Corporate Social Responsibility and Value Creation -
Evidence from the United Kingdom
HARRY W. SUFFIELD*
January 2017
ABSTRACT
This thesis examines the relationship between corporate social responsibility (CSR) and value creation of U.K. listed firms over the period of 2001-2015, and aims to advance CSR as a dimension of traditional financial analysis. On a FTSE 100 sample, I test the impact of environmental, social and governance (ESG) factors on accounting and market multiples of value. I also test the impact on the underlying value drivers in the firm: return on invested capital (ROIC), weighted average cost of capital (WACC) and organic growth (OGROWTH). I analyse the outcomes in aggregate, at sector level and over time to identify which value driver is most influenced by CSR, and which aspect of CSR is most financially material. I present evidence that CSR is significant to maximising income, but is ineffective at reducing capital costs or boosting growth. Social factors – such as human capital management – are the key to achieving the long-term competitive advantage that CSR can deliver in cash flows and earnings.
Keywords: CSR, ESG, SRI Value Creation, United Kingdom Word count: 14,942
1. Introduction
This thesis researches the relationship between corporate social responsibility (CSR) and value creation of U.K. listed firms. The mainstay of existing CSR literature examines the relationship with traditional measures of financial performance, such as stock returns and profitability, yet this continues to produce inconclusive results. The inconsistency of the data should motivate further research into the individual components of profitability – the drivers of value creation. This thesis investigates the impact of CSR on key metrics of corporate profitability and aims to offer a guide where to best apply CSR considerations to traditional financial analysis.
assets three years ago), however, sends a signal to traditional investors that SRI’s ‘structural underperformance’ is quickly becoming a thing of the past and that client attitudes are changing (Breinlinger 2016). The remaining scepticism may be largely rooted in a time when SRI consisted of simply excluding stocks from one’s universe on the grounds of ethics. The SRI backdrop of this thesis encounters many different semantics and so it is noted at this stage that CSR and ESG will be used interchangeably throughout. In practice these are the umbrella terms adopted by firms and investors respectively (Chadwick 2013).
In terms of methodology, I investigate the relationship between CSR and value creation using several Ordinary Least Squares (OLS) regression specifications, each adapted from the CSR-firm-value model of Gregory et al. (2014). I explore which core CSR-firm-value driver (ROIC, WACC, OGROWTH) is most influenced by CSR and which aspect of CSR (environmental, social, governance) is the most financially material. Results are processed both on aggregate and at sector level, and robustness tests are carried out to evaluate the role of time effects (period fixed and random specifications) and industry effects (dummy variables and sample splits). The results are interpreted in two ways. Firstly, to assess statistical significance, I refer to econometric diagnostics and explanations. Secondly, to assess economic significance, I refer to the implications for investors and their valuations. More specifically, I refer to core financial statements – the balance sheet and income statement – which are the universal foundations of stock valuation models.
Advanced accounting standards (IFRS), corporate governance, shareholder rights, and rule of common law (Nobes 1998) lend themselves to reliable data and information. Furthermore, the U.K. equity market is the largest in Europe by capitalisation and is the most internationally diverse in terms of stock listings and investor-base, with 54 percent foreign ownership at the end of FY2014 (Office for National Statistics 2014). This has the potential to give it wider representativeness than one may initially give it credit for.
The main result is a statistically significant positive relationship between CSR and value creation for our total sample of U.K. listed firms. Upon the decomposition of these two elements (ESG factors and ROIC, WACC, OGROWTH), I find the strongest positive relationship between social factors and the ROIC. I find the strongest negative relationship between environmental factors and OGROWTH. WACC exhibits the weakest magnitude of results and these show CSR to be a negative influence. Further robustness tests find evidence that the positive relationship is more pronounced both in period fixed effects and industry dummy specifications. Economically, and certainly for the purposes of corporate valuation, CSR is shown to be mainly relevant to maximising income. CSR is ineffective at reducing capital expenses or boosting growth but does have risk-management properties – as shown by its strong inverse relationships with measures of idiosyncratic and systematic risk.
This thesis makes two key contributions to the existing body of literature on SRI and CSR. Firstly, it attempts to offer more practical guidance than previous investor-centric papers. Accordingly, this thesis will provide more clarity on financial materiality and economic significance of ESG to the process of value creation – that which lies at the heart of a fundamental investment approach. Earlier studies (Gregory et al. 2014; Jo et al. 2011) have investigated the impact of ESG on firm value, although these typically devote more attention to decomposing ESG than value creation. Secondly, it is set purely in a U.K. context, often overlooked by academic research that tends to prefer American samples and datasets (Aguilera et al. 2006).
2. Literature Review
2.1 Corporate social responsibility 2.1.1. CSR definitions
CSR and, by extension, corporate social performance (CSP) encapsulates a myriad of different objectives, activities and stakeholders (Heinkel et al. 2001; Schroder 2014). In defining CSR, as many as thirty-seven alternatives exist (Dahlsrud 2006). This paper follows the definition put forward by Bénabou et al. (2010) that CSR is prosocial behaviour by a firm. It can be understood as a set of strategies to reduce the differential between the social and private costs of a business. This involves firms exceeding regulatory and legal minimums in managing positive and negative externalities (McWilliams et al. 2001). This commonly manifests itself in the form of (often delegated) community philanthropy, charitable initiatives or participation in social action programs, and typically culminates in the provision of a public good. Or, as Cowton et al. (2000) put it, “financing the social economy”. CSP measures these endeavours relative to market expectations and to the actions of other firms. The best performing firms on CSP are those who prove able to integrate sustainability and ethical considerations within every operating stage of their business. CSP is universally defined within the categories of environmental, social and governance considerations. A significant section of existing literature (Brown et al. 2009; Jo et al. 2011; Hiraki et al. 2003) suggests that governance is regarded to be the most financially material category of CSR, followed by environmental and social categories respectively. The reasoning for this is intuitive – firms with well-designed control, decision-making, incentive and ethical structures will be in the strongest position to design and implement proactive social and environmental strategies. This forms the first of our three hypotheses:
Hypothesis I: governance is the most financially material aspect of CSR, i.e. reports the strongest coefficients.
2.1.2. CSR motivations
relation). Strategic (profit) motivations are found to be the most prominent and the most effective. The risk of greenwashing, therefore, is likely overstated. Largely driven by image concerns, greenwashing seeks to raise participation costs for rival firms and placate regulators. It attempts to mimic genuine CSR through a mix of high profile charity initiatives and overtly positive CSR reporting. Genuine (strategic and altruistic) and greenwashing forms of CSR can exhibit many of the same apparent behaviours – but only the former two are doing so in a calculated move to enhance long-term profits. Kitzmueller and Shimshack (2012) are also more optimistic, asserting that whilst both may yield positive outcomes, purely short-term CSR will be found out by the capital and labour markets alike. Consistently sustainable firms will retain the advantage of attracting human and investor capital (Heinkel et al. 2001) that has aligned skills and motivations. An activist investor base is a proven means to eliminating CSR pursuits that are either superfluous or outright value destroying (Cohn et al. 2016).
