Sources of Stakeholder Salience in the Responsible Investment Movement: Why Do Investors Sign the Principles for Responsible Investment?

Since its inception in 2006, the United Nations-backed Principles for Responsible Investment (PRI) have grown to over 1300 signatories representing over $45 trillion. This growth is not slowing down. In this paper, we argue that there is a set of attributes which make the PRI salient as a stakeholder and its claim to sign the six PRI important to institutional investors. We use Mitchell et al.’s (Acad Manag Rev 22:853–886, 1997) theoretical framework of stakeholder salience, as extended by Gifford (J Bus Eth 92:79–97, 2010). We use as evidence confidential data from the annual survey of signatories carried out by the PRI in a 5-year period between 2007 and 2011. The findings highlight pragmatic and organizational legitimacy, normative and utilitarian power, and management values as the attributes that contribute most to the salience of the PRI as a stakeholder.


Introduction
As recurring financial crises and financial market instability are prompting a reconsideration of how we invest (Woods and Urwin 2010), there is increased interest from academics and practitioners in responsible investment strategies. Socially responsible investment (SRI) includes environmental, social and governance (ESG) factors in investment decision making. If responsible investment is part of the answer to the troubles of the financial system, it is important to know what causes investors to take this strategy on board. Institutional investors as a group of financial actors with distinct motivations (Jansson and Biel 2011) and barriers (Guyatt 2006) to adopting ESG are worthy of scrutiny. They control the majority of total shareholdings -84 % in the UK (Mallin 2007)-and in most countries account for the overwhelming majority of SRI assets (Jansson and Biel 2011). A wide adoption of ESG by institutional investors would mean real momentum behind the ESG movement (Sandberg 2011).
The United Nations-backed Principles for Responsible Investment (PRI) is a global organization that advances responsible investment with over 1300 signatories representing over $45 trillion (as of Dec. 2014). Because of its size, prominence, and first-mover status, the PRI is likely the most important global responsible investment initiative in existence today (Sandberg 2013;Woods and Urwin 2010). The combined impact of the PRI's activity has firmly put responsible investment on the map (Sievanen et al. 2013). As Sophia Grene wrote in the Financial Times in 2009: ''That sustainability is no longer a niche concept, sitting in the corner with the church groups and green evangelists, can be demonstrated by figures from the UN PRI.'' Why have over 1300 institutional investors decided over the last 9 years to sign the PRI? Why dedicate resources that would otherwise be used to satisfy other stakeholders' claims, to the annual reporting, signatory fees and implementing the six PRI principles? To analyze this question, we use the framework of institutional theory which unpacks the way institutions respond to external pressures beyond their internal mandate (DiMaggio and Powell 1983;Dorado 2005;Greenwood and Hinings 1996;Greenwood et al. 2011;Lounsbury 2007). We will explore the stakeholder relationship between the PRI and investors to find out what makes the PRI's claim salient to these investors. For that purpose, we use a set of attributes first identified by Mitchell et al. (1997) and then expanded by Gifford (2010) as to what makes stakeholders and their claims salient to organizations.
While there has been much speculation, there are currently no other empirical studies as to why investors have chosen to sign these principles. It is the motivation of this paper to shed some light on that question, using an important confidential dataset obtained directly from the PRI drawn from annual surveys over a 5-year period (2007)(2008)(2009)(2010)(2011) of their signatories. This confidential dataset has been provided to us exclusively for the purpose of this study Fig. 1.
To tackle the question why institutional investors sign the PRI, we have to think of the investment firm as an organization that has stakeholders with competing claims. This is a new way to look at investors, as the academic literature to date, usually describes them as stakeholder's of the firm (Brower and Mahajan 2013;Verbeke and Tung 2013). The PRI with its objective of pushing SRI into the mainstream via its signatory body is affected by the institutional investor's adoption or rejection of the principles, making the PRI a stakeholder to investors according to Freeman's (1984) definition of stakeholder. Observing the principles and the commitment to them is the claim that the PRI makes of all institutional investors. Therefore, signing the PRI for the purpose of this paper counts as prioritizing a stakeholder's claim based on its salience. Competing stakeholder claims would be e.g., the clients' or beneficiaries' lack of expressed interest in RI, competing reporting frameworks that would claim the resources otherwise dedicated to the PRI annual reporting and implementation of the principles, or the investment team's entrenchment in their existing mainstream investment approach.
We code 5 years of signatories' responses to the question of why they signed the principles and what they perceive to be the benefits of signing, in search of evidence of the attributes of power, legitimacy, urgency, management values, coalition building, and relative economic size (Gifford 2010;Mitchell et al. 1997) contributing to the salience of the PRI's claim in the eyes of investors.
The paper will first introduce the conceptual background, exploring existing literature on institutional theory and stakeholder relationships in the context of responsible investment. Next, the theoretical framework and its application will be detailed. In the following parts, we describe the dataset and methodology. The final parts of the paper present and discuss the findings. The paper closes with research limitations and conclusions.

