Does the Underlying Design of Environmental, Social, and Governance (ESG) Indices Affect Investor Reactions? The Role of Legitimacy and Reputation Effects

Abstract

The growing importance of socially responsible investments is causing a rapid increase in the number of various ESG indices. This raises the question of whether the index design matters to stock market investors. The purpose of the article is therefore to analyze the impact of ESG index design on investor decisions motivated by announcements of index reconstitutions. It was assumed that information about company additions to, or deletions from, an index—signaling an improvement in or deterioration of its CSR standards—may be differently interpreted by investors depending on the context provided by the index design. This study used data on the reconstitutions of two ESG indices. One of them, FTSE4Good US, is based on negative screening. Due to its design, membership in it is strongly associated with legitimacy. The other index, DJSI North America, is a best-in-class index which confers a reputation effect. We have applied the event window methodology, which identifies the economic effects of an event by estimating its impact on share prices as reflected in the rate of return. Analysis encompassed 691 events concerning American listed companies in the years 2009–2019, of which 441 were additions and 250 were deletions. It was found that significant investor reactions were triggered only by reconstitutions of the index generating a reputation effect (DJSI). These results indicate that index design does matter. The reactions of investors were positive only when they associated a company’s social commitment with the creation of intangible resources contributing to its competitive advantage. Our results suggest that inclusion in a best-in-class index is more beneficial for a company than in an index based on negative screening.

1. Introduction

The performance of companies in the field of corporate social responsibility (CSR) has attracted growing interest among investors. Between 2016 and 2022, the value of assets managed according to the idea of socially responsible investing (SRI) increased from USD 22.9 trillion to USD 30.3 trillion [1]. Companies may signal their CSR efforts via their websites, sustainability reports, press releases, advertising, cause-related marketing, and other means. While investors may get their information from all those sources, such voluntarily published communications are of limited use for them, as they are incomplete, sometimes inconsistent with third-party intelligence, and of uncertain reliability. Therefore, investors seek verified and trusted information, such as that provided by institutional intermediaries—operators of sustainability indices that evaluate the CSR efforts of companies and make decisions about listing and delisting them. The companies to be included in an index must meet a set of sustainability performance criteria, and so an addition to or presence in an index is seen by investors as evidence of high CSR involvement, while a deletion from an index is viewed as deterioration in that involvement in relative or absolute terms. Announcements of index reconstitutions provide reliable signals to investors, strongly affecting the way they perceive a company’s commitment to sustainability [2]. As a result, an addition to or deletion from a sustainability index may elicit a response from investors if they deem this event to be causing an increase or decrease in the company’s attractiveness.

According to Hirigoyen and Poulain-Rehm [3], studies on market reactions to sustainability index reconstitutions belong to one of the two main strands of research on the relationship between CSR and corporate financial performance (CFP). Existing research on reactions to company additions to or deletions from sustainability indices is inconclusive. While some authors have reported positive reactions to additions and negative reactions to deletions [4,5,6,7], others have found negative reactions to additions and positive to deletions [8,9], negative reactions to deletions combined with no definite response to additions [10,11], or negative reactions to both additions and deletions [12,13]. Finally, some studies have reported no significant reactions to index reconstitutions at all [14,15,16] or they have reported mixed results [17,18]. One of the reasons for these discrepancies is the presence of additional factors modifying investor evaluations of company sustainability, and consequently their reactions to it. According to the literature, such factors include time and space [15], overall market conditions [2,18], the institutional environment [19], the properties of specific industrial sectors [7], as well as the characteristics of companies themselves: their financial performance [20], leverage ratio [12], and expenditures on R&D (Research and Development) and advertising [21]. However, no study to date has investigated the effects on investor reactions of sustainability index design. The present paper fills that gap.

ESG indices may be considered a form of self-regulation combined with certification [22]. The three longest-standing indices are the Domini 400 Social Index, the Dow Jones Sustainability Index (DJSI), and the FTSE4Good Index. While all of them have been used in research on investor reactions to index reconstitutions, most studies have investigated reactions to changes in one index only, or in several indices of similar design. However, index design may vary considerably, and a clear distinction should be made between negative and positive screening indices, with some of the latter using best-in-class selection systems [23]. The fundamental difference between negative screening and best-in-class indices is the mechanism of company selection, both in terms of the pool of candidates and the nature of evaluation, with ramifications for addition/deletion decisions. In contrast to best-in-class indices, negative screening indices exclude a priori companies which are active in sectors controversial from the standpoint of broader social interest. Furthermore, in their case company evaluation outcomes are considered in absolute terms, which means that all companies meeting the minimum requirements defined by the index operator will be included. In turn, best-in-class indices interpret evaluation results in relative terms, and so they only include industry leaders which maintain higher sustainability standards as compared to other companies in their sectors.

In considering differences in ESG index design and their ramifications for the way investors assess the signals generated by company additions to and deletions from a given type of index, we drew on legitimacy theory [24,25,26,27] and reputation theory [28,29,30,31]. Both legitimacy and reputation can be understood, on the one hand, as social judgement, and, on the other hand, as a resource. However, there are some significant differences between them in both of these cases. Legitimacy and reputation as social judgement differ in terms of the nature of social appraisal (and its reference frame), as well as the effects of that judgment on key reinforcement mechanisms and the stratification of the evaluated companies. Due to these differences, the mechanisms of company selection used in ESG indices of different design reflect the aspects of social judgment more aligned with either legitimacy or reputation. As a result, the design of an index is critical to the way investors interpret its reconstitutions. For some indices, additions/deletions will be treated as signals of an increase/decrease in company legitimacy; for other indices, they will be taken as signals of an improvement/deterioration in company reputation. The context in which company additions/deletions are considered may also affect investor decisions due to the properties of legitimacy and reputation as resources. Legitimacy is a resource that is valuable, imperfectly imitable, and not readily substitutable, but it is not rare. While it is necessary for company survival, it is not a source of competitive advantage. Reputation is also valuable, imperfectly imitable, and not readily substitutable, but it is rare; so, it does give rise to competitive advantage [32], which may be sustained over time [33]. As a company’s pool of resources determines its strategy [34], a good reputation has been recognized as a factor differentiating the company [35], which may enable it to achieve superior performance in the long run [36]. Thus, improved reputation increases company attractiveness to investors to a far greater degree than legitimacy.

