2006 - Gouldson - Corporate Environmentalism Tracing The Links Between Policies and Performance Using

Business Strategy and the Environment
Bus. Strat. Env. 16, 1–11 (2007)

Published online 16 May 2006 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/bse.543
Corporate Environmentalism: Tracing

the Links between Policies and
Performance Using Corporate Reports
and Public Registers
Andy Gouldson1* and Rory Sullivan2
1
Sustainability Research Institute, School of Earth and Environmental Sciences,
University of Leeds, UK
2
Investor Responsibility, Insight Investment, London, UK
ABSTRACT
Within this paper, we consider whether it is possible to trace the links between the
procedural and the substantive dimensions of corporate environmentalism using
information that is in the public domain – most notably in corporate reports and in
pollutant releases and transfer registers (PRTRs) such as the US Toxics Release Inven-
tory (TRI) and the EU Polluting Emissions Register (EPER). Based on an analysis of
firms in the oil and gas sector, and specifically of the environmental performance of
oil refineries, we find that corporate reports are of very limited value when seeking to
compare and contrast levels of environmental performance at the site level, but that
a significant body of useful information is provided by public registers such as the
US TRI and the EU EPER. Drawing upon these data, we find significant variations in
corporate environmental performance across the US and the EU, and we note the
existence of significant correlations between higher levels of emissions and lower
levels of employment and income in the areas where industrial facilities are located.
We then discuss the relevance of our findings to broader debates on corporate envi-
ronmentalism and corporate social responsibility. Copyright © 2006 John Wiley &
Sons, Ltd and ERP Environment.
Received 1 November 2005; revised 21 January 2006; accepted 21 February 2006

