Sustainability Performance Relation to Financial Performance PDF Free Download
1 / 69/69
100%
Sustainability Performance
Relation to Financial
Performance
A quantitative study of companies in the textile industry within the European and
North American markets
Bachelor’s degree project
Thesis within: Business Administration
Number of credits: 15
Program of Study: International Management
Authors: Cajsa Malmström, Lovis Ekström
Jönköping, 19th of May,2022
Acknowledgements
We would like to give appreciation to each individual that has been involved in the process of
writing this study. The participants in the seminar sessions as well as the tutors have provided
us with feedback continuously. The insights from our supervisor Emma Stendahl and co-
supervisor Jonas Dahlqvist have helped us to develop and accomplish this thesis. It would not
have been possible without their engagement and support for the development of this thesis.
Lastly, we would like to thank Anders Melander, Associate Professor of Business
Administration at Jönköping University Business School, for his guidance and instructions
throughout the whole writing process.
____________________________ ________________________________
Cajsa Malmström Lovis Ekström
Bachelor’s degree Project in Business Administration
Title: Sustainability Performance Relation to Financial Performance - A quantitative study of
companies in the textile industry within the European and North American markets
Authors: Ekström. L & Malmström. C
Tutor: Stendal. Emma
Date: 2022-05-19
Key terms: Sustainable Development; Textile Industry; Sustainability Performance; ESG
reporting; ESG scores; Financial Performance; Return on Invested Capital; Stakeholder Theory
Abstract
Background: Light has been shed on the textile industry as one of the leading industries when
it comes to economic growth, global warming and sustainable development. The increasing
demand for sustainable activities from stakeholders has led to the importance of measuring the
sustainable development of companies. ESG reporting is a common tool used to indicate a
company’s sustainability performance. Prior research has tended to focus on cross-sectional
industries, and therefore a gap was identified for industry specific research.
Purpose: The purpose of this research is to explain the relationship between sustainability
performance and financial performance in the textile industry in the European and North
American markets to see if companies that invest in sustainability activities benefit financially.
Method: This research has followed a positivistic paradigm, with deductive reasoning and a
quantitative approach. A probability sampling approach was performed by conducting
secondary data from Thomson Reuters DataStream of companies in the textile industry in
Europe and North America. This resulted in a final sample of ESG scores and ROIC of 44
companies. The data was later analysed in the SPSS software program by following the
estimation method Ordinary Least Squares (OLS).
Findings: The literature review developed two hypotheses to address the research purpose and
questions. The two hypotheses were analysed through two regression analyses that were
satisfied through the OLS estimation method. The result showed that there was a significant
relationship between the aggregated ESG score and ROIC which supported the first hypothesis.
The second hypothesis of the multiple regression model showed that each component of ESG
is correlated to ROIC, however, the environmental factor was not statistically significantly
related.
Conclusion: The thesis showed that there is a positive relationship between ESG performance
and ROIC in this study. This implies that companies that invest in sustainable development
increase their financial performance. The aggregated ESG score as well as the social factor and
the governance factor had the highest impact on ROIC, which is supported by the stakeholder
theory as there has been an increasing demand on social and governance activities in the textile
industry. This further supports that sustainability performance impact on financial performance
is industry specific.
Table of Contents
LIST OF ABBREVIATIONS .........................................................................................................................
1. INTRODUCTION .................................................................................................................................... 1
1.1 BACKGROUND .............................................................................................................................................. 1
1.2 PROBLEM DISCUSSION ................................................................................................................................ 3
1.3 PURPOSE AND RESEARCH QUESTIONS ........................................................................................................ 5
1.4 DELIMITATIONS ........................................................................................................................................... 5
2. LITERATURE REVIEW ......................................................................................................................... 7
2.1 METHOD FOR LITERATURE REVIEW ........................................................................................................... 7
2.2 SUSTAINABLE DEVELOPMENT IN THE TEXTILE INDUSTRY ....................................................................... 7
2.3 MEASURING SUSTAINABLE DEVELOPMENT THROUGH SUSTAINABILITY PERFORMANCE ....................... 9
2.3.1 ESG Reporting ................................................................................................................................... 10
2.4 THE LINKAGE BETWEEN SUSTAINABILITY PERFORMANCE AND FINANCIAL PERFORMANCE ................ 11
3. THEORETICAL FRAMEWORK ......................................................................................................... 13
3.1 THE STAKEHOLDER PERSPECTIVE ........................................................................................................... 13
3.2 STUDY HYPOTHESES .................................................................................................................................. 14
4. METHODOLOGY AND METHOD ...................................................................................................... 16
4.1 RESEARCH PARADIGM ............................................................................................................................... 16
4.2 RESEARCH APPROACH ............................................................................................................................... 16
4.3 RESEARCH STRATEGY ............................................................................................................................... 17
4.4 DATA COLLECTION .................................................................................................................................... 17
4.5 SAMPLING APPROACH ............................................................................................................................... 18
4.6 DATA ANALYSIS ......................................................................................................................................... 19
4.7 SECONDARY DATA ANALYSIS ................................................................................................................... 19
4.8 MEASUREMENTS AND VARIABLES ............................................................................................................ 20
4.8.1 Independent variables: ESG Scores .................................................................................................. 21
4.8.2 Dependent variable: ROIC ................................................................................................................ 22
4.9 REGRESSION ANALYSIS ............................................................................................................................. 22
4.10 OLS REGRESSION MODEL ...................................................................................................................... 24
4.11 RESEARCH QUALITY ............................................................................................................................... 27
4.11.1 Validity and reliability ..................................................................................................................... 27
4.11.2 Ethical consideration ....................................................................................................................... 28
5. EMPIRICAL RESULTS & ANALYSIS ................................................................................................ 30
5.1 DESCRIPTIVE STATISTICS ......................................................................................................................... 30
5.2 CORRELATION BETWEEN THE VARIABLES ............................................................................................... 31
5.3 OLS MODEL ASSUMPTIONS ...................................................................................................................... 32
5.3.1 Linearity ............................................................................................................................................. 32
5.3.2 Random Sampling of Observations ................................................................................................... 34
5.3.3. A Conditional Mean of Zero ............................................................................................................. 35
5.3.4 Absence of Multicollinearity .............................................................................................................. 35
5.3.5 Spherical Errors: Homoscedasticity and Non-Autocorrelation ....................................................... 36
5.3.6 Normally distributed Error Terms ..................................................................................................... 37
5.4 RESULTS FROM THE REGRESSION ANALYSES .......................................................................................... 38
5.4.1 ESG Performance Relation to ROIC ................................................................................................ 38
5.4.2 ENV, GOV, and SOC Performance Relation to ROIC ..................................................................... 39
6. DISCUSSION & CONCLUSION ........................................................................................................... 43
6.1 THEORETICAL CONTRIBUTIONS ............................................................................................................... 43
6.2 MANAGERIAL IMPLICATIONS ................................................................................................................... 46
6.3 LIMITATIONS ............................................................................................................................................. 47
6.4 FUTURE RESEARCH .................................................................................................................................... 48
6.5 CONCLUSIONS ............................................................................................................................................ 49
7. REFERENCES ....................................................................................................................................... 51
Table of Figures
Figure 1: Research model for H1 in this study. Source: Created by the authors for the study’s purpose. ............ 15
Figure 2: Research model for hypotheses H2abc in this study. Source: Created by the authors for the study's
purpose. .................................................................................................................................................................. 15
Figure 3: How ROIC is calculated ......................................................................................................................... 22
Figure 5.4: Normal Q-Q Plot of ENV Score 2020: expected values against the observed values of the
independent variable ENV. .................................................................................................................................... 33
Figure 5.5: Normal Q-Q Plot of GOV Score 2020: expected values against the observed values of the
independent variable GOV. ................................................................................................................................... 33
Figure 5.6:Normal Q-Q Plot of SOC Score 2020: expected values against the observed values of the
independent variable SOC. .................................................................................................................................... 34
Figure 5.7: Normal Q-Q Plot of ESG Score 2020: expected values against the observed values of the
independent variable ESG. .................................................................................................................................... 34
Figure 5.8: Normal Q-Q Plot of ROIC 2021: expected values against the observed values of the dependent
variable ROIC. ....................................................................................................................................................... 34
Figure 5.9: Histogram illustrating the distribution of the residuals of the dependent variable ROIC 2021. ......... 35
Figure 5.10: Scatterplot of the dependent variable’s Residuals vs Predicted values of the regression. ................ 36
Figure 5.11: Scatterplot between ESG Score 2020 and ROIC 2021 including R squared. ................................... 39
Figure 5.12: Scatterplot illustrating the relation between ENV Score 2020 and ROIC 2021 ............................... 41
Figure 5.13: Scatterplot illustrating the relation between GOV Score 2020 and ROIC 2021. .............................. 42
Figure 5.14: Scatterplot illustrating the relation between SOC Score 2020 and ROIC 2021. .............................. 42
Figure 15: Research model for H1 in this study. Source: Created by the authors for the study’s purpose. .......... 43
Figure 16: Research model for hypotheses H2abc in this study. Source: Created by the authors for the study's
purpose. .................................................................................................................................................................. 43
Table of Tables
Table 5.1: Descriptive Statistics of the variables used in the regression analyses. ............................................... 31
Table 5.2: Pearson Correlation of ENV, GOV, SOC and ROIC ........................................................................... 32
Table 5.3: VIF statistics test to discover collinearity of the multiple regression model. ...................................... 36
Table 5.4: Durbin Watson test for autocorrelation of the dependent variable ROIC and the independent variable
ESG ........................................................................................................................................................................ 37
Table 5.5: Durbin Watson test for autocorrelation of the dependent variable ROIC and the independent variable
ENV, SOC, and GOV. ........................................................................................................................................... 37
Table 5.6: Test of Normality, Shapiro-Wilk tested at 95% confidence intervals .................................................. 38
Table 5.7: Pearson Correlation of ESG and ROIC ................................................................................................ 39
List of abbreviations
ESG - Environmental, Social, and Governance Performance
ENV - Environmental Performance
SOC - Social Performance
GOV - Governance Performance
ROIC - Return on invested Capital
OLS - Ordinary Least Squares
1
1. Introduction
__________________________________________________________________________________________
This chapter presents a background to the subject for the research, followed by a problem
discussion which contributes to the given research purpose and the research questions of this
study. Lastly, the research delimitations are presented.
___________________________________________________________________________
1.1 Background
The topic of global warming has become an evident threat to our existence, where the
requirement of mitigating climate change and global warming are two of the biggest concerns
of today (Luján et al., 2020; Rajah et al., 2018; Eriksen et al., 2011). It has rapidly expanded
during the last decade and this has made governments and business managers progressively
more concerned for the planet's future. The increased concern has contributed to increased
demand on reducing global warming on all parts; governments, companies, and individuals
(Kolk & Van Tulder, 2010).
The increasingly unstable environment has for decades recognised companies’ as the primary
contributors to global warming (Wood et al., 2021; Diener & Habisch, 2020; Rajah, 2018).
Yet, the growing trend of incorporating sustainable development suggests that companies are
also vital players as part of the solution to decrease climate change (Kolk & Van Tulder, 2010).
Hence, to acquire sustainable progress, companies can make a huge positive impact by
addressing sustainable development strategies and practices (Wood et al., 2021). Furthermore,
research has shed light on the textile industry regarding unsustainable practices and more
particularly as one of the major contributors to fostering water pollution across the world
(Paraschiv et al., 2015). Therefore, the subject of sustainability has become a prominent
concept to address these issues for both governments and companies (Wood et al., 2021;
Eriksen et al., 2011). It is a widespread concept that depends on context and situation, but a
broadly accepted definition stated by Brundtland Report (1987, p. 37), is “Sustainable
development is a development that meets the needs of the present generation without
compromising the ability of future generations to meet their own needs”. However, to give a
more suitable definition of sustainability in the business context, Van Marrewijk (2003, p. 8),
explains it as “demonstrating the inclusion of social and environmental concerns in business
2
operations and interactions with stakeholders”. This is the definition that will be used in this
study from now on.
Because of the textile industry’s major influence on global warming and economy, companies
in this industry play a significant role in impacting the environmental, social, and economic
activities globally (Luján et al., 2020; Moorhouse & Moorhouse, 2017). Subsequently, this has
led to major improvements regarding fostering sustainable development and performance in
the industry. For example, using less water; adopting environmentally friendly manufacturing
techniques; utilizing less energy in manufacturing processes, and implementing the 3 Rs -
Reduce, Reuse, and Recycle (Muthu, 2017). Therefore, companies in the textile industry are of
particular interest as the requirement of adopting sustainability strategies and practices are
pressured and under progress in this industry.
It is broadly acknowledged that in order to determine a company’s sustainability performance,
it needs to be measurable (Özdemir et al., 2011; Epstein & Rejc Buhovac, 2014). The
requirements from stakeholders on reporting on economic performance has increased, as well
as the demand on environmental and social practices, which have motivated the development
of different corporate sustainability tools (SRTS) (Özdemir et al., 2011). The three-dimensional
metric ESG which stands for environmental, social, and governance reporting has developed
as one of the most common ways of obtaining and rating how a company performs
environmentally, socially, and in governance activities (Zubeltzu‐Jaka et al., 2018). This metric
is of particular interest to understand the relationship between sustainability performance and
financial performance (Özdemir et al., 2011). Along with the growing social pressure on
sustainable development in the textile industry, the adoption of ESG activities has subsequently
increased (Diener & Habisch, 2020). This is supported by the Survey of Sustainability
Reporting (KPMG, 2020), which identified that the biggest companies worldwide that report
on sustainability have increased from 51% in 2013 to 80% in 2020. Hence, this wider market
trend enhances that the business environment is changing rapidly due to the importance of
sustainability (KPMG, 2020). Additionally, new regulations are developing in the European
Union (EU). These will require legislation on ESG reporting on all large and all publicly traded
companies including small and medium-sized enterprises (SMEs) that operate within the EU
(European Parliament, 2021). The emerging legislation within European boundaries increases
the requirements of reporting on ESG activities in Europe, which contributes to that this market
is interesting to investigate as it is a leader when it comes to transparency and reporting of ESG.
