Master thesis
Corporate social performance and financial performance link:
Mediating effect of firm’s innovation investments in the IT sector
Hieu Le
7th December 2015
University of Groningen Newcastle University Faculty of Economics and Business Business School
MSc International Business and Management MA Advanced International Business Management
Student number: S2858606 Student number: B4062249
Supervisor Supervisor
2
ABSTRACT
3
ACKNOWLEDGEMENTS
4
CONTENTS
Chapter 1. Introduction ... 7
1.1 Corporate social and financial performance ... 7
1.2 Controversies in the CSP-FP literature ... 8
1.3 Research approach ... 8
1.4 Main findings ... 8
1.5 Structure of the thesis ... 9
chapter 2. Research framework ... 10
2.1 Corporate social performance ... 10
2.2 Shareholder theory versus stakeholder theory ... 10
2.3 Resource-based view ... 11
2.4 Relevance of innovation investments and CSP in the IT sector ... 12
Chapter 3. Literature review and hypotheses ... 14
3.1 Empirical results in the previous literature ... 14
3.1.1 Negative association ... 14
3.1.2 Positive association ... 15
3.1.3 Mixed association ... 16
3.1.4 No association ... 16
3.2 Recent stream of CSP-FP literature ... 17
3.2.1 Mediated association ... 17
3.2.2 Moderated association... 18
3.3 Literature review discussion and research gap... 18
3.4 Research question ... 19
3.5 Hypotheses ... 22
3.5.1 From the shareholder theory to the stakeholder theory ... 22
3.5.2 Active engagement ... 22
3.5.3 Passive engagement ... 23
3.5.4 From the resource-based view ... 23
3.5.5 Innovation capability as one of essential intangible resources ... 24
3.5.6 CSP and innovation investments ... 24
Chapter 4. Methodology ... 26
4.1 Sample... 26
4.2 Operationalisation of the constructs ... 26
4.2.1 Corporate financial performance... 26
4.2.2 Corporate social performance ... 27
4.2.3 Innovation investments ... 28
5 4.3 Time periods ... 29 4.4 Statistical method ... 30 4.5 Validity check ... 30 4.5.1 Linearity ... 30 4.5.2 Collinearity ... 30 4.5.3 Multicollinearity... 30 4.5.4 Heteroscedasticity ... 31 4.5.5 Endogeneity ... 31 Chapter 6. Findings ... 32 6.1 Descriptive statistics ... 32 6.2 Hypotheses testing ... 32
6.2.1 Step 1: Overall CSP-FP link ... 32
6.2.2 Step 2: CSP and innovation investments link ... 33
6.2.3 Step 3: Innovation investments and FP link ... 34
6.2.4 Step 4: Mediation effect of innovation investments on CSP-FP link ... 35
6.3 Robustness check ... 36
Chapter 7. Discussion ... 38
7.1 Results discussion ... 38
7.2 Limitations ... 40
Chapter 8. Conclusion ... 42
8.1 Research objective and outcomes ... 42
6
LIST OF ABBREVIATIONS Abbreviation Explanation
CSP Corporate social performance
CFP Corporate financial performance
IT Information technology
R&D Research and development
KLD Kinder Lydenberg Domini
RBV Resource-based view
OLT Organisational learning theory
ROA Return on assets
ROE Return on equity
ROI Return on investments
ROS Return on sales
7
CHAPTER 1. INTRODUCTION
1.1 Corporate social and financial performance
Unlike in the past, firms are now pressured not only by traditional industry-based competitive forces described by Porter (1980), but also by non-financial concerns, such as environmental issues, changes in regulations, customer expectation shift, employee’s health and safety (Prahalad and Hamel, 1994). Henceforth, many companies are trying to display their commitment to socially responsible activities nowadays (Perrini et al., 2011). Meanwhile, researchers also contend that organisations are required to invest in these activities and address the commercial as well as social demands of society in order to success (Van Beurden and Gössling, 2008). This raised a controversial discussion whether higher corporate social performance (CSP) lead to better financial performance (FP) (Tang et al., 2012). CSP can be defined as a measure of firm’s socially and environmentally responsible and responsiveness actions, principles, policies, and programmes (McWilliams and Siegel, 2000, Wood, 2010).
The CSP-FP link was mentioned for the first time by Moskowitz (1972) in the Business and Society Review journal. Several decades have passed since then and a large body of literature was dedicated to this topic, but the results have been inconclusive (Margolis and Walsh, 2003, Orlitzky et al., 2003). In general, 4 relationship types were identified: positive, negative, non-significant, and mixed (Soana, 2011). From their meta-analysis of the past papers, Margolis and Walsh (2003), Orlitzky et al. (2003), and Frooman (1997) concluded that empirical findings support a modest positive relationship between CSP and FP. Aguilera et al. (2007) called for a closure of the discussion, arguing there is sufficient evidence of a positive association between these 2 constructs. Therefore, one may question the need for additional research on this topic (Simpson and Kohers, 2002).
8
1.2 Controversies in the CSP-FP literature
Following this line of reasoning, this research was set out to analyse the CSP-FP relationship which appears to be still unclear. Several controversies and inconsistencies were discovered after a review of 27 articles. The results of CSP-FP association were indeed mixed. Authors of older papers tested only a simple and direct effects of CSP on FP, whereas more recent studies in the review have shifted towards a mediated and moderated CSP-FP link. The majority of literature was focused on the United States or Europe, in spite of contextual influences (Soana, 2011). Few researchers actually examined the CSP impacts on FP in a specific industry. No unified measurements of CSP and FP exist. Similar sources for CSP ratings seem to have been repeated in many articles. Therefore, these points served as main reasons for a further exploration of the CSP-FP relationship.
1.3 Research approach
Two theories were employed in this thesis. Through the lens of the stakeholder theory this study was more inclined to a positive relationship between CSP and FP. By adding the resource-based view, the impacts of CSP on FP were argued to be indirect and mediated by innovation investments. The rationale for focusing on innovation investments is because of their essential roles in organistaions’ ability to keep up with the technological progress in the industry they operate in. The scope of the thesis was narrowed down to the information technology (IT) sector due to different industry influences on the CSP-FP link (Griffin and Mahon, 1997, Moore, 2001) and also because innovation investments are assumed to be more relevant in the high-tech sector (Chen and Chen, 2010). The main research question of the thesis then was formulated as: ‘Do innovation investments mediate the relationship between corporate social performance and financial performance in the context of the IT sector?’
1.4 Main findings
9
1.5 Structure of the thesis
10
CHAPTER 2. RESEARCH FRAMEWORK
This chapter begins with the conceptualisation of corporate social performance. The shareholder theory and stakeholder theory are then elaborated in order to present the contradicting perspectives of scholars on the financial outputs from the socially responsible activities. The resource-based view is then introduced to provide a justification for a mediated relationship between CSP and FP. The last part of the chapter is dedicated to the justification for putting the emphasis on the IT sector.