2.1.3. CSR measurement
2.2. Linking CSR to corporate financial performance and value creation 2.2.1. How a firm creates value
2.2.2. The CSR – value creation link
To evaluate the connection between the value creation theory above and CSR, we first turn to the existing meta-analyses that test CSR against upside performance and downside risk management measures. From this we draw our second of three hypotheses:
Hypothesis II: CSR is predominantly an upside-capturing exercise rather than a downside protection exercise, i.e. the relationship is stronger with ROIC than with WACC.
Aupperle et al. (1985) review publications throughout the 1970s and report that most performance measures detect a positive relationship, with the study of Vance (1975) proving the only exception. For instance, positive relationships are found between CSR and stock returns by Moskowitz (1972), ROE (Bragdon et al. 1972; Heinz 1976), and net income, margin, and earnings per share growth (Parket et al. 1975; Sturdivant et al. 1977). Alexander et al. (1978), however, report no significant relationship, as do Abbott et al. (1979) when comparing CSR with ten-year bond yields. Aupperle et al. (1985) conduct their own analysis using the previous method of one of their co-authors (Carroll 1979), whereby an “elaborate, forced-choice instrument” survey is administered to corporate CEOs. This novel technique requires CEOs to assign a weight budget to twenty sets of statements concerning economic, legal, ethical and discretionary business factors. From their responses, they conclude that no significant relationship can be ascertained between CSR and CFP, or, more specifically, that “varying levels of social orientation were not found to correlate”. Thus, in the words of the authors, “assessing profitability is a relatively clear-cut process, but assessing CSR is not”. They criticise a lack of measurement consistency in the academic literature and observe a similar lack of firm-risk controls in models – advice that this thesis acts upon. Notably, Aupperle et al. (1985) also make an early reference to the disclosure bias, which is now central to the CSP-CFP debate from the eyes of the investor.
Griffin et al. (1997), Margolis et al. (2007) and Revelli et al. (2014) offer more recent meta-analyses on the CSP-CFP link. Margolis et al. (2007) cover 192 effects revealed in 167 different studies over a 35-year period, and find a small positive overall effect of CSR. Interestingly, the authors find that the association is weakest when assessed through third party audits (i.e. firm disclosure bias) and mutual fund screens. This introduces the concept that a firm can perform deliver strong CSP and CFP, yet deliver lacklustre stock returns for investors. Brammer et al. (2016) point out that by studying only stock returns, socially responsible companies do not appear to be superior performers. The reason for this, however, is not their weak earnings – but rather their elevated valuations (price-earnings and price-book ratios). Furthermore, if high-scoring CSR firms have high valuations and low returns it means investors consider them less risky, i.e. they are assigned a low discount rate (cost of equity). Therefore, we must compare both market and accounting measures before exploring a full range of underlying value drivers. This forms our third and final hypothesis:
Hypothesis III: CSR shares a stronger relationship with a market multiple of value than its accounting counterpart, i.e. market ‘hype’ or overvaluation.
2.3. Sustainable and responsible investing (SRI) 2.3.1. SRI: theory and practice
ecology, equity and futurity”. Whilst these guiding principles have stood the test of time, modern investors have since given detail to the practical implications. Dimson et al. (2013) assert that SRI “seeks to accomplish both financial and social benefits” by incorporating ESG considerations into the investment process. Krosinsky et al. (2008) stress the need for ESG integration, stating that “sustainable investment contains a commitment to systematically integrate ESG within the valuation, the choice of assets and the exercise of ownership rights and duties”. They too make the critical distinction between sustainable investing and ethical investing, quoting Hudson (2006) who stated that ethical investing is “an approach driven by the value system of the key investment decision-maker”. To this end, we should view SRI as a broad spectrum, with pure non-profit philanthropy on one end, ethical investing in the centre and sustainable investing on the other. Overlaps between ethical and sustainable investing are very common, but they are not one and the same. In practical terms, any strategy on the SRI spectrum is constructed using a set of established sub-strategies, which I explain in Table A.I (Appendix). These can be used exclusively or in combinations. SRI operates on the belief that enhanced ethics and performance are positively interlinked in the long-term, in lieu of added private costs (Diltz 1995).
2.3.2. SRI: performance and prospects
does not hurt returns (Phillips et al. 2007; Hale 2016). Whilst a meta-analysis of twenty-five studies from 1996-2008 by Mercer et al. (2007) reports that an overwhelming two-thirds do not find a significant relationship, more recent research by Asmundson et al. (2001) and Sjölström (2011) covering twenty-one studies from 2008-2010 gives SRI a clearer direction and even a slight positive tilt. It is also proven that the increased costs associated with SRI funds are not passed on in the form of management fees, and so this is not a factor in any potential investor performance drag (Gil-Bazo et al. 2010). Performance is nonetheless conditional on asset class, management style, process, geography, and vehicle. The SRI picture, therefore, is still mixed.
2.4. Cultural contexts and the United Kingdom
Scholtens et al. (2007) highlight the major differences in business ethics in different cultural environments and this developed market versus emerging market relativity in CSR must always be understood. CSR studies have been performed all over the globe – on international and national datasets. For example, the CSR-CFP link has been investigated in South Korea (Choi 2010), Greece (Karagiorgos 2010), Canada (Makni et al. 2009) and Japan (Hiraki et al. 2003). A clear bias exists, however, towards U.S. based studies where data quality and availability is the greatest. Bechetti et al. (2005) study the CSP-CFP relationship between a panel of U.S. listed companies on a sample size (100) and time-period similar to our own (13 years). Using Domini Social Index scores they find that strong CSR is shown to significantly increase total sales per employee, but also to reduce ROE for large, research and development intensive firms. They conclude that CSR implies “on the one side, decisions leading to higher cost of labour and intermediate output, but may, on the other side, enhance involvement, motivation and identification of the workforce with company goals with positive effects on productivity”. This verdict was also reached by Greening et al. (2000), who claimed that “firms with strong social performance have an easier time attracting desirable employees”. These papers suggest a ‘longevity’ advantage for socially responsible firms. Inevitably, returns will be prone to reductions in the short term as higher standards are created and maintained. Over the long-term, however, the rewards can be reaped by firms and their investors. Staff turnover, total risk and systematic risk are all found to be lower.
around the world. HM Government is currently tabling plans to link executive pay with that of the average employee and for a regular worker to sit on every board in the FTSE 100. Both have profound implications for corporate governance and evolution of the theory of the firm. Although this paper decouples itself from the U.S., the importance of transatlantic comparisons (Keffas et al. 2011) is duly recognized. Cortez et al. (2011) compares the performance of U.S. and European SRI mutual funds and makes several key findings that are useful to us. Firstly, that funds exhibit a home-bias in stock selection (supporting our choice of a single geography) and, secondly, evidence is found of time-varying betas, but not time-varying alphas (supporting our robustness test of period effects). Aguilera et al. (2006) offer a comparison at the firm-level. Comparing firms and environments in the U.S. and the U.K., they find evidence that CSR differences are primarily driven by corporate governance arrangements in either country, for example board structure, board size, ownership rights, type and distribution. I make a similar case – that in terms of research outcomes, geography has significant implications. The U.K. is the archetypal Anglo-Saxon model of corporate governance and neoliberal macro-economy. Its trademark one-tier board structure leads to faster and more direct decision making and accountability. Under these circumstances, we would expect to see that CSR policies, when implemented, have a more immediate and accentuated effect on the business than may be the case for Germanic systems in continental Europe. This thesis builds on a limited existing U.K. CSR literature body (Brammer et al. 2006) by learning from the greater range of approaches conducted in the U.S. (Gregory et al. 2014).
