Table of Contents
Type: Full Masters Dissertation | Subject: Finance | Level: Masters | Word Count: ~14,000 words
This model Full Masters Dissertation was produced by an Essays UK specialist as reference material for learning purposes only. For support in this field, see our finance dissertation writers.
Overview: This model MSc Finance dissertation examines whether environmental, social and governance (ESG) ratings influence the share price performance of FTSE 100 companies. Using an illustrative panel dataset covering 2015 to 2024, it applies fixed-effects panel regression to test the relationship between composite and pillar-level ESG scores, shareholder returns and firm valuation. All figures, tables and results in this document are illustrative and have been constructed solely as a learning reference; they are not drawn from any live dataset and should not be cited as empirical evidence.
The relationship between corporate sustainability and shareholder value has become one of the most contested questions in modern finance. This dissertation investigates whether environmental, social and governance (ESG) ratings are associated with the share price performance of large, publicly listed UK companies, using the constituents of the FTSE 100 index as the empirical setting. Drawing on stakeholder theory, agency theory, signalling theory and the efficient market hypothesis, the study develops four hypotheses linking a composite ESG rating, and each of its three underlying pillars, to two measures of market performance: annual total shareholder return and Tobin’s Q as a proxy for market valuation.
Adopting a positivist philosophy and a deductive, quantitative design, the study analyses an illustrative balanced panel of eighty-six non-financial FTSE 100 firms observed annually between 2015 and 2024, yielding eight hundred and sixty firm-year observations. Fixed-effects panel regression is employed after formal specification testing, controlling for firm size, leverage, profitability, book-to-market ratio, systematic risk and dividend yield. The analysis finds that the composite ESG rating is positively and significantly associated with firm valuation but has no statistically significant association with contemporaneous annual returns. Decomposing the composite score reveals that the governance pillar carries the strongest and most consistent effect, while the environmental and social pillars are economically small and statistically weak once governance and firm fundamentals are controlled for.
These results suggest that the market capitalises credible governance quality into price levels rather than rewarding sustainability with persistent abnormal returns, a pattern consistent with information being impounded into prices and with agency-cost reasoning. The findings carry implications for investors constructing ESG-tilted portfolios, for corporate managers allocating scarce resources across sustainability initiatives, and for policymakers seeking to standardise ESG disclosure. The study contributes a pillar-level, UK-focused analysis to a literature dominated by composite scores and US samples, while acknowledging limitations arising from ESG rating divergence and the illustrative nature of the data used for this model document.
Over the past two decades, the way in which capital markets evaluate listed companies has broadened well beyond conventional accounting earnings and cash flow. Investors, regulators and civil society increasingly expect firms to demonstrate responsible conduct across environmental, social and governance dimensions, collectively abbreviated as ESG. What began as a niche concern associated with ethical or faith-based investment screens has become a mainstream feature of institutional asset management, embedded in mandates, benchmarks and fiduciary expectations. Trillions of pounds of assets are now managed under strategies that reference ESG criteria in some form, and the providers who score companies on their sustainability conduct have grown into influential gatekeepers whose ratings shape index inclusion, capital allocation and corporate reputation.
The United Kingdom occupies a distinctive position within this movement. London remains one of the world’s deepest equity markets, and the FTSE 100 index, comprising the largest companies by market capitalisation listed on the London Stock Exchange, is both a barometer of domestic economic sentiment and a globally held investment universe. UK policy has been comparatively assertive on sustainability disclosure: the adoption of mandatory climate-related financial disclosures aligned with the recommendations of the Task Force on Climate-related Financial Disclosures, the strengthening of stewardship expectations for asset owners, and the evolution of the corporate governance code have all raised the salience of non-financial performance for premium-listed firms. Against this backdrop, FTSE 100 companies face intense scrutiny of their ESG conduct from a sophisticated and concentrated investor base.
Yet the central question for a finance audience is not whether ESG matters ethically, but whether it matters financially, and specifically whether it is reflected in share prices. If strong ESG performance creates value, that value should eventually be visible in valuations and returns; if it is merely a cost with no offsetting benefit, the opposite should hold; and if the market efficiently prices ESG information as it arrives, any value should be capitalised into price levels without generating persistent abnormal returns. Distinguishing between these possibilities has significant consequences for how investors build portfolios, how managers justify sustainability expenditure, and how regulators design disclosure regimes.
Despite an extensive and rapidly growing body of empirical research, the evidence on the ESG–performance relationship remains strikingly inconsistent. Meta-analyses summarising thousands of studies report that the majority find a non-negative relationship, yet the dispersion of results is wide, and a meaningful minority report negative or insignificant associations. Several factors explain this lack of consensus. First, ESG is measured in many different ways, and the major rating agencies disagree substantially with one another, so that the same company can appear strong under one provider and mediocre under another. Second, studies differ in their outcome variables, mixing accounting measures of profitability, market valuation ratios and realised stock returns, which capture conceptually distinct ideas. Third, much of the literature relies on United States samples, with comparatively less attention to the UK market despite its regulatory leadership on sustainability.
A further unresolved issue concerns the level of aggregation. The great majority of studies use a single composite ESG score, which obscures the possibility that the three pillars operate through different channels and with different strengths. Governance quality, for example, bears directly on the agency conflicts that finance theory identifies as a core driver of value, whereas environmental and social performance may influence value more diffusely through reputation, risk mitigation and the cost of capital. Treating ESG as a monolithic construct risks averaging out effects that point in opposite directions. The research problem addressed here, therefore, is the absence of clear, pillar-level evidence on how ESG ratings relate to the share price performance of large UK firms, within a framework that distinguishes valuation levels from realised returns.
The overarching aim of this dissertation is to evaluate the extent to which ESG ratings are associated with the share price performance of FTSE 100 companies over the period 2015 to 2024. To achieve this aim, the study pursues the following objectives:
Consistent with these objectives, the study is organised around one principal research question and three subsidiary questions:
Drawing on the theoretical framework developed in Chapter 2, the study tests four directional hypotheses. Each is stated in the alternative form, with the corresponding null hypothesis being that of no association.
The study makes several contributions. Empirically, it offers a UK-focused, pillar-level analysis of a question that has been studied predominantly through composite scores and US data, thereby addressing a specific gap in the literature. Methodologically, it adopts a longitudinal panel design with firm fixed effects, which controls for unobserved, time-invariant firm characteristics that cross-sectional studies cannot address and that may otherwise bias estimates of the ESG effect. Practically, by distinguishing valuation levels from realised returns and by separating the pillars, the study provides more actionable guidance than a single headline coefficient: investors can better understand whether ESG tilts are likely to generate outperformance or simply reflect already-priced quality, and managers can prioritise the dimensions most closely associated with value. The study also engages directly with the problem of rating divergence, treating it not as a nuisance but as a substantive feature of the ESG landscape with implications for interpretation.
It is important to delimit the study clearly. The investigation is confined to the constituents of the FTSE 100 index over the decade from 2015 to 2024, and to the non-financial firms within it; financial institutions are excluded because their capital structures and regulatory requirements render measures such as leverage and Tobin’s Q non-comparable. The study examines the association between ESG ratings and two market-based outcomes—valuation and realised return—rather than accounting profitability, operating efficiency or broader measures of social impact, which lie beyond its scope. It relies on the ESG ratings of a single provider, both for internal consistency and because the use of multiple providers would introduce the divergence problem into the dependent construct; the consequences of this choice are addressed in the discussion. The analysis is quantitative and does not seek to capture the qualitative, firm-specific narratives behind particular ratings. Finally, and most importantly for a model document, the data employed are illustrative: they are constructed to be realistic and internally consistent for demonstration purposes and do not represent the actual reported figures of any named company. These delimitations define the boundaries within which the findings should be read and generalised.
The remainder of the dissertation is organised into five chapters. Chapter 2 reviews the conceptual foundations of ESG, the competing theoretical perspectives on the sustainability–value relationship and the relevant empirical evidence, culminating in the identification of the research gap and the development of a conceptual framework. Chapter 3 sets out the research methodology, including the philosophical position, research design, sample, data, variable definitions and econometric specification. Chapter 4 presents the empirical findings, beginning with descriptive statistics and correlations before reporting the panel regression results and robustness checks. Chapter 5 discusses the findings in relation to theory and prior research and draws out their implications. Chapter 6 concludes, summarising the contribution, acknowledging limitations and recommending avenues for future research.
This chapter critically examines the body of knowledge relevant to the relationship between ESG ratings and share price performance. It begins by clarifying the concept of ESG and the manner in which it is measured, since conceptual and measurement ambiguity underlie much of the inconsistency in the empirical record. It then develops the theoretical framework, drawing on the long-standing debate between shareholder and stakeholder conceptions of the firm and incorporating agency, signalling and efficient-markets perspectives that bear directly on how sustainability might be reflected in prices. The chapter proceeds to synthesise the empirical evidence under three themes—ESG and financial performance, the construction and divergence of ESG ratings, and ESG in valuation and returns—before situating the discussion in the UK and FTSE 100 context. It closes by articulating the research gap and presenting the conceptual framework and hypotheses that guide the empirical analysis.
