Corporate Philanthropy as an Ideological Tool for Advocacy A DISSERTATION SUBMITTED TO THE FACULTY OF THE UNIVERSITY OF MINNESOTA BY Haram Seo IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY Advisor: Aseem Kaul and Jiao Luo May 2021 © Haram Seo 2021 i ACKNOWLEDGEMENTS I could not have completed this dissertation and my PhD without the help of so many wonderful people. I would first like to thank all of the faculty members in the Strategic Management and Entrepreneurship department of Carlson School of Management, who have been instrumental to my development as a scholar for the last six years. I would especially like to thank the members of my dissertation committee: Myles Shaver and Carrie Oelberger, who have been irreplaceable with their feedback and encouragement throughout the dissertation process. Most importantly, I would like to thank my wonderful co-advisors, Aseem Kaul and Jiao Luo, for being a constant source of support and inspiration through the endless ups and downs of graduate school. I could not ask for better mentors and role models, and I consider myself truly blessed to have had the opportunity to be your student. I would also like to thank my friends and family, who were always there for me when I needed them, and who kept me grounded during what at times felt like a very long journey. ii ABSTRACT This dissertation explores a phenomenon of corporate advocacy giving: philanthropic giving by corporations to 501c3 advocacy nonprofits (e.g., activist groups, think tanks) via their affiliated foundations. Corporate philanthropy is often conceptualized as the voluntary redistribution of corporate wealth, channeled through service nonprofits. However, another important function of nonprofits is advocacy, which leads to the question of whether corporate giving is channeled toward advocacy nonprofits and what role it plays in the repertoire of firms’ nonmarket strategies. Given the lack of prior knowledge on this phenomenon, throughout three chapters of my dissertation, I develop an initial yet systematic understanding on this underexplored phenomenon of corporate advocacy giving. Based on a novel database on corporate advocacy giving from 2003 to 2015, I find that it is not only an empirically significant phenomenon, but also plays a distinctive role in the repertoire of firm nonmarket strategies. Especially, my findings suggest that it can be understood as an underexplored channel of corporate activism pursued to advance firm employees' issue-related ideological preferences, which can often be homogeneous and extreme. In doing so, I present a more nuanced and less optimistic take on the role of corporate philanthropic giving in society, thus raising the need for effective monitoring on the practice. iii Table of Contents List of Tables v List of Figures vi Overview 1 Chapter 1: Corporate Philanthropy as a Channel of Sociopolitical Activism 8 1.1. Introduction 9 1.2. Data 15 1.3. Empirical Prevalence of Corporate Advocacy Giving 17 1.4. Distinctiveness of Corporate Advocacy Giving from Service Giving 21 1.4.1. Are they different? 21 1.4.2. Where does the difference come from? 24 1.5. Distinctiveness of Corporate Advocacy Giving from Lobbying 35 1.4.1. Are they different? 35 1.4.2. Where does the difference come from? 38 1.6. Discussion and Conclusion 45 Chapter 2: Employee Ideology as A Driver Of Corporate Activism: Evidence From Corporate Giving To Advocacy Issues 56 2.1. Introduction 57 2.2. Literature Review 63 2.3. Ideological Attributes of Firms as Antecedents of Advocacy Engagement 65 2.4. Methodology 71 2.4.1. Research Context and Data 71 2.4.2. Variables 75 2.5. Results 80 2.5.1. Main Results 80 2.5.2. Moderation Analyses 83 2.5.3 Supplementary Analyses: Alternative Explanations 85 2.6. Discussion 89 iv Chapter 3: Competitive Spillover effects of Corporate Activism from Heterogeneity in Across-Firm Ideology in Follower Population 98 3.1. Introduction 99 3.2. Literature Review 104 3.2.1. Corporate Political Ideologies and Advocacy Engagement 104 3.2.2. Spillover effects of ideological firms’ contributions 106 3.3. Possibility of Bifurcated Response from Followers 108 3.4. Asymmetry in the Response between Ideological Allies vs. Opponents 112 3.5. Methodology 116 3.5.1. Research Context and Data 116 3.5.2. Variables 117 3.5.3. Specification 122 3.6. Results 123 3.6.1. Main Results 123 3.6.2. Robustness Check 126 3.6.3. Mechanism analyses (1): Role of Employees 129 3.6.4. Mechanism analyses (2): Ideological Competition 135 3.6.5. Mechanism analyses (3): Recipient choices within each issue area 139 3.7. Discussion 142 General Conclusion and Discussion 156 References 159 Appendices 174 v List of Tables Table 1.1. Summary statistics for focus in advocacy and service giving………..………52 Table 1.2. Summary statistics for differentiation in advocacy and service giving………52 Table 1.3. Regression of firm characteristics on the choices of advocacy and service giving…………………………………………………………………………………….53 Table 1.4. Regression of firm characteristics on the choices of advocacy giving and lobbying………………………………………………………………………………….55 Table 2.1. Summary statistics and correlation tables…………………………………….93 Table 2.2. Main results……………………………………………………………….......95 Table 2.3. Moderation analysis………………………………………………………......96 Table 2.4. Supplementary analysis………………………………………………………97 Table 3.1. Summary statistics and correlation tables……………………………….......149 Table 3.2. Main results……………………………………………………………….....151 Table 3.3. Robustness analyses………………………………………………………....152 Table 3.4. Mechanism analysis (1): Role of employees…………………………..…....153 Table 3.5. Mechanism analysis (2): Ideological competition...…………………..….....154 Table 3.6. Mechanism analysis (3): Recipient choices within each issue area................155 vi List of Figures Figure 1.1. Firm-year level distribution of advocacy giving ratio………………….........50 Figure 1.2. Recipient-year level distribution of dependence ratio on corporate funding…………………………………………………………………………………...50 Figure 1.3. Cause priority in corporate advocacy vs. service giving……………….........51 Figure 1.4. Average correlation between advocacy and service amount across different cause areas………………………..……………………………………………………...51 Figure 1.5. Average correlation between advocacy and service amount across different firms…………………………………………………………………………...................52 Figure 1.6. Population of advocacy givers and lobbying firms among fortune 1000……………………………………………………………………………………....54 Figure 1.7. Scatterplot of advocacy - lobbying amount at firm-year level………………54 Figure 1.8. Top advocacy giving and lobbying industries……………………………….54 Figure 2.1. Individual employee-level distribution of ideological positions……….........92 Figure 2.2. Firm-level distribution of ideological intensity……………………………...92 Figure 3.1. Firm-level distribution of ideological position.…………………………….147 Figure 3.2. Summary of moderating analyses…………….……………………………148 1 OVERVIEW Philanthropy has long been a tool used by society to address its needs with the help of private actors. It is particularly important in American society, where the role of government is traditionally limited, and a substantial portion of public goods provision is left to the hands of constituents in the private sphere. As elite power holders in society, firms are often expected to do their share, thus carrying expectations not only to enhance the wealth of their shareholders, but also to share some of their wealth with diverse members of society. Corporate philanthropy is a primary avenue by which firms execute such voluntary redistribution of their wealth, channeled through service nonprofits (e.g., soup kitchens, homeless shelters) that deliver various social goods and services (e.g., food, shelter) to those in need. However, potential recipients for corporate philanthropy include not just service nonprofits, but also advocacy nonprofits that specialize in framing public discourse by influencing society’s conceptions about which issues matter and how they should be dealt with, especially for the most salient and contentious social issues of our time (e.g., global warming, LGBTQ rights, gun control). For example, The Recycling Partnership, a nonprofit specialized in advocacy on global warming, raised $10 Mil from firms from 2013 to 2017, which accounted for 85% of its entire revenue. While prior work has paid scant attention to this phenomenon of corporate advocacy giving, largely equating it with giving that goes to service nonprofits, the fundamentally different function of advocacy nonprofits points to a distinct motivation behind firms’ decision to fund them. Their unique function presents the possibility that firms could use advocacy giving to actively 2 shape what the public care about, in an effort to align public interests with firm members’ internal interests, while the prior literature assumes that firms give to the areas that the public care about. Notably, this is a qualitatively different and more nuanced take on the role of corporate giving in society, raising important questions about the appropriate limits on powerful business actors’ participation in civic discourse, and the use of tax- exempt philanthropic giving to those ends. Despite the theoretical and practical importance of this phenomenon, little scholarly attention has been given to this matter. Thus, throughout three chapters of my dissertation, I seek to develop an initial yet systematic understanding on this underexplored phenomenon. To this end, I built a novel database on corporate advocacy giving from 2003 to 2015 by collecting detailed information on not only corporate givers, but also their 48,325 recipient nonprofit organizations. This enables me to track which issues and geographic areas a specific corporate grant goes to, and also which function of nonprofit organizations it supports, allowing me to break down giving to advocacy vs. service nonprofits at an individual grant level. The resulting database includes over one billion dollars of advocacy grants from 441 corporations. The dissertation is organized as the following three chapters. First chapter of dissertation: What is corporate advocacy giving? The first chapter of my dissertation focuses on filling this gap by documenting key stylized facts to provide a baseline for further inquiry into this understudied phenomenon. Specifically, this essay centers on the following two questions. First, is corporate advocacy giving an empirically significant phenomenon? Second, how is it 3 distinct from alternative nonmarket strategies, such as traditional service giving and lobbying? To this end, I built a novel database on corporate advocacy giving from 2003 to 2015 by collecting detailed information on not only corporate givers, but also their 48,325 recipient nonprofit organizations. This enables me to track which issues and geographic areas a specific corporate grant goes to, and also which function of nonprofit organizations it supports, allowing me to break down giving to advocacy vs. service nonprofits at an individual grant level. The resulting database includes over one billion dollars of advocacy grants from 441 corporations. According to my data, the majority of corporate givers (72%) give to advocacy nonprofits, directing about 7% of their entire giving to these organizations. From the side of advocacy nonprofits, corporate giving accounts for over 40% of these recipients’ external organizational contributions. Also, when giving to advocacy nonprofits, firms support causes different from those in service giving, and their advocacy giving is not only more focused and differentiated than service giving, but also more closely connected to employees’ preferences. These patterns suggest that, while service giving is largely reactive to external societal demands as understood in the literature, advocacy giving is more closely linked to firms’ own internal preferences. Moreover, I find that a substantial number of advocacy givers do not engage in lobbying at all, and even among firms that engage in both, the correlation between the two is modest. Also, compared to firms that lobby exclusively, advocacy givers are more employee-focused, suggesting that advocacy giving may target broader ranges of issues that are not necessarily directly related to 4 firms’ core business operation yet may elicit strong emotions and reactions among their internal stakeholders. Overall, the first chapter of my dissertation shows that corporate advocacy giving is not only an empirically significant phenomenon, but also a distinct form of corporate giving that is pursued in ways systematically different from traditional corporate service giving, thus calling for more scholarly attention on this understudied phenomenon. Notably, the systematic pattern of differences between advocacy giving and alternative forms of nonmarket strategies provides initial evidence for the possibility that corporate advocacy giving may be understood as a channel of employee activism, whereby a firm’s employees may use philanthropic resources to shape the nature of public discourse around issues that they care about to their private preferences. Second chapter of dissertation: Why do firms engage in advocacy giving? In the second chapter of my dissertation, I further explore this possibility by investigating the role of firms’ organizational members’ ideological interests on their decision to engage in advocacy giving. While the first chapter highlights firms’ internal interests as the potential driver of advocacy giving, what remains unclear is the nature of such interests. One possibility of interest is that those interests represent highly ideological preferences held by CEOs and other employees of the firm (i.e., strong liberals or conservatives), who are motivated to use its philanthropic resources to influence civic discourse around advocacy issues of their concern. While nonmarket scholars have long focused on external drivers of corporate social and philanthropic activities, scholars have recently started to highlight intrinsic and normative motivations 5 of a firm’s internal organizational members as an underexplored driver of those activities. According to these studies, firms often develop a strong ideological leaning among their membership via a sorting mechanism (i.e., individuals are attracted to and stay with the firms whose employees exhibit similar preferences to themselves) and such ideological clustering of like-minded individuals shape the pattern of corporate social actions. While these existing studies have demonstrated the role of such organizational ideological preferences in the context of traditional service issues (e.g., providing food and shelter to people in need), there are reasons to expect that its role will not only be more pronounced around advocacy issues, but also play out in a different manner. Most existing studies focus on the relationship between the direction of the firm’s ideological leaning among their workforce and the level of its social engagement, by showing that liberal firms are willing to invest more in CSR than conservative firms. Unlike traditional service issues, which focus on provisioning necessary social goods and services to those in need and thus appeal disproportionately to liberals who care more about egalitarianism, advocacy issues are unique in that they tend to attract attention from people of both ideological leanings. Therefore, it remains unclear which ideological characteristics within the organization would shape firms’ engagement around advocacy issues of contentious nature, and in what ways. This paper addresses this gap by uncovering multiple ideological attributes of the organization as important antecedents that jointly affect firms’ willingness to engage in corporate activism, thus departing from the existing focus on the direction of ideological leaning. First, I examine the role of the intensity of employees’ ideological preferences, 6 which refers to the importance of ideology to average organizational members’ self- concept. After all, not all employees are ideologically motivated, and this variation matters, as only those with higher ideological intensity are willing to put a high price tag on the firm’s action. Indeed, While controlling for the effect of the average ideological position of the firm’s workforce, I find that firms with a higher ideological intensity, proxied as firms with a higher number of employees that make campaign political contributions, do give more to advocacy issues. I also consider additional ideological attributes of the organization as moderators, finding that the positive effect of ideological intensity holds only for those whose members think not only similarly to each other but also hold extreme preferences. Overall, the main finding of the second chapter—ideological preferences of firms’ organizational members as the main driver of corporate advocacy giving—is highly consistent with the main idea behind my dissertation: firms’ internal stakeholders may use giving to shape what society wants for their own internal preferences, rather than using such giving as a public tool to fulfill the needs of society. Third chapter of dissertation: What happens afterwards? In the second chapter, I answered the question of why firms engage in advocacy giving, demonstrating the effect of a focal firm’s ideology on its own social actions. However, we are yet to study its broader spillover effects: specifically, how ideological firms’ initial advocacy actions affect the actions of other firms in the population. Traditional accounts of how firms influence others’ philanthropic actions have focused on homogeneity and isomorphism, showing that firms tend to give more when 7 others give more. This finding suggests that initial giving by ideological firms may encourage their ideological allies to give, thus inviting widespread support for the issue. In this way, initiating ideological firms may act as “issue entrepreneurs” that put an issue into play by raising its salience in their field. Yet, what remains underexplored is the possibility that initiating firms’ actions may invite competition and countermobilization from their ideological rivals, who may join in the debate with a different or even opposing issue position. In the second chapter of my dissertation, I focus on this tension, positing that competitive response from ideological opponents may dominate the response from ideological allies. This is not only because the perceived costs for inaction are likely to be higher for organizational members of ideological opponents, but also because ideological opponents can strategically differentiate themselves with such counter-response in the eyes of their employees. My analyses show that while ideological firms’ giving encourages their ideological allies’ giving in the issue area of their choice, it also invites their ideological opponents’ giving, with the response from the opponents often dominating the response from the allies. Such selective counter-response from ideological opponents is stronger in swing states, and among firms with radical preferences, consistent with my theory that this response reflects the dynamics of within-issue competition between firms that lie at the opposite ends of the ideological spectrum. These findings suggest that ideological firms’ initial action to promote the cause may backfire by inviting a disproportionately stronger response from their ideological opponents compared to their allies, thus contributing to increased conflict and polarization around the issue. 8 CHAPTER 1: CORPORATE PHILANTHROPY AS A CHANNEL OF SOCIOPOLITICAL ACTIVISM 9 1.1. Introduction In the nonmarket literature, corporate social responsibility (CSR) and philanthropic giving are often viewed as strategic tools used to fulfill the demands of external stakeholders. Specifically, firms can fulfill the demands of the public and other external stakeholders, who want to see firms give to social areas of need, by channeling donations to service nonprofits that provide social goods and services (e.g., food, shelter) (Porter and Kramer, 2002; Bénabou and Tirole, 2010; Kaul and Luo, 2018). Scholars have shown that this voluntary social commitment and resulting goodwill from a variety of external stakeholders can translate into various strategic benefits: higher willingness to pay by customers (Elfenbein and McManus, 2010; Lev, Petrovits, and Radhakrishnan, 2010), more lenient assessment by investors and regulators (Mackey, Mackey, and Barney, 2007; Wang and Qian, 2011), and effective impression management and reputation repair in the face of activist attacks (McDonnell and King, 2013; Ingram, Yue, and Rao, 2010; Godfrey, 2005; Luo, Kaul, and Seo, 2018). Importantly, what these explanations have in common is that they see firms undertaking philanthropy in response to exogenous stakeholder pressures and social needs, such that firms can address demands of stakeholders by funding service nonprofits operating in the issues that are dear to those stakeholders. However, the population of potential recipients for corporate philanthropy includes not just service nonprofits, but also advocacy nonprofits such as activist groups and think tanks (McCarthy and Zald, 1977; Jenkins and Halcli, 1999; Kimberlin, 2010). For example, The Recycling Partnership, a 501(c)3 advocacy nonprofit specialized on 10 matters of environmental degradation and global warming, raised $10 million from firms from 2013 to 2017, which accounted for approximately 85% of its entire revenue (O’Neil, 2018). While the extant literature has paid scant attention to this phenomenon of corporate advocacy giving, largely equating it with service giving, the fundamentally different organizational function of advocacy nonprofits points to a distinct motivation that may drive firms’ decision to fund them. Unlike service nonprofits, advocacy nonprofits specialize in framing public discourse, influencing society’s conceptions about what issues matter and how they should be dealt with, and thus, they are often viewed as the moving force behind major cultural changes (King and Soule, 2007; Hiatt, Sine, and Tolbert, 2009). This suggests not only that stakeholders’ preferences and demands are endogenous, but also, more importantly, that firms may use their giving to advocacy nonprofits as a means of shaping the process of institutional change to their organizational members’ internal preferences and interests (Oliver and Holzinger, 2008). Firms’ use of nonmarket strategies for their internal interests and benefits is well acknowledged in the broader nonmarket literature, especially in the arena of public politics that study how firms engage in lobbying and campaign contributions to influence key public stakeholders such as policymakers and regulators (Hillman, Zardkoohi, and Bierman, 1999; De Figueiredo and Tiller, 2001; De Figueiredo and Silverman, 2006; Bonardi and Keum, 2005; Jia, 2014; Choi, Jia, and Lu, 2015; Richter and Werner, 2017). Extending this knowledge, scholars have recently shown that firms can also use CSR (Werner, 2015; McDonnell and King, 2018) and corporate philanthropy in particular (Bertrand et al., 2018; 2020) as a means to influence policy issues of direct strategic 11 relevance for their operation (e.g., taxation, subsidy). Yet, what remains relatively underexplored is the possibility that firms may use their philanthropic donation as a