Environmental performance and executive compensation: An integrated agency-institutional perspective by Pascual Berrone | Papers by Pascual

Academy of Management Journal 2009, Vol. 52, No. 1, 103–126. ENVIRONMENTAL PERFORMANCE AND EXECUTIVE COMPENSATION: AN INTEGRATED AGENCY-INSTITUTIONAL PERSPECTIVE PASCUAL BERRONE IESE Business School LUIS R. GOMEZ-MEJIA Arizona State University Relying on institutional theory, agency rationale, and environmental management research, we hypothesize that, in polluting industries, good environmental performance increases CEO pay; that environmental governance mechanisms strengthen this linkage; that pollution prevention strategies affect executive compensation more than end-of-pipe pollution control; and that long-term pay increases pollution prevention success. Using longitudinal data on 469 U.S. firms, we found support for three hypotheses. Contrary to our expectations, firms with an explicit environmental pay policy and an environmental committee do not reward environmental strategies more than those without such structures, suggesting that these mechanisms play a merely symbolic role. Because environmental issues are now a major social concern, companies in polluting industries face tight governmental regulations, increased media attention, and strong environmental activism. Firms respond to these external pressures by implementing strategies that help promote good environmental performance and reduce negative environmental impact (Dowling & Pfeffer, 1975; Hoffman, 2000). The resulting environmental legitimacy lowers liability exposure, enhances corporate reputation, improves access to resources, and strengthens stakeholder relations (Bansal, 2005; Hart, 1995; Russo & Fouts, 1997; Shrivastava, 1995a). Thus, it makes sense that firms should reward their executives for environmental actions that confer greater legitimacy and may directly or indirectly improve firm performance. Only recently have scholars started to analyze the relationship between environmental performance and executive pay (Coombs & Gilley, 2005; Russo & Harrison, 2005; Stanwick & Stanwick, 2001), and, contrary to the inference drawn above, some of these studies suggest that firms either penalize their managers for environmental initiatives or reward them for poor environmental performance. In this article, we offer a hybrid framework that draws on institutional theory, agency theory, and environmental management research to better explain the link between executive pay and environmental performance in polluting industries. We argue that firms that operate in environmentally sensitive sectors1 but have good environmental performance enjoy enhanced social legitimacy and organizational survival capabilities, and reward their CEOs accordingly. That is, they include environmental performance as a criterion in incentive schemes for chief executives. Our framework also suggests that firms reward pollution prevention strategies more heavily than “end-of-pipe” pollution control strategies, as the former approach confers greater legitimacy within polluting industries than the latter. We further hypothesize that corporate governance structures play an important role in this process—specifically, that the CEO pay– environmental performance relation is stronger when a firm has in place an environmental pay policy and an environmental committee within its board of directors. Lastly, we hypothesize that CEO long-term pay enhances future environmental per- The authors would like to thank the former editor-inchief, Sara Rynes, and four anonymous reviewers for their insightful comments and suggestions during the review process. The first author also thanks Andrea Giuliodori, Judith Walls, Isabel Gutierrez, Josep Tribo, ´ and Manuel Nunez Nickel for their encouragement and ´˜ advice. 103 These are firms that manufacture or process more than 25,000 pounds and/or use at least 10,000 pounds of substances from a list of chemicals deemed hazardous by the Environmental Protection Agency (EPA). These firms must report their emissions to the EPA’s Toxics Release Inventory program. 1 Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express written permission. Users may print, download or email articles for individual use only. 104 Academy of Management Journal February formance, particularly pollution prevention results, and that the positive effect is greatest where it is needed the most: where industry pollution intensity is at its highest. Our analysis of longitudinal data (1997–2003) on 469 publicly traded firms from polluting industries in the United States supports all these hypotheses except the one concerning governance structures. Our work contributes to three different branches of business research. First, we combine elements of institutional and agency theories to bring a fresh perspective into executive compensation research, which traditionally has overwhelmingly concerned financial performance (Barkema & Gomez-Mejia, 1998; Gomez-Mejia & Wiseman, 1997). Second, we add to environmental research by recognizing the importance of executive compensation in promoting good environmental behavior, and we reinforce the consensus in the field that pollution prevention strategies tend to be valued more (as reflected in higher compensation) than end-of-pipe pollution control strategies. Lastly, we contribute to the corporate governance literature by analyzing whether specialized environmental board committees and CEO pay policies serve to monitor, guide, and reward environmental actions and thus affect the relation between environmental performance and CEO compensation. THEORETICAL FRAMEWORK AND HYPOTHESES The main thesis of institutional theory is that organizations enhance or protect their legitimacy (Scott, 1995) by conforming to the expectations of institutions and stakeholders (Aldrich & Fiol, 1994; DiMaggio & Powell, 1983). By adhering to institutional prescriptions, firms reflect an alignment of corporate and societal values (Meyer & Rowan, 1977). Thus, concern over legitimacy forces firms to adopt managerial practices that are expected to have social value (Deephouse, 1999; Scott, 1995). Legitimate actions are defined within a firm’s institutional field, which transcends the industry in which the organization directly competes, and which establishes idiosyncratic rules, belief systems, and practices deemed to be legitimate (DiMaggio & Powell, 1991; Scott, 2005). Firms in polluting industries are all subject to the same regulatory framework and arguably face similar media attention, scrutiny from activists, community concerns, and changes in consumer preferences; the institutional theory prediction is that companies in this strong institutional field will gain legitimacy by exhibiting good environmental performance (Bansal, 2005; Bansal & Clelland, 2004; Hoffman, 2001). Environmental legitimacy brings several advantages. For instance, legitimate companies have better exchange conditions with partners and better access to resources (DiMaggio & Powell, 1983), which the firms can allocate to improve environmental performance even further by, for example, hiring experts or reacting rapidly in the case of an environmental mishap (Bansal & Clelland, 2004). Legitimacy also enables firms to innovate with less risk of loss (Sherer & Lee, 2002); thus, good environmental performers can take better advantage of new market opportunities created by the increased demand for green products and services. Moreover, environmentally legitimate firms run less risk of environmental mishap and thus of legal sanctions, costly penalties, high insurance premiums, and significant environmental remediation costs (Godfrey, 2005; Khanna & Damon, 1999; Sharma & Vredenburg, 1998; Shrivastava, 1995b). Limiting impact on the natural environment also isolates a firm from stakeholder scrutiny and reduces the risk of social sanctions (Oliver, 1991), including negative press and boycotts by environmental activists. Environmentally legitimate firms can attract and retain better partners, customers, and employees than poor performers (Buysee & Verbeke, 2003; Henriques & Sadorsky, 1999; Sharma & Henriques, 2005; Turban & Greening, 1997), and thus have less employee turnover and fewer unproductive associations. Additionally, firms that have implemented environmental due diligence may incur less risk of community opposition to corporate actions such as the construction of new plants. Lastly, environmental legitimacy reduces idiosyncratic firm risk. Bansal and Clelland (2004) showed that environmentally legitimate firms incur lower unsystematic stock market risk than less legitimate firms, so they have a lower cost of capital. In sum, firms are likely to recognize the value of conformity to environmental expectations, as the resultant legitimacy reduces the probability of organizational failure (Scott, 1995; Singh, Tucker, & House, 1984) and may enhance financial performance (King & Lenox, 2002; Klassen & McLaughlin, 1996). Hence, firms in polluting sectors should motivate their CEOs to engage in strategies to improve environmental performance. From the perspective of managers, however, the link between environmental actions and financial performance is not straightforward (Bansal, 2005; Sharma, 2000). For instance, to reduce or eliminate pollution emissions, executives may need to implement technologies that may fail or may cause quality problems or unforeseen costs, all issues for 2009 Berrone and Gomez-Mejia 105 which managers are likely to be held responsible (Klassen & Whybark, 1999; Russo & Fouts, 1997). Pollution reduction strategies are also hard to put into practice, since they require production redesign, new equipment, and cross-functional employee coordination. Moreover, good environmental performance may take time to come to fruition, increasing uncertainty about outcomes (Aragon-Correa, 1998; AragonCorrea & Sharma, 2003; Hart, 1995; Khanna & Damon, 1999). Managers may therefore avoid environmental strategies and allocate resources to more conservative investments. Even when managers recognize the importance of good environmental performance for their firm and its stakeholders, they may be tempted to focus on the actions that are easiest to observe, “as a way to ‘provide cover’ for poor emissions performance by appearing to take steps in the right direction” (Russo & Harrison, 2005: 588). Thus, to motivate CEOs to enhance environmental performance, an incentive system should reward evidence of environmental strategies with potentially high, though uncertain, benefits. Integrating Institutional and Agency Perspectives: Environmental Performance and CEO Compensation Institutional theory suggests both that firms benefit from conforming to societal expectations and that managers have the capacity (“internal power”; Oliver [1991]) and the motivation (“fear of novelty”; Oliver [1997]) to resist these institutional pressures to the extent that there is ambiguity in financial returns. Firms therefore need incentive mechanisms to dissuade managers from avoidance. CEOs who exhibit good environmental performance should be rewarded with higher pay because on the one hand they are enhancing their firms’ chances of survival, and on the other hand higher pay will make them less reluctant to engage in environmental actions with uncertain economic benefits. Use of environmental criteria in executive pay schemes is consistent with findings by institutionalist scholars (Peng, 2004; Staw & Epstein, 2000) that investors, boards of directors, and their compensation committees do use evidence of managerial actions believed to procure legitimacy to assess the effort and value of their top executives, and not just observed economic performance, particularly if the link between actions and results is blurred. “When there is so much ambiguity in attributing the causes of organizational outcomes such as performance, outside observers often rely on positively valued behavior as a signal in making their judgment of a firm’s management” (Peng, 2004: 458). The institutional prediction discussed above is consistent with an agency perspective. Agency scholars argue that to deal with the classical agency problems of moral hazard (unobserved actions by an agent that may be damaging to the principal’s interests) and adverse selection (hidden information), principals may acquire information about the agent’s performance and design a compensation system based on that information (Eisenhardt, 1989; Fama, 1980; Jensen & Meckling, 1976; Lambert, Larcker, & Weigelt, 1993). Three interrelated factors should be considered in designing such systems. The first is “informativeness” (Holmstrom, 1979), or the extent to which the performance measure actually reflects the agent’s contribution to the principal’s welfare. By tying CEO pay to environmental performance, the principal, through the board of directors, can use that indicator to gauge whether or nor the CEO is acting in the best interests of shareholders. Given their complexity, environmental strategies are likely to be beyond the expertise of board members, and they may use pay tied to environmental performance as a substitute for close vigilance over difficult-to-monitor actions. The second factor is “risk bearing” (Baker, 1990; Bloom & Milkovich, 1998; Gomez-Mejia, TakacsHaynes, Nunez-Nickel, Jacobson, & Moyano, 2007; Gray & Cannella, 1997; Miller, Wiseman, & GomezMejia, 2002), or the extent to which an agent may incur potential losses in pursuit of performance targets (such as lower reputation or high employment risk [Larraza-Kintana, Wiseman, Gomez-Mejia, & Welbourne, 2007]). As agent risk bearing increases, compensation should increase accordingly (Bloom & Milkovich, 1998). Environmental investments are risky, since “there is little reason to believe that this investment will result in