Keywords: corporate environmental performance; corporate environmental reporting; oil and gas; PRTRs; TRI; EPER;
environmental justice
* Correspondence to: Professor Andy Gouldson, Director, Sustainability Research Institute, School of Earth and Environmental Sciences, Uni-
versity of Leeds, Leeds LS2 9JT, UK. E-mail: a.gouldson@see.leeds.ac.uk & rory.sullivan@insightinvestment.com
Copyright © 2006 John Wiley & Sons, Ltd and ERP Environment
10990836, 2007, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/bse.543 by Gavle University College, Wiley Online Library on [16/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 A. Gouldson and R. Sullivan
Introduction
T
HE PAST TEN YEARS HAVE SEEN A RAPID GROWTH IN THE NUMBER OF COMPANIES THAT HAVE
made explicit policy commitments to ‘corporate responsibility’ or ‘corporate social responsibil-
ity’ (hereafter referred to as ‘CSR’), and that have subsequently developed and implemented
management systems to deliver on these commitments and published reports that outline their
activities and include some measures of performance. Amongst the largest firms with stock market
listings, levels of participation in such initiatives are high. In their survey of corporate responsibility
reporting, KPMG (2005) find that 52% of companies in the Global Fortune 250 index publish corporate
responsibility reports, and that 64% include information on corporate responsibility in their annual
financial reports. However, levels of participation in corporate responsibility initiatives vary dramatically
according to company size, industrial sector and country of origin.
CSR initiatives, in most western companies at least, usually start with some sort of ‘top level’ policy
commitment, for example to ‘supporting the goals of sustainable development’ or ‘contributing to the
protection of human rights’. Thereafter, CSR initiatives have both procedural dimensions, relating to
policies and processes, and substantive dimensions, relating to performance and outcomes. To date, the
bulk of corporate activity has related to the procedural dimensions of CSR; companies have developed
policies, management systems and audit protocols and they have started to open up and engage with
stakeholders. Recently, however, more emphasis is being placed on the substantive implications of
corporate activities. Stakeholders are increasingly keen to know not only whether companies have CSR
policies and procedures in place, but also whether these have gone on to influence the social and
environmental outcomes that are experienced by different social groups in the diverse contexts in which
companies operate. Answering these questions requires that credible information is available to allow
company performance to be assessed both against appropriate standards of acceptability and against the
performance of other different companies.
It is these questions that this article examines. More specifically, we consider whether it is possible
to trace the links between the procedural and the substantive dimensions of CSR using information that
is in the public domain – most notably in corporate reports and in government-led pollutant releases
and transfer registers (PRTRs) such as the US Toxics Release Inventory (TRI) or the EU Polluting Emis-
sions Register (EPER) – and, if so, to consider what the available data tells us about different levels of
CSR performance in different settings. While our primary objective is to inform debates on the focus
and influence of CSR initiatives, as a secondary goal we seek to contribute to related debates on envi-
ronmental justice. A central claim that is often made within these debates is that levels of environmental
protection are uneven and that poorer areas and populations are likely to be exposed to higher levels of
emissions and lower levels of environmental quality (Agyeman, 2002; Bullard and Johnson, 2000;
Dobson, 1998; Shrader-Frechette, 2002; US EPA, 2003; Zarsky, 2002). By examining whether there
are any links between variations in corporate environmental performance and the socio-economic status
of the areas where industrial facilities are located, we can offer evidence that supports or refutes these
claims.
The discussion starts with a brief overview of the research project and the data sets used in the analy-
sis. We then move on to consider the main findings from the research. We find that the links between
policies, procedures and performance in the CSR sphere are either not well developed or are not fully
transparent. We find that corporate reports are of very limited value when seeking to compare and con-
trast levels of environmental performance at the site level, but that a significant body of useful infor-
mation is provided by public registers such as the US TRI and the EU EPER. Drawing upon this data,
we find significant variations in corporate environmental performance across the US and the EU, and
Copyright © 2006 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 16, 1–11 (2007)
DOI: 10.1002/bse
Tracing the Links between Policies and Performance 3
we note the existence of significant correlations between higher levels of emissions and lower levels of
employment and income in the areas where industrial facilities are located. We then discuss the rele-
vance of our findings to broader debates on CSR. We highlight some of the limits of existing approaches
to CSR, and we discuss the role that PRTRs can play in enabling a shift towards less procedural and
more outcome-oriented approaches to CSR.
Project Overview
This paper reports the results of collaborative research conducted between December 2003 and January
2005, which sought to examine the links between CSR and environmental justice.1 The main body of