3
Stakeholders' increasing demand on sustainable development has contributed to an increased
ESG reporting in the North American market as well. However, for now there are no mandatory
regulations stating that North American companies need to report on ESG but there is a
requirement that all public companies need to disclose information that could be connected to
the ESG risk (Silk & Lu, 2022). Even though the North American market is not as developed
when it comes to ESG performance and reporting as the European market, it is highly
developing (SiaPartners, 2022). Therefore, these two markets have established several ESG
activities and are compared to other markets more developed, which contributes to that these
two markets are of particular interest (Silk & Lu, 2022). The requirements of ESG reporting
is still in progress in both markets contributing to that not all companies’ in the textile industry
have established ESG reporting yet. This can therefore result in inadequate results (Siew,
2015).
1.2 Problem Discussion
Several research studies have been conducted exploring how the textile industry is developing
and have developed towards a sustainable industry, where literature argues that even though it
is one of the most polluting industries, it is also a major contributor to economic growth (Luján
et al., 2020; Moorhouse & Moorhouse, 2017; Wrap, 2020b; Patwary, 2020). Moreover,
research shows that major actions towards sustainable development have been made in the
industry, for instance environmentally friendly manufacturing techniques, circular materials,
and reduction of water use (Luján et al., 2020; Moorhouse & Moorhouse, 2017). The subject
of whether the textile industry is transitioning towards a sustainable industry is, however,
debated. Some scholars argue that many companies still follow a traditional linear economic
model of “take, make, dispose” (Luján et al., 2020). Other scholars mean that the industry is
transitioning towards a sustainable business model by implementing sustainability activities
due to the high pressure from society (Moorhouse & Moorhouse, 2017; Wrap, 2020b; Patwary,
2020).
In previous research that has examined the relationship between ESG performance and
financial performance of companies, the results have been conflicting. Several earlier studies
have investigated the relationship on a broader aspect, where they have conducted the research
based on cross-sectional industry investigations based on companies in cross-sectional markets
(Kaiser, 2020; Velte, 2017; Xie et al., 2019). These studies have examined whether there is a
relationship between companies ESG performance and their financial performance. However,
4
investigating cross-sectional industries can demonstrate inconclusive results since
sustainability is a complex concept and differentiates between industries (Wood et al., 2021;
Eriksen et al., 2011). This suggests a gap for investigating a specific industry and market with
a narrower estimation method to reduce the risk of insignificant results. Research shows that
the textile industry has particularly developed and increased their sustainability activities
(Luján et al., 2020; Moorhouse & Moorhouse, 2017; Wrap, 2020b; Patwary, 2020). The
industry has been heavily criticized, which has led to the textile industry has improved its
processes and strategies to maintain its relationship with stakeholders (Phan et al., 2020). This
indicates that research is needed in the textile industry in both Europe and North America due
to an increasing demand of ESG reporting as well as upcoming legalization in these markets.
Further, prior research has studied both accounting-based and market-based financial variables
to see which financial variable has the strongest relationship with ESG (Kaiser, 2020; Velte,
2017; Xie et al., 2019; Auer & Schuhmacher, 2016; Fiskerstrand et al., 2020; Duque-Grisales
& Aguilera-Caracuel, 2021; Landi & Sciarelli, 2019). Although there is literature based on
investigating accounting-based financial performance, studies tend to investigate variables
based on Return on Asset (ROA) and Return on Equity (ROE). This can be an issue as these
variables are influenced by the majority of a company's operations and may not show the reality
of the company’s financial performance (Hagel et al., 2013). This can contribute to more
general results that are not certainly connected to sustainability performance (Hagel et al.,
2013).
The majority of studies investigating sustainability performance and financial performance
have shown that there is a positive relationship between them, especially when considering
accounting-based dependent variables such as ROA and ROE (Kaiser, 2020; Velte, 2017; Xie
et al., 2019). These studies prove that companies' ESG performance is positively related to their
ROA or ROE. In these cases, this means that sustainable development practices lead to
increased financial performance. However, other studies have encountered a neutral (Auer &
Schuhmacher, 2016; Fiskerstrand et al., 2020), or even negative relationship (Duque-Grisales
& Aguilera-Caracuel, 2021; Landi & Sciarelli, 2019). The neutral results mean that they have
not found any benefits of having ESG performance when it comes to financial performance,
whereas the negative result means that the companies’ ESG performance has led to lower
financial performance. However, neutral and negative results have been found in particular
when measuring market-based dependent variables such as Tobin’s Q with ESG factors
(Kaiser, 2020; Velte, 2017; Xie et al., 2019; Auer & Schuhmacher, 2016; Fiskerstrand et al.,
5
2020; Duque-Grisales & Aguilera-Caracuel, 2021; Landi & Sciarelli, 2019). This shows that
financial market-based measurements have low, neutral or negative relation to ESG
performance. To fill in the research gap, there is a need for investigating the relationship
between ESG performance and financial performance based on a financial performance
estimator that takes investment into account. The variable Return on Invested Capital (ROIC)
is sufficient, as this is a multiplier that measures a company’s efficiency in returning profits
from invested capital (Cachon & Terwiesch, 2008). This is suitable since ESG performance of
a company is based on how much they invest in sustainability activities.
To see if sustainability performance indicates better financial performance, it is interesting to
investigate the relationship between ESG performance and ROIC of companies in the textile
industry in Europe and in North America. This will contribute to a more specific research of
one industry and two markets with similar industrial activities. To measure their sustainability
performance, ESG scores will be used whereas to measure their financial performance, the
accounting-based measurement ROIC will be used. To gain further knowledge in what affects
financial performance and to fill in the gap, each factor of ESG as well as the total ESG scores
relation to ROIC will be estimated.
1.3 Purpose and Research Questions
The aim of this study is to explain the relationship between sustainability performance and
financial performance in the textile industry in Europe and in North America to investigate if
companies’ investing in sustainability activities benefit financially.
To address this research’s purpose, we have formulated two research questions:
● What is the relationship between ESG performance and financial performance of
companies in the textile industry in Europe and in North America?
● What is the relationship between each of the ESG factors and financial performance in
the companies?
1.4 Delimitations
The complexity of sustainability with several different ways of reporting sustainable
performance throughout the world makes limiting the research to a specific industry and market
consequently important. When first retrieving data on textile companies in Europe, one could
see that the sample that could be found was inadequate, which contributed to the choice of
6
including the North American market as well. The similar progress in the two markets in terms
of ESG reporting, as well as a common ground and goals for sustainable strategies and practices
within the textile industry, enhance a better result. Hence, the decision to delimitate the study
to the textile industry, the European market, and the North American market was made to
decrease the risk of insignificant results. Moreover, to explore the relationship between ESG
performance and financial performance more, limiting the research to evaluate ROIC as an
accounting-based financial performance indicator was appropriate. This to obtain investment
results more directly linked to sustainability performance, which also reduces the risk of
insufficient results.
Lastly, the datasets of ESG scores and ROIC will be limited to a specific year, where the ROIC
will be time-lagged one year after the ESG scores. Including a time lag of one year to see the
relationship between ESG scores and financial performance is necessary since ESG scores will
not have a direct impact on the financial performance (Scholtens, 2008). It is also important in
order to see how ESG performance affects ROIC, and not the other way around to investigate
the correlation between the variables rather than a causal relationship. Therefore, the financial
performance will be measured as year t+1 with ESG scores from year t. This is made for several
reasons; to limit the research; accessibility; and relevance.
7
2. Literature Review
__________________________________________________________________________________________
The purpose of this chapter is to provide literature and theories that are relevant to address
the research purpose and questions. The topics covered are “Sustainable Development in the
Textile Industry; Measuring Sustainable Development through Sustainability Performance;
ESG Reporting and lastly; The Linkage between Sustainability Performance and Financial
Performance”. Moreover, the studied literature serves as a foundation for the study hypotheses
which are presented in the next chapter.
_________________________________________________________________________
2.1 Method for Literature review
To present relevant information from previous literature in the research area, the authors have
performed a literature search using a systematic approach. The first step included finding
journals and databases whereby Google Scholar, Scopus, and Business Source Premier have
been used. Within the area of study, the authors have used and combined the keywords;
Sustainable Development; Textile Industry; Sustainability Performance; ESG reporting; ESG
scores; Financial Performance; Return on Invested Capital; Stakeholder Theory. To secure a
relevant, trustworthy, and high-quality literature review, peer-reviewed literature from the
“Academic Journal Guide 2021” was collected. The year of publication of the literature has
also been considered to perform a study that is based on relative and durable research. This, as
the relevant topics covered in the literature review, have emerged and developed during the last
decades and are therefore time sensitive. Moreover, to identify the research gap and purpose of
this investigation, the literature review provided a foundation for this. Lastly, it was used as the
foundation when developing a method to address the purpose of the study.
2.2 Sustainable Development in The Textile Industry
To become more sustainable, companies need to incorporate sustainable development through
corporate governance along with strategies for utilizing social and environmental problems in
their operations (Ferrell et al., 2016). The concept of sustainable development has emerged
globally as a common goal to satisfy the current generations´ needs without distorting future
generation’s ability to meet their needs (Brundtland Report, 1987). The two terms
“Sustainability” and “Sustainable Development'' are commonly used as synonyms. However,
research asserts that the concept of sustainability more commonly prioritizes the environmental
perspective (Dresner, 2008; Robinson, 2004; Carley & Christie, 2000; Reid 2005; Dalal-
8
Clayton & Bass, 2002; Rogers et al., 2008; Rockström et al., 2009; World Footprint, n.d;
UNDP, 2019). Therefore, since this study aims at investigating a company perspective, which
leads to the concept of sustainable development is more suitable for this literature review.
When referring to sustainable development, Van Marrewijk’s (2003, p.8), the definition will
be used; “Demonstrating the inclusion of social and environmental concerns in business
operations and interactions with stakeholders”.
Responding to the growing climate crisis as well as to enhance an equal commitment to global
development, sustainable development goals were developed as the leading model by the end
of the 20th century by the organization United Nations (Carley & Christie, 2000; Reid 2005;
Dalal-Clayton & Bass, 2002; Rogers et al., 2008; Rockström et al., 2009; World Footprint, n.d;
UNDP, 2019). This to encourage global sustainability trends that contribute to a sustainable
relationship between the environment and the people (Kates & Parris, 2003). The textile
industry plays an important role as it is one of the main contributors to the global economy
(Luján et al., 2020; Moorhouse & Moorhouse, 2017; Patwary, 2020). However, it is also
considered the second biggest industry considering the destruction of the environment where
the industry commonly uses the production and consumption model of “take, make, dispose”
(Luján et al., 2020; Moorhouse & Moorhouse, 2017; Wrap, 2020b; Patwary, 2020). During the
last decades, policies and strategies have emerged to enhance sustainable development to
prevent a climate catastrophic disaster, which has put pressure on the industry as it plays a
significant role in economic, social, and environmental impacts globally (Luján et al., 2020).
To efficiently meet the sustainable requirements which focus on the three major elements;
social, environmental, and economic aspects, the term sustainability performance has emerged
to determine a company's sustainable development (Bi, 2011). The textile industry has a
complex global supply chain where many companies in the sector still follow a traditional
linear economic model (Luján et al., 2020). Seven key life cycle stages for textile production
have been identified; fiber production; textile production; clothing production;
commercialization; use; and end-of-life. During the last couple of years, several initiatives and
innovations have emerged in the different life cycles with the intention to transition the industry
towards sustainable production and consumption to enhance sustainability performance.
Literature shows that both public and private companies have increased their sustainable
development practices by adopting different strategies for the different life cycles of textiles to
achieve sustainability (Luján et al., 2020; Moorhouse & Moorhouse, 2017).
9
To further increase and promote sustainable development, The United Nations (UN, 2019)
established 12 common goals for all industries to achieve by 2030, as well as The European
Clothing Action plan was launched in May 2019 to motivate and push the textile industry and
scientists towards sustainability of textiles (Wrap, 2020a). The European Clothing action plan
has initiated actions that simultaneously focus on several life cycle parts to increase the
sustainability performance of the textile industry in Europe (Luján et al., 2020). It emphasizes
the importance of rethinking and reinventing the way products are designed and produced, and
the importance of reusing and recycling from the consumer perspective (Wrap, 2020a). For
instance, by using efficient garments recovery, the waste of textiles can be heavily reduced. In
the action plan, different stakeholders are involved such as suppliers, the public sector,
recycling organizations, and different retailer brands (Luján et al., 2020).
2.3 Measuring Sustainable Development through Sustainability Performance
The increased pressure on sustainable development practices in companies has contributed to
the high demand for measuring how well these practices perform (Epstein & Rejc Buhovac,
2014). By measuring sustainability performance, companies can see whether the implemented
changes and practices concerning sustainability have led to sustainable development (Epstein
& Rejc Buhovac, 2014). Moreover, this enhances the measurement of sustainable development
and transparency which can help companies with different approaches on how to overcome
barriers and trade-offs (Beckman et al., 2014). The concern for sustainability performance
measurement has increased over the last two decades (Devuyst 2000; Maas et al., 2016; Sala
et al., 2015), where the demand is based on creating transparency and fostering innovations for
companies (Burritt & Schaltegger, 2014). There are different ways to measure the performance,
where Maas et al (2016) describe it as, collecting, analyzing, and communicating sustainability
effects in business, society, and environmental decisions which is the process of sustainability
performance measurement. Schöggl et al., (2016), argues that the heart of sustainability
performance measurement is the indicator for capturing and consolidating performance
information. In addition to this, the different methods that have developed over time use their
own established indicators to measure sustainability performance (Özdemir et al., 2011). One
way of measuring this can be through different methods that contribute to economic, social,
and environmental activities in companies’ management which indicate whether they have
included sustainability in their business both in the short and long run (Searcy, 2012). Because
of the multiple corporate sustainability tools (SRTS), and a lack of standardization of how to
10
measure companies’ sustainable development, the process of this is more complicated (Siew,
2015).
2.3.1 ESG Reporting
A commonly accepted quantitative metric to measure sustainability performance in a company
is to use ESG scores (Ahi et al., 2018; Van der Vaart et al., 2018). Initially, ESG scores were
intended to be used as a legitimacy-seeking method, where companies could gain a better
understanding of how to deal with societal expectations (Schaltegger et al., 2017). ESG scores
measure three different elements of a company’s sustainability performance, that is
environmental, social, and governance (Ashwin Kumar et al., 2016). The environmental score
is based on how well a company manages its resources. For example, their energy use, waste,
pollution, treatment of animals, and natural resource conservation. The social factor explains
corporate social responsibility (CSR) which focuses on three areas: operations, supply chain
social impact, and community investment in the company. The last factor, governance,
measures how a company runs its business which includes structures, policies, processes, and
practices. In this part, companies include their goals, and how to handle their relationships
among the different stakeholders (Hedstrom, 2018).
