firm resources, quality signals and the determinants of ... - SSRN papers

British Accounting Review (2002) 34, 257ÿ282 doi:10.1006/bare.2002.0211, available online at http://www.idealibrary.com on

1

FIRM RESOURCES, QUALITY SIGNALS AND THE DETERMINANTS OF CORPORATE ENVIRONMENTAL REPUTATION: SOME UK EVIDENCE J. S. TOMS

University of Nottingham Using the example of the creation of corporate environmental reputation, the article offers a theoretical extension of the resource-based view of the ®rm to include quality signalling via the channel of accounting disclosure. The proposed framework is then tested via an empirical survey into the relationship between environmental disclosure and environmental reputation. The results suggest that implementation, monitoring and disclosure of environmental policies and their disclosure in annual reports contribute signi®cantly to the creation of environmental reputation. Diverse institutional share ownership and low systematic risk are also associated with positive environmental reputation. Prior ®nancial performance has no impact and there is no evidence that environmental reputation is created by a ®nancial halo effect or by the availability of slack ®nancial resources. # 2002 Elsevier Science Ltd. All rights reserved.

INTRODUCTION The article is concerned with how the annual report might promote the creation and management of environmental reputation. It examines and tests two general propositions. First that the creation of reputation based Financial assistance from the Chartered Institute of Management Accountants is gratefully acknowledged. I am grateful to the editors and to two anonymous reviewers for their constructive comments. I would like to thank Professor David Owen, and also the participants at the Fourth Financial Reporting and Business Communication Conference in Cardiff, in particular Professor Mike Jones, and a seminar held at Nottingham University Business School for their constructive comments. I would also like to thank Mike Brown for allowing me access to the MAC data. Please address all correspondence to: J. S. Toms, Business School, University of Nottingham, Jubilee Campus, Wollaton Rd, Nottingham NG8 1BB, UK. Tel: 44 115-951 5276; Fax: 44 115-956 6667; email: [email protected] Received September 2000; revised May 2002; accepted July 2002 0890±8389/02/$ÿÿsee front matter

#

2002 Elsevier Science Ltd. All rights reserved.

Electronic copy available at: http://ssrn.com/abstract=1714030

258

J. S. TOMS

intangible assets depends on investment in dif®cult to imitate projects and second that such investment will promote the use of the annual report as a quality signalling device. These relationships have not been extensively explored within the accounting literature, or in the ®eld of corporate social responsibility. Nonetheless much recent US based research has suggested that the resource-based view (RBV) of the ®rm provides an appropriate basis for investigating how ®rms create social and environmental reputation (Hart, 1995; Litz, 1996; Porter & Van Der Linde, 1996; Russo & Fouts, 1997; Waddock & Graves, 1997). The main theoretical contention of this article is that the RBV might be extended to consider the role of accounting disclosure as a signal of improved social and environmental conduct and hence reputation in those ®elds.1 Similarly the RBV might be used to extend the scope of positive accounting research in this area. To investigate the relationship between investment in intangible resources, voluntary environmental disclosures and the creation of corporate social reputation (CSR), the article synthesises propositions from the RBV and positive accounting theory. Hence signalling is a possible response to market failure where there is information asymmetry (Watts & Zimmerman, 1986, pp. 163±166). Signalling and agency theories are consistent with each other (Morris, 1987), and useful for explaining voluntary disclosure (Holthausen & Leftwich, 1983; Morris, 1987). An important motivation for the study is that besides producing empirically ambiguous results, investigations into environmental performance relationships have typically been under-theorised (Ulmann, 1985). The article therefore begins with an examination of the RBV and its relationship to ®nance and accounting based theories of signalling. The remainder of the article seeks to develop and test an empirical model. In doing so, as a corollary of the theoretical framework, it offers a qualitative variant on content analysis methodology. A speci®c test of this relationship is then conducted by examining corporate environmental management, as communicated via accounting disclosures, and corporate environmental performance as measured by environmental reputation. The implications of the model for empirical testing are then discussed. A further section describes the empirical model. Test results are then analysed. In the ®nal section conclusions are drawn. RESOURCES, REPUTATION AND REPUTATION SIGNALLING: A THEORETICAL FRAMEWORK This section develops a framework to explain management actions, how intangibles are created and why it is necessary to disclose information about them. Speci®cally intangible asset creation occurs through enhanced reputation and disclosure in¯uences the external perception of reputation. To examine these relationships it is necessary to offer a synthesis of three

Electronic copy available at: http://ssrn.com/abstract=1714030

UK EVIDENCE OF CORPORATE ENVIRONMENTAL REPUTATION

259

areas of theory. These are a theory of the ®rm (the RBV), aspects of corporate governance, particularly the role played by shareholders, and their interpretation of signals about managerial activity provided in the annual report. Four basic propositions follow from this. First, that RBV relies on ®rm speci®c, inimitable qualities. Second that the pattern of share ownership imposes heterogeneous pressures on managers. Third that managers will rely on quality signalling to respond to those pressures. Finally that accounting disclosure is potentially an important channel for the transmission of quality signals. Each is discussed in turn, with an emphasis on the crucial linkages. Reputation is arguably the most important of intangible assets. Such an interpretation is consistent with the RBV that explains competitive advantage in terms of inimitable qualities, including those encapsulated in reputation. The RBV as articulated by Barney (1991) is applicable to reputation in general (Caves & Porter, 1977; Hall, 1992). More recently it has been applied speci®cally to environmental management and competitive advantage (Hart, 1995; Litz, 1996; Porter & Van Der Linde, 1996; Russo & Fouts, 1997; Waddock & Graves, 1997). Physical resources can be easily acquired by competitors (Barney, 1991), whereas tacit or team based socially complex assets are less replicable (Hart, 1995). Thus whilst large ®rms exhaust the competitive advantages of scale economies, economies of scope remain exploitable (Chandler, 1990) particularly through the creation of intangible assets. Intangible reputation-based assets are becoming more important contributors to ®rm performance and such resources are more likely to be valuable and inimitable when society demands a cleaner environment (Russo & Fouts, 1997, p. 537). Empirical research suggests that improving the ®rm's reputation creates goodwill and ultimately has a positive effect on market value (Chauvin & Hirschey, 1994; Fombrun, 1996). Whilst there may be good internal reasons for management to pursue positive environmental programmes, there may also be externally driven pressures to which they must respond, particularly when operating in more environmentally sensitive sectors. These may be a function of managerial abilities and opportunities presented by the competitive environment (Barney, 1991; Conner, 1991), although there has been little discussion in this context of the constraints imposed on management by the system of governance. Notwithstanding managerial resource endowment, pressure from shareholders may affect managerial capacity to develop new strategies. This may be a function of the manager±shareholder agency relationship ( Jensen, 1986), or of managerial (in) abilities to act as the strategic situation dictates (Barker & Duhaime, 1997). Such a view recognises the moderating impact of corporate governance pressures, mainly applied to analysis of ®nancial and portfolio restructuring strategies (Bethel & Liebeskind, 1983; Gibbs, 1993; Hoskisson et al., 1994; Johnson, 1996), rather than reputation building strategies. There is also evidence that risk-averse institutional