channel of corporate activism to influence society around on the issues that have little direct relevance to the way firms run their businesses and are rather a part of bigger sociopolitical debate. Indeed, this possibility is consistent with the growing body of research showing that firms and executives are increasingly weighing in on the issue that is peripheral to their core business operation, essentially acting like activist themselves by mobilizing the public (Walker and Rea, 2014; Chatterji and Toffel, 2019; Hambrick and Wowak, 2020). Despite such novel theoretical possibility, our scholarly interest in this phenomenon of corporate advocacy giving has been limited. Given the lack of our prior knowledge on this phenomenon of corporate advocacy giving, including even the basic facts such as how many firms give to advocacy nonprofits, how much, where, and who they are, this paper takes an exploratory, fact- based approach that documents key stylized facts to provide a baseline for further inquiry into this understudied phenomenon (Helfat, 2007; Oxley et al., 2010; Bettis et al., 2014; James and Shaver, 2016; Moeen and Agarwal, 2017). I begin this paper by asking if corporate advocacy giving is an empirically significant phenomenon in the first place. Upon establishing the empirical significance of corporate advocacy giving, I then move on to exploring its strategic role in the repertoire of corporate nonmarket strategies, by asking if and how advocacy giving is distinctive from alternative forms of nonmarket strategies. To this end, I compare corporate advocacy giving to its two empirical and theoretical counterparts whose roles are relatively well understood in the nonmarket 12 literature: corporate service giving (Porter and Kramer, 2002; Bénabou and Tirole, 2010; Kaul and Luo, 2018) and lobbying (De Figueiredo and Silverman, 2006; Jia, 2014; Richter and Werner, 2017). The comparison to service giving helps to inform if and to what extent advocacy giving, unlike the traditional reactive form of service giving attending to unmet needs and demands of society, can be understood as a internally driven strategy based on the firm members’ private demands and preferences, which may significantly diverge from what’s objectively needed in society. The comparison against lobbying activities helps to evaluate the scope and nature of issues that firms are trying to influence via their advocacy giving and, in particular, if they target broader sociopolitical issues that extend beyond the traditional focus of lobbying activities centered around economic issues. Taken together, the systematic patterns of dissimilarity would present us with initial evidence that informs if corporate advocacy giving is a distinctive form of nonmarket strategy that deserves more scholarly attention, and specifically if it can be understood as a novel form of corporate activism driven by the firm’s internal organizational members’ private preferences. Based on about 0.4 million grant-level records of Fortune 1000 companies’ philanthropic donations through corporate foundations from 2003 to 2010, I find that about 72% of corporate givers give to advocacy, not just to service, and they spend $534 Mil in total on advocacy giving, accounting for about 7% of their entire donation spending. Also, corporate money that goes to advocacy nonprofits accounts for over 40% of these recipients’ entire organizational funding, pointing to the potential significant influence firms may exercise on those advocacy groups. I also find the systematic 13 differences that exist between corporate advocacy giving and the two alternative forms of nonmarket strategies: corporate service giving and lobbying. When giving to advocacy nonprofits, firms support causes that are often different from those that they support through service giving, and advocacy giving tends to be both more focused and differentiated than service giving. These patterns suggest that, while service giving is largely reactive to external stakeholder demands, advocacy giving is more closely linked to firms’ own internal needs, being pursued by firms in a more differentiated and focused manner, especially by those that have strong incentives to care about employees’ demands. The final analysis that compares advocacy giving to lobbying activities shows that, while many advocacy givers engage in lobbying, a substantial number of advocacy givers do not engage in lobbying at all, and even within the population of firms that engage in both activities, the correlation between the two is fairly modest. Importantly, advocacy givers, compared to exclusive lobbying firms, are more employee-focused, suggesting that advocacy giving may be understood as a channel of employee activism, whereby employees use their firm’s philanthropic resources as a vehicle for activism to influence social discourse around issues that they personally care about, and do not necessarily have direct economic relevance to the firm’s core business operations. This study contributes to the literature on nonmarket strategy in multiple ways. To the best of my knowledge, it is one of the first studies that systematically examines the phenomenon of corporate advocacy giving with large-scale data (Bertrand et al., 2018; 2020). In doing so, this research demonstrates the phenomenon’s empirical significance while also showing that it is pursued in ways systematically different from other 14 nonmarket strategies such as corporate service giving or lobbying, thus highlighting the unique role it may play in the broad repertoire of nonmarket strategies and providing a solid ground for future research on this phenomenon. Second, this study contributes to the corporate philanthropy literature by extending its scholarly conceptions of what corporate giving is. By the existing CSR accounts, corporate philanthropy is typically conceptualized as a form of delegated philanthropy, whereby firms give back to society in expectation of rewards from external stakeholders by funding of service nonprofits (e.g., soup kitchens, homeless shelters) that deliver social goods and services to people in need. While this view of corporate giving as a means of voluntary redistribution of corporate wealth and stakeholder management tool applies well to the traditional form of service giving, it does not hold in the case of corporate advocacy giving. My findings suggest that firms give to advocacy not because others expect them to give, but because firms’ organizational members have their own private, internal preferences to give to certain advocacy causes, which may deviate from what’s best for society. In doing so, I extend our view on why firms give and how firms can benefit from such giving, thus joining a growing body of research that examines firms’ proactive use of CSR and specifically tax-exempt corporate philanthropy for their own private benefits and its welfare implications (Werner, 2015; McDonnell and King, 2018; Barnett, 2019; Lyon et al., 2018; Kaul and Luo, 2018; Oelberger, 2018; Gehringer; 2021). Third, this paper joins an emerging body of literature that study how firms’ influence to shape the social and institutional environment is not limited to the limited scope of political and economic issues that have direct relevance to the firms’ business operations. While scholars have 15 shown that firms often engage in efforts to organize movements of their own by engaging the public (Walker, 2009; 2014, Walker and Rea, 2014), our knowledge on such proactive role that firms play outside the political arena of traditional lobbying activities is still relatively limited. This paper shows that corporate philanthropy can serve as another unexplored channel of corporate activism, with employees using their firm’s philanthropic resources to fund advocacy groups specialized in issues that they personally care about and thus to potentially shape the nature of broader sociopolitical debates to their private preferences. 1.2. Data To establish corporate advocacy giving as an empirically significant phenomenon and compare it to the alternative forms of nonmarket strategies, the first and foremost task is to build a reliable database on it. To this end, I draw on multiple datasets of different natures. The first dataset comes from the Foundation Center, a nonprofit organization that compiles information on U.S. philanthropic foundations and their grants. Its Foundation Directory Online (FDO) database provides the single most comprehensive coverage on U.S. based philanthropic foundations and their grants (Luo et al., 2018), compiling the information on over 3 million grants made by 28,851 foundations during my study period of 2003 to 2010. Its grant data are originally drawn from the IRS Form 990s (Return of Organization Exempt from Income tax), which philanthropic foundations are mandated to file annually. My sampling frame is the public firms in Fortune 1000, based on which I map the list of corporate foundations and direct giving programs in the FDO dataset that 16 are affiliated to these Fortune 1000. This leaves me with 426 companies and detailed information on about 0.3 Mil records of their grants totaling $9 Bil, including the recipient nonprofit’s name, location, and unique organizational identifier (EIN). The availability of EIN information allows a clean matching of the FDO database to the second main database of my study: Exempt Organizations Business Master File Extract (EO BMF). This is a master file on U.S. tax-exempt nonprofits that provides various information on their organizational characteristics. The most important information from this file is, to this study’s interest, a nonprofit’s primary organizational purpose, which I use to identify which of the grants correspond to advocacy giving. Following the previous studies (McCarthy and Zald, 1977; Minkoff and Agnone, 2010; Coke et al., 2009), I operationalize a given grant as advocacy giving if it goes to a nonprofit whose primary purpose is advocacy as defined by the National Taxonomy of Exempt Entities Core Codes (NTEE-CC; a standard classification system in the nonprofit sector similar to SIC in the for-profit sector)1. The details of the categorization scheme are reported in Appendix Table 1.1. This approach of identifying advocacy nonprofits has a huge advantage in terms of its comprehensiveness, especially compared to the popular alternative of media report-based sampling of advocacy organizations that is known to be biased towards oversampling organizations of certain size and location (Earl et al., 2004). The coverage of BMF file is based on the IRS forms required for any exempt 1 This mostly includes R (Civil Rights, Social Action, Advocacy) in NTEE major code, or 01 (alliances & advocacy) and 05 (public policy analysis) in NTEE subcode. 17 organizations with gross receipts of over $50,0002 and is thus extensive, providing the necessary information on over 1 Mil nonprofits during my study period. However, an important limitation with this approach should be noted. As the NTEE-CC categorization system categorizes nonprofits only by their primary mission, I cannot track advocacy giving that goes to nonprofits whose core mission is not advocacy but still engages in it as a secondary function. While the extant understanding is that most nonprofit advocacy is performed by core advocacy groups (Kimberlin, 2010), it does warrant that my observation about the prevalence of corporate advocacy giving, especially its relative prevalence compared to service giving, should be interpreted strictly as a conservative, lower-bound estimate. In addition to the donation data, this paper’s inquiry also requires the dataset on firms’ lobbying activities, which serve as the second empirical counterpart to the phenomenon of corporate advocacy giving. The data come from the firm-level lobbying database at lobbyview.org (Kim, 2018), which is originally generated from lobbying reports filed between 1999 and 2017 and recorded at the Senate Office of Public Records. I also draw on a variety of other sources that include Compustat and KLD Research & Analytics, etc., which provide additional firm-level financial and operational information. 1.3. Empirical Prevalence of Corporate Advocacy Giving In the first section, I focus on examining the empirical significance of corporate advocacy giving in my sample. I begin this section by calling attention to the fact that a 2 While the threshold was lowered to $25,000 as of 2009, I keep the $50,000 cutoff throughout the entire period to ensure the consistency in the sample of 501(c) nonprofit recipients. 18 substantial number of corporate givers give to advocacy nonprofits. In the entire sample of Fortune 1000 companies with affiliated foundations or direct giving programs, I find that 72% of them give at least once to advocacy nonprofits during the study period3. At the firm-year level, the proportion of firms that give to advocacy in each year is on average 54%. Clearly, corporate advocacy giving is an inherent part of philanthropic practices for the majority of corporate givers. The absolute dollar figure of advocacy grant that firms made during the sample period (2003 - 2010) is $534 Mil. In the years when they give, they give about 7% of their total giving to advocacy nonprofits; this portion is not small, especially given the significantly smaller presence of advocacy nonprofits in the entire recipient population (2% by income and 1% by asset size)4. I also find a significant variance in the relative magnitude of advocacy giving to service giving across firms. Figure 1.1 maps the distribution of the share of advocacy grant in total grant at the firm–year level. On the one hand, it is clear that a lot of companies dedicate only a very small portion of their giving to advocacy, with about half of them giving under 4%. On the other hand, the figure also illustrates that the ratio of advocacy giving does range widely from 0.1% to 100% (standard deviation is 10%). In other words, there exist companies whose 3 In the entire sample of Fortune 1000 firms, 38% of the firms are identified to have made at least one donation during the study period. 4 The relative magnitude of corporate advocacy giving is substantial compared to lobbying as well (the second counterpart), with the amount of advocacy giving representing around 8% of firms’ federal lobbying expenditures for those in my sample in 2010. If we apply a 6% advocacy-giving ratio to the reported amount of entire philanthropic giving made by corporations in 2010 (Giving USA, 2010), the implied dollar value of advocacy giving is $1.07 Bil, which amounts to about 26% of total lobbying spending in that year. 19 advocacy-giving ratio is far over the average value of 7%, with some of them dedicating their entire philanthropic expenditure to advocacy organizations. ***Insert Figure 1.1 about here*** Taken together, these figures imply that advocacy giving may indeed be an important component of corporate philanthropic practices, with the majority of firms allocating at least some portion of their giving to advocacy, and with some of them spending a lot more than others. Yet, a more important question is what corporate advocacy giving means from the side of nonprofit recipients—that is, if it is big enough to represent significant influence on recipient advocacy organizations. This question is especially important in light of the key potential motivation behind advocacy giving, such that firms may use philanthropic giving as a tool to influence advocacy organizations’ goals and activities. Thus, what matters is the relative significance of corporate contributions compared to recipient nonprofits’ alternative sources of funding. To this end, I look into how many advocacy nonprofits are the recipients of corporate philanthropy in the first place. I find that among 20,046 advocacy nonprofits that are identified on BMF file, 2,413 organizations receive funding from Fortune 1000 companies at least once during the sample period. This corresponds to about 12% of the entire advocacy nonprofit population, showing that a substantial portion of this population is subject to the influence from corporations at least to some extent. Importantly, a lot of these nonprofit recipients of corporate advocacy giving depend quite heavily on corporate money. Within the sample of corporate advocacy giving recipients, the average value of dependence on corporate funding at the recipient organization level 20 is 41%; that is, corporate giving accounts for a major portion of the total organizational funding these advocacy nonprofits collect in each year, which includes funding from the various sources captured in the FDO data—independent/family foundations, community foundations, government organizations, etc5. A closer look at the distribution, which is illustrated in Figure 1.2, shows an interesting pattern of heterogeneity in the level of dependence on corporate funding across these advocacy recipients. The standard deviation in this distribution is very high—39%, which is consistent with the bimodal distribution shown in Figure 1.2 with two peaks on the opposite ends of the spectrum; thus, advocacy recipients tend to show either a low level of dependence on corporate funding or a very high level of dependence. Interestingly, the peak in the right end of the dependence—i.e., 100% dependence on corporate grant—is the highest, a result which suggests corporate preference for recipients for which they can be the exclusive source of funding. ***Insert Figure 1.2 about here*** In summary, the descriptive analyses that investigate the significance of corporate advocacy giving from the sides of both corporate givers and advocacy recipients produce several initial insights. For one, corporate advocacy giving is an action taken by the majority of firms that give, suggesting that it is indeed an important component of firms’ 5 To be clear, it does not include other income sources of nonprofits such as individual giving or program revenue generated through nonprofits’ own operational activities. However, previous research does suggest that foundation funding has traditionally been the major source of funding for a lot of advocacy nonprofits (Jenkins and Halcli, 1999; Berry, 2015). For instance, Berry (2015) notes that, according to his interviews, roughly 1/3 of all the nonprofit advocacy groups that have begun in the past five years were obtaining at least half of their total income from foundations. 21 philanthropic portfolios. Also, the findings support the empirical significance of advocacy giving in terms of giving amount—especially in terms of what it means from the nonprofit recipient side. Not only does advocacy giving account for a substantial share of firms’ entire philanthropic expenditures, it represents an even bigger share of funding when it comes to advocacy nonprofits that are the recipients of corporate giving, implying the possibility that corporate funders’ influence on them can indeed be substantial. Finally, I observe a significant heterogeneity across firms in their engagement levels with advocacy giving. Some never give to advocacy, whereas a lot of companies do. Also, while many dedicate only a small portion of their philanthropic expenditures to advocacy, there exist corporate givers who give out substantial portions to advocacy—in some cases, the entire amount of philanthropic expenditure in that year. 1.4. Distinctiveness of Corporate Advocacy Giving from Service Giving 1.4.1. Are they different? Having established the empirical significance of the phenomenon of corporate advocacy giving, I turn to the next question: how distinctive is this phenomenon from the alternative forms of nonmarket strategies? This section focuses on the comparison with the first counterpart, by asking if corporate advocacy giving is different from service giving, and if so, how. After all, most studies in the nonmarket literature have long equated corporate philanthropic giving with service giving, which refers to donation to service nonprofits whose role focuses on provisioning various social goods and services to those in need (Porter and Kramer, 2002; Bénabou and Tirole, 2010; Kaul and Luo, 2018). Underlying this generalization in the 22 literature is the assumption that advocacy giving is not substantially different from service giving. This assumption serves as the null hypothesis of this section. Under this null hypothesis, corporate advocacy giving plays essentially the same strategic role with service giving; thus, the traditional understanding of how philanthropic giving can create strategic value to firms can be readily extended to the case of advocacy giving. According to this logic, firms give to advocacy nonprofits because external stakeholders want them to meet social needs. That is, there exist external stakeholders who care about advocacy issues (e.g., LGBTQ rights, racial equality issues, school reform, environmental protection), and firms can buy goodwill from those stakeholders by supporting advocacy nonprofits in those issues that are salient in stakeholders’ minds, which is then likely to form goodwill with multiple external stakeholders such as customers, investors, and regulators (Elfenbein and McManus, 2010; Lev et al., 2010; Mackey et al., 2007; Bode et al., 2015; Burbano, 2016; Wang and Qian, 2011). Overall, this scenario posits that firms give not just to service nonprofits, but also to advocacy nonprofits simply because it enables them to manage various demands and threats of external stakeholders better. To evaluate the validity of this claim, I now turn to empirical evidence that examines the patterns of similarity and dissimilarity between advocacy giving and service giving. I begin with the simplest evidence: the correlation between the amount of advocacy giving and service giving. I find that the amount of advocacy giving is highly correlated with the amount of service giving: the mean value of correlation at the firm– year level is 74%. Thus, it seems that, not surprisingly, big service givers are the ones 23 that also give a lot to advocacy. In other words, there does exist a significant similarity between the two phenomena in terms of who gives, indicating that service giving is indeed an important empirical counterpart to the phenomenon of advocacy giving. However, a closer look at the data starts to reveal systematic differences that exist between the two in other dimensions. For one, when I examine the data at a more granular level of firm–cause–year level, the average level of correlation between advocacy giving and service giving significantly drops down to 24%. In other words, firms tend to spend their advocacy giving in areas that are different from where they spend service giving. This points to significant differences that exist between advocacy and service giving in terms of how firms spend their money. Indeed, the cause priority structure in advocacy giving looks significantly different from that in service giving, as shown in Figure 1.3. This gap in issue priority structure between the two types of giving can be more systematically observed in the following Figure 1.4, which calculates the average level of correlation between advocacy and service giving at the cause level. On the one hand, in issue areas such as Community, Education, and Environment, the average correlation is over 40%, meaning that firms’ advocacy giving closely maps their service