environmental performance, environmental risk and risk management
TRANSCRIPT
Environmental Performance, Environmental Risk andRisk Management
Michael Dobler,1* Kaouthar Lajili2 and Daniel Zéghal21Dresden University of Technology, Germany
2University of Ottawa, Canada
ABSTRACTEnvironmental performance, environmental risk and risk management are of contemporaryinterest, but to date there is limited evidence on their relationships. This paper is the first toprovide detailed insights by adopting a content analysis approach and disaggregating firm-levelenvironmental risk into types related to regulations, operations and nature. For a sample of USfirms in polluting sectors, descriptive findings show that the level of risk and the likelihood ofactive risk management differ in the type considered. Environmental performance, risk and thelikelihood of risk management all differ across firms and industries. Multiple regressions reveala negative association between environmental performance and environmental risk, the extentof which depends on the type of risk. Results hold when controlling for active risk management,which is not found to contribute significantly to environmental performance. Our findings haveimplications for public policy and suggest that linkages to environmental risk and risk manage-ment are worth exploring in more differential ways and beyond industry-level assessments inenvironmental studies. Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment.
Received 25 April 2012; revised 05 July 2012; accepted 12 July 2012
Keywords: environmental policy; environmental risk; environmental risk management; environmental performance; content
analysis; United States
Introduction
ENVIRONMENTAL PERFORMANCE, ENVIRONMENTAL RISK AND RISK MANAGEMENT ARE INCREASINGLY IMPORTANT AND
crucial for business success (Gouldson, 2004; Ambec and Lanoie, 2008; Anderson and Anderson, 2009;Al-Najjar and Anfimiadou, 2012). Research reveals environmental risk and its corporate management askey factors in business relations with investors (Sharfman and Fernando, 2008; Aktas et al., 2011), banks
(Aintablian et al., 2007; Weber et al., 2008) and other stakeholders (Matten, 1995; Hofer et al., 2012). Findingssuggest that both lower levels of environmental risk and the implementation of environmental risk managementpractices are associated with economic benefits for firms. However, there is surprisingly little evidence on theirassociation with corporate environmental performance.
This gap in research likely stems from difficulties in assessing and measuring environmental risk faced by afirm. The few empirical studies addressing the association between environmental performance and environmental
*Correspondence to: Prof. Dr. Michael Dobler, Faculty of Business and Economics, Dresden University of Technology, 01062 Dresden, Germany.E-mail: [email protected]
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment
Business Strategy and the EnvironmentBus. Strat. Env. (2012)Published online in Wiley Online Library(wileyonlinelibrary.com) DOI: 10.1002/bse.1754
risk rely on a firm’s affiliation to a polluting sector as proxy for environmental risk (Semenova, 2010; also Kassinisand Vafeas, 2006; Clarkson et al., 2008; Semenova and Hassel, 2008). This does not capture environmentalrisk at the firm level, the types of risk it is composed of and the firm’s response to specific types of environ-mental risk by means of risk management. What types of environmental risk do firms face and what is theirlevel? Which of them are addressed by active corporate risk management? And how are both environmentalrisk and risk management associated with environmental performance? These are interesting questions thatseek empirical evidence.
This paper addresses the above questions by hand collecting environmental risk and risk managementinformation from 10-K filings and linking them to alternative environmental performance measures derivedfrom Toxics Release Inventory (TRI) data for US S&P 500 constituents from polluting sectors. Our contentanalysis approach to develop risk and risk management measures exploits disclosure mandated and enforcedby the Securities and Exchange Commission (SEC). Most relevant disclosure requirements concern contingenciesin the notes to financial statements, a description of business (item 101 of Regulation S-K), legal proceedings(item 103), the management’s discussion and analysis (item 303) and risk factors (item 503(c)) (Cho andPatten, 2008; Dobler et al., 2011), which have been supplemented by specific guidance regarding disclosurerelated to climate change (SEC, 2010). Disclosure in 10-K filings, however, is only required when material(Clarkson et al., 2008; SEC, 2010). This implies that we do not capture information on remote or minorenvironmental risk and its management. Thus, findings should be interpreted within the focus on materialrisk and related risk management. Our approach has two major advantages. It overcomes the limitation onindustry-level environmental risk and allows for a firm-level assessment of risk while controlling for inter-industrydifferences. Consistent with a conceptual view on environmental risk (Baum et al., 1983; Matten, 1995), it allowsus to distinguish different types of environmental risk a firm faces and may actively manage. This enables us toprovide new insights into the association between environmental performance and environmental risk faced andmanaged, respectively.
Background and Hypothesis Development
In business research, there is high consensus that risk can be seen as a combination of the probability ofoccurrence of a state or event and its consequences or impact (Kaplan and Garrick, 1981; Aven, 2012). Environ-mental research typically focuses on downside risk as opposed to upside risk or opportunity (Cuddihy, 2000;Böhm and Pfister, 2005). This focus restricts possible consequences to undesirable consequences by referenceto levels of acceptance set by regulation, technology or by corporate or social objectives, conventions or percep-tions. Typologies of environmental risk employed in psychology basically distinguish two categories linked toeither human activities or the natural environment as sources of risk (Baum et al., 1983; Böhm and Pfister,2005). Put in a business context, Matten (1995) argues that human business activities may result in damageto the environment which in turn threatens a firm with undesirable consequences. By reference to regulatoryor technological and firm-specific levels of acceptance, respectively, risk from regulations and risk from opera-tions can be distinguished. Risk arising from the natural environment relates to natural disasters or changesin the natural environment that are beyond the control of an individual firm but may have negative impact onit. In sum, we can distinguish three types of environmental risk – risk from regulations, from operations andfrom nature – where the last conceptually differs from the others. Firms may choose to accept environmentalrisk of a given type or to take preventive or remedial risk management actions such as to avoid, reduce ortransfer it (Anderson and Anderson, 2009; Anghelache, 2011).
With a focus on environmental risk and risk management, our study is broadly connected to the expansiveliterature on corporate social responsibility (Orlitzky et al., 2003; Campbell, 2007) and more particularly embeddedin research on environmental performance. One strand of environmental performance research examines theassociation between environmental performance and the volume or composition of corporate environmentaldisclosures. Findings are mixed and seem to be sensitive to the construct to measure disclosure. Studies thatemploy a disclosure index equal or similar to that developed by Wiseman (1982) document no (Freedman and
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Wasley, 1990; Hughes et al., 2001) or a negative relation (Bewley and Li, 2000; Patten, 2002). Using more sophisti-cated disclosure variables, Al-Tuwaijri et al. (2004) and Clarkson et al. (2008) find that environmental perfor-mance and disclosure are positively related. While all these studies do not individually assess specificdisclosure on environmental risk and risk management, there is evidence that firms provide environmentaldisclosure of these types (Chan and Welford, 2005; Dobler et al., 2011) and regulators strive to enhance them(Cho and Patten, 2008; SEC, 2010).
Another large strand of research examines the association between environmental and economic performance(Hart and Ahuja, 1996; Ambec and Lanoie, 2008; van Buerden and Gössling, 2008; Molina-Azorín et al., 2009).Findings imply that environmental performance is related to firm size, capturing environmental exposure, relativepollution propensity and corporate visibility (Hart and Ahuja, 1996; Bowen, 2000), to ownership and governance ofthe firm as proxies for stakeholder pressure and the ability to respond to it (Berrone et al., 2010; de Villiers et al., 2011),to board compensation (Berrone and Gomez-Mejia, 2009; Walls et al., 2012) and to the industry sector in which the firmoperates (King and Lenox, 2001; Clarkson et al. 2008). Despite being neglected inmany studies, industry affiliation seemsto be a key determinant of environmental performance because cross-sectional differences in emissions are likely toemerge from different operational practices, technologies and regulations employed. Consistent with these assessments,for example, several studies in the US document that the utilities sector is associated with relatively low environmentalperformance as measured by toxic releases (Kassinis and Vafeas, 2006; Clarkson et al., 2008). To the extent that industryaffiliation proxies for environmental risk, findings suggest a negative association between environmental performanceand environmental risk.
