DISENTANGLING CORPORATE REPUTATION: HOW MUCH DO ...

ASAC 2008 Halifax, Nova Scotia

Sujit Sur (Ph.D. Student) Carol-Ann Tetrault Sirsly (Ph.D. Student) John Molson School of Business Concordia University

DISENTANGLING CORPORATE REPUTATION: HOW MUCH DO TIME, FIRM AND INDUSTRY MATTER?

We decompose corporate reputation over time, and between firms and industry utilizing random coefficient modeling on a ten year sample of major US firms rated as Fortune’s Most Admired Companies. We find that the major variation (50%) in reputation is firm specific and performance accounts for only 19% of variation.

Introduction Firm reputation is an intangible asset that within a resource-based view can provide a sustainable competitive advantage (Barney, 1991; Carter & Ruefli, 2006; Fombrun, 1996; Roberts & Dowling, 2002). While reputation is built over time, it is highly vulnerable to being tarnished and risks being lost in no time at all (Carter & Ruefli, 2006; Davies, Chun, Vinhas da Silva & Roper, 2003; Hall, 1992 & 1993). One only has to think of the fall and subsequent demise of Arthur Andersen & Co. to see how rapidly a stellar reputation may be reversed. When considering how firm reputation may benefit the firm over time, the various advantages of a good reputation include cost savings as suppliers and employees seek to be associated with the firm, favourable access to capital given the perception of lower risk, as well as being able to charge premium prices to generate superior margins, (Carter & Ruefli, 2006; Fombrun, 1996 & 2001; Fombrun & Shanley, 1990; Roberts & Dowling, 2002). Attributed as key to firm success (Hall, 1992 & 1993), but only durable to the few (Carter & Ruefli, 2006) and rarely over a long time-frame (Wiggins & Ruefli, 2002), understanding the evolution of reputation might be considered as integral to firm strategy. Although the valuable nature of a good reputation is undisputed (Fombrun, 1996 & 2001; Fombrun & Shanley, 1990) and has been associated with enhanced market value (Black, Carnes & Richardson, 2000), the causality between reputation and financial performance is much less evident (Fombrun, 2001; Roberts & Dowling, 2002). The sustainable competitive advantage of a favourable firm reputation has captured much research attention (Carter & Ruefli, 2006; Kotha, Rajgopal & Rindova, 2001; Obloj & Obloj, 2006; Williams, Schnake & Fredenberger, 2005), however less is known about the roles of industry positioning (Shamsie, 2003) and time (Roberts & Dowling, 2002) on reputation. Considering the nesting of firm reputation within industry reputation (Shamsie, 2003), when assessing the evolution of reputation over time we question the relative contribution of the firm, industry and time to the change in firm reputation. Addressing methodological criticisms of traditional variance analyses techniques (Short, Ketchen Jr., Bennett & Du Toit, 2006), we utilize the variance decomposition technique of random coefficients modeling (RCM) to disentangle the effects of this nesting. The advantage of RCM for variance decomposition over the more traditional analysis of variance (ANOVA) methodology found in most reputation studies is RCM’s underlying assumption of heteroscedasticity (i.e. variance of error terms is not equal), as contrasted with the homoscedasticity assumption of ANOVA. As firms are nested within industries and over time, the variances in the error terms at multiple levels are inherently dependent and not constant. RCM appropriately analyzes the multi-level structure of 150

reputation, allowing us to examine the multi-level effects on corporate reputation over time, across firms and within/between industries. We consider this a necessary first step to understanding how much each level influences corporate reputation. We are then in a better position to parse out the relative importance of each level to then appropriately theorise and test the variables of interest to respond to Greve and Goldeng's (2004) suggestions for improved longitudinal analysis in strategic management research. We focus our research question on the evolution of firm reputation over time to decompose the effects of time, firm and industry. Following the literature review we describe our measures and research methodology. We then discuss the results, review limitations and expand on the direction for future research as well as practical applications to reputation management.

Decomposing the Constituents of Reputation Firm and Industry Reputation The most widely used management definition of reputation (Wartick, 2002) is "a perceptual representation of a company's past actions and future prospects that describes the firm's overall appeal to all of its key constituents when compared with other leading rivals" (Fombrun, 1996: 72). Reputation is thus rooted in stakeholders' comparisons of firms, rather than against the societal standard of legitimacy (Deephouse & Carter, 2005; Wartick, 2002). Reputation is enhanced in the eyes of stakeholders when they perceive the firm's actions to be consistent with their expectations and compare favourably against other firms (Logsdon & Wood, 2002; Mahon, 2002; Wartick, 2002; Whetten & Mackey, 2002). Miller noted "capability and reputation cycles reinforced one another", underlining the value of creating these "reputation resources" (2003: 970) over time. Like capabilities, reputation is an intangible asset embedded in the firm (Roberts & Dowling, 2002; Granovetter, 1985). The potential for value creation in an intangible asset such as corporate reputation is inherent in the difficulty to replicate a competitor’s reputation due to the ambiguity as to how it was derived (Roberts & Dowling, 2002). Thus a firm's reputation is the "overall estimation in which a company is held by its constituents" (Fombrun, 1996: 37), and will evolve over time (Hall, 1992 & 1993; Mahon, 2002). While the shortterm influence of advertising may enhance recognition in the marketplace, the esteem component of reputation with which a firm is held may only be built over the longer term (Hall, 1992 & 1993). However, a firm's individual reputation is also nested in that of the industries within which it is associated, linking it to not only the other firms within the industry but also how the overall industry reputation relates to other industries (Eccles, Newquist & Schatz, 2007). The relative influence of an excellent industry reputation on that of an average firm may infer free-riding reputation advantages, while a good firm may be held back, captive by a poor industry reputation (Mahon, 2002). Dolphin (2004: 85) notes an international banker’s views that “banking is a reputation business”, thus situating professional services as more vulnerable to reputation shocks. One of the most widely utilized measures of reputation is the Fortune Most Admired Companies annual rating (Fombrun, 1996; Harrison & Freeman, 1999). While Fryxell and Wang (1994) criticize the over reliance on financial performance, Davies and colleagues (2003) support reputation rankings albeit with the caution that they may not predict future performance as they reflect past accomplishments. The Fortune measure is inherently comparative within a given industry, thus nesting firm reputation within industry reputation, necessitating analysis at multiple levels. While there has been much critique of Fortune's annual survey of corporate reputations in their America's Most Admired Companies ratings, ranging from the halo effect of previous financial performance (Brown & Perry, 1994; Deephouse, 2000) to the choice of respondents and aggregation of 151