the research was divided into two phases. The first phase involved an examination of companies’ cor-
porate governance processes, which involved a qualitative evaluation of the ways in which corporate poli-
cies on CSR are translated into specific, site-level practices. This phase of the research focused on the
corporate environmental reports relating to 20022 and material presented on corporate websites between
January and March 2004. This information was evaluated using a set of questions that sought to trace
the links between policies, procedures and performance. These questions focused on the presence of
corporate environmental policies and reports and, within these reports, the transparency of the proce-
dures used to translate policies into practice, the inclusion of explicit standards to be complied with at
the local level, the provision of data on the environmental performance of specific sites and the provi-
sion of information on links between levels of emissions and environmental conditions in particular
areas.
The second phase involved an evaluation of corporate environmental performance, which sought to
link site-level practices with quantitative measures of environmental performance. More particularly, it
sought to establish whether there were any variations in corporate environmental performance at the
site level across the European Union (EU) and the United States (US) and to examine whether these
variations were related in any way to differences in the socio-economic conditions in the areas where
industrial sites were located. Geographically, we focused on the EU and the US as these were the main
areas where information on site-level environmental performance had been made available through the
publication of PRTRs (the TRI in the US and the EPER in the EU). To evaluate socio-economic condi-
tions in the areas where industrial sites are located, we used 2001 data from the US census, Eurostat
and the statistics agencies of the relevant EU member states. Data were collected at two spatial scales,
namely counties and the smaller districts within them. Whilst there were problems with gaining access
to consistent, reliable socio-economic data,3 comparable data from various national census agencies and
research bodies was collected. Although we used the most complete and the most recent data sets avail-
able at the time of the research, we have to recognize that the data sets were not perfectly comparable4
and that analysis based on more recent and more consistent data sets may paint a slightly different
1
The project was jointly funded by the Higher Education Innovations Fund at the London School of Economics and Insight Investment, the
asset management arm of HBOS plc, one of Britain’s largest banks.
2
Company CSR reports are generally published some six months following the (reporting or calendar) year end. Hence, at the time of com-
mencing the research (i.e. at the end of 2003), the most recent data that were available related to 2002.
3
Specifically, the geographical scope of the counties and districts considered varied both between the US and the EU and within the different
countries of the EU. Following conventions established by previous research, we took the areas under consideration to be broadly equivalent.
Furthermore, while data were available at both scales in the US, data were only available at the county level for all such areas in the EU. Dis-
trict level data were only available from certain national government agencies and research bodies, and so we narrowed the boundaries of this
part of the analysis so that it focused only on refineries in the UK, Germany, France and Italy.
4
The two data sets were not perfectly comparable; at the time of the research the most recent emissions data that were available were for 1999
in the US and for 2001 in the EU. Furthermore, the methods used to collect the data may vary slightly both between the US and the EU and
between the member states within the EU.
DOI: 10.1002/bse
picture. Nonetheless, it is our view that the implications of these restrictions in data availability/consis-
tency are unlikely to be significant and that the results of the analysis still offer useful insights into com-
parative levels of corporate environmental performance. More information on the methods used to
collect and analyse the data is presented at the relevant points in the discussion that follows.
Empirically, the project focused on the nine companies within the oil and gas sector that were in the
top 250 of the companies listed on the London stock exchange. To allow benchmarks of corporate envi-
ronmental performance to be developed, the project focused on one relatively generic process within
this sector, namely oil refining. This focus was adopted for a number of reasons. Perhaps most impor-
tantly, oil refineries are relatively homogenous across the world and it is, therefore, possible to make
meaningful comparisons between the environmental performance of different facilities. In addition,
because of their scale and environmental impacts, oil refinery operators are generally required to make
a significant amount of information publicly available, through both corporate reports and PRTRs,
thereby allowing comparisons between facilities. Because of public concerns about the environmental
impacts of oil refining (and of the oil and gas industry more generally), the bigger companies at least
have invested significant resources managing environmental and social issues and these companies have
been to the forefront of many of the discussions on corporate responsibility. Because of this leadership
position, our view was that this sector could provide something of a litmus test that could inform the
development of CSR in other sectors.
Key Research Findings
Phase 1: Corporate Governance Processes