More specific to our study, Semenova and Hassel (2008) and Semenova (2010) take approaches more focusedon environmental risk. Both papers use samples of listed US firms and GES Investment Services risk rating data.Semenova and Hassel (2008) show that environmental risk differs significantly across industries. The extent ofindustry-level environmental risk is positively associated with financial performance and negatively associatedwith market value. Firm-specific GES measures are significantly and positively associated with the economicposition, but this relation does not hold for firms operating in highly polluting industries. In examining relationsbetween environmental rating dimensions across databases, Semenova (2010) documents a positive associationbetween industry-level environmental risk and the firm-specific GES measures. Results incorporating firm-specific GES measures should be interpreted with care because they capture environmental reputation andmanagement rather than firm-specific environmental risk. Findings imply that a firm’s environmental (risk)management efforts are more pronounced in environmentally risky industries and are rewarded in economicand reputational regards.
Prior empirical research has provided limited direct and rather industry-level evidence on a negative associationbetween environmental performance and environmental risk and falls short on the components of environmentalrisk. We posit that the association depends on the type of environmental risk focused on. Arguments in priorliterature emphasize risk from regulations (Karpoff et al., 2005; Sarkis et al., 2010). Firms with poor environ-mental performance, for instance as indicated by high levels of emissions or releases, are likely to face high riskof violating environmental regulations and being fined. Consistent with political cost theory (Posner, 1974;Stephan, 2002), firms with poor environmental performance are also likely to face increased risk from costlyregulatory intervention such as stricter regulation. A negative association between environmental performanceand environmental risk is also likely to prevail for environmental risk from operations. For example, Konar andCohen (1997, 2001) argue that high levels of emissions indicate increased risk of operational accidents and spills.Similarly, Chatterji et al. (2009) suggest that higher environmental risk from operations predicts increased levelsof emissions. More distinctive to risk from regulations, poor environmental performance may be associated withunfavorable actions from stakeholders such as activists’ intervention that could potentially put operations atfurther environmental risk. Our first hypothesis thus states the following.
H1: There is a negative association between the level of environmental risk and environmental performance.
Weighing the costs and benefits, firms facing environmental risk can decide whether to engage in active riskmanagement (Heinkel et al., 2001; Siegel, 2009). Several strands of theory, including stakeholder theory, agencytheory, signaling theory and legitimacy theory as well as resource-based views or a Schumpeterian perspective,
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
imply that firms engaged in respective management activities are rewarded by competitive advantages (Delmas andToffel, 2008; Kock et al., 2012; Delmas and Montiel, 2009; Bansal and Clelland, 2004; Hart and Dowell, 2011;Hofer et al., 2012). For example, literature has linked various beneficial effects to the adoption of instruments ofenvironmental management (Klassen and McLaughlin, 1996; Sarkis et al., 2010), to environmental managementaccounting (Burritt et al., 2002; Ferreira et al., 2010) and more particularly to environmental risk management(Matten, 1995; Sharfman and Fernando, 2008). The beneficial effects assumed include both economic and environ-mental performance (Melnyk et al., 2003; Zhu and Sarkis, 2004). This would imply a positive association betweenenvironmental performance and environmental risk management which is consistent with basic intuition: goodenvironmental performance may be viewed as environmental risk management (King and Shaver, 2001; Sharfmanand Fernando, 2008). To date, however, evidence on various environmental management activities is lesscompelling than one would expect (Annandale et al., 2004; Qi et al., 2012).
Specific to the focus of our study, the positive association between environmental performance and environ-mental risk management of a firm is challenged for several reasons. First, firms in highly polluting industrieswith poor environmental performance are particularly likely to invest in environmental risk management, whichsuggests a negative association on a cross-sectional basis (Karpoff et al., 2005; Semenova and Hassel, 2008).Second, environmental risk management may not work effectively due to deficient implementation, complianceand commitment on various levels of the process (Iraldo et al., 2009; Sarkis et al., 2010). Third, there are inherentbarriers to environmental risk management, including particular problems in identifying and evaluating optionsto handle risk that may lead to inferior decisions and ineffective risk management (Gregory et al., 2006; Winnet al., 2011). Fourth, given multiple and potentially competing targets and effects of risk management, theremay be effective instruments of environmental risk management in place that do not or negatively impact oncertain measures of environmental performance (Gregory et al., 2006; Semenova and Hassel, 2008). Forexample, environmental risk management that is not preventive but remedial in nature is unlikely to directlyaffect environmental performance as measured based on waste or toxic releases. Given these arguments oursecond hypothesis states the following.
H2: There is no association between active management of environmental risk and environmental performance.
Data and Methodology
Sample Selection
The sample of this study consists of S&P 500 listed firms from four sectors that file both toxic release data to the USEnvironmental Protection Agency and 10-K filings to the SEC for the calendar year 2010. The four sectors focusedon are energy, industrials, materials and utilities, all of which are considered to have high pollution propensity(Semenova and Hassel, 2008; Clarkson et al., 2011). The US provide a highly regulated setting with regards toenvironmental and investor protection and related reporting (Barrieu and Sinclair-Desgagné, 2006; Dobleret al., 2011). In this setting, the use of TRI data is well established in environmental research (Gerde and Logsdon,2001; Sullivan and Gouldson, 2007). 10-K filings provide a common and reliable source of information exploitedin risk and environmental disclosure studies (Al-Tuwaijri et al., 2004; Cho and Patten, 2008). Since TRI datais released per calendar year we exclude firms with a financial year different from the calendar year to obtain aconsistent set of data.1
Our final sample contains 89 firms. The distribution across the sectors is as follows: 18 firms (20.2% of thesample) in the Energy sector, 24 (27.0%) in the Industrials sector, 27 (30.3%) in the Materials sector, and 20(22.5%) in the Utilities sector.
1By referring to the year 2010, we avoid possible impacts during the financial crisis (see, e.g., Ellis and Bastin, 2011) and are able to assess climatechange disclosure as guided by the SEC (2010).
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Main Variables
For the purpose of this study, three main variables have to be measured: environmental performance, environ-mental risk and environmental risk management. Measures for environmental performance (EP) are drawn fromthe TRI database. Since the database reports plant-specific data we aggregate relevant data at the sample firmlevel. In order not to rely on a single proxy, we calculate two relative environmental performance variables thathave been used in prior environmental performance studies (Patten, 2002; Al-Tuwaijri et al., 2004; Clarksonet al., 2008; Sharfman and Fernando, 2008). The first variable is the percentage of total waste that is treated,recycled or recovered as a percentage of total waste generated (%RECYC). The second is the ratio of total wastegenerated in gram per dollar of sales with reversed sign (�TRI/SAL). We reverse the sign of the second ratioto facilitate the interpretation of the results. In other words, the larger each variable, the better is the firm’senvironmental performance.
Table 1 reports descriptive statistics on environmental performance measures of our sample by sector. Onaverage, sample firms recycle 69.7% of waste generated and have 0.430 gram of toxic release per dollar sales.Consistent with prior evidence, there are considerable inter-industry differences.2 Firms in the Energy andIndustrials sectors seem to have similar and relatively good environmental performance, as indicated by highvalues of our EP proxies. t and Mann–Whitney tests do not indicate a difference between sample firms inthese sectors. Likewise, the Materials and Utilities sectors have similar and relatively poor environmentalperformance, as indicated by low values of our EP proxies. Significant at the 10% level, only %RECYC islower in the Utilities sector. The latter is largely consistent with prior studies that identify firms in theUtilities sector as poor environmental performers in terms of toxic releases (Kassinis and Vafeas, 2006;Clarkson et al., 2008). Apart from the two industry pairs, there are significant cross-sector differences foralmost all other combinations.