The term ESG denotes a set of non-financial criteria used to assess the sustainability and ethical conduct of an organisation. The environmental dimension captures a firm’s interaction with the natural world, encompassing greenhouse gas emissions, energy and resource efficiency, waste and pollution, and exposure to climate-related transition and physical risk. The social dimension concerns relationships with employees, customers, suppliers and communities, including labour standards, health and safety, diversity, product responsibility and human rights across the supply chain. The governance dimension addresses the structures and processes by which a firm is directed and controlled, including board composition and independence, executive remuneration, shareholder rights, audit quality and the management of conflicts of interest.
ESG is conceptually related to, but distinct from, earlier notions such as corporate social responsibility and socially responsible investment. Whereas corporate social responsibility is often framed as a normative or voluntary commitment, ESG has evolved into a more systematic, data-driven and investor-oriented construct, operationalised through scores and ratings intended to be comparable across firms and over time. This operationalisation is both the source of ESG’s practical influence and the origin of many of its measurement problems, since the translation of complex, multidimensional conduct into a single numerical score inevitably involves subjective judgements about scope, weighting and materiality. As Gillan, Koch and Starks (2021) observe in their review of the corporate finance literature, the proliferation of ESG metrics has outpaced consensus on what they actually capture, leaving researchers to grapple with constructs that are measured differently across studies.
The question of whether ESG performance should be associated with shareholder value sits at the intersection of several theoretical traditions, which offer competing and sometimes complementary predictions.
The classical position, most famously articulated by Friedman (1970), holds that the social responsibility of business is to increase its profits within the rules of the game, and that managers who divert corporate resources to social causes are in effect spending other people’s money. From this perspective, expenditure on ESG initiatives that does not directly enhance cash flow represents an agency cost, a misallocation that should, all else equal, reduce shareholder value. A strict reading of shareholder theory therefore predicts a negative or, at best, neutral relationship between ESG ratings and performance, insofar as high ratings reflect discretionary spending beyond what value maximisation requires.
The principal counterweight is stakeholder theory, developed by Freeman (1984), which argues that a firm’s long-run success depends on its ability to create value for all parties with a stake in its activities—employees, customers, suppliers, communities and the environment—not only its shareholders. On this view, attending to stakeholder interests is not a diversion from value creation but a precondition for it: satisfied employees are more productive, loyal customers generate repeat revenue, reliable suppliers reduce operational risk, and a firm trusted by its community enjoys a stronger licence to operate. Stakeholder theory thus predicts a positive relationship between genuine ESG performance and value, mediated through improved operating performance and reduced risk.
Agency theory, as formalised by Jensen and Meckling (1976), concerns the conflicts of interest that arise when ownership and control are separated, and the costs incurred in monitoring and bonding management to act in shareholders’ interests. This framework is especially relevant to the governance pillar. Strong governance—independent boards, aligned remuneration, robust audit and transparent reporting—reduces agency costs by constraining managerial self-interest, limiting value-destroying empire-building and improving the quality of capital allocation. Agency theory therefore predicts that the governance dimension of ESG should be the most directly and reliably linked to value, because it operates on the very mechanism that finance theory identifies as central to the principal–agent problem. It also offers a subtler prediction for the environmental and social pillars: where these reflect managerial private benefits rather than value creation, they may signal weak rather than strong governance.
Signalling theory, originating with Spence (1973), explains how parties with superior information convey credible information to less-informed parties through costly, observable actions. Applied to ESG, voluntary and verified sustainability disclosure can function as a signal of managerial quality, long-term orientation and lower risk, reducing information asymmetry between the firm and the market. Because high-quality firms can bear the cost of credible disclosure more easily than low-quality firms, strong ESG ratings may be interpreted by investors as a signal that lowers the required rate of return and raises valuation. Signalling theory thus predicts a positive relationship, but one that operates through the cost of capital and valuation rather than necessarily through realised abnormal returns.
Finally, the efficient market hypothesis, articulated by Fama (1970), provides a crucial lens for interpreting any observed relationship. In a semi-strong efficient market, all publicly available information—including ESG ratings and sustainability disclosures—is rapidly impounded into prices. The implication is profound: even if ESG performance creates real value, that value should be reflected in price levels and valuation ratios rather than in persistent future abnormal returns, because rational investors will have already bid prices up to reflect the information. On this view, a positive association with valuation is entirely consistent with market efficiency, whereas a positive association with future risk-adjusted returns would be harder to reconcile with it and might indicate either a priced risk factor or a market inefficiency. This distinction between valuation levels and realised returns is central to the interpretation of the present study’s findings.
Two further perspectives complement the core framework. Legitimacy theory holds that organisations seek to ensure that their activities are perceived as appropriate within the norms and values of the society in which they operate, and that they use voluntary disclosure, including sustainability reporting, to maintain or repair this social licence to operate. On this account, ESG conduct and its disclosure are instruments by which firms manage their legitimacy, and the market rewards credible legitimacy because firms that lose it face regulatory, reputational and operational costs. Legitimacy theory thus reinforces the stakeholder prediction of a positive relationship, while warning that disclosure may at times be symbolic rather than substantive—a caveat directly relevant to the greenwashing critique considered later in this chapter.
The resource-based view offers a parallel rationale rooted in competitive strategy. It contends that sustainable competitive advantage derives from resources and capabilities that are valuable, rare, difficult to imitate and well organised. Strong ESG capabilities—an engaged workforce, a trusted brand, resilient supply chains and sophisticated risk management—can constitute precisely such intangible resources, generating rents that accrue to shareholders over time. From this standpoint, ESG performance is not a cost to be minimised but a capability to be cultivated, and its value should be visible in the firm’s market valuation as the capitalised worth of those intangible assets. Together, legitimacy theory and the resource-based view enrich the framework by linking ESG to both the preservation of the firm’s operating environment and the accumulation of value-generating intangibles, further motivating the expectation of a positive valuation effect.
The empirical literature linking ESG to financial performance is vast and, taken as a whole, points towards a predominantly positive but far from uniform relationship. The most frequently cited synthesis is the meta-study by Friede, Busch and Bassen (2015), which aggregates the findings of more than two thousand empirical studies and reports that approximately ninety per cent find a non-negative relationship between ESG and corporate financial performance, with a clear majority reporting positive results. This finding is often invoked as settling the debate in favour of a business case for sustainability, yet the authors themselves caution that the relationship varies by region, asset class, outcome measure and the way ESG is operationalised. The headline figure conceals substantial heterogeneity and should not be read as implying that higher ESG reliably produces higher returns in every setting.
A particularly influential refinement is the materiality perspective advanced by Khan, Serafeim and Yoon (2016), who distinguish between sustainability issues that are financially material to a given industry and those that are not. Using an industry-specific materiality classification, they find that firms performing well on material sustainability issues outperform, while performance on immaterial issues bears little relation to returns and may even detract from value by consuming resources. This result is important because it reframes the debate: the question is not whether ESG matters in general, but which ESG issues matter for which firms. It implies that composite scores, which blend material and immaterial factors, may understate the true relationship for the factors that matter and overstate it for those that do not.
Evidence on specific channels lends further nuance. Edmans (2011) demonstrates that employee satisfaction, an element of the social pillar, is associated with superior long-run stock returns, interpreting this as evidence that intangible human-capital value is not immediately recognised by the market. On the environmental side, El Ghoul and colleagues (2011) find that firms with stronger social and environmental conduct enjoy a lower cost of equity capital, consistent with the signalling and risk-reduction channels, while Chava (2014) shows that firms with environmental concerns face higher costs of both debt and equity, implying that investors demand compensation for environmental risk exposure. Pedersen, Fitzgibbons and Pomorski (2021) provide a theoretical and empirical framework—the ESG-efficient frontier—showing how the relationship between ESG and returns depends on investor preferences and the extent to which ESG information predicts fundamentals, reconciling seemingly contradictory findings within a single model.
Taken together, this body of work supports the view that ESG can be associated with value, but that the association is conditional, channel-specific and sensitive to how performance is measured. It also exposes a persistent tension between studies that document a relationship with valuation or the cost of capital and those that search, with more mixed success, for a relationship with realised abnormal returns—precisely the tension that the efficient market hypothesis would lead one to expect.
A particularly robust finding across this literature concerns risk rather than return. A substantial number of studies report that firms with strong ESG profiles experience lower return volatility, smaller maximum drawdowns and a reduced frequency of severe negative events such as environmental disasters, product-safety scandals, regulatory sanctions and governance failures. The intuition is that good ESG management functions, in part, as a form of operational and reputational risk control: firms that monitor their environmental footprint are less exposed to pollution liabilities and transition risk; firms that treat employees and customers well are less prone to strikes, boycotts and litigation; and firms with strong governance are less likely to suffer the fraud and misallocation that destroy value abruptly. This downside-protection channel implies that ESG may matter most not in the average, calm year but in the tails, during crises and shocks when poorly governed, high-risk firms are most likely to suffer catastrophic losses.