enhanced short-term profits” (Hart, 1995: 998). Indeed, although there is empirical evidence that good environmental performance improves long-term economic performance (e.g., King & Lenox, 2002; Klassen & McLaughlin, 1996), studies have also shown that environmental performance can impair financial results, especially in the short term (Hart & Ahuja, 1996; Sarkis & Cordeiro, 2001). If CEOs are not compensated for the increased risk associated with environmental investments, they will presumably allocate capital into less uncertain alternatives, maintaining their firms’ current levels of environmental impact and possibly impairing the firms’ legitimacy. Also, a good CEO may leave for a job in a sector less environmentally sensitive, where the additional effort and risk are not demanded. The third factor is “controllability” (Antle & 106 Academy of Management Journal February Demski, 1988; Demski, 1994; Demski & Feltham, 1978), or the extent to which an agent can exert some influence over a performance criterion. This means that an “agent should be evaluated and rewarded by a performance measure, if he or she can control or significantly influence that measure” (Franco, 2007: 51). Because polluting industries are considered turbulent sectors (Klassen & Whybark, 1999) in which firms face high governmental scrutiny, are subject to media attention, and are recurrent targets of environmental activism, financial performance may experience greater variance than in other sectors, and executives may have less influence over it. In this case, incentives should be closely tied to managerial effort and strategic actions (Balkin, Markman, & Gomez-Mejia, 2000). Because environmental strategies are decided by managers (Aragon-Correa, Matıas-Reche, & Senise-Barrio, ´ 2003; Sharma, 2000) and are likely to benefit an organization (as we argued earlier), environmental performance may be a reasonable pay criterion.2 We therefore propose the following hypothesis: Hypothesis 1. Environmental performance has a positive effect on CEO total pay. The Impacts of Different Environmental Strategies on CEO Compensation As we noted earlier, the environmental management literature has identified two general types of environmental strategies: pollution control and pollution prevention (Christmann, 2000; Hart, 1995; Klassen & Whybark, 1999; Russo & Fouts, 1997; Sarkis & Cordeiro, 2001). End-of-pipe technologies capture, treat, and dispose of pollutants and waste at the end of a manufacturing process. Focusing on compliance, hazard control, and remediation, end-of-pipe technologies involve the addition of equipment and devices as the last stage of production processes and thus do not “require the firm to develop expertise or skills in managing new environmental technologies” (Russo & Fouts, 1997: 538). In contrast, pollution prevention strategies minimize or eliminate the creation of toxic chemical agents during the various stages of production and thus require structural investments in new, cleaner technologies (Klassen & Whybark, 1999; Russo & Fouts, 1997). At the same time, research has shown that pollution prevention efforts provide organizations with unique advantages (Christmann, 2000; Hart, 1995; Klassen & Whybark, 1999; Russo & Fouts, 1997) and may even increase manufacturing performance because they require a fundamental rethinking of products and processes that can create opportunities for improvements and innovation. Klassen and Whybark (1999) found empirical evidence indicating that pollution prevention strategies have a positive impact on manufacturing performance, while the opposite was true for end-of-pipe efforts. Pollution prevention strategies can also reduce costs through better use of inputs, reduction of waste disposal costs, and removal of unnecessary steps in production processes. And given that pollution prevention strategies reduce and eliminate waste generation, they can potentially cut emissions below required levels and thus reduce compliance costs and legal liabilities. Empirically, Christmann (2000) found that, in the presence of complementary assets, pollution prevention technologies create a cost advantage. Under the rubric of the resource-based view of the firm, various authors (Aragon-Correa et al., 2003; Christmann, 2000; Hart, 1995; Klassen & Whybark, 1999; Russo & Fouts, 1997) have argued that because pollution prevention is specific to particular production processes and requires people-intensive structures, it is not easily imitable by competitors and thus represents a distinct source of competitive advantage. End-of-pipe solutions, on the other hand, are relatively inexpensive off-the-shelf technologies that can be obtained in the open market, and thus competitors can readily copy and implement them. We do not consider the impact of environmental performance on each pay form singly. Rather, we analyze the influence of environmental performance on CEO total compensation, which includes salary, annual bonuses, and long-term pay income. This decision follows the agency rationale that in its total CEO compensation package, a firm will most likely try to reward its CEO for good past behaviors in order to reinforce those actions and assure their continuance in the future (Gomez-Mejia & Wiseman, 1997). That is, the firm looks back to assess the CEO’s decisions and rewards those actions believed to add value to the firm in order to align interests and create a “common fate” between CEO and firm. Our analysis is also consistent with the institutional rationale that boards attempt to justify compensation levels (including stock options awards) on the basis of past executive decisions deemed legitimate. In our case, we expect a board to consider past environmental activities among the criteria for assessing a CEO’s contributions to their firm and/or as a criterion to justify pay levels. This environmental performance-reward linkage, in turn, creates an expectation that in future the CEO will receive greater compensation (including all forms of pay) if environmental performance remains acceptable or improves. That is, rewarding past performance (for instance, through larger stock option awards) and hoping for better future performance (for instance, through stock appreciation) are not incompatible objectives but actually complementary. We explicitly recognize the incentive role of long-term pay in Hypotheses 4a and 4b. 2 2009 Berrone and Gomez-Mejia 107 In sum, there is a fair degree of consensus among environmental scholars that pollution prevention strategies are more valuable than end-of-pipe solutions. To the extent that pollution prevention efforts provide greater benefits than end-of-pipe technologies, agency theory (via the informativeness principle discussed earlier) suggests that pollution prevention success should be compensated more than end-of-pipe pollution control. At the same time, pollution prevention strategies are more complex and risky than end-of-pipe strategies. They are technologically complex because they require changes in systems, processes, and products (Aragon-Correa & Sharma, 2003); socially complex because they involve diverse stakeholders at different levels (Russo & Fouts, 1997); and structurally complex because they require managerial commitment and cross-functional coordination (Aragon-Correa, 1998). Moreover, their systemic approach requires risky investments in low-impact technologies, product innovation, and source reduction processes. Thus, from an agency perspective principals should be more inclined to reward pollution prevention rather than end-of-pipe results. The same proposition holds from an institutional perspective. It has been argued that “the appearance rather than the fact of conformity is often presumed to be sufficient for the attainment of legitimacy” (Oliver, 1991: 155). This suggests that CEOs may try to secure compliance and meet minimal standards of environmental performance through end-of-pipe strategies, which may be more visible than pollution prevention, in order to manage impressions and fulfill their obligations to external constituencies. However, some authors (Ashforth & Gibbs, 1990; Staw & Epstein, 2000; Suchman, 1995) have argued that exceeding minimum requirements confers greater legitimacy, so that “once minimal standards are met, corporations are likely to continue working . . . to be the best or the most admired” (Staw & Epstein, 2000: 526), and that firms’ constituents prefer more definitive responses (Suchman, 1995). According to this logic, CEOs who are committed to environmental excellence (through pollution prevention strategies) should receive higher pay than those who merely meet minimum requirements. Moreover, achieving legitimacy with more substantive, though less visible, strategies (such as pollution prevention) may be easier in a strong institutional field, where objective measures are made public and institutional pressures are steady. Thus, Hypothesis 2. Evidence of pollution prevention strategies has a greater impact on CEO total pay than evidence of end-of-pipe pollution control strategies. The Moderating Role of Environmental Governance One way to reward top management for desirable behaviors is through a compensation policy that permits considering the value of strategic actions, not just financial results (what Baysinger and Hoskisson [1990] referred to as “strategic controls”). Institutional pressures are likely to influence the presence of such a compensation policy (GomezMejia & Wiseman, 2007; Lubatkin, Lane, Collin, & Very, 2007). Boards may consider environmental performance implicitly or explicitly when designing an executive compensation policy; by making an environmental pay policy explicit, a firm assumes a public commitment and clearly signals its beliefs (Peng, 2004). Gross deviations from that policy would be perceived as hypocritical and thus would be likely to impair legitimacy. However, an explicit environmental pay policy may not be enough to guarantee that a CEO will value environmental performance. Implementing strategic controls also requires additional information—much more than is needed for implementing financial controls (Baysinger & Hoskisson, 1990). Incentives based on executives’ behavior depend on a board’s knowledge of that behavior (see also Boyd, 1994; Dalton, Daily, Ellstrand, & Johnson, 1998; Fama & Jensen, 1983; Jensen & Meckling, 1976; Lorsch & MacIver, 1989; Mizruchi, 1983). Conyon and Peck (1998) argued that a compensation committee within a board of directors is an important tool for evaluating CEO performance and designing appropriate rewards for top executives. Baysinger and Hoskisson (1990) also argued that board composition may influence the assessment of the strategic value of executive decisions. And although forming committees related to certain social issues may be a response to institutional pressures (Luoma & Goodstein, 1999), these committees may also improve firm performance (both socially and financially) (Greening & Gray, 1994). It is reasonable to expect that when environmental oversight responsibilities are explicitly and formally delegated to a subgroup of a board (that is, an environmental committee),3 the board is in a better position to assess executive performance on the environmental dimension (for instance, tracking relevant pollution data and judging the extent to which executive choices reduce pollution) and to 3 Each firm has its own name for this committee, e.g., “safety, health and environmental affairs” or “environmental compliance.” For simplicity, we just call it an “environmental committee.” In all cases, we are referring to committees within boards of directors. 108 Academy of Management Journal February consider this assessment in CEO pay decisions. In the parlance of agency theory, delegating environmental issues to a committee made up of knowledgeable board members should reduce the information asymmetries between principal and agent, allowing for a more accurate assessment of the executive’s environmental performance and a tighter linkage between that performance and total pay. We therefore expect firms that have a CEO environmental compensation policy and a specialized environmental board committee to tie executive pay more strongly to environmental performance. And given the arguments underlying Hypothesis 2, we expect this moderating effect to be stronger for evidence of pollution prevention strategies than for evidence of end-of-pipe pollution control. Thus, Hypothesis 3a. The positive effect of environmental performance on CEOs’ pay is higher for firms that have environmental governance mechanisms in place. Hypothesis 3b. Firms with environmental governance mechanisms in place put greater emphasis on pollution prevention than on end-ofpipe pollution control as a criterion for CEO pay. The Influence of Long-Term Pay on Subsequent Environmental Performance Long-term pay forms, like stock options, are explicit incentives, since their final value is contingent on future performance (Murphy, 1999); they are intended to align the interests of managers and shareholders (Fama & Jensen, 1983; Jensen & Meckling, 1976; Jensen & Murphy, 1990) and induce managers to actively seek business opportunities and make strategic investments. In polluting industries, managers who receive long-term pay contingent on the future value of their companies are likely to perceive the potential value of green practices more easily, because good environmental behaviors are widely believed to have an enduring impact on performance (Hart, 1995; Russo & Fouts, 1997). Klassen and McLaughlin (1996) used event study methodology to gauge investor reactions to news about environmental performance awards and environmental crises. These authors found significant, positive returns for firms with strong environmental management and significant, negative