September 21, 2007
Determinants of Governance Structure Among Companies: A Test of Agency Theory Predictions
By Frankforter, Steven A Davis, James H; Vollrath, David A; Hill, Vanessa
This study tests the model of Davis, Schoorman, and Donaldson (1997) that proposed determinants of a company's governance structure. In particular, we focus on the agency theory aspects of the model and its ability to predict the presence of an opportunist- orientation CEO at publicly listed U.S. companies. Using survey based data obtained from CEOs and directors of 100 companies in a match-pair design, we identified three variables that predicted the occurrence of agency-oriented behaviors by the company's CEO. These results lend support for the model's ability to predict the conditions under which opportunists become CEOs. Agency theory is an oft-used tool researchers apply in exploring the effects and conditions of governance relations between corporate CEOs and directors. However, the successful application of agency theory has proven problematic over time, with spotty performance regarding its predictions in areas such as incentives, monitoring, and corporate performance. Eisenhardt (1988) criticized agency theory as having few testable hypotheses and yielding mixed findings. Applying a meta- analysis methodology, Dalton, Daily, Certo and Roengpitya (2003) concluded that agency theory does not predict relationships between ownership structure and performance, nor does it reveal insights concerning organizational complexities. Such criticisms lend credence to questions about agency theory's ability to predict leadership and governance structure in corporations.Since its inception, agency theory research has almost exclusively focused on the use of secondary data. While the use of primary data might reveal new insights, with few exceptions (Tosi, Brownlee, Silva, & Katz, 2003), little research has investigated the factors that lead to the choice of either a stewardship- or an agency-oriented corporate governance structure. This paper applies the Davis, Schoorman, and Donaldson (1997) model that, in part, predicts when an agency theory-oriented corporate governance structure will occur. Applying a matched-pair methodology with surveys of CEOs and directors at the same U.S. companies, we find agency theory's core assumption regarding the assumptions of agent opportunism hold when certain CEO psychological and company situational factors are present.
Agency theory views the firm as a nexus of contracts (Jensen and Meckling, 1976) with the role of the CEO as a catalyst to initiate and manage relationships with similarly opportunistic stakeholders (Hill & Jones, 1992). The result is unstable, short-term marriages of convenience that exist as long as transacting parties reap short- term profits. Uzzi (1997) referred to such relationships as being a part of an atomistic market, typified by arm's-length transactions motivated by self-interest, with impersonal relationships, while Yan, Zhu, and Hall (2002) characterized agency-oriented principal- agent relationships as project based, and reflecting mutual mistrust that requires the formation of explicit agreements and expectations into formal contracts. However, in spite of its popularity and seemingly common sense approach to curbing the problems of opportunism or shirking, there has been a paucity of research providing clear support for agency theory (Dalton et al, 2003; Eisenhardt, 1988).
Hypotheses: Testing the Davis, Schoorman, and Donaldson Model
We test the agency theory aspects of the model Davis et al (1997) proposed to distinguish between corporate governance structures based on agency and stewardship theories. The model proposes determinants of governance structures. For a more comprehensive discussion of the theory underlying the hypotheses that follow, refer to Davis et al (1997).
Agency theory emphasizes the role of extrinsic rewards in motivating an opportunistic CEO, employing a tit-for-tat perspective in which rewards are contingent on effort and tangible results (Jensen & Meckling, 1976). Additionally, managers pursue lower order needs, fulfilled primarily through material reward and the acquisition of wealth, rather than through the pursuit of higher order needs such as socialization, esteem, and self-actualization (Davis et al, 1997).
Hypothesis 1: CEOs with higher order needs motivation will be negatively associated with an agency orientation.
Hypothesis 2: CEOs with extrinsic motivation will be positively associated an agency orientation.
Managers may externalize organizational problems to avoid blame or shift responsibility to others and not identify with the organization (D'Aveni & MacMillan, 1990). Under the agency perspective, managers view the organization in instrumental terms, considering only how using it may personally benefit him or her (Jensen & Meckling, 1976), lacking an organizational perspective. Absent identification with the organization, an individual will tend to view it as a tool with which opportunistic ends can be pursued (Yan et al (2002).