data (Wartick, 2002), it is none the less a popular measure of reputation (Dolphin, 2004; Wartick, 2002). Fortune has been tabulating reputation ratings of some 300 US companies since 1982. The respondents to the Fortune survey include industry executives and market analysts reporting within their own areas of expertise, where they conceivably have ample access to information about the focal firm as well as the industry. For a more detailed description of the Fortune reputation rating and the methodology used for collecting the data see the Fortune Datastore (2008). Ownership and Governance Considerations At the firm level, ownership structure and governance mechanisms have had their impact on firm reputations, as evidenced most recently with highly publicized trials such as that of Conrad Black (Khalil & Magnan, 2007). Some of the issues that haven given rise to reputational damage include excessive executive compensation (i.e. Robert Nardelli at Home Depot), dual-class shares where founders exploit other shareholders via super-voting rights (i.e. Frank Stronach at Magna) and dysfunctional boards in disarray (i.e. Hewlett-Packard under Patricia Dunn’s chairmanship). Incoherent or divergent decisions, such as demanding concessions from unions while protecting executive pensions when Donald Carty was CEO at American Airlines (Eccles et al., 2007), also reflect on the firm’s reputation. More independent directors have been proposed (Khalil & Magnan, 2007; Murray, 2003) to enhance governance and accordingly reduce reputational risk, however empirical research (Dalton, Daily, Ellstrand & Johnson, 1998) is less conclusive on the benefits of director independence. The chief executive officer (CEO) may be seen as the most visible face within the firm’s industry and to the investment community. The distinction as to whether the same individual chairs the board as well as runs the firm has often been termed as duality, as in wearing both hats. In providing a sole leadership the firm reputation may be seen to be advantaged by duality, avoiding multiple signals that may be sent when two individuals steer the firm. Specific CEO characteristics that may be viewed to positively promote firm reputation include CEO tenure with the firm (Carter, 2006) and when the CEO was promoted from within the firm. Whether ownership is diffused or concentrated, be it by a founder, blockholder or institution, is possibly a further element in the reputational puzzle. The transference of a founder’s personal reputation to the corporate reputation is a distinct possibility, particularly given the executive role often maintained by founders. At the opposite end of the ownership spectrum, a widely held firm falls under the scrutiny of securities analysts, viewed as “reputational intermediaries” (Chen, 2008: 191), and as such may prioritize reputation management as a mechanism for protecting share price. Fombrun and Shanley (1990) found institutional ownership to provide a positive effect on firm reputation, suggesting the informed screening of institutions when becoming shareholders favourably influenced public opinion. The Temporal Factor The enduring quality of reputation that has been referred to as sticky (Kraatz & Love, 2006; Roberts & Dowling, 2002; Schultz, Mouritsen & Gabrielsen, 2001) highlights the importance of studying the change over time in a firm’s reputation. The resilience of reputation, particularly where a strong reputation has been cultivated (Fombrun, Gardberg & Barnett, 2000) indicates that a lag effect may exist before reputation assessments are revised. Fombrun (2001: 294) refers to the “winner-take-all” nature of the reputational markets, where attention is given to the most noticeable firms, making it more difficult for competitors to challenge and to improve their reputations. However he also notes how corporate reputation may plunge when a crisis strikes and how recovery will depend on both the pre-crisis reputation and how well the firm is perceived to deal with the situation. This may be viewed as a long uphill battle to build reputation with well received deeds, but a slippery slope to destroy reputation with a misstep. 152

Rumelt (1991) was the first to identify the temporal effect on performance not explained by firm and industry effects. More recently, Short and colleagues (2007) found time to account for the majority of performance variation, followed by firm specific effects, with industry having the least effect. Using a 25 year analysis of 40 industries, Wiggins & Ruefli (2002) identify the rarity of sustained superior performance, particularly over any significant number of years. Thus we utilize a multilevel analysis of corporate reputation as it allows us to model all three levels of interest simultaneously – a statistical advance over previous methods (Short et al., 2006) - and allows us also to capture and model the determinants of each firm’s and each industry’s unique year-to-year growth in reputation. In contrast to previous reputation research, we are thus able to measure and partition the variation in slopes and in intercepts of the reputation growth model over time, firms and industries, rather than assume that these are constant across the sample of firms. Roberts and Dowling (2002) examined 15 years of Fortune reputation ratings using the underlying premise of the stability of reputation over time to link a good reputation to profit persistence. However Carter and Ruefli's (2006) ordinal time series analysis over 11 years shows a gradual building of reputation over time, while a more accelerated pace is noted for erosion. Short and colleagues (2006) suggest the influence of macro economic conditions be considered. Referring to Figure 1, the curvilinear trend of the S&P 500 index over time can be expected to be reflected in the trend of the firms and industries within these macro conditions.