returns for firms with weak environmental management. Dowell, Hart, and Yeung (2000) also found that large companies that adopt strict global environmental standards are rewarded with higher stock market performance. More recently, King and Lenox (2002), using a sample of 614 publicly traded U.S. manufacturing firms, found strong evidence that a waste prevention strategy leads to financial gains measured with Tobin’s Q. Thus, to the extent that stocks appreciate if firms avoid actions with negative environmental impacts, and/or the executives believe this is the case, CEOs should make decisions that reduce pollution. In addition, executives have an incentive to reduce pollution in the future to the extent that new stock options may be awarded if environmental performance improves. Furthermore, Sanders and Hambrick (2007) reported that stock option pay was positively associated with greater levels of investment in risky long-term projects such as R&D, capital equipment purchases, and acquisitions. Since good environmental performance, particularly pollution prevention, requires a multiyear commitment to demanding and risky environmental strategies, to the extent that long-term pay reinforces those types of behaviors, it should improve environmental performance. Given that pollution prevention is more valuable to firms than end-of-pipe and should be rewarded accordingly (as per Hypothesis 2), we would expect that executives receiving long-term income would tend to devote more attention to improving pollution prevention results than to improving end-of-pipe results. Hypothesis 4a. Long-term pay has a positive effect on subsequent environmental performance. This positive effect is greater on pollution prevention than end-of-pipe results. The more polluting a firm’s industry is, the more likely it is that environmental initiatives will bring strategic advantages. Other things being equal, the value of long-term pay for the executive should increase in tandem with those environmental efforts. That is, when executives operate in highly polluting industries, then long-term incentives are more likely to have a beneficial effect on their firms’ environmental performance. And, given our previous arguments, CEOs should place greater emphasis on obtaining high pollution prevention marks than on end-of-pipe success. Thus, Hypothesis 4b. Long-term pay has a greater effect on subsequent environmental performance as a firm’s presence in polluting industries increases. This positive moderating effect is greater for pollution prevention than end-ofpipe results. 2009 Berrone and Gomez-Mejia 109 METHODS Sample and Data Collection Data on an institutional field should represent firms facing similar institutional pressures (Hoffman, 2001). Given that government is one of the most prominent sources of such pressures, we chose to focus on firms from industries subject to reporting under the EPA’s Toxics Release Inventory, a program that requires facilities exceeding a threshold level to report their emissions. These firms are all subject to the same regulatory framework and arguably face similar media attention, scrutiny from activists, community concerns, and changes in consumer preferences. We collected data on these firms for CEO pay, CEO tenure, and CEO duality (in which the CEO doubles as the board chair) from the ExecuComp database, a product of Standard & Poor’s that includes several CEO compensation items for companies included in the S&P 1500 index and that is extensively used in compensation research. We initially identified 1,071 companies with data for the period 1997– 2003. After subtracting companies with missing values for some of these variables, cross-referencing to the Compustat database for information on firm size, financial performance, and industry, and matching with the Investor Responsibility Research Center (IRRC) database for information on board structure, family firm status, and CEO and directors’ ownership, we were left with 823 companies. We then gathered environmental governance information from proxy statements reported to the Securities and Exchange Commission (SEC). As noted later, we used firm fixed-effects estimation models, which require mean-deviating the sample (i.e., adjusting each observation by its within-firm mean). Thus, we needed data on each firm for at least two years; otherwise, units would be dropped out of the fixed-effects estimator (Wooldridge, 2002). This requirement reduced the sample to 762 firms. Furthermore, only firms that manufacture or process more than 25,000 pounds and use at least 10,000 pounds of any of the EPA’s listed chemicals are required to report their emissions to the TRI program. In the end, we found environmental data for at least two years for 469 publicly traded companies, which constituted our final unbalanced panel and represented 2,088 firm-year observations.4 Ap- pendix A provides a detailed description of the number of firms, years, and observations. Measures Compensation measures. Our dependent variable for Hypotheses 1–3, CEO total pay, consisted of the sum of salary, bonus, and all long-term components of CEO pay (stock options, restricted stock, and other long-term compensation). Total longterm pay was used as an independent variable in testing Hypotheses 4a and 4b. In all cases, we used the natural logarithm of compensation to reduce heteroscedasticity. In keeping with much of the executive compensation literature, we valued stock options at the time they were awarded. To value stock options, two general methods are available. One is the Black-Scholes method, which uses a sophisticated model to estimate the value of this type of compensation. Many studies of executive compensation (Balkin et al., 2000; Finkelstein & Boyd, 1998; Gomez-Mejia, Larraza-Kintana, & Makri, 2003; Henderson & Fredrickson, 1996, 2001) have used a second method, proposed by Lambert and colleagues (1993), that values stock options at 25 percent of their exercise value. We selected the latter method over the former for several reasons. First, in some cases, Black-Scholes values were not reported in ExecuComp even when stock options were granted. Second, in our sample the correlation between the values yielded by the two methods was very high (.97), as it has also been in previous research (Finkelstein & Boyd, 1998; Lambert et al., 1993). Lastly, and most important, separate analyses using Black-Scholes values yielded results almost identical to those reported here. Environmental performance. Use of TRI data is well established in management research on environmental impacts of corporations (e.g., King & Lenox, 2002; Klassen & Whybark, 1999; Russo & Harrison, 2005). Under EPA’s Emergency-Right-toKnow Provision, industrial facilities with ten or more full-time employees that release any listed toxic substance in excess of the minimum reporting threshold via any of four different media (air, water, land, or underground injection) are required to sample (92%) belonged to either the high or medium group. We followed a similar comparison procedure using the classification made by Smarzynska and Wei (2001), who used a combination of abatement costs and emission data to classify industries at the three-digit SIC code level. Approximately 80 percent of our firms then belonged to the high- and medium-polluting sectors. These results suggest that our sample was representative of polluting industries in the United States. 4 To analyze how representative of polluting industries our sample was, we ranked industries defined at the two-digit SIC code level on total emissions and divided them into three groups (high-, medium-, and low-polluting). Four hundred and forty-nine of the 469 firms in our 110 Academy of Management Journal February report the type and amount of emissions to the EPA (EPA, 2002). Early studies simply used the sum of annual emissions of all TRI substances released in a given year as a proxy for a firm’s potential harm to human health or the environment (e.g., Khanna & Damon, 1999). However, total quantity of emissions is not the same as exposure to harm, for either the public or the ecosystem (EPA, 2002). Recently, some studies have tried to mitigate this shortcoming by weighting emissions, using “reportable quantities” (RQs) to account for the toxicity levels of chemical agents (e.g., King & Lenox, 2000, 2002; Russo & Harrison, 2005). But Toffel and Marshall (2004) argued that RQs are problematic for several reasons. First, the discrete scale of RQs reduces their precision. Second, there is only one RQ value for each chemical agent, regardless of the medium in which it is released (air, water, land). Third, it is hard to determine the bases on which the toxicity level for a particular chemical was established (Toffel & Marshall, 2004). To address these drawbacks, we weighted chemicals using the “human toxicity potential factor” (HTP) developed by Hertwich and colleagues (Hertwich, Mateles, Pease, & McKone, 2001), which measures toxicity in terms of benzene equivalence (for carcinogens) or toluene equivalence (for noncarcinogens). This method assigns each chemical separate HTP values for different media of release. The results are more closely associated with actual risks to human health. The HTP procedure is more precise than that for RQs, and HTP results are highly correlated with those obtained with more sophisticated weighting methods such as the EPA’s “risk-screening environmental indicator” (r .73) and the “ecoindicator99” (r .92) (Toffel & Marshall, 2004). Thus, before calculating our environmental measures, we weighted the quantity of each chemical generated over a given year by its correspondent HTP value. (See Appendix B, part 1, for the formal calculation procedure.) In accord with previous research (King & Lenox, 2000, 2002, 2004), we considered only chemicals that were consistently reported on over our period of analysis and were included in the HTP list. The HTP covers 268 chemicals that EPA has consistently reported on and that represented 79 percent of the TRI releases to air reported in 1997 (Hertwich et al., 2001). Of our two environmental performance measures, the first, pollution prevention strategies, has been traditionally measured as the difference between a predicted value and some actual pollution level (King & Lenox, 2000, 2002).5 Accordingly, to create our pollution prevention measure, we estimated total waste generation levels and later contrasted this estimate with real values. Given that facilities must report their production ratios for a current reporting year as compared to the previous reporting year (i.e., the ratio of the production volume in t 1 to the production volume in t), we used these values to estimate waste generation following several steps.6 First, we weighted each chemical by its HTP value (as described in Appendix B, part 1); second, we aggregated the results across chemicals at the facility level (see Equation B1); third, we multiplied these results by their corresponding production ratios; fourth, we aggregated results by parent company; and finally, we compared these results against real values. The formal procedure for calculation can be found in Appendix B, part 2. Because the HTP method offers cancer and noncancer values, we calculated the formulas given in Appendix B, part 2, using these values separately and obtained two different pollution prevention indexes. Given that these variables 6.88e 10; s.d. were highly skewed (mean 1.69e 12; 1 31.02 for carcinogen values, and mean 2.22e 13; s.d. 7.01e 14; 1 38.47 for noncarcinogen values), we log-transformed them to achieve normality. Later, we calculated their reliability and, given their high Cronbach alpha value ( .96), standardized and averaged both measures to create our final pollution prevention variable. Our second environmental performance measure was end-of-pipe pollution control. Following previous environmental research (e.g., King & Lenox, 2001, 2002, 2004; Sarkis & Cordeiro, 2001), we defined this variable as a ratio in which the numerator was the sum of chemicals recycled, treated on-site, and transferred to other locations for further treatment, and the denominator was the total waste generated by a firm. This calculation is consistent with the definition given by the EPA (1997), which defines end-of-pipe technologies as methods used to remove already formed pollutants from a stream of air, water, waste, product, or similar medium. We weighted chemical figures by the average HTP values before calculating the ratio. Again, we computed the reliability between cancer and noncancer figures. Because the Cronbach alpha coefficient was fairly high ( .79), we combined those two figures into a single end-of-pipe composite by standardizing and averaging them. We were not able to replicate King and Lenox’s measure because facility-level data were not available. 5 Production ratio values often vary around 1. For instance, a ratio of 1.1 would indicate a 10 percent increase in production. 