260

J. S. TOMS

investors are likely to respond positively to improving social performance (Graves & Waddock, 1994). Meanwhile the impact of governance structures on environmental strategy has been explored in the stakeholder literature (Roberts, 1992) but without fully engaging governance based theories of accounting disclosure (Watts & Zimmerman, 1986) and quality signalling (Akerlof, 1970). Corporate governance issues arise from agency problems associated with the separation of ownership and control and incomplete contracts that cannot specify all possible future outcomes (Schleifer & Vishny, 1997). Therefore principals rely on either direct pressure on agents (voice) or the threat of ending their association with the company (exit) (Hirschman, 1970). According to one view of corporate governance, the availability of a liquid share market is a suf®cient condition for shareholders to discipline managers. In such conditions, if shareholders are dissatis®ed they simply exit the market by selling their shares. It is unlikely, however that these circumstances will be conducive to extensive voluntary environmental disclosure, since value or proprietary disclosures are likely to dominate decision sets of investors attempting to manage portfolios with minimum monitoring cost. Empirical research con®rms that investors usually see CSR disclosures as important or useful, but less so than the main ®nancial variables (Benjamin & Stanga, 1977; Firth, 1978; Harte et al., 1991; Milne & Chan, 1998). In other words, environmental disclosures are more likely where investors proactively and continually monitor using voice based governance mechanisms. Some commentators have suggested that institutional investors have failed to perform such a role in recent decades (Hutton, 1995). However, increased ownership in the shares of UK companies by institutional investors at the expense of families and management insiders (Scott, 1997) accompanied rapid increases in the level of disclosure during the 1980s and 1990s (Gray et al., 1995; Toms, 2000). It is possible that these trends were related. Theoretical and empirical evidence from the USA suggests that where shareholding becomes more diverse, re¯ecting a broader investor base there is greater pressure on managers to disclose social responsibility issues (Ullmann, 1985; Roberts, 1992). In the 1990s in particular there is evidence that British institutional investors were more active monitors of their investee companies (Black & Coffee, 1994; Holland, 1998; 2001). Finally agency theory ( Jensen & Meckling, 1976) predicts that where managers and other insiders own fewer blocks of shares managers have less incentive to act in the interests of outside investors and that accounting disclosure is likely to increase. An important contention of this article is that accounting disclosure, incorporating the signalling hypothesis, forms a crucial link between the resource and governance perspectives. There has been little development of this economic perspective in CSR research (but see Ness & Mirza, 1991; Mak, 1991), precisely because its positivist tenets are contested by leading researchers into corporate social responsibility (Gray et al., 1995, p. 51).


261

Nonetheless signalling theory is important since it potentially draws together separate strands of empirical work; those concerned with disclosure (Spicer, 1978; Jaggi & Freedman, 1992) and the resource-based and governance perspectives outlined earlier. This is because there are adverse selection and moral hazard problems where management engages in risky but undisclosed activities (for example under-investing in clean technology). In other words, the investor perceives the risk (and therefore risk adjusted expected return) to be similar across ®rms, whereas it is actually different. This may be because managers do not share certain information (adverse selection), or because some of their actions are unobservable (moral hazard). Under these conditions, according to the so-called `signalling hypothesis' managers have an incentive to signal if there are higher returns to disclosing companies (Watts & Zimmerman, 1986, pp. 164±165). This argument originally derives from the problem of information asymmetry identi®ed by Akerlof (1970). In certain cases this can lead to full disclosure equilibrium (Grossman, 1981). However, this will not be achieved where there are proprietary costs of disclosure (Dye, 1985). Where costs are proprietary, signalling should in theory be a strong predictor of managerial behaviour. The signalling hypothesis suggests that where certain conditions are ful®lled, true signals will be believed and false ones rejected. These are that management must have an appropriate incentive to disclose, the signal must be dif®cult to imitate, for example by a competitor ®rm attempting to make the same claims, there must be an observable relationship, and ®nally it must be cost effective. These ideas follow from Spence (1973) and their subsequent extension and application to the dividend decision as a signalling device (Ross, 1977; Bhattacharya, 1979). They also have a high degree of congruence with key aspects of RBV, especially the assumption that the ®rm possesses inimitable assets. So, if management have made investments in reputation building activities, then clearly they have appropriate and compelling incentives to inform their capital market monitors, although this will be moderated by variations in governance structures. Further, if claims about reputation building activities are at risk of contradiction by subsequent experience then management is also likely to be dissuaded from the use of rhetoric and other low cost unsubstantiated disclosures. In terms of environmental disclosures, it follows that speci®ed, quanti®able and veri®able information will be perceived to be of higher quality. A ®rm whose management is not actually pursuing environmentally responsible policies would ®nd it more dif®cult to imitate a genuine competitor if that ®rm followed such a quality signalling strategy. Meanwhile as suggested earlier, the RBV literature when applied to intangible assets equates reputation with credibility, trustworthiness, reliability and responsibility (Fombrun, 1996). These factors, particularly in the form of dif®culty of imitation provide a crucial link with RBV. At the same time, awareness of disclosure options among outsiders is a vital part of the disclosure expansion process. This is especially true where

262

J. S. TOMS

the public knows the information is available to the company and that its disclosure is unlikely to damage its competitive position (Horngren, 1957). As with RBV, these arguments can be directly related to the mediating in¯uences of governance structure. The signalling hypothesis depends on the relationship between managerial agents and investors. For example, larger, institutionally controlled ®rms might be expected to face greater scrutiny from analysts and fund managers. These accountability processes are similar to those suggesting an incentive structure for ®rms to submit themselves voluntarily to external audit (Jensen & Meckling, 1976), and now this may also include specialised environmental audits. The obvious place for signalling disclosures is the annual report. However other signalling devices are available, for example quality kite-marks, as are channels to transmit the signal, for example press releases. There are thus many permutations of device and channel. In selecting them, managers must bear in mind the chief elements of the above framework; the inimitable resources possessed by the ®rm and the optimum method of convincingly transmitting the details of intangible asset investments to a potentially sceptical capital market. The relationships described above are summarised in Figure 1. At stage 1 managers take action to implement environmental policy. Where nonmandatory, it is assumed that their motivation is to create ®rm-speci®c Stage 1 Management environmental strategy

Stage 2 Disclosure strategy: Signal and channel

Mediating variables: * Shareownership * Financial resources * Risk * Firm size

Stage 3 Environmental performance/ Reputation Asset

Figure 1. Development of corporate environmental reputation


263

reputation assets that in accordance with RBV and stakeholder theory will create competitive advantage (Russo & Fouts, 1997). It is possible that such strategies simultaneously allow managers to satisfy personal interests (McWilliams & Siegel, 2001, p. 119). Stage 2 assumes that the value of investment in intangibles cannot be realised unless signalled to stakeholders. This might be achieved through advertising (McWilliams & Siegel, 2001), but also through the annual report. Indeed the annual report is a central corporate document, which speaks about the organisation as a whole (Gray et al., 2001; p. 350). Stage 3 is the outcome of the signalling process in terms of reputation. Whilst reputation will be in¯uenced by other factors, such as the impact of unforeseen events, the assumption is that reputation can be created and managed through the disclosure process. The value of the intangible reputation asset will be a function of uniqueness and dif®culty of imitation. Thus if managers are able to create inimitable assets at stage 1, they will be able to make disclosures at stage 2 which competitors will also ®nd impossible to imitate without making false declarations. The process is mediated by investor expectations. It is assumed that some, although not necessarily all pressures from lobby groups and regulators can be transmitted via the stock market and can therefore affect stock market value. Meanwhile, to pursue investment in environmental projects, managers require access to ®nancial resources and this is governed by investor expectations. Active shareholders are more likely to promote this process. In addition, political pressures may directly in¯uence managerial strategy either through ideology and/or rational ex ante pro®t calculations and may also in¯uence environmental performance and reputation through changes in consumer demand etc. These indirect pressures suggest that the pro®tability of the industry in which the ®rm operates, its products and processes, and its size, though diversity and exposure to lobbying, will mediate its exposure to environmental risks and opportunities. Investment in environmental projects may be seen as risk reducing and therefore the level of general risk to which the ®rm is exposed may increase the pressure on managers to adopt such projects. Finally, active shareholders may demand investment in reputational assets either as risk reduction strategies or from recognition that it may bring superior returns through competitive advantage.