During the past decade, stakeholders have increased their demand and pressure on ESG
reporting which has led to the sustainability rating agencies' (SRAs) role have become more
critical in the process of rating (Rajesh, & Rajendran, 2020). Because of the increased demand
from stakeholders, and to be able to give increased benefits to the investors, companies have
introduced more and more ESG activities in their management (Hübel & Scholz, 2020).
Moreover, the upcoming legislation on ESG reporting in Europe will increase the number of
companies that have to report on ESG which will enable a more transparent view of companies’
sustainability performance (Luján et al., 2020; Moorhouse & Moorhouse, 2017). Additionally,
the increased demand for ESG reporting has been established through the 12 common
sustainability goals for all industries (UN, 2019) to enhance the trend towards sustainable
development. This together with the increased demand from stakeholders in North America
implies that this market will follow the trend of sustainability reporting in Europe to sustain in
the future.
It is important to acknowledge that the validity and ethical question of how to trust the ESG
scores and ratings have lately been questioned since there are no common measurement
11
methods used among the different SRAs (Ester & Perkins, 2021). The different SRAs collect
data from both third parties and directly from companies’ reports (Drempetic et al., 2020), to
provide stakeholders with ratings on the three different ESG factors (Ester & Perkins, 2021).
In addition, the SRAs provide companies with an overall ESG rating based on their
performance. The ratings are established through their own developed methods which can
differ between SRAs and this is also the reason for it being questioned (Drempetic et al., 2020).
In this research, the Thomson Reuters DataStream database will be used to gather ESG data. It
is a trusted database that is commonly used and undertaken by many researchers in recent years
(Gallego-Alvarez, & Quina-Custodi, 2017; Garcia et al., 2017).
2.4 The linkage between Sustainability Performance and Financial Performance
Sustainability performance measured through ESG scores in relation to financial performance
has been widely debated and has been examined in various studies (Fischer & Sawczyn, 2013;
Hussain et al., 2018; Velte, 2017; Wang & Sarkis, 2017; Fatemi et al., 2018; Xie et al., 2019;
Duque-Grisles & Aguilera-Caracuel, 2019). In this context, financial performance is defined
as the financial visibility of a company and as a measure of how they achieve its economic
goals (Orlitzky et al., 2003). By measuring financial performance, one can predict the financial
health of a company to get a better understanding of its development. In addition, it can be used
to compare companies in and between industries, as well as markets (Manogna & Mishra,
2021). Several different measurements indicate the financial performance, where one of them
is the ratio Return on Invested Capital (ROIC). This is an economic multiplier that measures
how efficiently a company invests its capital to be able to gain profitable investments (Cachon
& Terwiesch, 2008). This is an accounting-based measurement that can be used to analyze the
return on sustainable investments in a company (Cachon & Terwiesch, 2008). ROIC is the
financial ratio measure that will be used in this study as the measure of financial performance.
There is a lack of coherency in the literature of the relation between ESG performance and
financial performance where both positive, neutral, and negative results have been
encountered. According to Walsh et al., (2003) and Friede et al., (2015), the majority of
research that has been conducted presents a positive relationship between ESG scores and
financial performance, mostly when using the accounting-based measurement for the financial
performance (Fischer & Sawczyn, 2013; Hussain et al., 2018; Velte, 2017; Wang & Sarkis,
2017). Furthermore, some studies have studied the relationship between the individual factors
of ESG and their relation to financial performance (Clark et al., 2015), whereas a study by
12
Velte (2017) shows that there is a positive relationship between the individual ESG factors and
financial performance. However, even though the subject is researched, some authors argue
that there is no clear evidence about the relationship (Eccles & Viviers, 2011). Irregular
findings in the relationship can be in terms of how the ESG performance is used and what type
of financial performance measurement the study is looking at (Ashwin Kumar et al., 2016;
Hussain et al., 2018; Wang et al., 2016).
13
3. Theoretical Framework
_____________________________________________________________________________________
This chapter will outline the theoretical framework that will be used in this study to support
the research purpose and questions. The theoretical framework presents the stakeholder
perspective and the study hypotheses.
______________________________________________________________________
The adoption of the stakeholder perspective will enable the understanding of how companies
that are concerned about their stakeholders, are more commonly active in ESG activities
(Freeman et al., 2010; Hörisch et al., 2014). Moreover, study hypotheses were conducted to
address the research purpose and questions with help of using the stakeholder perspective to
support the stakeholder’s impact on the sustainable development and performance.
3.1 The Stakeholder Perspective
The term stakeholder has many different definitions, but the most common and general
definition stated by Freeman (1994, p.25; Freeman et al., 2010, p.9) is “those groups and
individuals who can affect or be affected” by a company’s activities linked to how trade and
value are created. In this research, the stakeholder theory will help to explain the relationship
between a company and its stakeholders, which have emerged to understand and mitigate
interconnected business issues (Freeman et al., 2010). There are explicit reasons why the
stakeholder perspective is an emphasized theory in this study, since it supports that the more
concerned a company is about the relationship with its stakeholders, the easier it will be for the
company to succeed (Freeman et al., 2010; Freeman, 1994). Therefore, adopting a stakeholder
lens when investigating the relationship between ESG performance and financial performance
supports the hypothesis that companies that are concerned about their stakeholders have higher
sustainable development.
It is important to acknowledge that many different versions of the stakeholder theory have
emerged which have contributed to simultaneous developments and improvements of the
theory (Hörisch et al., 2014). However, the original theory established by Freeman and co-
authors referred to as the “normative stakeholder theory” remains the most relevant,
appropriate, and comprehensible theory for this study since it includes multiple linkages to
sustainability management (Hörisch et al., 2014). Focusing on the perspective of “managing
stakeholder relationships” in the theory relates to the development of creating mutual interests
between various stakeholders instead of putting an emphasis on trade-offs. This is one of the
14
core components to establish value for all stakeholders connected (Freeman et al., 2010;
Hörisch et al., 2014). Moreover, the theory encourages management in which companies can
maximize value to their stakeholders through their activities. Edmans (2011); Deng Kang, et
al., (2013), argues that a company that has efficient management is more likely to be engaged
in ESG activities, which favours the stakeholder perspective. To enhance sustainable
development, the stakeholder theory is efficient to interpret as it argues that the more concerned
businesses are with their stakeholders, the more successful will they be (Freeman et al., 2010;
Freeman, 1994). The theory is highly linked with sustainable management and utilizes
sustainable development practices by considering stakeholders' impact and value (Freeman et
al., 2010; Hörisch et al., 2014). Moreover, earlier research has adopted the stakeholder
perspective into the relation between ESG performance and financial performance (Duque-
Grisales & Aguilera-Caracuel, 2021). This is logical as the theory implies that there is a positive
relationship between a company with a stakeholder perspective and implemented ESG
activities (Edmans, 2011; Deng Kang et al., 2013). Since stakeholders benefit from better
financial performance (Freeman et al., 2010), the relation between ESG activities and financial
performance is supported by the stakeholder theory and is suitable for this study.
3.2 Study hypotheses
To address the research questions, two study hypotheses have been developed to test if
companies in the textile industry in Europe and in North America’s ESG score impacts their
ROIC. In addition, taking the stakeholder perspective is relevant to see the relationship between
ESG scores and financial performance. This is because companies that are concerned about
their stakeholders, generally focus more on ESG activities (Ferell et al., 2016). Moreover,
literature shows evidence that each ESG factor has a diverse relationship and impact on
financial performance (Friede et al., 2015). Therefore, each of the elements' relationship to
financial performance is equally important to examine, which is the Environmental (ENV),
Social (SOC), and Governance (GOV) relationship to a corporation’s financial performance,
ROIC.
To investigate the first research question on how the overall ESG performance of a company
is related to ROIC, hypothesis H1 was conducted:
● H1: ESG performance has a positive relation to ROIC
15
Secondly, to investigate how each factor of the ESG performance relate to ROIC, the
hypotheses H2abc was conducted:
● H2a - ENV performance has a positive relation to ROIC
● H2b - GOV performance has a positive relation to ROIC
● H2c - SOC performance has a positive relation to ROIC
Figure 1: Research model for H1 in this study. Source: Created by the authors for the study’s purpose.
Figure 2: Research model for hypotheses H2abc in this study. Source: Created by the authors for the
study's purpose.
16
4. Methodology and Method
_____________________________________________________________________________________
This section contains parts of research philosophy, research approach, and research strategy,
as well as the selected technique and research methodology. A description of the data
collection methodology, sampling approach, and data analysis are also included to meet the
study's objectives. Secondly, it will present the variables and estimation method that was used
in the study. Finally, the research quality elements; validity, reliability, and ethics are
examined to assess trustworthiness of this study.
______________________________________________________________________
4.1 Research paradigm
The research paradigm explains how the research is structured and conducted, as well as how
the knowledge is produced which defines the nature of the study (Collis & Hussey, 2014). Two
common research paradigms are interpretivism and positivism which outlines the two spectra
when conducting a study. Positivism is a philosophy that interprets and explains the world by
logical interpretations. Comparing it to interpretivism, which defines it as there is no objective
reality and the social phenomenon comes from subjective reality. We were interested in
performing a study with the aim to explain a relationship. This developed into a purpose and
research question where we, therefore, chose to follow the positivism paradigm to investigate
the relationship between the independent and dependent variables. Knowledge and evidence
about the relationship investigated were observed and analysed through objective
measurements (Collis & Hussey, 2014). When investigating the relationship between
sustainability performance and financial performance, we needed to have objective
observations that were measurable which in this case was secondary data of the companies in
the textile industry. Moreover, adopting the positivistic approach allowed us to interpret and
explain the data to achieve generalizations about the industry. Our positivistic approach during
the process helped us to stay objective as well as interpreting the quantitative data.
4.2 Research approach
There are typically two general research approaches, deductive or inductive. According to
O’Reilly (2012), deductive reasoning is where the researcher first develops a hypothesis from
an existing theory and thereafter collects data to analyse and test the hypothesis stated. In
contrast, an inductive approach is when the researcher develops a theory out of the data,
meaning that the theories are developed from the observations and analyses of the data. This
17
research went from specific to general by testing empirical observations which are in line with
the study’s positivistic approach to logical interpretations. Therefore, deductive reasoning was
suitable to investigate the hypotheses.
4.3 Research strategy
The research strategy is constituted on the study’s research philosophy and research approach,
where two major approaches on how to collect data exist: quantitative or qualitative.
Depending on the research philosophy and research approach, one can obtain how the
researcher will investigate the research topic. The quantitative approach collects data from a
distance whereas the qualitative approach collects data more closely to do in-depth case studies
(Collis & Hussey, 2014), which often refers to a more exploratory study (Saunders et al., 2016).
Based on this study approach and research philosophy, the quantitative approach was suitable
to test the hypotheses and to answer the research questions, where secondary data on ESG
scores and ROIC was collected and analysed. Furthermore, since this research followed a
descriptive approach with the aim of finding a relation between the variables, it was an
additional indicator to address a quantitative approach. Lastly, since the stakeholder theory is
used to explain the relationship between the variables in this quantitative research, there is no
need of developing a new theory.
4.4 Data collection
According to Boslaugh (2007), data that is collected by other people for a specific purpose is
secondary data that can be used in a secondary analysis. In this research, secondary data in the
form of ESG scores and ROIC was collected from the database Thomson Reuters DataStream.
This is to be able to investigate the relationship between ESG scores and ROIC, to see how
sustainable performance is related to financial performance. When collecting the data,
Thomson Reuters Eikon Excel was used as a tool, whereas a time series request was performed
to gather data over a specific date. The following step of the process included selecting a
random sample of companies in Europe and North America in the category of product goods.
Moreover, in the category of product goods, companies within the textile industry were
sampled. Thereafter, ENV, SOC, GOV, and the aggregated ESG score, as well as ROIC were
the chosen datatypes in the time series request. After this, the start and end dates of 2020 and
2021 were applied. This process led to an Excel file where the data was organized and finalized
to be used for the analysis. The sample included observations from 160 companies, where it
was 108 companies from Europe and 52 from North America, out of which the companies that
18
had ESG scores for the studied year were selected. The difference in the sample sizes between
the two markets is due to the number of listed companies and the accessibility of the Thomson
Reuters DataStream. This led to a sample of 69 companies for the studied period. However, it
was not possible to obtain ROIC for 2021 from 11 of the companies as they had not released
their financial reports, yet which led to a reduction to 58 companies. Lastly, after the removal
of outliers and influential values were made, this resulted in a final sample of 44 companies.
4.5 Sampling approach
When conducting research, two common approaches on how sampling can be performed are
probability sampling and non-probability sampling. The main difference between the two
sampling approaches is that probability sampling is a known biased sample, whereas non-
probability sampling is an unknown and unbiased sample (Saunders et al., 2016). In the sample
of this research, data were retrieved from one industry: the textile industry. Choosing one
industry increases the possibilities of improved results due to similar sustainable practices and
major-specific similarities between the companies, which was beneficial when analysing the
data. Moreover, to minimise the risk of uncontrolled variance, we used an industry specific
analysis due to similar industry regulations and common sustainable strategies by targeting the
population of textile companies in Europe and North America. This also contributed to a
reduced sample size. This is where the two units of analysis ESG scores and the financial
performance measurement ROIC were collected. Because of this, the most suitable sampling
approach for this research was probability sampling.
Furthermore, quantitative studies are generally conducted based on a large sample to increase
the level of certainty (Boslaugh, 2007). In this case, when choosing a specific industry, the
sample was reduced by nature, which could have had an impact on the level of certainty in this
study. However, Osborne & Costello (2004) argues that various types of studies need different
sample sizes to generate sufficient results. Based on results from earlier literature (Velte, 2017),
we concluded that the similar sustainable practices and major-specific similarities between the
companies in the textile industry would generate a more accurate result in the analysis. To
conduct a relevant study, ESG scores were taken from the year 2020 and the financial
performance measurement ROIC was taken from the year 2021. The time lag was of great
importance as ESG investments rather have a long-term impact on financial performance than
the short-term (Sholtens, 2008). However, including a time lag led to a reduction in the sample
size because 11 companies had not yet presented their financial performance for the year 2021.