This risk framing has important implications for the present study. If the primary financial benefit of ESG is the reduction of tail risk rather than the enhancement of average returns, then studies focused on mean returns in ordinary conditions may understate the value of ESG, because the benefit manifests as the avoidance of rare, large losses that are difficult to capture in annual return regressions. It also helps explain why a positive valuation effect can coexist with a null average-return effect: the market may award high-ESG firms a lower cost of capital and hence a higher valuation precisely because they are perceived as less risky, even though this does not translate into higher expected returns—indeed, lower risk should, all else equal, be associated with lower expected returns. This reasoning strengthens the case for examining valuation and returns separately, as the present study does.
Any study that relies on ESG ratings must confront the fact that these ratings are constructed rather than observed, and that different providers construct them very differently. Rating agencies differ in the scope of indicators they consider, the way they measure each indicator, the weights they attach to indicators in aggregating them, and their treatment of missing data and firm size. The consequence is substantial disagreement across providers. Berg, Kölbel and Rigobon (2022) document this phenomenon systematically, reporting an average correlation of approximately 0.54 among the ratings of six major providers—far below the near-unity correlation observed among credit ratings. They decompose this divergence into scope, measurement and weight contributions, finding that measurement divergence, where providers disagree on the value of the same attribute, is the largest single source.
This divergence has serious implications for research and practice. For researchers, it means that empirical results may be sensitive to the choice of rating provider, so that two studies using different providers could reach different conclusions about the same firms. For investors, it undermines the comparability that ratings are supposed to provide and complicates the construction of ESG-tilted portfolios. For firms, it blunts the incentive effect of ratings, since improving one’s standing with one provider may have little effect on another. Critics have further argued that ESG ratings can be backward-looking, biased towards larger firms with the resources to disclose extensively, and vulnerable to a form of greenwashing in which disclosure volume is rewarded over substantive performance. These critiques do not negate the usefulness of ratings, but they do require that any single-provider study be interpreted with appropriate caution and that the robustness of findings to the choice of measure be considered. The present study acknowledges this limitation explicitly and treats rating divergence as a substantive feature of the research context rather than mere noise.
A distinct strand of the literature focuses specifically on how ESG enters equity valuation and return generation. Giese and colleagues (2019) propose three transmission channels through which ESG characteristics affect valuation and risk: a cash-flow channel, in which high-ESG firms are more competitive and generate higher profitability and dividends; an idiosyncratic-risk channel, in which such firms experience fewer severe incidents and lower tail risk; and a systematic-risk channel, in which they exhibit lower exposure to systematic risk factors and hence a lower cost of capital. The valuation and systematic-risk channels imply that ESG should be reflected in price levels and valuation multiples, while the cash-flow channel implies effects on fundamentals that the market may or may not anticipate fully.
The empirical results on returns are correspondingly mixed. Some studies document positive risk-adjusted returns to high-ESG portfolios, particularly where the ESG signal captures information not yet reflected in fundamentals, consistent with Edmans (2011). Others find that once ESG information becomes widely available and priced, any return premium dissipates, consistent with market efficiency. A recurring theme is that strong ESG performance is associated with lower risk—smaller drawdowns, fewer controversies and lower volatility—more robustly than with higher returns, suggesting that the primary financial benefit may be risk mitigation rather than return enhancement. This distinction motivates the present study’s use of two separate outcome variables: a valuation measure, which should reflect capitalised value, and a realised return measure, which under efficiency should show little association with publicly available ESG ratings.
It would be incomplete to review this literature without acknowledging the growing critique of ESG investing, which has intensified as the field has matured. Sceptics advance several arguments. The first, echoing Friedman (1970), is that ESG mandates risk diverting managerial attention and corporate resources away from value maximisation towards objectives that are properly the province of governments and individuals, and that fiduciaries who sacrifice return for social goals may breach their duties to beneficiaries. The second is empirical: critics contend that much of the apparent outperformance of ESG strategies reflects incidental sector and style tilts—an underweighting of energy and an overweighting of large, profitable, low-volatility growth stocks—rather than any causal effect of sustainability itself, so that controlling for these factors, as the present study does, substantially weakens the measured relationship.
A third strand concerns greenwashing and the reliability of the ratings themselves. If firms can improve their scores through disclosure volume and presentational sophistication rather than substantive change, and if ratings diverge as sharply as the evidence suggests, then the informational content of ESG scores is diluted and the incentives they create may be distorted. These critiques do not overturn the balance of evidence, which remains predominantly favourable, but they impose discipline on interpretation: they warn against attributing causation too readily, they emphasise the importance of controlling for conventional risk factors, and they caution that a single rating is a noisy proxy for the underlying construct. The present study takes these warnings seriously by controlling extensively for firm fundamentals and risk, by distinguishing valuation from returns, and by treating its single-provider ratings as an imperfect measure whose limitations are made explicit.
Most of the evidence reviewed above derives from United States samples, yet the UK context differs in ways that may affect the ESG–value relationship. The UK operates a comply-or-explain corporate governance regime that places strong emphasis on board independence, the separation of chair and chief executive roles, and shareholder engagement, so that governance quality among premium-listed FTSE 100 firms is relatively high and relatively homogeneous by international standards. Ownership is dominated by institutional investors, including large asset managers and pension funds with stewardship obligations, who are well placed to price ESG information and to engage with companies on it. The regulatory environment has moved early towards mandatory climate-related disclosure, raising the quality and comparability of environmental information for large firms.
These features suggest two competing possibilities. On one hand, the sophistication and concentration of UK institutional ownership imply an efficient market in which ESG information is rapidly priced, so that any value is capitalised into levels rather than returns. On the other hand, the relative homogeneity of governance among FTSE 100 firms may compress the cross-sectional variation in governance scores, making it harder to detect an effect. The social and environmental pillars, by contrast, may exhibit greater variation across the diverse sectors represented in the index—from extractives and energy to consumer goods, pharmaceuticals and financial services—offering more scope to detect associations. The FTSE 100 therefore constitutes a theoretically interesting setting in which the governance pillar may be well priced but narrowly dispersed, while the environmental and social pillars vary more widely.
Three gaps emerge from this review. First, the literature is dominated by composite ESG scores, with comparatively few studies decomposing the relationship by pillar in a way that connects each pillar to a distinct theoretical channel; yet theory strongly suggests that governance should operate differently from environmental and social performance. Second, the UK market is under-represented relative to the United States, despite its regulatory leadership and distinctive ownership structure, leaving open the question of whether US-based findings generalise to FTSE 100 firms. Third, studies frequently conflate valuation and returns, reporting a single relationship without distinguishing whether ESG is capitalised into price levels or generates abnormal returns—a distinction that is central to interpreting results through the lens of market efficiency. This dissertation addresses all three gaps by conducting a pillar-level analysis of FTSE 100 firms using two separate outcome measures within a panel framework.
Synthesising the theoretical perspectives yields a conceptual framework in which ESG ratings influence share price performance through valuation and risk channels, moderated by market efficiency and conditioned on the pillar in question. The composite ESG rating is expected to be positively associated with valuation, reflecting the capitalisation of stakeholder value and signalling benefits, but only weakly associated with realised returns, because publicly available ratings should already be priced in an efficient market dominated by sophisticated investors. This reasoning supports H1 while anticipating that the effect will be stronger for valuation than for returns.
At the pillar level, agency theory predicts that governance should exhibit the strongest and most reliable association with value, because it operates directly on the principal–agent mechanism; this underpins H4. The environmental and social pillars are expected to show positive but weaker associations, operating more diffusely through reputation, risk mitigation and the cost of capital, and potentially attenuated by the materiality considerations of Khan, Serafeim and Yoon (2016) and by the homogeneity of large-firm disclosure; this underpins H2 and H3 while anticipating weaker effects. The framework thus predicts a hierarchy of effects—governance strongest, composite intermediate, environmental and social weakest—and a divergence between valuation and return outcomes consistent with market efficiency.
The framework can be visualised as a sequence of channels connecting ESG conduct to market outcomes. Genuine ESG performance improves fundamentals (through the cash-flow and resource-based channels), reduces risk (through the idiosyncratic- and systematic-risk channels), and signals quality (through the signalling and legitimacy channels). Each channel raises the intrinsic value of the firm. An efficient market, observing the publicly disclosed ratings, capitalises this higher intrinsic value into the share price, raising the valuation multiple. Because the repricing is complete and rapid, no persistent abnormal return follows. The strength of each channel, however, varies by pillar: governance acts most directly on the agency-cost determinants of value and is measured most objectively, so its effect is strongest and most reliably detected; environmental and social performance act more diffusely, are conditioned on materiality, and are measured with more divergence, so their effects are weaker and noisier. This chain of reasoning yields the testable predictions—a significant valuation effect, an insignificant return effect, and a governance-dominated pillar hierarchy—that the empirical chapters evaluate.