3. Methodology
This thesis examines the relationships between CSR and value creation – as determined by an accounting-multiple and a market-multiple. Furthermore, it aims to examine the same relationship at a more detailed level than in previous studies, specifically by decomposing CSR into its underlying components (ESG scores) and by decomposing value creation into its underlying components (ROIC, WACC and OGROWTH). By clarifying these relationships, I hope to re-establish the starting point for integrating CSR into corporate valuation, both in terms of discounted cash flow model variants and in the use of multiples analysis. Similarly, my U.K.-specific sample acts as a counterbalance to the U.S. data bias prevalent in the literature.
of environmental, social and governance Asset4 scores – which are already adjusted to reflect strengths and weaknesses. This thesis also employs a greater number of measures for value. Independent variables in the form of ESG pillar scores and control variables for size (natural log of total assets), leverage (ratio of total debt to total assets), total risk (standard deviation), systematic risk (stock beta) and market-to-book, are regressed against rotating dependent variables in the form of the two value multiples and the three core value drivers. With regards to controlling for size, taking the log of total assets is necessary because of the heterogeneity in the size of firms in the dataset. My model specification, therefore, is as follows:
Valuei,t = b1 CSRi,t + b2Controlsi,t + ei,t (1)
Where Value represents the respective multiple or value driver, CSR represents the total, environmental, social or governance score, Controls represents the firm characteristics and e represents the residual error, i.e. the disturbance term to capture the effect of omitted variables in the model. Each fixture in the model is two-dimensional, i.e. represented by indices for firm, i, and time, t. A detailed overview of the model is provided in the Appendix.
The model is estimated using panel OLS. Robustness tests are first, in the form of industry effects (with accompanying sample-split industry breakdowns), and second, in the form of period fixed effects. The use of period fixed effects allows for an exploration of whether time plays a decisive role in the data. Cross-sectional effects are bypassed, as the differences emerging between our sample firms are those which are of most interest. Non-linearity adaptations are not explored by this thesis, although this is an area noted for future research. Only limited academic studies suggest that value creation is maximised at a specific level of CSR, i.e. that it is “curvi-linear” (Barnett et al. 2006; 2012). For the purposes of this research, I assume linearity.
4. Data and Descriptive Statistics
4.1. Sample selection
time at which they occur, provided only that they share the same Industry Classification Benchmark (ICB) classification as the original company. Owned by FTSE Russell, the ICB is a globally recognised standard by investors for classifying equity-related securities. It consists of ten industries, nineteen supersectors, forty-one sectors and one hundred and fourteen subsectors. This structure allows for both top-down and bottom-up investors to analyse markets using their preferred balance of aggregation or granularity (ICB 2016) and allocations are made to closely reflect the primary revenue source and nature of a business (FTSE Russell, 2016). Several of the additions made over time are large firms listed on foreign exchanges – this is accounted for by translating the data where necessary into sterling terms using corresponding monthly foreign exchange rates. Entities that either fail or are taken private are left in – however their effect on the data naturally reduces to zero. The sole exclusion made relates to Royal Dutch Shell, which would otherwise be double counted owing to its dual share classes. No other data filters are used.
This FTSE 100 sample is selected as it offers a representative reflection of the U.K. stock market and contains a sufficiently diverse set of readily investable companies across a range of sectors and sizes. After liquidity screening, index weighting is applied by way of the companies’ free float (trading) values. Accordingly, larger companies have a stronger bearing on index movements than smaller ones. After M&A additions, adjustments, and one exclusion, my final sample consists of monthly data for 107 firms dating from December 2001 to December 2015. This equates to the following numbers of monthly observations for each round of regression testing: 11,794 for EBITDA/TA, 11,728 for EV/EBITDA, 11,794 for ROIC, 11,851 for WACC and 11,848 for OGROWTH. The numbers of observations vary slightly due to missing datapoints in part of the sample. Presented in Table A.II is the sample sectoral distribution on the FTSE 100 for the period. Financial data used to construct the value metrics are sourced from Datastream. ESG score data used to construct the CSP proxies are sourced from Thomson Reuters Asset4. This database has a base date of 2001 and this is reflected in the chosen time-period for this thesis. The ten industry groups we use, and their corresponding numbers, are defined by the ICB (Table A.II).
4.2. CSR variables
impact on living and non-living natural systems. It reflects how well a company uses best management practices to avoid environmental risks and capitalize on environmental opportunities.” The social score measures a company's “capacity to generate trust and loyalty with its workforce, customers and society, through its use of best management practices.” Finally, the corporate governance score measures a company's “systems and processes, which ensure that its board members and executives act in the best interests of its long-term shareholders. It reflects a company's capacity, through its use of best management practices, to direct and control its rights and responsibilities through the creation of incentives, checks and balances.” A full description of the CSR scores and other variables is provided in Table A.IV.
4.3. Value multiples
I use two multiples as proxies for value ~ one that is accounting-based, and one that is market-based. The former is EBITDA/TA, more specifically the pure earnings generated for each pound of total assets. This is a percentage term but can be directly translated into an investor multiple. The latter is EV/EBITDA, which shows the enterprise value (market value of debt and equity) awarded to each pound of earnings before interest, tax, depreciation and amortization. The purity of different earnings measures can essentially be seen in two ways. EBITDA is selected to feature in both because it depicts earnings close to their source (McKinsey et al. 2005) and is intended to improve upon the choice of EBIT/TA by earlier studies (Goss et al. 2010). Earnings before interest and tax (EBIT) or net operating profit less adjusted taxes (NOPLAT) are conventionally ‘purer’ measures of earnings, but are more distanced from the source. Enterprise value (EV) is important for our study as it is used to capture the ‘market effect’ (Hoskisson et al. 1994). It represents the sum of all creditor and shareholder claims on the business, at market value. This ‘market effect’ gives a wider assessment as to the value of intangible assets and also signals the value placed on the firm’s discounted future cash flows by market prices.
4.4. Value drivers
adjustments by Damodaran (2007), such that the invested capital is lagged by one period to the return on capital. WACC is a calculation of a firm’s cost of capital in which each source of financing in the capital structure is proportionally weighted, i.e. equity versus debt. It is typically employed as the discount rate for firm projects and reinvestment, as this is the cost level the firm must overcome to generate profit. Employing WACC allows us to evaluate the cost of debt capital (Oikonomou et al. 2014) and the cost of equity capital (Tanke 2013) simultaneously. Earlier research provides evidence that additional layers of responsibility screening contributes to lower cost (El Ghoul et al. 2011), more secure loan portfolios (Goss et al. 2010) and better creditor relationships. WACC is constructed manually using components from Datastream and KPMG’s historic bi-annual market premia (2016). OGROWTH (i.e. internally derived) growth can be defined in a multitude of ways, the most common of which is revenue growth, which is employed in this thesis and deemed to be the least likely to be manipulated in accounting terms. Free Cash Flow per share and Earnings Per Share growth were trialled and found to be unsuitably volatile. OGROWTH clearly encounters a major challenge in measurement as turnover can dramatically change due to M&A, cyclicality and aggressive expansion – this is kept in mind when drawing conclusions from the analysis. There is evidence, for example, to suggest that firms with strong corporate governance engage in less M&A (Fu et al. 2011).