2.1 Corporate social performance
McWilliams and Siegel (2000) looked at CSP simply as a measure of socially and environmentally responsible activities. Wood (1991) defined CSP as observable outcomes from firm’s configuration of socially responsible principles, social responsiveness processes, policies, and programs, as they relate to firm’s societal relationship with stakeholders. By combining these 2 definitions, CSP is conceptualised here as a measure of all socially and environmentally responsible and responsiveness actions, principles, policies, and program that organisation makes. Many authors support the idea that CSP is a multi-dimensional construct (Roman et al., 1999, Wartick and Cochran, 1985, Griffin and Mahon, 1997, Swanson, 1999, Rowley and Berman, 2000, Baird et al., 2012), however, its measurement has not been unified. Soana (2011) summarised measurement types in previous studies into 5 categories: content analysis, questionnaire surveys, reputational measures, one-dimensional indicator, and multi-construct ethical rating. This research employs multi-multi-construct ethical ratings, as they take the multi-dimensional nature of CSP into account.
2.2 Shareholder theory versus stakeholder theory
11
The stakeholder theory can be divided into normative and instrumental streams. In normative view, stakeholder groups of each firm are identified, whereas in instrumental view, the relationship of firm and its stakeholders and the outcomes from their mutual interactions are investigated (Donaldson and Preston, 1995). In line with the instrumental traditions, by satisfying the stakeholders’ demands, meeting their requirements, and being responsive to their expectations organisations can enjoy better financial results (Donaldson and Preston, 1995). This gave researchers the main drive for studying the CSP-FP link (Carroll and Shabana, 2010).
In accordance to the stakeholder theory, companies are perceived as open systems which interact on a constant basis with a broader external environmental system consisted of many different groups, or stakeholders (Perrini et al., 2011). Each exchange between firm and its stakeholders can be regarded as a single contract, or in other words, one portion of an ongoing serious of recurring contracts (Ruf et al., 2001). Henceforth, building relational contracting with stakeholders will reduce explicit transaction costs (Williamson, 1975). Along this line, Hillman and Keim (2001) argued that by converting the transactional nature of the firm-stakeholders relationship into a relational one, this relationship can bring or be transformed into intangible resources giving firms better competitive advantage, leading to better FP (Porter, 1980). At the same time, the number of companies disclosing their CSP information has been growing, as they try to show their commitment to social and environmental causes by presenting transparent and verifiable data in order to gain legitimacy and approval from their stakeholders (López et al., 2007). Therefore, through the stakeholder theory it is argued that CSP is positively related to FP in general.
2.3 Resource-based view
12
that tangible resources cannot provide competitive advantage because they can be purchased and sold in the factor market. In contrast, intangible resources can be translated in competitive advantage over a period of time by following an appropriate path (Dierickx and Cool, 1989).
In light of the RBV theory, when companies engage in socially responsible activities they might develop intangible resources that correspond to the four criteria mentioned above, leading to competitive advantage creation in both internal (human capital, innovation, and organisational culture) and external (trust, reputation, and relational contracts) dimensions (Surroca et al., 2010, Russo and Fouts, 1997, Saeidi et al., 2015, Hillman and Keim, 2001). Once firm gains a competitive advantage over its competitors, it can enjoy superior profitability (Porter, 1980). To elaborate, the intangible resource in the focus of this thesis is firm’s innovation investments, since they increase innovative capabilities, one of the most essential assets of firms among many intangible factors (Gómez-Mejía et al., 2005). Organisation’s innovative capability is closely bounded to its success and survival because of self-reinforcing effects residing on its ability or inability to keep up with the technological progress of the industry in which they operate (Cohen and Levinthal, 1990). Based on the RBV, it is assumed that firms inspiring to be more socially and environmentally responsible invest more in their innovative capabilities (innovation investments) to achieve more product and process innovations (McWilliams and Siegel, 2000), which were proven to be associated with greater financial results (Surroca et al., 2010, Lichtenberg and Siegel, 1989, Guerard Jr et al., 1987).
2.4 Relevance of innovation investments and CSP in the IT sector
13
14
CHAPTER 3. LITERATURE REVIEW AND HYPOTHESES
This chapter starts with an overview of the past literature on the CSP-FP topic. From a total amount of 27 randomly selected articles, a distinction between older and more recent ones was made based on 6 results streams. Afterwards, a discussion is provided to highlight the evidence of a positive CSP-FP link, controversies, and inconsistencies in the CSP-FP literature. Through the lens of the stakeholder theory, the CSP-FP relationship is then hypothesised as positive. Building on the RBV, this relationship is argued to be indirect and mediated by innovation investments.
3.1 Empirical results in the previous literature
The debate on the CSP-FP link was started by Milton Moskowitz (1972), who identified 14 companies with high ethical ratings and suggested that they are good investments without any conclusive evidence and disclosure of criteria for selecting these 14 firms. Later, he noted that stock prices of these firms rose higher compared to the increase of Dow-Jones, New York Stock Exchange, and Standard and Poor’s Industrials indices over the period of 6 months. However, when Stanley Vance (1975) analysed the stock price of the 14 Moskowitz-recommended firms from 1972 to 1975, he concluded that these companies performed far below the Dow-Jones, New York Stock Exchange, and Standard and Poor’s Industrials indices. Since then, a great number of scholars set out to test the relationship between CSP and FP. However, several decades have passed, yet the findings remain inconclusive (Margolis and Walsh, 2003, Orlitzky et al., 2003). Soana (2011) grouped past results into 4 categories of association: negative, positive, mixed, non-significant (See the Table 1).
3.1.1 Negative association
15
López et al. (2007) tested 55 companies included in the Dow Jones Sustainability Index against their 55 peers in Dow Jones Global Index in the time frame 1998 - 2004, with the use of 7 financial measures of firm performance - profit before tax, growth in revenue, capital, profit margin, return on earnings, return on assets (ROA), cost of capital - and recorded a negative relationship in short-term and no relationship in long-term. Baird et al. (2012) used Domini 400 Social index (part of KLD) for CSP, intrinsic value calculated by forecasts of earnings and stock price obtained from Value Line, Inc. for FP, and Linear Mixed Method to analysis a sample of 5,0573 firm-year observations from 2001 to 2008, then reported a negative association in overall.
3.1.2 Positive association
According to Waddock and Graves (1997), there are several arguments for a positive CSP-FP relationship. First, the stakeholder theory contends that there is a tension between company’s implicit and explicit costs, thus by attempting to lower implicit costs through ethical misbehaviours, firm will receive higher explicit costs. Second, the actual cost of socially responsible activities and programs are minimal compared to potentially great benefits generated afterwards. Last, ‘good management’ can do most things well.
16
In their review of 21 empirical studies from 1972 to 1992, 12 articles with a positive link, 8 with no significant relationship, and only 1 with a negative link, Pava and Krausz (1995) suggested that there is overwhelming evidence of socially responsible firms performing at least as well as the other companies. Griffin and Mahon (1997) summarised the empirical results of the CSP-FP link over a 25-years period, from 1972 to 1997, and found a support for a positive relationship. Frooman (1997) conducted a meta-analysis of 27 event studies on the association of stock market reactions with firm’s ethical misbehaviours and registered a negative reaction of the market to companies that acted socially irresponsible or illegally, thus showing an evidence for a positive CSP-FP relationship. Margolis and Walsh (2003) also conducted a meta-analysis on 160 articles that examined CSP-FP link and found 55% with a positive association, 22% with no association, 18% with a mixed association, and 4% with a negative association, giving further support for an overall positive FP-FP relationship. Another meta-analysis of 52 previous papers on CSP-FP topic was done by Orlitzky et al. (2003), who provided a modest evidence of a positive CSP-FP link.