Conceptual Background
The PRI is a global investor association founded in 2005 with the backing of the United Nations and the objective of promoting the six PRI and helping institutional investors who sign to implement them. Institutional investors are defined as organizations that manage and invest on behalf of clients and beneficiaries. This includes pension funds, banks, asset managers, and insurers among other types of organizations (Sandberg 2011). Devised by the investment community itself, the principles form a framework for incorporating ESG issues into investment decision making. They are based on the conviction that ESG factors have an impact on the performance of investments (Table 1).
The PRI's voluntary and aspirational nature means that there is a large heterogeneity of ESG advancement among its one thousand-odd signatories. Therefore, being a signatory to the principles is not necessarily synonymous with being a responsible investor. However, the mere act of signing the principles remains worthy of scrutiny, as, in the words of PRI founder James Gifford: ''The important thing is to get people in the tent, for whatever reason. Then once they are in, you can start to inspire change.'' (Gifford 2014, pers. comm., 15 June) This mechanism of gradual decrease with time in the gap between declared policy and practice is documented in academic literature on decoupling (Boxenbaum and Jonsson 2008;Bromley and Powell 2012;Edelman 1992;Hallett 2010;Sauder and Espeland 2009;Scott 2008;Tilcsik 2010), most relevantly in Haack et al.'s (2012) study on the adoption of the Equator Principles that finds decoupling to be a transitory phenomenon. The PRI is not an isolated phenomenon in the contemporary institutional landscape. In the area of responsible investment regional, Social Investment Forums (SIFs) and the United Nations Environmental Program Finance Initiative (UNEP FI) are also important organizations.
Institutional theory has made many valid attempts to systematize the way we look at industry initiatives of similar nature (Dumas and Louche 2011;Gond et al. 2011). Especially the literature on private regulation and multistakeholder initiatives (MSIs) has contributed significantly to this effort (Bartley 2007;Hardenbrook 2007;Mena and Palazzo 2012). It is worth highlighting that the PRI differs from the organizations studied by the MSI literature in that it does not have regulatory characteristics per se as the six Principles in and of themselves are not overly prescriptive.
However, despite their differences in perspective, institutional theory, and stakeholder theory overlap significantly and share a common interest: explaining how organizations ensure survival and growth (Chen and Roberts 2010). External pressures explored in institutional theory and questions of power, legitimacy, and other sources of influence (Markowitz et al. 2008) are also incorporated into the study of stakeholder relationships. In contrast to institutional theory, stakeholder theory with its focus on the firm emerged in opposition to the shareholder view of the firm, according to which the only relevant stakeholder of the firm is the shareholder. Stakeholder theory argues that the tension or alignment between the interests of managers and shareholders is not the only relationship that is relevant to a corporation (Freeman 1984;Donaldson and Preston 1995). From there, stakeholder theory developed and established itself throughout the 1990s with contributions from Goodpaster (1991), Clarkson (1994Clarkson ( , 1995, Donaldson and Preston (1995), Rowley (1997), andFrooman (1999) being among the most prominent ones. Because of its firm-level unit of analysis, it has been used in studies in the field of responsible investment to examine stakeholder relationships between actors, for instance NGOs and the investment community (Guay et al. 2004). Kaler (2006) names one of the functions of stakeholder theory in the field of business ethics as being a tool for understanding CSR. One of the characteristics that make it suitable for that application is that it is managerial (Donaldson and Preston 1995) in the sense that it is useful in directing managers toward the serving of particular interests. For as long as the interests served by organizations, the financial interests of shareholders and owners were primarily explaining that status quo did not present a challenge. However, in an increasingly mobilized society (Power 1997;Strathern 2000) where NGOs, employees, suppliers, and governments are more active in presenting their claims of organizations, how do managers decide which of those claims are worthwhile in terms of dedicating limited resources of time, labor, and capital to them? While in the past investors were seen as a stakeholder, this paper argues that investors themselves are increasingly the object of other stakeholder claims. How do they choose between these competing demands? Here the notion of stakeholder salience theory becomes useful. And it is a variation of that same question that this paper addresses: why do institutional investors choose to satisfy the claim of signing the PRI among many other stakeholder claims addressed at them?
In this paper, we choose to apply Mitchell et al.'s (1997) stakeholder salience model, following Gifford (2010), as one of the most influential frameworks in the literature (Mainardes et al. 2012). There are important advantages to Mitchell et al.'s model that explain its widespread adoption. It is political, acknowledging the organization's position at the intersection of conflicting and unequal interests. It is practical and easily operationalized, offering versatile and relevant categories applicable to varied stakeholder relationships (Tashman and Raelin 2013). It is also dynamic, recognizing that stakeholder relationships are transient and although not exploring that aspect of the relationships extensively, it leaves space for its inclusion (Friedman and Miles 2002;Myllykangas et al. 2010;Mainardes et al. 2012).
Mitchell et al.'s model has been applied extensively to examine an organization's relationship with different stakeholders (Harvey and Schaefer 2001;Parent and Deephouse 2007;Mainardes et al. 2012;Brower and Mahajan 2013;Chang et al. 2014), but it has rarely been applied empirically to focus in detail on a particular stakeholder relationship, which this paper undertakes. This framework has also been further developed by Gifford (2010) specifically in the context of responsible investment, making the extended version a highly adequate Table 1 The six principles for responsible investment (1) We will incorporate ESG issues into investment analysis and decision-making processes (2) We will be active owners and incorporate ESG issues into our ownership policies and practices (3) We will seek appropriate disclosure on ESG issues by the entities in which we invest (4) We will promote acceptance and implementation of the Principles within the investment industry (5) We will work together to enhance our effectiveness in implementing the Principles (6) We will each report on our activities and progress toward implementing the Principles