4.5. Control variables
reflect the severity and positioning of the outliers shown in the respective distributions. For CSR data, anomalies are not found that would justify winsorizing. For leverage, it is intuitive to exclude negative observations (i.e. those where there is net cash on the balance sheet) by setting said values to zero. A full overview of the variables is provided in Table A.IV.
Descriptive Statistics
Table I. presents the descriptive statistics summary for the full sample.
Table I. Full sample descriptive statistics for U.K. FTSE 100 firms, 2001-2015
Table I. reports full sample descriptive statistics for 100 (+7) FTSE 100 listed (FTSE 100 originating) firms. These are categorised into CSR, value creation and firm characteristics. CSR characteristics are determined by Thomson Reuters Asset4 ratings: CSR is the overall, equally-weighted score,
ENVIRONMENTAL SCORE is the environmental pillar score, SOCIAL SCORE is the social pillar score
and GOVERNANCE SCORE is the corporate governance pillar score. Value creation and firm characteristics (controls) are determined by Datastream and Worldscope financial metrics: EBITDA/TA is the multiple of earnings before interest, tax, depreciation and amortization to total assets, EV/EBITDA is the multiple of enterprise value to earnings before interest, taxes, depreciation and amortization, ROIC is the return on invested capital, WACC is the weighted average cost of capital and OGROWTH is the organic growth rate of the firm, i.e. net revenues per share growth. LEVERAGE is the gearing level of the firm, i.e. debt to total capital ratio, MARKET TO BOOK is the ratio of market equity value to book value, SYSTEMATIC RISK is the firm’s beta factor to market, TOTAL RISK is the standard deviation of the firm’s returns and SIZE is the natural logarithm of total assets. (M) denotes a multiple and (%) a percentage. To control for outlier effects, percentile winsorizing is utilised for all dependent variables at a 4% level and on the leverage and total risk control variables at a 2% level.
Mean Median Min. Max. Std. Dev. Obs.
Corporate social responsibility
CSR Total Score (CSR) 85.65 91.00 17.80 98.28 14.05 11,417
Environmental Score (ENV) 80.29 87.81 17.24 97.24 17.19 11,417
Social Score (SOC) 83.10 88.71 11.71 98.78 15.68 11,417
Governance Score (GOV) 77.53 82.99 6.92 97.33 16.77 11,417
We can observe from the outset that the FTSE 100 performs strongly on all aspects of CSR for the period, boasting mean values of 85.65 (CSR), 80.29 (ENV), 83.10 (SOC) and 77.53 (GOV). This is substantiated by the median and maximum values. The standard deviation in the CSR statistics, however, underlines the large range in the data. GOV scores, for example, fall to as low as 6.92 (W. M. Morrison supermarket ca. 2003). Although momentary, this is a level unacceptable to any socially responsible investor and would widely warrant portfolio and/or universe exclusion. In terms of value creation, EBITDA/TA and EV/EBITDA multiples average at 11.38 and 10.78 respectively. This means that, on average, the FTSE 100 companies are generating 11.38 percent of EBITDA on their assets. The same firms exhibit enterprise values, i.e. market-driven values, of 10.78 times earnings.
Month-on-month ROIC ranges from -15.66 to 56.33 percent, whilst WACC is relatively steady at an average of 6.09 percent. It is important to note that the full sample WACC values are likely to be dampened by the inclusion of Financials, and, in reality, can be expected to be slightly higher. Due to the nature of their businesses, conventional WACC calculations are largely redundant for our banking stocks (the single largest group in the index). OGROWTH is the most volatile of the value drivers with a standard deviation of 13.39 percent, which is heavily driven by cyclical stocks. In an average month, a firm in our sample has an organic growth rate of 4.62 percent (unweighted and non-annualised).
average of 0.93 and a median of 0.90. This means that firms move almost in perfect tandem with the broader U.K. equity environment, which we would expect given their collective sizes constitute the great majority of the market. The minimum and maximum values, 0.25 and 2.18, represent Barratt Developments (Basic Materials) and ARM Holdings (Technology). These are stock-specific cases and, whilst Technology firms report higher betas, Barratt is not representative of its industry. In general, Consumer Services, Consumer Goods and Utilities are the most market risk-neutral of our firms and Technology and Healthcare are the riskiest. Total risk (systematic plus idiosyncratic) shows a stable average (7.83 percent), and a peak at approximately double this (15.42 percent). Total assets average at £7.22 billion, of which the largest are characteristically banks (especially leading up to the financial crisis). Due to index selection criteria, the size range is narrow with a standard deviation around the mean of just 0.78 percent.
Table II. Full sample pairwise correlation matrix for U.K. FTSE 100 firms, 2001-2015
Table II. reports the full sample, pairwise correlation matrix for the variables employed in the empirical analysis. These are categorised into CSR, value creation and firm characteristics. CSR characteristics are determined by Thomson Reuters Asset4 ratings: CSR is the overall, equally-weighted score, ENVIRONMENTAL SCORE is the environmental pillar score, SOCIAL SCORE is the social pillar score and GOVERNANCE SCORE is the corporate governance pillar score. Value creation and firm characteristics (controls) are determined by Datastream and Worldscope financial metrics: EBITDA/TA is the multiple of earnings before interest, tax, depreciation and amortization to total assets, EV/EBITDA is the multiple of enterprise value to earnings before interest, taxes, depreciation and amortization, ROIC is the return on invested capital, WACC is the weighted average cost of capital and OGROWTH is the organic growth rate of the firm, i.e. net revenues per share growth. LEVERAGE is the gearing level of the firm, i.e. debt to total capital ratio, MARKET TO BOOK is the ratio of market equity value to book value, SYSTEMATIC RISK is the firm’s beta factor to market,
TOTAL RISK is the standard deviation of the firm’s returns and SIZE is the natural logarithm of total assets.
5. Results and Robustness Tests
5.1. Results
Summaries of the main empirical results are presented in charts I and II below the regression outputs. These correspond with the full sample results for the OLS regressions displayed in the three tables that follow. Table III shows results for the two value multiples, EBITDA/TA and EV/EBITDA. Table IV shows results for ROIC and OGROWTH and, finally, Table IV (cont.) shows the results for WACC. To counteract potential multicollinearity, each CSR variable is tested in isolation for each estimation, resulting in four columns per dependent variable. The standard errors of the estimations are shown in parentheses beneath the beta coefficients.
Table III. CSR and Value Multiples for U.K. FTSE 100 firms, 2001-2015
Table III. reports the OLS regression results for the full sample U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time-period. The results shown display the tested relationship between CSR, firm characteristics and our chosen investor ‘value’ multiples. Four tests are carried out for each, taking ENV, SOC and GOV scores in isolation to account for potential multicollinearity in the full test [1]. Full definitions for each variable can be found in both the earlier descriptive outputs and in the Appendix. The number of observations refers to the total number of monthly datapoints in the (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
The value drivers present a mixed story. ROIC shares the strongest and most positive relationship with CSR, which supports Hypothesis II. Social scores are the strongest CSR component in this regard, such that each point increase is associated with a 10.30 percent increase in ROIC. The adjusted R-squared values supporting these regressions are similarly amongst the strongest, with explanatory power of up to 24.5 percent. Social dimensions of CSR, therefore, contribute to more efficient profit generation relative to the net operating capital and fixed assets on the balance sheet. WACC shares the weakest relationships with CSR, which is in line with the findings of our main reference study (Gregory et al. 2014). Environmental and social scores are both found to increase the cost of capital by 2 percent, implying a cost to socially responsible and sustainable practices. Governance, however, is associated with a 1 percent reduction in the cost of capital. This latter result is intuitive and corroborates the findings of Wu et al. (2013) – better governed firms have better access to funding from equity and credit markets. The statistical significance of the coefficients, coupled with the high R-squared values (44.5–45 percent) reinforce these economic inferences. OGROWTH tests achieve little explanatory power (4.8 percent) but interestingly show that environmentally friendly firms grow at a slower rate (-3.9 percent) than they would otherwise.