------------------------------------insert Figure 1 about here -------------------------------------

Given these observations on the role of time, firm and industry on corporate reputation, we propose the following broad hypotheses:

Hypothesis 1: There will be a significant variance in corporate reputation over time, between firms and between industries. Hypothesis 2: There will be a positive relationship between present and prior firm performance and corporate reputation over time. Hypothesis 3: Corporate reputation will follow a curvilinear trend over time Hypothesis 4: Board independence will have a positive impact on reputation at the firm level. Hypothesis 5: Diffused ownership will have a positive impact on reputation at the firm level. Hypothesis 6: Institutional ownership will positively impact reputation at the firm level. Hypothesis 7: Industry reputation will follow a curvilinear trend over time.

Methodology Sample and Data Sources 153

Using the Fortune Most Admired Companies data from the magazine’s 1998 to 2007 issues (which reported the reputation score for 1997 to 2006 period), we compiled the overall reputation ratings for all of the firms in all of the industries, as covered in each of those years. Any firm not present for at least three years during the time period was removed from the sample, as were firms that merged or were acquired by another firm being reported. This resulted in a final sample size of 2204 firm-years reputational data for 293 major public firms across 58 industries (as identified by the GCIS industry description that generally match to the Fortune industry definitions). All financial data came from the COMPUSTAT database. The Board data was compiled from the RiskMetrics database, while the ownership data was compiled from the Osiris database from Bureau van Djik Electronic Publishing. Measures Dependent Variables Firm overall reputation (REP_SCORE) as rated by Fortune's Most Admired Companies is our dependent variable. This rating from zero to ten is based on an annual survey of senior executives and board directors of Fortune 1000 companies, as well as financial analysts, where they rate the major companies in their industry on eight reputation drivers. These include the quality of management, product and service quality, innovation, use of corporate assets, personnel development and ability to attract talent, long-term investment value, financial performance and social responsibility. The most recently published 2007 ratings were based on fourth quarter, 2006 survey responses from over 3,300 executives, directors and analysts representing over 600 companies in almost 70 industries. Fombrun and Shanley (1990: 245) found “that the eight attributes elicited from respondents were components of an underlying and stable construct of reputation”, further supported by Roberts and Dowling’s (2002) longitudinal examination of the link of reputation to persistent superior financial performance. Kraatz and Love (2006) note the importance of using reputation measures which are significant to firms, singling out Fortune’s ratings for special consideration. One example of how the Fortune placing forms part of the strategic goals may be found in General Electric’s most recent proxy statement where the CEO’s compensation related to risk and reputation management is assessed as “GE remains one of the most admired companies...Fortune (#1)” (GE, 207: 19). Independent Variables Time Level Variables: Three time-varying independent variables were selected: Return on Assets (ROA), 3 years Market Value CAGR (MKT_V3GR), and 1 year Total Returns (TRTN_1YR). These are well-recognized accounting and market measures in linking reputation to financial performance (Carter & Ruefli, 2006; Hall 1992 & 1993; McMillan & Joshi, 1997). As literature (Roberts & Dowling, 2002) points to previous financial performance having a lingering effect on reputation, we also included the lagged ROA measures going back up to six years prior to the reputation score. Firm size effects were controlled by selecting the Net Sales (SALES) measure. Time was coded (YR_INTGR) using 0 for 1997 and incrementing this value by 1 for each year thereafter to assess the temporal trend. To assess the curvilinear trend, we also included the squared term of the year integer (YRITG_SQ). Firm Level Variables: At the firm level, a time invariant (i.e. firm specific) factor that impacts reputation is organizational slack. The strong link to innovation of slack resources (Damanpour, 1991) as well as their necessity to firm survival over the longer-term (Sharfman, Wolf, Chase & Tansik, 1988) reinforces the contribution of slack to enhancing reputation. Lawson (2001) notes the variety of forms that slack may take, such as unused capital as found in financial leverage as well as a variety of other financial ratios (Singh, 1986). Thus we use the smoothened Financial Leverage (FI_LEVER), i.e. the median value of the financial leverage of the firm over the years. Corporate governance was considered in the proportion of independent outside directors on the board, shown as Board Independence (INDEPD_P). Ownership characteristics were assessed as the total number of shareholders reported for 154

the firm, shown as Ownership Diffusion (NO_SHRDR), while institutions were calculated as the total percentage of shares held by financial institutions like pension funds, mutual funds, banks, insurance companies and any other financial company, shown as Institutional Ownership (INSTI_OW). Industry Level Variables: At the industry level, we utilize the Industry Munificence (NEWMUN) and Industry Dynamics (NEWDYN) as commonly used in IO economics (Castrogiovanni, 2002, Dess & Beard, 1984) to assess the industry specific effect on reputation. Additionally, we include industry discretion (Cordeiro & Rajagopalan, 2003; Finkelstein & Boyd, 1998; Hambrick & Abrahamson, 1995) as a factor of capital intensity (i.e. property, plant and equipment divided by employees) and firm sales growth.