In order to determine whether the characteristics of an index influence investor reactions, the present study examined indices with different designs. One of them, the FTSE4Good US Index, is based on negative screening, and so it exhibits a greater affinity to the concept of legitimacy. The other one, the DJSI North America Index, follows a best-in-class approach, which means that it generates a positive reputation effect for the companies included. The event window methodology was used to analyze investor responses in the U.S. market to 242 additions and 98 deletions from FTSE4Good US as well as 199 additions and 152 deletions from DJSI North America. In total, analysis comprised 691 events from the years 2009–2019. To the best of our knowledge, this is the first study of its kind to examine such a large number of events involving markedly different indices.

The results show that index design does matter. Investors reacted to company additions and deletions only when information was regarded as being reputation-relevant (unsurprisingly, reputation improvement elicited positive investor reactions and vice versa). These findings indicate that investors believe CSR activity to be a source of value growth when associated with the enhancement of reputation, perceived as a resource contributing to the company’s competitive advantage. In situations when CSR policies do not give rise to such resources, CSR expenditures are not deemed to affect future company value.

The present paper makes several contributions to the literature. It expands knowledge about the factors shaping investor reactions to announcements of company additions to and deletions from sustainability indices. Previous research has examined reactions to reconstitutions in one index only or in several indices of similar design, and so it has not elucidated the effects of index design, including company selection mechanisms, on investor reactions. Our findings show that index design provides a context that influences the way investors interpret the signals sent by index reconstitutions. Thus, it may be treated as a new, previously undescribed factor modifying investor behavior, explaining the heterogeneity of results reported from previous research on investor reactions to announcements of company additions/deletions. We also identified the mechanisms underpinning differential investor reactions to the reconstitutions of indices with different designs. In this context, the distinction between legitimacy and reputation effects appears to be of the essence. Negative screening indices confer a legitimacy effect, but not a reputation effect; in turn, the latter is associated with best-in-class indices. Previous research concerning CSR effects on legitimacy and reputation in the context of company additions to and deletions from ESG indices [10,37] has not differentiated between those effects or linked them to index design. The present paper also affords a better understanding of the differences between the economic significance of legitimacy and reputation. To date, both of them have been treated as sources of economic benefits [26,38,39,40], but neither has been the subject of comparative analyses. This study indicates that investors do not treat legitimacy as a factor that could impact business performance sufficiently enough to alter the perceived levels of company attractiveness. However, positive/negative reputation changes do matter for investors, which may indicate that reputation exerts a stronger effect on company performance, and so that it is linked to greater economic benefits as compared to legitimacy.

The paper is organized as follows. The section on literature review and hypothesis development gives an outline of existing publications on investor reactions, legitimacy, and reputation; it also explores the relationship of legitimacy and reputation with ESG index design and sets forth the research hypotheses. This is followed by a description of the methodology and dataset used in the study. Subsequently, the findings are presented alongside a discussion of their implications. The paper ends with conclusions.

2. Theoretical Background and Hypothesis Development

2.1. ESG Indices and Investor Reactions

The development of sustainability indices is associated with the rise of socially responsible investing (SRI). The first index of the type was the Domini 400 Social Index, established in May 1990. Ever since, the popularity of sustainability indices has grown considerably. With respect to their design, most indices fall into two broad categories of negative and positive screening, with some in the latter category using a best-in-class selection strategy [23]. In addition, a few indices employ a combination of different approaches. In negative screening, companies from sectors considered controversial from the standpoint of general social interest are excluded a priori (the exact list of such sectors varies between indices and may change over time). A priori exclusion is not applied in positive screening indices, which include companies with above-average levels of social commitment regardless of their sector. Such indices may also follow a best-in-class approach, listing only those companies which score highest within their respective sectors.

Irrespective of the selection mechanism underlying a given index, an addition means that the company in question has been favorably evaluated by an independent body monitoring corporate social performance (CSP). In the case of each individual index (or family of indices), such evaluation is based on a different set of specific criteria addressing the three fundamental areas of environmental, social, and corporate governance (ESG). Similarly to the list of ineligible sectors in negative screening indices, these criteria may be modified over time to reflect the evolution of CSP-relevant issues driven by changes in the sphere of social norms and values determining expectations about companies [41]. A good case in point is environmental protection, which has been reflected in abolishing the coal industry from the family of FTSE4Good indices. The elimination of controversial sectors and the accommodation of changes in the hierarchy of social norms and values enables ESG indices to play a legitimizing role as the addition of a company implies that it conforms to the social standards prevalent at a given time. Depending on index design, evaluation results may be of absolute or relative nature. In the former case, an index addition means that the company has met the minimum criteria defined by the index operator, while, in the latter case, it indicates that the company represents higher ESG standards than its competitors. Due to these differences in selection mechanisms, indices differ in their potential effects on company legitimacy and reputation.

Despite some criticism [42], sustainability indices may play an important role for investors and companies alike, providing the former with additional synthetic information about firms and enabling the latter to communicate their social commitment [43]. To investors, information about index reconstitutions signals changes in the level of that commitment, which can be either ignored or acknowledged and interpreted. Thus, such signals may either provoke no reaction or, depending on interpretation, elicit favorable or unfavorable responses reflected in positive or negative abnormal returns on the company’s stock. As a result, analysis of abnormal returns can be used to determine whether the market treats CSP as relevant and whether it perceives social commitment as beneficial or disadvantageous. This kind of analysis has been applied in event studies designed to identify the short-term financial impacts of CSP-related announcements.

The results of previous research concerning investor responses to sustainability index additions and deletions are inconclusive. Existing discrepancies imply the presence of some additional factors affecting investor reactions. It has been noted that investor evaluations of sustainability are influenced by time and space [15], the overall market conditions [2,18], the risk associated with the institutional environment [19], the properties of the sector in which the company is active [7,12], and the characteristics of the individual company itself [20,21,44]. In this study, it was assumed that investor reactions may be also affected by the design of the ESG index, which provides the context for interpreting the signals generated by its reconstitutions. Indeed, depending on index design, investors may consider changes in the evaluation of a company’s CSP, reflected by an index addition/deletion, either from the perspective of improving/deteriorating legitimacy or in terms of reputation enhancement/damage (Figure 1). The social commitment of a company plays a major role both in acquiring legitimacy and building reputation [45].