The first part of the study involved an analysis of the corporate governance processes of the nine largest
UK-listed oil and gas companies. This analysis involved a literature review of the companies’ published
information, as well as one-on-one interviews with the HSE managers of two major oil companies. The
results of the research were subsequently presented to BP, BG, Shell and the International Petroleum
Industry Environment and Conservation Association’s (IPIECA) Social Responsibility Working Group.
These presentations and the associated discussions provided some additional information that was incor-
porated into the final report.
The results of this phase were mixed: we found that the three larger oil and gas companies (BP, BG
and Shell) stood out in many respects from the six smaller firms in the sector that were also listed in
the FTSE250 (Abbott, Cairn, Palladin, Premier, Tullow and Wood). The major differences related to the
level of development of the policies, procedures and reporting frameworks adopted by the firms in areas
relating to the environmental aspects of CSR and to the amount of information on these aspects that
was placed in the public domain. Given these discrepancies, the decision was taken to focus on the
leaders – BP, BG and Shell – rather than the laggards.
Each of the three larger companies within the sector had well established corporate policies relating
to the environmental aspects of CSR. These policies related not only to their core activities, but also to
their subsidiaries and to any joint ventures with which they were associated. Each of the firms set out
some aspirations and targets for corporate environmental performance and they offered some detail on
the procedures used to translate corporate policies into site-level practices. They also provided informa-
tion on how site-level activities were monitored, and how their performance was audited and verified.
In addition, the firms made explicit commitments relating to some of the procedural issues that under-
pin CSR – for example, they each had clear statements on transparency, engagement, equity and human
rights. Their commitments on transparency were reflected in the publication of corporate reports that
provided externally verified data on their aggregated environmental impacts.
DOI: 10.1002/bse
With regard to their policies and procedures, each of the companies stated that all of their sites had
to contribute to corporate aspirations and targets and to comply with all relevant laws and regulations.
Although in some instances they were guided by international standards,5 decisions on site-specific stan-
dards tended to be devolved to the site level. In the case of BP, for example, site-specific standard setting
was based on deliberations between local managers and their consultants or certifiers on the significant
environmental aspects to be addressed by environmental management systems such as ISO14001. As
well as presenting very limited levels of transparency and no opportunities for direct stakeholder
involvement, such a devolved decision-making process does not necessarily ensure consistency. All of
the companies suggested that levels of environmental performance varied from site to site. It is impor-
tant to recognize that such variations do not necessarily imply failures of management, as some varia-
tion would be expected anyway as a consequence of local legal requirements and local air quality issues.
Furthermore, given that none of the companies committed to ‘best practice’ across all facilities, some
variation in emissions performance could be expected.
With regard to the provision of data on site-level environmental performance, there was considerable
variation between the three firms. One of the three (BP) routinely provided such information through
site-level environmental reports, one (Shell) provided site-level reports for particular ‘hot spots’ and one
(BG) did not publish site-level data. Although BP and Shell stated that they used a common reporting
framework to release such site-level data, only BP’s site-level reports allowed the environmental perfor-
mance of individual sites to be evaluated and tracked over time or the performance of different sites to
be compared and contrasted. Even then, performance could only be compared with other BP sites. Con-
sequently, it was impossible for the environmental aspects of CSR to be compared and contrasted on a
consistent basis across the sector. Furthermore, within the corporate and site level reports, little or no
reference was made to key outcomes such as levels of local air quality or the health of local populations.
Thus, even amongst the leaders in the sector, corporate governance processes focused on policies and
procedures and on aggregated, corporate-level performance measures. Although all of the companies
acknowledged that both site-level standards and performance varied, the absence of clear, consistent and
reliable data meant that it was impossible for stakeholders to compare and contrast the performance of
different sites. The focus on emissions rather than, for example, local air quality, also meant that it was
impossible to evaluate social or environmental outcomes at the local level, thereby restricting the ability
of stakeholders to make informed decisions or to focus their engagement with the companies on indi-
vidual sites or on particular aspects of their performance.
Phase 2: Evaluating Corporate Environmental Performance