We develop our environmental risk and risk management variables based on a content analysis of each samplefirm’s disclosure in 10-K filings. A pretest of 12 randomly chosen filings reveals that the types of environ-mental risk derived in the second section are well reflected in actual disclosure and provide superordinatecoding categories.
• Risk from regulations (R) relates to environmental laws and regulations and puts firms at risk to penalties, litigationand enhanced environmental constraints when violating a legal or a regulator’s level of acceptance.
• Risk from operations (O) refers to technological failures and firm-specific levels of acceptance beyond regulations.Examples include environmental risk related to power generation or distribution, oil or gas drilling, storage anddistribution, waste and landfall disposal.
• Risk from nature (N) relates to the impact of nature that is beyond an individual firm’s control. It includes risk ofclimate change, natural disasters, seasonality and weather conditions.
For each of these categories, information disclosed has to be hand collected along three dimensions: riskexposure, risk consequence and risk management. Detailed pretest coding of 12 filings reveals surprisinglyrich disclosure, but in each dimension a limited number of categories is relevant and operational. Riskexposure (RE) is coded 2 for high probability, 1 for low probability of occurrence and 0 otherwise. Riskconsequence (RC) is coded 5 for catastrophic, 3 for major, 1 for moderate impact and 0 otherwise. The cate-gorization is consistent with typical risk map approaches used to assess, quantify and communicate environ-mental risk (see, e.g., Kennedy, 2001). Risk management (RM) is a binary variable coded 1 for firms indicatingactive risk management, and 0 otherwise, for each type of environmental risk (RM_R, RM_O and RM_N).Procedures yield coding 0 in absence of specific disclosure. There is only one exception. For the sake ofconsistency, RC is coded 1 (moderate impact) if there is no specific disclosure on RC for a type of envi-ronmental risk the firm has disclosed being exposed to (RE>0). In some cases, a firm indicates beingexposed to a type of environmental risk but does not explicitly mention the level of exposure. RE is thenimplicitly assessed by the relative rank in the list of risk factors in 10-K filings (Item 503(c) of RegulationS-K; Dobler, 2008). To foster consistency and reproducibility, one single coder (an experienced graduate student)
2Detailed inter-industry comparisons of the EP proxies and the other main variables are shown in the appendix.
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
conducted the coding of the full sample under the supervision of one of the authors. The coder replicated ourpretest coding. This yielded minimal differences, indicating a high level of inter-coder reliability.
In order to develop measures for environmental risk, we multiply the scores for risk exposure and riskconsequence for each type at the firm level. This is consistent with approaches taken in environmental riskmanagement practice (see, e.g., Kennedy, 2001) and yields RISK_R, RISK_O and RISK_N as proxies for envi-ronmental risk in each category with values between 0 and 10. Although theory and intuition do not imply therisk from nature category to be associated with our environmental performance measures, we collected data onthis category to complement the descriptive insights and the sensitivity analyses. Adding RISK_R and RISK_Oyields RISK_RO as a measure for total environmental risk. To construct a variable for total environmental riskmanagement we use the sector median of the ratio of aggregate RISK actively managed across the types toaggregate RISK across categories as a cutoff. Similar approaches have been used to obtain binary variablesin prior environmental research (e.g., Clarkson et al., 2008). For the two key types of risk (from regulationsand from operations), this treatment turns out to be equal to assigning RM_RO= 1 to firms that actively manageenvironmental risk whenever they face it.
Table 2 provides a per sector overview of coding results constituting the resulting risk and risk managementvariables. We observe 88 firms (98.9% of the sample) with material risk from regulations, 40 (44.9%) withmaterial risk from operations and 61 (68.5%) with material risk from nature. This rank pattern holds true acrossall sectors. Observations indicate that environmental risk from regulations is most frequently associated withlarge probability and major impact (as denoted by No RE = 2 and No RC = 3) while low probability and catastrophicimpact dominate the type of risk from operations (as denoted by No RE = 1 and No RC = 5). In the total sample, themean aggregated risk measure is 7.663. Mean risk from regulations is 5.169 and significantly higher than eachother type. This pattern holds true across all sectors. Active risk management seems prevalent for risk fromregulations (75.3% of the sample) but is scarcely observable for risk from nature (6.7%). Concerning risk fromclimate change in particular, the latter finding is consistent with limited instruments available to business firms(Winn et al., 2011).
Inter-industry analyses reveal that sample firms in the Energy and Utilities sectors have similar and ratherhigh environmental risk (mean RISK_RO is equal to 9.667 and 9.850, respectively). In turn, sample firmsin the Industrials and Materials sectors have similar and rather low environmental risk (mean RISK_RO isequal to 5.208 and 6.889, respectively). t and Mann–Whitney tests presented in the appendix do not indicatea significant difference between these industry pairs for any type of environmental risk. For all other industrypairs and almost all types of risk we find differences that are significant at the 5% level. Active management of
Mean Median Min Max St. dev.
Total sample (N = 89)%RECYC 0.697 0.817 0.000 0.998 0.321–TRI/SAL �0.430 �0.030 �15.499 �0.000 1.665Energy (N = 18)%RECYC 0.776 0.953 0.000 0.998 0.364–TRI/SAL �0.015 �0.007 �0.068 �0.000 0.020Industrials (N = 24)%RECYC 0.850 0.900 0.139 0.998 0.196–TRI/SAL �0.055 �0.005 �1.093 �0.000 0.222Materials (N = 27)%RECYC 0.665 0.811 0.000 0.998 0.333–TRI/SAL �0.943 �0.202 �15.498 �0.001 2.942Utilities (N = 20)%RECYC 0.486 0.557 0.000 0.884 0.277–TRI/SAL �0.564 �0.461 �1.730 �0.033 0.448
Table 1. Descriptive statistics on environmental performance variables
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Cod
ingresults
Descriptivestatistics
No
RE>0
No
RE=1
No
RE=2
No
RC=1
No
RC=3
No
RC=5
No
RM=1
RM Mean
RMMedian
RISK
Mean
RISK
Median
RISK
Min
RISK
Max
RISK
St.d
ev.
Totalsam
ple(N
=89
)Re
gulatio
nstype
(R)
880
8818
691
670.753
15.169
60
101.779
Ope
ratio
nstype
(O)
4030
103
631
250.281
02.49
40
010
3.255
Naturetype
(N)
6149
1215
3214
60.06
70
2.528
30
102.66
3Ag
gregated
(RO)
8956
0.62
71
7.66
36
120
4.034
Energy
(N=18)
Regu
latio
nstype
(R)
180
181
161
130.722
16.00
06
210
1.372
Ope
ratio
nstype
(O)
1210
20
210
70.389
03.66
75
010
3.218
Naturetype
(N)
1410
41
85
30.167
03.556
00
102.727
Aggregated
(RO)
1810
0.556
19.66
710
620
3.819
Indu
stria
ls(N
=24
)Re
gulatio
nstype
(R)
230
238
150
170.70
81
4.417
60
62.125
Ope
ratio
nstype
(O)
53
23
11
20.08
30
0.792
00
102.322
Naturetype
(N)
119
22
72
00.00
00
1.70
80
010
2.528
Aggregated
(RO)
2415
0.62
51
5.20
86
116
3.148
Materials(N
=27)
Regu
latio
nstype
(R)
270
277
200
200.741
14.96
36
26
1.786
Ope
ratio
nstype
(O)
84
40
35
20.07
40
1.92
60
010
3.419
Naturetype
(N)
1917
29
82
10.037
01.815
10
102.149
Aggregated
(RO)
2715
0.556
16.889
62
163.955
Utilities
(N=20
)Re
gulatio
nstype
( R)
200
202
180
170.850
15.60
06
26
1.231
Ope
ratio
nstype
(O)
1513
20
015
140.70
01
4.250
50
102.936
Naturetype
(N)
1713
43
95
20.100
03.550
30
102.911
Aggregated
(RO)
2016
0.80
01
9.850
102
163.40
7
Table2.