Theoretical Framework
Mitchell et al.'s stakeholder salience model is a tool both for the identification and the prioritization of stakeholders and their claims (Neville et al. 2011). In this paper, the underlying assumption is that the PRI is a stakeholder from the perspective of investors, with its mission of 'understanding the implications of sustainability for investors and supporting signatories to incorporate these issues into their investment decision making and ownership practices' (PRI website 2014). The PRI's stakeholder claim that management prioritizes over others or not is that an investment institution should sign the principles. As stated in the introduction, it is more common for the investor to be considered as a stakeholder from the perspective of the firm (Kaler 2003). While this is the natural approach in questions relating to corporate strategy, when discussing investment, stakeholder relationships from the investor's perspective become relevant. From the investor's perspective, it is the PRI that is a stakeholder (Freeman 1984). Either its claim is salient to investors and they become a signatory or it is not and they do not sign, following Eesley and Lenox's (2006) action-based definition of salience. Taking this approach contributes to the work within stakeholder theory that has been undertaken to expand our views of stakeholders and stakeholder relationships (Crane and Ruebottom 2011;Rowley 1997). It also follows the lead of scholars who have applied stakeholder theory to different types of institutions, for instance Knox and Gruar (2007) who take the perspective of a nonprofit organization and examine stakeholder relationships from its perspective. A more recent example is a paper by Sobczak and Harvard (2015) discussing the influence of NGOs, government, members, and other stakeholders on labor unions. The theoretical framework is applied to identify to what degree the PRI's stakeholder claim possesses the salience-producing attributes as predicted by Mitchell et al. These attributes are power, legitimacy and urgency.
Mitchell and his co-authors adopt Etzioni's (1964) typology of power. Power concerns the coercive, utilitarian, or normative means that a stakeholder has at their disposal to exert influence on management. Coercive power relates to the use of force, restraint, or violence to achieve the actor's desired outcome despite resistance (Weber 1947). Utilitarian power is based on material resources e.g., shareholders exercise utilitarian power by means of financial reward or punishment i.e., investment or divestment. Using normative power is linked to symbolic resources, such as media attention or reputation.
Legitimacy and urgency apply to how the claim itself is viewed by the management. 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' (Mitchell et al. 1997). Mitchell et al. divide it into individual, organizational, and societal legitimacy where the degree of legitimacy is tied to the perception of an individual, the whole organization, or society's expressed endorsement of the claim as legitimate. Gifford (2010) introduces another moderating factor, which is pragmatic legitimacy. It refers to the business case perspective on the stakeholder's claim. It is determined by the strength of the arguments presented by the stakeholder, and the amount of new information they present to the management.
The final factor presented by Mitchell et al. is urgency. It refers to the degree to which the claim is perceived by management as calling for immediate attention. According to Agle et al. (1999), urgency can be a crucial factor in achieving maximum salience. Urgency has two sources: time sensitivity and criticality, time sensitivity coming from time pressure and criticality the importance attached to the claim by the stakeholder. For example, in Gifford (2010) stakeholders signal criticality by being persistent, assertive and by dedicating significant resources to advancing their claim.
Gifford (2010) expands Mitchell's et al. model, based on its application to a shareholder-company relationship in an engagement context. Besides the additions already mentioned, he expands the framework with four moderating factors of pragmatic legitimacy, management values, the relative economic size of the stakeholder, and coalition building. Gifford also adds a temporal dimension to the relative importance of these factors, in that different sources of salience are added to the equation during the interaction with the stakeholder, rendering the model even less static. This is also the way institutional theory has evolved from its early static view of institutions to more dynamic models that take into account the existence of competing logics (Meyer and Hollerer 2014).
Pragmatic legitimacy is the perceived legitimacy of the stakeholder claim as well as the value and relevance of the new information that the claim brings to the attention of managers that they would not otherwise have considered.
Management values are defined as the degree of overlap between the values expressed through the stakeholder claim and the values of managers. It can moderate salience independently of the attributes proposed by Mitchell et al. (1997).
The size of the stakeholder relative to the company contributes to salience through the increased power and legitimacy that stem from it. For example, a larger shareholder is likely to hold a more significant stake in a smaller company and have more access to governance power as a result. Likewise, they are likely to be a more legitimate and important actor on the market.
The final moderating factor-coalition-building refers to the pooling of resources by stakeholders. A stakeholder coalition has the combined size, resources, legitimacy, etc. of its participants (e.g., coalitions of shareholders, policy makers, or NGOs). It is therefore a moderating factor of power, legitimacy, and urgency.

Applied Framework
The application of stakeholder salience theory to the research question 'Why do investors sign the PRI?' is complicated by the fact that for the purpose of this research, there is no single stakeholder whose claim's salience has specific attributes from the Mitchell and Gifford frameworks. Therefore, the analysis focuses on the salience of the claim rather than a particular stakeholder. The claim to sign the PRI presented to the investor can and is advanced by the PRI itself, but in many cases, it is also advanced by other stakeholders e.g., the trustees, an NGO, a senior manager, and it is in that context that it possesses some of the attributes that contribute to its salience. Freeman (1984) specified that perceived salience can result from the attributes of both a claim and a stakeholder group, as Eesley and Lenox (2006) confirm empirically in a study of stakeholder action on environmental issues. In the case discussed in this paper, therefore, the subject is the claim to sign the PRI but its salience may be attached to different stakeholders. Eesley and Lennox (2006) test the legitimacy and urgency attributes as applied to stakeholder claims in a data driven paper and obtain positive results. However, academic literature has generally focused on the salience of stakeholders rather than their claims. One of the contributions of this paper is to further explore the perceived salience of a stakeholder claim through empirical research.
Below we address each attribute from the Mitchell et al.'s (1997) and Gifford's (2010) framework, and explain how it is applicable to the PRI-investor relationship. Each of these attributes is a potential source of salience of the claim to sign the PRI and in coding our dataset we look for evidence that backs this intuition.
As a voluntary and aspirational framework the PRI itself does not use the means of threat or coercion to influence an investor's decision to sign. Coercive power as a means of increasing the salience of the claim to sign the PRI would have to come from other sources such as regulatory bodies or trustees. A utilitarian, financial incentive is most likely to come from an investor's clients. The desire to attract new clients or to satisfy the demand from existing ones by signing the PRI would be examples of utilitarian power as a source of salience. The reputational benefits and the signaling of an ESG capability involved in becoming a PRI signatory are examples of normative power, as non-material, symbolic incentives to sign the PRI.
Urgency, as the degree to which a stakeholder's claim calls for immediate attention (time sensitivity), has its source in the increasing visibility of the PRI in the investment space and the pressure on investors to join the growing trend of responsible investment. The critical aspect can be found in the persistence of stakeholders asking investors to sign the PRI, and the amount of resources they dedicate to advancing their claim. Legitimacy in an investor-PRI interaction can stem from the legitimacy of the PRI as an organization (organizational legitimacy), of an individual (individual legitimacy) or from the perceived endorsement of the principles by society (societal legitimacy).
Gifford adds to the above the relative economic size of the stakeholder, which in this context becomes the size of the PRI and the growing weight of the AUM of the existing signatories; coalition building, i.e., stakeholders building coalitions to advance the claim of signing the PRI more effectively, for example, the PRI securing the endorsement of the UN; management values which in the case of some investors may already be aligned with those represented by the Principles before signing; and pragmatic legitimacy in the form of a perceived business case behind the stakeholder claim, in this case the materiality of ESG issues in investment.
The temporal dimension of the model highlighted by Gifford is observed both in the importance of different attributes in each year of the sample and in the importance of different attributes over time to signatories who joined in each of the sample years. But empirical evidence is needed to deepen our understanding.