2.2. Legitimacy and Reputation Effects of ESG Index Design

Legitimacy may be considered from three perspectives: as a property (a resource or asset), as a perception (an evaluation or social judgement), or as a process [27]. The last case, however, does not concern legitimacy as such, but rather the process of acquiring or losing it (legitimization/delegitimization). According to the classical definition by M.C. Suchman, legitimacy is “a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions” [26] (p. 574). It follows then that legitimacy can be acquired as a result of conduct conforming to acceptable social norms and values, which constitute the only criteria in company evaluation. In this sense, legitimacy is of a non-competitive nature, as it does not require comparison between companies. The entire process of (de)legitimization takes place between the company and the evaluator(s).

The literature offers various typologies of legitimacy [46], with the most popular one making a distinction between moral, pragmatic, and cognitive legitimacy [26]. The first type is based on normative evaluation and concerns the contribution made by the company to the common good as well as the perceived compliance of its conduct with ethical standards. In turn, pragmatic legitimacy arises from evaluation as to how well the company serves the interests of its stakeholder groups. Finally, cognitive evaluation is associated with perceived compliance of the company’s image with expectations about it as a specific kind of organization, facilitating an understanding of its purpose. According to Aldrich and Fiol [47], the highest form of cognitive legitimacy is being taken for granted. Among the three aforementioned types, moral legitimacy requires the broadest consensus since, as noted by Koppell, “without agreement among those who would judge an institution, moral legitimacy is effectively impossible” [48] (p. 182).

The addition of a company to an ESG index bolsters its moral legitimacy, because of the implication that its conduct has been assessed to be consistent with socially acceptable values and norms. However, the strength of the legitimacy effect differs depending on the design of the index, which has a bearing on the extent to which it signals the company’s compliance with the fundamental conditions of moral legitimacy. Melé and Armengou [49] indicate four such conditions: (1) the intended effects of the organization’s activity contribute to the common good; (2) the means and procedures employed by the organization are ethical; (3) the organization engages in ethical evaluation of its activity accommodating stakeholder concerns and needs; (4) the organization engages in ethical evaluation of the foreseeable consequences of its activity and minimizes the associated possible damage or risks. Indices with a design that can more successfully signal the included companies’ conformity with these criteria generate a stronger legitimacy effect.

An important factor reinforcing legitimacy is isomorphism, understood as the growing similarity of organizations in a given sector [50]. This can take place as a result of mimetic processes [51], which reveal a connection between legitimacy and homogenization. Depending on their design, ESG indices may enhance or discourage such processes by strengthening or weakening the impact of index addition information on legitimacy. Legitimacy effects will be more pronounced for indices whose design facilitates imitation, and weaker for those whose design does not allow companies to use imitation as a strategy for remaining in the index.

The design of sustainability indices has important implications not only for generating a legitimacy effect, but also for generating a reputation effect. The latter is quite distinct from the former due to the different natures of the underlying constructs. Literature reviews of reputation definitions [29,52] indicate that, similarly to legitimacy, it may be analyzed from different perspectives: as awareness (visibility, prominence in the collective perception), as perceived predictability, as an assessment (judgement), or as an asset (resource). Although some of these perspectives (assessment and asset) are also present in legitimacy theory, this does not imply an affinity between the two concepts (constructs). As far as assessment is concerned, legitimacy and reputation are dissimilar in terms of their nature, context, and effects. Reputation arises from an evaluation of the organization and its activity in different areas based on a consistent pattern of those activities [53], enabling inferences about the organization’s characteristics [31,54]. This evaluation is relative [55], with the point of reference being other organizations and their performance. Consequently, a reputational assessment determines the organization’s place within an established hierarchy—the better the reputation, the higher the position [56]. Thus, reputation is by nature competitive [57] because it is acquired at the cost of other organizations defeated in the struggle for advancing in the hierarchy. This gives rise to some critical differences between legitimacy and reputation. The former requires only two actors: an organization and an evaluator, while the latter requires three actors: an organization, an evaluator, and other organizations as points of reference. A legitimacy judgment is absolute in the sense that it mostly concerns the compliance of the organization’s conduct with social values, norms, and ethics. As such, it is not competitive and does not engender hierarchy. Stronger legitimacy effects will be found for those sustainability indices whose design does not involve comparisons between companies, while stronger reputation effects will be generated by indices which include only the top-rating companies from a hierarchy developed as a result of the evaluation of companies conducted in the process of index revision.

Legitimacy and reputation differ not only when considered in terms of evaluation, but also in terms of the other shared perspective—that is, as an asset (resource). Good reputation generates support [58], and so it constitutes a significant component of the company’s intangible assets [59], providing a foundation for a sustainable competitive advantage [33]. Indeed, reputation is characterized by all the features given by Barney [32] for sources of competitive advantage: it is valuable, rare, imperfectly imitable, and not readily substitutable. In addition, it cannot be purchased in the market [60]. The role of reputation in building a sustainable competitive advantage makes it distinct from legitimacy. The latter cannot contribute to such an advantage as it is not rare; on the contrary, it is needed by all organizations to survive [61]. Reputation is also important in that it can be employed in the company’s strategy, which is determined by the pool of its resources [34]. In this context, good reputation is considered to be a factor enabling the company to differentiate [35], in contrast to legitimacy, which is conducive to homogenization due to its imitative mechanism. A synthetic characterization of differences between legitimacy and reputation is given in

Table 1. Comparison of legitimacy and reputation.

Perspective	Differentiating Factor	Legitimacy	Reputation
Legitimacy and reputation as an evaluation	Point of reference (context)	Norms, values, ethics (contribution to the common good)	Other organizations and their performance
	Type of evaluation	Absolute	Relative
	Stratification	Does not lead to a hierarchy	Place in the hierarchy is decisive
	Key mechanism	Conformism	Competition
Legitimacy and reputation as a resource	Features of the resource	Valuable, imperfectly imitable, and not readily substitutable, but not rare	Valuable, imperfectly imitable, not readily substitutable, and rare
	Ability to confer a competitive advantage	Does not confer a competitive advantage	Confers a competitive advantage
	Market effect (strategic consequences)	Leads to homogenization	Enables differentiation

.