Results and Discussion Consistent with a recent trend (Misangyi, Elms, Greckhamer and Lepine, 2006; Short et al., 2007), we use hierarchical linear modeling (HLM) to explore firm and industry reputation over time, as HLM allows us to model all three levels of interest simultaneously – a statistical advance over previous methods (Short et al., 2006). HLM can also capture and model the determinants of each firm’s and each industry’s unique year-to-year growth in corporate reputation. A further feature of HLM is that it “allows unequal numbers of spacing and time points and thus larger sample sizes in many cases” (Short et al., 2006: 277), thus HLM offers a distinct advantage over structural equation modeling when considering missing data in the Fortune ratings. We utilize the HLM 6.03 package (Raudenbush, Bryk, Cheong, Congdon & du Toit, 2004) to model time varying factors at level 1, firm variables at level 2 and the industry variables at level 3 (see Short et al. (2006) for an explanation of this modeling approach). Table 1 gives the descriptive statistics of the variables at each level of analysis. We first ran the fully unconditional random coefficient model i.e. the model without predictor variables at any level, to assess the decomposition of the variance in reputation at each level of analysis. The results are reported at Table 2.

------------------------------------insert Tables 1 & 2 about here -------------------------------------

As can be seen from the results, the variations in corporate reputation are distributed over time, firm and industry levels. Time accounts for 45.67% of the variation, firm accounts for 50.06%, while industry accounts for only 5.08%. The chi square values were significant at p< 0.000 and p< 0.005 levels respectively. Thus hypothesis 1 that predicted significant variation over time and across firms is strongly supported and marginally supported for variation between industries. We thereafter ran a series of models adding the variables as suggested by literature at appropriate levels. We started by specifying the coefficients as randomly varying (Raudenbush et al., 2004: 82), and then respecified appropriately. We also compared each model to the null model to check the goodness of fit ( Ȥ2 ) value to identify the model with the best fit. The result for the complete three-level model with the best fit, along with the equation sets associated with the model, is reported in Table 3. ------------------------------------insert Table 3 about here ------------------------------------155

As the results show, time variant firm performance indicators like 1 Year Total Return, 3 Years Market Value CAGR, Return on Assets are all positively and significantly related to corporate reputation, as is firm size indicator of Net Sales. However, the lagged ROA relationship was found to be more complex as 1 Year Lagged ROA was not significant, 2 Years Lagged ROA was positive and significant, while 3 Year Lagged ROA was not significant, but was negative. None of the 4, 5 and 6 Years Lagged ROA were significant, however their inclusion significantly deteriorated the model fit. Financial Leverage showed no significant relationship with reputation and actually degraded the model fit. Thus hypothesis 2 predicting a positive relationship between reputation and firm’s present and past performance is only partially supported. The Year Integer Indicator and the squared term were both significant, however the year indicator was negatively related while the squared term was positively related to corporate reputation. Both the firm level reputation linear change rate, as well as the curvilinear change rate had significant chi square values (p< 0.000), thus hypothesis 3 predicting a curvilinear relationship between reputation and time is strongly supported. At the firm level of analysis, the Proportion of Independent Directors on the board is not significantly related to reputation, thus hypothesis 4 is not supported. The Number of Shareholders is positively and significantly related to reputation, thus hypothesis 5 predicting a positive relationship between reputation and diffused ownership is strongly supported. Finally, the percentage of shares held by Financial Institutions was significantly related to reputation, however the relationship was in the opposite direction, thus hypothesis 6 was not supported. A post-hoc rationalization might be that as this hypothesis was inspired by Fombrun and Shanley’s (1990) study using 1985 data, institutions were then seen as the gold standards of respectability. However, given the scandals and rogue trading that have plagued institutions more recently, the results are less surprising and do warrant further study. At the industry level of analysis, both Industry Munificence and Industry Discretion were negatively and significantly related to reputation, while Industry Dynamism had no significant relationship. Also, the chi square values for the industry level linear as well as curvilinear change rates were significant (p< 0.006 & p< 0.002). Thus hypothesis 7 predicting a curvilinear relationship between reputation and industry is also strongly supported.

Limitations and Future Research Our results (Tables 2 & 3) show that the vast proportion of the variation in the reputation scores for this large sample of US firm-years occurs within firms over time as a result of changes in sales, market value and ROA over time. However, the lagged performance variables have mostly non-significant coefficients, suggesting that the financial performance ‘halo’ effect on reputation is short-lived. Also the largest component of variation is between firms rather than between industries. We also find that there is significant variation in both intercepts and slopes across the three levels studied, and that the largest variance component is explained by the firm’s initial reputation rating. As can be seen from Table 4, the time varying performance variable explains only 40.91% of the variation at the time level of analysis, and thus only 18.68% (40.91% of 45.67%) of total variation in corporate reputation. Therefore about 27% (45.67% less 18.68%) of total variation that is attributable to the time level of analysis is still unexplained. We hope to capitalize on our findings by modeling 156

potential time-varying economic influences such as growth in GNP and the overall stock market performance in future research. Future studies can also theorise and investigate the theoretically grounded possible time varying influences on corporate reputation.