giving in these areas. On the other hand, in areas such as Recreation, Public safety, and Science, the average correlation stands below 15%, showing that how much firms spend in advocacy and service giving in these areas is largely independent of each other. I find an even more significant variance at the firm level, which is shown in Figure 1.5. While there are some firms that spend their advocacy and service giving largely in the same areas, more firms spend their advocacy giving in areas very different from 24 where they spend service giving. Indeed, a substantial number of firms even show a negative correlation in how much they give to advocacy and service giving in a given cause area, again, demonstrating that advocacy and service giving mean two separate choices to many of the firms in the population. This observed difference between corporate service giving and advocacy giving is not a simple artifact of the differences in the distribution of advocacy organizations across different causes in the nonprofit population. Appendix Figure 1.1 shows how the pattern of corporate advocacy giving shows a significant divergence from this baseline distribution. As shown in Appendix Figure 1.2, I also find a substantial difference in the pattern of advocacy giving that exist among corporate foundations and general, non- corporate foundations. ***Insert Figures 1.3, 1.4 and 1.5 about here*** 1.4.2. Where does the difference come from? The evidence so far is inconsistent with the null hypothesis, indicating that there does exist significant differences between advocacy and service giving. Importantly, this suggests that the existing model of corporate philanthropy cannot be easily extended to explain why firms give to advocacy. The logical next question is then, where does this difference between the two phenomena come from? How should we understand advocacy giving in a different way? One possibility of interest is that firms give to advocacy not to fulfill the externally given demands of stakeholders in their institutional environment as assumed in the case of service giving, but rather to change such demands. That is, firms may give not 25 because stakeholders interested in certain issues want them to give, but rather because firms themselves want to make stakeholders more (or less) interested in those issues. This conjecture is based on the understanding that the core function of advocacy nonprofits as “issue entrepreneurs” is centered around changing the perceived importance and public understanding of various social issues in society (McCarthy and Zald, 1977; p. 1215). This important role of advocacy organizations as an agent of institutional change has been well noted in the existing literature. These activist groups have historically been powerful advocates of social and cultural changes, playing a critical role in reshaping society’s perceptions about minority rights, need for environmental protection, and most relevantly, the social responsibilities of corporations regarding those issues as well (Baron, 2001; King, 2008; King and Soule, 2007; Hiatt et al., 2009; Soule, 2009). Thus, corporate advocacy giving may be internally driven attempt by the firm’s organizational members to make use of this main role of advocacy nonprofits for their benefits: shaping people’s shared perception on what issues matter and how they should be dealt with. For one, the firm’s organizational members may want to use its philanthropic resources to shape the issue priority structure among stakeholders to their private preferences, either by playing up the saliency of the issue area that they personally care a lot about, or by keeping a certain issue area off the public discourse that they don’t care as much (i.e., making people care less about the issue). Also, firms’ advocacy giving can be used to shape the nature of public discourse regarding those issues to make sure that society reaches a type of solution that would benefit them. For example, LGBTQ employee groups within Fortune 500 companies often used their organizations to collectively 26 sponsor political events such as annual Gay Rights parades and to fund nonprofit groups that advocate for gay and lesbian rights (Briscoe and Safford, 2010; Briscoe and Gupta, 2016). In these scenarios, the public understanding and saliency of social issues are no longer taken as given—rather, they become the subject of endogenous change to advance their organizational members’ internal interests. Importantly, if this is the main consideration behind firms’ decision to give to advocacy, we should observe significant differences in the pattern of firms’ philanthropic giving to advocacy nonprofits vs. service nonprofits. In particular, I focus on two criteria to compare the pattern of allocation in service and advocacy giving: the level of focus and differentiation. In terms of focus, a firm can vary its philanthropic portfolio by either diversifying—i.e., spreading their giving across a wide variety of cause areas—or specializing—i.e., concentrating on a single or a few selective causes (Su and Tsang, 2015; Dorobantu, Henisz, and Nartey, 2017). Firms’ giving approaches can also vary in the level of differentiation, which is about how similar a given firm’s giving portfolio is to those of other corporate givers (Kaul and Luo, 2018). Which approach to giving would make more sense depends critically on what strategic goals a firm is trying to achieve by its philanthropic activities. First, let’s consider how the expected level of focus would vary between service and advocacy giving, if the two are indeed driven by the distinct strategic goals as discussed above. In service giving, firms give to fulfill what stakeholders want. The question is how firms would allocate their giving to achieve this goal. The chances are that firms are better off spreading their giving across a variety of causes. This is due to 27 the high level of stakeholder heterogeneity in philanthropy market, where people tend to disagree on which social areas should be prioritized (Mackey et al., 2007). Under this condition, diversification can be a superior strategic choice, as it allows the firm to appeal to stakeholders with heterogeneous preferences and thus to maximize the share of stakeholders who would value its philanthropic efforts. This is especially true because the extant research suggests that external stakeholders lack in both their capability and willingness to closely monitor and evaluate the quality of a firm’s philanthropic contribution to a given issue, thus caring only about whether or not a firm gives to the issue that they care about, but not much about how much it gives (Su and Tsang, 2015; Seo, Luo, and Kaul, 2019). For this reason, a relatively lower level of focus is likely to be preferred for firms’ service giving in general, the main strategic goal of which is to meet the given external demands of stakeholders. In contrast, if firms give to advocacy to achieve a different goal—to change the nature of issue-related public discourse to their internal organizational members’ preference—, a higher level of focus may make more sense. This is because the range of social and environmental advocacy issues that the firm’s organizational members feel passionate about is likely to be narrower than what broader external stakeholders of the firm would care about. Moreover, unlike many other types of external stakeholders such as customers or investors, employees—as internal stakeholders of the firm—have the capacity to effectively observe the firm’s donation practices, develop a lot deeper knowledge about the firm’s donation practices. Employees may also be more motivated to monitor firms’ giving practices, since they have a stronger stake in the effectiveness of 28 such efforts, with their identity often at stake in the companies’ policies and reputation. This difference means that firms may be incentivized to target their advocacy giving to the a few areas that is of primary concern to its employees. Thus, if advocacy giving is indeed driven by the strategic goal of a different nature—internally driven rather than externally driven—to service giving, firms would concentrate their advocacy giving to more selective areas of interest. Empirically, this means that firms’ advocacy giving portfolios are likely to show a significantly higher level of focus than their service giving portfolios. Another dimension of giving that helps to evaluate the qualitative difference between advocacy and service giving is the level of differentiation. As noted, in service giving, firms are incentivized to match their giving portfolios to the external preferences and demands of stakeholders. This would naturally result in a relatively low level of differentiation among firms that face similar institutional environments—that is, a lot of firms’ giving portfolios would look largely similar in terms of where and how much they give. Such a high level of similarity can result organically when companies’ efforts to map their giving portfolios to the demand structure of the environment are aggregated at the higher level (Dimaggio and Powell, 1983). Or, it can also occur by some firms’ intentional efforts to imitate other firms’ philanthropic practices. This is particularly likely when the observed actions of elite institutional actors have evolved into a shared normative and cognitive model of giving, thus exerting strong institutional pressures (Marquis, Glynn, and Davis, 2007; Marquis and Lee, 2013; Marquis and Tilcsik, 2016). On the other hand, firms’ advocacy giving, if it reflects firms’ proactive efforts to shape 29 the environment to the internal, private preferences of their organizational members, is likely to exhibit a higher level of differentiation. As discussed, firms that use their giving to shape the external environment have incentives to customize it to their distinctive profiles and preferences of their workforce. Importantly, the composition of the firm’s labor force on which firms would base the decision of how to structure their giving portfolios—which is essentially a dimension of firms’ resources and capabilities—are unevenly distributed (Barney, 1991). Even among firms within the same market environment, there exists a significant heterogeneity in the profile of their organizational membership, such that given social issues of the same nature can represent a key concern to some firms while eliciting little interest from others. Thus, contrary to the case of service giving when firms give in accordance with the external pressures of institutional stakeholders, firms’ advocacy giving portfolios would show a substantially higher level of variance when based on the goal of shaping the environment to their unique needs. Table 1.1 summarizes the descriptive statistics of the level of focus in firms’ advocacy and service giving. Comparison based on the simple count measure shows that firms’ advocacy giving goes to 4 causes on average, whereas their service giving goes to 10. The findings at the recipient level are consistent as well, with the difference growing bigger. The same pattern shows up with a more systematic measure of focus; I find that the mean value of the Herfindahl index for the distribution of advocacy giving is 0.63 at the cause level, and 0.53 at the recipient level, which compare to the corresponding 30 figures of 0.28 and 0.24 for service giving6; in other words, firms are clearly more targeted in their advocacy giving than in service giving. The following Table 1.2 tabulates the mean and median values of differentiation for advocacy and service giving. I measure the level of differentiation as the vector distance between a firm’s giving portfolio across cause areas and the average giving portfolio in the broader population. When the value of differentiation is calculated against the population mean, the average in advocacy giving is 0.17 and the median is 0.16. These figures compare to the mean value of 0.12 and median of 0.04 in service giving, showing that firms’ advocacy giving portfolios are indeed more differentiated than their service giving portfolios. Importantly, I find the same pattern with the level of differentiation calculated against the industry mean. In this result that narrows down the comparison point of firms’ philanthropic portfolios to companies in the same industry, the level of differentiation is still significantly higher in advocacy giving compared to service giving; the median level of differentiation in advocacy giving is 0.1, whereas the corresponding value is 0.04 in service giving. These systematic differences that advocacy and service giving exhibit in both the level of focus and differentiation find strong statistical support as well (p=0.000). 6 While one may wonder if this comparison is misleading given the difference in the sheer amount of giving that exists between service and advocacy giving, the correlation that exists between the amount of advocacy giving and the level of focus in my data is fairly modest, ranging from 8% (recipient-level focus) to 15% (cause-level focus), indicating that the significant gap in the level of focus between service and advocacy giving cannot be explained away as being largely associated with the differences in the absolute amount of giving itself. 31 Overall, the results point to the significant differences in how firms structure their giving portfolios when they give to advocacy vs. service nonprofits; that is, firms are significantly more targeted and differentiated in their advocacy giving than in service giving. Not only does this evidence clearly show that advocacy and service giving are qualitatively different, but it also points to the nature of the difference between the two; unlike the traditional form of philanthropic giving understood to be mostly a reactive strategy driven by external institutional pressure, advocacy giving may indeed be firms’ proactive choice directly linked to their internal needs and unique organizational profiles. ***Insert Tables 1.1 and 1.2 about here*** Importantly, this finding lends credibility to the aforementioned idea of this paper that firms may be giving to advocacy not to meet external demands of stakeholders but rather to shape such demands endogenously to suit the unique interests of their organizational members. If this is the case, we can expect that firms who have strategic incentives to care a lot about their organizational members’ demands would have the most incentive to engage in advocacy giving. To explore this possibility, I now turn to the last evidence in this section; I investigate if and how Fortune 1000 firms’ sensitivity to employees (measured by the amount of staff-related annual expense of the firm, scaled by total sales) explain their choice of advocacy giving vs. service giving. Included in the list of additional firm-level and industry-level variables as controls are: firm size (total asset), performance (ROA), operating characteristics (the amount of expenditure on advertising 32 expense, CAPEX, and R&D expense scaled by total sales), KLD strengths and concerns7 and market competition (the Herfindahl index of firms’ market shares in a given industry)8. Specifically, I run the multinomial logistic regression to investigate if firms’ social performances, controlling for other firm- and market-level characteristics, explain their decision to engage in 1) both advocacy and service giving, 2) only service giving, over not giving at all (which serves as a base outcome in the specification)9. Industry (NAICS 2 digit) and year dummies are included in the specification. To be clear, the purpose of this regression is not to test a specific prediction that implies causal effects; rather, this analysis is designed to explore which firms are more likely to give to advocacy than others, by invoking only the cross-sectional variance. The results are reported in Table 1.3. M1a explains firms’ decision to give both to advocacy and service nonprofits in a given year, whereas M1b explains the decision to give only to service nonprofits. My expectation is that firms’ decision to give both to advocacy and service, compared to the exclusive service giving decision, would be more strongly associated with their employee sensitivity, if advocacy giving reflects firms’ 7 I calculate KLD strengths and weaknesses in five major dimensions of environment, diversity, community, governance and product. Note that I exclude the philanthropy-related scores (a sub-category in KLD community dimension) in the calculation of KLD rating scores to make sure that the measured social performances are not already capturing the variation in firms’ service or advocacy giving. Also, given the high correlation between firm size, KLD strengths and KLD weaknesses, I include the orthogonalized measure of these variables in the specification, though the use of raw measures produces the same findings. 8 Appendix Table 2 reports summary statistics and correlation of variables. In all specifications, mean VIF and individual VIF for any of the variables does not exceed 2, which is below the threshold that signals a multicollinearity issue (Greene, 2012). 9 As there does not exist enough number of firm-year observations when firms engage only in advocacy giving to warrant statistical power (N: 5), these cases are excluded from the regression. 33 attempts to shape stakeholder demands to unique internal needs of their employees. The results are consistent with such expectation; as shown in M1a, employee expense-to-sales ratio is very strong and positive predictor (β=8.4178, p=0.0000) of Fortune 1000 firms’ decision to give both to service and advocacy. In contrast, M1b shows that the effect of this variable is not significantly associated with the decision to exclusively engage in service giving at conventional levels. Such stronger effect of employee sensitivity on their advocacy giving decisions persists in the M2a and M2b, where I further control for the (logged) total amount of giving and re-run the estimations. narrow down the sample of analysis to Fortune 1000s that make donation at least once during the study period10. In fact, employee sensitivity is one of the only two main regressor that has an asymmetrically stronger effect on the firms’ choice to give both to service and advocacy nonprofits, compared to the choice of exclusive service giving, across both sets of estimations. Another firm characteristic that is shown to uniquely predict firms’ decision to also give to advocacy nonprofits is KLD concern (β=0.1378, p=0.0001 in M1a; β=0.0976, p=0.0379 in M2a) but not the decision to exclusively fund service nonprofits, whereas KLD strength are the significant predictor of both giving choices. This finding that advocacy givers turn out to be the firms that have not just advantages but also disadvantages in various dimensions of social and environmental performances makes sense. As noted, if an advocacy giver’s philanthropic and social practices are primarily 10 This narrows down the baseline sample for comparison from the firm-year observations of no giving by the entire Fortune 1000 firms (N: 6,057) to observations of no giving by Fortune 1000 donor firms that give at least once during the study period (N: 1,269). 34 driven by internal organizational members’ interests, the firm is likely to be a lot more selective in which issue areas they choose to excel at, and focus their efforts on those specific areas; not surprisingly, this focused and differentiated strategy comes with the trade-off of relatively limited investment and lower performances in other areas that garner less attention from their workforce. ***Insert Table 1.3 about here*** Taken together, the collective evidence suggests that the answer to the main question of this section is yes; that is, corporate advocacy giving indeed seems to be a phenomenon distinctive from service giving. Importantly, the systematic pattern of dissimilarity we observe between the two phenomena also allows us to understand the nature of the difference between the two and the unique strategic role that advocacy giving may play in firms’ broad repertoires of nonmarket strategies. I find that, while firms that spend a lot of money in service giving also spend a lot in advocacy, the way they spend it—in terms of where and how they give—is significantly different between the two types of giving. Not only do firms allocate their advocacy giving to causes that are different from service giving, but they also spend their advocacy giving in a more targeted and differentiated manner than they do with their service giving, suggesting that advocacy giving is a proactive form of strategy that is more internally driven (based on what they want) and should thus be distinguished from a traditional reactive from of philanthropic giving that is known to be externally driven (based on what external stakeholders want). I also find that advocacy givers, compared to exclusive service givers, exhibit a far higher sensitivity to their employees; that is, advocacy givers are the 35 ones that have the most incentive to attend to unique preferences of their organizational members, who may want to change the nature of civic discourse to their private preferences by using advocacy nonprofits as an agent of such institutional change. 1.5. Distinctiveness of Corporate Advocacy Giving from Lobbying 1.5.1. Are they different? Now, I turn to the comparison of advocacy giving against its second counterpart: corporate lobbying activities. After all, the empirical evidence presented so far produced initial evidence that advocacy giving, at least to some extent, might reflect firms’ internally driven attempts to shape external stakeholder demands to their employees’ preferences. Importantly, this suggests that corporate lobbying can indeed be a close counterpart to the phenomenon of advocacy giving, in the sense that they can both be understood as nonmarket strategy of proactive nature whereby firms attempt endogenous changes to the institutional structure. In this section, I thus focus on the following question: if and to what extent corporate advocacy giving can be viewed as a phenomenon parallel to lobbying. That is, how similar and dissimilar is it to lobbying? Consistent with the previous section, I begin this section by considering the null hypothesis that posits that the two phenomena are essentially the same. According to this hypothesis, advocacy giving should share the same strategic goal to lobbying; that is, it can be understood as an effort by firms to influence institutional environments: in particular, public politics (Hillman et al., 1999; De Figueiredo and Tiller, 2001; De Figueiredo and Silverman, 2006; Bonardi and Keum, 2005; Jia, 2014). Thus, this scenario posits that firms use advocacy giving as a complementary tactic to traditional 36 lobbying activities to maximize their success rate of shaping public policies to their favor. Indeed, recent papers of Bertrand et al. (2018; 2020) find empirical evidence consistent with this conjecture, showing that firms allocate a portion of their philanthropic giving to the nonprofits that are closely connected to key policymakers for their strategic interests. Yet, it is still an open question if the primary audience whose preferences firms are targeting to shape via advocacy giving is limited to those institutional stakeholders in the policy arena of traditional lobbying activities. The first evidence I examine to evaluate the validity of this claim is how many advocacy recipients are lobbying organizations: i.e., if firms prefer to give their money to advocacy nonprofits that lobby. After all, even though political activities of tax-exempt 501(c)3 advocacy nonprofits are strictly limited by law, they can still legally engage in some extent of lobbying11, as long as not a “substantial” part of its activities is devoted to it (usually, the threshold is about 20% of an organization’s total annual expenditures). In light of this possibility, I check how many recipients of corporate advocacy giving engage in lobbying at all, and when they do, how much they spend. The findings are clear; advocacy nonprofits that receive philanthropic donations from companies are not lobbying organizations in general. On average, only 7.4% of the recipients engage in any 11 Some tax-exempt nonprofits – notably, 501(c)4 social welfare organizations – can engage in politics without many constraints. The political roles of these nonprofits have especially garnered much attention in recent years since the Supreme Court’s 2010 Citizens United v. FEC decision, which stipulated that 501(c)4 nonprofits are no longer required to disclose their donors and thus made them an ideal channel for covert influence of private money into the political arena (Werner and Coleman, 2014; Werner, 2017). Yet, the vast majority of corporate foundation giving does not go to these so-called “dark money” 501(c)4 nonprofits. The data show that a mere 0.2% of the entire corporate foundation grant goes to 501(c)4 organizations. The rest of the grants from corporations go to the 501(c)3 nonprofits, which are the focus of this study and are subject to much more stringent legal constraints in their political engagements. 