The differences between legitimacy and reputation judgments are crucial in terms of the effects exerted by ESG indices of different designs. Among the three basic index types, i.e., negative screening, positive screening, and best-in-class, the first category exhibits the strongest legitimacy effects as its design has the greatest affinity to the characteristics of that social construct. Negative screening leads to a priori exclusion of companies in controversial sectors, such as tobacco, alcohol, gambling, arms, coal, ect., whose activity is perceived to violate social norms and values by at least part of the public; from a social perspective, they at best do not contribute to the common good, and may detract from it in the worst-case scenario. The design of the other types of indices does not create a barrier to such companies. In contrast to best-in-class indices, the selection mechanism of negative screening indices is based on the results of absolute evaluation of companies in terms of ESG criteria, with all candidates exceeding a certain threshold being included. As a result, the design of such indices does not give rise to hierarchy and their composition is not determined by competition between companies. To be added to such an index, it is sufficient to imitate the companies that have already been included. On the other hand, the strongest reputation effect occurs for companies belonging to best-in-class indices. Their selection mechanism is based on comparing ESG evaluation results between candidates, which gives rise to a hierarchy, with the company’s position being determined by the obtained score. Best-in-class indices include only the top-ranking companies that have outperformed their rivals.

These mechanisms are clearly visible when one compares the designs of ESG indices belonging to the two main families, that is, the FTSE4Good Index Series (including FTSE4Good US), based on negative screening, and Dow Jones Sustainability Indices (including DJSI North America), which use best-in-class methodology. In the former case, company addition decisions are based on two factors: ESG criteria and a lack of involvement in controversial sectors. Thus, all candidate companies from a given pool will be added to the index as long as they pass a certain threshold (for FTSE4Good US it is 3.3 on a scale from 0 to 5) and meet the second criterion. In turn, companies will be deleted when they cease to meet either condition [62]. The design of DJSI indices is very different. In this case, company additions and deletions are conditional on the outcome of RobecoSAM’s corporate sustainability assessment process [63]. As part of this process, a total sustainability score is calculated for all participating companies, which is the basis for their ranking within their respective sectors. Depending on the index, between 10% and 30% of the top-rated companies are included per sector (in the case of DJSI North America, it is 20%). The indices are revised annually, and so companies may be deleted if they fall behind their competitors over time. A synthetic characterization of the relationship between index design and legitimacy and reputation effects is given in

Table 2. Relationships between index design and legitimacy and reputation effects.

Index Type	ESG Index Used in Our Research	Effects Generated by Indices
Panel A
Indices based on a negative screening	FTSE4Good US Index	Legitimacy effect
Companies from sectors considered as not contributing to the common good and/or those that violate current social norms and values are excluded a priori.	Companies manufacturing tobacco, weapons, and coal are excluded from the index a priori, as they are companies identified as being embroiled in significant controversy.	Contributing to the common good and behaving consistently with socially acceptable values and norms boosts moral legitimacy.
The selection mechanism is based on the results of absolute evaluation of companies in terms of ESG criteria; ratings obtained by other companies are irrelevant.	The ESG rating level required of each included company is 3.3 or more; companies’ scores are not cross-compared.	As companies are not benchmarked against each other, the design of such indices does not give rise to a hierarchy.
The selection mechanism is not based on competition between companies—all candidates exceeding a certain threshold are included.	All companies that pass the eligibility criteria are automatically included in the index; the number of constituents is not limited—additional inclusions pose no threat to companies already in the index; exclusions occur only when a company engages in restricted activities, becomes the subject of significant controversy, or its ESG assessment results decline below the threshold.	The lack of a competition mechanism promotes mimetic processes.
Panel B
Indices following a best-in-class approach	DJSI North America Index	Reputation effect
Companies are not excluded a priori on account of the profile of their business.	The universe of companies under consideration consists of the largest 600 US and Canadian companies included in the S&P Global BMI, and therefore also includes entities operating in controversial sectors.	Neither norms nor values are a reference point for pre-selection.
The selection mechanism is based on the cross-comparison of companies’ scores within their respective sectors, and so it is relative in nature.	The Total Sustainability Score is calculated for each company under an annual Corporate Sustainability Assessment; this score is the basis for ranking companies within their sectors. Therefore, a company’s position in the hierarchy depends not only on the results of its evaluation, but also on the scores of other companies in the sector.	Comparing assessment results makes evaluation relative and gives rise to a hierarchy.
At the core of the selection mechanism is competition, as only the top scorers within their respective sectors are included in the index.	Only 20% of companies with the highest ratings within a sector are included in the index; at the next revision of index composition, a company maintaining the same score, which previously secured it a place in the index, may be excluded from it if its competitors’ assessments have improved.	The selection mechanism is based on competition driven by constant rivalry for a position within the hierarchy.

.

Company inclusions in such an index signal an increase in their resource pool, while deletions signal a reduction in that pool. However, investor reactions to these signals will depend primarily on whether they perceive that particular resource as a substantial factor in the company’s superior performance in the future. Therefore, for investors, the differences between legitimacy and reputation as resources play a pivotal role. Not being a rare resource, the former cannot contribute to competitive advantage. Additionally, the mechanism of its acquisition, based on conformism and imitation, stands in contradiction to the possibility of radical differentiation from competitors. Thus, information about an index addition generating a legitimacy effect will be difficult to interpret for investors in terms of its impact on the future value of the company. Similarly, a company’s deletion from the index, while weakening its legitimacy, does not lead to delegitimization jeopardizing its survival.

In contrast, reputation is a rare resource, and as such it may provide the basis for a differentiation strategy. Resources giving rise to a sustainable competitive advantage are of crucial importance to investor assessment of the company’s future value, as such an advantage enables the company to reap higher profits in the long term [36]. Companies with a good reputation are characterized by superior financial performance [64], and the ability to maintain it over a long period of time [65]. Consequently, information about the reconstitution of indices generating stronger reputation effects will be more straightforward and more readily interpretable for investors.

H1. 

The design of ESG indices differentiates the reactions of investors, who respond to reputation-relevant reconstitution information, but not to announcements affecting company legitimacy. Investor reactions to the additions/deletions of companies are reflected in abnormal returns on their stocks in the case of best-in-class indices, but not in the case of negative screening indices.