Cod
ingan
ddescrip
tivestatisticson
riskan
dris
kman
agem
entvaria
bles
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
risk from operations and aggregated environmental risk is most prevalent in the Utilities sector (70.0% and80.0% of the sub-sample), which is consistent with the overall regulatory environment in this sector (Kassinisand Vafeas, 2006).
Since environmental risk differs significantly, we use the ranks of the risk measures within each industry asour main risk variables per type (RRISK_R and RRISK_O) and as aggregated (RRISK_RO) in our regressionsin order to control for industry differences in the level of risk. This treatment is consistent with Lang and Lundholm(1996) and Clarkson et al. (2008), and maintains that the larger the risk variable the higher is the risk faced by the firmin each sector.
Regression Model
To test our hypotheses we employ two empirical models. Model (1) focuses on the ranked risk variables. Model (2)additionally addresses active environmental risk management. We run both regressions including either aggregatedenvironmental risk and risk management variables (aggregate models (1A) and (2A)) or separate variables for theregulations and operations types (separate models (1S) and (2S)). In consequence, there are four sets of modelsregressing each of our alternative environmental performance variables on the environmental risk and riskmanagement variables and the control variables:
EP ¼ a0 þ a1RRISK þ a2ROAþ a3SIZ þX
ja3þjINDj þ e (1)
EP ¼ a0 þ a1RRISK þ a2RM þ a3ROAþ a4SIZ þX
ja4þjINDj þ e (2)
where
EP environmental performance variable, i.e. %RECYC (percentage of total waste that is treated, recycled orrecovered as a percentage of total waste generated) or –TRI/SAL (ratio of total waste generated in gramsper dollar of sales with reversed sign),
RRISK industry ranked environmental risk variable, aggregated (RRISK_RO) or separated per risk category(RRISK_R and RRISK_O),
RM dummy variable that equals 1 for firms actively managing risk, and 0 otherwise, aggregated (RM_RO) orseparated per risk category (RM_R, RM_O),
ROA return on assets,SIZ natural logarithm of waste generated by the firm in grams,INDj industry dummies that equal 1 if the firm is in sector j (Energy, Industrials, Materials), and 0 otherwise.
Our hypotheses predict a negative estimate of a1, while leaving the coefficient on risk management variables,which we expect to be insignificant, to be answered by evidence. We include ROA, SIZ and INDj as control variables.ROA controls for the association between financial performance and environmental performance in our sampleyear (Orlitzky and Benjamin, 2001; Semenova, 2010). SIZ controls for size, pollution propensity and relatedenvironmental visibility of a firm (Bowen, 2000; Walls et al., 2012).3 Finally, we include industry dummies tocontrol for differences in environmental performance revealed in the preceding section (Kassinis and Vafeas, 2006;Semenova and Hassel, 2008).
3If we alternatively use the natural logarithm of a firm’s total assets for SIZ, our results are qualitatively unchanged but the adjusted R² decreases.
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Research Results
Correlation Results
Table 3 presents summary statistics and Pearson correlations among the variables. The correlation between%RECYC and –TRI/SAL is equal to 0.213 and significant at 5%, implying that both variables measure environ-mental performance but the overlap is modest. The ranked risk and the risk management measures, respectively,are positively but insignificantly correlated between the regulation and operation types. Each of them is positivelycorrelated at the 1% level with the aggregated measure for both types. This suggests that individual risk type variablescan capture differential effects.
We find that correlations between environmental performance and risk or risk management variables are allnegative. With the exception of the regulations type, the negative correlations between environmental performanceand environmental risk measures are all significant as predicted by H1. Inconsistent with H2, however, correlationsbetween our environmental performance variables and active management of risk from operations (RM_O) areequal to �0.316 or �0.219 and significant. ROA is not correlated with any of our main variables, while SIZ is signif-icantly correlated with environmental performance only. These univariate results suggest that financial performanceas well as size and pollution propensity of firms are not associated with environmental risk from regulations andoperations and respective management.
Additional analyses (untabulated) reveal that correlations between environmental performance and risk or riskmanagement variables in the nature categories, respectively, have inconsistent signs but neither of them is signif-icant at the 10% level. This is consistent with our expectation and our decision to exclude the type of risk fromnature from our main regressions.
Regression Results
Table 4 reports the results of multivariate regressions of environmental performance measures on the environ-mental risk and risk management variables and the control variables. For both dependent variables, the estimatedcoefficients on RRISK_RO in aggregated risk model (1A) are negative and significant (p = 0.006 for %RECYC;p = 0.012 for –TRI/SAL). This result supports H1 and implies that firms facing higher environmental risk havea poorer environmental performance. When we disaggregate total risk into risk from regulations and risk fromoperations in model (1S), we find that coefficients on risk variables all are negative but only coefficients on riskfrom operations are significant (p = 0.011 for %RECYC; p = 0.016 for –TRI/SAL). This suggests that the types ofrisk are not incrementally informative to each other with respect to environmental performance. Rather, risk fromoperations seems to be the determining factor for the negative association between environmental performanceand environmental risk.
These results hold true when we include the risk management variables in models (2A) and (2S). Again, findingsare consistent for both dependent environmental performance variables. In each specification, coefficients on therisk management variables are negative but insignificant (p≥0.121). This is consistent with H2 predicting no asso-ciation between environmental performance and environmental risk management.
Concerning the control variables, Table 4 consistently indicates a negative but insignificant associationbetween environmental performance and financial performance as measured by ROA (p≥0.274). While sig-nificant across the board (p≤ 0.050), the sign of the association between environmental performance andthe firm size and pollution propensity (SIZ) differs between the dependent variables. Both results are consis-tent with our univariate findings discussed in the preceding section. They particularly imply that firms that arelarger and generate more total waste are relatively more engaged in recycling but still have more releases perdollar of sales.
While our %RECYC regressions have higher adjusted R² and higher F values compared with the –TRI/SALregressions, the respective specifications yield consistent results. In sum, our results support H1 and H2 but implythat the association between environmental performance and environmental risk and risk management, respec-tively, is differential in the types of environmental risk linked to environmental performance.
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Varia
bles
Correlatio
ns
Mean
St.d
ev.No
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
%RE
CYC
0.69
70.321
[1]0.213*
*�0
.193
*�0
.299
***
�0.327
***
�0.067
�0.191
*�0
.184
*�0
.161
�0.316
***
�0.159
0.09
30.255*
*
–TR
I/SA
L�0
.430
1.66
5[2]
1�0
.052
�0.281
***
�0.250
**�0
.066
�0.282
***�0
.273
***
�0.111
�0.219
**�0
.101
0.04
9�0
.283
***
RISK
_R5.169
1.779
[3]
10.217*
*0.616*
**0.571*
**�0
.018
0.327*
**0.113
0.197*
0.191*
�0.151
0.07
9RISK
_O2.49
43.255
[4]
10.90
3***
0.09
10.782*
**0.69
8***
0.08
80.64
6***
�0.134
�0.052
0.116
RISK
_RO
7.66
34.034
[5]
10.178*
0.62
3***
0.70
7***
0.121
0.62
6***
�0.024
�0.108
0.128
RRISK_
R4.99
53.674
[6]
10.00
40.42
9***
0.014
�0.199
*0.06
70.012
�0.026
RRISK_
O6.46
17.813
[7]
10.812*
**0.037
0.356*
**�0
.317
***
�0.017
0.113
RRISK_
RO8.328
6.611
[8]
10.144
0.309*
**�0
.164
�0.033
0.031
RM_R
0.753
0.434
[9]
10.126
0.746*
**0.00
80.122
RM_O
0.281
0.452
[10]
10.273*
**�0
.190
*0.052
RM_R
O0.62
70.48
6[11]
1�0
.024
�0.005
ROA
0.053
0.04
8[12]
1�0
.082
SIZ
21.643
2.813
[13]
1
Table3.