This chapter has established that ESG is a multidimensional, imperfectly measured construct whose relationship with shareholder value is theoretically contested and empirically inconsistent. Shareholder theory predicts a negative or neutral relationship, while stakeholder, signalling and agency perspectives predict positive relationships operating through distinct channels, and the efficient market hypothesis cautions that any value should appear in prices rather than in persistent returns. The empirical record supports a predominantly positive but conditional relationship, sensitive to materiality, measurement and the choice of outcome. Rating divergence emerges as a significant complication. The identified gaps—the dominance of composite scores, the under-representation of the UK market, and the conflation of valuation and returns—motivate the pillar-level, UK-focused, dual-outcome design pursued in the remainder of this dissertation.
This chapter sets out the methodological choices that operationalise the research aim and enable the hypotheses to be tested. It explains the philosophical position underpinning the study, the research approach and design, the definition of the population and sample, the sources and nature of the data, the specification of the dependent, independent and control variables, and the econometric models and diagnostic tests employed. It concludes by addressing ethical considerations and the limitations inherent in the chosen approach. Throughout, the chapter emphasises the alignment between the research question and the methods selected, following the principle that method should follow from purpose rather than the reverse.
The study adopts a positivist research philosophy. Positivism holds that social phenomena, including the behaviour of financial markets, can be studied through objective, observable and measurable evidence, and that the researcher stands apart from the object of study. This position is appropriate here because the research question concerns a measurable association between quantified ESG ratings and quantified market outcomes, amenable to statistical testing and in principle replicable by other researchers. The alternative interpretivist philosophy, which emphasises subjective meaning and context, would be ill-suited to a question that is fundamentally about numerical relationships in large samples. The ontological assumption is that the relationship between ESG and performance exists independently of the researcher’s perception of it, and the epistemological assumption is that it can be known through empirical observation and hypothesis testing.
Consistent with positivism, the study follows a deductive approach. It begins with established theory—stakeholder, agency, signalling and efficient-markets perspectives—from which specific, testable hypotheses are derived, and it then collects and analyses data to test those hypotheses. This contrasts with an inductive approach, which would build theory from observation. The deductive approach is appropriate because a substantial theoretical and empirical literature already exists, providing clear predictions that the present study can confirm, refine or refute. The hypotheses articulated in Chapter 1 and developed in Chapter 2 represent the deductive propositions under test.
The study employs a quantitative, longitudinal panel design. A quantitative design is dictated by the nature of the variables and hypotheses, which concern statistical associations among measured quantities. A longitudinal panel design—observing the same firms repeatedly over several years—is chosen in preference to a cross-sectional design for two reasons. First, it greatly increases the number of observations and hence the statistical power available to detect associations. Second, and more importantly, it permits the use of firm fixed effects, which control for unobserved, time-invariant characteristics of each firm—such as industry, culture, business model or managerial philosophy—that might otherwise confound the relationship between ESG and performance. Because such omitted characteristics are a central threat to inference in cross-sectional ESG studies, the panel design offers a substantial improvement in internal validity. A secondary data strategy is adopted, drawing on established ratings and market databases rather than primary data collection, which is appropriate given that the required variables are routinely compiled by specialist providers.
Several alternative quantitative designs were considered and rejected. An event-study design, which measures abnormal returns around specific ESG-related announcements such as index inclusions or controversy disclosures, would offer sharper causal identification for discrete events but could not address the broader question of how the level of a firm’s ESG rating relates to its valuation and returns over time, which is the focus of this study. A portfolio-sorting design, which forms portfolios on ESG scores and compares their risk-adjusted returns, is well suited to detecting a return premium but is less able to control for firm-level fundamentals or to decompose effects by pillar at the firm level, and it discards the cross-sectional and time-series variation that the panel exploits. A purely cross-sectional regression, finally, would be unable to control for unobserved time-invariant firm heterogeneity and would therefore be vulnerable to the omitted-variable bias that is the principal threat to inference in this setting. The longitudinal panel design with firm and year fixed effects was selected because it best matches the research question, maximises the use of the available variation, permits the pillar decomposition at the firm level, and most effectively addresses the central identification concern.
The population of interest is the set of large, premium-listed UK companies, operationalised as the constituents of the FTSE 100 index. Because index membership changes over time as companies are promoted and relegated, the study uses the constituents as at the start of the sample period and tracks them over the full window, supplemented by firms that remained in the index for the majority of the period, in order to mitigate survivorship bias while maintaining a stable panel. Financial firms—banks, insurers and other institutions—are excluded, following standard practice in corporate finance, because their capital structures, accounting conventions and regulatory capital requirements differ fundamentally from those of non-financial firms and would distort measures such as leverage and Tobin’s Q. Firms with insufficient ESG or financial data are also excluded.
These criteria yield an illustrative final sample of eighty-six non-financial firms observed annually over the ten years from 2015 to 2024, producing eight hundred and sixty firm-year observations in a balanced panel. The composition of the sample by broad sector is summarised in Table 1. The sample spans the principal sectors represented in the index, providing cross-sectional variation in ESG performance while holding constant the broad characteristics of large, mature, UK-listed firms.
| Sector | Number of Firms | Firm-Year Observations | Share of Sample |
|---|---|---|---|
| Consumer goods and retail | 20 | 200 | 23.3% |
| Industrials | 15 | 150 | 17.4% |
| Energy and utilities | 12 | 120 | 14.0% |
| Healthcare and pharmaceuticals | 11 | 110 | 12.8% |
| Basic materials | 10 | 100 | 11.6% |
| Telecommunications and technology | 9 | 90 | 10.5% |
| Consumer services | 9 | 90 | 10.5% |
| Total | 86 | 860 | 100% |
The study draws on secondary data from two broad categories of source. ESG ratings—comprising a composite score and separate environmental, social and governance pillar scores, each expressed on a nought-to-one-hundred scale—are of the kind compiled by established providers of sustainability data, in which higher scores denote stronger assessed performance. Market and accounting data—share prices, dividends, market capitalisation, total assets, total debt, net income and book value of equity—are of the kind available from standard financial databases. For the purpose of this model dissertation, the specific values used in the analysis are illustrative and have been constructed to be internally consistent and broadly plausible for a sample of large UK firms; they are not drawn from any live data feed and are presented solely to demonstrate the method and the form of the expected results. A single rating provider is used throughout for consistency, with the implications of rating divergence addressed in the discussion and limitations.
The study employs two dependent variables, four explanatory ESG variables and six control variables, defined in Table 2. Two dependent variables are used in order to distinguish, as the literature review motivated, between realised return and capitalised valuation. The first is annual total shareholder return, computed as the sum of the capital gain and dividend yield over the financial year, which captures realised performance from the perspective of an investor. The second is Tobin’s Q, defined as the market value of equity plus the book value of debt divided by total assets, which captures the market’s valuation of the firm relative to the replacement cost of its assets and is widely used as a forward-looking valuation measure. The explanatory variables are the composite ESG score and its three pillar scores. The control variables are standard determinants of return and valuation drawn from the asset-pricing and corporate finance literatures.
| Variable | Symbol | Definition and Measurement |
|---|---|---|
| Total shareholder return | TSR | Annual capital gain plus dividend yield, in per cent (dependent variable 1). |
| Tobin’s Q | TQ | (Market value of equity + book value of debt) ÷ total assets (dependent variable 2). |
| Composite ESG score | ESG | Overall ESG rating on a 0–100 scale. |
| Environmental pillar | ENV | Environmental pillar score, 0–100. |
| Social pillar | SOC | Social pillar score, 0–100. |
| Governance pillar | GOV | Governance pillar score, 0–100. |
| Firm size | SIZE | Natural logarithm of total assets. |
| Leverage | LEV | Total debt ÷ total assets. |
| Profitability | ROA | Net income ÷ total assets. |
| Book-to-market | BM | Book value of equity ÷ market value of equity. |
| Systematic risk | BETA | Capital asset pricing model beta estimated on monthly returns. |
| Dividend yield | DY | Dividend per share ÷ share price, in per cent. |
The hypotheses are tested using panel regression. The baseline models relate each dependent variable to the composite ESG score and the control variables. For firm i in year t, the valuation model is specified as:
TQit = α + β1ESGit + β2SIZEit + β3LEVit + β4ROAit + β5BMit + β6BETAit + β7DYit + μi + λt + εit
The return model replaces the dependent variable with annual total shareholder return and retains the same right-hand side. In these equations μi denotes firm fixed effects, which absorb unobserved time-invariant firm characteristics; λt denotes year fixed effects, which absorb market-wide shocks common to all firms in a given year, such as macroeconomic cycles or market-wide sentiment; and εit is the idiosyncratic error term. To test the pillar-level hypotheses, a second set of models replaces the composite score with the three pillar scores entered simultaneously, allowing the marginal contribution of each pillar to be assessed while holding the others constant.
Three estimators are considered: pooled ordinary least squares, the random-effects estimator and the fixed-effects estimator. Pooled ordinary least squares ignores the panel structure and is reported only as a benchmark. The choice between random and fixed effects is determined by the Hausman test, which examines whether the firm-specific effects are correlated with the regressors; a rejection of the null favours the fixed-effects estimator. Given the strong theoretical likelihood that unobserved firm characteristics are correlated with both ESG ratings and performance, fixed effects are anticipated to be preferred, and the discussion of results focuses on that specification.