19
The alternative to retrieve data from earlier years was neither suitable for this study since ESG
scores for many companies were lacking for those years. In addition, due to the scarcity of
additional independent variables, imputation was not considered. As a consequence, we chose
to proceed with the gathered sample. Lastly, there was no need to narrow down the sample to
a specific currency as there is no importance when investigating ratios.
4.6 Data analysis
It is important to use an appropriate technique to analyse the data from a secondary analysis.
According to LoBiondo-Wood & Haber (2016), the two most common ways to analyse data
are either descriptive statistics or inferential statistics. Analysing the data to be able to test
hypotheses and answer research questions indicates that it is inferential statistics, where the
aim is to compare variables to find a correlation and significance. This is where the result of
the sample taken will be a generalisation for the whole population, whereas a larger sample
size will indicate a more accurate result. This allowed us to test the hypotheses using a
probability sample to understand the relation between the variables. In comparison, descriptive
statistical analysis is where the researchers are describing and summarizing the data that
summarize the characteristics of the data set (LoBiondo-Wood & Haber, 2016). Therefore,
testing the relation between the variables ESG scores and ROIC indicates that an inferential
statistical analysis was performed. However, since the sample size in this study was limited
due to lack of reporting one can argue that generalizations that were made are not that
compelling. As all available data on companies in the textile industry was sampled from the
database, we believe that even though it is not counted as a large sample it generated an accurate
result. This resulted in a generalization of how sustainable performance affects the financial
performance of the companies in the textile industry in Europe and North America.
4.7 Secondary Data Analysis
There are several benefits with performing research based on secondary data, for example when
doing a longitudinal analysis, a subset analysis, or a cross-cultural analysis, a better result can
be achieved (Bryman & Bell, 2011). The disadvantage of using secondary data is the lack of
familiarity with the complex data where you do not have control over the quality. Choosing a
reliable source for gathering the data is therefore of high importance where using as few sources
as possible creates more reliability. To find secondary data that was reliable, the researchers
had to investigate various available sources and compare them to see where data could be
retrieved most reliably. In addition, comparing what sources had been used in similar studies
20
helped to gain knowledge and understanding of which secondary data generated the highest
quality and the most trustworthy information. This led to the choice of Thomson Reuters
DataStream. Other sources that were investigated, such as Sustainalytics, presented the ESG
scores by a risk rating instead of an aggregated score, and in addition, they are not providing
the financial data ROIC. Therefore, is Thomson Reuters DataStream a better choice since
collecting data from only one source is adding reliability. This is also since the database is well-
known and accessible through Jönköping University.
When performing secondary data analysis, a research-question-driven approach or a data-
driven approach can be followed. The research question-driven approach is substituted by
finding a suitable dataset for the research question and hypothesis (Cheng & Phillips, 2014).
In contrast, in the data-driven approach, the researcher examines a dataset to decide what kind
of questions that could be retrieved from the accessible data. When collecting the data sample
from the database Thomson Reuters DataStream, we wanted to collect a sample that was
suitable for the research question and for the hypotheses in this research, which indicate a
research question-driven approach. The database Thomson Reuters DataStream was selected
due to its objectivity, reliability, and transparency which also is indicated in previous studies
(Eccles et al., 2014; Garcia et al., 2017; Velte, 2017). In addition, it is reliable to get all the data
from a limited number of sources (Cheng & Philips, 2014).
To analyse the secondary data efficiently, the software program SPSS was chosen. The
program SPSS is a software that can be used for quantitative analysis and is owned by a
corporation that exercises complete control over the maintenance (Collis & Hussey, 2014).
This is a software that uses its own language to create various programs and macros. The
advantage of SPSS is that it produces a syntax where you can upload your excel file as a record
(Falk & Pritikin 2018). The choice of using SPSS was natural due to the researcher's earlier
experience and knowledge of the program.
4.8 Measurements and Variables
The variables should be seen as observable and measurable attributions to the study where it is
important to see the relationship between the variables (Collis & Hussey, 2014). When the
variables were identified for this study, we wanted to be certain that they were related to the
research questions and the study's aim. To explain the relationship between sustainability
21
performance and financial performance, we needed to research how to investigate the variables
in the best efficient way. Moreover, a comparison to similar previous research that had given
significant results was made. Most of the earlier studies used ESG scores to measure the
sustainable performance and ROA and ROE to measure the financial performance (Eccles et
al., 2014; Garcia et al., 2017; Velte, 2017). The measures ROA and ROE of financial
performance include the majority of a company’s operations and therefore might not indicate
the reality of the company's financial health. Minimizing the risk of a general result that was
not strongly related to sustainable performance, therefore, led to the choice of not considering
the measurement of ROA and ROE (Hagel et al., 2013). These factors led to the choice of using
ESG scores and the financial performance measurement ROIC. In addition to this, earlier
research has undertaken different control variables when using broad variables to minimize the
risk of insignificant results (Kaiser, 2020; Velte, 2017; Xie et al., 2019; Auer & Schuhmacher,
2016; Fiskerstrand et al., 2020; Duque-Grisales & Aguilera-Caracuel, 2021; Landi & Sciarelli,
2019). However, control variables were not used, since this study has already minimized the
risk of insignificant results by the choice of the textile industry, the European market, and the
North American market.
4.8.1 Independent variables: ESG Scores
Thomson Reuters DataStream Refinitiv was used when collecting ESG scores to measure the
companies’ sustainability performance. Their ESG metrics are established on 186 industry-
specific indicators that are based on data availability, impact, relevance, and comparability,
where the three factors; ENV score, SOC score, and GOV score are formed. Additionally, the
database provides an aggregate ESG score based on the three factors. The Thomson Reuters
DataStream’s scores of ENV, SOC, and GOV are overall scores based on the self-reported
information in the environment, social and corporate governance pillars. The scores are in a
range from 0-100, where a higher score argues for better sustainable performance. Furthermore,
the scores are relative to the company’s industry peer which means that the company’s ESG
scores depend on the other companies in the same industry (Refinitiv, 2020). This way of
scoring is an advantage since this research targets a specific industry and it is therefore
beneficial that the scores are depending on the company's industry peers. Another additional
benefit of using this database is that the numbers are updated every second week which
indicates the high quality of the numbers that can be retrieved. In this research, the individual
22
factors ENV, SOC, and GOV and the total ESG score were retrieved and used for the empirical
analysis as the independent variables.
4.8.2 Dependent variable: ROIC
There are several different financial performance measurements, where this study used the ratio
ROIC as the dependent variable. The database Thomson Reuters DataStream Refinitive was
used to gather ROIC information on the companies. ROIC is a profitability ratio that shows to
what extent the amount of capital invested appears back as a return or a loss. The ratio covers
all activities of investment and has four important components. The first component is the
operating income, the second is the tax adjustment to the operating income, followed by the
book values for the invested capital, and lastly the timing difference (Damodaran, 2007). As
discussed in previous sections, this is an accounting-based measurement that was used to
analyse the return on sustainable investments in a company (Cachon & Terwiesch, 2008).
According to Koller et al., (2015), ROIC is a beneficial way to measure a firm’s financial
performance as it is measuring the operating performance in a ratio and is therefore comparable
across different firms. As mentioned earlier, previous research has looked at ROA and ROE
where we chose not to consider these variables due to that ROIC is considered a superior
measurement for a company's profitability. On the other hand, it is important to recall that
ROIC will not inform which segment in the company has generated the value (Koller et al.,
2015). However, since this study’s aim is not to specifically explain which segment generates
the value this is not of importance for the researcher to consider.
Figure 3: How ROIC is calculated
4.9 Regression Analysis
This study aimed at answering hypotheses and therefore, the requirement of a method to
analyse the hypotheses was essential. To measure the relationship between the independent and
dependent variables to see if there was a relationship, how strong it was, and if it was positive
or negative, two linear regression analyses were appropriate (Studenmund, 2014). A regression
analysis allows the study to find relationships among the variables which is done by comparing
changes in one variable with one or several other variables. In this case, the independent
variables are the different ESG factors; ENV, SOC, and GOV as well as the aggregated ESG
23
factor, whereas the dependent variable that the analysis wanted to explain is ROIC. Moreover,
since this study’s hypotheses wanted to explore a relationship rather than causality between the
variables, a regression model was suitable. This was further supported as earlier research
indicates that a common way to see the relationship between the variables was by using
regression analysis (Velte, 2017, Garcia et al., 2017, Han et al., 2016). To investigate the
correlation in the linear regression analysis, the Pearson Correlation test is a common
measurement (Studenmund, 2014). The Person correlation coefficients have a range from -1 to
+1, where a value close to -1 or +1 indicates a strong relationship between the variables. A
correlation of +1 signifies that when one variable increases by 1 unit, the other variable
increases by 1 unit meaning that the coefficients are strongly positively related to one another.
On the other hand, a negative number close to -1 indicates that when one variable increases,
the other variable decreases, representing strongly negative coefficients (Studenmund, 2014).
In this study, delimitations to a specific industry and markets contributed to minimizing the
chances of insufficient results, and therefore the choice of not considering control variables
was made. Therefore, this study investigated a simple linear analysis between the independent
variable ESG and the dependent variable ROIC, as well as a multiple linear analysis between
the different independent variables ENV, SOC, and GOV and the dependent variable ROIC.
This was appropriate as the aim was to explain how each of the independent variables relates
to ROIC.
The simple linear regression model is constituted as follows:
𝑦!,
=β0 +β1x1,i+
𝜀!
The model has one coefficient β, which explains the effect the independent variable (xi) has
on the dependent variable (
𝑦!,
). This implies that when xn,i rise with 1 unit,
𝑦!,
will rise with
βn units. The intercept of the dependent variable (
𝑦!,
) considering the independent variable is
0, is explained by β0 (Studenmund, 2014). However, it is, in reality, rare that the independent
variables can account for the whole value of y which is why the error term
𝜀!
is added. The
error term shows the effect that other factors have on y which the variables included in the
model do not explain.
24
Therefore, to investigate the relationship between aggregated ESG performance and ROIC,
the following simple linear regression model was established:
𝑅𝑂𝐼𝐶!,#$
1=β0 +β1ESGi,t+ε
Moreover, to investigate each independent variables ENV, SOC, and GOV in relation to ROIC,
a multiple linear regression model was formed as well:
𝑅𝑂𝐼𝐶!,#$
1=β0 +β1ENVi,t +β2SOCi,t+ β3GOVi,t+ε
The variable ENVi,t is the variable explaining the environmental factor of ESG, SOCi,t is the
social factor of ESG, whereas GOVi,t is the governance factor of ESG.
When performing regression analyses, it is important to have in mind that statistical errors
might occur (Moore et al., 2016). Type I and type II errors are the most common if the data
sample collected is not sufficient enough to reflect the population. Admitting a type I error
refers to when a null hypothesis is rejected when it is in reality true. In this study, this would
mean that in reality, there is not a significant relationship between the variables but due to a
type I error, the result would be that there is a falsely significant relationship between the
variables. If having a lower significance level, the chances of committing a type I error increase
as the requirement of empirical evidence is lower for rejecting the null hypothesis.
Controversially, admitting a type II error means accepting the null hypothesis when it is, in
reality, false, meaning in this case that the variables have a relationship in reality in the
population. To minimise the risk of a type II error, a higher significance level is sufficient.
Therefore, to minimise the risk of committing a type I or type II error, this study used three
significance levels; 90%, 95%, and 99% to obtain one that is low enough to prevent a type I
error, as well as a high enough to prevent a type II error (Moore et al., 2016). This is supported
by previous research in the field, which also has used both significance levels (Velte, 2017,
Garcia et al., 2017, Han et al., 2016).
4.10 OLS Regression Model
When choosing an estimation method, it was of high importance that it was suitable for the
data and the purpose of the study (Collis & Hussey, 2014). Different techniques could be used
when analysing a regression model, where Studenmund (2014) argues that Ordinary Least
Squares, (OLS), is a useful method to explain a relationship between the independent and
25
dependent variables. In this study, this estimation technique was used to investigate the
relationship between the chosen variables, ESG scores, and ROIC to fulfil the research purpose.
OLS is a linear regressions model that analyses quantitative relationships by calculating the
coefficients that minimize the sum of the squared residuals. Using an estimated regression
equation will present a result that is as close as possible to the observed data. This is possible
due to the summed squared residuals being minimized by using the OLS method (Studenmund,
2014). This indicates that the regression model in this study has explained the relationship
between the variables as closely as possible for the companies in the textile industry. To assure
that OLS was an efficient and reliable method for the data in this study, certain assumptions of
the data had to be fulfilled as the implication of using this estimation method was that if they
were not fulfilled, it would have generated an inadequate regression model. Moreover, the
advantage of using this as the estimator method is that if the assumptions are met, it is the best
linear unbiased estimator (BLUE). This means that it is the most efficient estimator method
that is also linear and unbiased to analyse the data. There are six assumptions to consider
establishing an efficient OLS model. Most importantly are the assumptions of linearity, random
sampling, and the absence of multicollinearity as it is otherwise not possible to set up the OLS
estimation (Studenmund, 2014).
The first assumption to enhance an optimal OLS estimation of the regression coefficients is the
requirement that the regression models need to be linear in parameters. To ensure linearity, the
models were formed based on the OLS linear regression parameters. Moreover, the linearity of
the error terms can be estimated by plotting the residuals of the variables against the fitted
values of the variable, to observe if there is a difference between the observed values and the
predicted, fitted values (Newbold et al., 2013). The predicted values illustrate the mean values
of the chosen variable in the regression model. If the model is linear, the residuals will show
no pattern around the fitted values of the dependent variable, but rather be stable over the range
of predicted values of the dependent variables (Newbold et al., 2013).
The second assumption of OLS regression is that the sample used in the linear regression has
to be collected from the population randomly. Moreover, the OLS estimators need to have fixed
independent variables, that is the independent variables should affect the dependent variable
and not the other way around as correlation is expected and not a causal relationship
(Studenmund, 2014).
26
The third assumption in the OLS regression is that the expected value of the mean error terms
needs to be zero. This ensures that the error terms do not depend on the independent variables
(Studenmund, 2014).