Pearsoncorrelations
%RE
CYC
,percentageof
totalw
astethatistreated,
recycled
orrecoveredas
ape
rcen
tage
oftotalw
astegene
rated;
–TR
I/SA
L,ratio
oftotalw
astegene
ratedin
gram
spe
rdo
llarof
saleswith
reversed
sign
;RISK
,measure
foren
vironm
entalris
kfrom
regu
latio
ns(RISK_
R),from
operations
(RISK_
O)an
daggregated
forbo
thtype
s(RISK_
RO);RR
ISK,
indu
stry
ranked
measure
foren
vironm
entalriskfrom
regu
latio
ns(RRISK
_R),from
operations
(RRISK
_O)an
daggregated
forbo
thtype
s(RRI-
SK_R
O);RM
,dum
myvaria
blethat
isequalto1forfirm
sactivelyman
agingen
vironm
entalriskan
d0otherwise,
forris
kfrom
regu
latio
ns(RM_R
),from
operations
(RM_O
)an
daggregated
forb
othtype
s(RM_R
O);RO
A,returnon
assets;S
IZ,n
aturallogarithm
ofwaste
gene
ratedby
thefirm
ingram
s.Num
bero
fobservatio
ns:89.
***/
**/*
Sign
ificanceat
1%,5%
and10%,respe
ctively.
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Dep
endent
varia
ble:
%RE
CYC
%RE
CYC
%RE
CYC
%RE
CYC
–TR
I/SA
L–TR
I/SA
L–TR
I/SA
L–TR
I/SA
L
Mod
el:
(1A)
(1S)
(2A)
(2S)
(1A)
(1S)
(2A)
(2S)
Intercep
t(significance)
�0.618
**�0
.645
**�0
.536
**�0
.573
**3.054*
*2.552*
*3.60
0**
3.86
8**
(0.013)
(0.011)
(0.033)
(0.027)
(0.040
)(0.020
)(0.020
)(0.015)
RRISK_
RO(significance)
�0.012
***
�0.013
***
�0.066
**�0
.071
***
(0.006
)(0.003)
(0.012)
(0.007
)RR
ISK_
R(significance)
�0.010
�0.009
�0.026
�0.020
(0.259)
(0.312)
(0.630)
(0.709
)RR
ISK_
O(significance)
�0.010
**�0
.008
**�0
.053
**�0
.050
**
(0.011)
(0.047)
(0.016)
(0.038)
RM_R
O(significance)
�0.085
�0.511
(0.142
)(0.149
)RM
_R(significance)
�0.100
�0.177
(0.121)
(0.469
)RM
_O(significance)
�0.048
�0.190
(0.565)
(0.312)
ROA(significance)
�0.496
�0.481
�0.468
�0.452
�3.839
�3.543
�3.670
�4.112
(0.418)
(0.436)
(0.442
)(0.465)
(0.306
)(0.350)
(0.325)
(0.274)
SIZ(significance)
0.053*
**0.055*
**0.052*
**0.055*
**�0
.138
**�1
.993
**�0
.140
**�0
.141
**
(0.000
)(0.000
)(0.000
)(0.000
)(0.033)
(0.050)
(0.029
)(0.030)
Indu
stry
controls
yes
yes
yes
yes
yes
yes
yes
yes
Adj.R²
0.370
0.362
0.379
0.380
0.123
0.111
0.135
0.140
Fstatistic
(significance)
9.60
4***
8.146*
**8.66
5***
8.743*
**3.054*
**2.555*
*2.957*
**2.593*
*
(0.000
)(0.000
)(0.000
)(0.000
)(0.009
)(0.019)
(0.008
)(0.011)
Table4.
Regression
results
Varia
bles
arede
fine
din
Table3.Num
berof
observations:8
9.***/
**/*
Sign
ificanceat
1%,5%
and10%,respe
ctively.
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Sensitivity Analysis
To assess whether results of our main analyses are robust, we conduct sensitivity analyses on the effects of the type ofrisk from nature and firms with zero risk in any category. Main regressions excluded the nature risk category since itshould not be meaningfully related to the environmental performance measures used in our study. Including thenature risk measures in separate models and in total risk and risk management variables used in aggregatemodels does not alter our results. As expected, neither coefficient on the additional category loads significantlyin separate model specifications.
To address the effect of firms with no disclosure in either the regulation or operation categories, we replicatedour analyses without these firms. For the subsample of 39 firms, correlations between main variables are consistentwith those presented in Table 3 except for the operations type variables. Notably, the negative correlations ofenvironmental performance measures with RRISK_O are significant at 10% only, and with RM_O insignificant.While uncorrelated in the total sample, the correlation between RRISK_R and RRISK_O is now equal to 0.597 andsignificant at 0.1%. When running %RECYC regressions for the subsample, neither coefficient on the risk or riskmanagement variable is significant in the separate models but all of them are negative. In aggregate models,however, RRISK_RO remains significant at a level decreased to 10%. In sum, our main results seem qualitativelyrobust but may to some extent be sensitive to firms that misleadingly claim not to face material environmental riskof the respective types.
Discussion and Conclusions
We adopted a content analysis approach to develop environmental risk and risk management variables in order toprovide various new insights into the association between environmental performance and environmental riskand its management, respectively. To the best of our knowledge, this is the first study exploring this associationbeyond environmental risk at the industry level.
Based on a conceptual perspective on risk, we disentangle three types of environmental risk related to regula-tions, to operations and to nature. Concurrent with strong financial disclosure regulation in the US, our contentanalysis reveals that these types of environmental risk and the firms’ responses to them are surprisingly wellreflected in disclosure in 10-K filings in US polluting industries. This implies that detailed disclosure on environ-mental risk and to a lesser extent its management are more prevalent than assumed by financial disclosureregulators (Cho and Patten, 2008; SEC, 2010). Descriptive findings indicate that the level of risk and thelikelihood of active risk management differ in the type of environmental risk considered. As a consistent patternacross industries, risk from regulations is the most frequent and largest scale type, risk from operations is the typewhere catastrophic impact upon the firm is most likely and risk from nature is the type that is least likely to besubject to active risk management. Cross-sectional comparisons reveal that environmental performance, envi-ronmental risk and the likelihood of its management all differ significantly between polluting industries. Suchdifferences need to be controlled for in regression analyses. This is important to note, since the still widespreadlack of industry controls is likely to neglect differences in the production processes and business environmentsacross industries and thus to affect inferences drawn in cross-sectional environmental studies (Semenova andHassel, 2008; Clarkson et al., 2011).
Controlling for inter-industry differences, our results reveal a negative association between environmentalperformance and environmental risk at the firm level. This is consistent with theory and industry-level assess-ments (Sarkis et al., 2010; Semenova, 2010). More particularly, however, results indicate that this negative asso-ciation depends on the type of environmental risk. The relationship is significant for total environmental riskand strongest for risk from operations. However, risk from nature seems to be unrelated to relative measuresof toxic releases, as expected. These findings suggest that different types of environmental risk have differenteffects on the firms’ environmental performance. While neglected in existing environmental research, the disag-gregation of environmental risk could be a fertile extension to environmental performance studies. Futureresearch could shed more light on the extent of environmental risk of different types during a longer period of
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
time. By lengthening the time period, future studies could further investigate the links between environmentalperformance and environmental risk management by refining the different components of environmental riskand their respective management strategies.