Phase 1 showed that even the leading companies in the sector examined rarely disclosed site-level data
and, where they did so, these data were not provided on a consistent basis or in a common format.
However, this does not mean that no such information is in the public domain. Increasingly data on
corporate environmental performance are being made available by governments and regulatory agen-
cies through the publication of PRTRs (see OECD, 1996). Typically, such registers publish on-line data
on the regulated emissions from industrial sites, thereby establishing an opportunity for stakeholders
to find out about the environmental performance of different industrial sites and to make judgements
about their impacts on health and the environment. Starting with the publication of the Toxic Releases
5
For example, both BP and BG claimed in interviews that if, for example, local air quality exceeded WHO or EU limits, this would be picked
up in local level reviews and tighter emissions standards would be set. However, the extent of their local environmental monitoring activities,
and the nature of the decision-making processes that might be triggered by a breach of minimum standards, were not discussed in any of
their corporate reports.
DOI: 10.1002/bse
Variables Units Data sources
Refinery capacity Barrels per annum Oil and Gas Journal (OGJ)
Refinery complexity Type 1, 2 or 3 OGJ
Refinery emissions Sulphur dioxide and benzene emissions per EU – EPER, US – TRI/NEI (1999)
annum (metric tons)
Home county/district GDP per capita (expressed as purchasing power EU – Eurostat (NUTS 3) and national
characteristics parity), unemployment (indexed against national agencies (LAU2)
averages), population density (people per km2) US – census (counties and congressional
districts)
Table 1. Data included in the quantitative study: 88 EU refineries, 144 US refineries (all data are for 2001 unless otherwise stated)
Inventory in the US in the late 1980s, such registers have spread and are now in place in Canada,
Australia, Japan and, most recently, across the EU as a result of the creation of the European Polluting
Emissions Register (EPER) in 2003. Such developments may be significant: previous research has found
that the provision of such access to information can have a significant effect both on the governance
processes that exist around industrial sites (i.e. the nature and influence of the relationships between
regulators, stakeholders and firms) and on the environmental performance of these sites (see Konar and
Cohen, 1997; Terry and Yandle, 1997; Khanna et al., 1998; Stephan, 2002, Gouldson, 2004).
Using the data released by PRTRs as the basis for Phase 2 of the project, we developed benchmarks
of environmental performance for all of the oil refineries in the EU and the US (see Table 1). These data
allowed us to compare and contrast the aggregated, annual emissions from approximately 240 refineries
across the EU and the US. We focused particularly on two pollutants – sulphur dioxide and benzene.
These are two of the key emissions from oil refineries with impacts on local and regional air quality and
on public health, and data regarding emissions of these pollutants are available for all refineries in the
US and EU. For these pollutants, we developed two indicators. The first of these was an eco-efficiency
indicator, which focused on emissions per unit of capacity. This is an important indicator for analysts
interested in the relative efficiencies of refineries. The second indicator was for total emissions of each
pollutant from each refinery. This is an important indicator for local communities, which are more inter-
ested in the absolute levels of emissions that might influence local health and environment than in emis-
sions per unit of production.
The results of the benchmarking study showed considerable variations in corporate environmental
performance, with wide variation in pollutant emissions from the different refineries. Even when focus-
ing on variations across the inter-quartile range (i.e. the refineries at the top and bottom of the central
50%), the ratio of the highest to the lowest emissions was significant: for sulphur dioxide and benzene,
total site emissions varied by factors of 7.4 and 7.3 respectively, whilst eco-efficiency measures varied by
factors of 5.4 and 5.3 respectively. Although to some extent these variations reflected differences in the
complexity of the refineries, they existed despite the presence of regulations in both the EU and the US
that oblige refinery operators to adopt the best available technologies and techniques and/or to comply
with particular emissions limits.
In addition to finding significant variations across the range of the refineries, there were also dra-
matic variations between levels of environmental performance in the US and the EU. For sulphur dioxide
and benzene, average total site emissions in the EU were, respectively, 2.4 and 4.4 times higher than
those in the US, while average levels of eco-efficiency in the EU were 1.9 and 2.5 times worse than those
in the US. In summary then, the benchmarking study found substantial variations in the environmen-
tal performance of oil refineries, and that these variations exist both across the range of refineries, and
DOI: 10.1002/bse
between the EU and the US. Such findings support the results of previous studies, which have also high-
lighted the variability of emissions from oil refineries (see EDF, 1999; PDC, 2004).
As well as evaluating variations in corporate environmental performance, we were also able to collect
data on the social and economic characteristics of the communities that lived in the areas where the
refineries were located. This enabled us to examine some of the key aspects of environmental justice –
namely whether emissions from oil refineries were higher in more deprived areas. As indicators of social
deprivation, we used data relating to unemployment, per capita GDP and population density (see Table
1). We examined the relationship between environmental performance and the socio-economic status
of local communities at two scales by collecting socio-economic data on both the counties and the dis-
tricts within which the refineries were located. A simple GIS model was used to align the emissions
data with the socio-economic data at different scales (counties and districts).
The results of the investigation into the links between corporate environmental performance and the
socio-economic status of local communities generated some significant results. While statistically
significant relationships (based on a series of bi-variate regressions) between levels of emissions and the
socio-economic indicators were not found at the county level, such relationships were found at the dis-
trict level. The results revealed some significant relationships between total site emissions and local
socio-economic conditions. For example, they indicated, with 99% confidence, that higher levels of emis-
sions of both sulphur dioxide and benzene were correlated with higher levels of unemployment in local
communities. For eco-efficiency, the results consistently suggested, with at least 90% confidence, that
lower levels of environmental performance were correlated with lower levels of per capita income, higher
levels of unemployment and lower levels of population density. Although the correlations reported were
all statistically significant, it is important to note that they did not explain all of the variation in emis-
sions. At most, 13% of the variability in measures of eco-efficiency could be explained with reference to
the socio-economic indicators. This is not surprising, as other factors such as the age of the refineries
or the content of crude oil inputs are likely to be more significant determinants of emissions, but there
are no data available for these.
Furthermore, while the results highlighted the presence of some significant correlations, it is impor-
tant to note that they do not infer anything about causality. Based on insights from previous research
(see Gouldson, 2004), we can speculate that these correlations might, potentially, be based upon one or
more of the following groups of factors: first, variations in the intensity of governmental, civil and self-
regulation; second, patterns of economic development and the dynamics of the labour market, and third,
patterns of social change and the dynamics of the housing market. More details on these groups of
potential influences are given in Table 2. Such causal linkages, and the socio-economic processes that
underpin them in different contexts, are undoubtedly complex and need to be explored much more fully
in future research.
Discussion
Results of the Study