37 form of lobbying. Even among the ones who do lobby, the average ratio of the expenditure spent on lobbying is merely 1.1%, and the median is 0.5%, showing that even those that do engage in lobbying are not organizations whose main organizational function is lobbying activities. The legal constraints on lobbying activities are not binding—not even close. Yet, it still does not rule out the possibility that firms can use advocacy organizations as a channel to influence public politics, albeit in a more indirect way12. One way to look at whether this is happening or not is to examine the pattern of similarity between firms’ advocacy giving and their traditional lobbying activities. The first evidence I start with is the similarity that exists in the population group who engages in corporate advocacy giving and lobbying activities—that is, similarity in terms of who does each activity. My sample frame is the entire Fortune 1000 companies during the period of 2003 to 2010. Within this sample, I compare the population of firms that give to advocacy to those that lobby. By their choice to engage in each type of activities, firms are categorized into three categories: 1) firms that engage in both advocacy giving and lobbying; 2) firms that only do advocacy giving; and 3) firms that only do lobbying. The breakdown of Fortune 1000 firms into each category is illustrated in Figure 1.6. The main two takeaways of this figure are the following. The first takeaway is that there exists a significant overlap between the population of firms who do advocacy giving and 12 Although 501(c)3 are not permitted to expend a significant portion of their activities on lobbying, they are able to advocate before administrative agencies and initiate litigation. Also, 501(c)3 groups are permitted to communicate with the Congress in an educational capacity, which lets them provide "technical" assistance (Berry, 2015). 38 lobbying; that is, a lot of firms do both (32%). This points to the relatedness that might indeed exist between the two activities, which is consistent with findings of recent studies (Bertrand et al., 2018; 2020) that a part of philanthropic giving is politically motivated. Yet, on a different note, the figure also shows that there exist a lot of firms (68% of the entire population) who choose to do one or the other, with 51% of them only doing lobbying and 17% only doing advocacy giving. Furthermore, even within the population of firms that do both, the correlation between the amount of advocacy giving and lobbying at firm-year level is fairly modest (21%), which is illustrated in Figure 1.7. Thus, firms’ choice of advocacy giving seems to be largely independent of the choice of lobbying for many firms in the population. These findings serve as an initial set of evidence against the null hypothesis that advocacy giving and lobbying are qualitatively the same phenomena. That is, advocacy giving may not simply be an alternative channel of lobbying. Rather, the correct conclusion so far seems to be that the two phenomena are significantly related, but distinctive from each other at the same time. ***Insert Figures 1.6 and 1.7 about here*** 1.5.2. Where does the difference come from? Then the next logical questions are, where does this similarity and dissimilarity come from? How should we understand corporate advocacy giving differently, especially in relation to lobbying? An important consideration that may help to distinguish the two phenomena theoretically is that, unlike what is often assumed in the public politics literature, changing government regulations and policies in a way that favors firms’ business models and operations (e.g., taxation, government subsidies) is not the sole 39 target of firms’ proactive efforts to shape the nature of institutional environment. Missing from the traditional scholarship on firms’ influence on the institutional environment built around lobbying activities is the potential influence of firms on the issues that have little direct relevance to the way firms run their businesses and are rather a part of bigger political and social debate. Yet, an emerging literature on corporate activism and related movement literature pays attention to how firms and executives are increasingly weighing in on the issue that is peripheral to their core business operation, essentially acting like activist themselves by mobilizing the public (Walker and Rea, 2014; Chatterji and Toffel, 2019; Hambrick and Wowak, 2020). Oftentimes, the target of such corporate activism is not limited to a narrow range of actors in the public policy arena such as policymakers or regulators, but rather includes other stakeholders in the private arena: activists, customers, investors, and the public in general. In other words, the influence of business to shape public agenda to their favor may extend far beyond the legislative agenda to the broader values held by those diverse constituents of society. In 2016, more than a hundred CEOs signed a public letter in opposition of proposed legislation in North Carolina that would limit transgender individuals’ access to public restrooms (Human Rights Campaign, 2016). In 2017, thirty CEOs took out a full-page ad in the Wall Street Journal criticizing President Trump’s decision to withdraw from the Paris climate accord (Winston, 2017). Sometimes, firms also engage in collaboration with social movement groups, which is shown to have not only cooptation effects on direct recipients (McDonnell, 2015) but also broader effects on the other groups in the field (Odziemkowska and McDonnell, 2019), thus influencing the nature of broader 40 movements. Importantly, what these recent findings suggest is consistent with the possibility that corporate advocacy giving might be a tool to not only influence laws and legislations around the issues with direct relevance to the firms’ core business model (which is the target of most lobbying activities), but also shape the nature of sociopolitical discourse around many advocacy issues that are less directly related to the firms’ operation yet highly salient to their stakeholders. In particular, the aforementioned finding that employee sensitivity is what distinguishes advocacy givers from exclusive service givers suggests that those advocacy issues might be what their employees are particularly concerned about; in other words, corporate advocacy giving can be, at least a part of it, understood as a channel of employee activism whereby employees use firms’ resources to express their stance on the issue and potentially shape the nature of issue- related discourse in society to their private preferences (Briscoe and Safford, 2010; Briscoe and Gupta, 2016). After all, one of the key organizational characteristics that define 501(c)3 advocacy nonprofits (that are the recipients of corporate advocacy giving) is, as already noted, clear legal constraints on their political engagement, making this possibility all the more likely. Such transformative capabilities of firms in private politics matter not only as an outside lobbying strategy for certain firms’ eventual goal of shaping laws and policies (Kollman, 1998; Walker and Rea, 2014), but also matters on its own, allowing firms to alter civic discourse and norms to their advantage, which constitutes the fundamental pillar of any institutional environment (Scott, 2001; Hiatt et al., 2009). To evaluate this possibility, I look at the industry-level variation of both advocacy giving and lobbying activities. Figure 1.8 plots which industries are top advocacy givers 41 and lobbying firms. On the one hand, there definitely exists a significant level of similarity between the two phenomena in terms of which industries are top givers. In fact, the following four industries are ranked among the top five giving industries in both advocacy giving and lobbying, albeit in slightly different order: Finance and Insurance (NAICS 52), Computer and Electronics (NAICS 33), Information (NAICS 51), Petroleum and Chemical Manufacturing (NAICS 32). A defining characteristic of these industries is that they are heavily regulated or concentrated, the importance of which has been already well documented in the lobbying literature (De Figueiredo and Richter, 2014). Clearly, the importance of regulation intensity is shared by the case of corporate advocacy giving as well. This shows that corporate advocacy giving does share the same target audience with lobbying, at least to some extent: institutional actors in public politics, whose decisions shape laws and regulations. On the other hand, the rest of the list shows quite a bit of the difference. For example, Food and Apparel (NAICS 31) is the top 5th industry that accounts for 7% of the entire advocacy giving, whereas the portion of lobbying attributable to this industry is significantly smaller at 3%. The difference is even bigger in the case of the industry sector General Merchandise Stores (NAICS 45); while this industry ranks as the top 6th corporate advocacy giving field, accounting for about 6% of advocacy giving, it only accounts for 1% of lobbying activities. Similarly, Couriers and Messengers industry (NAICS 49) does a lot more of advocacy giving than lobbying; their portion of advocacy giving is 6%, whereas it is below 2% in lobbying. These fields of disproportionately higher levels of advocacy giving than lobbying share one important characteristic that 42 they are so-called service industries whose bottom line critically depends on employees’ motivation and commitment toward customer satisfaction13. In other words, these industries have in common that they exhibit higher sensitivity to the demands and preferences of their internal stakeholders (i.e., employees) around sociopolitical issues. Taken together, the overall industry-level results provide an initial evidence that, while a significant portion of corporate advocacy giving is about changing laws and policies directly related the firm’s core operation (just like lobbying), some portion of it may indeed target changes in the nature of broader sociopolitical discourse around the issues that firms’ employees personally care about, with the goal of changing the issue- related preferences of broader audiences in the society. In other words, this finding is consistent with the idea that the key distinction between corporate advocacy giving and lobbying expenditures may lie in the differences in target audience, albeit with some level of overlap. ***Insert Figure 1.8 about here*** To get closer to the answer to the question of where the similarity and dissimilarity between these two phenomena come from, I further look at firm-level variation. In this section, I explore which firm characteristics are most strongly correlated with the choice of advocacy giving versus. lobbying. Included as additional regressors are the same set of firm-level and industry-level variables to those used in the previous 13 Of course, another key stakeholder that critically matters in these industries is customers, which raises the possibility of alternative explanations that advocacy giving is driven by customer preferences, compared to employee preferences. Yet, this pattern is inconsistent with the finding in the comparison between service vs. advocacy giving. Also, in the next section that analyses firms’ choice of lobbying vs. advocacy giving based on more granular firm-level variation, I look at varying roles of employee sensitivity and consumer sensitivity. 43 section: firm size, performance, operating characteristics, social performances, and industry competition. I run the multinomial logistic regression with industry and year dummies, which investigates which firm characteristics makes firms more likely to engage in: 1) both advocacy giving and lobbying, 2) only advocacy giving or, 3) only lobbying, over doing neither of these two (which serves as a base outcome in the specification). Again, the purpose of this regression is not to test or imply any causal relationship; rather, it is designed to explore which firm characteristics are associated with the choice of advocacy giving relative to lobbying in a more systematic manner. The results are presented in Table 1.4. One important takeaway is that advocacy giving firms and lobbying firms share a lot of key characteristics. The results show that firm size is, not surprisingly, the strong and positive predictor of both advocacy giving and lobbying activities, showing that the bigger the firm is, the more likely it is to spend money on both activities (De Figueiredo and Richter, 2014). Interestingly, the ratio of expenditure on advertising to sales, which can be understood as a proxy for firms’ sensitivity to another type of stakeholders in the private arena (i.e., customers), is shown to be another characteristic of firms that advocacy givers and lobbying firms share. As shown in Table 1.4, its effect is significant and positive in all models, indicating that firms that engage in either advocacy giving or lobbying, compared to those that do neither, spend significantly more money on advertising. Another set of firm characteristics shared by advocacy givers and lobbying firms are firms’ social performances. I find that both KLD strengths and concerns are highly significant and positive predictors of lobbying and advocacy giving, which indicates that good and bad 44 firms—the firms at the opposite ends of the social performance spectrum—are the main participants of both activities. Not only does this finding support the idea of both lobbying and advocacy giving as internally driven activities as previously noted (Delmas et al., 2016), but it is also consistent with the finding in the previous section that compared the choice of advocacy giving vs. service giving, thus adding more confidence to the overall reliability of the results. On the other hand, the second important takeaway from the result is that there do exist firm characteristics that seem to distinguish advocacy givers and lobbying companies. One important characteristic that clearly distinguishes lobbying-only firms from advocacy givers is, consistent with the pattern of results comparing service vs. advocacy giving choices, the level of employee sensitivity proxied by employee expense- to-sales ratio. Indeed, employee expense-to-sales ratio is significant and positive both in M3a (β=7.7515, p=0.0000) and M3b (β=8.9527, p=0.0402), which explains firms’ choices to engage in advocacy giving either along with lobbying or on its own. In contrast, this measure of employee sensitivity loses its significance completely in M3c whose estimation is based on the sample of firms who exclusively engage in lobbying (β=-2.2503, p=0.2316). This unique role of employee sensitivity as a predictor of firms’ advocacy giving decisions shows that advocacy giving matters among firms whose performance is highly dependent on employee motivation and loyalty, and thus tend to spend a lot more on employee-related expenses. Importantly, this suggests that, unlike lobbying, which is often understood as an activity to pursue firm-level strategic interests by changing government regulations and policies directly related to the firm’s core 45 businesses, the firm’s advocacy giving activities may indeed be driven by employees’ unique private preferences around social and environmental issues that may have little to do with firms’ core operations and business model. Overall, this pattern of findings goes back to the idea that the main target of corporate advocacy giving is not limited to the policy issues that have direct impact on corporate business model and operations. Rather, a significant portion of corporate advocacy giving is likely to target issues of fundamentally different nature that have little direct relevance to the firm’s short-term bottom line yet elicit strong interests and reaction from its internal stakeholders, suggesting that it may indeed be understood as another channel of employee activism by which employees use their firm as a vehicle for activism to communicate their ideological views on the issues of concern. ***Insert Table 1.4 about here*** 1.6. Discussion and Conclusion My paper is, to my best knowledge, one of the first systematic studies on the phenomenon of corporate advocacy giving. Taking a fact-based approach (Helfat, 2007; Oxley et al., 2010; Moeen and Agarwal, 2017), the goal of this paper is to provide a set of key empirical facts around the understudied phenomenon of corporate advocacy giving. This empirical exercise is centered around essentially one key question: is corporate advocacy giving a significant and distinctive phenomenon? The findings of this paper indicate that the answer to the main question is yes, for both. First, corporate advocacy giving is indeed an empirically significant phenomenon, in the sense that it is an action taken by the majority of firms, and their amount of spending is likely to serve as a 46 significant influence on the part of recipient advocacy organizations. Second, corporate advocacy giving is a phenomenon qualitatively different from its two nearest counterparts (service giving and lobbying activities), suggesting that it may play a unique strategic role in the broad repertoire of firms’ nonmarket strategies. In particular, the systematic patterns of dissimilarity that I find in the comparison against each counterpart are consistent with the view that corporate advocacy giving may be understood as an internally driven activities of firms to shape stakeholder demands in broad society to the unique and private preferences of their employees. For one, compared to service giving, advocacy giving is pursued by firms in a more differentiated and focused manner, especially by those that have strong incentives to care about employees’ demands. This importance of internal organizational members is also shown in the comparison against lobbying, with the result suggesting that firms seem to use advocacy giving to influence civic discourse around many advocacy issue areas of high salience to their employees’ intrinsic and private interests, unlike lobbying that is targeted to influence specific legislations and policies that are directly related to firms’ operations and business models. Of course, the findings of this paper are exploratory and thus far from conclusive; after all, the goal of this study was not to imply or test any causal claims about the role of corporate advocacy giving. However, what this paper does suggest is that corporate advocacy giving seems to be indeed a unique phenomenon that deserves a lot more scholarly attention, with the view of it as a tool to influence public discourse being one of the most promising theoretical possibilities. I hope that the key findings and multiple sets 47 of empirical facts I provide throughout this paper provide a useful ground for future empirical and theoretical research on this matter. This paper also contributes to the corporate philanthropy literature by challenging existing conceptions of what corporate giving is and how it benefits firms. By the existing accounts, firms are understood to give because their external stakeholders want them to, in expectation of rewards from those stakeholders who would appreciate firms’ philanthropic efforts (Bénabou and Tirole, 2010; Kaul and Luo, 2018). That is, firms give to make stakeholders happy by supporting the social issues that are dear to those stakeholders. However, what is missing from this line of literature is the fact that the saliency and public awareness of social issues are not exogenous to firms, and more importantly, that firms can use their philanthropy as a tool to endogenously shape the issue priority structure. This means that, while the extant view of corporate giving as a reactive tool of stakeholder management applies well to the traditional form of service giving, it does not to the case of advocacy giving, where firms can fund and use advocacy nonprofits as “issue entrepreneurs” (McCarthy and Zald, 1977). Although both forms of giving have the potential to benefit companies strategically, the findings of this study suggest that why firms give and how they benefit as a result of such giving is fundamentally different between the two types of giving, thus extending our understanding on how firms create and appropriate value from their philanthropic giving and broader CSR activities. In doing so, this study joins a recently growing body of literature that explores firms’ proactive use of CSR for their own private and political benefits (Werner, 2015; McDonnell and King, 2018). Moreover, the argument that firms 48 can shape stakeholder demands to fit their unique interests and thus influence society for better or worse connects this work to the recent literature that critically examines the welfare implications of CSR efforts (Barnett, 2016; Kaul and Luo, 2018). Finally, this paper informs the body of nonmarket research on firms’ use of proactive strategies and its implications. Firms’ strategic use of proactive strategy has traditionally been the focus in the public politics literature, with scholars studying how firms influence public officials to shape government legislations and policies in their favor (i.e., policymakers and regulators) via lobbying and campaign contributions (De Figueiredo and Tiller, 2001; De Figueiredo and Silverman, 2006; Jia, 2014). However, public actors are not the only type of stakeholders that matter to firms; included in the long list of diverse stakeholders whose roles matter are those in the private arena such as activists, customers, and the public in general, which connects to firms’ strategic need to shape demands and preferences of these private stakeholders to their favor. Indeed, the finding of this paper suggests that corporate advocacy giving may reflect attempts of firms to shape stakeholder demands in broader society and, unlike what has been suggested in a few recent papers (Bertrand et al., 2018; 2020), may be more than yet another political tool to influence the government and legislations. In particular, advocacy giving may reflect firms’ efforts to influence the nature of movements and public discourse around issues salient in broader sociopolitical discourse, which garner a lot of attention from many private stakeholders in society: in particular, the firm’s own employees. This finding suggests that firms’ proactive role in the arena of private politics may indeed be more prevalent than previously thought (Walker and Rea, 2014), thus 49 broadening our understanding of the repertoire of firm strategies used to deal with demands of private stakeholders—notably, activists. While private politics scholars have long focused on understanding firms’ defensive performance repertoire in times of activist attacks (McDonnell and King, 2013), this study shows that firms can use corporate philanthropy as a channel of proactive influence on advocacy groups and the direction of broader movements, thus joining a recent stream of research that challenges traditional scholarly focus on the reactive role of firms in the private politics literature (McDonnell, 2015; Odziemkowska and McDonnell, 2019). 