Our results do not support the intuitive view that active management of environmental risk is associated withgood environmental performance. Results rather imply a negative yet insignificant relationship. Complementingcritical perspectives on environmental management (Welford, 1998; Ählström et al., 2009), this seems to suggestthat active risk management is no guarantee for good environmental performance. Our finding could be due toissues in the implementation, compliance and effectivity of environmental risk management discussed in literature(Gregory et al., 2006; Iraldo et al., 2009). However, findings do not necessarily imply that the environmental riskmanagement activities lack effectivity. Some risk management action taken may be effective but does not directlyaffect our alternative environmental performance variables. Examples include remedial types of risk managementsuch as actions addressing stakeholder reactions upon environmental incidents rather than taking actions to preventthe incident or reduce its direct impact on the environment. This can be seen in support of a differential relationshipbetween environmental performance and types of environmental risk management.
Relying on environmental disclosure in 10-K filings, our study is subject to limitations of the disclosure provided.First, disclosure allows us to disentangle types of environmental risk and whether action is taken in response to eachtype, but it does not allow us to distinguish types of risk management action taken. This prevents us fromperforming a more detailed analysis of effects of risk management. Second, disclosure is required for materialinformation. As our results show, we do not capture remote or minor risk and its management. Thus, findingsshould be interpreted within the scope of material risk and risk management. Finally, a more general concern couldarise over the extent to which incentives self-servingly bias environmental risk disclosure. Recent environmentaldisclosure research in the US finds that environmental performance and environmental disclosure are positivelyrelated, as predicted by discretionary disclosure theory (Al-Tuwaijri et al., 2004; Clarkson et al., 2008). This suggeststhat incentives would work against the direction of our results when assuming that environmental performance ispositively related to the likelihood of disclosure of material risk and risk management information. Adoptingdiscretionary disclosure theory to risk-related disclosure in particular indicates partly countervailing incentives(Dobler, 2008; Dobler et al., 2011). So do recent analytical results on ’greenwashing’ through environmentaldisclosure by Lyon and Maxwell (2011). However, we consider any potential bias imposed to be minor in ourstudy. This is due to the highly regulated and enforced disclosure regime in the US that limits disclosure flexibilityin 10-K filings (Hughes et al., 2001; SEC, 2010). Disclosure incentives are likely to be a more substantial concernwhen adopting our content analysis approach in countries with less sophisticated risk disclosure regulation orpoor disclosure enforcement or to poorly enforced environmental disclosure, such as on corporate websitesanalyzed, e.g., by Jose and Lee (2007) or Morhardt (2010).
Considering the informativeness of corporate environmental risk disclosure and the differential nature of relation-ships, our study suggests that the extent of environmental risk implied by a firm’s disclosure can be used as a noisyinverse indicator for its environmental performance as measured by TRI data and vice versa. Prior research concludesthat corporate environmental disclosure is inferior to data provided by public registers in assessing environmentalperformance (Gouldson and Sullivan, 2007; Sullivan and Gouldson, 2007). However, our results indicate that environ-mental risk disclosure in recent 10-K filings seems to complement information provided by public registers. This hasan important implication for public policy makers. Concurrent with assumed disciplining effects of the release of TRIdata on a firm’s polluting behavior (Konar and Cohen, 1997; Stephan, 2002), mandated and enforced environmentalrisk disclosure can be seen as a complementary regulatory instrument. As a means of transparency, it could be furtheremployed to induce firms to reduce and manage environmental risk in order to mitigate interventions by investors andother stakeholders upon the release of high levels of environmental risk or lack of respective risk management. Mostparticularly, it could address risk from operations which has been identified as a key determinant of environmentalperformance beyond risk from regulations in our study. Best practice and guidance on how to report on this type ofenvironmental risk might be warranted and will help researchers study in more depth the linkages between environ-mental performance, risk and risk management in the future.
Acknowledging efforts necessary and still limited access to adequate data, we conclude that linkages to environ-mental risk and respective risk management are worth exploring in more differential ways and beyond industry-levelassessments in environmental studies.
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Acknowledgements
The authors gratefully acknowledge the financial support of the CGA Accounting Research Centre at the University of Ottawaand wish to thank the editor (Richard Welford) and the anonymous reviewers for their valuable comments and suggestions.
References
Ählström J, Macquet M, Richter U. 2009. The lack of a critical perspective in environmental management research: distortion in the scientificdiscourse. Business Strategy and the Environment 18(5): 334–346. DOI: 10.1002/bse.592
Aintablian S, Mcgraw PA, Roberts GS. 2007. Bank monitoring and environmental risk. Journal of Business Finance and Accounting 34(1/2):389–401.
Aktas N, de Bodt E, Cousin J-G. 2011. Do financial markets care about SRI? Evidence from mergers and acquisitions. Journal of Banking andFinance 35(7): 1753–1761.
Al-Najjar B, Anfimiadou A. 2012. Environmental policies and firm value. Business Strategy and the Environment 21(1): 49–59. DOI: 10.1002/bse.713Al-Tuwaijri SA, Christensen TE, Hughes KE II. 2004. The relations among environmental disclosure, environmental performance, and
economic performance: a simultaneous equations approach. Accounting, Organizations and Society 29(5/6): 447–471.Ambec S, Lanoie P. 2008. Does it pay to be green? A systematic overview. Academy of Management Perspectives 22(4): 45–62.Anderson DR, Anderson KE. 2009. Sustainability risk management. Risk Management and Insurance Review 12(1): 25–38.Anghelache C. 2011. Management of the environmental risk – an economic–social priority. Theoretical and Applied Economics 18(3): 117–130.Annandale D, Morrison-Saunders A, Bouma G. 2004. The impact of voluntary environmental protection instruments on company environmental
performance. Business Strategy and the Environment 13(1): 1–12. DOI: 10.1002/bse.390Aven T. 2012. The risk concept – historical and recent development trends. Reliability Engineering and System Safety 99: 33–44.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(1): 93–103.Barrieu P, Sinclair-Desgagné B. 2006. On precautionary policies. Management Science 52(8): 1145–1154.Baum A, Fleming R, Davidson LM. 1983. Natural disaster and technological catastrophe. Environment and Behavior 15(3): 333–354.Berrone P, Gomez-Mejia LR. 2009. Environmental performance and executive compensation: an integrated agency-institutional perspective.
Academy of Management Journal 52(1): 103–126.Berrone P, Cruz C, Gomez-Mejia LR, Larraza-Kintana M. 2010. Socioemotional wealth and corporate responses to institutional pressures: do
family-controlled firms pollute less? Administrative Science Quarterly 55(1): 82–113.Bewley K, Li Y. 2000. Disclosure of environmental information by Canadian manufacturing companies: a voluntary disclosure perspective.
Advances in Environmental Accounting and Management 1: 201–226.Böhm, G, Pfister H-R. 2005. Consequences, morality, and time in environmental risk evaluation. Journal of Risk Research 8(6): 461–479.Bowen FE. 2000. Environmental visibility: a trigger of green organizational response? Business Strategy and the Environment 9(2): 92–107.
DOI: 10.1002/(SICI)1099-0836(200003/04)9:2<92::AID-BSE230> 3.0.CO;2-Xvan Buerden P, Gössling T. 2008. The worth of values – a literature review on the relation between corporate social and financial performance.