ACCOUNTING DISCLOSURE AND CORPORATE ENVIRONMENTAL REPUTATION Reputation measures To measure reputation, this study uses the corporate ratings for community and environmental responsibility (CER) as published in the Management Today survey of Britain's Most Admired Companies (MAC). In both the

264

J. S. TOMS

USA and the UK this is a potentially useful measure of social and environmental reputation. Notwithstanding its availability, this measure has not been used in US surveys of the social disclosure and social performance relationship (Ullman, 1985, table 2). MAC ratings nonetheless offer the option to survey a wide range of industries. An apparent weakness of the MAC survey methodology is that it may confound the true factors underlying superior performance. However, there is some evidence that CER behaves differently to the other management quality and ®nancial performance variables in the MAC data (Fryxell & Wang, 1994), which otherwise demonstrate a high degree of co-linearity. For the purposes of the current study, the advantages of the CER measure outweigh the disadvantages of other indices. For example, some studies measure the actual environmental performance and reputation of US corporations by reference to Council on Economic Priorities (CEP) ratings (Freedman & Jaggi, 1982; Wiseman, 1982). However, in the UK there is no equivalent system of rating polluting activities by corporations. Another disadvantage of CEP style measures is that they are con®ned to measuring pollution, rather than the wider aspects of environmental management and therefore only relevant to certain industries. Notwithstanding the availability of CEP data, some US studies have nonetheless used other variables. Reputation based measures are principal among these (Fry & Hock, 1976; Preston, 1978; Abbot & Monsen, 1979). However, these are typically not available for the analysis of a wide cross section of UK companies. Disclosure measures It follows from the earlier theoretical discussion that volume of disclosure alone is not in itself a suf®cient condition for the creation of reputation. In most studies, disclosure is measured by volume-based content analysis (Abbot & Monsen, 1979; Ingram & Frazier, 1980; Wiseman, 1982). Nonetheless, there has been recognition that reliance on mere number of disclosures may be misleading (Cowen et al., 1987, p. 121). In the UK, where environmental disclosures are subject to less regulation than in the US, it might be expected that voluntary disclosures by companies might be more self-congratulatory and be less reliable. Recent studies (e.g. Gray et al., 1995) have begun to combine details of disclosures with theoretical underpinnings (Mathews, 1997, p. 496), and this can be taken further using the quality signalling RBV framework. None of the research discussed in the previous section suggests that quality signalling depends on the volume of information. Rather it is the credibility of the signal that is important. Some empirical studies recognise this. For example, institutional investors took information into account if it was quanti®ed and focused on speci®c issues and obtained from disinterested parties (Teoh & Shiu, 1990). Similarly, Robertson & Nicholson (1996) suggest that there is a hierarchy of environmental commitment. To test this view further, a pilot questionnaire

265


survey was sent to fund managers and analysts.2 The response distribution and details of mean scores are set out in Table 1. This shows a hierarchy of importance from the low rating given to `non-quanti®ed information' to the high rating for èxternally monitored environmental report'. The hierarchy was much more pronounced for ethical fund managers. Non-quanti®ed information has a relatively low value scoring signi®cantly worse than all other disclosure strategies to which responses were invited. Quantity of information appears to add value to environmental disclosures where nonquanti®ed data is used. Speci®cation of policies was signi®cantly more important than non-quanti®ed data. It was also more important than a general statement of policy, although not signi®cantly so. The use of external monitoring sits at the top of a hierarchy of qualitative disclosure strategies with a signi®cantly higher average score than other categories. Their responses suggested a hierarchy of disclosures, along the lines suggested by Robertson and Nicholson (1996) each of which increased the credibility of the information. This is set out in Table 2. The hierarchy suggested in Tables 1 and 2 con®rms and develops the views from prior studies referred to above and is symmetrical with the concept of signalling investments in inimitable resources suggested in the preceding theoretical discussion. As can be seen, the hierarchy penalises rhetoric and rewards the quanti®able and the veri®able. These varying degrees of disclosure form the basis for empirical testing below. It is possible that the allocation of scores using Table 2 could unbalance the results if ®rms are accredited with a high quality disclosure score on the basis of activities that relate to a relatively insigni®cant part of their business. However, following from the earlier discussion on signalling theory, it seems reasonable to suppose at this stage that where ®rms have made genuine and signi®cant environmental investments, management is more likely to offer TABLE 1 Investment professionals' perceived of importance of qualitative environmental disclosures Disclosure type

1. Externally monitored environmental report 2. Quanti®ed environmental performance in annual reports 3. Speci®ed policies 4. Publication of an environmental policy 5. Volume of information available in reports 6. Non-quanti®ed information

Mean

% response distribution 5

4

3

2

1

40351 36964

28 21

57 53

9 11

2 7

4 8

37368 36842 34561 31404

21 21 21 8

43 43 30 28

26 23 28 45

8 9 13 11

2 4 8 8

Respondents (N 59) asked to rate relative importance of each category on a 1 not important to 5 very important scale. Mean differences within the hierarchy of disclosures all signi®cant at p , 01 using Mann±Whitney±Wilcoxon, asymptotic two-tailed signi®cance.

266

J. S. TOMS

TABLE 2 Rating method for corporate environmental disclosures Rating scale/environmental policy 0. No disclosure 1. General rhetoric

2. Speci®c endeavour; policy only

3. Speci®c endeavour; policy speci®ed 4. Implementation and monitoring; use of targets, results not published 5. Implementation and monitoring; use of targets, results published

Typical example Reuters is an environmentally aware organisation. Reuter journalists cover environmental issues as part of their regular activity (Reuters). Ìt is the Group's policy . . . to minimise the potential for adverse effects of its operations through systematic and rigorous application of high standards and practices' (BOC). Òur policy is to comply with all relevant environmental standards . . . We regularly recycle spent radioactive sources . . . ' (Amersham International). To establish baseline data and help set targets and objectives, working parties have been set up to study our position . . . ' (Boots). Òur capital investment on pollution abatement in 1993 was about £250 m, compared with £350 m in 1992. In 1994 and 1995 we expect the sums to be about £290 m and £250 m respectively' (BP).

Source: Sample company annual reports, 1993. For further examples, see Robertson and Nicholson (1996). p. 1101.

the strongest possible quality signal. Several prominent cases in the USA suggest that the reputation of the whole company will suffer even though environmental litigation is in respect of relatively small subsidiary companies. Conversely, it might be expected that the exemplary behaviour of one business unit would re¯ect favourably on the whole organisation. If the investment is material, there is little point in engaging in mere rhetoric. Also, even if investments are immaterial to larger ®rms, they may still be of suf®cient extent to constitute an entry barrier to smaller ®rms. If ®rms do however remain mis-classi®ed, it is unlikely that such errors will be nonsystematic across the sample. Bearing these issues in mind, similar reasoning underpinned the allocation of scores to other variables. Variation of disclosure and channel In addition to the hierarchy in Tables 1 and 2, ®rms may send quality signals to the markets by other means. For example, they may use a separate environmental report to underline their commitment to reputation building initiatives. As with the annual report however, management teams may