A good reputation provides a company with a number of benefits arising from its impact on stakeholder behavior. It has been shown to affect customers’ intention to buy [40], acceptance of higher prices [39,66,67], lower sensitivity to price increases [68], positive attitudes towards CSR advertisement [69] and higher acceptance of advertising claims [70], as well as attractiveness for job applicants [71]. In addition, a good reputation reduces the company’s risk by lowering the probability of adverse events that could harm its value and by limiting any resulting damage. Importantly, the greatest negative impact is brought about by events whose causes are internal and arise from the company’s ill will, since responsibility for them may be fully attributed to the company [72]. A good reputation reduces the probability of such events, as it modifies the structure of incentives, discouraging opportunistic behavior. In other words, an initial investment in good reputation—an intangible asset—deters behaviors which could be detrimental to it, and which could cause the loss of capital invested in its creation [73].

A good CSR reputation also protects the company and its value via the halo effect [74], whereby responsibility for negative events tends to be attributed to factors beyond the control of the company [75]. The non-attribution of guilt lessens the negative impact of such events on the company, as stakeholders are less inclined to punish it or take actions that would harm it. Thus, a good reputation serves as a protective shield or buffer and can be treated by investors as a form of insurance [76,77].

On the other hand, in accordance with the theory of asset stock accumulation [78], the creation of resources ensuring a sustainable competitive advantage (such as a good reputation) requires appropriate prior flows, meaning consistent investment in CSR. It should be noted that further flows will be needed after those resources are already in place. In a situation where CSR is a playing field for companies vying for a position in the hierarchy of social recognition, a good reputation cannot be maintained without continuously advancing the organization’s CSR policy. This leads to a Red Queen effect, whereby corporations must run just to stand still [79], which may raise concerns about the increasing costs of maintaining a good reputation.

H2. 

Investors attach greater importance to the profits arising from a good reputation than to the costs entailed by the need to reinforce it. They will exhibit positive responses to company additions to a best-in-class index and negative responses to deletions from such an index, as measured by abnormal returns.

3. Materials and Methods

Investor reactions to the reconstitutions of ESG indices affecting company legitimacy and those generating a reputation effect were investigated using the FTSE4Good US Index, which uses negative screening, and the DJSI North America Index, which follows a best-in-class approach, respectively. In both cases, it was examined whether and how investors reacted to information about company additions and deletions. In order to avoid systematic error due to the impact of current market trends on investor decisions, data were collected from the years 2009–2019. The study period began after the greatest turbulence of the 2008 financial crisis had already subsided and ended before the onset of the COVID-19 pandemic turmoil. The indices included in the study differ not only in their structure, but also in the territorial scope of markets: FTSE4Good US includes listed companies only from the United States, while DJSI North America comprises listed companies from both the United States and Canada. Therefore, in the latter case, information on companies from Canada was excluded in order to ensure data comparability between the indices.

In the current analysis, we follow a methodological approach proposed in previous research [19], and involving the event window method, which has been widely applied in studies examining investor reactions to information concerning both entire markets and individual companies [80]. To date, event windows have been used to study the impact of both negative [81] and positive information [82] of either financial [83,84,85] or non-financial nature, concerning, e.g., company reputation [86] as well as corporate social responsibility or irresponsibility [87]. Event window analysis has also been applied to company additions and deletions from sustainability indices [10,11,21,88,89,90,91]. Such a wide range of research proves the usefulness of the method and indicates that it can be successfully employed to verify the presented hypotheses.

The event window method identifies the economic effects of an event by estimating its impact on share prices as reflected in the rate of return. In the present study, similarly to other research [2,90], an event was defined as an announcement of company addition or deletion from the considered ESG indices. The dataset encompassed all additions and deletions from the 42 periodic reviews carried out by the index operators throughout the study period. In the years 2009–2019, the FTSE4Good index was reconstituted 31 times, while DJSI was reconstituted 11 times (the former index is revised quarterly, and the latter one is revised annually). A total of 744 events were identified. Company deletions due to bankruptcy and mergers, and those for which historical data could not be found were rejected (50 cases in total). Moreover, three data points were excluded as outliers; in those cases, the cumulative abnormal return on a company’s shares was more than four times greater than the standard deviation of the cumulative abnormal returns calculated for all companies within a given event class (addition/deletion) and a given index (FTSE4Good US Index/DJSI North America). The final sample consisted of 691 events, including 441 company additions and 250 deletions. Detailed information on the number of events is presented in

Table 3. Number of data points by index.

Index	Number of Index Reconstitutions	Total Number of Events a	Number of Ineligible Events b	Number of Outliers	Percentage of Rejected Events	Number of Eligible Events	Number of Company Additions	Number of Company Deletions
FTSE4Good US	31	357	15	2	5%	340	242	98
DJSI North America	11	387	35	1	9%	351	199	152
Total	42	744	50	3	7%	691	441	250

^a Excluding companies listed in the Canadian market. ^b Firms deleted for reasons other than failure to meet the ESG criteria (e.g., M&A, delisting) and missing data.

.

We checked whether the two indices happened to add or delete the same companies in particular periods. Over the entire 11-year period of analysis, only 19 such instances were identified (16 additions and 3 deletions). In the next step, we determined the length of the event window. Although the shortest possible windows are generally recommended to minimize the impact of confounding events on share prices [92], previous studies vary greatly in terms of event window length. Furthermore, while most authors have adopted symmetrical windows, asymmetrical ones have also been used. In situations where the analyzed events involved new market information and immediate changes in asset prices resulting from investor reactions to this information, the day preceding official announcements was often included to account for possible information leaks. Given the absence of a single, commonly shared view on both event window length and shape, the present study adopted an asymmetrical five-day window starting on the day preceding announcements of sustainability index reconstitutions (<−1; +3>). Such a window should be deemed short enough to eliminate confounding variables [92], while allowing information about index reconstitutions to reach investors and influence their decisions, triggering a reaction.