Journal of Business Ethics 82(2): 407–424.Burritt RL, Hahn T, Schaltegger S. 2002. Towards a comprehensive framework for environmental management accounting – links between
business actors and environmental management accounting tools. Australian Accounting Review 12(2): 39–50.Campbell JL. 2007. Why would corporations behave in socially responsible ways? An institutional theory of corporate social responsibility.
Academy of Management Review 32(3): 946–967.Chan JC, Welford R. 2005. Assessing corporate environmental risk in China: an evaluation of reporting activities of Hong Kong listed enterprises.
Corporate Social Responsibility and Environmental Management 12(2): 88–104. DOI: 10.1002/csr.88Chatterji AK, Levine DI, Toffel MW. 2009. How well do socially ratings actually measure corporate social responsibility? Journal of Economics and
Management Strategy 18(1): 125–169.Cho CH, Patten DM. 2008. Did the GAO get it right? Another look at corporate environmental disclosure. Social and Environmental Accountability
Journal 28(1): 21–32.Clarkson PM, Li Y, Richardson GD, Vasvari FP. 2008. Revisiting the relation between environmental performance and environmental disclosure:
an empirical analysis. Accounting, Organizations and Society 33(4/5): 303–327.Clarkson PM, Li Y, Richardson GD, Vasvari FP. 2011. Does it really pay to be green? Determinants and consequences of proactive environmental
strategies. Journal of Accounting and Public Policy 30(2): 122–144.Cuddihy T. 2000. Environmental liability risk management for the 21st century. The Geneva Papers on Risk and Insurance – Issues and Practice
25(1): 128–135.Delmas M, Montiel I. 2009. Greening the supply chain: when is customer pressure effective? Journal of Economics and Management Strategy
18(1): 171–201.Delmas MA, Toffel MW. 2008. Organizational responses to environmental demands: opening the black box. Strategic Management Journal
29(10): 1027–1055.
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Dobler M. 2008. Incentives for risk reporting – a discretionary disclosure and cheap talk approach. The International Journal of Accounting 43(2):184–206.
Dobler M, Lajili K, Zéghal D. 2011. Attributes of corporate risk disclosure: an international investigation in the manufacturing sector. Journal ofInternational Accounting Research 10(2): 1–22.
Ellis L, Bastin C. 2011. Corporate social responsibility in times of recession: changing discourses and implications for policy and practice.Corporate Social Responsibility and Environmental Management 18(5): 294–305. DOI: 10.1002/csr.254
Ferreira A, Moulang C, Hendro B. 2010. Environmental management accounting and innovation: an exploratory analysis. Accounting, Auditingand Accountability Journal 23(7): 920–948.
Freedman M, Wasley C. 1990. The association between environmental performance and environmental disclosure in annual reports and 10 Ks.Advances in Public Interest Accounting 3: 183–193.
Gerde VW, Logsdon JM. 2001. Measuring environmental performance: use of the Toxics Release Inventory (TRI) and other US environmentaldatabases. Business Strategy and the Environment 10(5): 269–285. DOI: 10.1002/bse.293
Gouldson A. 2004. Risk, regulation and the right to know: exploring the impacts of access to information on the governance of environmentalrisk. Sustainable Development 12(3): 136–149.
Gouldson A, Sullivan R. 2007. Corporate environmentalism: tracing the links between policies and performance using corporate reports andpublic registers. Business Strategy and the Environment 16(1): 1–11. DOI: 10.1002/bse.543
Gregory R, Failing L, Ohlson D, McDaniels TL. 2006. Some pitfalls of an overemphasis on science in environmental risk management decisions.Journal of Risk Research 9(7): 717–735.
Hart SL, Ahuja G. 1996. Does it pay to be green? An empirical examination of the relationship between emission reduction and firm perfor-mance. Business Strategy and the Environment 5(1): 30–37. DOI: 10.1002/(SICI)1099-0836(199603)
Hart SL, Dowell G. 2011. A natural-resource-based view of the firm: fifteen years after. Journal of Management 37(5): 1464–1479.Heinkel R, Kraus A, Zechner J. 2001. The effect of green investment on corporate behavior. The Journal of Financial and Quantitative Analysis
36(4): 431–449.Hofer C, Cantor DE, Dai J. 2012. The competitive determinants of a firm’s environmental management activities: evidence from US manufac-
turing industries. Journal of Operations Management 30(1/2): 69–84.Hughes SB, Anderson A, Golden S. 2001. Corporate environmental disclosures: are they useful in determining environmental performance?
Journal of Accounting and Public Policy 20(3): 217–240.Iraldo F, Testa F, Frey M. 2009. Is an environmental management system able to influence environmental and competitive performance? The
case of the eco-management and audit scheme (EMAS) in the European Union. Journal of Cleaner Production 17(16): 1444–1452.Jose A, Lee S-M. 2007. Environmental reporting of global corporations: a content analysis based on website disclosures. Journal of Business Ethics
72(4): 307–321.Kaplan S, Garrick BJ. 1981. On the quantitative definition of risk. Risk Analysis 1(1): 11–27.Karpoff JM, Lott JR Jr, Wehrly EW. 2005. The reputational penalties for environmental violations: empirical evidence. Journal of Law and Economics
48(2): 653–675.Kassinis G, Vafeas N. 2006. Stakeholder pressures and environmental performance. Academy of Management Journal 49(1): 145–159.Kennedy MJ. 2001. The management of environmental risk in a global industrial company. Corporate Environmental Strategy 8(2): 177–185.King AA, Lenox MJ. 2001. Does it really pay to be green? An empirical study of firm environmental and financial performance. Journal of
Industrial Ecology 5(1): 105–116.King AA, Shaver JM. 2001. Are aliens green? Assessing foreign establishments’ environmental conduct in the United States. Strategic
Management Journal 22(11): 1069–1085.Klassen RD, McLaughlin CP. 1996. The impact of environmental performance on firm performance. Management Science 42(8): 1199–1214.Kock CJ, Santaló J, Diestre L. 2012. Corporate governance and the environment: what type of governance creates greener companies? Journal of
Management Studies 49(3): 492–514.Konar S, Cohen MA. 1997. Information as regulation: the effect of community right to know laws on toxic emissions. Journal of Environmental
Economics and Management 32(1): 109–124.Konar S, Cohen MA. 2001. Does the market value environmental performance? The Review of Economics and Statistics 83(2): 281–289.Lang MH, Lundholm RJ. 1996. Corporate disclosure policy and analyst behavior. The Accounting Review 71(4): 467–492.Lyon TP, Maxwell JW. 2011. Greenwash: corporate environmental disclosure under threat of audit. Journal of Economics and Management Strategy
20(1): 3–41.Matten D. 1995. Strategy follows structure: environmental risk management in commercial enterprises. Business Strategy and the Environment
4(3): 107–116. DOI: 10.1002/bse.3280040302Melnyk SA, Sroufe RP, Calantone, R. 2003. Assessing the impact of environmental management systems on corporate and environmental
performance. Journal of Operations Management 21(3): 329–351.Molina-Azorín JF, Claver-Cortés E, López-Gamero MD, Tarí JJ. 2009. Green management and financial performance: a literature review.
Management Decision 47(7): 1080–1100.Morhardt JE. 2010. Corporate social responsibility and sustainability reporting on the internet. Business Strategy and the Environment 19(7):
436–452. DOI: 10.1002/bse.657Orlitzky M, Benjamin JD. 2001. Corporate social performance and firm risk: a meta-analytic review. Business and Society 40(4): 369–396.Orlitzky M, Schmidt FL, Reyes SL. 2003. Corporate social and financial performance: a meta-analysis. Organization Studies 24(3): 403–441.Patten DM. 2002. The relation between environmental performance and environmental disclosure: a research note. Accounting, Organizations
and Society 27(8): 763–773.