Table IV. CSR and Value Drivers for U.K. FTSE 100 firms, 2001-2015
Table IV. reports the OLS regression results for the full sample U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time-period. The results shown display the tested relationship between CSR, firm characteristics and our chosen value drivers. Four tests are carried out for each, taking ENV, SOC and GOV scores in isolation to account for potential multicollinearity in the full test [1]. Full definitions for each variable can be found in both the earlier descriptive outputs and in the Appendix. The number of observations refers to the total number of monthly datapoints in the (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Table IV (Continued). CSR and Value Drivers for U.K. FTSE 100 firms, 2001-2015
Table IV. (continued) reports the OLS regression results for the full sample U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time-period. The results shown display the tested relationship between CSR, firm characteristics and our chosen value drivers. Four tests are carried out for each, taking ENV, SOC and GOV scores in isolation to account for potential multicollinearity in the full test [1]. Full definitions for each variable can be found in both the earlier descriptive outputs and in the Appendix. The number of observations refers to the total number of monthly datapoints in the (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Chart I. Summary of the analysis on CSR and value multiples (U.K. FTSE 100 firms, 2001-2015): ESG impact (%)
Chart I. represents the relationships between CSR aspects and value multiples, as determined by their coefficient values in the regression results (Table III). These values are shown in percentage terms below, and are all statistically significant to the 1% level.
Chart II. Summary of the analysis on CSR and value multiples (U.K. FTSE 100 firms, 2001-2015): ESG impact (%)
Chart II. represents the relationships between CSR aspects and value drivers, as determined by their coefficient values in the regression results (tables IV and IV Continued). These values are shown in percentage terms below, and are all statistically significant to the 1% level, except for WACC and Environmental Scores (statistically insignificant) and WACC and Social Scores (5% significance level).
7.00 -3.00 10.10 2.60 3.60 2.50 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
EBITDA / Total Assets EV / EBITDA
GOV SOC ENV
7.60 2.00 -3.90 10.30 2.00 3.20 5.70 -1.00 2.90 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 ROIC WACC OGROWTH
5.2. Robustness tests
Subsequent to the main empirical results, two robustness tests are conducted in this paper. Firstly, the tests are repeated with industry effects and for each of our ten ICB industry groups (for CSR total score only because of the earlier multicollinearity finding). Secondly, the tests are repeated with the inclusion of period fixed effects (tables A.X, A.XI and A.XI cont.). Visual summaries of the main coefficients from the empirical results are presented in charts III and IV.
Table V. Industry Effects Specification:
CSR and Value Creation for U.K. FTSE 100 firms, 2001-2015
Table V. reports the OLS regression results for the industry effects specification, i.e. the original model with the inclusion of ICB Industry dummy variables. The number of observations refers to the total number of monthly datapoints in the (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
EBITDA/TA EV/EBITDA ROIC WACC OGROWTH
C Intercept 29.486*** -25.086*** 36.508*** 10.666*** 30.206*** (0.904) (1.230) (1.231) (0.219) (1.675) CSR Total Score 0.124*** 0.001 0.148*** -0.003** 0.014 (0.005) (0.007) (0.007) (0.001) (0.009) Leverage 0.055*** 0.068*** 0.073*** -0.028*** -0.048*** (0.003) (0.004) (0.004) (0.001) (0.005) Market to Book 0.238*** 0.087*** 0.426*** -0.002 0.097*** (0.010) (0.014) (0.014) (0.003) (0.019) Systematic Risk 0.417** -0.395 0.877*** 3.598*** 2.530*** (0.183) (0.250) (0.249) (0.045) (0.339) Total Risk -0.527*** 0.577*** -0.234*** 0.033*** -0.045 (0.037) (0.050) (0.051) (0.009) (0.070) Size -4.270*** 3.588*** -6.226*** -0.783*** -3.421*** (0.116) (0.157) (0.158) (0.028) (0.214)
Industry effects Yes Yes Yes Yes Yes
R-Squared 0.414 0.138 0.301 0.477 0.053 Adjusted R-Squared 0.413 0.137 0.300 0.476 0.051 S.E. of regression 7.094 9.597 9.733 1.725 13.253 S.D. dependent variable 9.257 10.329 11.634 2.383 13.608 F-statistic 553.624*** 124.920*** 340.400*** 718.567*** 43.817*** Durbin-Watson statistic 0.078 0.112 0.081 0.050 0.174 Observations 11,794 11,728 11,875 11,851 11,848
in purely statistical terms, our sample (FTSE 100 constituents) is effectively ‘randomly’ drawn from a larger population (the entire UK equity market). Period random effects account for heterogeneity and still capture time invariant variables. Furthermore, as there are no dummy variables or transformations involved (i.e. reduced parameters), degrees of freedom are saved which enhances estimation efficiency (Brooks 2014). However, by conducting Hausman diagnostic tests, it emerges that random effects are unsuitable. This is because the Chi-squared statistics are large and statistically significant, i.e. the composite error term is found to be highly correlated with other explanatory variables. The solution, therefore, is to employ period fixed effects to analyse the role of time in our data. These create new model constants for each month of the dataset and, opposite to the random effects explanation above, remove time invariant variables. The advantage we do gain from using fixed effects is that we can control for potential omitted variables that vary only over time, which is especially useful given the chosen avoidance of cross-sectional fixed effects in the main empirical testing.
Chart III. Summary of the analysis on CSR and value multiples (U.K. FTSE 100 firms 2001-2015): ESG impact with period fixed effects (%)
Chart III. represents the relationships between CSR and value multiples with the inclusion of period fixed effects, as determined by their coefficient values in the regression results (Table A.X). These values are shown in percentage terms below, and are all statistically significant to the 1% level.
Chart IV. Summary of the analysis on CSR and value multiples (U.K. FTSE 100 firms 2001-2015): ESG impact with period fixed effects (%)
Chart IV. represents the relationships between CSR aspects and value drivers, as determined by their coefficient values in the regression results (tables A.XI and A.XI (Continued)). These values are shown in percentage terms below, and are all statistically significant to the 1% level, except for WACC and Environmental Scores (statistically insignificant) and WACC and Social Scores (5% significance level).
