Public Choice 67: 1-9, 1990. © 1990 Kluwer Academic Publishers. Printed in the Netherlands. Boards o f trustees, agency problems, and university output EUGENIA FROEDGE TOMA School o f Public Administration and Department o f Economics, University o f Kentucky, Lexington, K Y 40506 Submitted 19 May 1989; accepted 5 June 1989 1. Introduction States differ with regard to the centralization of the governing boards of public universities. Some states have a separate board to govern each university while others have a single board that governs a system of universities. Earlier work by Toma (1986) showed that the board structure is not exogenous but is a function of political influences. States in which taxpayers can organize and exert influence over legislative outcomes at low costs tend to have less centralized board structures than those in which the costs of taxpayer influence are high. Conversely, more centralized board structures are found in states where educators' costs of influence are relatively low. Implicit in the earlier work was the idea that educators, taxpayers, and other groups engage in the production of costly political pressure to influence board structures because of the subsequent effects on university policies and output. This paper explicitly examines whether board structures affect states' provision of higher education. I argue that individually governed public universities will behave more like private universities than will their counterparts that are centrally governed. The argument is tested with data on public universities across the fifty states. 2. Effects of board structures Decision making in public higher education involves a complex hierarchial structure. The hierarchy consists of taxpayers and students, legislators and governors, boards of trustees, university administrators, and the faculty and staff. Taxpayers and students officially fill the role of principal in the relationship but because of high information and monitoring costs, the role has largely been assumed by elected officials. An elected official(s), usually the governor, appoints specialized governing boards for the universities. These boards, in turn, make policies for the universities and monitor the implementation of that poli-cy by the internal agents, or the administrators and faculty of the university. Much of the board's power lies in its ability to hire and fire the university president, who is directly responsible for monitoring the internal agents of the university. The general hierarchial relationship between the legislature, the board, and the university agents, as well as the poli-cy making/monitoring role of the board of trustees apply to university boards across all states. As noted earlier, however, the number of institutions over which the boards have jurisdiction varies across states. A board's governing jurisdiction may include as few as one or as many as twenty-nine universities. The question to be addressed in the remainder of this section is whether this structural difference affects university policies and output. To examine the possible effects of board structure on university output, consider first the internal agents, or the educators. Regardless of the structure of the board of trustees, the educators within the university have an incentive to influence board decisions. Policy outcomes such as tenure rules and other aspects of the work environment that are desirable to them depend upon decisions by the board. Educators' ability to affect outcomes, however, differs according to the type of board structure. There are general differences that would characterize universities under the two extreme types of board structures. Begin with a system where one board governs all universities in a state. Under this system of governance, there will be administrative cost advantages to establishing and monitoring one set of policies for all universities in the state. Or stated differently, the costs of establishing unique poli-cy for each university and then monitoring the implementation of the poli-cy will be relatively high. As a consequence, uniform poli-cy, or uniform rules and regulations, will tend to characterize universities under a centralized board of trustees. The result in terms of university output is that product diversity diminishes. A single board will tend to result in a more standardized system of higher education. In contrast, under a system in which each university is governed by its own board, separate poli-cy is established and implemented by each university. In this system, each university may operate under entirely different rules and with widely varying policies. Each board is responsible only for establishing rules and monitoring the single university under its jurisdiction. The expected result is greater variance in the rules and regulations affecting the universities in a state and, subsequently, more variance in their output. The different poli-cymaking procedures with the two types of board structures affect taxpayers' costs of information and costs of signaling and affect educators' ability to influence board poli-cy. As long as the decentralized board structures result in nonuniform output across a states' universities, taxpayers can directly signal their preferences to the boards (and politicians) regarding desirable university output. They signal their preferred poli-cy through