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Posner RA. 1974. Theories of economic regulation. Bell Journal of Economics and Management Science 5(2): 335–358.Qi G, Zeng S, Li X, Tam C. 2012. Role of internalization process in defining the relationship between ISO 14001 certification and corporate
environmental performance. Corporate Social Responsibility and Environmental Management 19(3): 129–140. DOI: 10.1002/csr.258Sarkis J, Gonzalez-Torre P, Adenso-Diaz B. 2010. Stakeholder pressure and the adoption of environmental practices: the mediating effect of
training. Journal of Operations Management 28(2): 163–176.Securities and Exchange Commission (SEC). 2010. Commission Guidance Regarding Disclosure Related to Climate Change, Release Nos.
33–9106, 34–61469, FR-82. http://www.sec.gov/rules/interp/2010/33-9106.pdf [05 July 2012].Semenova N. 2010. Corporate Environmental Performance: Consistency of Metrics and Identification of Drivers, Working Paper SIRP 10–09,
Åbo Akademi University. http://swoba.hhs.se/sicgwp/abs/sicgwp2010_009.htm [05 July 2012].Semenova N, Hassel LG. 2008. Financial outcomes of environmental risk and opportunity for US companies. Sustainable Development 16(3):
195–212.Sharfman MP, Fernando CS. 2008. Environmental risk management and the cost of capital. Strategic Management Journal 29(6): 569–592.Siegel DS. 2009. Green management matters only if it yields more green: an economic/strategic perspective. Academy of Management Perspectives
23(3): 5–16.Stephan M. 2002. Environmental information disclosure programs: they work, but why? Social Science Quarterly 83(1): 190–205.Sullivan R, Gouldson A. 2007. Pollutant release and transfer registers: examining the value of government-led reporting on corporate
environmental performance. Corporate Social Responsibility and Environmental Management 14(5): 263–273. DOI: 10.1002/csr.148de Villiers C, Naiker V, van Staden CJ. 2011. The effect of board characteristics on firm environmental performance. Journal of Management 37(6):
1636–1663.Walls JL, Berrone P, Phan PH. 2012. Corporate governance and environmental performance: is there really a link? Strategic Management Journal
33(8): 885–913.Weber O, Fenchel M, Scholz RW. 2008. Empirical analysis of the integration of environmental risks into the credit risk management process of
European banks. Business Strategy and the Environment 17(3): 149–159. DOI: 10.1002/bse.507Welford RJ. 1998. Corporate environmental management, technology and sustainable development: postmodern perspectives and the
need for a critical research agenda. Business Strategy and the Environment 7(1): 1–12. DOI: 10.1002/(SICI)1099-0836(199802)7:1< 1::AID-BSE132> 3.0.CO;2–7
Winn MI, Kirchgeorg M, Griffiths A, Linnenluecke MK, Günther E. 2011. Impacts from climate change on organizations: a conceptual foundation.Business Strategy and the Environment 20(3): 157–173. DOI: 10.1002/bse.679
Wiseman J. 1982. An evaluation of environmental disclosures made in corporate annual reports. Accounting, Organizations and Society 7(1): 53–63.Zhu Q, Sarkis J. 2004. Relationships between operational practices and performance among early adopters of green supply chain management
practices in Chinese manufacturing enterprises. Journal of Operations Management 22(3): 265–289.
M. Dobler et al.
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse
Appendix. Inter-Industry Differences
Energy vs.Industrials
Energy vs.Materials
Energy vs.Utilities
Industrials vs.Materials
Industrials vs.Utilities
Materials vs.Utilities
%RECYC Mean difference(significance t test)
�0.074 0.111 0.290** 0.185** 0.364*** 0.179*
(0.401) (0.208) (0.010) (0.021) (0.000) (0.051)Z value (significanceMann–Whitney test)
�0.559 �2.086** �3.304*** �2.246** �4.573*** �1.907*
(0.576) (0.037) (0.001) (0.025) (0.000) (0.052)–TRI/SAL Mean difference
(significance t test)0.040 0.928* 0.549*** 0.888 0.509*** �0.379(0.388) (0.093) (0.000) (0.147) (0.000) (0.515)
Z value (significanceMann–Whitney test)
�0.686 �3.939*** �5.145*** �4.189*** �5.185*** �1.567(0.493) (0.000) (0.000) (0.000) (0.000) (0.128)
RISK_R Mean difference(significance t test)
1.583*** 1.037** 0.400 �0.546 �1.183** �0.637(0.009) (0.043) (0.353) (0.323) (0.033) (0.177)
Z value (significanceMann–Whitney test)
�2.551** �1.984** �0.934 �0.959 �2.098** �1.357(0.011) (0.047) (0.351) (0.338) (0.036) (0.175)
RISK_O Mean difference(significance t test)
2.875*** 1.741* �0.583 �1.134 �3.458*** �2.324**
(0.002) (0.091) (0.565) (0.177) (0.000) (0.016)Z value (significanceMann–Whitney test)
�3.105*** �2.078** �0.814 �0.925 �3.676*** �2.699***
(0.002) (0.038) (0.416) (0.355) (0.000) (0.007)RISK_N Mean difference
(significance t test)1.847** 1.741** 0.006 �0.106 �1.842** �1.735**
(0.031) (0.030) (0.995) (0.872) (0.033) (0.031)Z value (significanceMann–Whitney test)
�2.420 �2.297** �0.180 �0.874 �2.550** �2.362**
(0.016) (0.022) (0.857) (0.382) (0.011) (0.018)RISK_RO Mean difference
(significance t test)4.459*** 2.778** �0.183 �1.681 �4.642*** �2.961***
(0.000) (0.024) (0.877) (0.103) (0.000) (0.009)Z value (significanceMann–Whitney test)
�3.819*** �2.785*** �0.679 �1.183 �4.030*** �2.897***
(0.000) (0.005) (0.497) (0.237) (0.000) (0.004)RM_R Mean difference
(significance t test)0.014 �0.019 �0.128 �0.032 �0.142 �0.109(0.924) (0.894) (0.348) (0.801) (0.275) (0.376)
Z value (significanceMann–Whitney test)
�0.097 �0.136 �0.952 �0.256 �1.104 �0.895(0.922) (0.892) (0.341) (0.798) (0.270) (0.371)
RM_O Mean difference(significance t test)
0.306** 0.315*** �0.311* 0.009 �0.617*** �0.626***
(0.016) (0.009) (0.057) (0.905) (0.000) (0.000)Z value (significanceMann–Whitney test)
�2.360** �2.558** �1.900* �0.122 �4.186*** �4.430***
(0.018) (0.011) (0.057) (0.903) (0.000) (0.000)RM_N Mean difference
(significance t test)0.167 0.130 0.067 �0.037 �0.100 �0.063(0.039)** (0.141) (0.561) (0.351) (0.118) (0.394)
Z value (significanceMann–Whitney test)
�2.051** �1.480 �0.599 �0.943 �1.568 �0.864(0.040) (0.139) (0.549) (0.346) (0.117) (0.388)
RM_RO Mean difference(significance t test)
�0.069 0.000 �0.244 0.069 �0.175 �0.244*
(0.661) (1.000) (0.111) (0.623) (0.214) (0.084)Z value (significanceMann–Whitney test)
�0.448 0.000 �1.597 �0.498 �1.252 �1.730*
(0.654) (1.000) (0.111) (0.618) (0.210) (0.084)
Variables are defined in Table 3.***/**/* Significance at 1%, 5% and 10%, respectively.
Environmental Performance, Environmental Risk and Risk Management
Copyright © 2012 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. (2012)DOI: 10.1002/bse