Hypothesis 3: CEOs who have low identification with the organization will be positively associated with an agency orientation.
Opportunists tend to place importance on material gain rather than values, lacking commitment or allegiance to organizations. Their relationship with the firm is one of using it as a device with which to achieve self-interest focused objectives (Jensen & Meckling, 1976). Agency theory views value commitment as lacking economic utility, being an irrelevant factor in determining the behavior of an opportunistic individual (Davis et al, 1997).
Hypothesis 4: CEOs with low value commitment will be positively associated an agency orientation.
Gibson, Ivancevich and Donnelly (1991) classified power as either personal or institutional. Opportunists will tend to employ coercive or institutional power which is derived by virtue of their formal authority in the organization, rather than forms of power based on relationships and trust, such as expert or referent power (Davis et al, 1997).
Hypothesis 5: CEOs with low personal power will be positively associated with an agency orientation.
Lawler (1986) contrasted management philosophies as control- oriented versus involvement-oriented. Under involvement-oriented approaches, managers undertake challenges and responsibility where they develop self-control over their behaviors (Davis, et al, 1997). Control-oriented managers remained detached from the organization, avoiding vulnerability and risk (Mayer, Davis, & Schoorman, 1995), leading to diminished involvement in the affairs of the company. The control-oriented approach is consistent with agency theory's view that opportunistic mangers use a firm to pursue their own self- interest.
Hypothesis 6: CEOs at companies with low involvement situations will be positively associated with an agency orientation.
Davis et al (1997) employed Hofstede's (1991) individualism - collectivism dimension as a cultural component to distinguish between agency and stewardship orientations in managers. Hofstede (1991) described an individualism-collectivism dichotomy where an individualism-focused culture emphasizes personal goals over group goals, consistent with theories of opportunism, and collectivist cultures where the self is defined as a part of the group, where one's group memberships are an important statement of identity and achievement. Situations reflecting an individualism-focused culture will tend to nurture agency-oriented, opportunistic behaviors among its employees (Davis et al, 1997).
Hypothesis 7: CEOs at companies with individualistic cultures will be positively associated with an agency orientation.
Power distance is defined as "the extent to which less powerful members of institutions and organizations within a country expect and accept that power is distributed unequally" (Hofstede, 1991: 28). In an organization with high power distance, there is an acceptance of large intergroup differences in authority, salary, and perquisites. Davis et al (1997) argued that low power distance situations are associated with stewardship and that high power distance is associated with agency theory.
Hypothesis 8: CEOs at companies with high power distance will be positively associated with an agency orientation.
Measures and Method
We created and mailed a survey to 500 randomly selected publicly listed U.S. corporations, targeting both the CEO and the boards of directors of these companies in 1997. The result of the initial mailing and follow-ups is that 161 directors and 135 CEOs responded. Statistical tests comparing the mean response for the variables between waves of data collection were insignificant, suggesting that later respondents were similar to earlier respondents. We yielded 102 matched pairs of CEO and director responses. Missing data reduced the number of companies in our analysis to 100. This study meets the minimum acceptable response rate of 20% for survey-based research (Hitt, Hoskinson, Johnson, & Moesel, 1996). Unmatched responses were used for factor analysis, but not included in hypothesis testing.
Dependent Variable We measured the dependent variable with director surveys, developing a four-item scale measuring the agency orientation they recognize in their CEOs behavior. A principal- components factor analysis determined whether this set of components could be used. The threshold established in advance for the selection of factor items were a factor loading of .50 or greater and at least a .20 difference between the item's loading with its factor and each of the other factors. The analysis of the component items indicated a one-component solution, agency, which explained 66.8% of the variance. Thus, the items were totaled for the subsequent analysis.
The independent variables were measured with CEO surveys. A 19- item scale developed by Frankfurter, Davis, & Vollrath (2001) measured high order and extrinsic motivation. A 10-item scale developed by Schechter (1985) measured value commitment. We employed an 11-item scale developed by Frost and Stahelski (1988) that measured personal power. The scale is based on French and Raven's (1959) bases of social power and corrects several scale format confounds found in previous research. Referent and expert power subscales were totaled to form a scale representing personal power. Coercive, legitimate, and reward power subscales were totaled to form a scale representing power distance.