7.5 -2.7 10.3 3.2 4 3 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
EBITDA / Total Assets EV / EBITDA
GOV SOC ENV
8.00 2.00 -2.00 10.40 3.00 4.00 6.00 -1.00 7.50 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 ROIC WACC OGROWTH
6. Conclusion
6.1. Concluding remarks
This thesis researches the relationship between CSR and value creation of U.K. listed companies. Controlling for size, risk and valuation, I explore the impact of environmental, social and governance on value multiples (market versus accounting) and underlying value drivers (ROIC, WACC and OGROWTH). Robustness tests are conducted to offer insight into these relationships across industries and across time. In doing so, this thesis makes two key academic contributions. Firstly, it adopts an investor-centric perspective that can offer practical guidance for SRI processes and valuation models. Secondly, it adds to the literature covering CSR in the U.K. equity market – which is all too often overshadowed by the U.S.
The research question, i.e. how does CSR impact on the value creation of U.K. listed firms? is answered by testing three hypotheses. The overall finding is a series of statistically significant, positive relationships between CSR and value creation. The first hypothesis, that governance is the most financially material aspect of CSR, is disproved by the empirical tests and in fact, social factors are found to be the most economically significant. The second hypothesis, that CSR is an exercise best applied to maximising ROIC as opposed to minimising WACC, is strongly supported by the regressions. ROIC shares highly significant positive associations with all aspects of CSP, and under all tested environments and parameters. WACC results are generally muted and in most cases, negative. The final hypothesis is disproved by the empirical analysis. In fact, a market multiple of value (EV/EBITDA) will be more impacted by CSR than its accounting near-equivalent (EBITDA/TA). The results show that statistical significance is far stronger for the accounting value multiple, and CSR coefficients (all positive) are between 1.1 and 10 percent higher than for the market value multiple. Given the notion that socially responsible firms are associate with stronger earnings, this would suggest a potential undervaluation of these firms by the market.
6.2. Practical guidance for investors
6.3. Limitations and future research
I reconstruct the FTSE 100 to accurately reflect the U.K. equity market, as it is a readily tradeable index and recognizable by investors and readers. It is large enough to offer statistical validity, whilst being small enough to allow for manual collection of precise data and careful sense checks where necessary. It is not, however, the full U.K. equity market. The logical extension, therefore, is to expand on this thesis to include the entire U.K. equity market – some 2,286 listed companies. This would give greater statistical and economic meaning to the existing purposes of this CSR research.
Many different proxies for CFP have been employed in empirical literature. The same is true for what I have defined as value creation. My choice of value drivers is reasoned in that they regularly feature in corporate valuation models. Other value drivers exist, however, and it would be interesting to see these used in future studies. Furthermore, it would be interesting to see the study conducted purely based on multiples – whilst I use two (one accounting-based and one market-based (Hoskisson et al. 1994)), many more exist that explain all kinds of performance characteristics in the firm. With its investor-centric viewpoint in mind, this thesis has also made clear why Asset4 weighted ESG pillar scores have been used to proxy for CSR. That said, I also reference scores and ratings that are available from elsewhere – a study that combines multiple providers, and therefore methodologies, could compare their respective accuracies and biases. This would also serve a valuable purpose for investors. Lastly, my control variables cover the major firm characteristics but cannot completely rule out exogenous forces that bear influence over margins and value creation aspects. For this reason, it would be beneficial for them to be expanded on further – especially in terms of proxies for risk (such as cash holdings as a percentage of total assets). In terms of methodology, my greatest assumption is linearity in the CSR-value creation relationship. A variety of econometric approaches should be explored that challenge this, such as quartile sample splits or squared independent variables. Reversing dependent and independent variables could also aid the distinction between correlation and causality that I attempt to make.
7. Appendix
The detailed regressions used in the empirical analysis are listed below in their order of occurrence:
(R1) EBITDA/TAi,t = b0 + b1CSRi,t + b2ENVi,t + b3SOCi,t + b4GOVi,t + b4SIZEi,t + b5LEVi,t +
b6TOTRISKi,t + b7SYSRISKi,t + b8MTBi,t + ei,t
(R2) EV/EBITDAi,t = b0 + b1CSRi,t + b2ENVi,t + b3SOCi,t + b4GOVi,t + b4SIZEi,t + b5LEVi,t +
b6TOTRISKi,t + b7SYSRISKi,t + b8MTBi,t + ei,t
(R3) ROICi,t = b0 + b1CSRi,t + b2ENVi,t + b3SOCi,t + b4GOVi,t + b4SIZEi,t + b5LEVi,t +
b6TOTRISKi,t + b7SYSRISKi,t + b8MTBi,t + ei,t
(R4) WACCi,t = b0 + b1CSRi,t + b2ENVi,t + b3SOCi,t + b4GOVi,t + b4SIZEi,t + b5LEVi,t +
b6TOTRISKi,t + b7SYSRISKi,t + b8MTBi,t + ei,t
(R5) OGROWTHi,t = b0 + b1CSRi,t + b2ENVi,t + b3SOCi,t + b4GOVi,t + b4SIZEi,t + b5LEVi,t +
b6TOTRISKi,t + b7SYSRISKi,t + b8MTBi,t + ei,t
The variables used in the empirical analysis are calculated as follows:
ROIC = !"#$%&'() +(,-.# (012)!"#$%&'() +(,-.#
456 (A1)
where Operating Income is Earnings Before Tax and Interest (EBIT) and T is corporate tax.
For the purposes of this thesis, WACC and OGROWTH are calculated universally as follows:
WACC = 7#8&9:;<'&=7#8& [?7#8& 1 − B +7#8&9:;<'&=:;<'&= ?:;<'&= (A2)
where rDebt = net interest expenses * (1-T); rEquity = rfree + b (rmarket – rfree); and T is
corporate tax. b represents the market beta (systematic risk) of the firm relative to the benchmark (FTSE All-Share).
Table A.I. Eurosif SRI classifications / strategies
Exclusions Exclusion lists are widespread and involve prohibiting investment into select companies
that engage in business sectors considered reprehensible or in breach of legislation, compacts or treaties. For example, firms involved with cluster bomb munitions, tobacco, pornography and gambling are frequently subject to exclusion. These lists vary greatly between investors, many of whom set thresholds for how much revenue (in percentage terms) a firm must derive from these businesses before they are excluded. Norms-based
screening
Norms-based screening is a negative screening process that allows more flexibility than an exclusion list. It draws on a limited number of ESG criteria and screens out companies that clearly violate norms established for their respective sectors.
Best-in-class selection
Best-in-class is a positive screening process. It is a basic method but frequently very effective. Similarly to norms-based screening, it establishes scores for the investment universe, from which the manager then applies a floor. For example, a minimum Asset4 CSR total score rating of 65.
Sustainability themed
Funds may employ two themes in this context: sustainability or ethical. Sustainability is less focussed around a moral basis but more concerned with how companies are proactively improving their sustainability profile. Ethical funds (including religious mandates) typically underperform their sustainable fund counterparts. There is significant overlap between sustainable firm characteristics and ethical firm characteristics, but they are not one and the same.
ESG integration
ESG integration is the most sophisticated form of SRI, and is still not widely employed in its true sense. This involves integrating sustainability factors and convictions into the investment process. Material ESG factors are used to identify additional downside and upside value for companies and integrating these into valuation can have the potential to significantly alter the price target and perception of company value.
Engagement and voting
Engagement involves opening and maintaining dialogues with investee companies in order to identify weaknesses in processes which can be improved upon. Voting (by proxy) is an essential part of active ownership and is generally widespread.