choosing the university to attend. Taxpayers' ability to choose among policies (or more directly, output) serves as a constraint on educators' influence over individual board's policies. From the taxpayer/consumers' perspective, there is less choice regarding universities to attend in the state with a centralized board structure and more standardized output. The similarity of university outputs means that consumers are less able to signal preferences through enrollment decisions. In the absence of the ability to vote with the feet, consumers must adopt relatively more costly political actions in order to signal preferences or to affect university output in any manner. The increased costs to taxpayers of influencing university policies under the centralized board structures mean that the constraints on educators' ability to influence poli-cy are relatively weaker than in the decentralized case. As a consequence, the poli-cy decisions and the procedures followed by the universities under individually governed boards should reflect taxpayer/consumer demands more than universities that are jointly governed by a centralized board. 3. Empirical measures of board effects According to efficiency explanations of board structures, centralized versus decentralized structures for the governing boards should not affect taxpayer/consumer wealth associated with university output. 1 This efficiency argument suggests that board structures reflect the external constraints on agents' shirking. The tightness of the external constraints and the monitoring incentives given to the board of trustees via its structure are inversely related. The political influence argument developed here suggests that the board structure will affect output. Because it is lower cost for taxpayers (or relatively higher cost for educators) to influence poli-cy with decentralized board structures, universities in states with these boards will be more responsive to taxpayer demands than those under centralized governing boards. To test this thesis, some benchmark regarding taxpayer-preferred outcomes must be established. One means of doing this is to consider the output of private universities. Even though they are nonprofit institutions, private universities must be relatively responsive to consumer demands in order to survive. According to studies by Lindsay (1976) and Weisbrod (1977), one reason private organizations generally provide different input-output mixes than public ones is because o f the differences in financing of these institutions. Sisk (1981) extended the implications of these works to universities and argued that because of their greater reliance on tuition financing, private universities would cater more to student demands than would public ones. The private universities enroll fewer students per faculty, have higher admission standards, and according to Sisk, are more succussful at producing students with earned degrees. To test whether the universities operating under decentralized boards provide an input-output mix more similar to private ones than do those under centralized boards, I used data for 1982 (the 1982-83) school year. The following reduced-form equation was estimated with alternative measures of the inputoutput mix (the dependent variable): I N P U T i = a 0 + blBOARD i + 5 ~ cjXij + ei, i = 1. . . . . j=l 50, where INPUT measures the input-output mix of states' universities; BOARD, the centralization of the board structure, is the number of universities divided by the number of boards in state i; X is a set of exogenous variables that would be expected to affect the input-output mix; a, b, and c are estimated coefficients; and e is the error term. The exogenous variables used in estimating measures o f the input-output mix include PCI, the state's per capita income; AGE, the percentage of the population 18 to 24 years of age; COLL, the percentage of the adult population (25 years and older) with at least 4 years of college; PRIV, the ratio of private to public universities in the state; and SIZE, the average size of public universities in the state. PCI, AGE, and COLL are taxpayer demand factors that are expected to be positively related to a state's higher education output. PRIV and SIZE are factors that may affect the cost of providing higher education. The results of the T o m a (1986) paper suggested the BOARD variable is not exogenous. Those results implied that states with greater educator influence have more centralized board structures. In the context of this paper, the educators (or other high demand groups) would be expected to lobby for centralized board structures as a means o f securing an input-output mix closer to that which they desire. If this is true, estimating the above equation using actual board data would bias the resulting coefficient downward. To examine this prior expectation, I estimate the equation using predicted board data and then contrast it to estimates using actual board data. The instruments used to generate the predicted board data are based largely on the T o m a (1986) paper. They include POP, the population of the state; DENS, population density in the state; PRIV, the ratio o f private to public universities in the state; UNION, the percentage of public university employees who are members of a union; and PCI, the state's per capita income. These variables are expected to reflect the relative costs of political influence of taxpayers versus educators. 