Investor reactions to information about ESG index reconstitutions were examined by measuring the cumulative average abnormal return (CAAR). Assuming informational market efficiency, the occurrence of an abnormal rate of return should be attributed to the analyzed events. For the purpose of calculating CAAR, an abnormal rate of return was estimated for each added/deleted company for each of the five days comprising the event window. The expected value was estimated by the Capital Assets Pricing Model (CAPM) [93,94]. The base interest rate of the Federal Reserve System was adopted as the risk-free rate, the S&P500 index was used as a reference market portfolio, while coefficient β was estimated on the basis of data from 60 days preceding the event. The obtained daily abnormal rates of return were added up to compute cumulative abnormal rate of return for each company throughout the event window. In the last step, we calculated the arithmetic means of rates of return separately for the groups of companies added to and deleted from sustainability indices. In the first stage of the study, devoted to the verification of H1, we examined whether or not investors reacted to information about the reconstitutions of either index, but without determining the character of their reactions (positive/negative). The occurrence of investor reactions to information about company additions and deletions was verified by evaluating the significance of differences in CAAR values between added and deleted companies. Two-tailed tests, both parametric (t-test, Levene’s F-test) and non-parametric (Mann–Whitney’s U-test), were used to check whether the CAAR for the added companies was equal to that for the deleted ones, with the alternative hypothesis being that they differed. The results of those calculations are given in

Table 4. Tests for significance of differences in CAAR values between companies added to and deleted from ESG indices.

FTSE4Good US Index	DJSI North America Index
n1 = 242 (added companies)	n1 = 199 (added companies)
n2 = 98 (deleted companies)	n2 = 152 (deleted companies)
Levene’s F = 19.998 ***	Levene’s F = 0.094
t = −1.464	t = 2.154 **
Mann–Whitney’s U = 0.490	Mann–Whitney’s U = −1.945 *

*** p < 0.001; ** p < 0.01; * p < 0.05.

.

The objective of the next stage of the study was to verify H2 by determining how investors reacted to information about the reconstitutions of an index generating a reputation effect (DJSI North America). The occurrence of positive reactions to company additions and negative reactions to deletions was verified using a one-tailed test to check for CAAR values that would be significantly higher than zero in the former case and lower than zero in the latter case. Also, the parametric t-test and the non-parametric z-test were used, similarly as in other studies on investor reactions to sustainability index reconstitutions [2,4]. The results are shown in

Table 5. Comparison of CAAR values for companies added to and deleted from the DJSI North America Index between 2009 and 2019.

Added Companies (n = 199)				Deleted Companies (n = 152)
CAAR	t-Test	z-Test	Standard Deviation	CAAR	t-Test	z-Test	Standard Deviation
0.32	1.712 **	1.914 **	2.618	−0.28	−1.372 *	−1.298 *	2.474

** p < 0.01; * p < 0.05.

.

Further analyses were conducted in order to ensure the robustness of results. Similarly as in other studies concerning investor reactions to sustainability index reconstitutions [2,4,12], an alternative method of determining CAAR was applied. In estimating the expected rate of return, which is the basis for CAAR calculations, the CAPM model was replaced with the market model (MM) [93], which has been used quite extensively by other authors [7,11,12]. The sensitivity of the presented results to changes in the parameters of the CAPM model was tested by replacing a broad market index (S&P500) with a narrower one, covering only blue chips (the Dow Jones Industrial Average, DJIA). Furthermore, the estimation period for β was extended by increasing the number of observations for that coefficient. Additionally, to elucidate how our results would be affected by different event window lengths and to identify any tendencies and patterns—if any—in investor reactions, we supplemented our analysis with two shorter and two longer event windows: <−1; +1>, <−1;+2> and <−1;+4>, <−1;+5>. Following those modifications, the statistical analysis involved tests for the significance of differences in CAAR values between companies added to and deleted from each of the analysed ESG indices (

Table 6. Tests for significance of differences in CAAR values between companies added to and deleted from ESG indices (based on the recalculated abnormal rate of return).

	FTSE4Good US Index	DJSI North America Index
Type of Modification	n1 = 242 (Added Companies) n2 = 98 (Deleted Companies)	n1 = 199 (Added Companies) n2 = 152 (Deleted Companies)
MM	Levene’s F = 20.225 *** t = −1.415 Mann–Whitney’s U = 0.713	Levene’s F = 0.217 t = 1.874 * Mann–Whitney’s U = −1.456
DJIA	Levene’s F = 17.282 *** t = −1.365 Mann–Whitney’s U = 0.535	Levene’s F = 0.197 t = 2.494 ** Mann–Whitney’s U = −2.452 **
Beta	Levene’s F = 16.954 *** t = −1.483 Mann–Whitney’s U = 0.538	Levene’s F =0.180 t = 2.085 ** Mann–Whitney’s U = −1.812 *
Event window
<−1; +1>	Levene’s F = 17.335 *** t = −1.247 Mann–Whitney’s U = 0.705	Levene’s F = 0.043 t = 0.712 Mann–Whitney’s U = −0.351
<−1; +2>	Levene’s F = 26.458 *** t = −1.350 Mann–Whitney’s U = 0.228	Levene’s F = 0.048 t = 1.272 Mann–Whitney’s U = −1.039
<−1; +4>	Levene’s F = 20.444 *** t = −1.389 Mann–Whitney’s U = 0.491	Levene’s F = 0.191 t = 1.775 ** Mann–Whitney’s U = −1.417
<−1; +5>	Levene’s F = 30.699 *** t = −1.125 Mann–Whitney’s U = 0.503	Levene’s F = 0.000 t = 1.875 ** Mann–Whitney’s U = −1.494

*** p < 0.001; ** p < 0.01; * p < 0.05.

). Finally, CAAR values were compared for companies added to and deleted from DJSI North America (

Table 7. Comparison of CAAR values for companies added to and deleted from the DJSI North America Index (based on the recalculated abnormal rate of return).

	Added Companies (n = 199)				Deleted Companies (n = 152)
Type of Modification	CAAR	t-Test	z-Test	Standard Deviation	CAAR	t-Test	z-Test	Standard Deviation
MM	0.36	1.828 **	0.496	2.772	−0.18	−0.879	−1.298 *	2.575
DJIA	0.39	2.092 **	1.772 **	2.623	−0.30	−1.487 *	−0.3244	2.520
Beta	0.30	1.606 *	0.922	2.632	−0.27	−1.377 *	−1.460 *	2.478
Event window
<−1; +1>	0.06	0.373	0.496	2.184	−0.11	−0.622	−0.162	2.167
<−1; +2>	0.17	1.025	0.922	2.405	−0.16	−0.792	−0.162	2.484
<−1; +4>	0.24	1.125	1.489 *	3.069	−0.33	−1.388 *	−0.487	2.954
<−1; +5>	0.26	1.105	0.780	3.331	−0.42	−1.512 *	−1.136	3.425

** p < 0.01; * p < 0.05.

).