In relation to corporate governance processes, it is clear that, at least amongst the leaders in the oil and
gas sector, corporate level activities on CSR are comparatively well developed. However, the processes
used to translate broad, corporate commitments into site specific standards are either not particularly
well developed or are not that transparent. Specifically, the principles established to guide site-specific
standard setting processes are not clear, and standards are commonly set through devolved decision-
making processes where there is considerable discretion and little transparency. Thereafter, data on site
DOI: 10.1002/bse
Group 1. Variations in the intensity of governmental, civic and self-regulation

• Oil refineries may be more likely to be built or expanded in poorer areas.
• Oil refineries may be likely to be regulated less intensively in poorer areas.
• Oil refineries may be likely to be put under less social pressure in poorer areas.
• Oil refineries may be less likely to adopt voluntary environmental management initiatives in poorer areas.
Group 2. Patterns of economic change and the dynamics of the labour market
• The areas around oil refineries may be likely to be become poorer as new employers are dissuaded from moving in.
• The areas around oil refineries may be likely to be become poorer as existing employers move out.
• The areas around oil refineries may be likely to be become poorer as the refineries employ fewer people from the local area.
Group 3. Patterns of social change and the dynamics of the housing market
• The areas around oil refineries may be likely to be become poorer as richer populations move out.
• The areas around oil refineries may be likely to be become poorer as poorer populations move in.
• The areas around oil refineries may be likely to become poorer as a consequence of relocation programmes designed to
create buffer zones around the facilities.
Table 2. Potential reasons for variations in environmental performance
level performance levels are rarely available, and where it is available it is not provided in a common,
consistent format. These factors make it virtually impossible to compare and contrast performance, both
from company to company and over time. This, in turn, makes it difficult to draw conclusions about
the links between CSR policy and CSR performance.
While the data that are available through PRTRs indicate that there are variations between facilities,
it is difficult to determine the factors that influence these variations. For example, significant variations
in emissions were found, both across the range of refineries and between the EU and the US. At the
local level, lower levels of environmental performance were correlated with lower levels of per capita
income, higher levels of unemployment and lower levels of population density in the districts where the
refineries were located. This finding lends some support to some of the claims associated with the
concept of environmental justice, namely that poorer communities experience lower levels of protection
and worse environmental conditions than richer communities, and seems to run counter to some of the
social expectations of the corporate responsibility debate. However, it is important to emphasize that
our analysis did not reveal anything about the causal factors that might lead to these correlations between
emissions and socio-economic conditions.
The lack of clarity around how companies translate their policy commitments into ‘on the ground’
performance – even in a sector such as oil and gas, which is generally recognized as a leader on CSR –
presents a significant issue for stakeholders such as local communities (i.e. those impacted by company
operations) and those who are interested in the environmental performance of such companies (e.g.
investors, regulatory agencies). The fact that performance varies so much seems to imply that corporate
policy commitments do not provide any guarantees regarding environmental performance, i.e. chal-
lenging the proposition that corporate responsibility policies can enhance environmental performance.
Does this mean that encouraging companies to strengthen the quality of their CSR policy commitments
and strengthening the implementation of these policies is a futile exercise? Unfortunately, firm answers
cannot be offered to this question, because of the manner in which CSR policies are conceived, with a
primary focus on procedures and the absence of specific performance measures. The only firm conclu-
sion that can be drawn is that the procedurally focussed approaches to CSR that characterize the oil and
gas sector – and, indeed, virtually all companies with commitments to CSR – do not seem to provide
assurance regarding the levels of environmental performance that will be achieved.
DOI: 10.1002/bse
The Role of PRTRs