We developed a six-item scale to measure identification, a three- item scale to measure involvement, and a five-item scale to measure culture. We used factor analysis with the CEO surveys, employing the same decision rules with these three independent variables as we did with the dependent variable. The analysis of the 14 items indicated a three-component solution. The first component (identification) had six items and explained 23.4% of the variance, the second (involvement) had three items and explained an additional 16.9% of the variance, and the third (culture) had five items and explained an additional 15.4% of the variance.
We included several control variables commonly found in corporate governance research; Profitability, capital expenditures, CEO stock ownership, duality, firm size, long-term compensation, and CEO board tenure. We measured profitability as the difference between firm and industry returns on equity, based four-digit SICs. Capital expenditures were computed as a firm's capital expenditures divided by its total assets and subtracting the equivalent industry benchmark, again based on four-digit SICs. CEO stock ownership was the percent of common stock owned. Duality was coded 1 when the CEO also served as board chair and O when he/she did not. Firm size was measured as the log of total employees. We computed long-term compensation as the percent of long-term compensation divided by total compensation. Board tenure was measured as the year the CEO was first appointed to the board (e.g., "84" would represent 1984). For all the firms examined, the CEO served on the board of directors. We obtained profitability, capital expenditure, and firm size data from Research Insight. Proxy statements provided CEO stock ownership, duality, contingent compensation, and CEO board tenure data.
The data for all of the dependent, independent, and control variables were continuous. Therefore, we conducted hypothesis testing with multiple regression analysis.
Descriptive statistics, alphas, variation inflation factors (VlFs), and the correlation matrix are reported in Table 1. Reliability was assessed by computing alpha scores and resulted in values of .50 or greater for all of the survey-based variables. These alphas all exceeded the minimum standard of .50 for experimental variables and were deemed acceptable (Nunnally, 1967). We observed no multi-collinearity problems that would affect the results, with only one variable (high order needs) exhibited a VIF that exceeded 2.0, far from the critical limit of 10 (Netter, Wasserman, & Kutner, 1989).
Multiple regression results for agency-orientation analysis appear on Table 2. Model 1 possessed no explanatory power over the independent variable, agency, with an insignificant adjusted R2 of .04. Model 2 examined the explanatory power of just the independent variables. This model reports a significant adjusted R2 of .23. Model 3 displays results for both the independent and control variables, with a significant adjusted R2 of .23.
Table 1. Descriptive Statistics, Alphas, Variation Inflation Factors (VIFs), and the Correlation Matrix
Hypotheses 1, 2, and 6 were supported. CEOs pursuit of higher order needs (Hypothesis 1) had a negative association with agency and extrinsic motivation (Hypothesis 2) had a positive association. In fact, both results were highly significant. Involvement (Hypothesis 6) was negatively associated with agency. One control variable, board tenure, was positively associated with agency at a statistically significant level. Overall, the agency model yielded good explanatory power explaining highly significant levels of variance in Models 2 and 3.
Table 2. Multiple Regression Results for Agency Orientation Analysis
Davis et al (1997) argued that stewardship- and agency-focused relationships result as a matter of choice between managers and principals. Our study finds the CEO-reported antecedents for agency associated with director perceptions of CEO agency orientation, consistent with cell 1 of the Davis et al (1997) principal-manager choice model. From our results, we conclude that directors tend to view a CEO as agency-oriented when the CEO reports psychological and situational factors consistent with that archetype.
The findings of this article are significant in three additional areas. First, while Williamson (1985) argued that one cannot know ex ante which managers will act opportunistically and so agency controls should be universally applied, Uzzi (1997) provide arguments and/or evidence to the contrary. Our findings show that directors correctly perceived agency behaviors exhibited by the CEO when the CEO reported psychological and situational factors consistent with the predictions of Davis' et al (1997) model. Given this evidence, we conclude that directors recognize an agency- orientation in CEOs when psychological and situational factors support opportunism.
Second, this study directly addresses the psychological basis for agency theory. Such research tends to use secondary data and yields inconsistent findings (Dalton, et al, 2003). This gap in the body of agency theory research has led to criticisms of the predictions and results of agency-oriented research over the years (Eisenhardt, 1988). Our findings confirm the essential core of agency theory with respect to CEO motivation.