267

choose to ®ll such reports with unveri®able rhetoric at relatively low cost. Any information thus disclosed is therefore likely to add credibility only if the information is veri®able and has not been disclosed elsewhere. Credibility can be further underpinned by voluntarily submitting the corporation's activities to environmental audit. Again this signals to the market that steps have been taken to verify the claims made via public disclosure channels. Kite-mark accreditation is another mechanism that ®rms can use to signal environmental initiatives. These are costly to acquire, dif®cult to imitate and management would clearly have appropriate incentive to signal their acquisition to the market. Indeed studies have shown that ISO 14000 is designed to enhance competitive advantage (Miles & Russell, 1997; Miles et al., 1997). Mediating in¯uences An important contention in the earlier theoretical discussion was that governance structures would play a mediating role in the creation of corporate reputation. In general, agency costs increase with outside capital therefore make disclosure more likely (Jensen & Meckling, 1976; Leftwich et al., 1981). More speci®cally, the annual report is the main source of information for smaller shareholders, who otherwise cannot monitor cheaply (Raffournier, 1995). Some empirical surveys have suggested that management are more responsive where shareholdings are dispersed, since ethical investors and ethical funds are more likely to have input into the decision making process (Keim, 1978; Roberts, 1992). It has also been suggested that institutional investors value CSR information provided by management and are therefore likely to increase their holdings in these ®rms (Graves & Waddock, 1994). As suggested earlier, there is evidence that UK fund managers have become more proactively involved in the monitoring and managerial incentive processes of corporate governance (Black & Coffee, 1994; Holland, 1998, 2001). For these reasons a proxy variable for institutional shareholder power was used in the empirical study described below. Previous empirical surveys have incorporated other determinants of reputation, which may mediate the in¯uence of disclosure. Firm size and industry grouping provide a proxy for the degree of pressure. Larger ®rms are usually subject to greater public scrutiny (Fombrun & Shanley, 1990). Low systematic risk and high ®nancial performance may also predispose monitors to regard ®rms more favourably (Fombrun & Shanley, 1990). In another survey, prior ®nancial performance is associated with subsequent perceptions of ®rm quality reported in the Fortune MAC results (McGuire et al., 1990). Conversely, it has been suggested that management quality in¯uences subsequent ®nancial performance (see McGuire et al., 1990 for a review). Finally, CSR has been related to both prior and future ®nancial

268

J. S. TOMS

performance (Waddock & Graves, 1997). It has been noted that good environmental practice is in any case confounded with good management per se and is thus impossible to separate from a `halo' effect (McGuire et al., 1988; Brown & Perry, 1994). These empirical studies have concentrated on the creation of overall reputation rather than speci®cally environmental reputation, and hence have not necessarily controlled for industry membership. In this case, following the literature on environmental disclosure (Patten, 1991; Trotman & Bradley, 1981; Wiseman, 1982; Cowen et al., 1987; Adams et al., 1998), an industry grouping variable would seem appropriate in order to capture the differing pro®les of ®rms involved in environmentally sensitive and non-sensitive activities. DETERMINANTS OF CORPORATE ENVIRONMENTAL REPUTATION: AN EMPIRICAL SURVEY In the light of the above discussion, an empirical survey of the reputation and disclosure relationship was designed using data from large quoted UK ®rms. To be included in the sample a ®rm had to have been included in the UK survey of `Most Admired Companies' for 1996 or 1997. Also the ®rm had to respond positively to a request from the researcher for a copy of its annual report for 1995 and 1996. Firms that were taken over during the period between disclosure and reputation rating were excluded. Certain banking and ®nance companies were also excluded if required ®nancial performance variables were not available from Datastream. This produced a pooled sample of 126 ®rms for 1997 and 89 ®rms for 1996. In the subsequent discussion these variables and the independent variables referred to in model (1) below, are referred to as 1997 and 1996 data respectively. Data were collected for three groups of variables for all the companies in the sample. These were reputation ratings, environmental disclosures, and mediating variables. The objective of this part of the study is to empirically test the model (1) set out below. CERt b0 b1 DISCtÿ1 b2 PSHt b3 KITEtÿ1 b4 ENVRtÿ1 b5 ENVAtÿ1 b6 BETAtÿ1 b7 PROFITt;;tÿ2 b8 SIZEtÿ1 b9 INDt 1 where CER

corporate environmental reputation as measured by the community and environmental responsibility rating for the Management Today survey of `Britain's Most Admired Companies'.


DISC

PSH KITE ENVR ENVA BETA PROFIT SIZE IND

269

disclosure score, ranging from 0 no disclosure to 5 high quality disclosure. Each disclosure level was tested separately and as part of a sub-group of disclosure levels using a dummy variable set to 1 if that level of disclosure was attained, 0 otherwise. power of shareholders; the total percentage of shareholder groups with a stake of >3% plus directors shareholdings. 1, if the company has obtained an environmental quality kitemark; 0, otherwise. 1, if the company publishes a separate environment report; 0, otherwise. 1, if the company has been subject to an environmental audit; 0, otherwise. systematic risk as measured by the company's beta factor. Average return on equity (ROE) for the previous 3 years (t, . . . , t ÿ 2). Company size measured by the natural logarithm of sales turnover. 1, if the company is in a Department of the Environment monitored industry group; 0, otherwise.

Dependent variable For the dependent variable, data was collected from the UK MAC survey for 1996 and 1997. Each annual survey contains all the FTSE100 British companies and, on average, 90% of the top 200 companies by market capitalisation. The sample companies are the largest by market capitalisation from each of 26 sectors. Each year Britain's MAC survey asks senior executives from 260 British companies and senior specialist business analysts to give a rating of the performance of each company, other than their own in the case of executives, within their industrial sector. They provide a score of 0 ( poor) to 10 ( excellent) for each of nine characteristics that impact on the major stakeholders, including CER, the variable of interest for this study. Other variables Disclosures were identi®ed with reference to the content of the whole annual report. Where the Annual Report included a cross-reference to a supplemental environmental report, the separate report was treated as part of the annual report. Environmental disclosures were de®ned as direct or indirect statements about the impact of the company's activities on the environment. Scores were allocated according to the best available example for each company. Only the best example was used to score the signal of each company.3 This is because where disclosure is quanti®ed, imitation is

270

J. S. TOMS

dif®cult where commitment to environmental programmes is not genuine. Where quanti®ed, disclosures are more likely to represent actual activities. Also it is rational for accurate disclosures to be made in such circumstances. In contrast, lower level rhetoric type disclosures can be made without equivalent commitment. Because rhetoric disclosures are cheap, they can potentially be made in large volume, leading to higher scores using standard content analysis. In situations where a company spends cash on environmental projects that management believe will be dif®cult to imitate, rhetoric based disclosures in whatever volume, are unlikely to add further credibility to the higher quality disclosures. Where the rhetoric is imitated by competitors without the same level of commitment, quanti®ed disclosures will represent the important and distinguishing difference between them. Whilst this approach follows from the above theoretical justi®cations, it is possible that the scoring system in Table 2 might misrepresent the overall quality of disclosure. This may well be the case where management discloses quanti®ed and veri®able information concerning a relatively small part of their business but produces inaccurate rhetoric concerning the substantive part of their activities. As noted above, this problem would apply to all studies based on content analysis and such potential therefore must be borne in mind when interpreting any conclusions. Hence a possible direction for future research might be to develop schemes employing weightings in order to capture both the quality and quantity of disclosure. In concentrating only on the former, the present study aims only to complement prior studies that have relied on the latter. Shareholder power is measured using the total percentage controlled by block shareholders. These blocks typically represent family holdings, trusts or other companies. Their absence is a proxy for the collective in¯uence of institutional investors, who typically hold shares below the disclosure threshold, also re¯ecting the low level of individual share ownership in the modern UK economy. As a proxy for membership of environmentally sensitive industries, the level of environmental capital expenditure for that industry was used. According to a Department of the Environment (1996, p. 37) survey 69% of environmental capital expenditure was accounted for by six industries. These were: chemicals (22%), food processing (7%), paper and pulp (8%), minerals processing, taken for the purposes of this study to include building and aggregates, (13%), energy supply, for the purposes of this study including water and all utilities (7%), metals manufacture (7%) and rubber/ plastics (5%). For the purposes of the current study, these industries were de®ned as environmentally sensitive (ES). Given the levels of expenditure, it is also expected that the level of disclosure should also be high for these industries. Each company in the sample was analysed by sales of product and SIC code taken from Datastream. If any of the company's product sales fell under an ES SIC code, ES was set equal to 1 and to 0 otherwise.