Moving on to the fourth assumption, for the OLS estimator to be efficient in the analysis, no
multicollinearity should be visible. This study is based on one simple linear regression model
and one multiple linear regression model, where multicollinearity can only occur in the
multiple regression model. As in the simple linear model, there is only one independent
variable and therefore this assumption automatically holds (Studenmund, 2014). When testing
for multicollinearity in the multiple linear regression model, the independent variables ENV,
SOC, and GOV should not have a linear relationship. A common way to discover
multicollinearity is by looking at correlation among the independent variables or performing a
Variance Inflation Factor (VIF) test. To ensure that there is no multicollinearity issue, the VIF
values should not exceed 10 as this would indicate a high correlation among the independent
variables which can cause statistical errors (Newbold et al., 2013). Furthermore, some authors
imply that a VIF value above 3 might also cause concerns for multicollinearity (Studenmund,
2014).
The fifth assumption applied to OLS is that the error terms in the regression should exhibit
homoscedasticity and no autocorrelation. To ensure homoscedasticity, the error terms should
have constant variance over time (Newbold et al., 2013). Otherwise, there will be
heteroscedastic errors in the linear regression model contributing to that the standard errors of
the OLS estimate cannot be fully trusted and can hence result in incorrect estimations.
Moreover, autocorrelation between the residuals means that the error terms in the different
observations are correlated with each other which is violating the efficiency of OLS. To assure
no autocorrelation, the error terms should be independent and identically distributed (IID).
Durbin-Watson autocorrelation tests are a common measurement to provide evidence of no
autocorrelation. The Durbin-Watson test has an outcome range between 0-4 where an outcome
close to 2 means no autocorrelation. A value less than 2 indicates positive autocorrelation
whereas a value higher than 2 means negative autocorrelation. When interpreting the Durbin-
Watson test, values between 1.5 and 2.5 indicate no autocorrelation (Newbold et al., 2013).
27
The last and sixth assumption of OLS is that the error terms should have a normal distribution
based on the independent variables. This is an optional assumption as there is no requirement
for the dependent variable to have a normal distribution for OLS to hold (Newbold et al., 2013).
However, by performing a goodness-of-fit to test for normality of the error terms of the
independent and dependent variables, one can assure a normal distribution and therefore
perform hypothesis testing based on parametric tests.
4.11 Research Quality
4.11.1 Validity and reliability
According to Heale & Twycross (2015), the quality of research can be measured and
accomplished by validity and reliability. To ensure that the research process is accurate in a
quantitative study that follows a positivistic approach, finding variables that are related to the
purpose of the study is crucial. In addition, one needs to be able to make inferences about the
data collected. When finding data and variables that reflect the area that is studied, researchers
must take into account errors such as misleading procedures, samples, measurements, and
incorrect information. Starting the process of finding suitable variables and processes for this
study, consideration of earlier researchers' processes helped to increase knowledge in the area
and in addition avoid making common mistakes. The choice of the ESG variables to measure
the sustainability performance of the companies was based on some important factors. First of
all, it is the most commonly used index (Ahi et al., 2018; Van der Vaart et al., 2018) to measure
sustainability performance as well as it is an index that explains all areas of sustainability
through an aggregated score. Furthermore, another factor due to the ESG’s reporting was
accessibility through the Jönköping University, which indicates its reliability. Therefore, to
explain and analyse sustainability performance, the ESG variables were the most suitable
choice for us. Finding a variable that could explain the financial performance part of the study,
we investigated what earlier studies had used and what type of result they gained from their
choice of variables. As this study aimed to explain sustainability performance in relation to
financial performance it was important that the financial performance could explain
sustainability operations. The majority of studies used ROA, ROE, and Tobin's Q as the
financial performance measurement (Kaiser, 2020; Velte, 2017; Xie et al., 2019), whereby
there was a chance of them not explaining the reality of the company’s financial performance
and a risk of a general result (Hagel et al., 2013). Using these measurements could therefore
lead to not fulfilling the reliability of explaining the relation between sustainability
28
performance and financial performance. This indicated the choice of using ROIC as the
variable explaining the financial performance.
To accomplish objectivity, Saunders et al., (2016) imply that one must have an unbiased
research process to avoid misinterpretations and unreliable analyses. To further ensure that
reliability and objectivity were met in the study, we, therefore, had to establish a process where
the empirical data could deliver consistent and reliable results. Here it was of importance to
choose a secondary data source that contributed to this. In addition, the process needed to be
clear and transparent so other researchers can perform it with the same result (Collis & Hussey,
2014). After investigating the different options, we could conclude that the Thomson Reuters
DataStream was the one most commonly used, frequently updated, and most undertaken by
other researchers which also were the reason for the choice. This was also one of the only
providers of an aggregated score of ESG which was desired in this study. Additionally, the
Thomson Reuters DataStream is one of the biggest providers of ESG reporting which adds
reliability to the collected data (Gallego-Alvarez, & Quina-Custodi, 2017; Garcia et al., 2017).
Moreover, as mentioned earlier, the choice of gathering secondary data from only one source
added both reliability and validity (Cheng & Philips, 2014). The data gathered on ROIC was
further controlled and re-calculated by reviewing the companies´ annual reports to ensure that
the data was reliable. The next step after the data collection was to analyse it. This was
performed by using the software program SPSS which allowed the researchers to present the
data in a reliable, transparent, and ethical way, where manipulation of data was eliminated
throughout the whole process. By following and fulfilling the OLS assumptions to obtain a
stable regression model further proves validity and reliability.
Lastly, this study was performed by two authors, allowing the research to have two
perspectives, resulting in a lower likelihood of individual bias. The researchers maintained a
consistent structure throughout the process by working collaboratively and constructing mutual
decisions to avoid distinctions, resulting in a more reliable study. By considering these areas
in the research one can consider that reliability and validity were attained through the selection
of methods and ways to analyse the data.
4.11.2 Ethical consideration
When conducting research, ethical considerations and moral standards need to be taken into
account when analysing and presenting the result. To reassure moral standards and ethics
29
throughout the process, guidelines from Bell & Bryman (2011) have been followed. First of
all, the well-being of researchers was assured by treating each other with respect and that
privacy was protected to reduce the risk of wrongdoings and discomfort. Moreover, honesty
and transparency were important to present an ethical study which was undertaken
continuously by arguing and explaining every step of the research.
This study was based on secondary data and did not involve any collection of primary data,
which resulted in that ethical questions about how to obtain primary data was not needed to be
considered. Bell & Bryman (2011) argues that when using secondary data one can ensure a
high degree of ethics since the risk of causing harm to a participant is eliminated. Nevertheless,
the usage of secondary data was presented with full transparency to make sure that the
presentation of the study is reliable and ethical.
30
5. Empirical Results & Analysis
_____________________________________________________________________________________
In the following section the empirical findings and analysis from the data will be presented and
interpreted. Firstly, the descriptive statistics and correlation will be presented, followed by the
assumptions that were made to establish an efficient OLS regression model. Thereafter, a part
including the empirical results from the two OLS regression analyses will be presented,
followed by a final summary to conclude on the results.
_______________________________________________________________________
5.1 Descriptive Statistics
To present the key variables used in the two regression analyses, descriptive statistics for all
independent variables and the dependent variable is presented in Table 5.1 below. The table
shows that the sample included 44 ESG ratings after eliminating outliers and influential values
where the ESG scale performance is measured on a range from 0-100. As can be seen, the table
includes minimum and maximum values, mean values, standard deviation, and skewness of
the variables. The factors ENV, SOC, GOV, and as well as an aggregated score of these factors,
the ESG score, display the companies´ sustainability performance during the year 2020. The
dependent variable ROIC explains the company's financial performance during 2021. The
company with the highest ESG score had 92.05 whereas the company with the lowest ESG
score had 22.23. This shows that the companies in the textile industry in Europe and North
America vary in sustainable performance where the mean ESG score of the companies is
59.6850 with a standard deviation of 17.90186. The ESG score is an aggregated score based
on the other independent variables; ENV, SOC, and GOV. Therefore, the individual ENV,
SOC, and GOV ratings have a higher span from the lowest to the highest score. One can see
that the SOC score has the highest mean value of 61.1175, followed by the ENV score which
has a mean value of 60.6895, and lastly, the GOV score has a mean value of 55.6120.
Moreover, 44 observations of ROIC were retrieved after delimiting influential values. Notably,
the dependent variable ROIC has a quite wide range from the lowest value of -6.9 to the highest
value of 29.88 which indicates that the financial performance of the companies varies. ROIC
has a positive mean value of 10.8691 with a standard deviation of 8.89957 indicating that in
general, the companies have increased their financial performance during 2021.
A high skewness of the variables can add complexity to the regression analyses as this would
imply that the distribution is asymmetrical, that is a non-normal distribution of the variables
31
(Studenmund, 2014). Since this study wants to explore linear relationships between the chosen
independent and dependent variables, it is therefore important that there is a relatively low
skewness of the variables. As can be seen in the table 5.1, neither ENV, GOV, SOC, ESG nor
ROIC exhibit a skewness statistic value that is higher than +1 or lower than -1, meaning that
none of the variables have a substantially skewed distribution.
Table 5.1: Descriptive Statistics of the variables used in the regression analyses.
5.2 Correlation between the Variables
Table 5.2 below illustrates the parametric test Pearson Correlation of the two regression models
for the chosen independent and dependent variables on the 5% level. This was made to show
the correlation between the coefficients. When interpreting table 5.2, it is visible that the
correlation results have a relatively large span of level of correlation to each other. However,
all variables are positively related. The high correlation between the independent variables
ENV, SOC, and GOV to ESG is expected as the ESG score is an aggregated score based on
these variables. An interesting observation was that the factors ENV and SOC are highly
correlated by 0.768 and have a statistically significant relation to each other on a 1% level. This
indicates that when a company has a high score in one of them, it can lead to a high score in
the other. Furthermore, an intriguing observation was that all the ESG components have a
visible positive correlation with ROIC.
In table 5.2, one can see that the aggregated ESG score has the highest positive correlation of
0.299 with ROIC. The SOC score and the GOV score have the second-highest positive
correlations of 0.287 with ROIC. Lastly, the ENV score obtains a positive correlation with an
ROIC of 0.165. Furthermore, the two-tailed significance tests show that the 5% level, the ESG
score is statistically significantly correlated to ROIC, whereas the GOV score and SOC score
32
are statistically significantly correlated to ROIC on the 10% level. However, one cannot
conclude that ENV is statistically significantly related to ROIC on neither the 5% or the 10%
levels.
Table 5.2: Pearson Correlation of ENV, GOV, SOC and ROIC
5.3 OLS Model Assumptions
Following the findings of the descriptive statistics and the correlation among the variables, this
part will present the classical assumptions of the OLS regression that were tested to establish
reliable linear regression models.
5.3.1 Linearity
At first, when the expected values of the independent variables were plotted against the
predicted values, outliers were spotted which had to be removed to obtain linearity of the
variables. Thereafter, the following QQ-plots in figures 5.4,5.5,5.6,5.7 of the independent
variables were formed. Additionally, in the first plot of the expected values against the
predicted values of the dependent variable, there was an issue of influential values. These were
removed to ensure linearity which led to figure 8. This resulted in the following five figures
which show normal QQ-plots of each variable where the residuals represent the difference
between the observed values and the fitted values of the variables (Newbold et al., 2013).
33
When looking at figure 5.4, illustrating ENVs difference between observed and fitted values,
one can see that it in general follows the 45 degree-line implying that the variable is linear.
However, the variable has a few observed values that differ from the expected values,
indicating that it has some excess kurtosis. This means that the ENV score is peaked and not
fully linear. Moreover, figure 5.5 shows that GOV has a distribution along the 45 degree-line
indicating linearity of the variable. Figure 5.6 of the variable SOC also follows the 45 degree-
line and therefore also holds for linearity. Subsequently, the aggregated variable ESG in figure
5.7s observed values have a linear relationship to the expected values. Lastly, figure 5.8 shows
that the dependent variable ROIC has a linear relationship between the observed and expected
values. Therefore, the assumption of linearity of the variables holds.
Figure 5.4: Normal Q-Q Plot of ENV Score 2020: expected values against the observed values of the
independent variable ENV.
Figure 5.5: Normal Q-Q Plot of GOV Score 2020: expected values against the observed values of the
independent variable GOV.
34
Figure 5.6:Normal Q-Q Plot of SOC Score 2020: expected values against the observed values of the
independent variable SOC.
Figure 5.7: Normal Q-Q Plot of ESG Score 2020: expected values against the observed values of the
independent variable ESG.
Figure 5.8: Normal Q-Q Plot of ROIC 2021: expected values against the observed values of the
dependent variable ROIC.
5.3.2 Random Sampling of Observations
This study followed a probability sampling method with a random selection of companies in
the textile industry in Europe and North America, which allows for making statistical
inferences about the whole industry. Moreover, the study had a higher number of observations
35
in the sample compared to the number of parameters that were estimated which is required for
the OLS regression. To analyse correlation and not causality, this was established by the
study’s modelling and hypotheses, as well as including a one-year time lag on the dependent
variable ROIC. When including a time-lag on the dependent variable, fixed independent
variables are established. Hence why the second assumption of OLS holds.
5.3.3. A Conditional Mean of Zero
In figure 5.9 below which illustrates the distribution of the error term of the dependent variable
ROIC in the form of a histogram, one can see that the residuals of the regression have a mean
of 4.68E-17 which is very close to 0. Therefore, the assumption of a conditional mean of zero
holds.
Figure 5.9: Histogram illustrating the distribution of the residuals of the dependent variable ROIC 2021.
5.3.4 Absence of Multicollinearity
The following table 5.3 illustrates collinearity statistics in the form of a VIF test among the
independent variables ENV, SOC, and GOV and the dependent variable ROIC. The variable
ENV has a value of 2.643, whereas GOV has 1.243 and lastly SOC has a value of 2.436. This
proves that none of the independent variables exhibit a higher value than 3, which indicates
that there is no multicollinearity problem in the regression model. Since the assumption of no
multicollinearity holds, the model is reliable and stable to estimate the regression coefficients
as there is no high correlation among the independent variables.
36
Table 5.3: VIF statistics test to discover collinearity of the multiple regression model.