3.1.3 Mixed association
Authors arguing for mixed relationship propose that the CSP-FP link is not linear and constant in time, but can take a ‘U-shape’ or inverted ‘U-shape’ (Soana, 2011). ‘U-shape’ means that the initial and final costs are extremely high and they are low only between medium to long term, whereas inverted ‘U-shape’ assumes an optimum level of CSP beyond which it will be costly to engage in socially responsible activities. Bowman and Haire (1975) applied content analysis to measure CSP and 5-years ROE as a proxy of FP of 82 food processing companies and found an existence of inverted ‘U-shape’ performance curve, with the firms having highest financial results in the middle range of CSP. A mixed CSP-FP relationship was also concluded by Sturdivant and Ginter (1977), who assessed 28 companies by using Moskowitz’s (1972) ratings for CSP and 10-years earning per share (EPS) growth for FP.
3.1.4 No association
17
When Alexander and Buchholz (1978) wanted to challenge findings of Vance (1975), they used the same sample and the same proxy for CSP as Vance’s (1975) together with 2-years and 5-years stock price increases for FP, then found no significant relationship. Abbott and Monsen (1979) developed a Social Involvement Disclosure (SID) scale from a content analysis of Fortune 500 firms to measure their level of CSP, then compared it to FP measured by10-year yield and reached a conclusion of no significant relationship. Arlow and Gannon (1982) analysed 7 previous studies on the CSP-FP link and stated that there is not a conclusive evidence of any relationship. Cochran and Wood (1984) provided results of weak support for a CSP-FP association after regressing Moskowitz’s (1972) ethical companies against their peers in corresponding industry groups with the use of the ratio of operating earnings to assets, the ratio of operating earnings to sales, and excess market valuation. Aupperle and colleagues (1985) found no support for the link between CSP and FP after employing their forced-choice survey based on Carroll’s (1979) construct, 1-year and 5-years ROA adjusted with risk, and a sample of 241 usable responses from Chief Executive Officers listed in Forbes 1981 Annual Directory. Similar to Simpson and Kohers (2002), Soana (2011) also tested the CSP-FP link in the banking sector but did not find any significant relationship. Unlike other scholars, the author employed three measures of CSP from unrelated agencies and both market-based and accounting-based ratios for FP.
3.2 Recent stream of CSP-FP literature
Based on the more recent papers, another 2 groups of findings can be added to the categorisation of Soana (2011). The first stream is built on exogenous factors and supports a mediated association between CSP and FP. The second stream is built on endogenous factors and argues for a moderated association between CSP and FP.
3.2.1 Mediated association
18
3.2.2 Moderated association
In their paper, Tang et al. (2012) found that the CSP-FP relationship is moderated by the way firms engage in socially responsible actions, after using KLD ratings for CSP and ROA for FP. Peng and Yang (2014) proposed that the CSP-FP link is moderated by ownership concentration in companies. They then hand-collected data on Taiwanese firms’ pollution control investments from 1996 to 2006 to measure CSP and used both market-based and accounting-based data for FP (Tobin’s Q and scores from a factor analysis with varimax rotation from cash flow to assets, ROE, ROA, and EPS).
3.3 Literature review discussion and research gap
The review of the past studies clearly illustrate a very mixed set of results. Although there is a higher number of papers supporting a positive association between CSP and FP, their theoretical model, operationalisation of CSP and FP, and context are inconsistent. To elaborate, Griffin and Mahon (1997), Margolis and Walsh (2003), McWilliams and Siegel (2000) argued that the evidence of a direct relationship between CSP and FP seems to be questionable, for many mediating and moderating factors might have been omitted from the theoretical model, resulting in inclusive findings. Only 4 out 29 reviewed articles moved towards the mediated and moderated CSP-FP link. Notably, these papers are also more recent. Peloza (2009) argued that one of the biggest research gaps is that scholars overlooked mediators in the CSP-FP link. To address this issue, the thesis inclines more toward a positive CSP-FP link which is, however, assumed to be indirect and mediated based on the RBV.
19
Furthermore, even though organisations in different sectors interact with different stakeholder groups, only 4 out of 29 reviewed articles were based on a single industry, which is similar to the claim of Griffin and Mahon (1997), who reported that the majority (78%) of the CSF-FP literature collected samples from multiple industries. Waddock and Graves (1997) argued that effects of CSP on FP can differ across industries. Rowley and Berman (2000) suggested that CSP research should be conducted in a specific industry due to diverse internal and external pressures. Tang et al. (2012) proposed that FP can be affected by industry growth and employed this construct as a control variable. According to Simpson and Kohers (2002), studies that focus on one industry tend to have a small sample size, for most of them were conducted in the settings of USA or UK (Soana, 2011). Therefore, another contribution of this paper is its large sample of firms in the IT sector spreading over many countries. The rationale for narrowing the scope of the thesis on the IT industry comes from its fast-growth and dynamic nature in which innovations play an important role.
Lastly, finding a good proxy for CSP has been a problematic issue. The literature review clearly shows how CSP measurements vary from study to study. As also reflected in the review, Kinder Lydenberg Domini (KLD) index and Fortune Reputation surveys belong to the commonly used third-party sources for CSP measures. This corresponds to the findings of Griffin and Mahon (1997). The shortcoming of these proxies is that they are solely based on US companies (Griffin and Mahon, 1997). Since they are employed in the majority of CSP-FP researches, the results from these studies are also bound to be similar, which might be the reason why many authors found support of a positive CSP-FP association (Orlitzky et al., 2003). Hence, in this paper CSP is captured by Bloomberg ESG scores, which have been very popular among practitioners, but have not been adequately used in the CSP-FP context or in CSP literature in general either (Grigoris et al., 2014)
3.4 Research question
Building up on the points mentioned above, the main research question is:
‘Do innovation investments mediate the relationship between corporate social performance and financial performance in the context of the IT sector?’
In order to answer the main research question, 4 sub-questions need to be addressed first:
1. Is CSP positively related to FP?
20
Author CSP proxy FP proxy Result Rationale
Vance (1975) Moskowitz's
(1972) ratings
Stock price change Negative High implicit
costs from socially responsible activities and management opportunism López et al. (2007) Dow Jones Sustainability index
Profit before tax, growth in revenue, capital, profit margin, return on earnings, return on assets, cost of capital Baird et al. (2012) Domini 400 Social
index (part of KLD)
Intrinsic value
Moskowitz (1972) Author's own ratings
Stock price change Positive High explicit
costs from ethical misconducts, greater benefits from socially responsible actions, and good management Bragdon and Marlin (1972) Council of Economics Priorities' pollution index Return on equity
Folger and Nutt (1975)
Government pollution indices
Common stock price
Preston and O’bannon (1997) Fortune's Reputation index Return on assets Stanwick and Stanwick (1998) Fortune's Reputation index Return on sales Simpson and Kohers (2002) Community Reinvestment Act ratings Return on assets, bank loans Ruf et al. (2001)
KLD ratings Growth in sales,
return on equity, return on sales Rettab et al.