271

KITES was given a value of one for any kite-mark obtained by any subsidiary within the group. In all cases these were either BS7750, ISO 14001 or in the ISO9000 quality management series. Where more than one subsidiary had a kite-mark, the group was still scored 1. The existence of ENVR was ascertained by reference to the main annual report of each company. Dealing with separate environmental reports is problematic from a measurement point of view as overlapping or discrete disclosures may be made in either the annual report or the environmental report. It is therefore assumed that the annual report is the main signalling device and separate reports are complements rather than substitutes. This imposes some limitation on the research by excluding cases where companies produce a separate report and make no reference to it in the annual report, in other words where the environmental report is a pure substitute. Hence the research only content analysed the environmental reports where they were cross-referenced and where hence clearly identi®able as an extension of the annual report. ENVA was de®ned as either internal or external audit. Only a small number of ®rms engaged in environmental audits of any kind, the majority being internal. It was therefore inappropriate to segment this variable further between internal and external audit. Information for ROEt, . . . ,t ÿ 2, PSH and SIZEt was obtained from Datastream for each company. BETAt was obtained from the London Business School Risk Measurement Service. IND was determined by reference to the ES of company activities. Table 3 shows the Pearson correlation co-ef®cients for all independent variables using the 1997 data. These correlations indicate that co-linearity is not present. The highest correlation coef®cient is between SIZE and PSH (ÿ0406). The variance in¯ation factors on these variables were low (1357 and 1289, respectively). Heteroscedasticity robust standard errors (White, 1980) were used in all models. DATA ANALYSIS AND DISCUSSION The main objective of the statistical analysis is to test the relationships implicit in model (1). Table 4 sets out four versions of the model using 1997 data that best summarise the relationships. Model 1.1 shows the results including all principal co-ef®cients de®ned above. Model 1.2 is a reduced form version showing only those co-ef®cients that were persistently signi®cant. Model 1.3 highlights the relationship between disclosure and reputation and model 1.4 the impact of the residual signi®cant variables independent of disclosure. In the light of the signi®cant performance of the DISC variable in all relevant models, further tests were conducted to test for the presence of important cut points in the 0±5 scale suggested in Tables 1 and 2 using Duncan and Scheffe post hoc contrasts. This procedure also acted as a further check against inter-rater unreliability.

272

TABLE 3 Descriptive statistics Mean

s.d.

A

B

C

D

E

F

G

H

I

J

A B C D E F G H I J

5500 2246 22087 0039 0103 0047 1006 17797 0523 14314

0788 1500 17834 0196 0305 0213 0193 15986 0501 1303

1000 0389 ÿ0225 0136 0211 0052 ÿ0221 0078 0209 0134

1000 ÿ0195 0184 0293 0113 ÿ0108 0153 0327 0313

1000 ÿ0143 ÿ0085 ÿ0173 ÿ0114 ÿ0042 0021 ÿ0406

1000 0332 0145 ÿ0085 ÿ0008 0112 ÿ0016

1000 0292 ÿ0146 0032 0219 0055

1000 ÿ0047 0180 ÿ0011 0167

1000 ÿ0139 ÿ0157 ÿ0074

1000 0146 0150

1000 0016

100

CER DISC PSH KITES ENVR ENVA BETA PROFIT IND SIZE

J. S. TOMS

Variable

273


TABLE 4 Determinants of community and environmental reputation for UK companies Dependent variable CER score Model (1.1) Independent variable CONSTANT DISC PSH KITES ENVR ENVA BETA PROFIT IND SIZE F R2 N

6624*** (615) 0160*** (317) ÿ 0009* (192) 0211 (0307) 0187 (053) ÿ0145 (049) ÿ0806** (207) 00000 (001) 0105 (065) ÿ0037 (056) 366 0230 126

(1.2) 6138*** (1495) 0173*** (404) ÿ 0008* (186)

(1.3) 5042*** (4214) 0204*** (455)

6775*** (1512)

ÿ 0011*** (241)

ÿ 0842*** (227)

934 0215 126

(1.4)

ÿ 1020*** (246)

2071 0151 126

629 0112 126

Numbers in parentheses are t-statistics based on White's (1980) heteroscedasticity consistent estimation matrix. Signi®cance levels (one-tailed test except intercept terms). *** p , 001. ** p , 005. * p , 01.

Quality of disclosure, institutional shareholder power and low risk are consistently associated with high CER in all models. Prior economic performance, size and industry membership, kitemark awards, environmental reports and environmental audit were consistently insigni®cant in all models tested. Replication of the tests on the 1996 data showed similar results in terms of signi®cant and non-signi®cant variables and co-ef®cient signs. The only difference was that the level of signi®cance for the disclosure variable declined marginally, with industry membership becoming

274

J. S. TOMS

signi®cant. These results are now discussed in detail, ®rst with reference to the disclosure variables then to the mediating variables. The positive and signi®cant co-ef®cient for the disclosure score variable in Table 4 suggests that an improvement in reputation might be generated from a move up the hierarchy of management policy and disclosure. However as discussed earlier, it is recognised that allocations to these groupings might be arbitrary. Further explorations using Duncan's post hoc contrasts suggested that there were two important steps. These were from generalised statements (DISC , 4) to quanti®ed statements (4 or 5) and from no disclosure at all (DISC 0) to some disclosure whether at the level of rhetoric or a generalised environmental policy statement (DISC 1, 2 or 3). This suggests that implementation and reporting of speci®c policies and quanti®able targets adds signi®cantly to reputation. Subsequent reporting of actual outcomes against targets (the additional step required to move from 4 to 5 on the scale) has little marginal bene®t in terms of reputation. However, if policies are to be stated, it is better to be speci®c. Generalised environmental policies do not appear to offer any more help than mere rhetoric in creating reputation. Nonetheless ®rms that change from no disclosure at all to the disclosure of at least an outline environmental policy statement can obtain a substantial increase in reputation. As the results here suggest, even general rhetoric appears to offer some bene®t. In the pilot questionnaire, 43% of all investment professionals in the sample considered environmental disclosure to be a least `fairly important'. Of the sample ®rms in the statistical tests reported above, 79% made at least some disclosure. Gray et al. (1995) show that the major increase in voluntary disclosure took place in the period 1985±1991, when the increase for their sample was from under 10% to over 70%. At the same time there was a change in the character of environmental reporting, in particular to set out objectives and statements of policy (Tonkin & Skerratt, 1991; Harte & Owen, 1991). The signi®cance of the 3, 2 and 1 levels of disclosure indicate that there is perhaps an expectation that some form of environmental disclosure, however limited, is emerging as a minimum standard. Taken together, these results offer support for the signalling/RBV hypothesis. In general, reputation is more likely to be created via signals that are less easy to replicate by competitors. Alternative styles and channels of disclosure appear to offer little direct or incremental assistance in the creation of reputation. This may be attributable to the relatively small numbers of companies using these strategies, in particular the use of kite-marks. Small sub-sample groupings prevented detailed investigation of these variables through model estimation or MANOVA tests. The numbers using separate environmental reports and environmental auditing were somewhat larger, but tended to be industry speci®c, with the majority of ®rms in the ES industries. Even then however, only the environmental report variable was signi®cant in terms of increased reputation.4 Given the relatively crude assignment of dummies and the cross