5.3.5 Spherical Errors: Homoscedasticity and Non-Autocorrelation
To detect if the dependent variable ROIC follows a homoscedastic pattern, a scatterplot
illustrating the variables´ standardized residuals against the standardized predicted values was
formed. When looking at figure 5.10, one can see that the residuals of the dependent variable
have homoscedasticity as the variance of the error terms randomly fluctuate around zero. In the
case of a heteroscedastic pattern, the error terms would have fluctuated unevenly around the
zero line.
Figure 5.10: Scatterplot of the dependent variable’s Residuals vs Predicted values of the regression.
The data sample is not based on time-series data which is one way of assuring low
autocorrelation. Furthermore, Durbin-Watson autocorrelation tests of the simple linear model
and the multiple linear model between the independent and dependent variables were
performed to provide evidence of no autocorrelation. When looking at table 5.4 showing the
Durbin-Watson test of the simple linear model one can see that the independent variable ESG
autocorrelation to the dependent variable ROIC is 1.397 which is relatively close to 2. It is
37
therefore a low positive autocorrelation between the variables. This means that when ESG
increases, this leads to an increase in the time lagged ROIC as well.
Table 5.4: Durbin Watson test for autocorrelation of the dependent variable ROIC and the independent
variable ESG
Additionally, table 5.5 of the multiple regression model shows that the independent variables
ENV, SOC, and GOV, and the dependent variable ROIC have an autocorrelation outcome of
1.467. This implies a very low positive autocorrelation between these independent variables
and the dependent variable. Since the values from the two Durbin-Watson tests are very close
to 1.5, the assumption of no autocorrelation still holds.
Table 5.5: Durbin Watson test for autocorrelation of the dependent variable ROIC and the independent
variable ENV, SOC, and GOV.
5.3.6 Normally distributed Error Terms
The previous scatter plots in figures 5.4,5.5,5.6,5.7,5.8 illustrate QQ-plots that the error terms
are generally normally distributed along the 45-degree line. Additionally, to ensure normality,
the goodness-of-fit test Shapiro-Wilk was performed on all the variables. Table 5.6 shows the
Shapiro-Wilk test of normality for each of the independent variables where a statistical value
close to 1 indicates a normal distribution. As can be seen, ENV has a value of 0.919, GOV has
a value of 0.0969, SOC has a value of 0.956, ESG has a value of 0.957 and lastly, ROIC has a
value of 0.982 given a 95 % confidence level. This means that we accept the null hypothesis
on each of the variables which states that the variable is normally distributed. According to the
QQ-plots in figures 5.4,5.5,5.6,5.7,5.8 and the tests of normality in table 5.6, one can assume
that the error terms are normally distributed.
38
Table 5.6: Test of Normality, Shapiro-Wilk tested at 95% confidence intervals
5.4 Results from The Regression Analyses
After analysing and interpreting the results from the descriptive statistics, correlation among
the variables, and as well as the OLS assumptions, the two regression analyses will be
performed to test the two study hypotheses. Since the OLS assumptions were obtained, the
simple linear regression model and the multiple linear regression model of this study can be
trusted and are reliable to test the hypotheses based on a parametric normal distribution. First
of all, the result from the simple linear regression analysis between ESG and ROIC will be
presented and explained in more depth. Thereafter, the result from the multiple regression
analysis between ENV, SOC, GOV, and ROIC will be examined and explained.
5.4.1 ESG Performance Relation to ROIC
This part will present the empirical results of the first hypothesis that involves the simple linear
regression model of the independent variable ESG and the dependent variable ROIC.
As previously mentioned, the simple linear regression model was to test the relationship
between ESG and ROIC and this resulted in the following alternative hypothesis H1:
● H1: ESG performance has a positive relation to ROIC
The Pearson Correlation test in table 5.7, resulted in a visible positive correlation between the
two variables ESG and ROIC of 0.299. The 2-tailed t-test showed a significance of 0.049 and
at a 95% confidence level, the p-value of 0.049 is less than 0.05. Therefore, the null hypothesis
was rejected at a 95% confidence level and one can conclude that there is a statistically
39
significant positive relationship between the two variables. Therefore, the alternative
hypothesis H1 is accepted.
Table 5.7: Pearson Correlation of ESG and ROIC
Moreover, to illustrate the positive relationship of the variables in the simple linear regression
model, a scatterplot was established in figure 5.11. The empirical results from figure 5.11 show
evidence that ESG and ROIC have a vague positive increasing curve where an increase in ESG
performance of the textile companies leads to an increase in their ROIC. The R Squared
illustrates the amount of the variance of ROIC that is explained by ESG in the regression model
which is 0.089. This means that 8.9% of the variance of the dependent variable ROIC is
explained by the model. This indicates that there are other major factors influencing the
performance of ROIC.
Figure 5.11: Scatterplot between ESG Score 2020 and ROIC 2021 including R squared.
5.4.2 ENV, GOV, and SOC Performance Relation to ROIC
Moving on to the second part which will present the multiple regression model’s hypothesis
testing and empirical results. The model includes the independent variables ENV, GOV, and
SOC and the dependent variable ROIC to investigate if there is a diverse relationship between
40
each of the ESG factors and ROIC. This resulted in hypothesis H2 with the three following
alternative hypotheses:
● H2a - ENV performance has a positive relation to ROIC
● H2b - GOV performance has a positive relation to ROIC
● H2c - SOC performance has a positive relation to ROIC
To find evidence of a relationship between the independent variables and the dependent
variable in this multiple linear regression model, a Pearson Correlation matrix was formed
which is found in table 5.2. The Pearson Correlation matrix was based on a 95% and a 99%
confidence interval. The table showed that each independent variable has a diverse relationship
to ROIC, where the social performance of a company, that is the SOC score, correlated to ROIC
of 0.287. Similarly, the governance performance of a company, the GOV score also correlated
to ROIC of 0.287. The environmental performance of a company, ENV, was shown to have
the lowest correlation to ROIC among the independent variables with a value of 0.165. This
shows that each of the independent variables is positively related to ROIC but with a diverse
impact.
When performing the hypotheses tests, one can conclude that none of the independent variables
are statistically significantly related to ROIC on a 95% confidence level. The p-values found
in table 5.2 show that SOC has a p-value equal to 0.059, GOV has a p-value equal to 0.059,
and ENV has a p-value equal to 0.285. Since all these p-values are higher than 0.05, we cannot
reject the null hypothesis on a 95% confidence level. However, it is visible that both the SOC
score and the GOV score have a significant relationship to ROIC based on a 90% confidence
level as their p-value is equal to 0.059 which is less than 0.1. One can therefore reject the null
hypothesis based on a 90% confidence level and prove that there is a significant relation
between SOC and ROIC, as well as between GOV and ROIC. This means that on a 90%
confidence level, H2b and H2c are accepted and proven. However, the ENV score cannot be
proven to have a significant relation to ROIC based on the confidence intervals 90%, 95%, or
99% as its p-value is equal to 0.285 which is higher than 0.1. Therefore, the alternative
hypothesis H2a cannot be proven.
Table 5.5 shows the R squared of the multiple regression model, which shows that the
independent variables ENV, GOV, and SOC together explain 0.144 of the dependent variable
ROIC. This means that 14.4% of the variance of the dependent variable ROIC is explained by
41
the model. Similar to the simple regression model between ESG and ROIC, this multiple linear
model cannot fully explain ROIC. This implies that there are other major factors influencing
ROIC. Moreover, based on the Durbin-Watson test in table 5.5, three individual scatter plots
were made to illustrate the relation between each of the independent variables and the
dependent variable in the multiple regression model.
By firstly looking at figure 5.12 below, illustrating a scatterplot of the individual relation
between ENV and ROIC, one can see that this independent variable has the lowest, yet positive
linear relationship with ROIC. Even though the positive relation is not statistically significant,
it is visible in the scatterplot that there is a weak positive relation between the variables. The R
squared value shows 0.027 indicating that 2.7% of the variance of ROIC can be explained by
ENV. This implies that the environmental factor is not as sufficient to explain ROIC as the
social and governance factors are.
Figure 5.12: Scatterplot illustrating the relation between ENV Score 2020 and ROIC 2021
Figure 5.13 shows a scatterplot of the individual relation between GOV and ROIC, which
indicates a slightly stronger positive relationship among the variables. One can see that the
variables follow a positive line with an R squared value of 0.082. This means that 8.2% of the
variance of ROIC can be explained by GOV.
42
Figure 5.13: Scatterplot illustrating the relation between GOV Score 2020 and ROIC 2021.
Lastly, figure 5.14 presents the individual relationship between SOC and ROIC where a
positive linear relationship is visible. This means that SOC has a positive impact on ROIC.
Moreover, the R squared value of SOC and ROIC is equal to 0.082, meaning that 8.2% of the
variance of ROIC can be explained by SOC.
Figure 5.14: Scatterplot illustrating the relation between SOC Score 2020 and ROIC 2021.
43
6. Discussion & Conclusion
_____________________________________________________________________________________
The purpose of this chapter is to discuss the results and analyses from chapter 5 in the light of
previous research, including theoretical contributions, managerial recommendations,
limitations and future research. This will be followed by a conclusion to specifically answer
the research questions of the study.
______________________________________________________________________
6.1 Theoretical Contributions
The empirical results and analysis from the two regression models based on the following two
models first presented in 3.2 show that each independent variable is positively correlated to the
dependent variable ROIC.
Figure 15: Research model for H1 in this study. Source: Created by the authors for the study’s purpose.
Figure 16: Research model for hypotheses H2abc in this study. Source: Created by the authors for the
study's purpose.
The aggregated ESG score of the companies shows the strongest positive relation to ROIC and
has a significant relationship on the 5% level which resulted in that H1 was accepted. This
means that textile companies that exhibit a high aggregated ESG score, that is performing well
on both environmental, social, and governance activities have a stronger return on invested
capital. This is further explained by the study's theoretical framework which suggests that
44
companies that are concerned about their stakeholders will invest in more sustainable activities
and this will benefit the companies and their stakeholders (Freeman et al., 2010). Since there
is a significant relation between ESG performance and ROIC, stakeholders benefit when
companies perform high on ESG. In this case, this means that when a textile company increases
its ESG performance, investors yield a higher return on invested capital. Moreover, the
literature review showed that the requirement of ESG reporting, as well as upcoming legislation
in the European market, suggests that stakeholders demand sustainable development from the
textile companies (Luján et al., 2020; European Parliament, 2021; Silk & Lu, 2022). This,
therefore, implies that companies that are concerned for their stakeholders have followed this
which signifies better financial performance.
The multiple regression analysis of the three ESG factors showed that SOC and GOV are
significantly positively related to ROIC on the 10% level, meaning that a company's
performance in social and governance activities has a stronger positive relation to ROIC
compared to its performance in environmental activities. This resulted in H2b and H2c being
accepted on a 90% certainty, whereas a positive significant relation between ENV and ROIC
that H2a assumed could not be proven. Interpreting the results from the analysis supports the
theory of companies that take the stakeholder perspective into account (Freeman et al., 2010).
The literature review examined that the textile industry has been under high social pressure
from its stakeholders to increase their sustainability (Özdemir et al., 2011), which supports that
the companies that have been more concerned with social activities contribute to a high SOC
score have increased their ROIC. Moreover, since the collected ESG scores are industry
specific (Refinitiv, 2020), this supports the stakeholder perspective further since the GOV score
has a significant relation to ROIC. This might be related to the legislation and regulations that
are upcoming within Europe and the regulations and legislation on ESG reporting that are
upcoming in Europe as well as the increased demand for ESG reporting in both Europe and
North America (Silk & Lu, 2022; European Parliament, 2021). This, therefore, implies that the
companies that have adopted a stakeholder perspective in the textile industry have increased
their governance performance and reporting which resulted in better financial performance.
This study’s result and analysis are based on the models above to investigate how the
independent variables ESG relate to the dependent variable ROIC by investigating correlation
among the variables. However, it is important to acknowledge that a model cannot theoretically
determine which variable is independent or dependent. This contributes to that it might be a
45
case of causality where the dependent variable ROIC can affect the ESG performance.
Companies that profitable, that are having a stable financial performance, can be more likely
to invest in sustainability activities, as they can afford this. However, companies that are
developing their business and have not become profitable yet, might be less likely to invest in
sustainability activities due to a lack of resources.
Previous research that has studied this subject has obtained mixed results, with some indicating
a positive relationship, (Kaiser, 2020; Velte, 2017; Xie et al., 2019), and some indicating a
negative relationship (Duque-Grisales & Aguilera-Caracuel, 2019; Landi & Sciarelli, 2019).
The earlier studies´ results are aligned with this study’s results, indicating that in general, ESG
performance has a positive relation to accounting-based measurements of financial
performance. As mentioned, this might be due to the increasing demand and legislation of ESG
activities and reporting that have developed rapidly during the last couple of years, and
therefore the data is very limited. Taking the results from the regression analyses into
consideration, even though there is a statistical positive relation between ESG and ROIC, it is
not particularly strong. This is in line with previous research studied on the topic (Kaiser, 2020;
Velte, 2017; Xie et al., 2019), where the weak positive relationship can be due to the lack of
data on the subject. ESG activities in the textile industry have been rapidly developed during
the last decade, and therefore, ESG reporting has previously started to increase. Since ESG
activities are implemented to meet long-term results, a time lag of one year of ROIC is therefore
only able to reflect short-term results from sustainable development.
Moreover, earlier studies have used a multiple regression model with several control variables
in their models to cover the factors influencing the financial performance (Kaiser, 2020; Velte,
2017; Xie et al., 2019; Auer & Schuhmacher, 2016; Fiskerstrand et al., 2020; Duque-Grisales
& Aguilera-Caracuel, 2021; Landi & Sciarelli, 2019), where their R squared have explained
the variability of financial performance better. Interpreting the result from the single regression
model as well as the multiple linear regression model, one can see that ESG, ENV, SOC, and
GOV cannot explain the variability of ROIC fully, which indicates that the textile companies
return on invested capital is affected by many other factors. Since most previous studies have
used cross-sectional industries, increased knowledge about the textile industry relation between
ESG performance and financial performance contributes to a more specific area and hence why
filled in the desired gap. However, this study aimed to examine the relation between ENV,
46
SOC, GOV, and ROIC of the textile companies´ in Europe and North America and therefore
control variables was not applied.