(2009)
Survey Survey
Pava and Krausz (1995) Review of 21 studies Review of 21 studies Griffin and Mahon (1997) Review of 51 studies Review of 51 studies Frooman (1997) Meta-analysis of 27 event studies
Stock price change
21 Bowman and
Haire (1975)
Content analysis Return on equity Mixed Non-linear
CSP-FP relationship Sturdivant and Ginter (1977) Moskowitz's (1972) ratings
Earning per share
Alexander and Buchholz (1978)
Moskowitz's (1972) ratings
Stock price change No
relation Too many intervening factors and equilibrium theory Arlow and Gannon (1982) Review of 7 studies Review of 7 studies Cochran and Wood (1984) Moskowitz's (1972) ratings
Operating earnings to assets, operating earnings to sales, excess market valuation Aupperle et al.
(1985)
Forced-choice survey
Return on assets
Soana (2011) Ethibel, Axia, EAL Return on average assets,
return on average equity, cost to income ratio, market to book value, price to book value, price earnings ratio Surroca et al.
(2010)
Sustainalytics scores Tobin’s Q Mediated
relation Indirect effects of CSP on FP through endogenous factors Saeidi et al. (2015) Survey based on Carroll’s (1979) constructs Balance Scorecard with 7 items:
market share growth, sales growth
return on equity, return on assets, return on investments, return on sales
Tang et al. (2012) KLD ratings Return on assets Moderated
relation
CSP-FP link influenced by exogenous factors Peng and Yang
(2014)
Hand-collected pollution control investments data
Tobin’s Q and
factor analysis with varimax rotation scores from return on equity, return on assets, earning per share, and cash flows to assets
22
3.5 Hypotheses
3.5.1 From the shareholder theory to the stakeholder theory
In the last few decades, many groups of stakeholders that have influence on firms’ business operations have asserted the need for firms to move toward more socially responsible practices (López et al., 2007) and there seems to be a consensus that by engaging in socially responsible activities, firms can collect greater financial rewards (Orlitzky et al., 2003). However, neoclassical economists, proponents of the shareholder theory, proposed that the main objective of companies and their managers is to maximise shareholders’ profits (Friedman, 1970). Engagement in socially responsible actions represents a violation of management’s responsibilities to shareholders, as they generate additional costs, while financial benefits are negligible (Friedman, 1970). On the other hand, scholars supporting the stakeholder theory suggested that this narrow view should be abandoned, as shareholders stand for only one group of stakeholders which interact with firms on a continuing basis (Freeman, 1984). Organisation’s success and survival depend on stakeholders’ satisfaction and the mutual relationship between them and their stakeholders (Freeman, 1984). Successful companies are those that have sustainable competitive advantage over their rivals, leading to superior profitability (Porter, 1980). Whereas, disappearance of firms from the market relates to the fact that they have competitive disadvantage compared to their competitors, resulting in negative financial returns and forcing them to exit from the market (Porter, 1980). This implicates that both success and survival are connected to the level of FP. That is why companies need to meet the requirements of stakeholders in order to achieve high financial performance.
3.5.2 Active engagement
23
their stakeholders into relational contracting, which reduces explicit costs and which can bring or be transformed into intangible resources providing firms competitive advantage (Ruf et al., 2001). Socially responsible activities were proved to be associated with relational contracting with stakeholders (Dentchev, 2004) and its product in form of intangible assets may be product and process innovations which occur after information and knowledge exchange between firms and stakeholders (Nidumolu et al., 2009).
3.5.3 Passive engagement
Apart from incorporating socially responsible principles into their business operations and actively engage into collaboration with its stakeholders, companies can also achieve legitimacy and stakeholders’ approval by demonstrating their strong commitment to social and environmental causes through disclosure of transparent and verifiable data on its socially responsible activities (Perrini et al., 2011). That is why the number of companies disclosing their social and environmental impacts have increased rapidly (Perrini et al., 2011). Furthermore, CSP disclosure can be employed not only to gain legitimacy and approvals, but also to mislead stakeholders’ attention from firm’s negative social performance (Lightstone and Driscoll, 2008). Lastly, companies which disclose data on their CSP are associated with better financial performance compared to their non-disclosure competitors (López et al., 2007). To summarise, firms can employ socially responsible practices as a tool, in both active and passive ways, to assist them in achieving legitimacy status and fulfilling stakeholders’ demands, resulting in greater financial outcomes. Thereby, the first hypothesis will be:
H1: In overall, CSP is positively related to FP. 3.5.4 From the resource-based view
24
bundles of resources that were collected during companies’ existence, hence, the differences in competitive advantage and financial performance depend on the characteristics of their resources (Barney, 1991, Dierickx and Cool, 1989, Wernerfelt, 1984). The reason for occurrence of differences comes from the fact that not all resources bring competitive advantage to organisations. Dierickx and Cool (1989) argued that tangible resources cannot give firms competitive advantage because they are tradable in the factor market. Whereas intangible resources can yield competitive advantage over a period of time and by following an appropriate flow (Dierickx and Cool, 1989).
3.5.5 Innovation capability as one of essential intangible resources
Among intangible resources, innovative capability is regarded as one of the most pivotal intangible assets (Gómez-Mejía et al., 2005). Firm’s level of innovative capacity also dictates its success and survival due to self-reinforcing effects proposed by Cohen and Levinthal (1990). The technologies in the industry that companies operate in keep evolving gradually, therefore, if companies do not possess sufficient innovative capabilities, they are not able to keep pace with the industry technology progress or recognise new opportunities hidden within this progress (Cohen and Levinthal, 1990). As an outcome, they end up with lower aspirations for further innovations developments and enter a vicious circle (Cohen and Levinthal, 1990). In contrast, companies that have adequate innovative capabilities are able to keep up with the industry technological advancements and tend to have high aspirations for more innovations (Cohen and Levinthal, 1990). Henceforth, they are more active in seeking new opportunities which then help them advance their innovative abilities even further, resulting in a positive circle (Cohen and Levinthal, 1990). Implicitly, as firm’s success and survival equal its financial outcomes, innovative capabilities is related to FP. Additionally, the ability to innovate new products, processes, and technologies in organisations in long-term can serve as a barrier for competitors and prevent them from imitation of their resources, maintaining more sustainable competitive advantage (Russo and Fouts, 1997).
3.5.6 CSP and innovation investments
25
to incorporate these attributes in their product and service development to make them more appealing and differentiated (McWilliams and Siegel, 2000). Besides that, many companies started promoting ‘going green’ motto because implementation of pollution and environmental impacts prevention during the R&D process not only enables them come up with new products and process innovations, but also helps them to differentiate from other competitors (Porter and Van der Linde, 1995). As a result, environmentally responsible programmes contribute to productivity gains induced by cutting down consumption of materials and energy as well as minimising waste from all business operations (Porter and Van der Linde, 1995).