275

sectional asymmetry within the sample, the poor performance of these variables is not surprising. Additional and more sophisticated measures to capture these signals more accurately are required for future research, although this will have to wait for more widespread adoption. Hence results are consistent with ®ndings elsewhere that kite-marks have yet to gain credibility and acceptance on a wide scale and have not achieved their objectives of improving environmental management practice (Krut & Gleckman, 1997). Also, the ®ndings suggest that auditing does not independently add to reputation. This conclusion is tentative, given the small sample and distortion probably caused here by variation in auditing practice across industry. Finally, a possible reason for the relatively poor performance of the separate environmental report is that it is seen as merely a spin off from the main report and the data usually contained therein. It is increasingly common practice for environmental reports to be published and circulated as part of an annual report package. The mediating variables suggested interesting associations. In particular, the shareholder power variable, PSH, was negative and strongly signi®cant in all models tested. In other words, companies with high reputations for CER tend to also have institutional shareholdings. The results therefore con®rm the importance of governance as a mediating variable, although the tests could be extended. For example they could examine whether managers respond to demands from institutional investors for improved environmental performance or whether poor monitoring leads to the accumulation of spare cash for spending on improved environmental reputation. In view of the lack of signi®cance for the lagged pro®t variable, the former view would appear more plausible. However, further research is required into the precise nature of the relationship. Systematic risk was also negative and signi®cant in all models tested. Thus the higher the risk, the more likely a company is to suffer poor environmental reputation. In other words, there is a potential bene®t to the creation of environmental reputation since this will tend to lower the cost of raising equity capital. Investors holding shares in such companies will require a lower risk adjusted rate of return. This was also the view of the investment professionals who responded to the pilot questionnaire. Industry effects were inconsistent, with a positive signi®cant association using 1996 data but an insigni®cant one for 1997. This suggests that industry membership perhaps has a neutral impact, with ®rms in ES industries having to work harder at creating and maintaining reputation through disclosure activities. The same might be said of size, which was insigni®cant in all models tested. There is no evidence from these results of signi®cant association between prior ®nancial performance and subsequent reputation, con®rming US evidence (Fryxell & Wang, 1994) that the CER variable in the Fortune surveys is the only one that appears to stand outside the ®nancial `halo' effect. However, the absence of signi®cant association between prior ®nancial performance and reputation is surprising in view of UK (Gray et al., 2001) and US (Roberts, 1992) evidence that prior ®nancial

276

J. S. TOMS

performance and disclosure are associated. One possibility is that ®nancial performance and reputation are indirectly related via disclosure. It is also clear from the results in Gray et al. (2001) that the association is time dependent and that therefore the absence of signi®cant results in this survey may be a function of only two years' of data being used. Further research into the longer-run relationship between ®nancial performance and reputation is therefore required. CONCLUSIONS The purpose of the empirical section of this study was to examine the determinants of corporate environmental reputation suggested by the threestage model in Figure 1. Of the three stages the link between stage 1, environmental policy-making and stage 2, disclosure strategy, has not been tested directly. Rather, it was supposed that because it is more dif®cult to make false declarations where there is quanti®cation and the possibility of veri®cation disclosures would resemble actual policy in such circumstances. However, empirical research is required into this relationship. Political pressures via lobbying and regulation and their direct effects on investor expectations, managerial strategy and the impact on corporate environmental reputation were also excluded from the empirical sections of the article. The principle empirical focus was to examine the relationship between disclosure strategy at stage 2 and the creation of environmental reputation at stage 3. The in¯uence of investor expectations on disclosure strategy was also examined through mediating variables. This group of variables incorporated those most commonly used in other studies examining ®rm speci®c in¯uences on disclosure such as pro®tability, size and industry membership (Adams et al., 1998; Gray et al., 2001). Governance and risk variables were also used to test the impact of monitoring and response to signals by equity investors. A wider range of variables (e.g. Roberts, 1992) based on multiple-stakeholder signalling and monitoring might have been used, but this is a matter for further research. The results provide strong support for the relationship between disclosure strategy and environmental reputation. There is some support for the view that accounting disclosure is an important conduit for signalling facts about environmental management. Also, as far as accounting disclosures accurately respond to reality, reputation is governed by the implementation and monitoring of environmental strategies. Because disclosure of performance against quantitative targets is dif®cult to imitate by companies not genuinely committed to environmental good practice, the information quality of such disclosures is high. Institutional share ownership and low systematic risk also tend to improve environmental reputation.5 There is a suggestion of diseconomies of scale and scope in environmental management activities. Actions beyond a certain point do not add signi®cantly to


277

environmental reputation. If corporate managers are implementing and monitoring performance against quantitative targets and are accountable for their actions through publication, auditing and kite-marks for accredited environmental management systems do little to add further credibility. A possible reason for lack of impact of these alternative signals is their limited use. Fewer than 20% of large UK plcs currently publish separate environmental reports and fewer than 10% use audits and kite-marks. Future research may ®nd these variables of greater signi®cance if investors' expectations change and their use becomes more widespread. The association of accounting disclosure with corporate reputation is signi®cant for two important reasons. Shareholder power, disclosure and their consequences for reputation are consistent with a story that suggests environmental policies have proprietary costs and bene®ts recognised via quality signals in a liquid market. However, further work is required to con®rm this view, particularly on the consequences of environmental policies for shareholder value. Second, there are signi®cant implications for further empirical tests. Reputation, as measured by the MAC surveys, is useful as a quantitative scale for measuring the pursuit of environmentally friendly policies by management. The association with disclosure con®rms their potential value as a non-®nancial proxy for environmental performance, although association with actual environmental policies requires further research. There are several caveats to the empirical ®ndings in this research. Environmental risk might as easily be speci®c rather than systematic, although this hypothesis was not tested directly. Also, there is the (currently unresolved) controversy over the role of beta in pricing risk (Fama & French, 1998). Perhaps the most important caveat is that the study has relied on judgmental disclosure scales in lieu of more extensive, volumebased content analysis. However, there are weaknesses with both approaches. The latter may tend to overweight verbosity and environmental rhetoric in relation to the actual activities of the business whilst the method used in this study may tend to place the whole ®rm in a halo where quanti®cation and veri®cation are occurring in only certain areas of activity. Although there is scope for further research using more complex variants of these approaches such as weighting schemes, the results from the current study might be usefully evaluated as complementary to the ®ndings of volume-based content analysis. These objections have also been dealt with in part by the consistency of the scoring system with the theoretical framework developed above and also by extensive testing of alternative groupings in the empirical section. Nonetheless the objection has some validity and should be borne in mind in the following summary of the principal ®ndings. The hypothesis that ®rms use environmental disclosures to signal inimitable ®rm speci®c resources is well supported by the results. Implementation, monitoring and disclosure of environmental policies and