------------------------------------insert Table 4 about here -------------------------------------

At the firm level, the board independence and ownership diffusion variables explain only 7.29% (14.57% of 50.06%) of the total variation in corporate reputation, leaving almost 43% (50.06% less 7.29%) of total variation in corporate reputation unexplained. There seems to be a major ‘stickiness’ aspect to reputation at the firm level that is not explained by the normally accepted governance indicators. Based on our findings, we urge future research to focus efforts at this level of analysis. We pointed earlier to the Founder and/or CEO reputation influencing the firm’s reputation. Also, the ‘brand’ or image of the firm might be the possible reason for the enduring reputation for a firm. These and other such possible factors that might influence reputation might be the most fruitful avenues for future reputational studies. Our finding of low industry effects is also intriguing. The theoretically derived variables explain only about 0.68% (13.48% of 5.08%). We speculate that the methodology of the Fortune survey might not enable between industries analysis of reputation. The respondents to the Fortune survey are restricted to industry insiders and thus do not take into consideration the perspective of a broad cross-section of stakeholders. The primary source of revenues determined the industry classification, thus diversified firms were compared with single industry companies. Furthermore, Fortune's ranking is limited to the top ten firms in any given industry. They publish ratings only for the largest publicly traded firms (i.e. the top ten in any given industry for a given year), thus this data has an extreme left censoring bias. Future work will examine this issue in greater detail. Disentangling the elements that contribute to the evolution of firm reputation is a starting point to future multi-level research that may also want to consider strategic group membership and other reputational spill-over possibilities, such as association with highly reputable investors, partners or other endorsements. For executives managing corporate reputation, insight into the salience of various contributors to firm reputation can fuel risk management and resource allocation strategies to sustain competitive advantage. In expanding our understanding of how the intangible asset of corporate reputation evolves, researchers and practitioners can develop proactive strategies to leverage corporate reputation. As it stands, this study enables clarity and helps us focus on the level of analysis that is of greatest importance, and allows us to take stock of how much we know about the mechanisms of corporate reputation, and hopefully motivates the direction of future research.

References Barney, J.B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17: 99-120. 157

Black, E., Carnes, T. & Richardson, V. (2000). The market value of corporate reputation. Corporate Reputation Review, 1: 31-42. Brown, B. & Perry, S. (1994). Removing the financial performance halo from Fortune's "most admired" companies. Academy of Management Journal, 37: 1347-1359. Carter, S.M. (2006). The interaction of top management group, stakeholder, and situational factors on certain corporate reputation management activities. Journal of Management Studies, 43(5): 1145-1176. Carter, S.M. & Ruefli, T.W. (2006). Intra-industry reputation dynamics under a resource-based framework: Assessing the durability factor. Corporate Reputation Review, 9: 3-25. Castrogiovanni, G. J. (2002). Organization task environment. Have they changed fundamentally over time? Journal of Management, 28(2), 129-150. Chen, Y. (2008). International dual listing: An analytical framework based on corporate governance theory. Journal of American Academy of Business, 12(2): 187-194. Cordeiro, J.J. & N. Rajagopalan. (2003). Industry discretion as a determinant of the mix and level of executive compensation: A multilevel analysis. Academy of Management Best Papers Proceedings. Dalton, D., Daily, D.M., Ellstran, A.E. & Johnson, J.L. (1998). Meta-analytic reviews of board composition, leadership structure, and financial performance. Strategic Management Journal, 19(3): 269-290. Damanpour, F. (1991). Organizational innovation: Academy of Management Journal, 34: 555-590.

A meta-analysis of effects of determinants.

Davies, G., Chun, R., Vinhas da Silva, R. & Roper, S. Competitiveness. London: Routledge.

(2003).

Corporate Reputation and

Deephouse, D.L. (2000). Media reputation as a strategic resource: An integration of mass communication and resource-based theories. Journal of Management, 26: 1091-1112. Deephouse, D.L. & Carter, S.M. (2005). An examination of differences between organizational legitimacy and organizational reputation. Journal of Management Studies, 42: 329-360. Dess, G. G., Beard, D. W. (1984). Dimensions of organizational task environments. Administrative Science Quarterly, 29(1), 52-73. Dolphin, R.R. (2004). Corporate reputation – a value creating strategy. Corporate Governance, 4(3): 77-92. Eccles, R.G., Newquist, S.C. & Schatz, R. (2007). Reputation and its risks. Harvard Business Review, February: 104-114. Finkelstein, S. & Boyd, B.K. (1998). How much does the CEO matter? The role of managerial discretion in the setting of CEO compensation. Academy of Management Journal. 41(2): 179-199. Fombrun, C.J. (1996). Reputation: Realizing value from the corporate image. Boston: Harvard Business School Press. Fombrun, C.J. (2001). Corporate reputations as economic assets. The Blackwell Handbook of Strategic Management, M.A. Hitt, R.E. Freeman & J.S. Harrison (eds): 289-312. Malden, MA.: Blackwell Publishers. Fombrun, C.J., Gardberg, N.A. & Barnett, M.L. (2000). Opportunity platforms and safety nets : Corporate citizenship and reputational risk. Business and Society Review, 105: 85-106. 158

Fombrun, C. & Shanley, M. (1990). What's in a name? Reputation building and corporate strategy. Academy of Management Journal, 33: 233-258. Fortune

Datastore. (2008). Fortune Corporate Reputation Industry Reports. http://www.timeinc.net/fortune/datastore/reputation/cr_report.html (Accessed January 25, 2008).