Third, this study adds to previous research by employing survey- based data provided by both directors and CEOs at the same company. Corporate governance research is largely driven by data availability and is therefore limited in its predictive ability. This is true of previous research of agency theory, which has been plagued by inconsistent findings over time. We believe that our study's use of primary data serves as an example of the benefits of using primary data over secondary data, providing a standard for the conducting of future agency theory research.
D'Aveni, R. A. & MacMillan, I. C. 1990. Crisis and content of managerial communications: A study of the focus of attention of top managers in surviving and failing firms. Administrative Science Quarterly, 35: 634-657.
Dalton, D. R., Daily, C. M., Certo, S. T., & Roengpitya, R. 2003. Meta-analyses of financial performance and equity: fusion or confusion? Academy of Management Journal, 46: 13-26.
Davis, J. H., Schoorman, F. D. & Donaldson, L. 1997. Toward a stewardship theory of management. Academy of Management Review, 22: 20-47.
Eisenhardt, K.M. 1988. Agency- and institutional-theory explanations: The case of retail sales compensation. Academy of Management Journal, 31:488-511.
Frankfurter, S. A., Davis, J. H., & Vollrath, D. A. 2001. Why implement the dual governance structure? Much ado about something. Central Business Review, 20 (1-2): 4-9.
French, J. R. P. & Raven, B. 1959. The bases of social power. In D. Cartwright (ed.), Studies in Social Power. Ann Arbor: University of Michigan, Institute for Social Research.
Frost, D. E. & Stahelski, A. J. 1988. The systematic measurement of French and Raven's bases of social power in workgroups. Journal of Applied Social Psychology, 18: 375-389.
Gibson, J.L., Ivancevich J.M., & Donnelly, J.H. 1991. Organizations. Homewood, IL: Irwin.
Hill, C. W L. & Jones, T. M. 1992. Stakeholder-agency theory. Journal of Management Studies, 29: 131-154.
Hitt, M. A., Hoskinson, R. E., Johnson, R. A., & Moesel, D. D. 1996. The market for corporate control and firm innovation. Academy of Management Journal, 39: 1084-1119.
Hofstede, G. 1991. Cultures and organizations: Software of the mind. London: McGraw-Hill.
Jensen, M. C. & Meckling, W. H. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3: 305-360.
Lawler, E.E. 1986. High involvement management. San Francisco: Jossey-Bass.
Mayer, R. C., Davis, LH. & Schoorman, F. D. 1995. An integrative model of organizational trust. Academy of Management Review, 20: 709- 734.
Netter, J., Wasserman, W. & Kutner, M. H. 1989. Applied regression models. Homewood, IL: Richard D. Irwin.
Nunnally, J. C. 1967. Psychometric theory. New York, McGraw- Hill, Inc.
Schechter, D.S. 1985. Value and continuance commitment: A field test of dual conceptualization of organizational commitment. Unpublished master's thesis, University of Maryland, College Park.
Tosi, H., A. Brownlee, P. Silva & L Katz (2003). An empirical exploration of decisionmaking under agency controls and stewardship structure. Journal of Management Studies, 40(8): 2053-2071.
Uzzi, B. 1997. Social structure and competition in interfirm networks: The paradox of embeddedness. Administrative Science Quarterly, 42: 35-67. Williamson, O. E. 1985. The economic institutions of capitalism. New York: Free Press.
Yan, A., Zhu, G.., & Hall, D. T. 2002. International assignments for career building: A model of agency relationships and psychological contracts. Academy of Management Review, 27:373-391.
Steven A. Frankforter
James H. Davis
University of Notre Dame
David A. Vollrath
University South Bend
University of Louisiana
Contact email address: [email protected]
Steven A. Frankfurter
Dept of Management and Marketing
SC 29733 USA
James H. Davis
Department of Management
University of Notre Dame
IN 46556 USA
David A. Vollrath
School of Business and Economics
Indiana University South Bend
IN 46634 USA
College of Business Administration
University of Louisiana Lafayette
LA 70504 USA
Copyright International Journal of Management Sep 2007
(c) 2007 International Journal of Management. Provided by ProQuest Information and Learning. All rights Reserved.