50 Figure 1.1 Firm-year level distribution of advocacy giving ratio Figure 1.2 Recipient-year level distribution of dependence ratio on corporate funding 51 Figure 1.3 Cause priority in corporate advocacy vs. service giving Figure 1.4 Average correlation between advocacy and service amount across different cause areas 0% 5% 10% 15% 20% 25% 30% 35% ci vi l r ig h ts ed u ca ti o n p u b lic b en ef it in te rn at io n al co m m u n it y en vi ro n m en t fo o d h ea lt h h u m an s er vi ce s em p lo ym en t m en ta l h ea lt h so ci al s ci en ce h o u si n g ar t p u b lic s af et y yo u th cr im e an im al d is ea se s m ed ic al r es ea rc h sc ie n ce re cr ea ti o n R B W Q S C K E P J F V L A M O I D G H U N % of advocacy giving to the area % of service giving to the area 54% 50% 48% 44% 37% 36% 36% 35% 31% 31% 29% 29% 28% 25% 25% 23% 20% 19% 15% 12% 12% 0.00 0.10 0.20 0.30 0.40 0.50 0.60 co m m u n it y e d u ca ti o n e n vi ro n m e n t p u b li c b e n e fi t h u m an s e rv ic e s h o u si n g h e al th in te rn at io n al yo u th cr im e ar t m e n ta l h e al th so ci al s ci e n ce d is e as e s e m p lo ym e n t m e d ic al r e se ar ch fo o d an im al sc ie n ce p u b li c sa fe ty re cr e at io n S B C W P L E Q O I A F V G J H K D U M N 52 Figure 1.5 Average correlation between advocacy and service amount across different firms Table 1.1 Summary statistics for focus in advocacy and service giving Advocacy giving Service giving Paired t-test Mean Median Mean Median P-value Cause level Count 4 3 10 10 0.0000 Herfindahl 0.63 0.58 0.28 0.38 0.0000 Recipient level Count 10 3 122 31 0.0000 Herfindahl 0.53 0.49 0.24 0.12 0.0000 Table 1.2 Summary statistics for differentiation in advocacy and service giving Advocacy giving Service giving Paired t-test Mean Median Mean Median P-value Against Population Mean 0.17 0.16 0.12 0.04 0.0000 Against Industry Mean 0.14 0.10 0.11 0.04 0.0000 53 Table 1.3. Regression of firm characteristics on the choices of advocacy and service giving M1a M1b M2a M2b DV: Both advocacy giving and service giving DV: Only service giving DV: Both advocacy giving and service giving DV: Only service giving Size 0.5477*** 0.2618*** 0.2854*** 0.1341*** (0.0000) (0.0000) (0.0000) (0.0002) ROA 0.0686 0.0016 0.0485 -0.0158 (0.2233) (0.9871) (0.5439) (0.8997) R&D-to-sales ratio 0.7687 -1.3475 1.3199 -0.9032 (0.2316) (0.1487) (0.1037) (0.3532) Advertising-to-sales ratio 4.6094** -1.8461 2.1587 -3.4180 (0.0017) (0.3723) (0.2720) (0.1290) Capex-to-sales ratio -0.6405 -1.2975* -0.6989 -1.0832+ (0.1963) (0.0194) (0.2765) (0.0570) Employee expense-to- sales ratio 8.4178*** -0.3799 6.5788*** -1.2995 (0.0000) (0.9029) (0.0006) (0.7009) KLD strength 0.4332*** 0.2241*** 0.2528*** 0.1185* (0.0000) (0.0000) (0.0000) (0.0102) KLD concern 0.1378*** 0.0073 0.0976* -0.0098 (0.0001) (0.8638) (0.0379) (0.8310) Market concentration -2.1756 3.1153 -3.1218 2.8791 (0.3736) (0.3695) (0.3830) (0.4353) Total amount of giving 0.2799*** 0.1805*** (0.0000) (0.0000) Constant -4.4398*** -4.4155*** -3.6125*** -3.9731*** (0.0000) (0.0000) (0.0001) (0.0001) N 1038 675 1038 675 Notes: p-values in parentheses. Results of the multinomial regression. Baseline for all the models is when firms do not give at all in the entire sample of Fortune 1000 – i.e., the firm-year observations when the amount of giving is zero (N=6057). All models include year and industry dummies. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 54 Figure 1.6. Population of advocacy givers and lobbying firms among fortune 1000 Figure 1.7. Scatterplot of advocacy - lobbying amount at firm-year level Figure 1.8. Top advocacy giving and lobbying industries Doing both 32% Only advocacy giving 17% Only lobbying 51% 3 9 % 1 3 % 9 % 9 % 7 % 6 % 6 % 3 % 3 % 2 % 1 % 1 % 0 % 0 % 0 % 0 % 0 % 0 % 0 % 0 % 0 % 0 % 1 4 % 1 9 % 1 1 % 1 6 % 3 % 1 % 2 % 1 % 4 % 1 % 1 % 1 6 % 0 % 4 % 1 % 3 % 0 % 1 % 2 % 0 % 0 % 0 % 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% % of advocacy giving by the industry % of lobbying expenditure by the industry 55 Table 1.4. Regression of firm characteristics on the choices of advocacy giving and lobbying M3a M3b M3c DV: Both advocacy giving and lobbying DV: Only advocacy giving DV: Only lobbying Size 0.7282*** 0.6166*** 0.3281*** (0.0000) (0.0000) (0.0000) ROA 0.0938 0.0548 -0.0829 (0.1458) (0.5685) (0.5958) R&D-to-sales ratio 4.6359*** -5.6852** 3.0973*** (0.0000) (0.0046) (0.0000) Advertising-to-sales ratio 5.9787** 9.4849*** 5.4556*** (0.0038) (0.0000) (0.0000) Capex-to-sales ratio -0.7752 -0.1009 -0.1048 (0.2355) (0.8854) (0.7116) Employee expense-to-sales ratio 7.7515*** 8.9527* -2.2503 (0.0000) (0.0402) (0.2316) KLD strength 0.7181*** 0.6157*** 0.4616*** (0.0000) (0.0000) (0.0000) KLD concern 0.5232*** 0.2315*** 0.4650*** (0.0000) (0.0001) (0.0000) Market concentration -0.8169 1.2049 6.8010** (0.8015) (0.7409) (0.0071) Constant -5.0042*** -23.7121 -3.0365*** (0.0000) (0.9960) (0.0000) N 614 424 2155 Notes: p-values in parentheses. Results of the multinomial regression. Baseline for this model is when Fortune 1000 firms do neither of the two activities – i.e., the firm-year observations when the amount of advocacy giving and lobbying is zero (N=4577). M5 runs the OLS panel regression. All models include year and industry dummies. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 56 CHAPTER 2: EMPLOYEE IDEOLOGY AS A DRIVER OF CORPORATE ACTIVISM: EVIDENCE FROM CORPORATE GIVING TO ADVOCACY ISSUES 57 2.1. Introduction From 2018 to 2019, more than 1,000 corporations and executives have taken a stand on a wide spectrum of socio-political issues that are hotly contested and seem to have little direct relevance to their core business operations (Hogan, 2019), such as abortion rights, LGBTQ rights, gun control, immigration, and global warming, etc. Examples include Dick’s Sporting Goods and Sales Force taking a stand against gun violence; Barbara Green, a co-founder of Hobby Lobby filing a lawsuit against the mandate that required corporations to provide emergency contraceptives; more than 2,000 companies including Airbnb, Best Buy, Cargill publicly stating opposition to Donald Trump’s 2017 decision to withdraw the US from the Paris Climate Agreement and signing the “We Are Still In” pledge letter. Rise in corporate activism in recent years have begun to garner attention from scholars, with recent studies starting to examine the consequences of such activism on both firms—whether it benefits firms’ financial performances (Burbano, 2019)—and society—whether such activism can effectively shape the public discourse (Chatterji and Toffel, 2019). However, a lot remains unknown about this relatively nascent phenomenon; especially, it is unclear why firms would engage in these forms of proactive activism in the first place. After all, the convention wisdom holds that firms prefer to stay under the radar when it comes to contentious issues that most of the activism is centered around (McDonnell, 2016). As a for-profit organization, the argument goes, firms do not have the informal source of power such as public trust and legitimacy to engage the public effectively. Unlike traditional service issues (e.g., educational scholarship, 58 children’s safety, natural disaster giving) whose need elicit broad agreement among stakeholders, stakeholders—especially those with varying ideologies—often hold multiple and heterogeneous preferences around advocacy issues, disagreeing over what firms should do about the issues. Thus, while firms may make a group of stakeholders happy by taking a stand that supports their position, the argument goes, it comes with the risk of alienating a different group of stakeholders who disagree with that position, which minimizes the expected benefits from getting involved with advocacy issues. The best course of strategic action for most firms is thus to stay neutral, which makes it all the more puzzling why we are observing a lot more cases of corporate activism (Chatterji and Toffel, 2019). In this study, I offer an initial answer to this question by exploring the possibility that firms’ activism is driven by internal ideological preferences of the firm’s employees. An emerging literature on corporate political ideology shows how firms often develop a distinctive ideological preference in their workforce, with employees being drawn to firms whose employees have already similar moral and ideological preferences to themselves (Gupta, Briscoe & Hambrick, 2017; Bermiss & McDonald, 2018; Gupta & Briscoe, 2019). This suggests that firms may face relatively homogeneous preferences among their workforce on how to engage with the advocacy issues, compared to ideological composition among other types of stakeholders (e.g., consumers); thus, firms who have their own employees as primary stakeholders may be uniquely positioned to engage in the advocacy issue in a way that is aligned with stakeholder preferences, while not facing the strategic risks of alienating stakeholders that firms without such 59 homogeneous employee base would otherwise face. Moreover, ideologically charged nature of many advocacy issues suggests that stakeholders’ ideological preferences play an especially salient role, suggesting that focus on employees’ ideological preferences might allow us to understand a growing phenomenon of corporate activism. Yet, there are reasons to expect that simple extension of what we know from existing studies on the role of organizational political ideology would be insufficient to explain corporate activism. For one, most of the existing studies focus on the relationship between the direction of the firm’s ideological leaning among their top executives or broader workforce and the level of its social engagement, by showing that liberal firms are willing to invest more in CSR than conservative firms (Chin, Hambrick & Trevino, 2013; Briscoe, Chin & Hambrick, 2014; Gupta & Briscoe, 2019). Yet, unlike traditional service issues that have been the context of these studies, which focus on provisioning necessary social goods and services to those in need and thus appeal disproportionately to liberals who care more about inclusiveness and egalitarianism, advocacy issues are unique in that they tend to attract attention from people of both ideological leanings. Therefore, it remains unclear which ideological characteristics within the organization would shape firms’ engagement around advocacy issues of contentious nature, and in what ways. This paper addresses this gap by uncovering multiple ideological attributes of the organization as important antecedents that jointly affect firms’ willingness to engage in corporate activism, thus departing from the existing focus on the direction of ideological leaning. First, I begin by examining the role of the intensity of employees’ ideological 60 preferences, which refers to the importance of ideology to average organizational members’ self-concept. After all, not all employees are ideologically motivated, as shown in the extensive literature on political psychology demonstrates a significant variance across individuals in the relative importance of ideological values to their self-identity (Huddy, 2001; Mason, 2018). This variation matters, as only those with higher ideological intensity are willing to put a high price tag on the firm’s action, as they care a lot about if their firm behaves around issue areas of their interests in the way aligned with their personal values. The higher average ideological intensity among their workforce, the bigger strategic benefits the firm can thus expect from their employees in return for taking a stand on the issue. Second, I also consider additional ideological attributes of the organization, by considering how the effect of ideological intensity varies by the level of ideological homogeneity and radicalness among the workforce. My expectation is that ideological firms whose employees hold homogeneous and extreme ideological preferences are uniquely positioned to benefit from their employees’ ideological interests and willingness to reward the firm for engaging in corporate activism. I test this theoretical claim by examining one of the channels by which firms can engage in activism: donation to advocacy nonprofits, which are viewed as the moving force behind most of the major movements and cultural changes (King and Soule, 2007; Hiatt, Sine, and Tolbert, 2009). Based on a novel database comprising about $1 Bil in advocacy donations, I demonstrate the role of employees’ ideological intensity on corporate advocacy giving. While controlling for the effect of the average ideological position of the firm’s workforce, I find that firms with a higher ideological intensity, 61 proxied as firms with a higher number of employees that make campaign political contributions, do give more to advocacy issues. I also confirm the moderating influences of both ideological homogeneity and radicalness, finding that the positive effect of ideological intensity holds only for those whose members think not only similarly to each other but also hold extreme preferences. In all the estimations, I further explore varying role of ideological intensity among senior executives vs. lower-level employees, and find that it is lower-level employees’ preferences that matters, which suggests that this reflects strategic, profit-maximizing decisions of the firm to meet demands of their employees (rather than an agency behavior to advance CEO and other top executives’ personal ideological agenda, which is a potential alternative explanation). This interpretation is also consistent with the findings from the supplementary analyses, where the effect of ideological intensity is pronounced for firms with stronger governance, and those facing stronger labor market competition. First and foremost, this paper contributes to a fast-growing body of literature on corporate activism. To my best knowledge, this is the first study that examines the antecedents of firms’ engagement in socio-political issues, based on novel, large-scale empirical dataset that systematically tracks not only multiple organizations’ advocacy engagement over time but also their engagement across the entire issue universe. Due to data limitations, most existing studies have employed experimental designs or observational studies sampling on a few selective advocacy issues, which may suffer from generalizability issues. Also, while scholarly attention on this understudied phenomenon of corporate activism has recently been growing (Chatterji and Toffel, 2019; 62 Burbano, 2019; Hambrick and Wowak, 2019), the fundamental question of why firms increasingly engage in such behavior—despite the strategic disincentives to do so— remains largely unanswered. In this study, I provide solid evidence on the role of ideological intensity and other multiple ideological attributes of firms’ employees in explaining the observed rise in corporate activism. Second, this paper contributes to the burgeoning literature on corporate political ideology. As noted, most scholarly attempts have been focused on the effect of the direction of ideological leaning, demonstrating an asymmetry in the level of participation in traditional service issues between liberal and conservative firms. However, an ideologically-charged nature of advocacy issues suggests that firms at both ends of the ideological spectrum (i.e., both conservatives and liberals). This difference suggests that the simple extension of existing knowledge on corporate political ideology is insufficient to account for the pattern of corporate activism, which is a gap I address in this paper by uncovering and empirically demonstrating multiple ideological attributes of the firm that deserve more scholarly attention. Extensive supplementary analyses also reveal how ideological preferences of employees, which are often conceptualized as intrinsic, normative drivers of CSR by scholars (Chin, Hambrick, and Trevino, 2013; Gupta and Briscoe, 2019), are factored into firms’ advocacy engagement only when they are expected to translate into firms’ strategic advantages, especially in their labor market competition. 63 2.2. Literature Review Organizational political ideology, which corresponds to a conservative-liberal continuum in the US context, refers to a deeply-held core values and preferences shared by organizational members about “how the social world operates, including convictions about what outcomes are desirable and how they should be achieved” (Simons and Ingram, 1997). While initial studies have focused on the variance in ideological preferences among CEOs and other top executives (Chin, Hambrick and Treviño, 2013; Briscoe, Chin and Hambrick, 2014; Chin and Semadeni, 2016; Gupta, Briscoe and Hambrick, 2018), recent studies highlight how firms can often develop a distinctive organizational ideological leaning within their broader workforce in one way or the other, showing how many for-profit corporations can be positioned on a conservatism- liberalism continuum in the U.S. context (Gupta, Briscoe and Hambrick, 2017; Gupta and Briscoe, 2019). As a result, some firms lean “red” (conservative), while other firms lean “blue” (liberal). Such distinctive ideological leaning at the firm-level is the consequence of an organic sorting process, which is widely known as attraction-selection-attrition (ASA) model (Schneider, 1987; Schneider, Goldstiein and Smith, 1995); individuals are attracted to and stay with the firms whose executives and employees exhibit similar preferences to themselves (Van den Steen, 2010; Bermiss and McDonald, 2018). Ideological orientation of the firm’s workforce is shown to shape various firm behaviors within ideologically-valenced strategic domains, including the pattern of firms’ social actions such as philanthropic donations, employee volunteering, etc. In particular, scholars have paid a lot of attention to the asymmetry in the willingness to invest in social 64 and environmental issues that is expected to arise between conservatives and liberals. The argument is that organizational members of liberal ideologies will value CSR more, as they have stronger preferences for social equality and highly value philanthropic and other forms of social efforts that are expected to increase equality among different socioeconomic groups represented among the firm’s constituents (Skitka and Tetlock, 1993). In contrast, people with conservative ideologies have fundamentally different moral assumptions and perspectives, believing in individual responsibilities and more accepting of inequality (Poole and Rosenthal, 1984; Graham, Haidt, and Nosek, 2009). These differences suggest that firms with liberal-leaning labor force can expect higher strategic rewards in their social engagement compared to firms with conservative-leaning membership, with empirical studies finding strong support for a disproportionate level of CSR engagement among liberal firms (Gupta, Briscoe and Hambrick, 2017; Gupta and Briscoe, 2019). However, these existing arguments do not consider the fact that not all corporate social engagement is focused on the goal of “giving back” to improve socioeconomic equality. Indeed, while many firms’ CSR investments do revolve around traditional so- called service issues whose goal is centered around providing social goods and services (e.g., food and shelter) to the underprivileged in need, an emerging body of literature on corporate activism shows an increasing participation of corporate actors in the civic discourse around many advocacy issues, which tend to elicit strong reactions from people of both ideological leaning (Chatterji and Toffel, 2019; Hambrick and Wowak, 2019). Contention around socio-political issues often follow along ideological fault lines, such 65 that differences in ideological preferences among people translate into opposing positions within those issues. For example, conservatives and liberals often fundamentally disagree about the approach that should be taken to address the issues such as LGBTQ rights, immigration, abortion rights, etc. Thus, the salience of contentious issues is likely to be higher for employees with stronger ideological preferences on both ends of the ideological spectrum. In fact, liberals’ growing interest in a given issue can be precisely why conservatives are motivated to care about it more, and vice versa. Importantly, this means that the direction of ideological disposition among the firm’s workforce is unlikely to account for the variation in the firms’ willingness to engage in advocacy issues, as both liberal and conservative firms have ideological motivations to engage with advocacy issues than non-ideological firms (Hambrick and Wowak, 2019); this suggests other ideological attributes at play in explaining the variation in the strategic incentives that firms face in considering their actions around advocacy issues. 