Unfortunately, it is not possible to differentiate between good and poor performers based on compa-
nies’ own reports. In some countries, the absence of corporate reporting on site specific standards is
compensated for by the publication of PRTRs, which provide site-level emissions data in the same format
for similar processes, thereby enabling comparative measures of corporate performance to be estab-
lished. However, in the course of our study we identified a number of problems with quality and scope
of the data, which limited our ability to conduct a complete analysis of the environmental performance
of oil refineries. First of all, PRTRs are simply not in place in many countries and, hence, corporate per-
formance cannot be scrutinized in these countries.6 Second, the PRTRs that have been established in
different countries are often not entirely compatible; there are inconsistencies in the data, and some
potentially valuable data are commonly not reported (for example on short term variations in emissions,
the numbers of incidents, links with air quality and health). One of the reasons for this lack of consis-
tency – and a specific issue that we found in analysing differences between US and EU refineries – is
that national PRTRs are developed for national policy purposes (OECD, 1996) and to make decisions at
the national level. It is difficult to extrapolate this information to international comparisons, unless the
PRTR has been explicitly developed for international purposes from the beginning (OECD, 1997).7 The
reasons include differences in the definition of terms, the scope of the PRTR, the release estimation
techniques utilized and the substances covered by the PRTR.8 Third, the data provided in PRTRs (i.e.
details of the geographic location of facilities and bulk totals of the different pollutant species emitted
over a year) are not sufficient to allow meaningful conclusions to be drawn about the impacts that cor-
porations may have. In order to understand the contribution of releases to environmental concentra-
tions of various substances, information would be required on releases from specific sources (e.g. stacks),
specific release conditions (e.g. flue gas velocity) and temporal variations in emissions (e.g. variations
on a daily, weekly, monthly or seasonal basis). In addition, the substances reported in a PRTR may only
be a subset of the substances of interest, and the manner in which PRTR substances are defined may
preclude the identification of all species of concern. For example, some PRTRs just require reporting
on bulk totals of metals (e.g. ‘lead and compounds’) because the metal portion of the compounds is
toxic, while other PRTRs collect information on some or all of the specific species (Kennedy, 1998).
Emission information would then need to be related to air quality data and ultimately to epidemiologi-
cal and eco-toxicological data. Reports and evaluations of CSR and corporate environmental performance
very rarely go into such depth. Moreover, critically, all of this information would need to be presented
in a format that is accessible to stakeholder groups. There are significant issues here relating to risk
communication and the public understanding of science, and previous research has found that the
impacts of PRTRs often depend upon the presence of ‘information entrepreneurs’ such as NGOs who
access and interpret the data and present it in a simpler form (i.e. league tables of local polluters) for
local communities (see Gouldson, 2004).
6
In relation to the research that underpins this article, it would have been interesting to extend the analysis to allow comparison between, for
example, developed and less developed countries (and, hence, to explore more fully the concept of ‘environmental justice’). The absence of
PRTRs in less developed countries meant that this line of enquiry could not be pursued.
7
For international inventories (most notably, national greenhouse gas emission inventories and inventories established for transboundary pol-
lution assessment and management), significant effort is invested to ensure consistency in the quality of the data reported to the inventory.
Consistency is ensured through measures such as specifying standard release estimation techniques, developing standard emission factors
and having specific quality assurance mechanisms in place.
8
In this context, the work of the OECD in developing guidance on release estimation techniques as well as various regional programmes in
North America and Europe will help to ensure that the data in national PRTRs can be used in international applications. See, in particular,
the Resource Centre for Release Estimate Techniques (http://www.oecd.org/env/prtr), which provides a clearinghouse for guidance manuals
and other documents on release estimation techniques.
DOI: 10.1002/bse
Notwithstanding these limitations, PRTRs may have an important role to play in facilitating dialogue
between companies and stakeholders, and in addressing the gap between corporate reporting and stake-
holder demands. In PRTRs, the data from reporting facilities are reviewed and quality assured by inde-
pendent agencies or government departments. This ensures that all reported data meet, at least, certain
minimum specified quality standards, and can lead to the PRTR being seen as an independent source
of information on pollutant releases and transfers. In many OECD countries, the establishment of a
PRTR system has been the first time that a comprehensive and consistent database on pollutant releases
and transfers has been made publicly available, providing an agreed framework for dialogue between
interested parties.
Concluding Comments
There are three broad conclusions that can be drawn from the material presented in this article. The
first is that corporate CSR commitments fall significantly short of what is being expected by stakehold-
ers. In particular, the absence of specific performance metrics makes it difficult for stakeholders to assess
the quality of policy implementation. The second is that there are significant limitations in the data that
are available, which makes it difficult to assess performance against accepted standards or against other
facilities or companies. While PRTRs go some way to addressing this issue, the data in PRTRs are not
necessarily comparable between countries. In addition, PRTRs do not gather all of the information that