Data and Method
The data for this study is sourced from a confidential dataset of annual questionnaires from 2007 to 2011 on the implementation of the PRI principles. This paper presents an analysis of responses from asset owners, asset managers, and insurers. In the first few years of this period (2007)(2008)(2009)(2010), the questionnaire was obligatory. 1 All responses from signatories who signed beginning in 2011 are voluntary.
The question coded for the purpose of this paper was 'Why did your organization join the PRI?' or 'Please describe the benefits you have enjoyed as a result of signing the PRI.' This question was optional and open ended. Table below gives some information on the question throughout the survey years and the makeup of the respondents ( Table 2).
The question only corresponds directly to the research question in the years 2008 and 2009. The questions from years 2009-2011 are only related, and they approach the decision to sign the principles from a backward looking perspective, evaluating the positive results of the decision. 2 The number of responses to the relevant question remains at around or slightly below 90 % of the total responses to obligatory questions and at around 60 % of the total signatory number throughout the sample years. The geographical distribution of the sample is still overweight in developed regions, with very similar numbers in the general signatory body.
The proportion of asset managers versus asset owners among the respondents gradually shifted over the years, in the sample analyzed as well as the rest of the signatory body. Among asset owners, 60 % of assets were managed internally in 2011. Around half of asset owners are noncorporate pension funds, followed by the second largest group of corporate pension funds accounting for 25 % of respondents in 2011. Among asset managers, mainstream investment managers are responsible for 77 % of the responses. They are followed by dedicated SRI Managers -12 %, and themed fund managers -9 %.
Listed equity and fixed income are the two dominant asset classes accounting for 36 and 49 % of respondents' assets in 2011, respectively. The third largest category (13 %) is cash, commodities, and other assets.
The analytic method used in this paper is content analysis. It is defined as a research technique used to systematically make inferences about the intentions, attitudes, and values of individuals by identifying specified characteristics in textual messages that are assumed to be objective for the purpose of the study (Morris 1994). It allows for the systematic, numeric analysis of a large amount of text in a manner customized to the research that is being conducted.
Content analysis has featured regularly in management literature (Bergh 1993;Butterfield et al. 1996;Buttner 2001;Davy et al. 1992;Molloy et al. 2011), accounting (Beck et al. 2010;Fischer 2010), marketing (Hite et al. 1988;Burnett et al. 1991;Rust and Cooil 1994), and business ethics where it has been often applied to analyze large volumes of company produced publications relating to their ESG policies and activities (Holder-Webb et al. 2009;Jose and Lee 2007;Lefebvre and Singh 1992;Sheth and Babiak 2010;Stray 2008). It is also the dominant research method for collecting empirical evidence in the field of social environmental accounting (Parker 2005). See Appendix 3 for a more detailed analysis of manual content coding.
Two reasons we chose manual coding over automated coding are specifically related to this PRI dataset. First, the data was collected from organizations from over 30 countries. The differences in wording between even a handful of countries can be hard to manage in computerized coding, so we decided it would not be effective in a study with such broad geographical scope to correctly identify the entire variety of phrases for automated coding.
Second, responsible investment is an emergent field and does not have an established vocabulary (Sandberg et al. 2009), which makes discrepancies even more likely to obscure results when using automated coding. Based on this, the authors agreed that to obtain the most accurate results, we should use manual coding where the categories answers fall into are decided on a case by case basis. As observed by George and Louise Spindler (1997), 'only the human observer can be alert to divergences and subtleties that may prove to be more important than the data produced by any predetermined categories of observation or any instrument' (p. 66-67). To ensure a high degree of reliability of the coding, it has been reviewed and discussed by James Gifford, founder and executive director of PRI until 2013, and author of extended theoretical framework used in this paper, as well as PRI staff members.
We use magnitude coding for each observation, following Wolfe et al. (1993) finding that counting each occurrence of an item in content analysis equally is a simplification and may produce misleading results. Units of text are coded on a scale: 0-no support, 1-weak support, 2-moderate support, 3-strong support, and 4-very strong support. However, only units coded 2-4 contribute toward the results presented in the analysis section of this paper to exclude instances where the coding might depend more on the coder's interpretation of the text than its explicit message. The coding method applied is also simultaneous, meaning that each unit of analysis can be coded as supporting multiple attributes in the theoretical framework (Saldana 2009). Following Gray et al.'s (1995) recommendation, the unit of analysis used is predominantly sentences; however, in cases where responses are poorly structured or unnecessarily long, multiple sentences or responses as a whole are an alternative. Morris (1994) also observed that coding based on entire units of text results in higher agreement between coders and human and computer analyses, which provides justification to depart from sentences as units of analysis if they are not singularly significant enough.