4. Results and Discussion

Investor reactions to ESG index reconstitutions were found to differ depending on index design, which indicates that investors in the U.S. market treated information about events affecting legitimacy and reputation differentially. Investor reactions to reconstitutions of the negative screening index were characterized by ambivalence: neither legitimacy-enhancing additions nor legitimacy-damaging deletions affected investor behavior in a definitive way. Accordingly, no statistically significant differences in CAAR values were found between companies added to and deleted from the FTSE4Good US Index (Table 4). On the other hand, reconstitutions of the studied best-in-class index generating a reputation effect were clearly reflected in investor decisions, as shown by the statistically significant difference in CAAR values between companies added to and deleted from the DJSI North America Index (Table 4). In other words, it appears that investors in the U.S. market responded only to index reconstitutions generating a reputation effect.

Besides verifying the occurrence of a reputation effect, the present study aimed to determine investor reactions to changes in reputation caused by company additions and deletions from a best-in-class sustainability index. The data in Table 5 show that investors in the U.S. market rewarded companies for improving their reputation and punished them for damaging it by reacting positively to company additions and negatively to deletions. In the case of DJSI North America, CAAR values were +0.32% for added companies and −0.28% for deleted ones, with the results being statistically significant. The CAAR values for added and deleted companies also indicate that investors responded more strongly to positive reputational data (additions to a best-in-class index) than to negative reputational data (deletions from such an index).

The robustness test showed that the present results are not sensitive either to the CAAR calculation method or to changes in the CAPM model parameters. The pattern of investor reactions to the reconstitutions of the two studied indices remained unaffected by replacing the CAPM model with a market model, using a narrower market index in the CAPM model, or extending the β estimation period. Also expanding our analysis to include additional event windows did not substantially modify the obtained investor reaction patterns. In the case of reconstitutions of the index generating a legitimacy effect, no statistically significant difference was observed between the CAAR values for included and excluded companies. On the other hand, such differences were found for the reconstitutions of the index generating a reputation effect in the <−1; +4> and <−1; +5> windows, which were longer than the basic <−1; +3> window adopted in this study. The absence of statistically significant results for the <−1; +1> and <−1; +2> windows may be explained by insufficient time after announcements (Table 6). In the case of DJSI North America, investors reacted favorably to information generating a positive reputation effect and unfavorably to reputation-damaging data. In the robustness test, CAAR values for companies added and deleted from DJSI North America continued to be positive and negative, respectively (Table 7). Moreover, similarly as in the main study, investors still reacted more strongly to positive reputational events.

The results of existing research on investor reactions to sustainability index reconstitutions are ambiguous. The considerable discrepancies in previous results have prevented the identification of definitive patterns of investor reactions to those events. The present paper shows that the discrepancies may be partially attributed to differences in investor responses arising from index design. It should be noted that few authors to date have dealt with negative screening indices. The study of Curran and Morran [89], covering a very short period of time (2000–2002), identified a positive investor response for companies added to the FTSE4Good index and a negative response for those deleted, albeit the results were not statistically significant. The paper by Doh et al. [10], encompassing a somewhat longer period (2000–2005), reported negative investor reactions to company deletions from the Calvert Social Index in the absence of a directional response to additions. These results are partially consistent with the present findings. In our study, spanning a much longer timeframe (2009–2019), it was found that neither additions to nor deletions from the FTSE4Good US Index elicited a definitively positive or negative investor response. Discrepancies in the results concerning deletions may be associated with differences in periods of study; while the years 2000–2005 were characterized by high volatility in the overall economic conditions, featuring both bull and bear market phases, favorable market conditions prevailed throughout the entire 2009–2019 period. This is significant as it has been suggested that both time [15] and economic conditions [2,18] may modify investor reactions to sustainability index reconstitutions.

Existing studies have made much more extensive use of best-in-class indices. Most of those studies identified positive reactions to company additions [2,37], often combined with negative reactions to deletions [4,5,20,44]. Negative investor reactions to company additions were found only by Joshi et al. [12] and Cheung and Roca [8], who also observed positive reactions to deletions. Two studies reported little or no response to index announcements [14,15], while Oberndorfer et al. [90] showed mixed results. This last study is of particular interest from the standpoint of the present paper as it compared investor reactions to the reconstitutions of two indices, i.e., DJSI World and DJSI Stoxx. Investors responded to changes in the composition of the former index only, which Oberndorfer et al. [90] attributed to the stronger signaling effect of DJSI World, which enjoys greater recognition than DJSI Stoxx. Those findings also indicate that investor reactions may be affected by the attributes of an index, which is consistent with the present results. However, while Oberndorfer et al. [90] examined two indices belonging to the same family and exhibiting a similar design (best-in-class), our study adopted a different approach affording a more comprehensive perspective on the properties of ESG indices. By selecting indices with fundamentally different designs, we were able to determine ways in which index design affects investor behavior and distinguish between legitimacy and reputation effects.

The present results, showing positive investor reactions to company inclusions in a best-in-class index, are in line with the results of most previous studies analyzing market responses to the reconstitutions of such indices. Similarly to Consolandi et al. [5], Cheung [4], Hawn et al. [20], and Stekelenburg et al. [44], we found that positive reactions to additions were accompanied by negative reactions to deletions. The discrepancies in results reported from the remaining studies involving indices of the same type [8,12,14,15,90] show that index design is not the only factor differentiating investor behavior. The study of Cheung and Roca [8] and Oberndorfer et al. [90] focused on companies listed in Asia Pacific exchanges and German exchanges, respectively, and so their results are likely to be influenced by the local circumstances. In turn, the studies by Durand et al. [14], Hawn et al. [15], and Joshi et al. [12] covered periods characterized by different overall market conditions, which may have affected their findings. Research on investor reactions to sustainability index reconstitutions has also involved some positive screening indices [6,16,17,21,95] and indices of a more complex design, combining negative screening with elements of the best-in-class methodology [7,11,18,96,97]. The legitimacy and reputation signals generated by such indices are more ambiguous, and so less readily interpretable by investors. In this case, investors may attach greater importance to the properties of those indices such as years active and the reputation of the index operator. Due to these properties, some of these indices may also generate a reputation effect despite not having a best-in-class design; in particular, this may hold true of indices of long standing, such as the Domini 400 Social Index, or those published by high-profile, reliable providers.