2.3. Ideological Attributes of Firms as Antecedents of Advocacy Engagement What incentivized firms to take a stand on advocacy issues? I argue that whether or not a firm engages in advocacy issues, and if so, how much, depend on the composition of the firm’s workforce in terms of their ideological characteristics. In particular, I define three ideological attributes of a firm’s internal workforce, which haven’t received much attention in the extant literature, and consider how they impact a firm’s advocacy engagement. First, ideological intensity of the workforce refers to how many employees of the firm have strong ideological preferences. After all, not every organizational members of 66 the firm are likely to have strong ideological convictions. While most individuals do have intrinsic ideological preferences that translate into varying positions with respect to many politically contested issues (Erikson and Tedin, 2003; Haidt, Graham & Joseph, 2009), not many people are willing to do anything about it, especially when promoting the position aligned with their own values entail investments of their time and money (Green, Kahneman, & Kunreuther, 1994; Halfpenny, 1999; Elfenbein & McManus, 2010; Fischbacher & Gachter, 2010). This can be attributed to the inherent nature of with many social issues as a public good (Olson, 1965); when an individual makes an active effort to promote the preferred issue position, it is hard to exclude those who don’t make such effort yet have similar ideological preferences from feeling good for knowing that their preferred issue position is being promoted. If a firm’s employees similarly exhibit such limited intention to invest their own time and money to advance the issue, the firm cannot expect significant strategic returns from their activism. Yet, extensive literature on political psychology documents a significant variance across different individuals in the perceived importance of ideological and political values to their self-identity, with some individuals who exhibit the highest interests in such values often willing to bear personal costs to advance the ideologically-charged causes (Oliver, Marwell & Teixeira, 1985; Huddy, 2001). When a firm has a sufficient number of these employees who are passionately concerned about the cause and thus willing to reward the firm by investing their own time and energy (e.g., stronger commitment and motivation to work, prosocial organizational citizenship behavior), the firm would have enough strategic incentives to engage in the issue in ways aligned with 67 their values (Chatman, 1989; Dineen & Noe 2009, Cable & DeRue 2002; Kutcher, Bragger & Masco, 2013). These employees may directly pressure firms to use their resources to advance their issue-related agenda, thus using their firm as a vehicle for activism to communicate their ideological views on the issues of concern. Even when they account for a small segment of the broader workforce, these employees with stronger preferences may play a significant role in driving firms’ activism by acting as a group that constitutes “critical mass” that is required to solve collective action problem (Hardin, 1982; Oliver, Marwell & Teixeira, 1985). By lowering interests necessary for subsequent issue-related actions by others, they may induce broader (albeit passive) participation of employees with lower interests who may act as movement allies (Lipsky, 1968), thus further facilitating firms’ engagement in activism14. H1: Firms with stronger ideological intensity give more to advocacy issues. Second, ideological homogeneity within the workforce is likely to act as one of the key contingencies that may moderate the positive relationship between ideological intensity among employees and firms’ activism behaviors. Ideological homogeneity refers to how similar the firm’s organizational members are in terms of their ideological preferences; put simply, whether they think similarly or not. The aforementioned ASA process at the organizational level is expected to promote a relatively high level of homogeneity among the workforce (Gupta, Briscoe & Hambrick, 2017; Gupta & Briscoe, 2019). Not only are individuals attracted to workplaces where existing employees and a 14 Research in political science has found that individuals who donate to political parties tend to be socially influential, ideologically minded, and motivated to promulgate their beliefs among friends and colleagues (Rosenstone and Hansen, 1993; Ansolabehere et al., 2003). 68 range of firm policies and behaviors suit them, but the firms are also more likely to select these similar-minded individuals who appear to have better fit with their organizations. Moreover, individuals with different ideologies to the prevailing ideologies in the workplace often leave the firm in search of other alternative workplaces that provide a better person–organization fit or value congruence (Bermiss and McDonald, 2018; McDonnell and Cobb, 2019). In fact, this unique clustering of similar-minded individuals within the organizations is why I consider employees’ ideological preferences as a key driver of corporate activism, compared to ideological preferences of other stakeholders such as consumers which are more likely to be characterized by a high level of divergence and heterogeneity15. Internal preferences of employees would be pronounced in the group of ideologically homogeneous firms whose employees think similarly to each other, which gives rise to a prevailing system of beliefs that would enable firms to reach internal consensus in their advocacy giving decisions as to which issues to support and in what ways (Schneider, 1987; Gupta et al., 2017). However, not all firms face skewed and homogeneous member profiles. Indeed, Gupta, Briscoe & Hambrick (2017) document the existence of many ‘purple’ firms whose employees are split in their ideological preferences, with some employees leaning blue, and others leaning red. In such an organization whose employees hold conflicting ideological preferences from each other, no clear consensus as to which issue stand to 15 Of course, this is not to suggest that the dynamics of ideological clustering may not arise among other types of stakeholders, such as consumers. A smaller, specialized firms such as Patagonia may indeed pursue an explicit differentiation strategy based on their ideological positions on salient divisive issues appeal to niche customer segments (Giarratana and Fosfuri, 2007; Barroso and Giarratana, 2013). The point is that the dynamics of ideological clustering may arise organically among the group of employees even without an explicit positioning strategy, via between- and within-firm organizational processes. 69 pick is likely to emerge. Corporate activism by such ideologically heterogeneous firm that face such challenge of preference aggregation may backfire severely by alienating a significant portion of their employee base by choosing to pick a side on the issue. The portion of employees who disagree with the firm’s chosen position may develop negative attitude toward their job and the organization because of their value incongruence with the firm, thus showing reduced engagement and motivation. Moreover, Burbano (2020) shows that this strategic penalty from different-minded employees can be disproportionately bigger than the strategic benefit the firm can expect from the portion of employees who do agree with the firm’s stance on divisive issues16. Thus, the firm characterized by a high level of within-firm ideological heterogeneity may be strategically better off staying neutral by not taking any sides, even when there exist many employees with strong ideological preferences who would like to see a firm to take an action on the cause in one way or the other. In other words, these arguments suggest that the positive relationship between higher ideological intensity in the firm’s workforce and its activism would hold only for ideologically homogeneous firms, but not for ideologically heterogeneous firms. H2: The positive association between ideological intensity and firms’ advocacy giving is strengthened by the level of ideological homogeneity. Third, I consider ideological radicalness of the firm’s employees as another ideological attribute that is expected to moderate the positive effect of ideological 16 According to Burbano (2020), this asymmetry can be partly explained a false consensus bias (Marks & Miller 1987), which refers to how individuals tend to overestimate the degree to which others share their preferences and beliefs, including the work contexts (Martinko & Gardner, 1987, Mowday 1981). 70 intensity on corporate advocacy engagement. Within the group of firms that hold ideologically aligned positions to each other, there is a range in the level of radicalness in their ideological positions (Abramowitz, 2010; Hall, 2015; Utych, 2020), such that some firms are placed on the extreme end of the ideological spectrum (i.e., ideological radicals), whereas others are placed more closely towards the middle (i.e., ideological moderates). People with extreme ideological preferences have more incentives to make substantial commitments to promote their agenda around the cause or defend against the competing positions advanced by others, because their preferred issue position is likely to be further away from the preference of the median. This means that those “ideologues” with radical ideological preferences are less likely to derive indirect utility from others’ action on the issue area, as such action is less likely to be aligned with their own, unique preferences; therefore, they are often willing to bear the costs of their action around the issue that they care about. Thus, the chances are that employees with extreme preferences are likely to put a high price tag on the firm’s action, should it choose to behave in ways aligned with their values by promoting the issue stand that they believe in. Ideological radicalness may also strengthen the willingness of the firm to act on the demands from employees with strong ideological intensity, as it amplifies the strategic benefits that the firm can expect from their activism by allowing firms to gain additional differentiation advantages. The firm that faces workforce with a more extreme ideology is uniquely positioned to take an extreme position on the issue of interest. In doing so, for example, a liberal-leaning firm with radical ideological membership can distance itself not only from the other firms that take conservative positions but also from 71 those that do take a relatively moderate liberal position. That is, ideological radicalness among the firm’s workforce allows it to occupy less crowded positioning and maximize its competitive advantages. While this notion that for-profit corporations can use ideological contention around divisive social issues as the basis for their strategic differentiation and competitive advantage is relatively new to our management and CSR literature, it harkens back to the classic model of competition between opposing parties in political science, where parties with varying ideological positions differentiate themselves by espousing different positions on advocacy issues of divisive nature (Cox, 1990; Burden, 2004; Carrillo and Castanheira, 2008; De Sio and Weber, 2014; Thomsen, 2014). H3: The positive association between ideological intensity and firms’ advocacy giving is strengthened by the level of ideological radicalness. 2.4. Methodology 2.4.1. Research Context and Data Finding a large-scale data that allows systematic tracking of multiple firms’ advocacy engagement across the entire issue universe is not easy. For this reason, most of the initial attempts made to understand the phenomenon of corporate activism have taken the approach of small-group experimental studies (Chatterji and Toffel, 2019; Burbano; 2019; Voegtlin et al., 2019), which may suffer from generalizability issues related to experimental conditions. Recent working papers have alternatively used few available archival records of firms’ performance around advocacy issues but the availability of such records limited the coverage of these studies typically to a single or a few issues 72 (Mohliver et al., 2020; Mohliver and Hawn, 2020), raising concerns on the generalizability of these findings in other issue space. Others have broadened their studies’ issue coverage based on the media report-based sampling (e.g., media articles, press release) to track corporate advocacy engagement across multiple issues (McKean and King, 2019), but this approach may suffer from the potential selection bias (Earl et al., 2004), especially the bias that the advocacy engagement that firms choose to actively advertise is more likely to be captured in the media search and thus overrepresented in the sample, whereas the actions of advocacy that do not get advertised—potentially because such activity is more geared towards meeting the demands of internal employees—are likely to be underrepresented. Finally, many media-based records of advocacy can reflect firms’ symbolic efforts rather than substantive efforts of corporate activism, i.e., a mere ‘talk’ that firms use to win the favor of stakeholders without significant outlay of resources (Hambrick and Wowak, 2019). I test my theoretical arguments by building a novel dataset which allows progress regarding several of these empirical issues. Specifically, I build a database of corporate advocacy giving: that is, firms’ support of 501c3 advocacy nonprofits via philanthropic donations. In the nonmarket literature, corporate philanthropy has primarily been conceptualized as the voluntary mechanism of corporate wealth redistribution—the way in which firms give back to society—by funding so-called service nonprofits (e.g., soup kitchen, homeless shelter) that provision various social goods and services such as food and shelter to those in need. However, the population of potential recipients of corporate giving include not only service nonprofits, but also advocacy nonprofits such as activist 73 groups and think tanks, as noted in the examples (e.g., Patagonia) above. Importantly, this suggests that corporate giving to advocacy nonprofits can be used as a novel channel to systematically track firms’ advocacy issue engagement at a large scale. As noted in previous studies (Luo, Kaul, and Seo, 2018), the main advantage using the donation records of corporations—especially, donation of corporate foundations—lies in its comprehensiveness; corporate foundations are legally required to file detailed information in their grant records annually via IRS Form 990s (Returns of Organization Exempt from Income Tax). This data is compiled from the Foundation Center, a nonprofit organization that compiles information on U.S. philanthropic foundations and their grants. Its Foundation Directory Online (FDO) database is understood to provide the single most comprehensive coverage on U.S. based philanthropic foundations and their grants, and I identify corporate foundations that are affiliated to U.S. public firms in this database and collect their grant records from 2003 to 2015. Of course, the key challenge for this study’s interest is identifying the sub-sample of corporate advocacy giving out of over 1 Mil individual records of entire corporate foundation giving made during this period, by tracking which grants specifically go to advocacy nonprofits. This requires collecting information on not just corporate grantmakers (which most of the previous studies on corporate giving focus on), but also a uniquely detailed information on the nonprofit recipients of corporate giving: most importantly, a nonprofit’s primary organizational function, based on which I can break down giving to advocacy vs. service nonprofits at an individual grant level. Following the previous studies (McCarthy and Zald, 1977; Minkoff and Agnone, 2010; Suárez, 2012), I 74 operationalize a given grant as advocacy giving if it goes to nonprofit whose primary purpose is advocacy as defined by National Taxonomy of Exempt Entities Core Codes (NTEE-CC; a standard classification system in nonprofit sector similar to SIC in for- profit sector)17. To this end, I hand collect the unique organizational identifier (EIN) on the 48,325 recipients of entire corporate foundation giving from 2003 to 2015, and then match this recipient list based on EIN to the second main database of this study: Exempt Organizations Business Master File Extract (EO BMF), which is a master file providing granular organizational characteristics on over 1 Mil U.S. tax-exempt nonprofits during my study period18. This matching process allows me to identify 22,812 records of corporate advocacy grant made to 4,021 advocacy nonprofits across 22 different issue areas from 2003 to 2015, which amount to around one billion dollars. Using this data, I can track not only a variation in multiple firms’ advocacy engagement across the entire issue universe, while minimizing the bias of sample selection by capturing the comprehensive records of corporate giving that flows into advocacy nonprofits. In addition to the data on corporate advocacy giving, this paper’s inquiry also requires the dataset on firms’ ideological characteristics. I follow the approach of identifying organizational ideologies of firms based on campaign contributions of 17 This mostly includes R (Civil Rights, Social Action, Advocacy) in NTEE major code, or 01 (alliances & advocacy) and 05 (public policy analysis) in NTEE subcode. The details of categorization scheme are reported in Appendix Table 1. 18 While the threshold was lowered to $25,000 as of 2009, I keep the $50,000 cutoff throughout the entire period to ensure the consistency in the sample of 501(c) nonprofit recipients. Also, an important limitation with this approach should be noted. As the NTEE-CC categorization system categorizes nonprofits only by their primary mission, I cannot track advocacy giving that goes to nonprofits whose core mission is not advocacy but still engages in it as a secondary function. While the extant understanding is that most of the nonprofit advocacy are performed by core advocacy groups (Kimberline, 2010), it does warrant that my observation about the prevalence of corporate advocacy giving, especially its relative prevalence compared to service giving, should be interpreted strictly as a conservative, lower-bound estimate. 75 individuals within a firm, whose empirical validity has been demonstrated in a considerable body of literature (Gupta et al., 2017; Bermiss and McDonald, 2018; Carnahan and Greenwood, 2018; Gupta and Briscoe, 2019; McDonnell and Cobb, 2019). The data on individuals’ campaign contributions are obtained from Bonica’s 2013 database on Ideology, Money in Politics and Elections (“DIME”)19. Building on the insight of recent papers (Gupta et al., 2017; Gupta and Briscoe, 2019) that firms’ CSR practices are often a reflection of the ideological characteristics exhibited by the entire body politic, not just only those of corporate elites, I collect political contributions of all individuals that report their main employer as a given firm in my sample to calculate the organization-level ideological characteristics20. Finally, a variety of other sources I draw on include Compustat, KLD Research & Analytics, lobbyview.org (Kim, 2018)21, and Labor Force statistics from the Current Population Survey, which provide additional firm-level financial and operational information. 2.4.2. Variables Dependent Variable The dependent variable of this study is the Amount of Advocacy Giving, which is measured as the dollar amount of advocacy giving that a focal firm made in a given year. To correct for skewed values, I log-transformed the variable. I also confirm the stability 19 Data.stanford.edu/dime. 20 In my analysis, I separate out the effect of organizational ideological intensity into that of ideological intensity by senior executives vs. ideological intensity by lower-level employees. 21 This is the firm-level lobbying database (Kim, 2018), which is originally generated from lobbying reports filed between 1999 and 2017 and recorded at the Senate Office of Public Records. 76 of my findings in the design using the dummy variable indicating whether a focal firm made donation to advocacy issues in a given year as an alternative dependent variable. Independent Variables Ideological Intensity of a given firm is measured as the total number of employees who make a positive amount of campaign contributions in each year as a measure of firm-level ideological intensity (note that the total number of employees is included as a control in estimations), which would capture how many employees are strongly ideologically motivated and thus likely supporters of the firm’s advocacy engagement. The Federal Election Commission (FEC) requires individuals to provide information on their employers for any campaign contribution of $200 or more, allowing me to identify individuals who identify themselves as an employee of the firms in my sample and comprehensive records of their political gift. This approach of using the number of individuals who are campaign contributors as a proxy for ideological intensity builds on a widely influential work of Ansolabehere, De Figueiredo & Snyder, 2003), which demonstrate that most of the campaign contributions, especially those of individuals, are primarily as a type of consumption good, whereby ideological individuals express their intrinsic preferences on various political and social issues. A recent paper of Bonica (2016) finds the same pattern of ideological and partisan giving in his study on the sample of political contributions by corporate decision makers. These findings suggest that organizational members’ relative amount of campaign contribution spending can proxy for the level of ideological intensity of individuals within a given firm. I also find the same pattern in my data on political contributions of individuals who are employed to 77 U.S. public companies in my sample; most of the employees who give exhibit patterns of partisan giving, with most of the employees giving to either highly liberal or highly conservative candidates, which is shown in Figure 2.1. Note that this measure of firm-level ideological intensity is time-varying, as it captures the variation in the count of employees that make campaign contributions from year to year. I also test the robustness of the results using the alternative measure of the aggregated amount of campaign contributions made by these employees in each year, which allows me to take advantage of further variation in how much these employees are willing to spend when they give. While there exists an impressive body of work has emerged that demonstrates the validity of the political ideology measure based on individual campaign contribution records, it is worth noting that the firm-level ideological intensity based on this approach is likely to correspond to a conservative, lower-bound estimate of how many employees are strongly ideologically motivated, and if so, how much. Not only do only a small percentage of those interested in ideologically- charged social and environmental issues actually make campaign contributions at all (after all, political giving is one of many channels in which people can express their ideological interests)22, but the cut-off used by FEC ($200) prevents me from capturing those who do make political donations but in an amount less than $200. Also, those who are willing to spend more than $200 are likely to be relatively senior employees who 22 According to OpenSecrets.org, “Only a tiny fraction of Americans actually give campaign contributions to political candidates, parties or PACs. The ones who give contributions large enough to be itemized (over $200) is even smaller… Just 0.51% of the United State population contributed more than two hundred dollars to federal candidates, PACs, parties and outside groups in 2015 - 2016.” (Source: https://www.opensecrets.org/elections-overview/donor-demographics?cycle=2016&display=A) 78 have enough discretionary income, which is consistent with the empirical regularity that individual campaign contributions come disproportionately from higher-income, more educated and older Americans (Pew Research Center, 2017)23. The second main ideological attribute of the organization of interest is Ideological Homogeneity of the firm, which is measured by the standard deviation of ideological positions among firm employees that made campaign contributions during my study period. Ideological scores of each firms’ individuals used to calculate this measure are derived from the common-space CFScore in DIME dataset (Bonica, 2013). Using the intuition that individuals with similar ideologies donate to similar political candidates, Bonica’s CFScore places individuals’ ideology on a scale of –2 (highly liberal) to +2 (highly conservative) based on their political contribution patterns. The third firm-level ideological attribute is Ideological Radicalness of the firm, which is measured as the absolute value of ideological scores of a focal firm’s employees who made campaign gift in a given year, averaged across the total number of employees who made a giving. Control Variables First, I control for the logged amount of Giving by Industry Peers, which indicates how much a focal firm’s industry peers made donation to advocacy issues in a given year, based on the existing studies that highlight the mimetic influences in firms’ philanthropic and social practices among firms that face similar institutional environment (Useem and Kutner, 1986; Bertels and Peloza, 2008; Marquis and Tilcsik, 2016). I define firms’ peers as those with common industry membership (NAICS 2 digit). I also control for Giving by 23 https://www.pewresearch.org/fact-tank/2017/05/17/5-facts-about-u-s-political-donations/ 79 Other Firms, which indicates whether or not firms outside the boundary of a focal firm’s industry gave to a given issue in a given year. Second, in light of the studies that highlight the effect of the direction of ideological leaning within the firm’s workforce, I include Ideological Position of a focal firm as a control, by calculating the average ideological score of the firm’s individual employees that made campaign contributions during my study period, with each individual’s ideological scores weighted by the frequency of their contributions. Third, I control for firm-level financial and operational attributes. I also include a firm’s R&D-to-Sales Ratio, Advertising-to-Sales Ratio, and Capex-to-Sales Ratio, measured by its expenditure in each area scaled by total sales, all of which have been shown to affect the firm’s philanthropic activities in previous studies (Surroca, Tribó and Waddock, 2010; Zhao and Murrell, 2016). Additional controls include a firm’s Size, which is a logged measure of its total revenue, Total Employees, which is the total count of firms’ employees, and a firm’s financial performance, measured by return on assets (ROA). I include the index of CSR rating to control for the potential role of firms’ other areas of social engagement on their advocacy giving decisions, which is measured as the firm’s net KLD score, i.e., its number of strengths minus number of concerns in four major dimensions (environment, diversity, employment, community) (Mattingly and Berman, 2006)24. I also control for Female Board Ratio, which is calculated by the number of a firm’s female board members scaled by its total board size, given the prior 24 Note that I exclude the philanthropy-related scores (a sub-category in KLD community dimension) in the calculation of KLD rating scores to make sure that the measured social performances are not already capturing the variation in firms’ service or advocacy giving. 