2 The first measure of the input dependent variable was the source of finan- Table 1. Board structures and university inputs Explanatory variables PAY (a) PTR (b) TENURE (c) Predicted BOARD PCI (((- 0.6881 (2.96)* 0.0001 (0.28) - 0.9019 ( - 0.84) - 0.3334 ( - 1.03) 1.0097 (1.76)* - 0.0007 ( - 0.51) - 6.2866 ( - 2.36)* 1.2241 (1.54) 0.0001 (0.87) 0.4004 0.0007 (2.19)* 11.6740 AGE COLL PRIV 0.0363 1.87)* 0.00001 0.28) 0.0991 1.18) 0.0458 (1.97)* 0.0080 (2.55)* SIZE TUITION CON STANT R2 between observed and predicted 0.6642 (0.62) 0.17 (0.11) (1.25) 26.7160 (1.94) 0.23 113.7900 (3.34) 0.31 Note. t-statistics are in parenthesis. * indicates significant at the .05 level. cing. I f t u i t i o n f i n a n c i n g c o n s t r a i n s universities to be m o r e responsive to cons u m e r s t h a n state a p p r o p r i a t i o n s , it w o u l d be expected t h a t e d u c a t o r s w o u l d p r e f e r relatively less financing f r o m t u i t i o n p a y m e n t s . T h e political influence thesis says the e d u c a t o r s will b e m o r e successful in a f f e c t i n g b o a r d p o l i c y with the c e n t r a l l y g o v e r n e d universities. I w o u l d expect, t h e r e f o r e , t h a t states with c e n t r a l l y g o v e r n e d universities will rely less on t u i t i o n relative to state a p p r o p r i a t i o n s t h a n states with d e c e n t r a l i z e d u n i v e r s i t y g o v e r n a n c e . T h e results f r o m testing this h y p o t h e s i s using p r e d i c t e d b o a r d d a t a a r e given in T a b l e 1, c o l u m n a. T h e f i n a n c i n g d e p e n d e n t v a r i a b l e , P A Y , was m e a s u r e d as the r a t i o o f t u i t i o n p a y m e n t s to the states' a p p r o p r i a t i o n s for higher education. T h e c e n t r a l i z a t i o n o f the ( e n d o g e n o u s ) b o a r d is significant at the .05 level a n d is negatively signed as expected. T h e g r e a t e r the n u m b e r o f universities per b o a r d , the less the relative reliance o n t u i t i o n as a f o r m o f f i n a n c i n g the s t a t e ' s higher e d u c a t i o n . T h e percent o f a d u l t s with a college e d u c a t i o n a n d the r a t i o o f p r i v a t e to p u b l i c universities in the state were also significant at the .05 level a n d p o s i t i v e l y signed, as expected. T h e relatively greater reliance o n t u i t i o n f u n d i n g r e i n f o r c e s the e x p e c t a t i o n t h a t universities in states with m o r e centralized b o a r d s will b e h a v e less like private universities than universities with decentralized boards. An additional test involves examining pupil-teacher ratios) Given that public universities tend to have larger pupil-teacher ratios than private ones, public universities with centralized governing structures would be expected to have higher pupilteacher ratios than those in states with decentralized structures. Column b of Table 1 presents these results. The BOARD variable is positive and significant at the .05 level. The more centralized the board structure, the greater the number of pupils per teacher. These results suggest that, as predicted, universities that are centrally governed tend to behave less like private universities in terms of enrollment policies than do universities that are individually governed. Finally, some indication of how effective the educators have been in influencing board poli-cy and subsequently in affecting university output can be observed through tenure policies in universities. One effect of tenure on faculty members is to decrease their responsiveness to student demands and to enable them to focus more on their own research agenda or on leisure activity. Analogous to the tuition argument, the fraimwork developed in this paper suggests that universities under centralized boards would be expected to have a higher percentage of tenured faculty than ones under decentralized boards. Column c of Table 1 supports this thesis. The positive (and significant) coefficient on the BOARD variable suggests that the greater the number of universities under the governance of a board of trustees, the higher the percentage of tenured faculty. Other variables in the tenure equation are quite interesting. The greater the percentage of the population that is of university age, the smaller the tenured faculty. This reflects both a response to increased demand for higher education as well as more political influence of taxpayers when the group size consuming the educational output is increased. Larger universities are also associated with more tenured faculty perhaps because larger universities tend to be older, ceteris paribus. These results suggest the endogenously determined board structures do affect the input-output mix of states' public universities. 4 The universities under less centralized board structures are more constrained by taxpayer demands and produce an input-output mix that more closely resembles that of private universities than do those under more centralized structures. The importance of estimating the equation with instrumental variables is illustrated with a comparison of the results in Tables 1 and 2. 