Furthermore, investments in proactive socially and environmentally responsible strategies that aim at creating relational contracting with stakeholders also lead to higher innovative capabilities (Sharma and Vredenburg, 1998, Thompson and Heron, 2006, Nidumolu et al., 2009). Relational contracts are sometime considered as intangible assets themselves (Hillman and Keim, 2001), or can be also viewed as relational capital (Thompson and Heron, 2006), which give firms competitive advantage over their rivals. For example, long-term relationship between firm and its suppliers has been shown to facilitate knowledge sharing (Vurro et al., 2009). Socially responsible practices that motivate mutual interactions between companies and their customers have also been proposed to increase customer-specific knowledge (Perrini et al., 2011). By having more knowledge about their customers and suppliers, organisations have higher chance to achieve new innovations (Simon and March, 1958). Lastly, investments and participation in community development projects have been proven to be associated with higher innovative capacity of firms (Perrini et al., 2011). In summary, adoption of socially and environmentally responsible policies and programmes are assumed to trigger investments into firm’s innovative capabilities (innovation investments) in order to incorporate the elements of such policies and programmes into its production processes and products. Without innovation investments organisations are less likely to be able to recognise, assimilate, and utilise knowledge acquired from their socially and environmentally responsible engagement. These investments also increase their innovative capabilities, leading to more production process and product innovations, which in turn lower costs and increase profits. The second, third, and fourth hypotheses will be:
H2: CSP positively affects innovation investments
H3: Innovation investments have positive impacts on FP.
26
CHAPTER 4. METHODOLOGY
This chapter opens with a short description of the sample used in the thesis. It continues with justifications for measurements of 5 variables: financial performance, corporate social performance, innovation investments, firm size, and firm risk. Afterwards, a short reasoning for lagging all variables, except for financial performance, by 1, 2 and 3 years is followed by an explanation of the statistical method which is based on Baron and Kenny’s (1986) mediation testing. 4 regression models were then derived and checked for validity through several tests addressing: linearity, collinearity, multicollinearity, heteroscedasticity, and endogeneity.
4.1 Sample
The initial sample of 10,465 firms was obtained from the Bloomberg Terminal and Bloomberg ESG, where environmental, social, and governance (ESG) ratings as well as other financial data were accessed. After filtering out other sectors and leaving only the IT sector, which is in the spotlight of the thesis, the number of companies reduced to 1,226. According to Sekaran and Bougie (2010), the sample size for a business research should be at least 30. Henceforth, the sample size in this paper highly exceeds this condition. Sectors categorisation, incorporated in the Bloomberg platform, is based on the Global Industry Classification Standard (GICS), mutually developed by Standard & Poor’s and MSCI in 1999 (Standard & Poor’s, 2006). In order to account for lagged effects of CSP and innovation investments on FP, the FP measures are set at the year 2014, while data entries for other variables are taken from the year 2013 for the main hypotheses testing, then 2012 and 2011 for the robustness check.
4.2 Operationalisation of the constructs
4.2.1 Corporate financial performance
27
also external macroeconomic factors (López et al., 2007). Lastly, stock prices are important for financial stakeholders but less relevant for other stakeholder groups (McWilliams et al., 1999). In contrast, accounting-based data can reflect internal efficiency more precisely and are less influenced by external trends (Orlitzky et al., 2003). McGuire et al. (1988) found that accounting-based ratios are more associated with CSP. Cochran and Wood (1984) proposed that accounting returns may be a better proxy for FP. In overall, there seems to be a preference of accounting data in the CSP-FP literature. The commonly used accounting-based measures are return on assets, return on equity, return on sales, net income, and asset age (Griffin and Mahon, 1997, Moore, 2001). Among these indicators, ROA and ROE are more suitable for the studies of long-term effects of CSP on FP, however, ROE should be applied in a timeframe longer than 3 years (Callan and Thomas, 2009). Considering the timespan used in the thesis, ROA will be used to capture FP. Furthermore, ROA is also less susceptible to corporate manipulation compared to other accounting returns (Aupperle et al., 1985). Data for ROA were taken from the Bloomberg Terminal.
4.2.2 Corporate social performance
28
To introduce a new dataset, Bloomberg ESG is employed as a proxy for CSP in this paper. Bloomberg ESG was launched in 2009 and by the end of the year 2014, it covers 11,000 firms in more than 100 countries. The main difference that Bloomberg ESG brings lies in the mixture of a multi-construct ethical ratings, content analysis based on firm’s social disclosure, questionnaire surveys. First, Bloomberg collects data from its surveys and company-sourced filings, for instance, CSP reports, annual reports, and company websites. All data entries (or data points) are fully transparent and can be traced back to their original source. Afterwards, data points are standardised and divided into 3 dimensions: environmental, social, and governance. Last, the final overall rating is aggregated from the scores of all 3 dimension, ranging from 0 for companies that do not disclose CSP related data to 100 for those which disclose data related to all of 219 data points required by Bloomberg, then aggregated into an overall score. Another unique feature of Bloomberg ESG is that the scores from data points are weighted in relation to the industry in which companies operate (Grigoris et al., 2014).
Although Bloomberg ESG ratings have been developed for commercial purposes and have not been applied adequately as a proxy for CSP in the CSP-FP literature, their credibility cannot be denied, since many scholars from other fields have used them in their studies (Grigoris et al., 2014, Zhihong and Joseph, 2013, Rahdari and Anvary Rostamy, 2015, Halbritter and Dorfleitner, 2015, Eccles et al., 2011). Bloomberg ESG is also ranked among the most credible ratings in the Rate the Raters survey (GlobeScan and SustainAbility, 2013). The only weakness of Bloomberg ESG is its limited transparency on more detailed criteria and methodology of the rating process due to the prevention from imitation by their rivals (Delmas and Blass, 2010).
4.2.3 Innovation investments
29
4.2.4 Size and risk as control variables
In the context of CSP-FP link, Margolis et al. (2009) found that the most common control variables are firm size, risk, and industry. As the focus of the paper is put on the IT sector, there is no need to control for industry effects, hence, two controls variables that are considered here are firm size and risk. There are also other factors which need to be controlled for, for instance, leverage, physical resources, financial resources, and firm age (Surroca et al., 2010). However, due to a limited access data and time availability, only the most repeated control variables are used in this study. Data for firm size and risk were collected from the Bloomberg Terminal.
Firm size has been recognised as a determinant of both CSP and FP (Ullmann, 1985). The main assumption here is that larger firms have more resources for socially responsible programmes and policies implementations and face higher pressure from the stakeholders to disclose CSP related information (Waddock and Graves, 1997). This assumption has been supported in a number of studies. Trotman and Bradley (1981) found that firm size, measured by sales volume and total assets, is significantly related to social disclosure. Stanwick and Stanwick (1998) concluded that firm size, operationalised by annual sales, is significantly associated with CSP. In the paper of Pava and Krausz (1995), socially responsible firms were larger than non-socially responsible peers. Arlow and Gannon (1982) noted that firm size is an important variable after reviewing 7 articles on the CSP-FP topic. Following Waddock and Graves (1997), firm size in this research will be measured by the number of employees.