278

J. S. TOMS

their disclosure in annual reports contribute signi®cantly to the creation of environmental reputation. There is a suggestion also that ®rms can improve reputation through making disclosures per se. This may re¯ect the high level of adoption of voluntary environmental disclosures amongst large UK plcs, to the extent that investors now regard them as a required norm. Diverse share ownership is associated with environmental reputation, suggesting that managers are held more closely to account on this issue by professional institutional investors than by insider or other block shareholders. Hence it is highly likely that ownership and capital market structures are important mediators in reputation building strategies. This is particularly likely since high reputation ®rms are also successful in reducing systematic risk and hence successful in gaining access to cheaper capital. Further research, is required however, since although quality disclosures have payback in terms of reputation, the ®nancial payback is more uncertain. To the extent that the ®rm has made investments in inimitable assets, it seems likely that it should be positive. NOTES 1 The word àccounting' is signi®cant in the sense that the article focuses on the annual report as the main channel of disclosure. Disclosures within the annual report measured by the survey data include both narrative and accounting numbers. The narrative may be of a non-accounting nature. Environmental performance is a sub-set of corporate social performance, which forms the context for much of the relevant literature discussed below. 2 The response rate from environmental fund managers was (11 out of 28, or 39%), but lower from conventional funds (5%). A total of 59 responses were received. Given the small population of ethical funds, the sample size here is similar when compared to studies that have concentrated exclusively on ethical funds. For example, Harte et al. (1991, 11 ethical trusts, questionnaire survey); Perks et al. (1992, 14 ethical trusts, review of published investment policies). This data was collected as part of a broader based questionnaire investigating the attitude of investment professionals to environmental policy and disclosure. 3 The ratings were checked for reliability between three coders using a sample of 60 1993 annual reports. The alpha reliability co-ef®cient (Krippendorff, 1980) was 0875, indicating a high level of agreement. 4 A mean difference test between the CER of ®rms using and not using environmental reports showed a signi®cant improvement 0544 (p , 005) for the former group. There were also improvements for ®rms using kitemarks and environmental auditing, but these were not signi®cant. 5 The survey referred to in note 2 above, also found that investment professionals perceived investment in environmental projects to be important in reducing risk.

REFERENCES Abbott, W.F. & Monsen, R.J. (1979). Òn the measurement of corporate social responsibility: Self-reported disclosures as a method of measuring corporate social involvement', Academy of Management Journal, 22(3), pp. 501±515. Adams, C.A., Hill, W.Y. & Roberts, C.B. (1998). `Corporate social reporting practices in western Europe: Legitimating corporate behaviour?' British Accounting Review, 30, pp. 1±21.


279

Akerlof, G.A. (1970). `The market for ``Lemons'': Quality uncertainty and the market mechanism', Quarterly Journal of Economics, 84(3), p. 488±500. Barker, V. & Duhaime, I. (1997). `Strategic change in the turnaround process: theory and empirical evidence', Strategic Management Journal, 18(1), pp. 13±38. Barney, J.B. (1991). `Firm resources and sustained competitive advantage', Journal of Management, 17, pp. 99±120. Benjamin, J. & Stanga, K. (1977). `Difference in disclosure needs of major users of ®nancial statements', Accounting and Business Research, 27, pp. 187±192. Bethel, J.E. & Liebeskind, J. (1993). `The effects of ownership structure on corporate restructuring', Strategic Management Journal, 14, pp. 15±31. Bhattacharya, S. (1979). Ìmperfect information, dividend policy and ``the bird in hand'' Fallacy', Bell Journal of Economics, Spring, pp. 259±270. Black, B. & Coffee, J (1994). `Hail Britannia: Institutional investor behaviour under limited regulation', Michigan Law Review, 92(7), pp. 1997±2087. Brown, B. & Perry, S. (1994). `Removing the ®nancial performance halo from Fortune's ``Most Admired'' companies', Academy of Management Journal, 37(5), pp. 1347±1359. Caves, R. & Porter, M.E. (1977). `From entry barriers to mobility barriers', Quarterly Journal of Economics, 91, pp. 421±434. Chandler, A.D. (1990), Scale and Scope: The Dynamics of Industrial Capitalism, Belnap Press, Harvard. Chauvin, K. & Herschey, M. (1994). `Goodwill, pro®tability and the market value of the ®rm', Journal of Accounting and Public Policy, 13, pp. 159±180. Conner, K. (1991). À historical comparison of resource-based theory and ®ve schools of thought within industrial organization economics: Do we have a new theory of the ®rm?', Journal of Management, 17, pp. 121±154. Cowen, S, Ferreri, L. & Parker, L. (1987). The impact of corporate characteristics on social responsibility disclosure: a typology and frequency based results', Accounting, Organizations and Society, 12, pp. 111±122. Department of the Environment (1996). Environmental Protection Expenditure by Industry, HMSO, London. Dye, R. (1985). `Disclosure of nonproprietary information', Journal of Accounting Research, 23, pp. 123±145. Fama, E. & French, K. (1998). `Value versus growth: The international evidence', Journal of Finance, 53(6) pp. 1975±1999. Firth, M. (1978). À study of the consensus of the perceived importance of disclosure of individual items in corporate annual reports', International Journal of Accounting Education and Research, 14(1), pp. 57±70. Fombrun, C. & Shanley, M. (1990). `What's in a name? Reputation building and corporate strategy', Academy of Management Journal, 33, pp. 233±256. Fombrun, C. (1996). Reputation: Realizing Value from the Corporate Image, Harvard Business School Press, Boston. Freedman, M. & Jaggi, B. (1982). `The SEC's pollution disclosure requirements ± Are they meaningful', California Management Review, 24(2), pp. 60±67. Fry, F. & Hock, R. (1976). `Who claims corporate social responsibility? The biggest and the worst', Business and Society Review, 18, pp. 62±65. Fryxell, C. & Wang, J. (1994). `The Fortune corporate ``reputation'' index: Reputation for what?' Journal of Management, 20(1), pp. 1±14. Gibbs, P. (1993). `Determinants of corporate restructuring: the relative importance of corporate governance, takeover threat, and free cash ¯ow', Strategic Management Journal, 14, pp. 51±68. Graves, S. & Waddock, S. (1994). Ìnstitutional owners and corporate social performance', Academy of Management Journal, 37, pp. 1034±1046.