Data Analysis and Discussion
We use descriptive statistics to demonstrate the importance of each factor and how it changed over time. We consider the absolute number of answers in support of an attribute in a given year and the number as a proportion of all the responses to the relevant question. We look at aggregate results from all signatories, as well as split into asset owners and investment managers, and by groups of signatories who signed in each sample year. The percentages in Tables 3, 4, and 5 below do not add up to 100 because they are a proportion of all survey answers submitted in the given year, and the percentage not mentioned would not have been coded in support of any of the attributes.

Power
In considering the claim of signing the PRI, power as a source of salience is closely linked to the institutional investors' relationships with their other stakeholders, such as clients, regulators, and society. Coercive power means that a stakeholder has the power to force an investor to sign the PRI, as for instance a legislator would. Normative power works through symbolic means such as the reputational consequences of rejecting or satisfying a stakeholder claim. Utilitarian power is exerted via material incentive, for example, the awarding and withdrawing of pension fund mandates.
Coercive power as a source of salience of the claim to sign the PRI is not documented in the data. This was in line with the researchers' expectations based on the voluntary and aspirational nature of the PRI framework (Richardson 2009;Scherer and Palazzo 2011) and the lack of formal mechanisms of power active in the PRI-investor relationship such as legislation.
The strongest evidence emerged for normative power. 217 respondents in 2011 submitted answers that the researchers coded as support for the attribute of normative power being a source of salience in the decision to sign the PRI. This number grew from 3 % in the first year of the sample, to 6 % in 2008, 16 % in 2009, and 18 % in 2010 before reaching 47 %, almost half of the sample, in 2011. Normative power started out more prominently in the asset owner sample and continued to grow steadily, whereas in the investment manager sample it started out weaker but caught up with the asset owners by 2010. This translates to asset owners such as pension funds and insurers experiencing the reputational and other symbolic pressures and benefits earlier in the sample period, before they became a source of salience in the eyes of investment managers, possibly with the former driving the latter.
Mainstream investment managers and corporate pension funds were more influenced by normative power than their SRI and public counterparts. In 2011, 40 % of corporate pension funds and 21 % of mainstream investment managers mentioned normative power in their responses, compared to 18 % of public pension funds, and 11 % of SRI managers. This would suggest that the impact of the PRI stamp on mainstream investment manager, corporate pensions image, and reputation is larger, perhaps because they start out with no responsible investment image while public pensions and SRI investors are more likely to have communicated an investment approach similar to the PRI principles even prior to signing. The proportion of responses referring to utilitarian power display a similar growth, although slightly smaller, starting at 1 % in 2007 and slowly growing to 9 % in 2009, then eventually reaching 32 % in 2011. Utilitarian power found more support among asset manager signatories, likely due to the power relationship between asset owners and investment managers where an ESG capacity is becoming a consideration in hiring managers. A mainstream Australian investment manager says 'We have been queried and reviewed by asset consultants and clients on our ESG activities' (PRI Association 2013). A mainstream UK fund explains: 'Becoming a PRI signatory has also enabled us to respond positively to asset owner's RFIs and RFPs (requests for information and requests for proposal respectively) that increasingly ask about status regarding the PRI' (PRI Association 2013). Therefore, investment managers signing the PRI are responding to the material incentive of gaining new clients or satisfying the clients they already work for.
Asset owners quote the support and interest of their members as incentives, e.g., 'Supports the interests/needs of our membership' as stated by a Canadian public pension fund (PRI Association 2013). These answers offer evidence that asset owners are becoming more confident about the place of ESG considerations in investment practice within a traditional fiduciary framework.
For added perspective, it is useful to look at the distinction between the proportion of mainstream versus SRI investment manager responses and public versus private pension fund responses coded in support of utilitarian power: 15 % of all mainstream investment manager responses to the questions were coded in support of utilitarian power in 2011, compared to 11 % of SRI managers. The proportion of mainstream investment manager responses remained higher throughout the sample, suggesting that this group is more financially driven than SRI managers.
A parallel pattern is clear among asset owners in 2011 where 20 % of corporate fund responses were coded in support of utilitarian power, compared to 12 % of public funds, again likely because of the more financially driven character of corporate pensions compared to public pensions.
It is also interesting to see that the growth of support for the power attributes are not as sharp among signatories Table 3 Summarizes the aggregate numbers of responses that supported each attribute from the framework in each survey year who signed the PRI in the first 2 years as it is among the later years. This indicates that these two groups are most driven by different attributes: the earlier signatories are less likely to have signed for reputational gains, or to attract new business, than their later counterparts who were perhaps looking for those benefits as the PRI accelerated the growth of ESG in the industry. This divide between the early and the later signatories is reinforced by the difference in their support for management values and pragmatic legitimacy and is consistent with Gifford's (2010) temporal aspect where different attributes have varying importance to the stakeholder relationship in time.
The dominant role of normative and utilitarian power as a factor contributing to ESG related behavior by organizations does not come as a surprise considering the degree to which it has been present in previous literature on the topic of CSR and SRI adoption. In terms of utilitarian power, Mackey et al. (2007) propose a model where ethical activity is beneficial to the organization conditional on a favorable supply and demand balance. This angle of analysis is seconded by Barnett (2007) and corresponds to investors signing the PRI to meet client demand, more pronounced among asset managers, as their asset manager clients are more likely to clearly communicate ESG requirements.
This source of salience is also theorized in the form of relational organizational identity orientation in Brickson (2007). Brickson distinguishes between collectivistic, relational, and individualistic organizations, each of which responds primarily to values-based, client-related, and material incentives, respectively. Investors signing the PRI in response to client expectations or in an effort to attract clients would be placed somewhere between relational and individualistic orientation, depending on how much it is utilitarian and normative power related to client relationship and reputation and how much purely business casetype pragmatic legitimacy that drives their behavior (Brickson 2007).
Normative power being an important attribute raises questions regarding a possible decoupling of policy and practice among PRI signatories (Crilly et al. 2012;Weaver et al. 1999). In this context, it would mean investors signing the PRI to satisfy stakeholders and for reputational or signaling benefits while limited effort is made within the organization to integrate the principles into investment practice, a legitimacy-seeking strategy described in detail by Long and Driscoll (2008). Based on our data, we cannot conclude if that is the case but it is a possibility that could be explored by further research.