Prior to estimation, the data are screened for completeness and continuous variables are winsorised at the first and ninety-ninth percentiles to limit the influence of extreme outliers. Several diagnostic tests support valid inference. Multicollinearity is assessed using variance inflation factors, with values below the conventional threshold of ten taken to indicate that collinearity is not a serious concern; this is particularly relevant to the pillar-level models, since the three pillars are expected to be positively correlated. Heteroskedasticity and within-firm serial correlation are addressed by reporting standard errors clustered at the firm level, which are robust to both. The Hausman test guides the choice of estimator, and the robustness of the findings is examined through the alternative specifications described in Chapter 4. Statistical significance is assessed at the one, five and ten per cent levels, with the five per cent level treated as the primary threshold for inference.
The credibility of a quantitative study rests on the reliability and validity of its measures and design. Reliability concerns the consistency and replicability of the measurement. The study enhances reliability by using systematically compiled ratings and audited financial data, by defining every variable precisely and transparently, and by employing standard, well-documented econometric procedures that another researcher could reproduce. Because the same rating methodology is applied consistently across all firms and years, measurement error, to the extent that it exists, is unlikely to be systematically related to the hypotheses in a way that would manufacture a spurious result; if anything, measurement noise in the ESG variables would bias coefficients towards zero, making the study’s significant findings conservative rather than inflated.
Validity concerns whether the study measures what it purports to measure and whether its inferences are sound. Construct validity is supported by the use of established measures—Tobin’s Q and total shareholder return for performance, widely used ratings for ESG, and conventional controls for fundamentals and risk—each with strong precedent in the finance literature. Internal validity is strengthened by the fixed-effects design, which removes time-invariant confounders, and by the lagged and robustness specifications that probe reverse causality and alternative explanations; nonetheless, as a non-experimental study it can establish association more securely than causation. External validity, or generalisability, is deliberately bounded: the findings pertain to large, premium-listed UK firms and should not be assumed to hold for smaller firms, private companies or markets with materially different governance and ownership structures. The study is best understood as offering internally valid evidence for a well-defined population, with generalisation beyond that population treated as a matter for future research rather than assumption.
Although the study uses secondary, publicly reported data and does not involve human participants, several ethical principles nonetheless apply. The research adheres to the standards of academic integrity, ensuring that all sources are appropriately acknowledged and that the methods and findings are reported transparently and honestly. Because the data for this model document are illustrative, this status is disclosed explicitly so that no reader is misled into treating the results as empirical evidence about specific companies. No individual firm is singled out for criticism, and the analysis is conducted at the level of aggregate relationships rather than naming particular companies in a manner that could be construed as a specific investment recommendation. The study complies with the data-use terms that typically govern commercial ratings and market databases, and it respects the intellectual property of the rating methodologies by describing them in general terms rather than reproducing proprietary detail. Ethical approval of the low-risk, secondary-data category would be obtained from the relevant institutional committee before data collection in a live study.
Several limitations should be acknowledged at the outset. First, reliance on a single ESG rating provider means that the findings may be sensitive to the choice of provider, given the documented divergence among providers; this is a feature of the ESG research landscape rather than a flaw unique to this study, but it constrains the generalisability of pillar-level results. Second, the analysis establishes association rather than causation; although firm and year fixed effects mitigate many sources of omitted-variable bias, the possibility of reverse causality—whereby better-performing firms can afford higher ESG investment—and of time-varying unobserved confounders cannot be fully excluded without a credible instrument or natural experiment. Third, the sample is confined to large UK firms, so the findings may not extend to smaller firms or to other markets. Fourth, for this model document the data are illustrative, which means the specific magnitudes reported demonstrate the method rather than constituting empirical evidence. These limitations are revisited in the concluding chapter.
This chapter presents the results of the empirical analysis. It begins with descriptive statistics that characterise the sample, followed by a correlation analysis that provides a preliminary view of the relationships among the variables and an initial check for multicollinearity. It then reports the main panel regression results for both dependent variables, first using the composite ESG score and then decomposing the effect into the three pillars, before summarising the formal specification tests and a series of robustness checks. The chapter concludes by relating the results to the four hypotheses. As stated throughout, the data underlying these results are illustrative and are presented to demonstrate the analytical approach and the form of the expected findings.
Table 3 reports summary statistics for the full sample of eight hundred and sixty firm-year observations. The mean composite ESG score is 62.4 out of one hundred, indicating that the large UK firms in the sample are, on average, assessed as above the midpoint on sustainability conduct, consistent with the expectation that FTSE 100 constituents disclose extensively and attract relatively favourable ratings. The governance pillar has the highest mean, at 68.1, and the narrowest dispersion, with a standard deviation of 11.3, reflecting the relatively uniform and high standard of governance among premium-listed firms under the UK code. The environmental pillar shows the widest dispersion, with a standard deviation of 18.7, consistent with the diverse sectoral composition of the sample, which ranges from carbon-intensive energy and materials firms to asset-light services. Mean annual total shareholder return is 7.8 per cent with substantial variation, while mean Tobin’s Q is 1.74, indicating that the typical firm is valued above the book value of its assets.
The dispersion patterns are themselves informative for the study’s hypotheses. The narrow spread of governance scores confirms the expectation that the UK code produces relatively homogeneous governance among premium-listed firms, which works against the detection of a governance effect by compressing the explanatory variation; that a significant effect nonetheless emerges in the regressions is therefore notable. Conversely, the wide spread of environmental scores provides ample variation to detect an environmental effect, so the weakness of that effect in the regressions cannot be attributed to a lack of variation and must instead reflect either a genuinely diffuse relationship or the attenuating influence of materiality and measurement divergence. The negative minimum values for total shareholder return and return on assets confirm that the sample spans both prosperous and difficult firm-years, as a decade-long window encompassing varied market conditions would require, lending the panel the temporal richness needed to identify within-firm relationships.
| Variable | Mean | Std. Dev. | Minimum | Median | Maximum |
|---|---|---|---|---|---|
| TSR (%) | 7.8 | 22.6 | −41.3 | 8.1 | 64.9 |
| Tobin’s Q | 1.74 | 1.02 | 0.61 | 1.48 | 5.37 |
| ESG composite | 62.4 | 14.1 | 28.0 | 63.5 | 91.0 |
| Environmental | 58.9 | 18.7 | 14.0 | 60.2 | 94.0 |
| Social | 60.7 | 15.9 | 22.0 | 61.0 | 92.0 |
| Governance | 68.1 | 11.3 | 35.0 | 69.0 | 95.0 |
| SIZE (ln assets) | 9.86 | 1.12 | 7.41 | 9.79 | 12.84 |
| LEV | 0.27 | 0.14 | 0.00 | 0.26 | 0.71 |
| ROA | 0.078 | 0.061 | −0.092 | 0.071 | 0.264 |
| BM | 0.52 | 0.34 | 0.08 | 0.46 | 1.63 |
| BETA | 0.97 | 0.31 | 0.34 | 0.95 | 1.78 |
| DY (%) | 3.6 | 1.9 | 0.0 | 3.5 | 9.2 |
To visualise the pattern that motivates the subsequent analysis, Figure 1 presents mean annual total shareholder return and mean Tobin’s Q across quintiles of the composite ESG score, from the lowest-rated quintile to the highest. The figure reveals a clear monotonic rise in valuation across ESG quintiles but only a weak and non-monotonic pattern in realised returns—a juxtaposition that foreshadows the central finding of the study.
Figure 1. Mean Tobin’s Q and Annual Return by ESG Quintile (Illustrative)
Tobin’s Q (valuation)
1.42
1.58
1.71
1.89
2.06
Annual total shareholder return (%)
6.9
7.6
7.3
8.2
8.4
Table 4 presents the pairwise correlations among the principal variables. The composite ESG score is positively correlated with Tobin’s Q (0.31) and only weakly correlated with annual return (0.08), consistent with the pattern in Figure 1. Among the pillars, governance exhibits the strongest correlation with both valuation (0.29) and return (0.12), while the environmental and social pillars show weaker correlations. As anticipated, the three pillars are positively correlated with one another—most strongly environmental with social (0.61)—which motivates the use of variance inflation factors in the pillar-level models. The correlations among the control variables are moderate, and none approaches a level that would signal severe multicollinearity. Firm size is positively correlated with the ESG score (0.34), reflecting the well-documented tendency for larger firms to achieve higher ratings, which underlines the importance of controlling for size.