6 2009 Berrone and Gomez-Mejia 111 Environmental governance. To define this variable (see Hypotheses 3a and 3b), we analyzed the proxy statements each company issued over our time period to identify whether each firm had an environmental pay policy and an environmental committee within its board. To determine the existence of an explicit environmental pay policy, we followed a three-step process. First, we defined a list of keywords representative of both pay and environmental issues. Our environmental word list included the keywords used in the work of Bansal (2005)—“sustainable development,” “environment,” “pollution,” and “toxic”—in addition to keywords from other papers cited in this article: “hazardous,” “waste,” “disposal,” “alternative energy,” “ecology,” and “contamination.” The pay word list included the terms “pay,” “compensation,” “salary,” “wage,” “reward,” “remuneration,” “incentives,” “bonus,” “stock,” and “income.” As a second step, we identified all paragraphs within the proxy statements that contained any word(s) from the environmental list plus any word(s) from the pay list. In all, 100 dyadic combinations were possible. Our large amount of data (5,053 text files, over 1.5 gigabytes) made this task more difficult and tedious than we originally anticipated it would be. We used the Find in Context, from iNetPrivacy Software (2004) and programmed it to search text blocks that contained combinations of words in both lists within a range of 500 characters. If a positive match was found, the software would extract a 5,000-character text block for further analysis. To maximize results and gain precision, we included variations of a stem (e.g., “environment,” “environmental,” “environmentally”). In the third step, we visually inspected each text block and coded either 1, for an explicit relationship between executive pay and environmental performance, or 0, for the absence of such a link. We then created a dummy variable with a value of 1 if a firm’s proxy statement contained at least one text block coded 1 and 0 otherwise. To identify the presence of an environmental committee, we followed an identical procedure. In this case, dyadic relationships were between the items in the environmental word list and the word “committee.” Paragraphs were individually inspected to determine whether or not a company had a committee responsible for environmental issues. We then created a dummy variable coded 1 if such a committee was present and 0 otherwise. We compiled our final environmental governance measure by adding the two dummy variables for environmental pay policy and environmental committee (so effectively this composite had a 0 –2 range for each firm, with a 2 indicating the presence of both an environmental pay policy and an environmental committee). All reports were coded by one of the authors. To check for potential bias derived from a single rater and to test for reliability, we randomly selected a sample of 300 paragraphs for coding by an independent researcher with knowledge in the area of executive compensation. We then compared this researcher’s codes with those of the original rater and calculated interrater reliability using Cohen’s kappa, obtaining highly satisfactory results: 0.90 for the case of environmental pay policy and 0.92 for environmental committee. Hence, we deemed it safe to use only the codes from the primary coder, which allowed us to code the entire sample consistently within resource constraints (Bansal, 2005). Control variables. We used a comprehensive set of control variables. The first two are the most widely recognized determinants of CEO pay, namely, firm size and firm performance (Tosi, Werner, Katz, & Gomez-Mejia, 2000). We captured firm size as the logarithm of a firm’s total assets (Bloom & Milkovich, 1998; Finkelstein & Boyd, 1998). We measured financial performance as a firm’s annual return on equity (ROE) (Finkelstein & Boyd, 1998; Sanders & Carpenter, 1998). We also measured market-based performance using Tobin’s Q, calculated by dividing the sum of firm equity value, book value of long-term debt, and current liabilities by total assets (Chung & Pruitt, 1994; Makri, Lane, & Gomez-Mejia, 2006). In keeping with relevant work on executive compensation, we controlled for two measures of CEO power and influence. The first was CEO ownership, or the percentage of ownership stake its CEO had in a firm, which some researchers have used as a proxy for CEO power (Finkelstein, 1992; McEachern, 1975). The second was CEO tenure, or the number of years an incumbent CEO had worked for the current firm as CEO, which some scholars have interpreted as a proxy for CEO entrenchment (Gomez-Mejia, Nunez-Nickel, & Gutierrez, 2001; Hill & ˜ Phan, 1991). Another set of four control variables accounted for governance structure. First, we controlled for the proportion of outside directors, measured as the ratio of outside board members to the total number of board members. This measure is often used as an indicator of board independence (Baysinger, Kosnik, & Turk, 1991; Westphal & Zajac, 1994). Second, we considered director ownership, a proxy for board and firm having a “common fate” (Boyd, 1994). This was measured as the percent ownership stake of directors, excluding the CEO’s share. 112 Academy of Management Journal February Third, we included CEO duality, the situation in which the same individual is both CEO and board chair, which purportedly reduces board independence (Boyd, 1994; Westphal & Zajac, 1994). This was measured as a dummy variable coded 1 if a CEO held both positions and 0 otherwise. Lastly, we controlled for family firm status, as this variable has been found to influence CEO pay (Gomez-Mejia et al., 2003; McConaughy, 2000). This dummy variable assumed a value of 1 if at least one board member of a firm had a family relationship with the CEO and was 0 otherwise. To control for the possibility that firms may improve their environmental performance simply by divesting from highly polluting sectors, we created an industry pollution position index that reflects the mix of sectors in which a firm operates and accounts for their pollution intensity. The procedure to calculate this index appears in Appendix B, part 3. We expected this index to be inversely related to our environmental measures. That is, firms with greater presences in highly polluting sectors should exhibit worse environmental scores. Lastly, we controlled for the number of reporting plants owned by a firm, as this might affect overall end-of-pipe and pollution prevention scores; this number was gathered from the TRI database.7 Hypotheses 4a and 4b have environmental performance as their dependent variable, so we included an additional set of control variables that might influence environmental performance; we also lagged all control variables for Hypotheses 4a and 4b by one year. The additional controls are as follows: Age of assets was approximated by the ratio between gross and net assets (Shameek & Cohen, 2000). Arguably, firms with older plants tend to pollute more and consequently exhibit worse environmental performance. Regulatory intensity was expected to positively influence environmental performance. Following Kassinis and Vafeas (2002), we estimated this variable as the total emissions of the state in which a firm had its headquarters, deflated by total employment in each state (from the U.S. Census Bureau), log-transformed and inverted. Estimation Methods Because our data set has observations of multiple firms over different points in time, the use of time series cross-sectional data analysis techniques was appropriate to test our hypotheses (Greene, 1993; Wooldridge, 2002). We used a fixed-effects model with White’s correction, which solves some heteroskedasticity problems (White, 1980). A fixedeffects model is equivalent to adding a dummy variable for each firm (Greene, 1993); it has the advantage of explicitly modeling features that are unobservable but stable over time and their possible correlation with explanatory variables. Fixedeffects models incorporate computational routines that effectively handle many potential statistical problems such as multicollinearity and protect against spurious results from the problematic error terms typically found in traditional crosssectional studies (Hsiao, 1985). They are considered conservative because only changes in independent variables within a firm can produce significant effects. We used a one-year-lag model because most of the firms based their incentives on annual results. In all cases, we used standardized variables. We included year dummy variables to eliminate year-specific heterogeneity. Thus, our models control for both firm-specific and time-specific effects. We applied several statistical methods to assure the robustness of our sample. First, given the unbalanced nature of our sample and the use of fixedeffects models, sample bias was a concern. To address this issue, we followed the method proposed by Wooldridge (2002), who argued that in a fixedeffects model with an unbalanced panel, sample bias is not a problem unless the selection is correlated to the idiosyncratic error term of the model. To test this assumption, we added a selection indicator with a one-period lag. This indicator reveals which time periods are missing for each firm. For each period, this variable assumes the value 1 if a firm is included in the estimation and 0 otherwise. Thus, the selection indicator models the presence or absence of firms in each time period. In our estimations, this selection indicator was not signif- To estimate the percentage of plants sold each year, we listed the facilities that exceeded the high-pollution threshold established by calculating the average emission level of all facilities and adding to it one standard deviation. Then we calculated the percentage of those facilities that remained in a firm over time. From year to year, an average of approximately 85 percent of the facilities remained under the control of the same firms (which means that from year to year, an average of 15 percent were either closed or sold). For the entire period of analysis (seven years), about 60 percent of the facilities classified as high emitters remained in the sample. This does not mean that high polluters sell plants at a faster rate than low polluters. To test that possibility, we conducted an unreported analysis comparing the selling rates of high and low polluters and found no statistically significant differences between these two groups (see the Discussion for more on this issue). 7 2009 Berrone and Gomez-Mejia 113 icant,8 indicating that imbalance in the sample did not lead to bias (Wooldridge, 2002). A second concern was the potential selection bias from omitting companies that pollute below the TRI threshold. To address this concern, we considered the larger sample (762 firms) and followed the two-stage procedure suggested by Heckman (1979). Results from this estimation showed no evidence of sample selection bias, and thus no corrections were needed in the models.9 When testing Hypotheses 4a and 4b (the impact of long-term pay on environmental performance), we calculated two models: one with our pollution prevention measure as dependent variable and the second with our end-of-pipe measure as dependent variable. When testing Hypothesis 4b (the impact of long-term pay on environmental performance as the industry polluting position increases), we included an interaction term that was the cross-product between industry pollution position and our environmental performance measures. For testing Hypotheses 4a and 4b, we adjusted a fixed-effects model as described above. As an additional analysis and in order to check again for potential sample selection bias, we used Tobit analysis, since these models have the advantage of using data from the nonreporting firms. To conduct this analysis, we considered our full sample of 762 firms (both reporting and nonreporting). Thus, we had to account for the fact that, for nonreporting firms, environmental values were missing. Following previous environmental research (Klassen & Whybark, 1999; Russo & Harrison, 2005), we decided to use the highest value of environmental performance for nonreporting firms. As a result, our dependent variables in this case (pollution prevention and end-of-pipe) were continuous but censored at the maximum values. Following Russo and Harrison (2005), we used a censored Tobit model, which applies when a dependent variable is continuous but bounded (Wooldridge, 2002). Fixedeffects Tobit models can be calculated, but estimates are inconsistent (Heckman & MacCurdy, 1980). Thus, we calculated a random-effects Tobit model, which is more adequate since it does give consistent estimates for betas (Robinson, 1982). RESULTS Table 1 reports descriptive statistics and correlations for the variables used in this study. Untransformed values for CEO total pay averaged $5,774,000, and long-term pay averaged $4,445,000. The highest correlation in Table 1 is between these two variables, indicating a strong linear relationship. Although this correlation value is clearly high, it is at levels obtained in other studies analyzing executive pay (Coombs & Gilley, 2005; Henderson & Fredrickson, 1996). Table 2 reports the results of the fixed-effects models used to test Hypotheses 1 and 2, regarding the influence of environmental performance on CEO pay and the importance of different environmental strategies for CEO pay, respectively. Model A presents our control variable results. As we expected, firm size, financial performance, and CEO duality were positively and significantly associated with CEO total pay. CEO ownership, however, was negatively related to CEO total compensation. Models B and C measure the impact of pollution prevention (PP) and end-of-pipe pollution control (EOP), respectively, on CEO total pay. Each had a positive and significant effect on CEO total pay (PP, p .01; EOP, p .05), providing strong support for Hypothesis 1, which predicted that environmental performance would be positively associated with CEO total pay. Model D includes both pollution prevention and end-of-pipe environmental performance measures in the same equation; only pollution prevention was statistically significant (p .01), in line with Hypothesis 2, which predicted that pollution prevention would have a higher impact on CEO pay than end-of-pipe pollution control. In addition, we conducted an analysis on the increments in R2, a measure of explained variance, to confirm this result. We tested the increment to R2 with the pollution prevention variable first and the end-of-pipe variable later. In the first case (pollution prevention), the increment was 2 percent, while in the second case (end-of-pipe), it was 1 percent. When including both measures in the model, the R2 increment was 2 percent, with the pollution prevention variable being significant and the end-of-pipe variable, nonsignificant. Together, these results suggest that adding the end-of-pipe measure to the model does not add sig- We also identified those firms for which there was a gap in the environmental data (that is, environmental data were missing for at least one year) and reran our models excluding them from the estimations. Results from these new estimations were fully consistent with those presented in this article. 