6.2 Managerial Implications
The increased knowledge obtained from this research allows companies and stakeholders that
are operating or interested in the textile industry in the European and North American markets
to see the benefits of sustainability performance. Moreover, hopefully, this knowledge will
encourage the companies to emphasize sustainable development and hence lead to a more
sustainable world.
The simple regression analysis showed a significant positive relation between ESG and ROIC
which contributes to the knowledge that companies that establish sustainable investments
increase their financial performance in the textile industry in the two markets. This gained
further understanding of the actual impact ESG scores have in the industry. Moreover, the
multiple regression analysis contributed to the findings of a visible difference between the
independent variables ENV, SOC, and GOV in relation to ROIC in the textile industry. ENV
showed the lowest positive and insignificant relation to ROIC which shows that the
environmental performance of the textile companies has less impact on their financial
performance.
According to earlier research (Kaiser, 2020; Velte, 2017; Xie et al., 2019; Auer &
Schuhmacher, 2016; Fiskerstrand et al., 2020; Duque-Grisales & Aguilera-Caracuel, 2021;
Landi & Sciarelli, 2019), this result is not coherent which suggests that the weak environmental
performance relation to financial performance is industry-specific. This might be due to that
investing in environmental activities in the textile industry yields more long-term results
compared to social and governance activities where a short-term positive result can be seen.
For the companies investing in ESG activities, the results show that there are short-term
financial benefits, which suggest that the long-term results will be sufficiently higher. This is
important for stakeholders and companies to take into account to further develop their
sustainability to achieve sustainable financial performance. Simultaneously, the stakeholder
perspective allowed the researcher to interpret how companies act based on stakeholders. The
increasing regulations and social pressure found in the literature imply that the textile
companies' stakeholders are more concerned about social and governance activities, and
47
therefore it has a stronger relation to financial performance. The result presented, which goes
in line with earlier studies presents evidence for the stakeholder perspective to hold which
concluded that companies should follow the theory to increase their relations with stakeholders
which in return give financial benefits.
The literature reflects that companies that are increasing their sustainability performance give
long-term financial results (Sholtens, 2008; UN, 2019), which implies that a time lag of
financial performance of one year used in this study may not be adequate. This means that
sustainable development is a performance that can be indicated in the financial performance in
the long run. The decision to have a time lag of one year is as mentioned the result of data
availability. However, the results from this study show a positive relation between ESG scores
and ROIC, which proves that sustainable investment can have an impact on financial
performance even in the short run. To conclude, the study has therefore provided evidence that
the companies that report well on ESG have taken the stakeholder perspective in the industry.
This gives guidance to the companies, stakeholders, and other researchers in the textile industry
to enhance sustainable development.
6.3 Limitations
This quantitative study encountered several limitations that need to be acknowledged to present
a comprehensive assessment. These are because of the fact that they were not possible to
execute or that they were not realized before the process started. However, the existence of the
limitations is not affecting the reliability and validity of this research. Firstly, it is critical to
understand the limitations of conducting a quantitative study with the use of secondary data. It
is important to find reliable sources that show reliability and validity. As presented in chapter
4, the data was retrieved from Thomson Reuters DataStream to add reliability to the research.
However, one needs to take into account that there are multiple providers of ESG reporting
with different methods on how to present the performance which could be a factor that similar
studies show different results. Another limitation is that the study chose to only include one
financial performance measurement, ROIC, whereas other studies have included more
variables. However, the purpose was to explain ESG’s relation to ROIC, and therefore the
result from the analysis is efficient.
Moreover, reporting on ESG is not mandatory in the two markets and is therefore not performed
by all the companies in the textile industry. This led to that it was not possible to obtain data
48
from all companies in the industry. When planning the method process, we could see that in
earlier years few companies had reported their ESG performance which also was the reason for
the chosen year 2020 was examined. In addition to this, the limitation of the time lag could be
discussed. The time lag of one year can affect the result in this study, as sustainability
investments are intended to give long-term results. None of the two regression analyses in this
study show a strong relationship, indicating that a short time lag of one year may not be
sufficient to address the impact on the financial performance given from sustainability
performance. The limited years of companies reporting on ESG made such investigation
impossible, contributing to that only a short-term result could be obtained. Lastly, the
researcher wants to mention the Covid-19 pandemic that has taken place over the chosen
research period, where the majority of companies in the textile industry have reported that they
encountered major obstacles due to the pandemic. As the pandemic has had a negative effect
on the world economy, textile companies´ financial performance may have been sufficiently
lower in 2021 compared to earlier years. Moreover, the companies might also encounter issues
with their work on sustainable development and performance.
6.4 Future research
The last part of the discussion presents suggestions for future research based on the limitations
of this study. This study aimed to investigate markets where there were similar sustainable
practices, where a suggestion for future research could be to include more markets to be able
to analyse if there are any differences between the markets regarding sustainability
performance and financial performance. If investigating this, control variables may need to be
considered as in the case of including more markets, these can differ significantly when it
comes to sustainability performance. This could help other researchers to increase their
knowledge on how sustainable performance and financial performance could be improved.
Furthermore, in the future, it would be interesting to consider a longer time lag on the dependent
variable, as sustainable development activities tend to have long-term financial results and not
short-term. Since this analysis only had a one-year time lag due to restrictions on available
ESG data from earlier years, a long-term analyse could not be performed. It can take time
before the financial performance of a company is affected when a company establishes
sustainability activities for long-term positive results. It would therefore be interesting to see
how the variables relate when including a time lag of five to 10 years.
49
Additionally, as mentioned in the limitations, the study only examined the relation to one
financial measurement ROIC whereas other studies examined more financial variables such as
ROA, ROE, and Tobin's Q. For future research, if considering multiple cross-sectional markets,
a suggestion is to take control variables and several financial performance variables into
account due to the differences in sustainable performance between developed and undeveloped
markets. This to understand what factors that influence ROIC.
The increasing demand and the upcoming legislation on ESG reporting for companies in the
European and North American markets will increase the amount of available data. This will
lead to that researcher can obtain data from more companies which will increase the possibility
to get more generalized results from the population. Lastly, it would be interesting to include
an external factor as an independent variable when investigating the relation to see if it has
affected the sustainability performance and financial performance. For instance, taking the
impact of Covid-19, one could investigate what the effect has been on ESG scores and ROIC.
This would be interesting since this study showed that the companies in the textile industry
obtained a relatively low financial performance during 2021, where the impact of the pandemic
could be a logical explanation for this (Craven et al., 2022). This would be interesting as many
industries have been affected by the pandemic and to see whether the sustainability
performance and financial performance in combination have been impacted.
6.5 Conclusions
The purpose of the study was to explain the relation between sustainability performance and
the financial performance of companies in the textile industry in Europe and North America.
This was performed by quantitative research investigating the relation between sustainability
performance and financial performance variables. The study’s research questions were the
following:
● What is the relationship between ESG performance and financial performance of
companies in the textile industry in Europe and North America?
● What is the relationship between each of the ESG factors and financial performance in
the companies?
The first question in the research is connected to the results for hypothesis H1 simple linear
regression analysis. The analysis showed that ESG scores are significantly positively related to
50
ROIC on the 5% significance level. This means that there is a positive relationship between
ESG scores and ROIC of companies in the textile industry in Europe and North America. The
results imply that companies should focus on sustainable development to achieve sustainable
financial performance and remain competitive in the market. This also indicates that the
companies benefit from considering the stakeholder perspective. The second research question
was answered by analysing hypothesis H2abc based on a multiple regression model. This gave
a mixed result. Each of the ESG factors ENV, SOC, and GOV showed a correlation to ROIC
but it was only SOC and GOV that had a significant positive relation to ROIC on the 10%
significance level. This concludes that companies in the textile industry in Europe and North
America that score high on social and governance reporting have a higher return on ROIC,
whereas a high score on environmental reporting is not significantly positively related to an
increased ROIC.
A final conclusion is that ESG scores and ROIC are correlated where ESG has a positive impact
on ROIC of the companies in the textile industry in Europe and North America. The different
ESG factors have a diverse relationship with ROIC which imply that sustainability
performance impact on financial performance is industry specific.
51
7. References
______________________________________________________________________
Ashwin Kumar, N. C., Smith, C., Badis, L., Wang, N., Ambrosy, P., & Tavares, R. (2016).
ESG factors and risk-adjusted performance: a new quantitative model. Journal of sustainable
finance & investment, 6(4), 292-300.https://doi.org/10.1080/20430795.2016.1234909
Auer, B. R., & Schuhmacher, F. (2016). Do socially (ir) responsible investments pay? New
evidence from international ESG data. The Quarterly Review of Economics and Finance, 59,
51-62. 10.1016/j.qref.2015.07.002
Beckman, T., Markham, D., & Khare, A. (2014). Greenwashing: A proposal to restrict its
spread. Journal of Environmental Assessment Policy and Management.
DOI:10.1142/S1464333214500306
Bi, Z. (2011). Revisiting System Paradigms from the Viewpoint of Manufacturing
Sustainability. Sustainability, 3, 1323–1340. https://doi.org/10.3390/su3091323
Boslaugh, S. (2007). Secondary Data Sources for Public Health (1st ed., pp. 2-10).
Cambridge University Press.
Brundtland Report. (1987). Report of the World Commission on Environment and
Development: Our Common Future. United Nations, New York. Retrieved 14 February 2022
from: Our Common Future (Brundtland Report) (PDF, 1 MB, 20.03.1987)
Burritt, & Schaltegger, S. (2014). Accounting for sustainability in production and supply
chains. The British Accounting Review, 46(4), 323–326.
https://doi.org/10.1016/j.bar.2014.10.006
Cachon, G., & Terwiesch, C. (2008). Matching supply with demand. McGraw-Hill
Publishing.
Cheng, H. G., & Phillips, M. R. (2014). Secondary analysis of existing data: opportunities
and implementation. Shanghai archives of psychiatry, 26(6), 371–375.
https://doi.org/10.11919/j.issn.1002-0829.214171
52
Carley, M., & Christie, I. (2000). Managing Sustainable Development; Earthscan: London,
UK.
Clark, G. L., Feiner, A., & Viehs, M. (2015). From the Stockholder to the Stakeholder: How
Sustainability Can Drive Financial Outperformance. Social Science Research Network.
https://doi.org/10.2139/ssrn.2508281
Collis, J. & Hussey, R. (2014). Business Research, a practical guide for undergraduate and
postgraduate students. Palgrave Macmillan higher education
Craven, M., Liu, L., Wilson, M., Mysore, M. (2022). Covid-19: Implications for business:
COVID-19: Briefing note #98, March 30, 2022. Retrieved 1 April 2022, from:
https://www.mckinsey.com/business-functions/risk-and-resilience/our-insights/covid-19-
implications-for-business
Dalal-Clayton, B. & Bass, S. (2002). Sustainable Development Strategies: A Resource Book;
Earthscan: London, UK. https://doi.org/10.4324/9781849772761
Damodaran, A. (2007). Information transparency and valuation: can you value what you
cannot see? Managerial Finance, 33(11), 877–892.
https://doi.org/10.1108/03074350710823836
Deng, X., Kang, J. K., & Low, B. (2013). Corporate social responsibility and stakeholder
value maximization: Evidence from mergers. Journal of Financial Economics, 110(1), 87–
109. https://doi.org/10.1016/j.jfineco.2013.04.014
Devuyst, D. (2000). Linking impact assessment and sustainable development at the local
level: the introduction of sustainability assessment systems. Sustainable Development 8(2),
67–78. https://doi.org/10.1002/(SICI)1099-1719(200005)8:23.0.CO;2-X
53
Diener, J., & Habisch, A. (2020). A plea for a stronger role of non-financial impact in the
socially responsible investment discourse. Corporate governance (Bradford),21(2), 294-306.
https://doi.org/10.1108/CG-01-2020-0039
Drempetic, S., Klein, C. & Zwergel, B. (2020). The Influence of Firm Size on the ESG Score:
Corporate Sustainability Ratings Under Review. J Bus Ethics 167, 333–360.
https://doi.org/10.1007/s10551-019-04164-1
Dresner, S. (2008). The Principles of Sustainability, 2nd ed.; Earthscan: London, UK.
https://doi.org/10.4324/9781849773249
Duque-Grisales, E., & Aguilera-Caracuel, J. (2021). Environmental, social and governance
(ESG) scores and financial performance of Multilatinas: Moderating effects of geographic
international diversification and financial slack. Journal of Business Ethics, 168(4), 1-
20.10.1007/s10551-019-04177-w
Eccles, N. S., & Viviers, S. (2011). The origins and meanings of names describing
investment practices that integrate a consideration of ESG issues in the academic literature.
Journal of business ethics, 104(3), 389-402. 10.1007/s10551-011-0917-7
Eccles, R., Ioannou, I., & Serafeim, G. (2014). The Impact of Corporate Sustainability on
Organizational Processes and Performance. Management Science, 60(11), 2835–2857.
https://doi.org/10.1287/mnsc.2014.1984
Edmans, A. (2011). Does the stock market fully value intangibles? Employee satisfaction and
equity prices. Journal of Financial Economics, 101(3), 621–640.
https://doi.org/10.1016/j.jfineco.2011.03.021
Epstein, M & Rejc Buhovac, A. (2014). Making Sustainability Work: Best Practices in
Managing and Measuring Corporate Social, Environmental, and Economic Impacts. In
Making Sustainability Work (2nd ed.). Berrett-Koehler Publishers, Incorporated.
54
Eriksen, S., Aldunce, P., Sekhar, B.C., D’almeida. M.R., Molefe, J. (2011). When not every
response to climate change is a good one: Identifying principles for sustainable adaptation.