| TSR | TQ | ESG | ENV | SOC | GOV | SIZE | ROA | |
|---|---|---|---|---|---|---|---|---|
| TSR | 1.00 | |||||||
| TQ | 0.19 | 1.00 | ||||||
| ESG | 0.08 | 0.31 | 1.00 | |||||
| ENV | 0.04 | 0.22 | 0.84 | 1.00 | ||||
| SOC | 0.06 | 0.24 | 0.81 | 0.61 | 1.00 | |||
| GOV | 0.12 | 0.29 | 0.66 | 0.38 | 0.41 | 1.00 | ||
| SIZE | −0.03 | −0.11 | 0.34 | 0.37 | 0.30 | 0.18 | 1.00 | |
| ROA | 0.21 | 0.46 | 0.14 | 0.09 | 0.12 | 0.17 | −0.08 | 1.00 |
Table 5 reports the regression of each dependent variable on the composite ESG score and controls, under pooled ordinary least squares, random effects and fixed effects. Attention focuses on the fixed-effects columns, which the specification tests described in Section 4.6 identify as preferred. For the valuation model, the coefficient on the composite ESG score is positive and statistically significant at the five per cent level under fixed effects (0.0068), indicating that a ten-point increase in the ESG rating is associated with an increase in Tobin’s Q of approximately 0.068, holding firm fundamentals, systematic risk and both firm and year effects constant. This is an economically meaningful effect given a sample mean Q of 1.74. For the return model, by contrast, the coefficient on the composite ESG score is small and statistically indistinguishable from zero (0.041, not significant), indicating no reliable association between the ESG rating and realised annual returns once controls and fixed effects are included.
| Variable | TQ (Pooled OLS) | TQ (Fixed Effects) | TSR (Fixed Effects) |
|---|---|---|---|
| ESG composite | 0.0121*** | 0.0068** | 0.041 |
| SIZE | −0.142*** | −0.098* | −1.21 |
| LEV | 0.214 | 0.187 | −4.83 |
| ROA | 4.86*** | 3.92*** | 61.4*** |
| BM | −0.638*** | −0.511*** | 9.72** |
| BETA | −0.093 | −0.077 | −3.14 |
| DY | −0.052** | −0.038 | 0.86 |
| Firm fixed effects | No | Yes | Yes |
| Year fixed effects | No | Yes | Yes |
| Observations | 860 | 860 | 860 |
| Within / adjusted R² | 0.38 | 0.29 | 0.17 |
Notes: Illustrative coefficients. Standard errors clustered at the firm level. *** , ** and * denote significance at the 1%, 5% and 10% levels respectively.
The control variables behave as theory would predict, which lends face validity to the specification. Profitability, measured by return on assets, is strongly and positively associated with both valuation and returns, consistent with its role as a fundamental driver of value. The book-to-market ratio is negatively associated with Tobin’s Q by construction, since high book-to-market firms are, almost by definition, valued closer to or below book value, while it carries the positive sign in the return model that the value premium would predict. Firm size enters negatively in the valuation model, consistent with the tendency for the very largest firms to trade at lower multiples, and systematic risk and dividend yield are not reliably associated with either outcome in the fixed-effects specification. The overall pattern—ESG significant for valuation but not for returns—is the headline result and is explored further below.
Table 6 reports the fixed-effects models in which the composite score is replaced by the three pillar scores entered simultaneously. Variance inflation factors for the three pillars range from 2.4 to 3.1, well below the conventional threshold of ten, indicating that although the pillars are correlated, multicollinearity does not undermine the separate estimation of their effects. The decomposition is revealing. In the valuation model, the governance pillar carries a positive and highly significant coefficient (0.0094, significant at the one per cent level), the environmental pillar a smaller and marginally significant coefficient (0.0031, significant at the ten per cent level), and the social pillar a coefficient that is positive but statistically insignificant. In the return model, only the governance pillar approaches significance (0.071, significant at the ten per cent level), while the environmental and social pillars are insignificant and economically small.
| Variable | Tobin’s Q | TSR |
|---|---|---|
| Environmental (ENV) | 0.0031* | 0.018 |
| Social (SOC) | 0.0022 | 0.015 |
| Governance (GOV) | 0.0094*** | 0.071* |
| Controls | Yes | Yes |
| Firm & year fixed effects | Yes | Yes |
| Observations | 860 | 860 |
| Within R² | 0.31 | 0.18 |
Notes: Illustrative coefficients. *** , ** and * denote significance at the 1%, 5% and 10% levels respectively.
Figure 2 summarises the valuation coefficients visually, making plain the hierarchy of effects anticipated by the theoretical framework: governance exerts the largest and most statistically reliable effect, the environmental pillar a smaller and weaker one, and the social pillar an effect that cannot be distinguished from zero at conventional levels. This hierarchy is the central empirical contribution of the pillar decomposition.
Figure 2. Valuation Effect by ESG Pillar (Illustrative Coefficients on Tobin’s Q)
0.0094***
0.0031*
0.0022 (n.s.)
To probe whether the governance–valuation relationship is uniform across the sample or concentrated in particular types of firm, the sample was partitioned in two ways and the valuation model re-estimated within each sub-sample. The first partition split firms at the median of firm size; the second split them at the median environmental-pillar score, separating firms in carbon-intensive activities, which tend to have lower environmental scores and greater exposure to transition risk, from the rest. The results, summarised in Table 8, indicate that the governance effect on valuation is present across all sub-samples but is economically larger among smaller firms within the index and among environmentally exposed firms. The larger governance effect among smaller firms is consistent with the view that governance quality is most consequential where monitoring by the market is less intense, so that the marginal benefit of strong internal governance is greater. The larger effect among environmentally exposed firms is consistent with investors placing a premium on sound oversight precisely where the stakes—in the form of environmental liabilities and transition risk—are highest, implying an interaction between the governance and environmental dimensions that composite scores obscure.
| Sub-sample | Governance coefficient | Observations |
|---|---|---|
| Smaller firms (below median size) | 0.0121*** | 430 |
| Larger firms (above median size) | 0.0068** | 430 |
| Environmentally exposed (low ENV) | 0.0110*** | 430 |
| Less exposed (high ENV) | 0.0071** | 430 |
Notes: Illustrative coefficients from firm- and year-fixed-effects models estimated within each sub-sample. *** and ** denote significance at the 1% and 5% levels respectively.
These heterogeneity results should be read as suggestive rather than definitive, given the reduced number of observations within each sub-sample and the illustrative nature of the data. Nonetheless, they reinforce the central message that governance is the dimension the market prices most reliably, and they hint that its value is contingent on context—greatest where oversight is otherwise weaker and where the consequences of poor oversight are most severe. This contingency is consistent with the materiality logic that recurs throughout the analysis and provides a bridge to the fuller interpretation offered in the discussion chapter.
The choice between the random-effects and fixed-effects estimators was formalised using the Hausman test. For both the valuation and return models, the test rejected the null hypothesis that the firm-specific effects are uncorrelated with the regressors at the one per cent level, confirming that the fixed-effects estimator is both consistent and preferred. This result is unsurprising given the strong theoretical expectation that unobserved firm characteristics—sector, business model, corporate culture—are correlated with both ESG ratings and performance, and it validates the decision to base inference on the fixed-effects specification. A test for the joint significance of the firm fixed effects comfortably rejected the null that they are jointly zero, confirming that pooled ordinary least squares is misspecified. Year fixed effects were likewise jointly significant, reflecting the influence of market-wide conditions over the sample period, which spanned a range of macroeconomic and market environments.
Several robustness checks were conducted to assess the stability of the findings. First, the dependent variables were re-measured—using buy-and-hold returns in place of annual total shareholder return and the market-to-book ratio in place of Tobin’s Q—and the pattern of results was unchanged: ESG remained significant for valuation and insignificant for returns, and governance remained the dominant pillar. Second, the ESG variables were lagged by one year to address the concern that contemporaneous ratings and performance might be jointly determined; the lagged specification yielded very similar coefficients, with the governance effect on valuation remaining significant at the one per cent level, which somewhat alleviates reverse-causality concerns although it cannot eliminate them. Third, the sample was split into the first and second halves of the period to examine temporal stability; the governance–valuation relationship was present in both sub-periods and was somewhat stronger in the later years, consistent with the rising salience of ESG among investors over the sample window. Fourth, re-estimating the models excluding energy and materials firms—the sectors with the most extreme environmental scores—left the conclusions intact. Collectively, these checks indicate that the central findings are robust to reasonable variations in measurement, timing and sample composition.
Table 7 summarises the outcome of the four hypothesis tests, distinguishing between the valuation and return outcomes. The composite ESG hypothesis (H1) is supported for valuation but not for returns. Among the pillars, the governance hypothesis (H4) receives the strongest support, being significant for valuation and marginally for returns; the environmental hypothesis (H2) receives weak support for valuation only; and the social hypothesis (H3) is not supported for either outcome. The overall picture is one of a value-relevant but efficiently priced relationship concentrated in the governance dimension.
| Hypothesis | Valuation (Tobin’s Q) | Return (TSR) |
|---|---|---|
| H1: Composite ESG positive | Supported (5%) | Not supported |
| H2: Environmental positive | Weakly supported (10%) | Not supported |
| H3: Social positive | Not supported | Not supported |
| H4: Governance positive | Supported (1%) | Weakly supported (10%) |
This chapter interprets the empirical findings in relation to the theoretical framework and prior literature developed in Chapter 2, and it draws out the implications for investors, managers and policymakers. The discussion is organised around the two principal patterns in the results: the divergence between the valuation and return outcomes, and the dominance of the governance pillar. It then considers the practical consequences of these patterns and situates them within the ongoing problem of ESG rating divergence.