9 Detailed explanation of the Heckman’s estimation and its results are available from the authors upon request. It should be noted that a limitation of this procedure is that the actual selection may occur at the facility level. However, owing to the lack of information at the facility level, analysis was performed at the firm level. In addition to this test, we conducted a separate analysis including a dummy variable that assumed the value 1 if a firm reported its emissions and 0 otherwise. As we expected, this variable was negative and significant. 8 114 Academy of Management Journal February TABLE 1 Descriptive Statistics and Correlationsa Variables Total payb Long-term payb Firm size (assets)b Firm financial performance (ROE) 5. Tobin’s Q 6. CEO duality 7. CEO ownership 8. CEO tenure 9. Proportion of outside directors 10. Director ownership 11. Family firm status 12. Pollution prevention 13. End-of-pipe pollution control 14. Age of assets 15. Environmental governance 16. Regulatory stringency 17. Industry pollution position 18. Reporting plants 1. 2. 3. 4. a b Mean s.d. 13.67 2.45 6.87 2.12 7.92 1.45 11.66 34.94 1.30 1.31 0.50 0.50 1.80 6.51 4.90 6.24 81.47 12.38 4.58 12.66 0.09 0.28 0.09 1.01 0.05 0.88 0.24 0.47 0.23 0.47 0.54 3.58 16.20 6.28 11.28 14.80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 .87 .47 .15 .25 .29 .15 .10 .15 .07 .14 .01 .03 .12 .11 .19 .07 .14 .38 .14 .31 .23 .16 .07 .11 .05 .12 .01 .05 .11 .06 .17 .06 .11 .09 .04 .09 .17 .18 .12 .11 .08 .14 .17 .08 .31 .06 .06 .46 .20 .03 .02 .06 .01 .01 .02 .02 .05 .11 .02 .01 .01 .00 .05 .03 .04 .07 .02 .06 .06 .21 .02 .13 .11 .07 .15 .07 .37 .11 .05 .06 .03 .08 .12 .08 .10 .02 .05 .22 .16 .14 .10 .08 .06 .01 .09 .06 .05 .11 .13 .06 .05 .07 .02 .03 .14 .13 .03 .14 .15 .10 .04 .08 .04 .14 .02 .03 .09 .25 .02 .01 .03 .10 .03 .09 .01 .02 .02 .03 .07 .02 .03 .01 .08 .01 .09 .18 .07 .17 .01 .12 .19 .21 .02 .03 .04 .06 .07 .05 .13 .29 .06 .09 .14 n 2,088. Correlations above .03 or below Logarithm. .03 are significant at the 5 percent level. nificantly more explanatory power than the pollution prevention measure by itself, providing support for Hypothesis 2. Following previous environmental studies (Aragon-Correa, 1998; Russo & Fouts, 1997), we conducted F-tests to examine the explanatory value of our independent variables. Results of these tests indicated that the increments in variance explained between the control model and the full models were all significant. This suggests that although the increase in variance was modest, the impact of environmental performance on CEO compensation was not due to random chance. In addition, it should be noted that the increments in explained variance are comparable to those obtained in other compensation studies (e.g., Bloom & Milkovich, 1998; Miller et al., 2002). Table 3 shows results for Hypotheses 3a and 3b, which predicted that environmental governance mechanisms would moderate the relationship between environmental performance and CEO pay and that this moderating effect would be greater for pollution prevention outcomes than for end-ofpipe results. Following recommendations of Aiken and West (1991) for reducing potential collinearity, we centered the environmental governance and environmental performance variables. We also calculated the variance inflation factor (VIF) after each regression to see whether results were subject to multicollinearity. Values were within acceptable limits, indicating that estimations were free of any significant multicollinearity bias. Contrary to Hypothesis 3a, findings in model C show that the interaction term for environmental governance mechanisms and pollution prevention strategies was negative and moderately significant (p .10). Also, contrary to our expectation, firms with environmental governance mechanisms in place do not rely on evidence of end-of-pipe pollution control strategies as a criterion for CEO pay. In all cases, R2 increments were below 1 percent, and results of F-tests indicated that the increments in variance explained between models were not significant. The exception was the increment in R2 between models B and C, which was moderately significant (p .10). As a whole these results fail to support both Hypothesis 3a and Hypothesis 3b.10 Tables 4 (with pollution prevention as dependent variable) and 5 (with end-of-pipe pollution control as dependent variable) summarize our analyses for testing Hypotheses 4a and 4b (the impact of long-term pay on subsequent financial performance). To account for the fact that it may take some time for managers to improve environmental performance, and for the probability that incentives already awarded may still convey incentive power in later years, we used the average value of longterm pay for two previous years. Model A of Table 4 includes only our control variables for a fixed- 10 We also calculated our models using two dummy variables, one for explicit environmental pay policy and another for environmental committee. Results of individual effects and three-way interactions are consistent with those presented in this article. 2009 Berrone and Gomez-Mejia 115 TABLE 2 Results of Fixed-Effects Panel Data Analyses on CEO Total Paya Variables Controls Firm size (log assets) Firm financial performance (ROE) Tobin’s Q CEO duality CEO ownership CEO tenure Proportion of outside directors Director ownership Family firm status Industry pollution position Reporting plants Main effects Pollution prevention End-of-pipe pollution control F R2 R2 F 23.15*** .34 Model A Model B Model C Model D 0.61*** (0.10) 2.89*** (0.72) 0.24*** (0.03) 0.09*** (0.02) 0.09*** (0.03) 0.09 (0.15) 0.02 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.03) 0.01 (0.04) 0.61*** (0.10) 2.87*** (0.71) 0.24*** (0.03) 0.09*** (0.02) 0.09** (0.03) 0.10 (0.15) 0.01 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.03) 0.01 (0.04) 0.59*** (0.10) 2.57*** (0.62) 0.26*** (0.03) 0.08*** (0.02) 0.05† (0.03) 0.08 (0.14) 0.02 (0.01) 0.00 (0.01) 0.01 (0.01) 0.01 (0.02) 0.01 (0.04) 0.62*** (0.10) 2.88*** (0.71) 0.25*** (0.03) 0.08*** (0.02) 0.09** (0.03) 0.08 (0.15) 0.01 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.02) 0.01 (0.03) 0.03** (0.01) 0.04* 20.15*** .36 .02 8.20** 22.80*** .35 .01 2.55* (0.02) 0.03** (0.01) 0.03 (0.02) 19.88*** .36 .02 11.06* a n 2,088. Table contains standardized regression coefficients. Standard errors are in parentheses. All models include controls for year that are not shown because of space constraints. R2 increments and their tests are with respect to model A. † p .10 * p .05 ** p .01 *** p .001 Two-tailed tests. effects model. Contrary to previous research (Grant, Jones, & Bergesen, 2002), firm size was not related to pollution prevention performance, but the reporting plants variable (also a proxy for size) was negatively and significantly related to pollution prevention as well as to the industry pollution position of a focal firm. Regulatory stringency, Tobin’s Q, and family firm status, on the other hand, were positive and slightly significant. Model B shows that long-term pay had a positive and highly significant impact on subsequent pollution prevention performance (p .01), strongly supporting Hypothesis 4a. Model C of Table 4 includes the interaction term between industry pollution position and long-term pay, which is positively and moderately significant (p .10). In accord with Hypothesis 4b, this finding suggests that the positive impact of long-term pay on subsequent environmental performance is greater for firms operating in the highest-pollution sectors. In models D, E, and F of Table 4 we ran the same models as before but used a Tobit model with random effects, which allowed us to control for potential selection bias (i.e., to account for the potential impact that omitting nonreporting firms might have on our analysis). Model D contains the control vari- ables and, as with the fixed-effects model, reporting plants, regulatory stringency, and family firm status were all significant, but the level of their significance increased considerably (p .001). As we expected, the influence of industry pollution position on pollution prevention performance was negative and highly significant. Model E again confirms Hypothesis 4a, showing a positive and moderately significant (p .10) coefficient for the effect of CEO long-term pay on pollution prevention performance. Lastly, the interaction between a firm’s industry pollution position and long-term pay (model F) is positive and highly significant for pollution prevention performance (p .01), confirming Hypothesis 4b. Again, VIF values were within acceptable limits, indicating that estimations were free of any significant multicollinearity bias. This result indicates that the impact of longterm pay on subsequent pollution prevention is stronger as a firm’s presence in highly polluting sectors increases, supporting Hypothesis 4b. Figure 1 visually depicts this relationship. Again, F-tests confirmed the strong significance of our explanatory variables. Turning our attention to Table 5 (which provides 116 Academy of Management Journal February TABLE 3 Results of Fixed-Effects Analysis of the Moderating Role of Environmental Governancea Variables Controls Firm size (log assets) Firm financial performance (ROE) Tobin’s Q CEO duality CEO ownership CEO tenure Proportion of outside directors Director ownership Family firm status Industry pollution position Reporting plants Main effects Pollution prevention End-of-pipe pollution control Environmental governance Interactions Pollution prevention environmental governance End-of-pipe pollution control environmental governance F R2 R2 F 19.88*** .36 18.90*** .36 0.00 0.64 Model A Model B Model C Model D 0.62*** (0.10) 2.88*** (0.71) 0.25*** (0.03) 0.08*** (0.02) 0.09** (0.03) 0.08 (0.15) 0.01 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.02) 0.01 (0.03) 0.62*** (0.10) 2.88*** (0.71) 0.25*** (0.03) 0.08*** (0.02) 0.09** (0.03) 0.08 (0.15) 0.01 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.02) 0.01 (0.03) 0.62*** (0.10) 2.90*** (0.71) 0.25*** (0.03) 0.08*** (0.02) 0.09** (0.03) 0.08 (0.15) 0.01 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.02) 0.01 (0.04) 0.62*** (0.10) 2.90*** (0.71) 0.25*** (0.03) 0.08*** (0.02) 0.09** (0.03) 0.08 (0.15) 0.01 (0.01) 0.01 (0.01) 0.02 (0.02) 0.01 (0.03) 0.01 (0.04) 0.03* 0.03 (0.01) (0.02) 0.03** (0.01) 0.03† (0.02) 0.02 (0.02) 0.03** (0.01) 0.03† (0.02) 0.01 (0.02) 0.03** (0.01) 0.03† (0.02) 0.03 (0.02) 0.02† (0.01) 0.02† 0.02 (0.01) (0.01) 18.01*** .37 0.01 2.78† 17.21*** .37 0.00 1.27 a n 2,088. Table contains standardized regression coefficients. Standard errors are in parentheses. All models include controls for year that are not shown because of space constraints. R2 increments and their test are with respect to the preceding model. † p .10 * p .05 ** p .01 *** p .001 Two-tailed tests. predictor information equivalent to that in Table 4, except that end-of-pipe performance replaces pollution prevention performance as the dependent variable), we find that the hypothesized main effects and interactions were weak and not statistically significant at the conventional (p .05) level. Only the interaction term in model C was slightly significant (p .10). This result is consistent with Hypotheses 4a and 4b, which predicted a stronger main effect and interaction in the case of pollution prevention (see Table 4) than end-of-pipe pollution controls (see Table 5). DISCUSSION AND CONCLUSIONS Given the increasing public awareness of environmental issues and the growing belief that environmental strategies can become a key source of competitive advantage in polluting industries, un- derstanding the relationships between CEO pay and environmental performance is crucial to the successful management of corporations. Consequently, the conclusions derived here have not only theoretical meaning, but also clear practical content. Implications for Research We found that environmental performance can be an important nonfinancial determinant of CEO pay within polluting industries, even after controlling for accounting and market-based measures of performance and other traditional determinants of executive compensation. These results suggest that CEOs are rewarded for pursuing environmental strategies because the outcomes associated with these strategies may provide intangible benefits that go beyond “hard-core” financial performance 2009 Berrone and Gomez-Mejia 117 TABLE 4 Results of Panel Data Analyses on Pollution Preventiona Fixed Effects Variables Controls Age of assets Regulatory intensity Firm size (log assets) Firm financial performance (ROE) Tobin’s Q CEO duality CEO ownership CEO tenure Proportion of outside directors Director ownership Family firm status Industry pollution position Reporting plants Model A Model B Model C Model D Random Effects (Tobit) Model E Model F 0.06 0.14† 0.12 0.01 0.05† 0.03 0.02 0.15 0.06 (0.05) (0.08) (0.10) (0.44) (0.03) (0.02) (0.02) (0.12) (0.07) 0.05 0.14† 0.16† 0.09 0.02 0.02 0.02 0.11 0.06 (0.05) (0.08) (0.10) (0.44) (0.03) (0.02) (0.02) (0.12) (0.07) 0.04 0.04 0.15† 0.42 0.04† 0.02 0.03 0.35* 0.03 (0.06) (0.06) (0.09) (0.44) (0.02) (0.02) (0.02) (0.12) (0.07) 0.14* (0.06) 0.27*** (0.04) 0.27*** (0.06) 0.57 (0.56) 0.16*** (0.05) 0.05 (0.03) 0.02 (0.02) 0.03 (0.04) 0.02 (0.06) 0.05 (0.03) 0.10*** (0.02) 0.06* (0.02) 0.32*** (0.04) 0.10 (0.07) 0.16*** (0.04) 0.33*** (0.07) 0.73 (0.61) 0.18** 0.02 0.01 0.01 0.03 (0.06) (0.03) (0.03) (0.06) (0.06) 0.02 (0.06) 0.17*** (0.03) 0.33*** (0.05) 0.72 (0.61) 0.17*** (0.05) 0.03 (0.03) 0.02 (0.02) 0.05 (0.04) 0.02 (0.05) 0.04 (0.03) 0.10*** (0.02) 0.20* (0.03) 0.33*** (0.03) 0.01 (0.03) 0.03* (0.02) 0.06** (0.02) 0.14*** (0.03) 0.01 (0.03) 0.03* (0.02) 0.06** (0.02) 0.14*** (0.03) 0.03 (0.02) 0.04* (0.02) 0.05** (0.02) 0.17*** (0.03) 0.05 (0.03) 0.11*** (0.02) 0.19*** (0.03) 0.35*** (0.04) Main effect Long-term pay 0.06** (0.02) 0.04† (0.02) 0.03† (0.02) 0.04 (0.03) Interaction Long-term pay industry pollution position 0.04† (0.02) 0.07** (0.03) n F Wald 2 R2 Log-likelihood (Tobit) F 1,714 27.28*** .13 1,714 26.49*** .15 4.08** 1,714 24.34*** 1,004.54*** .15 10.05** 2,837 989.09*** 2,060.3 2,837 934.64*** 1,925.4 3.28† 2,837 1,911.3 11.44** a Table contains standardized regression coefficients. Standard errors are in parentheses. All models include controls for year that are not shown because of space constraints. F-tests are with respect to the control model. † p .10 * p .05 ** p .01 *** p .001 Two-tailed tests. (social legitimacy, corporate reputation, stakeholder satisfaction, and so on). Previous findings of studies relating executive pay and social indicators (Coombs & Gilley, 2005; Russo & Harrison, 2005; Stanwick & Stanwick, 2001) are consistent with the so-called traditionalist view, in which a trade-off between environmental strategies and profitability is posited (Jaffe, Peterson, Portney, & Stavins, 1995; Walley & Whitehead, 1994). In contrast, our results are in line with the revisionist view that good environmental performance is beneficial for firms (Hart, 1995; Porter & van der Linde, 1995). The setting of our study can help shed some light on this apparent contradiction. We show that firms within polluting industries may achieve legitimacy in their institutional field by adopting environment-friendly processes, and their CEOs are rewarded accordingly. However, this might not be the case for broader samples or for nonpolluting industries (e.g., Coombs & Gilley, 2005). Future research should be aimed at identifying the specific settings in which environmental investments have positive, neutral, or negative relationship with compensation and other organizational features. Other design aspects of our study may explain the differences between our results and those that support the traditional view: we used an environmental performance measure with much greater range and variability than those used in past research (e.g., the KLD index); we accounted for board characteristics and ownership structure as control variables; and we used a sophisticated set of analytical techniques to tease out various effects with a relatively large sample. From an institutional perspective, linking compensation to environmental performance induces 118 Academy of Management Journal February FIGURE 1 Effect of the Interaction of Long-Term Pay and Industry Pollution Position on Pollution Preventiona 0.25 0.2 0.15 Pollution Prevention 0.1 0.05 0 2 2.2 2.4 2.6 2.8 3 3.2 Long-Term Pay 3.4 3.6 3.8 4 Firms in industries with average pollution Firms in industries with high pollution a Standardized values from the Tobit model (Table 4, model F) were used for calculations. The average value of polluting industries used in the graph is the mean value of the industry pollution position measure. managers to conform to institutional demands and discourages avoidance (Ashforth & Gibbs, 1990; Oliver, 1991). From an agency perspective, our results are in line with the “positivist” agency argument (Beatty & Zajac, 1994; Eisenhardt, 1989; Jensen, 1983) that when the link between strategies and performance is uncertain, a principal will use a criterion over which agents have more influence (as per the controllability principle discussed earlier) and that may improve financial performance (as per the informativeness principle noted before). Recently, agency theory has come under intense fire in the management literature on grounds that it fails to address the influence of the social and institutional environment surrounding a principalagent relation (see Aguilera & Jackson, 2003; Bruce, Buck, & Main, 2005; Fligstein & Freeland, 1995; Lubatkin et al., 2007). Our study is a response to these critics and expands agency theory by including insights from institutional theory, examining how a key institutional factor (i.e., legitimacy) intervenes in the agent-principal relationship. Our study is in accord with Eisenhardt’s (1989) recommendation to expand agency theory to richer and more complex contexts. It is also fully consistent with recent work by Gomez-Mejia and his col- leagues (Gomez-Mejia & Wiseman, 2007; GomezMejia, Wiseman, & Johnson, 2005; Wiseman, Cuevas-Rodriguez, & Gomez-Mejia, 2008) regarding the influence of institutional context on principalagent relationships. In short, our study suggests that institutional theory can reinforce rather than negate the basic tenets of agency theory. Our finding that end-of-pipe pollution control did not contribute significantly to CEO pay is consistent with much of the environmental management literature and with the idea that achieving legitimacy in a strong institutional field requires relatively substantial strategies. The environmental management literature has offered a variety of answers to the question, Why do firms go green? The importance of managers in this process seems to be widely recognized, but different researchers have offered varied explanations. Some have suggested that it is the strategic position of managers that determines their environmental performance (e.g., Aragon-Correa, 1998). Others have argued that their “greenness” depends on how sensitive they are to stakeholder pressures (e.g., Henriques & Sadorsky, 1999). Still others have sought explanations in their ethical values and stances (e.g., Bansal & Roth, 2000). The underlying reason suggested by our 2009 Berrone and Gomez-Mejia 119 TABLE 5 Results of Panel Data Analyses on End-of-Pipe Pollution Controla Fixed Effects Variables Controls Age of assets Regulatory intensity Firm size (log assets) Firm financial performance ROE Tobin’s Q CEO duality CEO ownership CEO tenure Proportion of outside directors Director ownership Family firm status Industry pollution position Reporting plants Main effect Long-term pay Model A Model B Model C Model D Random Effects (Tobit) Model E Model F 0.13 0.13† 0.18† 0.30 0.02 0.02 0.00 0.33† 0.03 (0.08) (0.08) (0.11) (0.52) (0.04) (0.03) (0.03) (0.16) (0.06) 0.12 0.13† 0.21* 0.23 0.00 0.03 0.00 0.30† 0.03 (0.09) (0.08) (0.11) (0.52) (0.04) (0.03) (0.03) (0.16) (0.06) 0.18† 0.04 0.19† 0.53 0.01 0.02 0.01 0.46** 0.01 (0.09) (0.07) (0.10) (0.62) (0.04) (0.03) (0.03) (0.16) (0.05) 0.21* (0.09) 0.25*** (0.05) 0.26*** (0.05) 0.33 (0.56) 0.15* 0.02 0.04 0.11** 0.09 (0.06) (0.03) (0.04) (0.04) (0.06) 0.14 (0.09) 0.23*** (0.04) 0.24*** (0.05) 0.43 (0.59) 0.14* 0.03 0.05 0.10** 0.07 (0.07) (0.03) (0.04) (0.04) (0.06) 0.18† (0.10) 0.23*** (0.04) 0.27*** (0.05) 0.25 (0.59) 0.13† 0.02 0.05 0.09** 0.07 (0.07) (0.03) (0.04) (0.04) (0.06) 0.01 (0.02) 0.01 (0.03) † 0.06 (0.03) 0.24*** (0.06) 0.01 (0.02) 0.01 (0.03) † 0.06 (0.03) 0.24*** (0.06) 0.05 (0.04) 0.01 (0.02) 0.06† (0.03) 0.05 (0.03) 0.35*** (0.05) 0.03 (0.04) 0.01 (0.04) 0.04 (0.03) 0.04 (0.04) 0.27*** (0.04) 0.00 (0.04) 0.04 (0.03) † 0.05 (0.03) 0.26*** (0.05) 0.05 (0.04) 0.00 (0.04) 0.05† (0.03) † 0.05 (0.03) 0.23*** (0.05) 0.04 (0.04) Interaction Long-term pay industry pollution position n F Wald 2 R2 Log-likelihood (Tobit) F 0.06† 1,714 4.21*** .07 1,714 4.13*** .07 1.13 1,714 3.89*** .10 (0.03) 2,837 1,071.18*** 2,822.7 2,837 896.31*** 2,666.7 4.71 0.02 2,837 836.75*** 2,568.5 1.02 (0.03) 5.65† a Standard errors are in parentheses. Regression coefficients are standardized. All models include controls for year that are not shown because of space constraints. F-tests are with respect to the control model. † p .10 * p .05 ** p .01 *** p .001 work is perhaps more rational in the economic sense: CEOs follow environmental strategies simply because they have economic incentives to do so. Of course, this does not rule out the above explanations. On the contrary, it suggests that being conscious about the environmental claims of stakeholders and having strong ethical principles can be of value for shareholders and managers. Industries differ dramatically in their emissions, and firms’ environmental performance depends largely on their industrial positions. A company’s expansion in one industry or selling off a facility in another sector can affect its environmental marks. This is an important research issue that has been ignored in many relevant environmental studies. In addition to including an index that accounts for changes in a firm’s industry position, we conducted several separate analyses to assess the potential effect of selling off plants. For instance, we calculated our pollution prevention and end-ofpipe measures using those facilities that reported in all periods and remained under the control of the same company for the entire period of analysis. Patterns of coefficients were similar to those reported, although the significance of main variables varied slightly in some cases.11 Moreover, we compared the selling rates of high and low polluters and found no statistically significant differences (with a 99 percent confidence level). This result suggests that even though selling off plants may occur, such divesting is not large enough to disturb our measures—and, more importantly, that highly polluting firms do not unload plants at a greater These changes in significance do not invalidate our results for two reasons. First, coefficients were significant at the .10 level or better. Second, statistical significance tests are not a necessary or sufficient condition for assessing the true importance of a variable as they suffer from many imperfections and thus are not conclusive (Schmidt & Hunter, 1995; Ziliak & McCloskey, 2004). 11 120 Academy of Management Journal February rate than low-polluting firms as a strategy to improve their environmental records. Future research could expand this line of inquiry, which may have significant policy implications. Although from a firm perspective it may be irrelevant how the reduction of overall firm emissions occurs (either by cleaning up the plants the firm retains or by getting rid of dirty plants over time), from a social welfare perspective these alternatives have clearly different consequences. The corporate governance literature has largely neglected governance mechanisms related to environmental issues. We did not find that environmental performance had a higher impact on CEO total pay in firms with both environmental pay policy and environmental committees. One potential explanation is that these mechanisms are symbolic rather than instrumental: firms that are unwilling to make the necessary investment to reduce or eliminate toxic emissions may instead adopt structures like explicit environmental pay policies and environmental committees to signal concern about the natural environment and appear to be taking the right steps to preserve it. This result is consistent with the logic that certain governance mechanisms are a response to institutional pressures (Luoma & Goodstein, 1999). Another way of looking at this is that environmental committees are cheaper and much more expedient in dealing with those pressures than actually redesigning plants to reduce pollution. Implications for Managers and Directors Structuring executive compensation around environmental performance can benefit firms in several ways. First, it can stimulate managers to deploy efforts and resources toward environmental initiatives that should build a resource vital to firm survival and success: legitimacy. Second, it holds them explicitly accountable for firms’ environmental behavior. Third, it can encourage CEOs to monitor environmental behaviors at lower organizational levels. Thus it can produce desirable outcomes for shareholders, who see firm survival secured; for managers, who increase their pay; and for society, as its noxious burden is alleviated. However, this “triple win” appears possible only as long as all stakeholders recognize the negative economic and social implications of poor environmental performance. Unfortunately, linking pay to environmental performance may also be open to manipulation. Although the TRI system sanctions firms for deficiencies in reporting, misrepresentation may still be possible. Therefore control policies and information systems are needed. However, our study suggests that environmental governance mechanisms are merely symbolic actions and casts doubts on the efficacy of environmental committees. Alternative monitoring mechanisms, such as external environmental audits, may be more effective. Our results also suggest that what firms actually reward is evidence of pollution prevention strategies rather than evidence of end-of-pipe strategies. Top executives should take note that while pollution prevention strategies are likely to be more demanding and risky, they should also bring greater rewards to a firm and therefore to its CEO. Finally, we found that long-term pay is an important incentive for pollution prevention, and it is more effective where it is needed the most—that is, in highly polluting industries. In sectors where polluting emissions are less of a concern, long-term pay is less likely to influence pollution levels than in sectors where the environmental impact is greater. Although we focused on the level of longterm pay, our results may suggest that a firm with poor environmental performance should increase the proportion of long-term pay in the CEO compensation package. However, this conclusion should be viewed with caution for two reasons. First, our random-effect models require the assumption of zero correlation between latent heterogeneity and included observed characteristics, which is particularly restrictive. Second and perhaps more important, this conclusion refers to the pay mix, an issue that was not examined in our study and that is probably deserving of its own study. Much theory building and research is still needed to understand how the relative proportions of various forms of pay in a compensation package may influence CEO decisions that affect subsequent environmental performance. This is a complex issue that would entail the analysis of factors such as risk transfer to agent, framing of problem, instant endowment, stock volatility, perverse incentives to manipulate stock values, whether or not the executive may be held responsible for past performance (for instance, whether the executive was recently appointed or has long tenure in the firm), fungibility of various pay forms, and the like. (For related discussions of some of these issues see Bloom and Milkovich [1998], Deya-Tortella, Gomez-Mejia, De Castro, ´ and Wiseman [2005], Gray and Cannella [1997], Miller et al. [2002], and Wiseman and GomezMejia [1998].) 