The primary argument for controlling for firm risk is that companies with high CSP level is regarded as better managed and less risky and low-risk firms are more likely to be able to engage in socially responsible activities (Roberts, 1992). Risk, operationalised by firm’s beta value, was found to be negatively related with social disclosure in the studies of Roberts (1992) and McGuire et al. (1988). Herremans et al. (1993) concluded that companies with higher reputation for CSP showed lower level of risk. Similar to the paper of Waddock and Graves (1997), total debt to total asset ratio will be used as a proxy for firm risk in this thesis.
4.3 Time periods
30
4.4 Statistical method
Baron and Kenny’s (1986) method is used to investigate the potential mediation effect of innovation investments on the CSP-FP link. The statistical procedure consists of 4 steps, in which there must be (1) a significant relationship between CSP and FP, (2) a significant relationship between CSP and innovation investments, (3) a significant relationship between innovation investments and FP, (4) a non-significant relationship between CSP and FP, when innovation investments are included. From these steps, 4 linear regression models are derived (See the Appendix E for explanations of the variables below)
1) ROAi = c0 + β1CSPi + β2Sizei + β3Riski + εi
2) R&Di = c0 + β1CSPi + β2Sizei + β3Riski + εi
3) ROAi = c0 + β1R&Di + β2Sizei + β3Riski + εi
4) ROAi = c0 + β1CSPi + β2R&Di + β3Sizei + β4Riski + εi
4.5 Validity check
4.5.1 Linearity
Looking at the results of the Pearson correlation test in the Appendix A, all predictor variables have a significant correlation with ROA, thus the linearity requirement for the regression models 1, 3, and 4 is fulfilled. As for the regression model 2, all predictor variables have a significant correlation with R&D expenditures, except for firm risk (Appendix A), thus the linearity assumption is not met. For that reason, a logarithmic transformation of the values of firm risk and R&D expenses was conducted and the linearity condition is not violated anymore after re-running the Pearson correlation test again (Appendix A).
4.5.2 Collinearity
Based on the figures in the Pearson correlation test (Appendix A), there is one suspicious correlation coefficient between firm size and CSP (r = 0.415) as its strength can be can be regarded as moderate (69 ≥ r ≥ 40) (Grimm, 1993). Nevertheless, the squared value of the correlation coefficient (r2 = 0.17) does not exceed the threshold 0.70. Therefore, it can be argued that there is no collinearity between the variables (Martin and Bridgmon, 2012).
4.5.3 Multicollinearity
31
4.5.4 Heteroscedasticity
The modified White’s test was conducted to check for heteroscedasticity. After regressing the unstandardized residual values on unstandardized predicted values and squared unstandardized predicted values from 4 linear regression models proposed above, the null hypothesis of homoscedasticity (or no heteroscedasticity) was rejected at the confidence level of 5%. To address the issue, heteroscedasticity-consistent standard error estimators from the paper of Hayes and Cai (2007) are employed in this thesis. Implicitly, a test for normality of residuals is no longer required.
4.5.5 Endogeneity
32
CHAPTER 5. FINDINGS
In this chapter, a short discussion of descriptive statistics of variables used in the thesis is followed by results evaluation from hypotheses testing in the statistical software SPSS. The hypotheses tests are divided into 4 steps: (1) overall CSP-FP association, (2) relationship between CSP and innovation investments, (3) impact of innovation investments on FP, and (4) mediation effect of innovation investments on FP. In all steps, data for CSP, innovation investments, size, and risk are taken from the year 2013, whereas figures for FP are from the year 2014. In order to ensure that the statistical results are robust, a robustness check is conducted in the last part of the chapter. This test contains entries for CSP, innovation investments, size, and risk from 2011 and 2012, while data for FP are still left at the year 2014.
5.1 Descriptive statistics
The Table 2 shows descriptive statistics of all variables which are used in the paper. On average, companies in the sample obtained relatively look ratings from Bloomberg. The minimum and maximum indicate that no firm achieved scores over 83.88 points or below 2.07. As for the R&D expenses, the range between minimum and maximum is extremely large. This justifies controlling for firm size and firm risk. Further support for using firm size and firm risk is because of the notable range between their minimum and maximum values. The mean and median of ROA are positive, meaning that on average firms in the sample are profitable. However, its standard deviation is relatively high, hence, there is a high variation in FP.
Variable Mean Median Maximum Minimum Std. Deviation
CSP2013 19.17 12.81 83.88 2.07 12.74
R&D2013 215,045,804 28,527,780 13,509,399,780 8,741 860,234,792
Size2013 10,395 1,932 431,212 13 30,761
Risk2013 18.5923 13.8669 799.0977 0.0022 30.5148
ROA2014 1.92 4.01 155.55 -397.28 21.09
Chapter 6. Table 2: Descriptive statistics
5.2 Hypotheses testing
5.2.1 Step 1: Overall CSP-FP link
As the first step, a linear regression analysis is used to test the hypothesis 1 which states that FP is positively influenced by CSP. The regression model 1 will be as follow:
33
Following Preston and O’bannon (1997) All of predictor variables are lagged by 1 year (t - 1), whereas the predicted variable is left without lagging (t), as it is proposed that it takes a certain period of time for the implementation of socially and environmentally responsible practices, programs, and policies to yield any value added to FP (Moore, 2001, Preston and O’bannon, 1997, Arlow and Gannon, 1982).
Variable Coefficient Constant 0.946 (1.65) CSP2013 3.808 *** (0.04) Size2013 2.904 *** (0.00) Risk2013 -1.583 *** (0.02) R2 0.019 Adjusted R2 0.016 F-statistics 9.768 *** *p<0.10; **p<0.05; ***p<0.01.
Chapter 5. Table 3: Overall CSP-FP link with ROA as a dependent variable.
The results in the Table 3 shows that CSP is positively (t-value = 3.808) and significantly (p-value < 0.01) related to ROA, thus the hypothesis 1 is supported. This means that CSP-FP link exists and in general firms with higher CSP achieve better FP. 1
5.2.2 Step 2: CSP and innovation investments link
In the second step, another linear regression model is used to test the hypothesis 2 which asserts that CSP positively affects innovation investments. The regression model 2 will be:
R&Di = c0 + β1CSPi + β2Sizei + β3Riski + εi
All variables are lagged by 1 year (t - 1) in order to keep the consistency with the previous and the next two steps, for innovation investments is assumed to occur at the same time frame as the investments in socially responsible programs, policies, practices, and activities.
1 ROA is also positively affected by firm size and negatively influenced by firm risk in a significant way.
Implicitly, larger firms tend to perform better than smaller firms. Meanwhile, high-risk companies produce worse financial outcomes compared to low-risk peers. The adjusted R2 illustrates that 1.6% of the variability of corporate
34 Variable Coefficient Constant 130.265 *** (0.06) CSP2013 7.890 *** (0.00) Size2013 5.248 *** (0.00) Risk2013 1.710 (0.04) R2 0.152 Adjusted R2 0.150 F-statistics 54.270 *** *p < 0.10; **p < 0.05; ***p < 0.01.