Urgency
We find no support in the data for urgency as a source of salience. Signatories make no explicit reference to the timing aspect of the claim to sign the PRI, or pressure resulting from the intensity of resources used by stakeholders in advancing it. This may be explained in part by urgency being part of the investor-client relationship where an organization considers joining the PRI for a long time and a conversation with a client who communicates that they are transitioning toward all-PRI-signatory managers delivers the final push (Gifford 2013, pers. comm.). It is likely that such instances would be reported as utilitarian incentive without mention of the urgency component.
Other studies in the area of stakeholder salience have reached conflicting conclusions on urgency being a contributing factor (Hautbois et al. 2012;Parent and Deephouse 2007;Neville et al. 2011;Weber and Marley 2012), and in this paper, we cannot establish its relevance based on the data despite considerable anecdotal evidence in its favor (Gifford 2013, pers. comm.).

Legitimacy
Of the legitimacy attributes, individual legitimacy is not documented in the data, as signatories do not make reference to individuals having a role in advancing the claim to sign the PRI. Individual legitimacy was also among the less salient attributes in Gifford's (2010) study of stakeholder engagement, but it has otherwise not been explored empirically in stakeholder theory literature. There is some discussion in the literature based on mainly qualitative The presence of organizational legitimacy in the survey responses has grown steadily over the sample period. Signatories refer to the legitimacy of the PRI as an organization adding to the salience of the claim to sign, and also to the PRI conferring legitimacy on their own organization and their SRI efforts of joining a leading initiative in the responsible investment space (Long and Driscoll 2008;Perez-Batres et al. 2012).
For instance, a mainstream European fund says signing the PRI 'grants credibility to its commitment to sustainability and corporate governance principles' (PRI Association 2011). A Canadian public pension fund quotes to be 'benefiting from the credibility of the PRI in the investment community as well as with companies we engage with' (PRI Association 2011).  Societal legitimacy defined as the perception of social support for a claim has had a small but constant presence in signatory responses. Investors see signing the PRI as a step toward aligning their goals with the goals of society, or being better attuned to society. It is difficult to hypothesize about why societal legitimacy was not an important factor to investors-other than investors feeling little pressure from the side of society as most ESG social pressure is focused on corporates. As observed by Baron (2009) who discusses societal pressures on organizations to engage is CSR, society is not explicit in manifesting its preferences and may therefore be difficult to quote as a source of salience. Support for societal legitimacy was slightly more pronounced among public pension funds and SRI managers than mainstream investment managers and corporate pensions-at an around 5 % and 1 % mean, respectivelyagain consistent with the generally more socially oriented character of the former (Blackburn 2006;Sethi 2005;Sievanen et al. 2013).

Extended Theoretical Framework
Of the Gifford additions to the stakeholder salience model are relative economic size and coalition building, but these attributes were not supported significantly by the data in this study population. Relative economic size in the form that Gifford introduces it to the model is reminiscent of institutional isomorphism in institutional theory (DiMaggio and Powell 1983), where the PRI as a growing trend in the industry would encourage larger and larger numbers of investors to sign. This may well be the case although it is not recognized by investors as a contributing factor to their decision to sign, or not reported as such. The relative size of assets is only reported as a significant factor in the context of coalition building opportunities for signatories among themselves via the clearinghouse. That is also the context in which coalition building is mentioned overwhelmingly by signatories, and also in literature, for instance as collectivistic organizational identity orientation in Brickson (2007) or membership of industry associations as a moderating factor of social responsibility of organizations in Campbell (2007). Neither coalition building nor relative economic size as defined in Gifford's extended model play a role in adding salience to the claim of signing the PRI itself.

Pragmatic Legitimacy
Pragmatic legitimacy in the sense of a business case or supplying management with new and relevant information is found in a growing number of responses over the sample period. At the end of the sample period, almost 50 % of investment manager signatories referred to the PRI as supplying relevant ESG integration methods and information, up to 20 percentage points from 2007. The proportion of asset owners grew in absolute numbers over the years but remained at around 30 % of responses proportionately. This is consistent with the fact that investment managers should be more responsive to the business case argument given that they are the most performance oriented and are in charge of investing the assets, whereas asset owners manage a minority of their assets internally.
Signatories see the value of the PRI in access to knowhow, best practice, research, and trends in the responsible investment. They see the PRI as a learning tool for their organizations in the transition toward an ESG integrated investor, via webinars, discussions with other investors, implementation support, collaborative engagements, and the association's research outputs. For instance, a French asset manager says: 'The PRI provides an ambitious roadmap for continuous improvement with clear signposts and long term objectives. Our organization has managed, through its involvement in the past year, to move towards greater ESG integration more broadly (additional asset classes) and more deeply (from SRI to mainstream)' (PRI Association 2011). A Brazilian investment manager describes the PRI as follows: 'We have found a framework that accommodated our pre-signing beliefs related to ESG issues and value creation, and helped us organize or reshape our internal analysis processes around the principles' (PRI Association 2011).
This finding is in line with much of the responsible investment literature that acknowledges the increasing realization among investors that SRI is a value-creating approach (Louche 2009) and that investors follow-up their investment beliefs with actions (Woods and Urwin 2010). It is also in line with Arjalies' (2010) social movement perspective on the development of SRI, where institutions seek to incorporate a new emerging logic into their existing institutional logic to create a new mainstream hybrid. The structure and know-how provided to its signatories by the PRI helps investors with this process, therefore pragmatic legitimacy contributes to the salience of PRI's claim.