280

J. S. TOMS

Gray, R., Javad, M., Power, D. & Sinclair, C.D. (2001), `Social and environmental disclosure and corporate characteristics: A research note and extensions', Journal of Business Finance and Accounting, 28(3/4), pp. 327±356. Gray, R., Kouhy, R. & Lavers, S. (1995). `Corporate social and environmental reporting: A review of the literature and a longitudinal study of UK disclosure', Accounting, Auditing & Accountability Journal, 8, pp. 47±77. Grossman, S.J. (1981). `The informational role of warranties and private disclosure about product quality', Journal of Law and Economics, 24(3), pp. 461±483. Hall, R. (1992). `The strategic analysis of intangible resources', Strategic Management Journal, 13, pp. 135±144. Hart, S. (1995). À natural resource-based view of the ®rm', Academy of Management Review, 20(4), pp. 986±1014. Harte, G. & Owen, D. (1991). Ènvironmental disclosure in the annual reports of british companies: A research Note', Accounting, Auditing & Accountability Journal, 4, pp. 51±61. Harte, G., Lewis, L. & Owen, D. (1991). Èthical investment and the corporate reporting function', Critical Perspectives in Accounting, 2, pp. 227±253. Hirschman, A. (1970). Exit, Voice and Loyalty: Responses to Decline in Firms, Organisations, and States, Cambridge, MA. Holland, J.B. (1998). `Private corporate disclosure, ®nancial intermediation and market ef®ciency', Journal of Business Finance and Accounting, 25(1/2), pp. 29±68. Holland, J.B. (2001). `Financial institutions, intangibles and corporate governance', Accounting, Auditing and Accountability Journal, 14(4), pp. 497±529. Holthausen, R. & Leftwich, R. (1983). `The economic consequences of accounting choice: Implications of costly contracting and monitoring', Journal of Accounting and Economics, 5, pp. 77±117. Horngren, C.T. (1957). `Disclosure: 1957', Accounting Review, 32(4), pp. 598±604. Hoskisson, R., Johnson, R. & Moesel, D. (1994). `Corporate divestiture intensity in restructuring ®rms: effects of governance, strategy and performance', Academy of Management Journal, 37, pp. 1207±1251. Hutton, W. (1995). The State We're In, London, Cape. Ingram, R. & Frazier, K. (1980). Ènvironmental performance and corporate disclosure', Journal of Accounting Research, 18, pp. 614±622. Jaggi, B. & Freedman, M. (1992). Àn examination of the impact of pollution performance on economic and market performance: Pulp and paper ®rms', Journal of Business Finance and Accounting, 19(5), pp. 697±713. Jensen, M. & Meckling, W. (1976). `Theory of the ®rm: Managerial behaviour, agency costs and capital structure', Journal of Financial Economics, 3, pp. 305±360. Jensen, M. (1986). `The agency costs of free cash ¯ow: Corporate ®nance and takeovers', American Economic Review, 76, pp. 323±329. Johnson, R. (1996). Àntecedents and outcomes of corporate restructuring', Journal of Management, 22, pp. 439±483. Keim, G. (1978). `Managerial behavior and the social responsibilities debate: An assessment of the enlightened self-interest model', Academy of Management Journal, 21, pp. 32±39. Krippendorff, K. (1980). Content Analysis: An Introduction to its Methodology, Sage Publications, Beverly Hills. Krut, R. & Gleckman, H. (1998). ISO14001: A Missed Opportunity for Global Industrial Development, Earthscan, London. Leftwich, R., Watts, R. & Zimmerman, J. (1981). `Voluntary corporate disclosure: The case of interim reporting', Journal of Accounting Research, 19(Suppl.), pp. 50±77. Litz, R. (1996). À resource-based view of the socially responsible ®rm: Stakeholder interdependence, ethical awareness and issue responsiveness as strategic assets', Journal of Business Ethics, 15, pp. 1355±1363.


281

Mak, Y.T (1991). `Corporate characteristics and the voluntary disclosure of forecast information: A study of New Zealand prospectuses, British Accounting Review, 23(4), pp. 305±28. Mathews, M. (1997). `Twenty-®ve years of social and environmental accounting research: Is there a silver jubilee to celebrate?', Accounting, Auditing and Accountability Journal, 10(4), pp. 481±531. McGuire, J., Schneeweiss, T. & Branch, B. (1990). `Perceptions of ®rm quality: A cause or result of ®rm performance', Journal of Management, 16, pp. 167±180. McGuire, J., Sundgren, A. & Schneeweiss, T. (1988). `Corporate social responsibility and ®rm ®nancial performance', Academy of Management Journal, 31, pp. 854±872. McWilliams, A. & Siegel, D. (2001). `Corporate social responsibility: A theory of the ®rm perspective', Academy of Management Review, 26, pp. 117±127. Miles, M. & Colvin, J. (2000). Ènvironmental marketing: A source of reputational, competitive, and ®nancial advantage', Journal of Business Ethics, 23, pp. 299±311. Miles, M. & Russell, G. (1997). ÌSO14000 total quality environmental management: The integration of environmental marketing, total quality management and corporate environmental policy', Journal of Quality Management, 2(1), pp. 151±168. Miles, M., Munilla, L. & Russell, G. (1997). `Marketing and environmental registration/ certi®cation: What industrial marketers should know about ISO14000', Industrial Marketing Management, 26(4), pp. 363±370. Milne, M. & Chan, C. (1998). `Narrative corporate social disclosures: How much difference do they make to investment decision-making'? British Accounting Review, 31, pp. 439±457. Morris, R. (1987). `Signalling, agency theory and accounting policy choice'. Accounting and Business Research, 18(69), pp. 47±56. Ness, K.E. & Mirza, A.M. (1991). `Corporate social disclosure: A note on a test of agency theory', British Accounting Review, 23(3), pp. 211±218. Patten, D. (1991). Èxposure, legitimacy, and social disclosure', Journal of Accounting and Public Policy, 10, pp. 297±308. Perks, R. Rawlinson, D. & Ingram, L. (1992). Àn Exploration of Ethical Investment in the UK', British Accounting Review, 24(1), pp. 43±66. Porter, M.E. & Van der Linde, C. (1996). `Green and competitive: Ending the stalemate', in Welford, R. and Starkey R. (eds). The Earthscan Reader in Business and the Environment, Earthscan, London. Preston, L. (1978). Ànalysing corporate social performance: Methods and results', Journal of Contemporary Business, 7, pp. 135±150. Raffournier, B. (1995). `The determinants of voluntary ®nancial disclosure by Swiss listed companies', European Accounting Review, 4, pp. 261±282. Roberts, R.W. (1992). `Determinants of corporate social responsibility disclosure: An application of stakeholder theory', Accounting, Organizations and Society, 17, pp. 595±612. Robertson, D. & Nicholson, N. (1996). Èxpressions of corporate responsibility in UK ®rms', Journal of Business Ethics, 15(10), pp. 1095±1106. Ross, S. (1977). `The determination of ®nancial structure: The incentive signalling approach', Bell Journal of Economics, Spring, pp. 23±40. Russo, M. & Fouts, P. (1997). À resource-based perspective on corporate environmental performance and pro®tability', Academy of Management Journal, 40(3), pp. 534±559. Schleifer, A. & Vishny, A. (1997). À Survey of Corporate Governance', Journal of Finance, 52, pp. 737±783. Scott, J. (1997). Corporate Businesses and Capitalist Classes, Oxford University Press, Oxford. Spence, A. (1973). `Job market signalling', Quarterly Journal of Economics, 87, pp. 355±379. Spicer, B. (1978). Ìnvestors, corporate social performance and information disclosure: An empirical study', The Accounting Review, LIII(1), pp. 94±111.

282

J. S. TOMS

Teoh, H. & Shiu, G. (1990). Àttitudes towards corporate social responsibility and perceived importance of social responsibility information characteristics in a decision context', Journal of Business Ethics, 9, pp. 71±77. Toms, J.S. (2000). Environmental Management, Environmental Accounting and Financial Performance, CIMA, London. Tonkin, D. & Skerrett, L.C. (1991). Financial Reporting, 1990±1: A survey of UK Reporting Practice, ICAEW, London. Trotman, K.T. & Bradley, G.W. (1981). Àssociations between social responsibility disclosure and characteristics of companies', Accounting, Organizations and Society, 6, pp. 355±362. Ullmann, A.A. (1985). `Data in search of a theory: A critical examination of the relationships among social performance, social disclosure and economic performance of US ®rms', Academy of Management Review, 10(3), pp. 540±557. Waddock, S. & Graves, S. (1997). `The corporate social-®nancial performance link', Strategic Management Journal, 18(4), pp. 303±319. Watts, R. & Zimmerman, J. (1986). Positive Accounting Theory, Prentice Hall, Englewood Cliffs. White, H. (1980). À heteroscedasticity-consistent covariance matrix estimator and a direct test for heteroscedasticity', Econometrica, 48, pp. 817±838. Wiseman, J. (1982). Àn evaluation of environmental disclosures made in corporate annual reports', Accounting, Organizations and Society, 7, pp. 53±63.