Chapter 5. Table 4: CSP and innovation investments link with R&D as a dependent variable.
The figures in the Table 4 demonstrate that CSP is positively (t-value = 7.89) and significantly (p-value < 0.01) related to R&D expenses, therefore, the hypothesis 2 is accepted. This means that higher CSP leads to more innovation investments which are required to assist firms in recognising, assimilating, and utilising knowledge acquired from their socially and environmentally responsible programmes, policies, and practices implementation. 2
5.2.3 Step 3: Innovation investments and FP link
The third step also includes a linear regression model for testing the hypothesis 3 which contends that innovation investments have positive impacts on FP. The regression model 3 will be as follow:
ROAi = c0 + β1R&Di + β2Sizei + β3Riski + εi
Similar to the first step, the independent and control variables are lagged by 1 year (t - 1), whereas the dependent variable is left without lagging (t), as it is assumed that it takes a certain period of time for innovations investments to turn into innovations which then enhance firm’s competitive advantage and financial performance (Gallego-Álvarez et al., 2011). Looking at the Table 5, it is clear that R&D expenses are positively (t-value = 3.679)
2 Firm size also has a positive relation with R&D expenditures, thus it is possible to say that larger firms invest
more in their innovative capabilities. Firm risk shows a positive, but non-significant association with R&D investments, meaning their mutual relationship cannot be interpreted. The adjusted R2 shows that 15% of
35
and also significantly (p-value < 0.01) related to ROA, giving support for the hypothesis 3. This implies that companies which invest more into innovations tend to have better financial outcomes. 3 Variable Coefficient Constant -3.128 *** (6.36) R&D2013 3.679 *** (0.87) Size2013 2.059 ** (0.000) Risk2013 -1.650 * (1.74) R2 0.028 Adjusted R2 0.026 F-statistics 8.951 *** *p < 0.10; **p < 0.05; ***p < 0.01.
Chapter 5. Table 5: Innovation investments and FP link with ROA as a dependent variable.
5.2.4 Step 4: Mediation effect of innovation investments on CSP-FP link
In the last step, the mediation effect of innovation investments on CSP-FP relationship is tested in the regression model 4, with all predictor variables lagged by 1 year (t - 1):
ROAi = c0 + β1CSPi + β2R&Di + β3Sizei + β4Riski + εi
According to Baron and Kenny (1986), when mediators are included in the last step, the significant relationship between dependent and independent variables in the first step needs to become insignificant in order to establish a full mediating effect, otherwise only a partial mediating effect can be accepted. In the Table 6, CSP is still positively and significantly related to ROA, even after inserting R&D expenses into the regression model, thus innovation investments are proposed to only partially mediate the CSP-FP link. 4
3 Identical to the findings in the first step, firm size has a significantly positive effect on FP, while firm risk is
significantly and negatively associated to FP. The R2 value in the Table 5 shows that 2.6% of variability in ROA
is explained by R&D investments, firm size, and firm risk. The regression model has sufficient prediction power (F = 8.951; p < 0.01).
4 Similar to the first step and the third step, firm risk still has a negative and significant impact on FP. However,
36 Variable Coefficient Constant -3.056 *** (6.28) CSP2013 3.032 *** (0.03) R&D2013 3.426 *** (0.84) Size2013 0.815 (0.00) Risk2013 -1.667 * (1.74) R2 0.031 Adjusted R2 0.028 F-statistics 7.359 *** *p < 0.10; **p < 0.05; ***p < 0.01.
Chapter 5. Table 6: CSP-FP link with ROA as a dependent variable and R&D as a mediator.
5.3 Robustness check
Following other authors (Gallego-Álvarez et al., 2011, Waddock and Graves, 1997, Roberts, 1992), except for the predicted variable ROA, all other variables are firstly lagged by 3 years (t - 3) to check of robustness of the regression models (Appendix B). As for the step 1, no differences have been spotted in the relationship between CSP and ROA. In the step 2, CSP still has positive and significant impacts on R&D investments. In the step 3, R&D expenditures continue showing a positive and signification relationship with ROA. As for the last step, the results are surprising. By inserting R&D expenses as a mediator in the regression model, the CSP-FP link has become non-significant, indicating that in a longer timeframe (3 years) innovation investments fully mediate the CSP-FP relationship. In contrast, in a shorter time horizon (1 year) innovation investments seem to have only a partial mediation effect on the CSP-FP link.
mediate the association between firm size and its financial outcomes. Considering the change in the p-value of firm size, it should be addressed in the future studies. R2 in the Table 6 shows that 2.8% of variability in ROA is
37
38
CHAPTER 6. DISCUSSION
In the first part of this chapter, the results from the hypotheses testing are related to the empirical findings from other studies as well as to the sub-research questions and the main research question. In the second part, the limitations of the thesis are stated.
6.1 Results discussion
The main goal of this thesis was to investigate the relationship between corporate social performance and financial performance in the context of the information technology sector. The CSP-FP link was hypothesised to be indirect and mediated by innovation investments. Following Baron and Kenny’s (1986) method for testing mediation effects, 4 linear regression models were derived to test 4 hypotheses. The comprehensive list of the variables applied in the hypotheses testing consists of: corporate social performance, financial performance, innovation investments, firm size, and firm risk. Bloomberg ESG scores were introduced as a proxy for CSP. Financial performance was measured by return on assets. Total R&D expenses were suggested to represent innovation investments but they did not meet linearity assumption in the third regression model, hence, logarithmic values of total R&D expenditures were used instead. Firm size was operationalised as number of employees. Firm risk was calculated as total debts to total assets ratio, however, this ratio did not fulfil the linearity requirements, thus its logarithmic values were employed.
39
The hypotheses 2, 3, and 4 were derived from the resource-based view which was used to demonstrate a mediation effect of innovation investments on the CSP-FP link. The proponents of the resource-based view propose that by participating in socially responsible activities companies can develop a set of intangible assets that give them a competitive advantage over their competitors, resulting in better financial outcomes (Barney, 1991, Wernerfelt, 1984, Dierickx and Cool, 1989). Following this logic, innovation investments were put into the focus of this paper, as they increase firms’ innovative capabilities, which fall among the most essential intangible resources that affect firms’ success (Gómez-Mejía et al., 2005). The main underlying idea is that socially and environmentally responsible engagement gives firms knowledge which can play a role in innovations creation, but innovation investments are required in order to recognise, understand, absorb, and utilise this newly acquired knowledge. Another rationale is that in order to become or appear more ethical companies need to dedicate more resources to their innovative capabilities which assist them incorporate socially responsible elements into their productions and products innovations.