80 evidence that documents its link to firms’ social actions (Marquis and Lee, 2013). I further control for a firm’s logged Amount of Lobbying, to account for the extent to which its advocacy giving is politically motivated and used as a complementary tactic for lobbying (Bertrand et al., 2020). The last set of controls is to account for the potential influence of external market conditions on firms’ advocacy giving decisions. Specifically, I account for the product- market competitive intensity in the firm’s main market by including the ratio of Market Concentration, measured by the Herfindahl index of market share distribution across firms in the same NAICS 2 digit code as the focal firm. In light of previous studies that highlight the importance of a firm’s geographic community in its CSR and corporate philanthropic practices (Marquis et al., 2007), I also include the size of Headquarter Population (in millions) and the Headquarter Poverty Ratio in the firm’s headquartered state. Table 2.1 reports summary statistics and correlations of variables. Mean VIF in all specifications is less than 2, with the highest individual VIF being 1.54, suggesting little reason to be concerned about multicollinearity. My sample frame is 505 U.S. public firms that are identified with at least one record of advocacy giving, and their 5,713 firm-year level observations. ***Insert Table 2.1 about here*** 2.5. Results 2.5.1. Main Results 81 Table 2.2 reports the results of a OLS regression analysis predicting the amount of a firm’s advocacy giving as its ideological intensity. In all estimations, I include year fixed effects to account for changes in the external environment around advocacy giving over time, and also firm fixed effects to alleviate concerns about unobserved firm heterogeneity. M1 reports the baseline regression with a full set of controls. It shows a highly significant and positive coefficient for Ideological Intensity (𝛽=0.0021, p=0.0010), controlling for the total number of employees. Based on the point estimate, this result means that a presence of additional 100 employees within the firm is associated with a 5% increase in the amount of advocacy giving over its mean value. In M2 and M3, I check the robustness of this finding to the alternative measures of Ideological Intensity. First, I use the logged total amount of campaign contribution made by a firm’s employees in a given year. This result making use of further variation in how much employees spent, when they give, is reported in M2. The result shows that the observed effect of Ideological Intensity (𝛽=0.1135, p=0.0179) stay to be significant and positive. Given that both this main regressor and the dependent variable of advocacy giving amount are logged, the coefficients in this model can be interpreted as an elasticity, suggesting that a 100% increase in the size of political gift made by a firm’s employees is associated with a 11.35% increase in the firm’s advocacy spending. Second, I divide the variable of Ideological Intensity among the entire workforce into Top Executive Ideological Intensity (i.e., the number of top executives who make political contribution) vs. Non-executive Ideological Intensity (i.e., the number of non- executive employees who make political contribution), and re-run the regression to see 82 ideological intensity of which group of employees is driving firms’ advocacy engagement. Specifically, individuals who are identified as a firm’s CEO and top executives in a given year (based on Execucomp) are coded as top executives in my sample, and the rest of the individuals are treated as lower-level employees. While the baseline measure of Ideological Intensity does not consider the rank of employees in an organization, if the observed positive effect of ideological intensity reflects firms’ efforts to satisfy their employees’ ideological interests in expectation of strategic returns from them as suggested in my theory, the effect should stay when we use the measure of Non- executive Ideological Intensity as the main regressor. The result of this analysis is shown in M3; it finds that the effect of Non-executive Ideological Intensity is indeed strong and positive (𝛽=0.0019, p=0.0026), whereas the effect of Top Executive Ideological Intensity is shown to be not significant, albeit positive. In economic terms, an addition of 100 employees who make campaign contribution to the firm’s workforce is associated with a 4.4 increase in advocacy giving over its mean value, suggesting that the positive relation between Ideological Intensity and advocacy giving amount in the baseline analysis (M1) came largely from ideological intensity among regular employees who are not executives. In M4, I test the robustness of the effect of Ideological Intensity to the use of alternative dependent variable: the dummy variable of advocacy giving (i.e., whether or not a firm made advocacy giving in a given year). Ideological Intensity is still shown to have a highly significant and positive coefficient (𝛽=0.0002, p=0.0003). Finally, in M5, I re-estimate my results using a system dynamic panel model (Arellano and Bover, 1995; Blundell and Bond, 1998). This is because the most salient 83 empirical threat to my estimation is the potential reverse causality—the direction of the influence from ideological intensity of the firm’s employees to its level of advocacy engagement can work in the opposite way; for example, when a firm takes a stand on an advocacy cause that appeals to liberals, the firm is likely to attract people of liberal leaning, while existing employees of conservative leaning may choose to leave the firm. While my main specification uses a lagged structure to alleviate this concern, the possibility of potential reverse causality and autocorrelation in my baseline model remains. Controlling for the recent past of a focal firm’s advocacy giving, this re- estimation still produces consistent results; the coefficient of Ideological Intensity is significant and positive, with the estimate highly similar with an estimate of the main M1 (𝛽=0.0022, p=0.0095). These overall results lend strong support for Hypothesis 1, showing that firms with higher ideological intensity—firms that have a higher number of ideologically engaged individuals as a part of their workforce—do give more to advocacy issues. ***Insert Table 2.2 about here*** 2.5.2. Moderation Analyses Having demonstrated support for my main conjecture, I now turn to consider the effect of other two ideological attributes of the organization as key contingent factors. these analyses would not only help to better understand the mechanisms underlying my observed main findings, but also to add credibility to the results. Based on the results of Model 3, which highlight varying effects of ideological intensity among executives vs. non-executives, I run all models with the measure of Non-executive Ideological Intensity 84 as my main regressor, with Top Executive Ideological Intensity as an additional control added to the full set of controls. The first ideological attribute we consider as a moderator is the level of ideological heterogeneity, which refers to whether organizational members who have strong ideological convictions think differently from each other. To test its moderating effect on the observed positive relationship between Non-executive Ideological Identity and advocacy giving, I run a split sample analysis based on the median value of within- firm ideological heterogeneity. As shown in Table 2.3, the results presented in M6a and M6b are consistent with my expectation. While the positive effect of Non-executive Ideological Intensity holds for both samples, its effect is not only stronger but also bigger in the sample of ideologically homogeneous firms with a lower level of ideological heterogeneity (𝛽=0.0038, p=0.0551), compared to the observed effect among ideologically heterogeneous firms (𝛽=0.0014, p=0.0937) that may face the risk of alienating a significant portion of their employees should they choose to take a stand on advocacy issues by making a donation. This observed pattern thus supports Hypothesis 2, showing that firms respond to the demands of ideologically motivated employees only when those employees hold similar preferences to each other. Now, I turn to test the moderating effect of the second ideological attribute: the level of radicalness in employees’ ideological preferences. To test this prediction, I split the sample into the group of ideological radicals (M7a) and ideological moderates (M7b), based on the median value of firm-level ideological position in each ideological camp of conservatives and radicals. Consistent with my expectation, the results show that having 85 ideologically engaged individuals among the workforce leads to more active advocacy engagement only when those employees hold relatively radical ideological preferences (𝛽=0.0020, p=0.0138), presumably because they are more likely to reward the firm for its action by committing their own time and energy (e.g., higher motivation and productivity). On the other hand, the effect of ideological intensity among lower-level employees is shown to be only marginally significant and positive at 10%. This finding is consistent with Hypothesis 3, suggesting that higher advocacy engagement by firms with ideologically motivated employee base is primarily coming from the group of firms with radical ideological positions, but not those with moderate positions. Note that, consistent with the previous results, the effect of Top Executive Ideology stays insignificant in all sub-samples, even in those whose employees hold homogeneous or radical preferences. In Appendix Table 2.1, I re-run these moderation analyses using the logged amount of traditional service giving as an alternative dependent variable. Interestingly, not only is the effect of Non-executive Ideological Intensity only marginally significant at 10%, I also find that both ideological attributes—ideological homogeneity and ideological radicalness—are shown to have no significant strengthening effect on the positive relationship between ideological intensity among the workforce and the firm’s service giving. If anything, interestingly, I find that the pattern is reversed, such that the effect of ideological intensity on service giving is relatively stronger for firms with employees who have heterogeneous (AM2a) and moderate preferences (AM3b). ***Insert Table 2.3 about here*** 2.5.3. Supplementary Analyses: Alternative Explanations 86 The findings so far support my posited theoretical argument that firms whose employees are ideologically motivated are more likely to engage in advocacy giving, with this effect pronounced for those whose employees hold not only homogeneous but also radical preferences. According to my theory, these observed positive relationship between ideological intensity among the firm’s workforce and its advocacy giving, and the way the effect of ideological intensity gets amplified by other ideological makeup of the workforce, reflect firms’ strategic considerations to do what’s best for them to win favor from their ideologically engaged employees and thus earn profits. However, one alternative explanation to these findings is that this observed sensitivity to the ideological preferences of internal organizational members reflects the firm’s intrinsically and normatively motivated behaviors, rather than extrinsically motivated attempts of profit-maximization. Indeed, Gupta, Briscoe and Hambrick (2017) argues how ideology-based ASA process can give rise to organization’s prevailing ethos and culture, which can be the basis of corporate social actions, with decision makers following the “rules of thumb” regarding acceptable and wise behavior. From this perspective, ideological preferences of the workforce can actually be a source of bias, with decision makers acknowledging the intrinsic merits of embracing stakeholders’ concerns in day-to-day organizational routines (Crilly and Sloan, 2012; Chin, Hambrick, and Trevino, 2013) and potentially justifying the instrumental merits of their decisions to suit their organizational values (Higgins and Molden, 2003). However, if this normative and intrinsic motivations are the central mechanism driving the positive effect of ideologically engaged organizational members on the firm’s 87 advocacy giving, we can expect that this effect would not systematically vary by the strategic incentives that make it necessary for certain firms to care and meet what their employees want. I explore this idea by running a split sample analysis based on the level of labor market competition (proxied by unemployment rate in a focal firm’s industry in each year); the intuition is that firms in a more competitive labor market environment have incentives to be more responsive to the ideological preferences of employees. The results shown in Model 8a and 8b of Table 2.4 report that the effect of lower-level employees’ ideological intensity do get stronger for firms in a higher level of competition. This evidence is consistent with the interpretation that strategic motivations are indeed the main driver behind the positive effect of ideological intensity. Another alternative explanation is that the observed effect of ideological intensity among the workforce may simply reflect agency behavior of top executives (Friedman, 1970; Ang, Cole and Lin, 2000; Tan and Tang, 2016). According to this scenario, CEOs and other top executive members may be using advocacy giving to pursue their private ideological preferences separate from other employees’ preferences, which may not necessarily be aligned with the firm’s strategic interests. Of course, this concern is already significantly alleviated by the consistent findings that demonstrate how the effect of Ideological Intensity is primarily driven by the effect of Non-executive Ideological Intensity; as shown above, Top Executive Ideological Intensity did not significantly predict a firm’s advocacy giving in most models. The observed difference in the effect of ideological intensity between top executives vs. broader workforce shows that what 88 matters is what employees—who are one of the key stakeholders that top executives have to try to satisfy—want, rather than personal preferences of CEOs and other executives. To add confidence to this result, I also run the analysis that tests how the positive relationship between ideological intensity and advocacy giving varies by the level of firm governance, which is proxied by the number of independent board members scaled by total board size. Inconsistent with the possibility that the effect of ideological intensity may reflect agentic behavior based on top executives’ personal preferences, M9a and M9b report that the observed main effect holds only for the firms with strong governance (i.e., firms with a higher ratio of independent directors), but not for those with weak governance. Of course, it is important to note that my results are correlational. As this is often the case for studies of organizational ideology, the primary concern is that the same underlying factors that may drive the ideological composition of the workforce may also drive the pattern of firms’ advocacy engagement. Specifically, the main concern is that advocacy engagement of the firm may actually lead to a change in the firm employees’ ideological attributes (as expected in the ASA process mentioned above). While I do find consistent results using a dynamic panel model that may partly alleviate these concerns, I do not have an identifying instrument to separate these effects and present causal evidence. Thus, note that my results can be interpreted as highlighting the importance of fit between the pattern of a firm’s strategy and its internal workforce. ***Insert Table 2.4 about here*** 89 2.6. Discussion First and foremost, this study makes contribution to a small but rapidly growing body of literature on corporate and CEO activism. Recently, many scholars have called for scholarly attention to the mounting evidence that corporate actors are increasingly engaging in social issues of contentious and political nature, which have long been understood to lie outside the traditional boundary of firms’ participation. Such growing trend in corporate activism is puzzling, the argument goes, especially given that these issues’ direct strategic connection to firms’ core businesses seems to be limited (Chatterji and Toffel, 2019) and firms inevitably risk alienating a significant portion of stakeholders who disagree with the firms’ position, again, given its contentious nature. Indeed, recent studies that investigated the consequences of corporate activism have found ambiguous or, in some cases, even negative performance implications of firms’ decisions to take a stand on these contentious issues (Burbano, 2019), making the need to answer the why question (i.e., why firms do this) all the more salient. This study provides an initial answer to this question by highlighting how firms characterized with unique ideological composition among their workforce are uniquely positioned to derive strategic benefits from engaging in activism, while minimizing risks and costs of potentially alienating those who disagree, thus joining few recent studies that examine the antecedents of corporate activism (Hambrick and Wowak, 2019; McKean and King, 2019). Empirically, this paper is one of the very first studies that provide systematic evidence on the antecedents of corporate engagement in contentious issues based on a large-scale dataset, 90 which captures variations in corporate engagement on a wide range of contentious issues, not only across firms, but also within firms over multiple years. Second, my study also contributes to the large body of literature on the role of corporate political ideology. The growing acknowledgement on the importance of internal values and beliefs of corporate actors has generated an impressive body of research on the role of actors’ ideology in recent years; numerous works have demonstrated that political ideology of corporate elites and more recently, that of entire organizational entity (including lower-level employees) shape firms’ willingness to contribute to social issues, shaping a wide variety of executive-level and firm-level behaviors and decisions related to social issues (Briscoe, Chin and Hambrick, 2014; Carnahan and Greenwood, 2018; Gupta, Briscoe, and Hambrick, 2018). I extend this research by exploring how multiple ideological characteristics of the organization jointly shape firms’ advocacy engagement. These moderation analyses not only allow a rich understanding of the mechanism behind the positive relationship between ideological intensity and the firm’s amount of advocacy giving, but they also present valuable additional evidence that shed further light on potential social implications of growing importance of ideology in the context of for-profit organizations, thus contributing to a small but growing body of work that questions the impact of firms’ social initiatives on society (Margolis and Walsh 2003, Luo et al. 2018). On the one hand, ideological actors’ active participation in advocacy issues and its cascade effect on others can be socially beneficial as it forces other firms to up their game, increasing the total supply level of corporate action around key socio-political 91 issues from which many firms would otherwise stayed away. In a sense, ideological actors can thus be understood as a group that constitutes “critical mass” that is required to solve collective action problem in the provision of social goods and induce broader participation of actors, as they are willing to pay the initial start-up costs of institutionalizing a given issue as a legitimate area of firms’ participation in general. On the other hand, ideological firms’ influence on others’ pattern of philanthropic provision may undermine social welfare, given that the norms they create might not be representative of the true needs of society. After all, ideological actors are likely to be the actors biased with their own beliefs and preferences; it is thus possible that the power that ideological actors get to exercise as agenda setters may result in distributional inefficiency in the supply of philanthropic support across various social issues; the chances are that certain social areas that are most salient to and favored by ideological actors would be over-provisioned with philanthropic support, leaving other areas under- provisioned. Moreover, my finding that corporate activism may be driven disproportionately by firms whose employees are not only ideologically engaged but also hold particularly homogenous and extreme preferences raises concerns that corporate participation in advocacy issues may result in increased gridlock and polarization (Abramowitz, 2010; Mason, 2018). 