Impact investing
Table A.II. FTSE 100 equity index by Industry Classification Benchmark (ICB)
Table A.II. reports the respective weighting allocation of the FTSE 100 equity index by industry, supersector and sector (the 114 subsectors are not included). The Industry Classification Benchmark (ICB) is purely owned by FTSE International and fully designed to best reflect the FTSE 100 index. Index weighting allocations are shown in terms of market capitalization. This table naturally does not include the additions made to the sample as a result of M&A activity, since these additions are frequently listed on other exchanges and indices.
Index weighting Industry Supersector
13.08% [0] Oil & Gas 0500 Oil & Gas
5.82% [1] Basic Materials 1300 Chemicals
1700 Basic Resources
7.54% [2] Industrials 2300 Construction & Materials
2700 Industrial Goods & Services
20.14% [3] Consumer Goods
3300 Automobiles & Parts 3500 Food & Beverage
3700 Personal & Household Goods
11.44% [4] Health Care 4500 Health Care
10.41% [5] Consumer Services
5300 Retail 5500 Media
5700 Travel & Leisure
Table A.III. Sample constituents
Table A.III. reports the sample constituents and their ICB industries []. This list comprises the firms in the FTSE 100 as of the start date, 26th December 2001. It also lists the seventeen new entities either formed or added as a result of M&A activity (shown in parentheses) below their origins.
[8] 3i Group [3] DSG International [5] Rentokil Initial Plc
[8] Abbey National [5] Electrocomponents Plc [0] Rio Tinto Plc
[8] (Banco Santander) [5] EMI Group Plc [2] Rolls-Royce
[4] Allied Domecq Plc [4] Glaxosmithkline [8] RSA Insurance Group
[5] (Pernod Ricard) [5] (GUS / Experian) [8] Royal Bank of Scotland
[4] Amersham Plc [1] Hanson Plc [0] Royal Dutch Shell
[2] (General Electric) [5] Hays Plc [3] SABMiller Plc
[0] Anglo American [8] HSBC Holdings [9] The Sage Group Plc
[9] Arm Holdings Plc [1] Imperial Chemical [5] J Sainsburys Plc
[4] Associated British [1] (Akzo Nobel) [8] Schroders Plc
[4] Astrazeneca [2] Innogy Holdings Plc [3] Scottish & Newcastle
[8] Aviva [2] (RWE) [3] (Heineken)
[5] BAA Plc [5] Intercontinental Hotels Group [7] Scottish Power Plc
[2] Babcock International [5] British Airways Plc [7] (Iberdrola)
[2] BAE Systems [7] International Power Plc [8] Segro Plc
[1] Balfour Beatty Plc [2] Invensys Plc [7] Severn Trent Plc
[8] Barclays Plc [2] (Schneider Electric) [4] Shire Plc
[1] Barratt Developments [8] Invesco Plc [4] Smith & Nephew Plc
[1] Berkeley Group [5] ITV Plc [2] Smith's Industries
[0] BHP Billiton Plc [5] Kingfisher Plc [7] SSE Plc
[1] BOC Group [8] Land Securities [8] Standard Chartered
[1] (Linde AG) [8] Legal & General Group [5] Tesco Plc
[0] BP Plc [8] Lloyds Banking Group [3] Unilever Plc
[5] Brambles Industries [9] Logica Ltd [7] United Utilities Plc
[5] (Brambles Industries (AT)) [9] (CGI (GIB)) [6] Vodafone Group Plc
[3] British American Tobacco [8] Man Group Plc [2] Wolseley Plc
[8] British Land Company [5] Marks & Spencer [5] WPP Plc
[6] BT Group Plc [5] Mitchells & Butlers [0] Enterprise Oil Plc
[6] Cable & Wireless Comms. [5] WM. Morrison Supermarkets [3] Imperial Brands
[5] Capita Plc [7] National Grid Plc [6] O2 Plc
[5] Carnival Plc [5] Next Plc [6] (Telefonica)
[5] New Carphone Warehouse [8] Northern Rock Plc [5] Ladbrokes Plc
[5] (Dixons Carphone) [8] Old Mutual Plc [8] Alliance & Leicester
[4] Celltech Group [5] Pearson Plc [8] Friends Life
[4] (UCB S.A.) [7] Pennon Group Plc [3] Gallaher Group
[7] Centrica Plc [7] Powergen Plc [3] (Japan Tobacco)
[5] Compass Group Plc [7] (E.ON) [1] Corus Group
[5] Daily Mail & General [8] Prudential Plc [1] (Tata Steel)
Table A.IV. Variable and variable component definitions
Table A.IV. provides the names, codes and descriptions for the variables and variable components used in the study. CSR descriptions are quoted verbatim from the Thomson Reuters Asset4 ESG data glossary and financial descriptions are quoted verbatim from the Worldscope (WS) definitions on Datastream.
Name Code Description
CSR Total Score
A4IR The Equal Weighted Rating reflects a balanced view of a company's performance
in all four areas, economic, environmental, social and corporate governance. Environmental
Score
ENVSCORE The environmental pillar measures a company's impact on living and non-living natural systems, including the air, land and water, as well as complete ecosystems. It reflects how well a company uses best management practices to avoid environmental risks and capitalize on environmental opportunities in order to generate long term shareholder value.
Social Score SOCSCORE The social pillar measures a company's capacity to generate trust and loyalty with
its workforce, customers and society, through its use of best management practices. It is a reflection of the company's reputation and the health of its license to operate, which are key factors in determining its ability to generate long term shareholder value.
Governance
Score CGVSCORE The corporate governance pillar measures a company's systems and processes, which ensure that its board members and executives act in the best interests of its long-term shareholders. It reflects a company's capacity, through its use of best management practices, to direct and control its rights and responsibilities through the creation of incentives, as well as checks and balances in order to generate long term shareholder value.
Return on Invested Capital
WC08376 (Net Income – Bottom Line + ((Interest Expense on Debt - Interest Capitalized) *
(1-Tax Rate))) / Average of Last Year's and Current Year’s (Total Capital + Short Term Debt & Current Portion of Long Term Debt) * 100. This calculation uses restated data for last year’s values where available.
Beta (Systematic Risk)
WC09802 Beta is a measure of market risk which shows the relationship between the volatility
of the stock and the volatility of the market. This coefficient is based on between 23 and 35 consecutive month end price percent changes and their relativity to a local market index. Only monthly data subsequent to the fresh start date is used for companies that have emerged from bankruptcy.
Total Debt % Total Capital (Leverage)
WC08211 (Long Term Debt + Short Term Debt & Current Portion of Long Term Debt) / (Total
Capital + Short Term Debt & Current Portion of Long Term Debt) * 100.
Total Assets WC02999 Total assets represent the sum of total current assets, long term receivables,
investment in unconsolidated subsidiaries, other investments, net property plant and equipment and other assets.
Enterprise Value
WC18100 Market Capitalization at fiscal year end date + Preferred Stock + Minority Interest
Table A.IV. (Continued) Variable and variable component definitions
Table A.IV. (Continued) provides the names, codes and descriptions for the variables and variable components used in the study. CSR descriptions are quoted verbatim from the Thomson Reuters Asset4 ESG data glossary and financial descriptions are quoted verbatim from the Worldscope (WS) definitions on Datastream.
Name Code Description
Market to Book ratio
MTBV The market value of common equity divided by the balance sheet value of common
equity in the company (Worldscope item 03501), at the security level.
