinomial distribution. When we take this approach, the effects of the sunshine variables in the final model remain the same, though some of them are less significant in the selection corrected version of model 2.
We also estimated the regression with state fixed effects, in which case the sunshine controls are all insignificant. Although there are good reasons to suspect why this might be the case, taken literally it means we are not able to distinguish the effect of open records laws from an unobserved, time-invariant state-specific factor.
We also considered a difference-in-difference approach, but this was infeasible because including a full set of state dummies, policy dummies, and interactions between these would require estimation of 361 parameters.
The main difference in the final model with selection controls is that age is now positive and significant.
When we use the age of the oldest college as an alternative instrument, the number of states with exemptions for non-finalists remains negative and significant, while the number with open records becomes insignificant. If we try to identify off of nonlinearity, all sunshine variables become insignificant although the numbers of states with open records and with exemptions for non-finalists remain close to 10% significance (p = .112 and p = .126, respectively).
References
Bai G (2014) University president compensation: evidence from the United States. Unpublished manuscript
Banker RD, Plehn-Dujowich JM, Xian C (2009) Organizational complexity and executive compensation: evidence on the compensation of university presidents. Unpublished manuscript
Cheng S (2014) Executive compensation in public higher education: does performance matter? Res High Educ 55(6):581–600
Ehrenberg RG, Cheslock JL, Epifantseva J (2001) Paying our presidents: what do trustees value? Rev High Educ 25(1):15–37
Estes N (2000) State university presidential searches: law and practice. J Coll Univ Law 26(3):485–509
Goldin C, Katz LF (2008) Transitions: career and family life cycles of the educational elite. Am Econ Rev Pap Proc 98(2):363–369
Grant DM (2004) Sunshine Laws in university presidential searches: do they really encourage openness? Unpublished manuscript
Katz RL (2009) Skills of an effective administrator. Harvard Business School, Cambridge, MA
Langbert M, Fox M (2013) The compensation and benefits of Private University presidents. J Acad Admin Higher Educ 9(2):45–56
McLaughlin JB (1989) The shady side of sunshine. Change 21(1):45–57
Mixon FG Jr, McKenzie RW (1999) Managerial tenure under private and government ownership: the case of higher education. Econ Educ Rev 18:51–58
Monks JW (2007) Public versus private university presidents pay levels and structure. Econ Educ Rev 26(3):338–348
Monks JW (2012) Job turnover among university presidents in the United States of America. J High Educ Policy Manag 34(2):139–152
Monks JW, McGoldrick K (2004) Gender earnings differentials among college administrators. Ind Relat 43(4):742–758
Muzzin LK, Tracz GS (1981) Characteristics and careers of Canadian University presidents. High Educ 10:335–351
Neumann R (2005) Doctoral differences: professional doctorates and PhDs compared. J High Educ Policy Manag 27(2):173–188
Lee CG, Ng PK (2015) Institutional determinants of university CEOs’ compensation – evidence from the U.K. Global Bus Econ Rev 17(2):151–163
O’Connell JF (2005) Administrative compensation in private nonprofits: the case of liberal arts colleges. Quart J Bus Econ 44(1/2):3–12
Parsons LM, Reitenga AL (2014) College and university president pay and future performance. Account Horiz 28(1):125–142
Peterson’s (2015) Colleges by state. Retrieved from the Peterson’s website: http://www.petersons.com/college-search/search-by-state.aspx
Pfeffer J, Ross J (1988) The compensation of college and university presidents. Res High Educ 29(1):79–91
Russell JD, Wellck AA (1940) The Administration of Higher Education. Rev Educ Res 10(4):347–350
Schmidt P (2001, October) Seeking the shade: public colleges and news media increasingly clash over sunshine-law issues. Chron High Educ 48(6):A21–A22
Shulman LS, Golde CM, Bueschel AC, Garabedian KJ (2006) Reclaiming Education’s doctorates: a critique and a proposal. Educ Res 35(3):25–32
Singell LD Jr, Tang HH (2013) Pomp and circumstance: university presidents and the role of human capital in determining who leads US research institutions. Econ Educ Rev 32:219–233
Tang TL, Tang DS, Tang CS (2000) Factors related to university presidents’ pay: an examination of private colleges and universities. High Educ 39:393–415
Walker DE (1979) The effective administrator: a practical approach to problem solving, decision making, and campus leadership. Jossey-Bass Inc., San Francisco, CA
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of Interest
The authors declare that they have no conflict of interest.
Rights and permissions
About this article
Cite this article
Singell, L.D., Stater, M. & Tang, HH. Let the Sunshine in: An Analysis of the Placement and Pay of University Presidents and the Effects of Open Records Statutes. J Labor Res 39, 405–432 (2018). https://doi.org/10.1007/s12122-018-9274-y
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12122-018-9274-y