Management Values
The salience of management values were strongly supported in the first 2 years of the PRI surveys, but support dropped dramatically in the subsequent years. In the years 2007, a staggering 90 % of asset owners and 60 % of asset managers explain their decision to sign the PRI by stating that the values of the PRI reflect the values of their organization, therefore, it is a natural step. In 2008, the asset owner number drops to 70 % while the asset manager number stays the same, but by 2009 evidence in both investors groups is only found in 10 % of responses. This number does not grow in the rest of the sample period. The groups of signatories that signed after year 2008 do not report on management values at all. It is the first signatories from 2006 to 2007 that drive the support for this attribute.
A possible explanation of the great prominence of management values in the first years of PRI's existence and why it was followed by its relative unimportance in the subsequent years is that the first wave of signatories would have been those investors who were already ESG minded. They made up most of the signatory body in the years 2007-2008, but by 2009 other factors such as the legitimacy of the PRI, utilitarian incentives, or pragmatic legitimacy attracted large numbers of mainstream investors who placed less emphasis on the alignment of values in their thinking about the benefits of signing the PRI. This is consistent with the social movement perspective on the evolution of responsible investment from activist and socially driven to profit driven (Markowitz et al. 2008). Other studies provide contrasting evidence, such as Jansson and Biel (2011) who find that institutional investors are not values driven as opposed to individual investors (McLachlan and Gardner 2004). It is, however, challenging to relate management values in the sense that they are manifested in the PRI-investor relationship to how they are understood elsewhere in the literature. The PRI principles convey little or arguably no moral values per se (Eccles 2010), while values-driven organizations as defined by Baron (2009) or Brickson (2007 take into account higher order values that are morally determined Figs. 2 and 3.
In summary, the attributes from the Mitchell et al. framework most supported by the data are normative and utilitarian power, and organizational and to a lesser extent societal legitimacy. Of Gifford's additions to the framework, the attributes of management values and pragmatic

Conclusion
In this paper, we have argued that there is a set of attributes that make the PRI as a stakeholder and its claim to sign the six PRI salient to institutional investors. The dataset provides evidence of the stakeholder relationship between investors and the PRI as defined by Freeman (1984) and according to Mitchell et al.'s (1997) stakeholder identification framework. The PRI has the legitimacy, power, and management values attributes that contribute to stakeholder salience. Based on an examination of 5 years of predominantly confidential survey data from PRI signatories spanning the period from 2007 to 2011, we find that the attributes that contribute most to the salience of the PRI, and its claim are organizational and pragmatic legitimacy, utilitarian and normative power, and management values. There is a high degree of variability in time and across groups of signatories who joined in different years, consistent with the temporal effect observed by Gifford (2010), and the dynamic view of stakeholder relationships (Fassin 2010;Sachs and Maurer 2009;Windsor 2010).
The strong presence of both utilitarian power and pragmatic legitimacy attributes in the responses indicates that an important source of salience for the claim to sign the PRI is the growing recognition of the materiality of ESG and its progressive mainstreaming reflected in demand for SRI (Louche 2009;Gifford 2010). This adds to the evidence already existing in the literature that investors see SRI as an avenue to value creation (Crifo and Forget 2013).
Normative power and organizational legitimacy as sources of salience highlight that signatories consider signing the PRI to be a way of communicating a certain image to clients and other stakeholders, and that the PRI has succeeded in gaining a considerable level of organizational legitimacy over its first 5 years that investors recognize and want to benefit from by association (Mitchell et al. 1997;Eesley and Lennox 2006). The prominence of normative power reinforces other findings such as that private equity funds adopt SRI strategies to differentiate themselves from competitors (Crifo and Forget 2013).
The dominant presence of normative power provides interesting input for the literature on decoupling, which documents how organizations may decouple policy from practice in an effort to respond to stakeholder pressures (Crilly et al. 2012;Weaver et al. 1999). The dataset available to us gives us limited opportunity to make inferences about the practical implementation of the principles and the closing of the gap between SRI policy and practice (Haack et al. 2012), as we do not track individual organizations over time. However, our analysis does highlight this as an interesting question and hopefully encourages further research in this area.
Management values are a unique attribute in that they appear to have been of major importance as a source of salience in the first 2 years in the sample, when the first wave of ethically and ESG-oriented investors signed up to the PRI. But they have been superseded by other factors in the later years when the values motivated signatory recruitment pool was largely exhausted and mainstream investors were drawn to the PRI in large numbers by its growing legitimacy and strong business case (Bromley and Powell 2012;Haack et al. 2012). Our findings support the proposition that the SRI movement transitions from socially driven toward profit oriented (Markowitz et al. 2008). Further research is required to more precisely define and measure the extent to which values play a role in the shift of institutional investment towards ESG integration.
Societal legitimacy found weak evidence in the data. Attributes of coercive power, individual legitimacy, urgency, relative economic size of stakeholder, and coalition building find little or no support in signatory survey responses.
This paper contributes to a better understanding of the stakeholder relationship between the PRI and investors. From the analysis, we gain an insight into how the sources of salience in this relationship are different from other contexts, in which the stakeholder salience model has been applied previously (Gifford 2010;Siltaoja and Lahdesmaki 2013;Hautbois et al. 2012;Parent and Deephouse 2007;Weber and Marley 2012). For instance, Parent and Deephouse (2007) find power to have the strongest influence on salience, followed by urgency and legitimacy. In the case of PRI signatories, legitimacy was more important than power, and urgency appeared to play no role at all. This paper also carries out a rare application of stakeholder theory to institutional investors as the organization managing its stakeholders, not as a stakeholder themselves, therefore extending stakeholder theory. It also sheds more light on the sources of salience of private regulatory initiatives in general, and of stakeholder claims as the primary unit of analysis.
The findings have practical implications for both institutions operating in the responsible investment space and investors themselves. A key take away for investors is that the gap between the perception of responsible investment as a legitimate investment strategy by clients and by investors practicing it for sustainable value creation is closing, as illustrated by the simultaneous rise in both utilitarian and normative power. This should encourage and reassure existing and potential responsible investors of the growing legitimacy of this investment approach. Institutions with similar objectives and stakeholder relationships to the PRI can better understand how the sources of stakeholder salience shift over time and transition towards a more business case, normative and legitimacy-driven character.
It is important to acknowledge the limitations of this research stemming both from the characteristics of the data and the method.
The data is self-reported, which has a potential distorting effect on the type of answers submitted. Moreover, the answers analyzed in the sample were not mandatory, meaning that the sample was likely to represent the more involved among the PRI signatory base, who dedicated resources above the minimum that is required to avoid delisting. If that is the case, the results might be only a biased reflection of the motivations of the entire signatory body. It is however reassuring that the geographical distribution and investor type break down is not too dissimilar between the signatories who answered the question annually and those who did not, which limits the risk of response bias. However a clear limitation is that geographical variation is not studied in this research, although it is acknowledged in the literature that responsible investment differs geographically (Sakuma and Louche 2008;Sievanen et al. 2013;Sparkes 2002). Detailed descriptive statistics comparing the two groups year on year can be found in Appendix 2.
The discretionary coding method is also a research limitation. It is hard to maintain consistency and lack of bias throughout the coding process for a human coder. The results are ultimately only one possible outcome of the analysis based on the interpretation of textual data by only one group of researchers.
The above research limitations are routinely found in survey based research (Agle et al. 1999;Valentine and Fleischman 2008). They are an obstacle to this type of study being precise and exactly representative of reality. Therefore the numbers quoted in the paper should be treated as a rough estimate of general trends among a sample of PRI signatories as reported by themselves, and as analyzed by a small team of researchers and not necessarily an accurate quantitative illustration of investor motivations.
This study opens up several avenues for further research. The temporal dimension of stakeholder relationships which is supported by our findings raises many questions around the possible determinants and regularities that may be found in this dynamic. There is also ample opportunity for further research insight into the importance of particular attributes in the salience of private regulatory processes to particular stakeholders. We touch upon this topic by distinguishing between SRI and mainstream investors and public and private pension funds but there remains much more complexity to explore.
Furthermore, the phenomena of conferring legitimacy (Long and Driscoll 2008;Perez-Batres et al. 2012) and decoupling of policy and practice (Boxenbaum and Jonsson 2008; Bromley and Powell 2012) that we highlighted in association with normative power may provide motivation for further studies in this area, as the concern regarding the true effectiveness of private regulatory processes in changing institutional behavior is widely shared among both practitioners and academics.
The attributes that find no support in our data also raise further questions. One notable absence from the results is individual legitimacy, despite there being qualitative evidence that individuals do play a role in promoting ESG in an institutional setting (Lewis and Juravle 2009;Waddock 2011). Urgency was also not mentioned by signatories, continuing a trend of mixed results in the literature (Hautbois et al. 2012;Parent and Deephouse 2007;Neville et al. 2011;Weber and Marley 2012). The question of whether or not these attributes contribute to the salience of private regulatory processes invites further investigation from future research.