Our findings provide new and valuable insights into the role of intangible resources as mediators of the CSP-CFP relationship [98]. According to reputation theory, a good reputation brings about economic benefits [64,65,99], while its deterioration leads to negative consequences [100]. In this respect, our results are in line with those of previous observations, but they do not support claims that investors pay much attention to company legitimacy. This is a new finding: although legitimacy theory posits that this intangible asset also confers economic benefits [38], they are not sufficiently important for investors to affect their decisions. This group of company stakeholders do not put legitimacy on an equal footing with reputation, attaching greater weight to the latter. As a result, reputation plays a far greater role as a mediator of the CSR-CFP relationship. This observation sheds new light on the mechanisms underlying the influence exerted by the operators of ESG indices on investor decisions. According to Doh et al. [10], this influence is attributable to the legitimacy-conferring function of these operators, while Lackmann et al. [2] and Ramchander et al. [7] put it down to the higher reliability and credibility of CSR performance information validated by the index operator’s decision. The present results do not bear out those claims. Despite the high levels of recognition and trust enjoyed by both index publishers, in our study investors reacted only to the reconstitutions of the best-in-class index, generating a reputation effect. Thus, investor behavior was not affected merely by the operator’s decision to include/exclude a company. Rather, the market responded only to information which, by virtue of index design, validated a company’s CSR leadership (or loss of it).

5. Practical Implications

Our study has some important practical implications for companies that wish to communicate their high CSR standards via participation in ESG indices. The proliferation of such indices raises the question of their effectiveness in CSR signaling. Managers are faced with the dilemma of whether their companies should strive to be featured in all indices or only some of them—and, if so, which ones. Slager and Gond [101] found that managers attempt to capitalize on the diversity of ratings. Absolute benchmarks such as the FTSE4Good Index are perceived as “easier to meet” compared to best-in-class indices. Our findings indicate, however, that inclusion in a best-in-class index is more advantageous for a company, as it generates a positive reputational effect. Good reputation is a resource that may contribute to the company’s competitive advantage, thus increasing its attractiveness to investors, who will perceive an index addition as a predictor of value growth. However, inclusion in a best-in-class ESG index is also associated with some risks since the applied company selection mechanism necessitates continuous competition. This gives rise to a Red Queen effect, whereby the company has to constantly further its CSR standards to outdo its sectoral rivals in a race to remain in the index. Maintaining social commitment at a constant level may not be sufficient and could lead to a deletion from the index if competitors improved their performance. That would in turn lead to a negative reputation effect, reducing the company’s attractiveness to investors.

Our results—based on longer event windows—also indicate that negative investor reactions to company deletions from the index generating a reputation effect grow and become stronger than positive reactions to company additions. This may imply that investors attach greater importance to the weakening of company reputation. This gives rise to new challenges. On the one hand, a superior reputation leads to higher expectations on the company, while on the other hand falling short of those expectations or the occurrence of events tarnishing company reputation present a greater jeopardy to the company, carrying an increased risk of losing value. As a result, the managers of companies signaling their superior reputation by being included in a best-in-class index must pay greater attention to the management of reputational risk.

6. Conclusions

The paper examines the influence of ESG index design on investor decisions, contributing to research on the short-term financial effects of CSR-related announcements. Additions to and deletions from sustainability indices signal changes in the social responsibility commitment of companies. Those signals are released by index operators, who enjoy an institutional mandate. Investor reactions to the signals depend on their interpretations, which are influenced by a number of factors identified in previous studies. The present paper identified an additional, previously unrecognized, factor affecting investor reactions, and so it sheds new light on the underlying causes of discrepancies in findings from other studies on investor reactions to ESG index reconstitutions. Other authors have reported both positive and negative investor reactions to company additions, negative reactions to company deletions, or the absence of any statistically significant reactions to reconstitutions. Most of those studies, however, involved only one sustainability index. Furthermore, no other study to date has tested its findings for robustness to change to change of the ESG index applied. Our results indicate that investors tend to react differently to information about company additions and deletions depending on index design. In the case of the studied negative screening index, both additions and deletions failed to prompt appreciable positive or negative investor reactions. On the other hand, reconstitutions of the best-in-class index did elicit significant responses from investors.

With respect to these findings, it seems plausible that the design of ESG indices, including the adopted mechanism of company selection, affects the way in which reconstitution information is processed. While additions to a negative screening index enhance company legitimacy, additions to a best-in-class index generate a positive reputation effect. The present results indicate that information about company legitimacy and reputation is treated differentially by investors. Legitimacy-relevant additions and deletions do not trigger investor reactions, while reputation-relevant ones do; that is, reputation-enhancing additions elicit favorable responses and reputation-damaging deletions are met with unfavorable responses.

The present study also has some limitations. The investigation was focused on establishing how investors in the U.S. react to information about ESG index reconstitutions. Consequently, it was not possible to determine whether the observed pattern of investor reactions is universal or specific to a particular market and timeframe. It is possible that, in some markets, investors pay more attention to legitimacy, responding more strongly to the reconstitutions of negative screening indices, while in others, they may attach greater importance to the reputation effect conferred by best-in-class index additions. Thus, an interesting direction for further research would be to examine reactions to legitimacy and reputation signals in other markets. It could be particularly enlightening to consider those characterized by institutional conditions different from the U.S., e.g., emerging markets or markets in countries with civil law systems. Another limitation is the fact that, throughout the analyzed period (2009–2019), the U.S. market exhibited an upward trend in share prices;, while it has been posited that economic conditions may influence investor reactions to sustainability index reconstitutions [2]. Therefore, in future investigations, the timeframe of analysis could be extended to encompass varied market conditions.