Journal of Climate and Development, (3),7-20. https://doi.org/10.3763/cdev.2010.0060
Ester, C., & Perkins, R. (2021). How Do Companies Respond to Environmental, Social, and
Governance (ESG) ratings? Evidence from Italy: JBE. Journal of Business Ethics, 171(2),
379-397. http://dx.doi.org/10.1007/s10551-020-04441-4
European Parliament. (2021). Corporate Sustainability Reporting Directive. Retrieved 15 of
February 2022 from:https://eur-
lex.europa.eu/legalcontent/EN/TXT/?uri=CELEX:52021DC0188
Falk, & Pritikin, J. (2018). Computer Programming in Quantitative Analysis. In The SAGE
Encyclopaedia of Educational Research, Measurement, and Evaluation
Fatemi, A., Glaum, M., & Kaiser, S. (2018). ESG performance and firm value: The
moderating role of disclosure. Global Finance Journal, 38, 45-64. 10.1016/j.gfj.2017.03.001
Ferrell, A., Liang, H., & Renneboog, L. (2016). Socially responsible firms. Journal of
Financial Economics, 122(3), 585–606. https://doi.org/10.1016/j. jfineco.2015.12.003
Fiskerstrand, S. R., Fjeldavli, S., Leirvik, T., Antoniuk, Y., & Nenadić, O. (2020).
Sustainable investments in the Norwegian stock market. Journal of sustainable finance &
investment, 10(3), 294-310. https://doi.org/10.1080/20430795.2019.1677441
Freeman, R. E. (1994). The Politics of Stakeholder Theory: Some Future Directions. Business
Ethics Quarterly, 4(4), 409–421. https://doi.org/10.2307/3857340
55
Freeman, R. E., Harrison, J. S., Wicks, A. C., Parmar, B. L., de Colle, S. (2010). Stakeholder
theory: The state of the art. Academy of Management Annals, 4(1),p. 403-445.
https://doi.org/10.1080/19416520.2010.495581
Friede, G., Busch, T., & Bassen, A. (2015). ESG and financial performance: aggregated
evidence from more than 2000 empirical studies. Journal of Sustainable Finance &
Investment, 5(4), 210-233. https://doi.org/10.1080/20430795.2015.1118917
Gallego-Alvarez I., & Quina-Custodio I.A., (2017). Corporate Social Responsibility
reporting and varied-ties of capitalism: An international analysis of state-led and liberal
market economies, Corporate Social Responsibility and Environmental Management, vol. 24,
no. 6, pp. 478–495.
Garcia A.S., Mendes-Da-Silva W., & Orsato R.J., (2017). Sensitive industries produce better
ESG performance: Evidence from emerging markets, Journal of Cleaner Production, vol. 150,
pp. 135–147.
Han, J-J., Hyun, J. K., & Yu, J. (2016). Empirical study on relationship between corporate
social responsibility and financial performance in Korea. Asian Journal of Sustainability
Social Responsibility, 1 (2), 61–76. https://doi.org/10.1186/s41180-016-0002-3
Heale, R., & Twycross, A. (2015). Validity and reliability in quantitative studies. Evidence -
Based Nursing, 18(3), 66. http://dx.doi.org/10.1136/eb-2015-102129
Hagel, J., S. Brown, J., & Lui, M. (2013). Success or struggle: ROA as a true measure of
business performance. Retrieved February 25, 2022, from:
https://www2.deloitte.com/content/dam/insights/us/articles/success-or-struggle-roa-as-a-true-
measure-of-business-performance/DUP505_ROA_vFINAL2.pdf
Hedstrom, G., S. (2018). Sustainability: What It Is and How to Measure It. De.
https://doi.org/10.1515/9781547400423
56
Hübel, B., & Scholz, H. (2020). Integrating sustainability risks in asset management: the role
of ESG exposures and ESG ratings. Journal of Asset Management, 21(1), 52-69.
http://dx.doi.org/10.1057/s41260-019-00139-z
Hussain, N., Rigoni, U., & Cavezzali, E. (2018). Does it pay to be sustainable? Looking
inside the black box of the relationship between sustainability performance and financial
performance. Corporate Social Responsibility and Environmental Management, 25(6), 1198-
1211. https://doi.org/10.1002/csr.1631
Hörisch, R., Freeman, E., & Schaltegger, S. (2014). Applying Stakeholder Theory in
Sustainability Management: Links, Similarities, Dissimilarities, and a Conceptual
Framework. https://doi-org.proxy.library.ju.se/10.1177/1086026614535786
Kaiser, L. (2020). ESG integration: value, growth and momentum. Journal of asset
management, 21(1), 32-51. https://doi.org/10.1057/s41260-019-00148-y
Kates, R.W., & Parris, T.M. (2003) Long-term trends and a sustainability transition. 100
(14), 8062–8067. https://doi.org/10.1073/pnas.1231331100
Kolk, A. & Van Tulder, R. (2010). "International business, corporate social responsibility and
sustainable development". International Business Review, 19(2), 119-125.
10.1016/j.ibusrev.2009.12.003
Koller, T., Goedhart, M. & Wessels, D. (2015). Valuation: Measuring and Managing the
value of Companies. 6th ed. Hoboken, New Jersey: John Wiley & Sons, Inc.
KPMG. (2020). The time has come, The KPMG survey of sustainability reporting. Retrieved
15 February 2022 from: https://assets.kpmg/content/dam/kpmg/xx/pdf/2020/11/the-time-has-
come.pdf
57
Landi, G., & Sciarelli, M. (2019). Towards a more ethical market: the impact of ESG rating
on corporate financial performance. Social Responsibility Journal, 15(1), 11-27.
https://doi.org/10.1108/SRJ-11-2017-0254
LoBiondo-Wood, G., & Haber, J. (2016). Nursing research (9th ed., pp. 310-333).
Luján, C., Guereca, L., Franco-García, M., & Heldgeweg, M. (2020). A Life Cycle Thinking
Approach to Analyze Sustainability in the Textile Industry: A Literature Review. Journal of
Sustainability 2020, 12(23), 10193. https://doi.org/10.3390/su122310193
Maas, K., Schaltegger, S., & Crutzen, N. (2016). Integrating corporate sustainability
assessment, management accounting, control, and reporting. Journal of Cleaner Production,
136, 237–248. https://doi.org/10.1016/j.jclepro.2016.05.008
Manogna, R.L., & Mishra, A. K. (2021). Measuring financial performance of Indian
manufacturing firms: application of decision tree algorithms. Measuring Business Excellence,
ahead-of-print(ahead-of-print). https://doi.org/10.1108/MBE-05-2020-0073
Moorhouse, D., & Moorhouse, D. (2017). Sustainable Design: Circular Economy in Fashion
and Textiles. The Design Journal, 20, 1948-
1959.https://doi.org/10.1080/14606925.2017.1352713
Moore, D.S., McCabe G.P., Alwan L.C., & Craig B.A. (2016). The Practice of Statistics for
Business and Economics. Fourth Edition. New York: W.H. Freeman and Company
Muthu, S.S. (2017). Evaluation of Sustainability in the Textile Industry. Springer.
Sustainability in the Textile Industry, 9-17.
58
Newbold, P., Carlsson, W. L. & Thorne, B. M. (2013). Statistics for Business and Economics.
8th edition. USA: Pearson Education, Inc.
O'Reilly, K. (2012). Key concepts in ethnography (1st ed., pp. 104-109). SAGE.
Orlitzky, Schmidt, F. L., & Rynes, S. L. (2003). Corporate Social and Financial Performance:
A Meta-Analysis. Organization Studies, 24(3), 403–441.
https://doi.org/10.1177/0170840603024003910
Paraschiv, D., Tudor, C., & Petrariu, R. (2015). The Textile Industry and Sustainable
Development: A Holt-Winters Forecasting Investigation for the Eastern European. Journal of
Sustainability, 7(2),1280-1291. https://doi.org/10.3390/su7021280
Patwary, S. (2020). Clothing and Textile Sustainability: Current State of Environmental
Challenges and the Ways forward. Textile and Leather Review, 3(3): 158-
173.https://doi.org/10.31881/TLR.2020.16
Phan, Tran, H. X., Le, T. T., Nguyen, N., Pervan, S., & Tran, M. D. (2020). The Relationship
between Sustainable Development Practices and Financial Performance: A Case Study of
Textile Firms in Vietnam. Sustainability (Basel, Switzerland), 12(15), 5930–.
https://doi.org/10.3390/su12155930
Rajah, R., Fatimah,K., Yuri, S., & Nazia,M. (2018). Climate change and sustainable
development issues: arguments and policy initiatives. Journal of the Asia Pacific economy,
(23), 187-194.
Rajesh, R & Rajendran, C. (2020). Relating Environmental, Social, and Governance scores
and sustainability performances of firms: An empirical analysis. Business Strategy and the
Environment, 29(3), 1247–1267. https://doi.org/10.1002/bse.2429
Refinitiv (2020). ENVIRONMENTAL, SOCIAL AND GOVERNANCE (ESG) SCORES
from:https://www.refinitiv.com/content/dam/marketing/en_us/documents/methodology/refini
tiv-esg-scores-methodology.pdf
59
Reid, D. (2005). Sustainable Development—An Introductory Guide; Earthscan: London, UK.
Robinson, J. (2004). Squaring the circle? Some thoughts on the idea of sustainable
development. Ecol. Econ. 2004, 48, 369–384. https://doi.org/10.1016/j.ecolecon.2003.10.017
Rockström, J., Steffen, W., Noone, K., Persson, A., Chapin, F.S., Lambin, E.F., Lenton,
T.M., Scheffer, M., Folke, C., & Schellnhuber, H.J. (2009). A safe operating space for
humanity. Nature. 461, 472–475. https://doi.org/10.1038/461472a
Rogers, P., Jalal, K., & Boyd, J. (2008). An Introduction to Sustainable Development;
Earthscan: London, UK. https://doi.org/10.4324/9781849770477
Sala, S., Ciuffo, B., & Nijkamp, P. (2015). A systemic framework for sustainability
assessment. Ecological Economics, 119, 314–325.
https://doi.org/10.1016/j.ecolecon.2015.09.01
Saunders, M., Lewis, P., & Thornhill, A. (2016). Research methods for business students
(7th. ed.). Pearson Education.
Schaltegger, S., Hörisch, J., & Freeman, R. E. (2017). Business Cases for Sustainability: A
Stakeholder Theory Perspective. Organization & Environment, 32(3), 191–212.
https://doi.org/10.1177/1086026617722882
Scholtens, B. (2008). A note on the interaction between corporate social responsibility and
financial performance. Ecological economics, 68(1-2), 46-55.
Schöggl, J. P., Fritz, M. M., & Baumgartner, R. J. (2016). Toward supply chain-wide
sustainability assessment: A conceptual framework and an aggregation method to assess
60
supply chain performance. Journal of Cleaner Production, 131, 822–835.
10.1016/j.jclepro.2016.04.035
Searcy, C. (2012). Corporate sustainability performance measurement systems: A review and
research agenda. Journal of Business Ethics, 107(3), 239–253.
https://doi.org/10.1007/s10551-011-1038-z
SiaPartners. (2022). Impact of Stricter ESG Reporting Requirements in the U.S. Retrieved 20
April 2022, from https://www.sia-partners.com/en/news-and-publications/from-our-
experts/impact-stricter-esg-reporting-requirements-us-0
Siew, RYJ. (2015). A review of corporate sustainability reporting tools (SRTs). Journal of
Environmental Management, 164, 180–195. https://doi.org/10.1016/j.jenvman.2015.09.010
Silk, D., & Lu, C. (2022). Environmental, Social, & Governance Law Report 2022 USA.
Retrieved 20 April 2022, from https://iclg.com/practice-areas/environmental-social-and-
governance-law/usa
Studenmund, A.H (2014). Using Econometrics: A practical guide. Pearson Edu- cation
Limited, pp. 5–6. ISBN: 978-1-292-02127-0.
UN. (2019). United Nations Development Programme. Human Development Report 2019.
Human Development Index. Retrieved 7 March 2022 from:
https://hdr.undp.org/sites/default/files/hdr2019.pdf
Van Marrewijk, M. (2003). Concepts and definitions of CSR and corporate sustainability:
between agency and communion. J. Bus. Ethics, 44(2–3), 95-105. 10.1023/A:1023331212247
61
Velte, P. (2017). Does ESG performance have an impact on financial performance? Evidence
from Germany. Journal of Global Responsibility, 8,(2), 169-178.
https://doi.org/10.1108/JGR-11-2016-0029
Walsh, J. P., Weber, K., & Margolis, J. D. (2003). Social Issues and Management: Our Lost
Cause Found. Journal of management, 29(6), 859-881. https://doi.org/10.1016/S0149-
2063_03_00082-5
Wang, Q., Dou, J., & Jia, S. (2016). A Meta-Analytic Review of Corporate Social
Responsibility and Corporate Financial Performance: The Moderating Effect of Contextual
Factors. Business & society, 55(8), 1083-1121. https://doi.org/10.1177/0007650315584317
Wang, Z., & Sarkis, J. (2017). Corporate social responsibility governance, outcomes, and
financial performance. Journal of Cleaner Production, 162, 1607-1616.
https://doi.org/10.1016/j.jclepro.2017.06.142
Wood, Pereira, V., Temouri, Y., & Wilkinson, A. (2021). Exploring and investigating
sustainable international business practices by MNEs in emerging markets. International
Business Review, 30(5), 101899–. https://doi.org/10.1016/j.ibusrev.2021.101899
World Footprint (n.d) Do we fit on the planet? Global Footprint Network Web site; Retrieved
4 March 2022 from:
http://www.footprintnetwork.org/en/index.php/GFN/page/world_footprint/
Wrap. (2020a). European Clothing Action Plan (ECAP) 2019. Retrieved 28 February 2022
from: http://wrap.org.uk
Wrap. (2020b). Taking Action; Textiles; Sustainable fashion and textiles. Retrieved 28
February 2022 from: https://wrap.org.uk/taking-action/textiles
Xie, J., Nozawa, W., Yagi, M., Fujii, H., & Managi, S. (2019). Do environmental, social, and
governance activities improve corporate financial performance? Business Strategy and the
Environment, 28(2), 286-300. https://doi.org/10.1002/bse.2224C
62
Zubeltzu‐Jaka, E., Andicoechea‐Arondo, L., & Alvarez Etxeberria, I. (2018). Corporate
social responsibility and corporate governance and corporate financial performance: Bridging
concepts for a more ethical business model. Business Strategy & Development, 1(3), 214–
222. https://doi.org/10.1002/bsd2.29
Özdemir, E.V., Härdtlein, M., Jenssen, T., Zech, D., Eltrop, L. (2011). A confusion of
tongues or the art of aggregating indicators – Reflections on four projective methodologies on
sustainability assessment. Renew. Sustain. Energy Rev., 15 (5), 2385-2396.
https://doi.org/10.1016/j.rser.2011.02.021