The most striking feature of the results is that the composite ESG rating is significantly associated with firm valuation but not with realised annual returns. This pattern is not contradictory; on the contrary, it is precisely what the efficient market hypothesis of Fama (1970) would predict. If ESG performance creates value—through the cash-flow, idiosyncratic-risk and systematic-risk channels described by Giese and colleagues (2019)—and if that value is publicly observable through ratings and disclosures, then rational investors will bid up the prices of high-ESG firms until their valuations reflect the capitalised value. Once this repricing has occurred, there is no reason to expect high-ESG firms to continue earning abnormal returns, because the good news is already in the price. A positive association with valuation levels alongside a null association with returns is therefore the signature of an efficient market that has already recognised and priced the value of ESG.
This interpretation reconciles two apparently conflicting strands of the literature. Studies that document a positive relationship between ESG and valuation or a lower cost of capital—such as El Ghoul and colleagues (2011)—are capturing the capitalisation of value into prices. Studies that fail to find persistent abnormal returns to high-ESG portfolios are observing the consequence of that capitalisation: once priced, the premium disappears. The exception identified by Edmans (2011), in which an element of the social pillar predicted returns, is explicable as a case in which the relevant information—employee satisfaction—was an intangible not yet fully recognised by the market; as such information becomes more widely disclosed and analysed, the scope for it to predict returns should diminish. The present finding that even the social pillar shows no return predictability in a sophisticated, heavily analysed market such as the FTSE 100 is consistent with this maturation of ESG information.
The practical corollary is important and frequently misunderstood. The finding that high-ESG firms are more highly valued does not imply that an investor can earn superior returns simply by buying them, because they are already expensive in valuation terms. The financial case for ESG, on this evidence, rests more on risk mitigation and on the avoidance of value-destroying controversies than on the generation of alpha—a conclusion aligned with the risk-focused strand of the literature and with the ESG-efficient-frontier reasoning of Pedersen, Fitzgibbons and Pomorski (2021), in which the return consequences of ESG depend on investor preferences and on whether ESG predicts fundamentals beyond what is already known.
The risk-based reading of the evidence, developed in the literature review, deepens this interpretation. If high-ESG firms command higher valuations partly because the market perceives them as less exposed to severe negative events—environmental liabilities, product-safety failures, labour disputes and governance scandals—then the valuation premium is, in effect, compensation for lower risk. Lower risk should, by the basic logic of asset pricing, be associated with lower, not higher, expected returns, which makes the null average-return result not merely consistent with efficiency but positively expected. On this account, the investor who buys high-ESG firms is purchasing a lower-risk asset at a correspondingly higher price, and should anticipate a smoother return path rather than a higher average return. The financial case for ESG integration therefore rests substantially on the shape of the return distribution—on the avoidance of large losses—rather than on its mean, a distinction that annual return regressions capture only imperfectly and that may explain why the return coefficient, though positive, is statistically indistinguishable from zero.
The second central finding is that, once the composite score is decomposed, the governance pillar carries the strongest and most consistent association with value, while the environmental and social pillars are weak or insignificant. This result is strongly consistent with agency theory as formalised by Jensen and Meckling (1976). Governance quality operates directly on the principal–agent mechanism that finance theory identifies as central to value: independent and effective boards, aligned remuneration, robust audit and transparent reporting reduce agency costs, constrain value-destroying managerial behaviour and improve the quality of capital allocation. Because this channel bears so directly on the determinants of firm value, it is unsurprising that governance should be the dimension most reliably priced.
The relative weakness of the environmental and social pillars admits several complementary explanations. The materiality perspective of Khan, Serafeim and Yoon (2016) suggests that the association between environmental or social performance and value depends heavily on whether the issues in question are financially material to the firm’s industry. Because the composite environmental and social scores aggregate material and immaterial issues indiscriminately, their average association with value is diluted. An energy firm’s environmental conduct may be highly material, while an asset-light services firm’s is not; averaging across such heterogeneous firms attenuates the measured effect. The relative homogeneity of the social pillar among large, well-resourced FTSE 100 firms—most of which maintain comprehensive human-capital and community programmes—further compresses the cross-sectional variation available to detect an effect.
It is also possible that some environmental and social expenditure reflects managerial private benefits rather than value creation, as a strict reading of Friedman (1970) and agency theory would caution. If a portion of high environmental and social scores arises from discretionary spending that does not enhance cash flow, its net association with value would be muted, offsetting the genuine value created elsewhere. The data cannot distinguish these mechanisms directly, but the pattern—governance strong, environmental and social weak—is consistent with a market that rewards the dimension most clearly tied to value discipline and is more sceptical of the dimensions where value creation and value destruction are harder to separate.
The sub-sample evidence reported in Chapter 4 adds a further layer to this interpretation. The finding that the governance effect is larger among smaller index constituents accords with the monitoring logic of agency theory: where external scrutiny from analysts and the market is less intense, internal governance mechanisms bear more of the burden of disciplining management, so their quality has a greater marginal effect on value. The finding that the governance effect is also larger among environmentally exposed firms suggests an interaction that composite scores cannot capture: for firms carrying substantial environmental liabilities and transition risk, the quality of oversight—the board’s capacity to anticipate, manage and disclose those risks—is especially consequential, so investors reward it more generously. This interaction implies that the pillars are not strictly independent; good governance may be the mechanism through which environmental and social risks are competently managed, so that part of the value conventionally attributed to the environmental and social pillars may in fact operate through governance. This is a conjecture that the present design cannot test directly, but it is a theoretically coherent explanation for the governance pillar’s dominance and a promising direction for future work.
The findings sit comfortably within the broad consensus summarised by Friede, Busch and Bassen (2015) that the ESG–performance relationship is predominantly non-negative, while adding important nuance. The predominantly positive relationship documented in the meta-literature is, on this evidence, concentrated in valuation rather than returns and in governance rather than the other pillars. This helps explain why the literature is so inconsistent: studies that use valuation measures and governance-heavy scores are more likely to find positive effects, while those that use return measures and composite scores are more likely to find weak or null effects. The present study’s dual-outcome, pillar-level design thus offers a partial resolution of the heterogeneity that has long troubled the field, by showing that the apparent inconsistency can arise from differences in what is measured rather than from genuine disagreement about the underlying relationship.
The results also echo the valuation-channel reasoning of Giese and colleagues (2019), in that the significant valuation effect is consistent with ESG affecting the cost of capital and the market’s assessment of risk and competitiveness. At the same time, the null return result cautions against the stronger claims sometimes made for ESG investing, aligning with the more measured position of Pedersen, Fitzgibbons and Pomorski (2021) that ESG’s return consequences are contingent rather than guaranteed.
For investors, the findings carry a clear message: ESG, and governance in particular, is value-relevant, but high-ESG firms are already priced to reflect it. An ESG tilt may be justified on the grounds of risk reduction, alignment with client preferences or long-term fundamentals, but it should not be expected, on this evidence, to generate systematic outperformance in an efficient, heavily analysed market. Investors seeking a return advantage from ESG would need to identify information not yet priced—for example, material issues that the market has underappreciated—rather than relying on headline ratings that are, by construction, public and widely followed.
For corporate managers, the pillar decomposition offers actionable guidance on the allocation of scarce resources. Investment in governance quality—board effectiveness, remuneration alignment, transparency and audit—appears to be the dimension most reliably associated with higher valuation, and it is therefore a priority with a defensible financial rationale. Investment in environmental and social performance should be guided by materiality: managers should concentrate on the environmental and social issues that are financially material to their specific industry, where the value case is strongest, rather than pursuing broad improvements in composite scores for their own sake. This targeted approach is more consistent with value creation than a generalised pursuit of higher ratings.
The pillar decomposition also has implications for corporate communication and disclosure strategy. Because the market appears to price governance quality reliably but to treat environmental and social claims more sceptically, managers are likely to be rewarded for governance improvements that are verifiable and structural—board refreshment, genuine independence, transparent remuneration—rather than for environmental and social assertions that are difficult to substantiate. This argues for an emphasis on credible, audited, decision-useful disclosure over promotional sustainability narratives, both because credible disclosure reduces the information asymmetry that depresses valuation and because unsubstantiated claims expose the firm to the reputational and regulatory risks associated with greenwashing. In an environment of heightened scrutiny of sustainability marketing, the disciplined, evidence-based communication of genuine governance and material environmental and social performance is more likely to be rewarded than an expansive but weakly supported ESG narrative.
For policymakers and standard-setters, the findings reinforce the case for improving the comparability and material focus of ESG disclosure. The value that the market places on governance quality suggests that continued attention to governance standards is well founded. The weaker measured effects of the environmental and social pillars, combined with the materiality evidence, argue for disclosure frameworks that distinguish material from immaterial issues, so that investors can price the information that matters rather than rewarding disclosure volume indiscriminately.