Fryxell, G. E. & Wang, J. (1994). The Fortune corporate "reputation" index: Reputation for what? Journal of Management, 20: 1-14. G.E. (2007). General Electric: Notice of 2007 annual meeting and proxy statement, February 28, 2007. Granovetter, M. (1985). Economic action and social structure: American Journal of Sociology, 91: 481-510.

The problem of embeddedness.

Greve, H.R. & Goldeng, E. (2004). Longitudinal analysis in strategic management. In D.J. Ketchen, Jr. & D.D. Bergh (Eds), Research methodology in strategy and management, Volume 1 (pp. 135163). Oxford: Elsevier Ltd Hall, R. (1992). The strategic analysis of intangible resources. Strategic Management Journal, 13: 135144. Hall, R. (1993). A framework linking intangible resources and capabilities to sustainable competitive advantage. Strategic Management Journal, 14: 607-618. Hambrick, D.C., & Abrahamson, E. (1995). Assessing managerial discretion across industries: A mutltimethod approach. Academy of Management Journal, 38(5): 1427-1441. Harrison, J.S. & Freeman, R.E. (1999). Stakeholders, social responsibility, and performance. Empirical evidence and theoretical perspectives. Academy of Management Journal, 42: 479-485. Khalil, S. & Magnan, M. (2007). Dual-class shares: Governance, risks, and rewards. Ivy Business Journal, May/June: 1-4. Kotha, S., Rajgopal, S. & Rindova, V. (2001). Reputation building and performance: An empirical analysis of the top-50 pure internet firms. European Management Journal, 19(6): 571-586. Kraatz, M.S. & Love, E.G. (2006). Studying the dynamics of reputation: A framework for research on the reputational consequences of corporate actions. Research methodology in strategy and management, D.J. Ketchen & D.D. Bergh (eds), 3: 343-383. Oxford, U.K.: Elsevier JAI. Lawson, M.B. (2001). In praise of slack: Time is of the essence. Academy of Management Executive, 15 (3): 125-135. Logsdon, J.M. & Wood, D.J. (2002). Reputation as an emerging construct in the business and society field: An introduction. Business and Society, 41: 365-370. Mahon, J.F. (2002). Corporate reputation: A research agenda using strategy and stakeholder literature. Business and Society, 41: 415-441. McMillan, G.S. & Joshi, M.P. (1997). Sustainable competitive advantage and firm performance: The role of intangible resources. Corporate Reputation Review, 1: 81-85. Miller, D. (2003). An asymmetry-based view of advantage: Towards an attainable sustainability. Strategic Management Journal, 24: 961-976. Misangyi, V.F., Elms, H., Greckhamer, T., & Lepine, J.A. (2006). A New Perspective on a Fundamental Debate: A Multilevel Approach to Industry, Corporate, and Business Unit Effects. Strategic Management Journal, 27: 571-590. Murray, K. (2003). Reputation: Managing the single greatest risk facing business today. Journal of Comunication Management, 8(2): 142-149. 159

Obloj, T. & Obloj, K. (2006). Diminishing returns from reputation: Do followers have a competitive advantage? Corporate Reputation Review, 9(4): 213-224 Raudenbuh, S., Bryk, A., Cheong, Y.F., Congdon, R. & du Toit, M. (2004). HLM 6: Hierarchical Linear and Nonlinear Modeling. Scientific Software International Inc. Lincolnwood, IL. Roberts, P.W. & Dowling, G.R. (2002) Corporate reputation and sustained superior financial performance. Strategic Management Journal, 23: 1077-1093. Rumelt, R. (1991). How much does industry matter? Strategic Management Journal, 13: 119-134. Schultz, M., Mouritsen, J. & Gabrielsen, G. (2001). Sticky reputation: Analyzing a ranking system. Corporate Reputation Review, 4: 21-41. Shamsie, J. (2003). The context of dominance: An industry-driven framework for exploiting reputation. Strategic Management Journal, 199-215. Sharfman, M.P., Wolf, G., Chase, R.B. & Tansik, D.A. (1988). Antecedents of organizational slack. Academy of Management Review, 13: 601-614. Short, J.C., Ketchen Jr., D.J., Bennett, N. & du Toit, M. (2006). An examination of firm, industry, and time effects on performance using random coefficients modeling. Organizational Research Methods, 9: 259-284. Short, J.C., Ketchen, Jr., D.J., Palmer, T.B. & Hult, G.T.M. (2007). Firm, strategic group, and industry influences on performance. Strategic Management Journal, 28: 147-167. Singh, J.V. (1986). Performance, slack, and risk taking in organizational decision making. Academy of Management Journal, 29: 562-585. Wartick, S.T. (2002). Measuring corporate reputation: Definition and data. Business and Society, 41: 371-392. Whetten, D.A. & Mackey, A. (2002). A social actor conception of organizational identity and its implications for the study of organizational reputation. Business and Society, 41: 393-414. Williams, R.J., Schnake, M.E. & Fredenberger, W. (2005). The impact of corporate strategy on a firm’s reputation. Corporate Reputation Review, 8(3): 187-197. Wiggins, R. & Ruefli, T.W. (2002). Sustained competitive advantage: Temporal dynamics and the incidence and persistence of superior economic performance. Organization Science, 13: 81-105.