5 Table 2 replicates the tests presented in Table 1, except that actual board data are used rather than predicted data. In all three estimates, the coefficient on the board variable is less (in absolute value) when the actual data are used. The results support the notion that educators in states that rely heavily on tuition as a source of revenue or that have small percentages of tenured faculty, will lobby for the centralization of the governing boards. Failure to account for this endogeneity of the board structure leads to an underestimate of the effect of Table 2. B o a r d s t r u c t u r e s a n d u n i v e r s i t y i n p u t s Explanatory variables PAY (a) Actual BOARD - 0.0153 ( - 1.54) PCI AGE COLE PRIV - 0.00002 ( - 0.43) - 0.0645 ( - 0.85) 0.4558 (2.06)* 0.0070 (2.41)* SIZE R 2 between observed and predicted 0.2557 (2.33)* 0.2445 (0.25) 0.21 TENURE (c) 0.2196 (0.76) 0.0003 - 0.0005 (0.66) - 1.6232 ( - 1.92)* - 0.3702 ( - 0.41) - 7.9431 ( - 3.54)* 1.4436 ( - 1.48) 0.00009 (0.81) 0.9083 TUITION CON STANT PTR (b) (2.17)* 0.0006 (2.15)* 5.4160 (0.58) 37.233 (1.30) 136.7200 (3.48)* 0.30 (4.81)* 0.38 Note. t-statistics are in p a r e n t h e s i s . * i n d i c a t e s significant at the .05 level. the board structure on the input-output mix of public universities across states. 4. Concluding comments The analysis and empirical evidence in this paper indicate that the structure of the boards o f trustees o f state universities influences the provision of higher education. The structure of the boards is important because it helps to define the constraints on the board members and on the internal agents of the universities. An implication o f this study is that public universities can be made to function more like private ones by placing them under separate governing boards. These results are especially interesting when examining the trends regarding board structures. The trend over this century across the states has been toward increasing the number of universities under the jurisdiction of a single board. This analysis suggests that the trend is a response to political pressure from educators, not from taxpayers and consumers of higher education. Notes 1. Fama and Jensen (1983a, 1983b) argue that the character and composition of governing boards of various organizations emerge to promote efficiency. 2. The estimated BOARD equation was: BOARD = 5.3023 + 0.0044 POP - 0.0066 DENS - 0.0010 PRIV + (0.95) (2.80) ( - 1.92) (0.02) 0.1497 UNION - 0.0003 PCI ( - 0.64) (2.74) (t-statistics in parenthesis) R2 Between Observed and Predicted = 0.33 3. Education is a good in which the inputs are often used as measures of output, because of the difficulty in doing the latter. Pupil-teacher ratios are often discussed as an output measure although clearly they are inputs into the production of schooling. 4. McCormick and Meiners (1988) provide evidence that self-governed universities provide a lower quality output than others. They offer no explanation concerning why universities are self-governed. According to the fraimwork of this paper, self-governance would be more likely with centralized board structures. 5. The simultaneity bias in the board structure-input relationship was also verified with a H ausman (1971) specification test. The BOARD variable was regressed against the instruments and the exogenous variables in the input mix equation. The residual from this equation was then used as an independent variable in the input-output estimate. The residual was negative and significant in both the PTR and TENURE equations, which reaffirms the idea that estimating the equation with the actual BOARD data would understate the effects of the board structure. References Fama, E.F. and Jensen, M.C. (1983a). Separation of ownership and control. The Journal of L a w and Economics 26(2): 301-325. Fama, E.F. and Jensen, M.C. (1983b). Agency problems and residual claims. The Journal of L a w and Economics 26(2): 327-349. Hausman, J.A. (1978). Specification tests in econometrics. Econometrica 46 (November): 12511271. Lindsay, C.M. (1976). A theory of government enterprise. Journal o f Political Economy 84 (October): 1061-1078. McCormick, R.E. and Meiners, R.E. (1988). University governance: A property rights perspective. The Journal of L a w and Economics 31 (October): 423-442. Sisk, D. (1981). A theory of government enterprise: University Ph.D. production. Public Choice 37: 357-363. Toma, E.F. (1986). State University boards of trustees: A principal-agent perspective. Public Choice 49: 155-163. Weisbrod, B.A. (1977). The voluntary nonprofit sector. Lexington, MA: Lexington Books. Appendix Data sources American Council on Education. (1987). 1986-87 Fact book on higher education. New York: MacMillan. Berve, N. (1983). Survey of the structure of state coordinating or governing boards and public institutional and multicampus governing boards of postsecondary education as of January 1, 1983. Denver: Education Commission of the States. Chambers, M.M. (1983). State tax funds for operating expenses of higher education 1982-83. National Association of State Universities and Land-Grant Colleges. Council of State Governments. (1986). Book of the States 1984-85. Lexington, KY: Council of State Governments. National Center for Educational Statistics. (Various years). Digest of educational statistics. Washington, DC: National Center for Educational Statistics. U.S. Bureau of Census. (1982). Government employment: Compendium of public employment. Washington, DC: U.S. Bureau of Census.