may be of interest to local communities (e.g. variations in emissions over time) or of those seeking to
conduct a meta-analysis such as that described here (where information such as facilities with
certification to ISO14001 would allow for a more nuanced analysis to be conducted). They may also fail
to present complex information on corporate environmental performance in a form that is readily
accessible to the lay public. Third, it is clear that current CSR policy commitments cannot be readily
related to the environmental outcomes that are achieved. That is, it is not possible to assess how the
process-focussed approach to CSR translates into specific outcomes such as site-level environmental
performance.
This article represents a significant challenge to current conceptions of CSR. While there is a general
consensus that companies will make commitments to social and environmental values, there is a clear
fault line between the process approaches that have been adopted by companies and the interests of
stakeholders such as local communities. From our analysis of the oil refining sector, it appears that
process-focussed approaches to CSR do not provide guarantees regarding the environmental outcomes
that will be achieved at the local level. Stakeholder criticisms of corporate environmental performance
will only be heightened by the lack of consistency in the manner in which companies approach these
issues, and the resulting variations in performance and the outcomes achieved. It is likely that this will
lead to pressure on companies to clearly explain how process and outcomes are related. That is, com-
panies will be expected to have clearer policy statements that explain what corporate policy commitments
mean in terms of on-the-ground performance, and to have stronger and more transparent processes for
the delivery of these policy commitments.
However, when we stray into the normative dimensions of these debates we immediately encounter
some complex questions. Should companies necessarily adopt the same standards wherever they
operate? If not, should they only comply with local laws? What should they do if these fail to provide
adequate protection for local communities? Should they then comply with certain minimum standards,
such as UN or WHO ambient quality standards? And what is it socially responsible for a company to
do when it contributes to but is not the sole cause of a breach of minimum standards? Whatever the
answers to these questions, the position of many is that companies should have explicit policies on such
DOI: 10.1002/bse
issues, that these should be accompanied by effective governance processes and that they should publish
clear, consistent and comprehensive reports on performance, including emissions and outcomes at the
site level. On these grounds, it seems that even the leading companies still have some way to go.
References
Agyeman J. 2002. Constructing environmental (in)justice: transatlantic tales. Environmental Politics 11(3): 31–53.
Bullard R, Johnson GS. 2000. Environmental justice: grassroots activism and its impact on public policy decision making.
Journal of Social Issues 56(3): 555–578.
Dobson A. 1998. Justice and the Environment. Oxford University Press: Oxford.
Environmental Defense Fund (EDF). 1999. Pollution Prevention Performance Ranking for Oil Refining Facilities. EDF web-site.
http://www.environmentaldefense.org/documents/1586_Facility%20Rankings.htm [11 February 2005].
Gouldson A. 2004. Risk, regulation and the right to know. Sustainable Development 12: 136–149.
Kennedy P. 1998. Comments on pollutant release and transfer registers from the viewpoint of environmental exposure asses-
sors. In OECD International Conference on Pollutant Release and Transfer Registers (PRTRs), Tokyo, 1998. OECD: Paris;
205–209.
Khanna M, Quimio W, Bojilova D. 1998. Toxic release information: a policy tool for environmental protection. Journal of Envi-
ronmental Economics and Management 36: 243–266.
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.
KPMG. 2005. International Survey of Corporate Responsibility Reporting. KPMG International. http://www.kpmg.com/Rut2000_
Prod/Documents/7/KPMG Intl CR Survey.pdf [24 January 2006].
OECD. 1996. Pollutant Release and Transfer Registers (PRTRs), a Tool for Environmental Policy and Sustainable Development:
Guidance Manual for Governments. OECD: Paris.
OECD. 1997. OECD Environmental Performance Reviews: a Practical Introduction, OECD/GD(97)35. OECD: Paris.
Petroleum Development Consultants (PDC). 2004. Study Contract for Investigation of EU Refineries’ Compliance for the Com-
bating of Air Pollution from Industrial Plants, report for the European Commission Environment Directorate. PDC: London.
Shrader-Frechette K. 2002. Environmental Justice: Creating Equality, Reclaiming Democracy. Oxford University Press: New York.
Stephan M. 2002. Environmental information disclosure programmes: they work, but why? Social Science Quarterly 83:
190–205.
Terry JC, Yandle B. 1997. EPA’s Toxic Release Inventory: stimulus and response. Management and Decision Economics 18:
433–441.
United States Environmental Protection Agency (US EPA). 2003. Moving Towards Collaborative Problem-Solving: Business and
Industry Perspectives on Environmental Justice. US EPA. Report Number EPA/300-R-03-003, available at http://www.
epa.gov/compliance/resources/publications/ej/annual_project-reports/study-industry-perspective-ej-permitting.pdf.
Zarsky L (ed.). 2002. Human Rights and the Environment: Conflicts and Norms in a Globalizing World. Earthscan: London.
DOI: 10.1002/bse

2006 - Gouldson - Corporate Environmentalism Tracing The Links Between Policies and Performance Using

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2006 - Gouldson - Corporate Environmentalism Tracing The Links Between Policies and Performance Using

Uploaded by

Copyright:

Available Formats

Business Strategy and the Environment

Bus. Strat. Env. 16, 1–11 (2007)

Corporate Environmentalism: Tracing

Received 1 November 2005; revised 21 January 2006; accepted 21 February 2006

Key Research Findings

Phase 1: Corporate Governance Processes

Phase 2: Evaluating Corporate Environmental Performance

Variables Units Data sources

Results of the Study

Group 1. Variations in the intensity of governmental, civic and self-regulation

Table 2. Potential reasons for variations in environmental performance

The Role of PRTRs

You might also like