2009 Berrone and Gomez-Mejia 121 Limitations and Future Research At least four limitations in our study could be rectified in future research. First, we concentrated our analysis on CEOs of relatively large companies in the United States, and our findings may not be generalizable to smaller companies or other geographical regions. Future research can extend the analysis to non-U.S. contexts, to other managerial levels, or to other organizational forms such as public organizations and family firms. Second, our measures considered only environmental data reported in the United States, but some firms locate their dirty operations in “pollution havens” abroad. Unfortunately, reliable pollution data on a global scale are almost nonexistent. This issue represents a challenge and an opportunity for future research. Third, the environmental performance measures we used may not necessarily be the ones boards of directors would use or may not be readily accessible to them. There is limited information about the sources boards use to appraise CEOs’ environmental performance; some board members may have expertise concerning environmental issues, or a board may consult environmental specialists. We recognize that our measures are proxies and may only partially capture boards’ evaluative processes. However, it would be very difficult to fully tap the cognitive process of the board of each firm. Makri et al. (2006), who used patent citations as a proxy for firm innovation quality, argued that “close personal involvement of the researchers in board deliberations . . . would not be very feasible and . . . in the best of cases, would yield very small samples with limited statistical power and generalizability” (2006: 1068). Future research could address this issue by examining alternative measures of environmental performance that might be more likely to draw board attention. Fourth, the topics of how environmental performance impacts compensation and how incentives affect subsequent environmental performance should, ideally, be analyzed simultaneously. Given the structure of our data (panel), fixed-effects models were appropriate, and we conducted our analysis in separate steps. However, future research should look at the complex interrelations discussed in this article and use more sophisticated methods (such as structural equation models) to analyze them simultaneously. A final avenue for future research concerns the role of environmental governance. For instance, the structure of an environmental committee, the characteristics of its members, and the fact that it can serve more as a consultant than as an inde- pendent overseer are issues to be pursued in future analyses. Final Remark Good environmental behaviors are important in achieving social legitimacy; firms should support environmental strategies by using environmental performance criteria to reward their CEOs. Such incentives can help shareholders, managers, and the general public benefit from good environmental performance. REFERENCES Aguilera, R. V., & Jackson, G. 2003. The cross-national diversity of corporate governance: Dimensions and determinants. Academy of Management Review, 38: 447– 465. Aiken, L. S., & West, S. G. 1991. Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Aldrich, H. E., & Fiol, C. M. 1994. Fools rush in? The institutional context of industry creation. Academy of Management Review, 19: 645– 670. Antle, R., & Demski, J. S. 1988. The controllability principle in responsibility accounting. Accounting Review, 63: 700 –718. Aragon-Correa, J. A. 1998. Strategic proactivity and firm approach to the natural environment. Academy of Management Journal, 41: 556 –567. Aragon-Correa, J. A., Matıas-Reche, F., & Senise-Barrio, ´ M. A. 2003. Managerial discretion and corporate commitment to the natural environment. Journal of Business Research, 57: 964 –975. Aragon-Correa, J. A., & Sharma, S. 2003. A contingent resource-based view of proactive corporate environmental strategy. Academy of Management Review, 28: 71– 88. Ashforth, B. E., & Gibbs, B. W. 1990. The double-edge of organizational legitimation. Organization Science, 1: 177–194. Baker, W. 1990. Market networks and corporate behavior. American Journal of Sociology, 96: 589 – 625. Balkin, D. B., Markman, G. D., & Gomez-Mejia, L. R. 2000. Is CEO pay in high-technology firms related to innovation? Academy of Management Journal, 43: 1118 –1129. Bansal, P. 2005. Evolving sustainably: A longitudinal study of corporate sustainable development. Strategic Management Journal, 26: 197–218. Bansal, P., & Clelland, I. 2004. Talking trash: Legitimacy, impression management, and unsystematic risk in the context of the natural environment. Academy of Management Journal, 47: 93–103. 122 Academy of Management Journal February Bansal, P., & Roth, K. 2000. Why companies go green: A model of ecological responsiveness. Academy of Management Journal, 43: 717–736. Barkema, H. G., & Gomez-Mejia, L. R. 1998. Managerial compensation and firm performance: A general research framework. Academy of Management Journal, 41: 135–145. Baysinger, B., & Hoskisson, R. E. 1990. The composition of boards of directors and strategic control: Effects on corporate strategy. Academy of Management Review, 15: 72– 87. Baysinger, B. D., Kosnik, R. D., & Turk, T. A. 1991. Effects of board and ownership structure on corporate R&D strategy. Academy of Management Journal, 34: 205–214. Beatty, R. P., & Zajac, E. J. 1994. Managerial incentives, monitoring, and risk bearing: A study of executive compensation, ownership, and board structure in initial public offerings. Administrative Science Quarterly, 39: 313–335. Bloom, M., & Milkovich, G. T. 1998. Relationship among risk, incentive pay, and organizational performance. Academy of Management Journal, 41: 283–297. Boyd, B. 1994. Board control and CEO compensation. Strategic Management Journal, 15: 335–344. Bruce, A., Buck, T., & Main, B. G. 2005. Top executive remuneration: A view from Europe. Journal of Management Studies, 42: 1493–1506. Buysee, K., & Verbeke, A. 2003. Proactive environmental strategies: A stakeholder management perspective. Strategic Management Journal, 24: 453– 470. Christmann, P. 2000. Effects of “best practices” of environmental management on cost advantage: The role of complementary assets. Academy of Management Journal, 43: 663– 680. Chung, K. S., & Pruitt, W. 1994. A simple approximation of Tobin’s Q. Financial Management, 23(3): 70 –74. Conyon, M. J., & Peck, S. I. 1998. Board control, remuneration committees, and top management compensation. Academy of Management Journal, 41: 146 – 157. Coombs, J. E., & Gilley, K. M. 2005. Stakeholder management as a predictor of CEO compensation: Main effects and interactions with financial performance. Strategic Management Journal, 26: 827– 840. Dalton, D. R., Daily, C. M., Ellstrand, A. E., & Johnson, J. L. 1998. Meta-analytic reviews of board composition leadership structure and financial performance. Strategic Management Journal, 19: 269 –290. Deephouse, D. L. 1999. To be different, or to be the same? It’s a question (and theory) of strategic balance. Strategic Management Journal, 20: 147–166. Demski, J. S. 1994. Managerial uses of accounting information. Boston: Kluwer. Demski, J. S., & Feltham, G. A. 1978. Economic incen- tives in budgetary control systems. Accounting Review, 2: 336 –359. Deya-Tortella, B., Gomez-Mejia, L. R., De Castro, J., & ´ Wiseman, R. M. 2005. Incentive alignment or perverse incentives? A behavioral view of stock options. Management Research, 3: 109 –120. DiMaggio, P. J., & Powell, W. W. 1983. The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48: 147–160. DiMaggio, P. J., & Powell, W. W. 1991. Introduction. In P. J. DiMaggio & W. W. Powell (Eds.), The new institutionalism in organizational analysis: 1–38. Chicago: University of Chicago Press. Dowell, G., Hart, S. L., & Yeung, B. 2000. Do corporate global environmental standards create or destroy market value? Management Science, 46: 1059 – 1074. Dowling, J., & Pfeffer, J. 1975. Organizational legitimacy: Social values and organizational behavior. Pacific Sociological Review, 18(1): 122–136. Eisenhardt, K. 1989. Agency theory: An assessment and a review. Academy of Management Review, 14: 57–74. EPA. 1997. Terms of environment: Glossary, abbreviations, and acronyms. Washington, DC: Environmental Protection Agency. EPA. 2002. The Toxics Release Inventory (TRI) and factors to consider when using TRI data. Report no. 60-F-02– 017: 1–29. Washington DC: Environmental Protection Agency. Fama, E. F. 1980. Agency problems and the theory of the firm. Journal of Political Economy, 88: 288 –298. Fama, E. F., & Jensen, M. C. 1983. Separation of ownership and control. Journal of Law and Economics, 26: 301–325. Finkelstein, S. 1992. Power in top management teams: Dimensions, measurement, and validation. Academy of Management Journal, 35: 505–538. Finkelstein, S., & Boyd, B. 1998. How much does the CEO matter? The role of managerial discretion in the setting of CEO compensation. Academy of Management Journal, 41: 179 –199. Fligstein, N., & Freeland, R. 1995. Theoretical and comparative perspectives on corporate organization. In J. Hagan & K. Cook (Eds.), Annual review of sociology, vol. 21: 21– 43. Palo Alto, CA: Annual Reviews. Franco, M. 2007. Performance measurement issues and incentive application. In L. R. Gomez-Mejia & S. Werner (Eds.), Global compensation: 47–58. London: Routledge. Godfrey, P. C. 2005. The relationship between corporate philanthropy and shareholder wealth: A risk management perspective. Academy of Management Review, 30: 777–798. 2009 Berrone and Gomez-Mejia 123 Gomez-Mejia, L. R., Larraza-Kintana, M., & Makri, M. 2003. The determinants of executive compensation in family-controlled public corporations. Academy of Management Journal, 46: 226 –238. Gomez-Mejia, L. R., Nunez-Nickel, M., & Gutierrez, I. ˜ 2001. The role of family ties in agency contracts. Academy of Management Journal, 44: 81–95. Gomez-Mejia, L. R., Takacs-Haynes, K., Nunez-Nickel, M., Jacobson, K. J. L., & Moyano, J. 2007. Socioemotional wealth and business risks in family-controlled firms. Administrative Science Quarterly, 52: 106 – 138. Gomez-Mejia, L. R., & Wiseman, R. M. 1997. Reframing executive compensation: An assessment and outlook. Journal of Management, 23: 291–374. Gomez-Mejia, L. R., & Wiseman, R. M. 2007. Does agency theory have universal relevance? A reply to Lubatkin, Lane, Collin, and Very. Journal of Organizational Behavior, 28: 81– 88. Gomez-Mejia, L. R., Wiseman, R. M., & Johnson, B. 2005. Agency problems in diverse contexts: A global perspective. Journal of Management Studies, 42: 1507–1520. Grant, D. S., Jones, A. W., & Bergesen, A. J. 2002. Organizational size and pollution: The case of the U.S. chemical industry. American Sociological Review, 67: 389 – 407. Gray, S. R., & Cannella, A. A. 1997. The role of risk in executive compensation. Journal of Management, 23: 517–540. Greene, W. H. 1993. Econometric analysis (2nd ed.). New York: MacMillan. Greening, D. W., & Gray, B. 1994. Testing a model of organizational response to social and political issues. Academy of Management Journal, 37: 467– 498. Hart, S. L. 1995. A natural-resource-based view of the firm. Academy of Management Review, 20: 986 – 1014. Hart, S. L., & Ahuja, G. 1996. Does it pay to be green? An empirical examination of the relationship between emission reduction and firm performance. Business Strategy and the Environment, 5(1): 30 –37. Heckman, J. J. 1979. Sample selection bias as a specification error. Econometrica, 47: 153–162. Heckman, J. J., & MacCurdy, T. E. 1980. A life-cycle model of female labor supply. Review of Economic Studies, 47: 47–74. Henderson, A. D., & Fredrickson, J. W. 1996. Informationprocessing demands as a determinant of CEO compensation. Academy of Management Journal, 39: 575– 606. Henderson, A. D., & Fredrickson, J. W. 2001. Top management team coordination needs and the CEO pay gap: A competitive test of economic and behavioral views. Academy of Management Journal, 44: 96 – 117. Henriques, I., & Sadorsky, P. 1999. The relationship between environmental commitment and managerial perceptions of stakeholder importance. Academy of Management Journal, 42: 87–99. Hertwich, E. G., Mateles, S. F., Pease, W. S., & McKone, T. E. 2001. Human toxicity potentials for life-cycle assessment and Toxics Release Inventory risk screening. Environmental Toxicology and Chemistry, 20: 928 –939. Hill, C. W., & Phan, P. H. 1991. CEO tenure as a determinant of CEO pay. Academy of Management Journal, 34: 707–717. Hoffman, A. J. 2000. Competitive environmental strategy: A guide to the changing business landscape. Washington, DC: Island Press. Hoffman, A. J. 2001. Linking organizational and fieldlevel analyses. Organization & Environment, 14: 133–252. Holmstrom, B. 1979. Moral hazard and observability. Bell Journal of Economics, 10: 74 –91. Hsiao, C. 1985. Benefits and limitations of panel data. Econometric Reviews, 4: 121–174. Jaffe, A. B., Peterson, S. R., Portney, P. R., & Stavins, R. N. 1995. Environmental regulation and the competitiveness of U.S. manufacturing: What does the evidence tell us? Journal of Economic Literature, 33: 132–163. Jensen, M. C. 1983. Organization theory and methodology. Accounting Review, 56: 109 –121. Jensen, M. C., & Meckling, W. 