160

Figure 1

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

Trend Line of the Value of the S&P 500 Index

161

Table 1 Descriptive Statistics for Reputation at Time, Firm and Industry Levels of Analysis LEVEL-1 (TIME LEVEL) DESCRIPTIVE STATISTICS VARIABLE NAME N MEAN SD MINIMUM MAXIMUM YR_INTGR 2204 4.45 2.67 0 9 YRITG_SQ 2204 26.92 24.46 0 81 REP_SCOR 2204 6.53 0.94 2.63 9.04 ROA 2203 5.62 9.5 -290.84 53.12 LAGROA_1 2203 5.68 12.11 -458.31 57.47 LAGROA_2 2202 5.95 6.88 -60.06 57.47 LAGROA_3 2197 6.04 6.83 -58.98 57.47 SALES 2189 19891.47 30874.66 833.1 335086 MKT_V3GR 2187 14.66 27.22 -52.82 267.75 TRTN_1YR 2196 25.85 436.92 -90.94 20388.89 LEVEL-2 (FIRM LEVEL) DESCRIPTIVE STATISTICS VARIABLE NAME N MEAN SD MINIMUM MAXIMUM REP_SCOR 293 6.55 0.85 3.73 8.75 FI_LEVER 293 3.43 29.62 -442.2 209.47 INDEPD_P 293 0.81 0.11 0.43 0.94 NO_SHRDR 293 49.17 34.29 1 111 INSTI_OW 293 53.63 18.56 0 111.46 LEVEL-3 (INDUSTRY LEVEL) DESCRIPTIVE STATISTICS VARIABLE NAME N MEAN SD MINIMUM MAXIMUM REP_SCOR 58 6.49 0.84 3.73 7.93 NEWMUN 58 0.09 0.13 -0.49 0.52 NEWDYN 58 0.12 0.13 0.01 0.93 INDDISCR 58 0 0.09 -0.06 0.36

162

Table 2 Estimation of Reputation - Fully Unconditional Model (Robust Standard Errors) Fixed Effects (Robust standard errors)

Coefficient

SE

d.f.

T-Ratio

p-value

6.471040

57

120.816

0.000

Random Effects

Std. Dev.

0.053561 Variance Compont 0.42138 0.46184 0.04690

d.f.

Ȥ²

p-value

Level 1 Variance (across time), etij 0.64914 Level 2 Variance (between firms), rij 0.67959 235 2056.86047 0.000 Level 3 Variance (between industries), uj 0.21657 57 88.77105 0.005 Variance Decomposition by levels Percentage of total variance across time 45.67% Percentage of total variance between firms 50.06% Percentage of total variance between industries 05.08% Random level-1 Reliability Random level- 2 Reliability coefficient estimate coefficient estimate Intercept 1, P0 0.867 Intercept 1/Intercept 2, B00 0.282 The starting values used data from 2204 level-1(firm year), 293 level-2 (firm), 58 level-3 (industry) records Standard Error of sigma squared = 0.01362 Deviance = 4997.176125 (Number of estimated parameters = 4) Summary of the model specified (in equation format): Reputation Score = G000 + R0 + U00 + E Level-1 Model: Y = P0 + E Level-2 Model: P0 = B00 + R0 Level-3 Model: B00 = G000 + U00

The model specified for the fixed effects was: Level-1 Coefficients INTRCPT1, P0

Level-2 Predictors

Level-3 Predictors

INTRCPT2, B00

INTRCPT3, G000

163

Table 3 Estimation of Reputation - Fixed Effects Model (Robust Standard Errors) Fixed Effects (Robust standard errors) Level 3 Time Level Variables Industry Munificence (NEWMUN) Industry Dynamism (NEWDYN) Industry Discretion (INDDISCR) Level 2 Firm Level Variables No. of Shareholders (NO_SHRDR) % Institutionally Owned (INSTI_OW) Proportion of Independent Outside Directors as Board Members (INDEPD_P) Level 1 Time Level Variables Linear Change Rate (YR_INTGR) Curvilinear Change Rate (YRITG_SQ) Return On Assets (ROA) 1 Year Lagged ROA (LAGROA_1) 2 Years Lagged ROA (LAGROA_2) 3 Years Lagged ROA (LAGROA_3) Net Sales (SALES) Market Value 3 Yr CAGR (MKT_V3GR) Total Return – 1 Yr (TRTN_1YR) Random Effects Level 1: Temporal Variation, E0 Level 2: Firm Initial Reputation, R0 Firm reputation Linear Change Rate, R1 Firm Reputation Curvilinear Change Rate, R2 Level 3: Industry Mean Reputation, U00 Industry Reputation Linear Change Rate, U10

Coefficient

S.E.

T-Ratio

pvalue 54 0.000

d.f.