40
Henceforth, the answer to the main research question would be: ‘Yes, innovation investments mediates the CSP-FP link in such a way that their partial mediation impact can be observed in the timeframe of 2 years or shorter and their full mediation effect can be seen in the time period of 3 years.’ Unfortunately, it is not possible to relate these findings to those of other studies for 3 reasons. First, most of researchers analysed innovations instead of innovation investments. Second, this thesis is restricted to the high-tech sector only. Last, as argued by Peloza (2009), one of the main research gaps in the CSP-FP literature is that scholars have overlooked the mediating factors in the association between CSP and FP. For that reason, there are not many articles in the top journals on the mediation effects of innovation investments as well as innovations. The only comparable paper could be the work of Surroca et al. (2010) who found a full mediation of intangible resources, including innovations, on the CSP-FP relationship after lagging only CSP by 1 period. Looking from another point of view, this could also indicate that a relevant research was tackled in the thesis.
6.2 Limitations
This study is not without limitations. That is why the results need to be interpreted in a careful manner. First, more innovation investments were presumed to automatically increase innovative capabilities, resulting in larger amount of innovations which then give firms competitive advantage over their rivals, leading to better financial outcomes (March, 1991, Kim et al., 2014). Nevertheless, the theoretical model would have been more comprehensive, if innovations had been included because innovation investments can be viewed as inputs, whereas innovations that organisations manage to achieve can be regarded as outputs.
Second, a one-way direction from CSP to FP was analysed here. However, Waddock and Graves (1997) proposed a concept called ‘slack resources’ in which FP-to-CSP impacts are also considered, assuming firms with better financial results are more likely to be able to invest in innovations and implement socially responsible principles and practices in their business operations. The theoretical model in this paper would have been more sophisticated if the FP-to-CSP direction had been added, including a mediation of innovation investments in it.
41
42
CHAPTER 7. CONCLUSION 7.1 Research objective and outcomes
It has already been a few decades since the positive effect of firms’ engagement in socially and environmentally responsible activities on their financial outcomes was first suggested by Moskowitz (1972). Since then a large number of scholars have set out to test this proposition, but have not managed to unify the findings (Orlitzky et al., 2003). The results in the CSP-FP literature often can be divided into 4 categories: positive, negative, mixed, or no association. The simple CSP-FP model has been developed further by including endogenous and exogenous factors as potential mediators or moderators in the most recent articles (Tang et al., 2012) or by inversing the relationship from FP to CSP in the concept of ‘slack resources’ (Waddock and Graves, 1997). Following the former stream of research, the CSP-FP link was considered indirect and mediated by innovation investments in this thesis and then tested with data from Bloomberg ESG and Bloomberg Terminal within the IT sector on a global scale.
Building on the stakeholder theory, the overall CSP-FP link was hypothesised as a positive one which was also supported by the statistical results. Through the application of the resource-based view, this link was proposed to be indirect and mediated by innovation investments. The results showed that there is only a partial mediation effect of innovation investments on CSP-FP relationship in 1-year and 2-years timeframe. However, in 3-years timespan, innovation investments fully mediate the CSP-FP association. Hence, it can be concluded that in the timeframe of 2 years or shorter some of potential benefits from firm’s social performance are directly linked to its FP, while some are reflected in FP through innovation investments. Whereas, in the time period of 3 years short-term advantages from socially responsible activities, programmes, and practices are no longer directly related to FP, but only those that have indirect impacts on FP through innovation investments remain.
7.2 Research contributions
43
7.3 Managerial implication
An important implication for managers is that socially responsible actions might incur implicit costs which, however, can be later offset by explicit costs reduction, greater financial benefits, and new acquired knowledge due to closer and stronger relationships with stakeholders. Socially and environmentally responsible practices, policies, and programmes implementation also requires firms to invest more in their innovative capabilities, which after a period of time will produce desired process and product innovations that also integrate socially and environmentally responsible elements. These innovations represent crucial intangible resources that can bring firms competitive advantage over their competitors, leading to better financial performance.
7.4 Further research
44
BIBLIOGRAPHY
ABBOTT, W. F. & MONSEN, R. J. 1979. On the measurement of corporate social responsibility: Self-reported disclosures as a method of measuring corporate social involvement. Academy of Management journal, 22, 501-515.
AGUILERA, R. V., RUPP, D. E., WILLIAMS, C. A. & GANAPATHI, J. 2007. Putting the S back in corporate social responsibility: A multilevel theory of social change in
organizations. Academy of management review, 32, 836-863.
ALEXANDER, G. J. & BUCHHOLZ, R. A. 1978. Corporate social responsibility and stock market performance. Academy of Management journal, 21, 479-486.
ANDERSON, J. C. & FRANKLE, A. W. 1980. Voluntary social reporting: An iso-beta portfolio analysis. Accounting Review, 467-479.
ARLOW, P. & GANNON, M. J. 1982. Social responsiveness, corporate structure, and economic performance. Academy of Management Review, 7, 235-241.
AUPPERLE, K. E., CARROLL, A. B. & HATFIELD, J. D. 1985. An empirical examination of the relationship between corporate social responsibility and profitability. Academy of
management Journal, 28, 446-463.
BAIRD, P. L., GEYLANI, P. C. & ROBERTS, J. A. 2012. Corporate social and financial performance re-examined: Industry effects in a linear mixed model analysis. Journal of
business ethics, 109, 367-388.
BARNEY, J. 1991. Firm resources and sustained competitive advantage. Journal of
management, 17, 99-120.
BARON, R. M. & KENNY, D. A. 1986. The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of
personality and social psychology, 51, 1173.
BOWMAN, E. H. & HAIRE, M. 1975. A strategic posture toward corporate social responsibility. California Management Review (pre-1986), 18, 49.
45
BRAMMER, S. & MILLINGTON, A. 2008. Does it pay to be different? An analysis of the relationship between corporate social and financial performance. Strategic Management
Journal, 29, 1325-1343.
BUCH, C. M., KOCH, C. T. & KOETTER, M. 2012. Do banks benefit from
internationalization? Revisiting the market power–risk nexus. Review of Finance, rfs033.
CALLAN, S. J. & THOMAS, J. M. 2009. Corporate financial performance and corporate social performance: an update and reinvestigation. Corporate Social Responsibility and
Environmental Management, 16, 61-78.
CARROLL, A. B. 1979. A three-dimensional conceptual model of corporate performance.
Academy of management review, 4, 497-505.
CARROLL, A. B. & SHABANA, K. M. 2010. The business case for corporate social responsibility: a review of concepts, research and practice. International Journal of
Management Reviews, 12, 85-105.
CHAN, J., PUN, N. & SELDEN, M. 2013. The politics of global production: Apple, Foxconn and China's new working class. New Technology, Work and Employment, 28, 100-115.
CHEN, K. H. & METCALF, R. W. 1980. The relationship between pollution control record and financial indicators revisited. Accounting Review, 168-177.
CHEN, M. L. & CHEN, K. J. 2010. Relationships among organizational innovation, learning and knowledge management in the information technology industry. African Journal of
Business Management, 4, 3191-3200.
CHEN, Y.-S., LAI, S.-B. & WEN, C.-T. 2006. The influence of green innovation performance on corporate advantage in Taiwan. Journal of business ethics, 67, 331-339.
CLEMENS, M. A., RADELET, S. & BHAVNANI, R. R. 2004. Counting chickens when they hatch: The short term effect of aid on growth. Center for Global Development working
paper.