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3: 305– 360. Jensen, M. C., & Murphy, K. J. 1990. Performance and top management incentives. Journal of Political Economy, 98: 225–264. Kassinis, G., & Vafeas, N. 2002. Corporate boards and outside stakeholders as determinants of environmental litigation. Strategic Management Journal, 23: 399 – 415. Khanna, M., & Damon, L. A. 1999. EPA’s voluntary 33/50 program: Impact on toxic releases and economic performance of firms. Journal of Environmental Economics and Management, 37: 1–25. King, A. A., & Lenox, M. J. 2000. Industry self-regulation without sanctions: The chemical industry’s responsible care program. Academy of Management Journal, 43: 698 –716. King, A. A., & Lenox, M. J. 2001. Lean and green? An empirical examination of the relationship between lean production and environmental performance. Production and Operations Management, 10: 244 – 256. 124 Academy of Management Journal February King, A. A., & Lenox, M. J. 2002. Exploring the locus of profitable pollution reduction. Management Science, 48: 289 –299. King, A. A., & Lenox, M. J. 2004. Prospects for developing absorptive capacity through internal information provision. Strategic Management Journal, 25: 331– 345. Klassen, R. D., & McLaughlin, C. P. 1996. The impact of environmental management on firm performance. Management Science, 42: 1199 –1214. Klassen, R. D., & Whybark, D. C. 1999. The impact of environmental technologies on manufacturing performance. Academy of Management Journal, 42: 599 – 615. Lambert, R. A., Larcker, D. F., & Weigelt, K. 1993. The structure of organizational incentives. Administrative Science Quarterly, 38: 438 – 461. Larraza-Kintana, M., Wiseman, R. M., Gomez-Mejia, L. R., & Welbourne, T. M. 2007. Disentangling compensation and employment risks using the behavioral agency model. Strategic Management Journal, 28: 1001–1019. Lorsch, J. W., & MacIver, E. 1989. Pawns or potentates: The reality of America’s corporate boards. Boston: Harvard University Graduate School of Business Administration. Lubatkin, M. H., Lane, P. J., Collin, S., & Very, P. 2007. An embeddedness framing of governance and opportunism: Towards a crossnationally accommodating theory of agency. Journal of Organizational Behaviour, 28: 43–58. Luoma, P., & Goodstein, J. 1999. Stakeholders and corporate boards: Institutional influences on board composition and structure. Academy of Management Journal, 42: 553–563. Makri, M., Lane, P. J., & Gomez-Mejia, L. R. 2006. CEO incentives, innovation, and performance in technology-intensive firms: A reconciliation of outcome and behavior-based incentive schemes. Strategic Management Journal, 27: 1057–1080. McConaughy, D. L. 2000. Family CEOs vs. non family CEOs in the family-controlled firm: An examination of the level and sensitivity of pay to performance. Family Business Review, 13: 121–131. McEachern, W. A. 1975. Management control and performance. Lexington, MA: Heath. Meyer, J., & Rowan, B. 1977. Institutional organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83: 340 –363. Miller, J. S., Wiseman, R. M., & Gomez-Mejia, L. R. 2002. The fit between CEO compensation design and firm risk. Academy of Management Journal, 45: 745– 756. Mizruchi, M. S. 1983. Who controls whom? An examination of the relation between management and boards of directors in large American corporations. Academy of Management Review, 8: 426 – 437. Murphy, K. J. 1999. Executive compensation. In O. Ashenfelter & D. Card (Eds.), Handbook of labor economics, vol. 3: 2485–2563. Amsterdam: North Holland. Oliver, C. 1991. Strategic responses to institutional processes. Academy of Management Review, 16: 145–179. Oliver, C. 1997. Sustainable competitive advantage: Combining institutional and resource-based views. Strategic Management Journal, 18: 697–713. Peng, M. W. 2004. Outside directors and firm performance during institutional transitions. Strategic Management Journal, 25: 453– 471. Porter, M. E., & van der Linde, C. 1995. Green and competitive: Ending the stalemate. Harvard Business Review, 73(5): 120 –134. Rawwas, M. Y. A., Swaidan, Z., & Oyman, M. 2005. Consumer ethics: A cross-cultural study of the ethical beliefs of Turkish and American consumers. Journal of Business Ethics, 57: 183–195. Robinson, P. M. 1982. On the asymptotic properties of estimators of models containing limited dependent variables. Econometrica, 50: 27– 41. Russo, M. V., & Fouts, P. A. 1997. A resource-based perspective on corporate environmental performance and profitability. Academy of Management Journal, 40: 534 –559. Russo, M. V., & Harrison, N. S. 2005. Organizational design and environmental performance: Clues from the electronics industry. Academy of Management Journal, 48: 582–593. Sanders, W. G., & Carpenter, M. A. 1998. Internationalization and firm governance: The roles of CEO compensation, top team composition, and board structure. Academy of Management Journal, 41: 158 –178. Sanders, W. G., & Hambrick, D. C. 2007. Swinging for the fences: The effects of CEO stock options on company risk taking and performance. Academy of Management Journal, 50: 1055–1078. Sarkis, J., & Cordeiro, J. J. 2001. An empirical evaluation of environmental efficiencies and firm performance: Pollution prevention versus end-of-pipe practice. European Journal of Operational Research, 135: 102–113. Schmidt, F., & Hunter, J. E. 1995. The impact of dataanalysis methods on cumulative research knowledge. Statistical significance testing, confidence intervals and meta-analysis. Evaluation & the Health Professions, 18: 408 – 427. Scott, W. R. 1995. Institutions and organizations. Thousand Oaks, CA: Sage. Scott, W. R. 2005. Institutional theory: Contributing to a 2009 Berrone and Gomez-Mejia 125 theoretical research program. In K. G. Smith & M. A. Hitt (Eds.), Great minds in management: The process of theory development: 460 – 485. London: Oxford University Press. Shameek, K., & Cohen, M. A. 2000. Why do firms pollute (and reduce) toxic emissions? Working paper, Vanderbilt University, Nashville. Sharma, S. 2000. Managerial interpretations and organizational context as predictors of corporate choice of environmental strategy. Academy of Management Journal, 43: 581–597. Sharma, S., & Henriques, I. 2005. Stakeholder influences on sustainability practices in the Canadian forest products industry. Strategic Management Journal, 26: 159 –180. Sharma, S., & Vredenburg, H. 1998. Proactive corporate environmental strategy and the development of competitively valuable organizational capabilities. Strategic Management Journal, 19: 729 –753. Sherer, P. D., & Lee, K. 2002. Institutional change in large law firms: A resource dependency and institutional perspective. Academy of Management Journal, 43: 102–119. Shrivastava, P. 1995a. The role of corporations in achieving ecological sustainability. Academy of Management Review, 20: 936 –960. Shrivastava, P. 1995b. Ecocentric management for a risk society. Academy of Management Review, 20: 118 –137. Singh, J. V., Tucker, D. J., & House, R. J. 1984. Organizational legitimacy and the liability of newness. Administrative Science Quarterly, 31: 171–193. Smarzynska, B. K., & Wei, S.-J. 2001. Pollution haven and foreign direct investment: Dirty secret or popular myth? Working paper series, 1–31, Washington, DC: World Bank. Stanwick, P. A., & Stanwick, S. D. 2001. CEO compensation: Does it pay to be green? Business Strategy and the Environment, 10(3): 176 –182. Staw, B. M., & Epstein, L. D. 2000. What bandwagons bring: Effects of popular management techniques on corporate performance, reputation, and CEO pay. Administrative Science Quarterly, 45: 523–556. Suchman, M. C. 1995. Managing legitimacy: Strategic and institutional approaches. Academy of Management Review, 20: 571– 610. Toffel, M. W., & Marshall, J. D. 2004. Improving environmental performance assessment: Comparative analysis of weighting methods used to evaluate chemical release inventories. Journal of Industrial Ecology, 8(1–2): 143–172. Tosi, H. L., Werner, S., Katz, J., & Gomez-Mejia, L. R. 2000. How much does performance matter? A metaanalysis of CEO pay studies. Journal of Management, 26: 301–339. Turban, D. B., & Greening, D. W. 1997. Corporate social performance and organizational attractiveness to prospective employees. Academy of Management Journal, 40: 658 – 672. Walley, N., & Whitehead, B. 1994. It’s not easy being green. Harvard Business Review, 72(3): 46 –52. Westphal, J. D., & Zajac, E. J. 1994. Substance and symbolism in CEOs’ long-term incentive plans. Administrative Science Quarterly, 39: 367–390. White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, 48: 817– 830. Wiseman, R. M., Cuevas-Rodrıguez, G., & Gomez-Mejia, ´ L. R. 2008. A social agency model. Michigan State University, Department of Management working papers series: 1– 47. Wiseman, R. M., & Gomez-Mejia, L. R. 1998. A behavioral agency model of managerial risk taking. Academy of Management Review, 22: 133–153. Wooldridge, J. M. 2002. Econometric analysis of cross section and panel data. Cambridge, MA: MIT Press. Ziliak, S. T., & McCloskey, D. N. 2004. Significance redux. Journal of Socio-Economics, 33: 665– 675. APPENDIX A Description of Sample TABLE A1 Sample Breakdown Number of Firms 90 89 62 62 79 87 469 Years of Data 2 3 4 5 6 7 Firm-Level Observations 180 267 248 310 474 609 2,088 APPENDIX B Calculation Procedures 1. Calculation of Weighted Waste Score Formally, we obtained a weighted waste score for each facility using the following formula: ww jt k l E klt f kl , (1) where Eklt is the emissions of chemical l to medium k in year t by facility j; and is the weighting factor corresponding to chemical l emitted to medium k. 126 Academy of Management Journal February 2. Calculation of Pollution Prevention Measure For firm i, we could formally state the pollution prevention measure as follows,a PP it where predicted_waste it waste it 1 j 1 assets of the firm) and then summed them at the firm level. The final measure we obtained for each firm can be formally represented with the following equation: n predicted_waste it 1 waste it 1 (2) Rj 1 Aj , AT (5) ww jt ww jt 1 , j PR jt 1 , (3) (4) 1 and j are the facilities that belong to firm i. If actual waste level is lower than the predicted level, Equation 2 would yield positive values, evidencing reduction of waste generation. Thus, bigger values are associated with better pollution prevention performance. 3. Industry Pollution Position Index The index industry pollution position uses data from the Compustat Business Segment database. For each firm, we gathered data on all business segments that a company operated and identified each segment’s SIC code. We then ranked each segment according to its “dirtiness,” established by ranking industries (categorized by two-digit SIC code) according to their total amount of toxic emissions, from the most to the least polluting sector. We later weighted the rank of each business segment by its economic importance to the firm (identifiable total assets of each segment divided by total We are thankful to an anonymous reviewer for suggesting this method. To gain confidence in this measure, we compared our pollution prevention measure with the log of on-site releases, and they were correlated at .52. We also calculated the inverse of weighted on-site emissions divided by firm sales, and using this measure in our estimations did not change our findings substantially. Results of estimations using this latter variable are available from the authors upon request. a where Rj is the pollution rank position of segment j, Aj is the total identifiable assets of segment j, AT is the total assets of the company, and n is the total number of segments of the firm. As a result, this measure accounts for the firm’s industry composition, its dirtiness, and the relative importance of each segment in which the firm operates. Pascual Berrone (pberrone@iese.edu) is an assistant professor of strategic management at IESE Business School. He received his Ph.D. in business administration and quantitative methods from Universidad Carlos III. His current research interests focus on various aspects of the interface between corporate governance and corporate social responsibility, with particular emphasis on the impact of incentive schemes on the environmental performance of firms. His research also examines sustainable innovation, family firms, and related corporate governance issues. Luis R. Gomez-Mejia (luis.gomez-mejia@asu.edu) is a Council of 100 Distinguished Scholar at Arizona State University (ASU) and holds the Horace Steel Chair there. He is a Regent’s Professor at ASU and has recently received the title of doctor honoris causa from Universidad Carlos III. He received his Ph.D. from the University of Minnesota. His current research interests are family business and macro compensation issues, including risk taking and executive compensation design.