6.517053

0.042627

152.885

-0.541540 -0.191658 -0.606930

0.042627 0.239684 0.239684

-1.834 -0.621 -2.532

54 54 54

0.072 0.537 0.015

0.003985 -0.007322

0.001286 0.002699

3.099 -2.713

289 289

0.003 0.007

-0.270773

0.336649

-0.804

289

0.422

-0.098381 0.013963 0.013863 0.000525 0.005500 -0.000154 0.000004 0.008099 0.000679

0.029457 0.003039 0.002072 0.003326 0.002055 0.002626 0.000001 0.000865 0.000266 Variance Compnt. 0.24901

-3.340 4.595 6.690 0.158 2.677 -0.059 3.409 9.364 2.557

57 57 2152 2152 2152 2152 2152 2152 2152

Ȥ²

d.f.

0.002 0.000 0.000 0.875 0.008 0.954 0.001 0.000 0.011 pvalue

Std. Dev.

0.49901 0.57744 0.24613 0.02318 0.15541 0.12746 Industry Reputation Curvilinear Change Rate, U20 0.01323 Random level-1 Reliability coefficient estimate Intercept 1, P0 0.745 Time (YR_INTGR), P1 0.386 Time Squared (YRITG_SQ), P2 0.317

1812.7078 209 0.000 0.06058 379.23281 212 0.000 0.00054 339.42127 212 0.000 0.02415 75.35060 54 0.029 0.01625 87.40527 57 0.006 0.00018 92.52629 57 0.002 Random level-2 Reliability coefficient estimate Intercept 1/Intercept 2, B00 0.213 YR_INTGR/INTRCPT2, B10 0.322 YRITG_SQ/INTRCPT2, B20 0.330 0.33343

* For starting values, data from 2136 level-1 and 270 level-2 records were used * Run-time deletion reduced the number of level-1 records to 2168 Note: Chi-square statistics and Reliability estimates reported above are based on 270 of the 293 firms with sufficient data for computation. Fixed effects and variance components are based on all the data. Sigma squared = 0.24901 Standard Error of Sigma squared = 0.00919 Deviance = 4150.752713 (Number of estimated parameters = 29)

164

Summary of the model specified (in equation format): Level-1 Model Y = P0 + P1*(YR_INTGR) + P2*(YRITG_SQ) + P3*(ROA) + P4*(LAGROA_1) + P5*(LAGROA_2) + P6*(LAGROA_3) + P7*(SALES) + P8*(MKT_V3GR) + P9*(TRTN_1YR) + E Level-2 Model P0 = B00 + B01*(INDEPD_P) + B02*(NO_SHRDR) + B03*(INSTI_OW) + R0 P1 = B10 + R1 P2 = B20 + R2 P3 = B30 P4 = B40 P5 = B50 P6 = B60 P7 = B70 P8 = B80 P9 = B90 Level-3 Model B00 = G000 + G001(NEWMUN) + G002(NEWDYN) + G003(INDDISCR) + U00 B01 = G010 B02 = G020 B03 = G030 B10 = G100 + U10 B20 = G200 + U20 B30 = G300 B40 = G400 B50 = G500 B60 = G600 B70 = G700 B80 = G800 B90 = G900

Model specified for the fixed effects: Level-1 Coefficients INTRCPT1, P0

Level-2 Predictors INTRCPT2, B00

Level-3 Predictors INTRCPT3, G000 % NEWMUN, G001 % NEWDYN, G002 %INDDISCR, G003 #%INDEPD_P, B01 INTRCPT3, G010 #%NO_SHRDR, B02 INTRCPT3, G020 #%INSTI_OW, B03 INTRCPT3, G030 %YR_INTGR slope, P1 INTRCPT2, B10 INTRCPT3, G100 %YRITG_SQ slope, P2 INTRCPT2, B20 INTRCPT3, G200 #% ROA slope, P3 # INTRCPT2, B30 INTRCPT3, G300 #%LAGROA_1 slope, P4 # INTRCPT2, B40 INTRCPT3, G400 #%LAGROA_2 slope, P5 # INTRCPT2, B50 INTRCPT3, G500 #%LAGROA_3 slope, P6 # INTRCPT2, B60 INTRCPT3, G600 #% SALES slope, P7 # INTRCPT2, B70 INTRCPT3, G700 #%MKT_V3GR slope, P8 # INTRCPT2, B80 INTRCPT3, G800 #%TRTN_1YR slope, P9 # INTRCPT2, B90 INTRCPT3, G900 '#' - The residual parameter variance for the parameter has been set to zero '%' - This variable has been centered around its grand mean 165

Table 4 Reputation: Level-wise Variance Decomposition and Explanation of Variance

Random Coefficient Model

Level 1: Time Level Temporal Variation, E0 Level 2: Firm Level Firm Initial Reputation, R0 Firm reputation Linear Change Rate, R1 Firm Reputation Curvilinear Change Rate, R2 Level 3 (Industry) Level Industry Mean Reputation, U00 Industry Reputation Linear Change Rate, U10 Industry Reputation Curvilinear Change Rate, U20 Total Variation

Variance Component

% Var. Decompo sition

Variance Compone nt

% Var. Decompo sition

% Variance Explained by incl. Variables

0.42138

45.67%

0.24901

36.40%

40.91%

0.33343

0.06058 0.00054

48.74% 8.85% 0.08%

14.57%

0.02415 0.01625 0.00018 0.68414

3.53% 2.38% 0.03% 100%

0.46184 50.06%

0.04690 5.08% 0.92262

100%

Fixed Effects Model

13.48% 25.85%

166