EBITDA WC18198 EBITDA represent the earnings of a company before interest expense, income taxes
and depreciation. It is calculated by taking the pre-tax income and adding back interest expense on debt and depreciation, depletion and amortization and subtracting interest capitalized.
Net Revenues WC01001 Net revenues represent gross sales and other operating revenue less discounts,
returns and allowances. It includes but is not restricted to: franchise sales when corresponding costs are available and included in expenses, consulting fees, service income, royalty income when included in revenues by the company, contracts-in-progress income and licensing and franchise fees.
Interest Expense on Debt
WC01251 The service charge for the use of capital before the reduction for interest capitalized.
If interest expense is reported net of interest income, and interest income cannot be found the net figure is shown. It includes but is not restricted to: interest expense on short term debt, interest expense on long term debt and capitalized lease obligations, amortization expense associated with the issuance of debt and similar charges. Interest
Expense Total
WC01075 Interest expense total represents the total amount of interest paid by a bank or other
financial company Interest
Income Total
WC01016 Interest income total represents income received from all earning assets such as
loans and investment securities. Net Interest
Income
WC01076 Net interest income represents the difference between the total interest income and
total interest expense of the bank.
Price P The official closing price. This is the default datatype for all equities. The ‘current’
price on Datastream’s equity programs is the latest price available from the appropriate market in primary units of currency (in the U.K., price is given in pence).
Common Shares Outstanding
WC05301 Common shares outstanding represent the number of shares outstanding at the
company's year-end. It is the difference between issued shares and treasury shares. For companies with more than one type of common/ordinary share, common shares outstanding represents the combined shares adjusted to reflect the par value of the share type identified in field 06005 - Type of Share.
Total Debt WC03255 Total debt represents all interest bearing and capitalized lease obligations. It is the
sum of long and short term debt.
Total Capital WC03998 The total investment in the firm, i.e. the sum of common equity, preferred stock,
Table A.V. CSR and EBITDA/TA for U.K. FTSE 100 firms, 2001-2015
Table A.V. reports the OLS regression results for the full sample and individual (ICB classification) sectors in the U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time period. The results shown display the tested relationship between CSR, firm characteristics and EBITDA / TA. The number of observations refers to the total number of monthly datapoints in each (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Full sample Oil & Gas Basic
Materials
Industrials Consumer
Goods
Health Care Consumer Services
Telecoms Utilities Financials Technology
Table A.VI. CSR and EV/EBITDA for U.K. FTSE 100 firms, 2001-2015
Table A.VI. reports the OLS regression results for the full sample and individual (ICB classification) sectors in the U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time period. The results shown display the tested relationship between CSR, firm characteristics and EV / EBITDA. The number of observations refers to the total number of monthly datapoints in each (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Full sample Oil & Gas Basic
Materials
Industrials Consumer
Goods
Health Care Consumer Services
Telecoms Utilities Financials Technology
Table A.VII. CSR and Return on Invested Capital (ROIC) for U.K. FTSE 100 firms, 2001-2015
Table A.VII. reports the OLS regression results for the full sample and individual (ICB classification) sectors in the U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time period. The results shown display the tested relationship between CSR, firm characteristics and ROIC. The number of observations refers to the total number of monthly datapoints in each (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Full sample Oil & Gas Basic
Materials
Industrials Consumer
Goods
Health Care Consumer Services
Telecoms Utilities Financials Technology
Table A.VIII. CSR and Weighted Average Cost of Capital (WACC) for U.K. FTSE 100 firms, 2001-2015
Table A.VIII. reports the OLS regression results for the full sample and individual (ICB classification) sectors in the U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time period. The results shown display the tested relationship between CSR, firm characteristics and WACC. The number of observations refers to the total number of monthly datapoints in each (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Full sample Oil & Gas Basic
Materials
Industrials Consumer
Goods
Health Care Consumer
Services
Telecoms Utilities Financials Technology
Table A.IX. CSR and Organic Growth (OGROWTH) for U.K. FTSE 100 firms, 2001-2015
Table A.IX. reports the OLS regression results for the full sample and individual (ICB classification) sectors in the U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time period. The results shown display the tested relationship between CSR, firm characteristics and OGROWTH. The number of observations refers to the total number of monthly datapoints in each (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
Full sample Oil & Gas Basic
Materials
Industrials Consumer
Goods
Health Care Consumer Services
Telecoms Utilities Financials Technology
Table A.X. CSR and Value Multiples for U.K. FTSE 100 firms, 2001-2015 (Panel Least Squares: Period Fixed Effects)
Table A.X. reports the OLS regression results for the full sample U.K. FTSE 100 equity index, comprising a total of 107 entities (firms) over the time-period. The results shown display the tested relationship between corporate social responsibility, firm characteristics and investor ‘value’ multiples – and include period fixed effects (i.e. Panel Least Squares method). Four tests are carried out for each, taking ENV, SOC and GOV scores in isolation to account for potential multicollinearity in the full test [1]. Full definitions for each variable can be found in both the earlier descriptive outputs and in the Appendix. The number of observations refers to the total number of monthly datapoints in the (unbalanced) panel. Standard errors are displayed in parentheses beneath their respective coefficients. ***, ** and * denote statistical significance at 0.01, 0.05 and 0.10 levels respectively.
EBITDA/TA EV/EBITDA [1] [2] [3] [4] [1] [2] [3] [4] C Intercept 38.664*** 43.844*** 41.605*** 41.956*** -13.778*** -13.078*** -14.618*** -15.941*** (0.788) (0.762) (0.772) (0.848) (1.004) (0.952) (0.973) (1.054) CSR Total Score 0.146*** 0.006 (0.006) (0.008) Environmental Score 0.075*** -0.027*** (0.005) (0.006) Social Score 0.103*** 0.032*** (0.005) (0.007) Governance Score 0.040*** 0.030*** (0.005) (0.006) Leverage 0.029*** 0.029*** 0.021*** 0.024*** 0.052*** 0.050*** 0.051*** 0.052*** (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Market to Book 0.263*** 0.260*** 0.265*** 0.264*** 0.102*** 0.104*** 0.101*** 0.099*** (0.011) (0.012) (0.012) (0.012) (0.015) (0.015) (0.015) (0.015) Systematic Risk 1.002*** 1.245*** 0.720*** 0.739*** 1.042*** 0.881*** 1.016*** 1.006*** (0.174) (0.179) (0.175) (0.178) (0.223) (0.225) (0.222) (0.222) Total Risk -0.783*** -0.824*** -0.809*** -0.881*** 0.600*** 0.566*** 0.625*** 0.615*** (0.036) (0.037) (0.036) (0.036) (0.046) (0.046) (0.046) (0.045) Size -5.107*** -4.903*** -4.852*** -4.081*** 2.183*** 2.519*** 1.981*** 2.214*** (0.105) (0.114) (0.105) (0.099) (0.134) (0.142) (0.133) (0.123)
Period Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
R-squared 0.293 0.269 0.279 0.259 0.081 0.082 0.083 0.083
Adj. R-squared 0.282 0.258 0.268 0.248 0.067 0.069 0.069 0.069
S.E. of regression 7.843 7.976 7.918 8.026 9.976 9.969 9.966 9.965
F-statistic 27.648*** 24,532*** 25.863*** 23.395*** 5.855*** 5.961*** 5.999*** 6.021***