The findings also speak to the stewardship role of institutional investors, which is especially prominent in the UK market. If governance is the dimension most reliably rewarded with higher valuation, then the engagement activities of large asset owners and managers—voting against poorly structured remuneration, pressing for independent and diverse boards, and holding directors to account—are not merely compliance exercises but potentially value-accretive interventions. The evidence that the governance effect is larger among smaller index constituents, where market monitoring is thinner, suggests that stewardship resources may be most productively directed towards firms where the governance deficit, and hence the scope for improvement, is greatest. This has practical significance for the design of stewardship programmes under the strengthened expectations placed on UK institutional investors, implying that a targeted, governance-focused engagement strategy may deliver more value than a diffuse effort spread thinly across all holdings and all ESG dimensions.
Finally, the results carry a cautionary implication for the rapidly expanding market in ESG-labelled investment products. Because the valuation premium attached to high-ESG firms appears to be largely priced, products that simply tilt towards high-rated firms are unlikely to deliver outperformance net of fees, and may expose investors to the sector and style biases that critics of ESG investing have highlighted. A more defensible proposition for such products is framed around risk reduction, values alignment and real-world impact rather than around a promise of superior returns—a framing that is both more consistent with the evidence and less vulnerable to the charge of mis-selling that has begun to attract regulatory attention.
Any interpretation of these results must be tempered by the rating-divergence problem documented by Berg, Kölbel and Rigobon (2022), who report an average correlation of approximately 0.54 among major providers’ ratings. Because the present study relies on a single provider, the pillar-level findings in particular may be sensitive to the choice of provider; a different provider, weighting the same underlying conduct differently, could produce somewhat different coefficients. This does not invalidate the findings, but it does bound their generalisability. The governance result is likely to be the most robust to provider choice, because governance attributes—board independence, chair–chief-executive separation, remuneration structures—are relatively objective and consistently measured, whereas environmental and social attributes involve more judgement and hence more divergence. The weakness of the environmental and social pillars may therefore reflect, in part, the noise introduced by measurement disagreement, which attenuates estimated coefficients towards zero. This observation suggests that improving the reliability of environmental and social measurement could strengthen the detectable relationship, and it underscores the importance of treating single-provider results with appropriate caution.
This dissertation set out to evaluate the extent to which ESG ratings are associated with the share price performance of FTSE 100 companies over the period 2015 to 2024, distinguishing between valuation and realised returns and decomposing the relationship by pillar. Using an illustrative balanced panel of eighty-six non-financial firms and fixed-effects regression with firm and year effects, the study reached three principal conclusions. First, the composite ESG rating is positively and significantly associated with firm valuation, measured by Tobin’s Q, but is not reliably associated with realised annual returns. Second, this relationship is driven overwhelmingly by the governance pillar, which is significant for valuation and marginally for returns, while the environmental pillar is weak and the social pillar insignificant once governance and fundamentals are controlled for. Third, these patterns are robust to alternative measures, lagged specifications, sub-period analysis and sample restrictions. Returning to the objectives set out in Chapter 1, the study has reviewed the relevant literature, constructed an appropriate panel design, estimated the ESG–performance relationship under both outcomes, decomposed it by pillar, and interpreted it through competing theoretical lenses.
The study contributes to theory in two ways. First, by distinguishing valuation from returns, it demonstrates that the ESG–performance relationship is best understood through the efficient market hypothesis: value is capitalised into prices rather than expressed as persistent abnormal returns, a framing that reconciles the conflicting findings of the empirical literature. Second, by decomposing the composite score, it provides direct support for the primacy of agency-theoretic reasoning in explaining which ESG dimension is priced: governance, the pillar most closely tied to the principal–agent mechanism, dominates, consistent with Jensen and Meckling (1976), while the environmental and social pillars operate more diffusely and are attenuated by materiality and measurement considerations. The study thus integrates stakeholder, agency, signalling and efficient-markets perspectives into a coherent account of how and where ESG enters prices.
For investors, the study cautions that ESG ratings, being public, are unlikely to generate outperformance in an efficient market and are better justified by risk management and preference alignment; any return advantage must come from information the market has not yet priced. For managers, it recommends prioritising governance quality, which carries the clearest value rationale, and targeting environmental and social investment at issues that are financially material to the firm’s industry rather than pursuing composite-score improvements indiscriminately. For policymakers, it supports continued emphasis on governance standards and argues for disclosure frameworks that distinguish material from immaterial issues and that improve the comparability and reliability of environmental and social data, thereby reducing the rating divergence that currently clouds interpretation.
The conclusions are subject to several limitations, beyond the illustrative nature of the data used for this model document. The reliance on a single rating provider means the pillar-level findings may not generalise across providers, given documented divergence. The design establishes association rather than causation; although fixed effects and lagged specifications mitigate omitted-variable and reverse-causality concerns, they do not eliminate them, and a definitive causal claim would require a credible natural experiment or instrument. The sample is restricted to large UK firms, so the findings may not extend to smaller companies, private firms or other markets with different governance regimes and ownership structures. Finally, the study uses contemporaneous and one-year-lagged ratings and cannot speak to very long-horizon effects that may unfold over decades.
Several avenues follow directly from these limitations. First, future work should replicate the pillar-level analysis using multiple rating providers, both to assess the sensitivity of the findings to provider choice and to exploit the divergence itself as a source of insight into which attributes the market prices. Second, researchers could pursue causal identification through natural experiments—such as regulatory shocks to disclosure requirements or governance reforms—that generate exogenous variation in ESG performance. Third, the materiality hypothesis could be tested directly within the UK setting by classifying environmental and social issues by industry materiality and re-estimating the models, which would clarify whether the weakness of these pillars reflects genuine irrelevance or the dilution of material effects by immaterial ones. Fourth, extending the analysis to smaller firms and to other European markets would establish the external validity of the governance result. Finally, event-study and cost-of-capital analyses could complement the present valuation and return approach, tracing more precisely the channels through which ESG information is impounded into prices.
The question of whether sustainability pays has too often been posed as though it admitted a single, universal answer. The evidence assembled in this model dissertation suggests a more nuanced conclusion: ESG is value-relevant, but chiefly through governance, chiefly in valuation rather than returns, and chiefly as information that an efficient market has already priced. For the large UK firms of the FTSE 100, the market appears to reward the discipline of good governance with higher valuations, to treat environmental and social performance more cautiously and conditionally, and to leave little room for ESG ratings alone to generate abnormal returns. Understanding this distinction—between what creates value, what is priced, and what remains to be earned—is essential for investors, managers and policymakers navigating the rapidly maturing landscape of sustainable finance.
Berg, F., Kölbel, J.F. and Rigobon, R. (2022) ‘Aggregate confusion: the divergence of ESG ratings’, Review of Finance, 26(6), pp. 1315–1344.
Chava, S. (2014) ‘Environmental externalities and cost of capital’, Management Science, 60(9), pp. 2223–2247.
Edmans, A. (2011) ‘Does the stock market fully value intangibles? Employee satisfaction and equity prices’, Journal of Financial Economics, 101(3), pp. 621–640.
El Ghoul, S., Guedhami, O., Kwok, C.C.Y. and Mishra, D.R. (2011) ‘Does corporate social responsibility affect the cost of capital?’, Journal of Banking & Finance, 35(9), pp. 2388–2406.
Fama, E.F. (1970) ‘Efficient capital markets: a review of theory and empirical work’, The Journal of Finance, 25(2), pp. 383–417.
Freeman, R.E. (1984) Strategic management: a stakeholder approach. Boston: Pitman.
Friede, G., Busch, T. and Bassen, A. (2015) ‘ESG and financial performance: aggregated evidence from more than 2000 empirical studies’, Journal of Sustainable Finance & Investment, 5(4), pp. 210–233.
Friedman, M. (1970) ‘The social responsibility of business is to increase its profits’, The New York Times Magazine, 13 September.
Giese, G., Lee, L.-E., Melas, D., Nagy, Z. and Nishikawa, L. (2019) ‘Foundations of ESG investing: how ESG affects equity valuation, risk, and performance’, The Journal of Portfolio Management, 45(5), pp. 69–83.
Gillan, S.L., Koch, A. and Starks, L.T. (2021) ‘Firms and social responsibility: a review of ESG and CSR research in corporate finance’, Journal of Corporate Finance, 66, 101889.
Jensen, M.C. and Meckling, W.H. (1976) ‘Theory of the firm: managerial behavior, agency costs and ownership structure’, Journal of Financial Economics, 3(4), pp. 305–360.
Khan, M., Serafeim, G. and Yoon, A. (2016) ‘Corporate sustainability: first evidence on materiality’, The Accounting Review, 91(6), pp. 1697–1724.
Pedersen, L.H., Fitzgibbons, S. and Pomorski, L. (2021) ‘Responsible investing: the ESG-efficient frontier’, Journal of Financial Economics, 142(2), pp. 572–597.
Spence, M. (1973) ‘Job market signaling’, The Quarterly Journal of Economics, 87(3), pp. 355–374.
Wooldridge, J.M. (2010) Econometric analysis of cross section and panel data. 2nd edn. Cambridge, MA: MIT Press.
You May Also Like
All work is written by human writers. 100% AI free, guaranteed.
100% money back guarantee if you find plagiarism in our work.
COMPANY DETAILS