Use of manual content analysis
Some of the following studies in particular establish a precedent for the choice of content analysis for coding selfreported textual data. In 1990, Jarvenpaa and Ives used it to analyze 649 letters to shareholders in annual reports to find out about the importance of IT to corporate strategy (Jarvenpaa and Ives 1990). Bravo et al. (2012) analyze selfdeclared CSR activities of Spanish banking institutions. Further, Meyer and Hollerer (2014) analyze the impact of framing contested issues in public discourse on shareholder value in Austria. The motivation for using content analysis in this particular paper is explained most accurately by the researchers' objective of using the method most effective in extracting maximum structured information from unstructured textual data.
Among the most prominent advantages of manual coding for content analysis is that the human coders brings in their experience and expertise into the coding process. Each of the authors of this paper has had experience of working in the PRI space and interacting with signatories, which helps interpret data correctly. Humans generally excel at analyzing unstructured text. It is for that reason that Deffner (1985) argues that computerized coding lacks validity by comparison, although it is effective as a method to maximize reliability. Following an experiment comparing automated and manual content analysis, Morris (1994) concludes that both methods can be equally effective. The fact that responsible investment still operates with an emergent vocabulary (Sandberg et al. 2009) makes the choice of manual content analysis a particularly suitable one since as it has been observed in previous academic literature, computerized content analysis takes away the complexity and context from language (Pennebaker and Lay 2002;Bligh et al. 2004b) preventing higher level insights into the material (Bligh et al. 2004a).
Moreover, as a methodology, content analysis is neither an unambiguously qualitative or quantitative method (Harwood and Garry 2003;Insch et al. 1997;Burnett et al. 1991). It can be described as qualitative in the first stage where the text is analyzed and the content divided into categories; quantitative when applied to quantifying the occurrence and relative importance of the categories or phenomena. This combination of qualitative and quantitative characteristics in content analysis allows us to present the data in a structured manner, and therefore communicate the findings more clearly.
The limitations are as follows. Manual coding is not very transparent, as opposed to automated coding following simple phrase or keyword identification processes; nor is it as consistent, because of fatigue, varying focus, and attention during the process and possible personal bias (Boritz et al. 2012;Morris 1994). A limitation that applies to management and finance research is that there is so far no specialized coding system (such as the Gottschalk-Gleser (1969) method in psychology). This limits the crosscomparability between different studies (Morris 1994  R percentage of respondents to question analyzed in given year belonging to each investor type, NR percentage of non-respondents to question analyzed in given year belonging to each investor type