92 Figure 2.1. Individual employee-level distribution of ideological positions Figure 2.2. Firm-level distribution of ideological intensity 93 Table 2.1. Summary Statistics and Correlation Tables VARIABLE MEAN SD MIN MAX 1 (Logged) amount of advocacy giving 4.304 5.502 0.000 17.009 2 Dummy of advocacy giving 0.392 0.488 0.000 1.000 3 Ideological intensity 78.914 193.561 0.000 2671.000 4 Executive ideological intensity 1.811 2.390 0.000 13.000 5 Non-executive ideological intensity 77.103 192.580 0.000 2668.000 6 Giving by industry peers 14.496 3.217 0.000 17.413 7 Giving by other firms 17.656 0.298 16.955 18.112 8 Ideological position -0.190 0.594 -1.595 3.074 9 Size 8.004 2.482 0.000 13.061 10 # of total employees 36.385 94.087 0.000 2100.000 11 ROA 0.043 0.083 -2.283 0.953 12 R&D-to-sales ratio 0.023 0.082 0.000 3.351 13 Advertising-to-sales ratio 0.011 0.028 0.000 0.371 14 Capex-to-sales ratio 0.059 0.111 0.000 2.955 15 Board female ratio 0.112 0.110 0.000 0.556 16 Lobbying amount 5.680 6.767 0.000 17.911 17 CSR rating 0.663 2.055 -7.000 13.000 18 Market concentration 0.027 0.030 0.011 0.276 19 Population in HQ state 13.075 10.239 0.567 39.250 20 Poverty rate in HQ state 13.076 2.753 6.800 25.800 94 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 1.00 2 0.97 1.00 3 0.18 0.13 1.00 4 0.13 0.09 0.42 1.00 5 0.18 0.13 1.00 0.41 1.00 6 0.09 0.08 0.07 -0.01 0.07 1.00 7 0.01 0.02 -0.04 -0.07 -0.04 -0.27 1.00 8 -0.04 -0.02 -0.05 0.11 -0.05 0.04 -0.17 1.00 9 0.27 0.22 0.32 0.30 0.32 0.02 0.10 0.11 1.00 10 0.21 0.15 0.35 0.09 0.35 -0.05 0.05 -0.04 0.37 1.00 11 0.02 0.01 0.01 0.03 0.01 -0.08 0.06 -0.03 0.16 0.07 1.00 12 0.01 0.00 0.05 0.03 0.05 0.08 0.03 -0.10 0.04 0.00 -0.08 1.00 13 0.08 0.06 0.00 -0.07 0.00 -0.02 0.07 -0.20 0.09 0.05 0.14 0.07 1.00 14 0.02 0.03 -0.04 0.05 -0.04 -0.10 0.17 0.07 0.06 -0.03 -0.10 0.26 -0.04 1.00 15 0.20 0.17 0.17 0.24 0.17 0.05 0.16 -0.02 0.35 0.15 0.12 0.02 0.12 0.03 1.00 16 0.18 0.14 0.18 0.29 0.18 -0.03 0.08 0.04 0.36 0.15 0.09 0.13 0.06 0.06 0.29 1.00 17 0.17 0.13 0.25 0.15 0.25 0.12 -0.06 -0.19 0.20 0.12 0.10 0.16 0.19 -0.08 0.34 0.18 1.00 18 0.00 -0.01 0.01 -0.03 0.01 -0.31 0.20 -0.10 0.07 0.37 0.07 -0.05 0.13 0.01 0.04 -0.02 -0.02 1.00 19 0.00 -0.02 0.14 0.00 0.14 0.03 0.05 -0.19 0.10 0.02 0.05 0.21 0.13 0.04 0.05 0.04 0.22 -0.03 1.00 20 0.02 0.00 0.13 0.04 0.13 -0.08 0.20 -0.03 0.13 0.11 -0.04 0.01 0.05 0.07 -0.03 0.06 0.03 0.05 0.40 1.00 Notes: Number of firm-year observations for all variables is 5,713. Missing values of all variables are recoded as zero. 95 Table 2.2. Main Results (M1) (M2) (M3) (M4) (M5) Main Alternative ideological intensity Alternative DV Dynamic panel Size of political gift Exec vs. non- exec gift Ideological intensity 0.0021*** 0.0002*** 0.0022** (0.0006) (0.0001) (0.0008) Size of political gift from firm employees 0.1135* (0.0478) Executive ideological intensity 0.0767 (0.0619) Non-executive ideological intensity 0.0019** (0.0006) Giving by industry peers 0.0050 0.0038 0.0060 0.0004 0.0332 (0.0395) (0.0393) (0.0395) (0.0036) (0.0504) Giving by other firms 0.4860 0.3610 0.5048 0.0496 2.2626 (1.0975) (1.0995) (1.0985) (0.1081) (1.3931) Ideological position 0.3637* 0.3823* 0.3386+ 0.0366* 0.6270** (0.1723) (0.1730) (0.1736) (0.0174) (0.2403) Size 0.1928** 0.1902** 0.1911** 0.0198*** 0.0268 (0.0627) (0.0623) (0.0624) (0.0058) (0.1029) Total employees 0.0007 0.0025 0.0010 -0.0003 -0.0189* (0.0041) (0.0042) (0.0041) (0.0003) (0.0080) ROA -2.6400*** -2.6508*** -2.6669*** -0.2474** -2.9785** (0.7867) (0.7868) (0.7816) (0.0768) (1.1554) R&D-to-sales ratio -1.7250* -1.5635+ -1.7164* -0.1526+ -1.6776 (0.8286) (0.8239) (0.8233) (0.0835) (1.5111) Advertising-to-sales ratio -9.0955 -9.2904 -8.9736 -0.8403 2.6372 (9.1869) (9.0659) (9.1579) (0.8693) (10.7846) Capex-to-sales ratio 1.4875 1.4375 1.4461 0.1089 -0.4353 (1.0497) (1.0447) (1.0431) (0.1045) (1.8276) Board female ratio 0.4005 0.4453 0.3926 0.0821 0.0561 (1.1995) (1.1837) (1.1986) (0.1128) (1.6889) Lobbying amount 0.0276 0.0260 0.0266 0.0013 0.0091 (0.0292) (0.0293) (0.0293) (0.0028) (0.0371) CSR rating -0.0196 -0.0159 -0.0201 0.0010 0.0716 (0.0551) (0.0552) (0.0551) (0.0050) (0.0699) Market concentration -5.6521 -6.2192 -6.0007 -0.4749 5.8902 (14.7604) (14.7381) (14.7164) (1.3880) (21.8010) Population in HQ state 0.2075 0.2140 0.2007 0.0233 0.1673 (0.1673) (0.1691) (0.1664) (0.0151) (0.1741) Poverty rate in HQ state 0.1293* 0.1303* 0.1236+ 0.0104+ 0.1631* (0.0638) (0.0640) (0.0643) (0.0061) (0.0740) Lag(amount of advocacy giving) 0.1989*** (0.0208) Constant -10.3575 -9.2786 -10.6445 -1.0766 -41.7548+ (19.1460) (19.1675) (19.1508) (1.8865) (24.2518) Year FE Y Y Y Y Y Firm FE Y Y Y Y Y R2 0.0804 0.0795 0.0808 0.0790 N 5713 5713 5713 5713 5713 Notes: Dependent variable is logged amount of advocacy giving of a focal firm in a given each year, except for M4 that uses dummy variable of advocacy giving as an alternative DV. Coefficients of OLS panel regressions. Standard errors in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 96 Table 2.3. Moderation Analyses (M6a) (M6b) (M7a) (M7b) Ideological heterogeneity Ideological radicalness High Low High Low Executive ideological intensity 0.1677 -0.0480 0.0266 0.1143 (0.1158) (0.0958) (0.0893) (0.0838) Non-executive ideological intensity 0.0014+ 0.0038+ 0.0020* 0.0016+ (0.0008) (0.0020) (0.0008) (0.0009) Giving by industry peers 0.0086 -0.0245 -0.0757 0.0743 (0.0591) (0.0481) (0.0461) (0.0605) Giving by other firms 1.3665 1.6030 -0.2374 0.9268 (2.1178) (1.9465) (1.3713) (1.7028) Ideological position 0.5547 0.4349 0.4980* 0.1104 (0.3734) (0.3564) (0.2160) (0.2829) Size 0.1971+ 0.1994* 0.1873* 0.2155* (0.1185) (0.0890) (0.0787) (0.1035) Total employees 0.0015 0.0075 0.0033 0.0010 (0.0042) (0.0094) (0.0098) (0.0047) ROA -1.7627* -4.1004* -3.9624*** -1.5136 (0.8464) (1.6282) (1.0955) (0.9200) R&D-to-sales ratio -2.1242 -1.8335 -1.3132 -8.0878** (1.4158) (2.6143) (0.8952) (3.1102) Advertising-to-sales ratio -26.7293* 2.2839 -15.0673 11.1326 (10.9986) (11.0947) (9.8397) (12.9256) Capex-to-sales ratio 2.3286 1.5332 0.6037 3.0392 (3.3622) (1.5450) (1.0365) (2.4156) Board female ratio 2.0991 1.3718 -0.3327 0.9486 (1.6779) (1.8501) (1.8113) (1.6001) Lobbying amount 0.0359 -0.0100 0.0546 -0.0037 (0.0393) (0.0419) (0.0443) (0.0393) CSR rating -0.0329 0.0349 -0.0615 0.0120 (0.0722) (0.0978) (0.0903) (0.0680) Market concentration 7.4827 -24.2295 -6.7799 -5.0746 (21.4074) (19.6370) (19.2048) (23.5797) Population in HQ state 0.0067 -0.0737 0.3586 0.1037 (0.1208) (0.3859) (0.2379) (0.1903) Poverty rate in HQ state 0.1282 0.1041 0.0934 0.1617+ (0.0922) (0.1096) (0.0919) (0.0893) Constant -23.5869 -25.1905 1.3498 -18.1532 (36.8856) (34.0490) (24.3319) (29.3435) Year FE Y Y Y Y Firm FE Y Y Y Y R2 0.0834 0.0592 0.1003 0.0714 N 2475 2507 2866 2847 Notes: Dependent variable is the logged amount of advocacy giving of a focal firm in a given each year. Coefficients of OLS panel regressions. Standard errors in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 97 Table 2.4. Supplementary Analyses (M8a) (M8b) (M9a) (M9b) Labor market competition Corporate governance High Low Strong Weak Executive ideological intensity 0.0944 0.0512 -0.0010 0.1002 (0.0822) (0.0902) (0.0765) (0.0999) Non-executive ideological intensity 0.0022** 0.0016+ 0.0017* 0.0026 (0.0008) (0.0009) (0.0007) (0.0017) Giving by industry peers 0.0699 -0.0567 -0.0256 0.0360 (0.0575) (0.0569) (0.0662) (0.0450) Giving by other firms 0.2510 4.8851 1.0256 0.0248 (1.5215) (3.6347) (1.7981) (1.3880) Ideological position 0.7102** 0.0730 0.3506 0.2524 (0.2313) (0.2624) (0.3219) (0.2032) Size 0.1548* 0.2373** 0.6665* 0.1701** (0.0760) (0.0879) (0.2625) (0.0622) Total employees 0.0014 -0.0047 -0.0026 0.0044 (0.0053) (0.0081) (0.0053) (0.0050) ROA -1.5023 -3.1491** -2.3187* -2.8300** (1.5296) (0.9519) (1.0479) (1.0844) R&D-to-sales ratio 2.1496 -2.6171* 0.5601 -1.7594** (2.2958) (1.2867) (2.8012) (0.6559) Advertising-to-sales ratio -7.1709 -11.3353 1.0639 -13.9328 (14.6315) (8.2536) (14.1018) (10.7828) Capex-to-sales ratio 1.1769 0.8555 3.5388 0.8913 (1.6440) (1.4591) (2.1545) (0.9495) Board female ratio -0.5119 1.1790 -0.6685 1.4537 (1.6247) (1.7295) (1.7898) (1.6474) Lobbying amount 0.0248 0.0497 -0.0192 0.0626 (0.0365) (0.0445) (0.0420) (0.0392) CSR rating -0.0009 -0.0835 -0.0649 0.0111 (0.0792) (0.0749) (0.0678) (0.0988) Market concentration -2.9469 -30.9570 -31.4737 24.0177 (26.6294) (18.9577) (20.5560) (19.1469) Population in HQ state 0.0935 0.1426 -0.0052 0.3161 (0.2720) (0.2400) (0.2549) (0.2541) Poverty rate in HQ state 0.1596+ 0.0570 0.2726** -0.0035 (0.0878) (0.1030) (0.0934) (0.0874) Constant -6.0965 -83.6283 -20.5709 -4.1473 (26.4420) (62.9039) (30.8758) (24.3886) Year FE Y Y Y Y Firm FE Y Y Y Y R2 0.0907 0.0832 0.0566 0.1349 N 3307 2320 2844 2869 Notes: Dependent variable is the logged amount of advocacy giving of a focal firm in a given each year. Coefficients of OLS panel regressions. Standard errors in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 98 CHAPTER 3: COMPETITIVE SPILLOVER EFFECTS OF CORPORATE ACTIVISM: CONSEQUENCES OF HETEROGENEITY IN ACROSS- FIRM IDEOLOGY IN FOLLOWER POPULATION 99 3.1. Introduction Recently, there have been increasing cases of corporate activism, with firms taking a stand on advocacy issues of contentious nature, such as global warming, immigration, abortion rights, gun control, LGBTQ rights, etc. In some cases, top executives and employees take a public stand on issues, as seen in a recent New York Times ad which CEOs of Ben & Jerry’s, Bloomberg, Twitter, and many others signed to oppose state abortion bans in 2019 (Chatterji and Toffel, 2019; Hambrick and Wowak, 2019). In other cases, firms collaborate with advocacy groups such as activists or think tanks specialized in framing civic discourse (Rich, 2005; Walker, 2014; McDonnell, 2016). For example, Patagonia funded $10 Mil to grassroots advocacy groups fighting climate change in 2018. According to a growing body of literature on corporate political ideologies (Chin, Hambrick and Treviño, 2013; Gupta, Briscoe, and Hambrick, 2017; Carnahan and Greenwood, 2018; Bermiss and McDonald, 2018; McDonnell and Cobb, 2019), instances of corporate activism reflect ideological preferences of the firm’s employees around advocacy issues of contentious nature, the contention around which follows ideological fault lines. Firms whose employees hold a strong ideology (i.e., strong conservatives or liberals) have internal incentives to engage in these ideologically- charged issues that their employees care about, as doing so promotes greater employee loyalty and commitment (Bode, Singh, and Rogan, 2015; Burbano, 2016; Flammer and Luo, 2017; Carnahan, Kryscynski, and Olson, 2017), even when clear external norms around such action are yet to emerge. 100 This rising trend in corporate activism led by ideological firms is often celebrated in the media as a catalyst for broader social change around the issue (Bersin, 2018; Murray, 2019). Given the extant focus in the literature on the direct effect of a focal firm’s ideology on its own social actions, however, we are yet to study the implications of these ideological firms’ actions on the broader field: in particular, if initiating ideological firms can indeed enable widespread issue support from other firms, essentially acting as “issue entrepreneurs” that put an issue into play by raising its salience in their field (McCarthy and Zald 1977, p. 1215). The existing accounts of how firms influence others’ social actions based on practice diffusion (Briscoe and Safford, 2008; Briscoe, Gupta, and Anner, 2015) and normative pressures (Marquis, Glynn, and Davis, 2007; Marquis and Tilcsik, 2016) support this possibility, suggesting that initial action of ideological firms would be closely followed by other firms in the same stakeholder environment—notably, industry peers—as a useful cue to guide their own social actions. However, what remains underexplored is the possibility that initiating ideological firms’ advocacy may open the door for not just their ideological allies, but also their ideological opponents, whose organizational members hold opposing issue positions. After all, polarizing nature of many advocacy issues points to this very possibility of ideological contestation (Meyer and Staggenborg, 1996; McCright and Dunlap, 2000; Barnett and Woywode, 2004; Blee and Creasap, 2010), suggesting that ideological opponents may respond to initiating firms’ actions by joining in the debate around the given issue with a different position (e.g., by funding nonprofits whose issue position is 101 aligned with their own preferences). Moreover, the imperatives of strategic differentiation (Porter, 1997; Baum and Haveman, 1997; Zhao et al., 2017) suggest a potential asymmetry in the response between the two groups, ideological opponents can gain additional strategic benefits of differentiation in the eyes of their organizational members by distancing themselves from original contributors in their issue position. These arguments suggest that ideological firms’ initial activism to promote the cause may backfire by inviting a disproportionately stronger response from their ideological opponents compared to their allies, thus contributing to increased gridlock and conflict around the issue. To test these theoretical arguments, I build a novel database on corporate advocacy giving, which corresponds to one of the few ways to systematically track corporate activism across multiple issues on a large scale. By collecting uniquely detailed information on the grant records of U.S. corporate foundations and their recipients from 2003 to 2015, I test my predictions using over 22,812 records of advocacy grants from 441 corporations made to 4,021 501(c)3 advocacy groups across 22 different issue areas, comprising about $1 Bil in donations. My analyses show that while ideological firms’ giving encourages their ideological allies’ giving in the issue area of their choice, it also invites their ideological opponents’ giving, with the response from the opponents often dominating the response from the allies. Such selective counter-response from ideological opponents is stronger among firms with homogeneous and radical preferences, and in swing states, consistent with my theory that this response reflects the dynamics of within-issue competition between firms that lie at the opposite ends of the 102 ideological spectrum. My findings survive my most stringent specifications with firm- issue pair fixed effects, issue-level time trends, and a battery of robustness checks. In the supplementary analyses, I further confirm the pattern of ideological clustering in firms’ recipient choices, finding an evidence that firms with different ideologies do fund different nonprofits within each issue area. First, this study contributes to an emerging body of literature on corporate activism and especially the line of scholarship that highlights the role of corporate ideology as the main driver for such action (Chin et al., 2013; Briscoe, Chin and Hambrick, 2014; Hambrick and Wowak, 2019; McDonnell and Cobb, 2019; Chatterji and Toffel, 2019; Burbano, 2019), by answering the open question of how ideological firms’ advocacy actions affect other firms in the broader field. An answer to this question matters not only resolves a theoretical ambiguity about the indirect spillover effects of corporate activism, but also may have important societal implications. While media accounts often celebrate corporate activism as a catalyst for social change that enables broader issue support in society, my paper offers a strategic differentiation-based account of corporate social counterpositioning (Porter, 1997; Zhao et al., 2017), which shows that ideological firms’ actions to promote a cause may backfire by creating a stronger economic incentive for opposing voices of their ideological opponents. In doing so, I challenge the current, narrow conceptualization of socio-political contention around social issues as a source of strategic liability that presents risks of alienating stakeholders, by highlighting how such contention can be leveraged by firms as a unique source of competitive advantage and value creation. 103 Second, my work extends our understanding of how firms influence others’ social actions. Most studies on this topic have focused on the normative pressures and homogeneity by showing how a uniform consensus emerges on an appropriate form of corporate social action (Marquis et al., 2007; Raffaelli and Glynn, 2014, Marquis and Tilcsik, 2016), paying limited attention to contested and fragile nature of the diffusion process. My paper addresses this gap by highlighting stakeholders’ heterogeneous preferences around many social issues shown in broader work in political psychology, based on which I argue that the firm’s initial activism can alienate stakeholders who disagree and create strong economic incentives for its rivals to take a different issue position. In doing so, this paper highlights that what may look like smooth mimicry on the surface can mask dynamics full of contention in the way that corporate social practices are diffused throughout the field. I also uncover the across-firm heterogeneity in the ideological position of firms’ organizational members as an underexplored source of a variation—and even a direct contention between firms—in the diffusion process, thus enriching our knowledge on how firms respond to external pressure in diverse ways (Oliver, 1991; Briscoe and Safford, 2008; Fiss, Kennedy and Davis, 2012). Further, my study points to additional ideological attributes of firms including ideological homogeneity and radicalness, and the level of institutional uncertainty as key organizational and environmental contingencies that moderate the posited asymmetry in the response between ideological allies and opponents. These moderation analyses not only allow a rich understanding of the mechanism behind the observed competitive response from ideological opponents, but they also present valuable additional evidence 104 that shed further light on potential social implications of this strategy of socio-political differentiation. Specifically, I find that competitive entry from ideological opponents is driven by those with extreme and homogeneous preferences, and also concentrated in swing states, which further raise concerns that this dynamic may indeed result in increased gridlock and polarization, thus contributing to a small but growing body of work that questions the impact of firms’ social initiatives on society (Margolis and Walsh 2003, Luo et al. 2018). 3.2. Literature Review 3.2.1. Corporate Political Ideologies and Advocacy Engagement From 2018 to 2019, more than 1,000 corporations and executives have taken a stand on a wide spectrum of advocacy issues that are hotly contested and have little direct relevance to their core business operations (Hogan, 2019). Such recent rise in corporate activism has begun to garner attention in the nonmarket literature, with scholars calling for studies that answer why a lot of firms are taking a stand on hot-button issues, the contention around which has long been understood to make firms stay away (Chatterji and Toffel, 2019; Burbano, 2019). One of the most salient explanations for this question can be found in the growing body of literature that focuses on the internal preferences of a firm’s organizational members: especially, the role of organizational political ideology (Chin et al., 2013; Briscoe et al., 2014; Gupta et al, 2017; Carnahan and Greenwood, 2018; Gupta and Briscoe, 2019). Defined as the core values and preferences shared by a collectivity about “how the social world operates, including convictions about what 105 outcomes are desirable and how they should be achieved” (Gupta et al, 2017; p. 1019), organizational political ideology refers to the notion that many organizations—including for-profit corporations—exhibit a strong ideological leaning among their membership, allowing themselves to be positioned on a conservatism-liberalism continuum in the U.S. context. Such distinctive ideological leaning at the firm-level is the consequence of a sorting process, whereby individuals are attracted to and stay with the firms whose executives and employees exhibit similar preferences to themselves (Van den Steen, 2010; Bermiss and McDonald, 2018). This ideological clustering of like-minded individuals shapes the pattern of corporate social actions, prompting firms to support social issues that are aligned with their members’ internal preferences (Chin et al., 2013; Briscoe et al., 2014; Gupta and Briscoe, 2019). Such role of corporate political ideology is pronounced in firms’ engagement in advocacy issues of contentious nature, contention around which often follow ideological fault lines (Hambrick and Wowak, 2019; McDonnell and Cobb, 2019). Differences in the ideological preferences among people translate into different and even opposing positions within those issues (Skitka and Tetlock, 1993; Jost, 2006; Haidt, 2012), with conservatives and liberals often disagreeing over how to approach the issues of global warming, gun control, LGBTQ rights, etc25. The salience of advocacy issues is thus higher for firms whose workforce hold strong ideological preferences; naturally, 25 According to 2018 Gallup survey, only 1% of the self-identified Democrats are skeptical about climate change, whereas 45% of the Republicans hold skeptical views on the issue. Similar differences are observed for other contentious advocacy issues such as same-sex marriage (supported by 83% of Democrats and only 44% of Republicans), gun control (supported by 87% of Democrats and 31% of Republicans), and abortion rights (supported by 39% of Democrats and only 12% of Republicans). 106 employees of these ideological firms may want to use corporate giving to advance their issue-related agenda, thus using their firm as a vehicle for activism to communicate their ideological views on the issues of concern (Briscoe and Safford, 2010; Briscoe and Gupta, 2016). These ideological firms have greater internal incentives to take a stand on advocacy issues that are dear to their employees, even when contentious nature of these advocacy issues prevent clear external norms from emerging and deter other firms’ participation. Such proactive advocacy engagement may allow firms to enjoy greater motivation and loyalty from their workforce who would appreciate such action (Bode et al., 2015; Burbano, 2016; Flammer and Luo, 2017; Carnahan et al., 2017; Bode and Singh, 2018). Ideological firms thus act as the main and early participants in the market of advocacy issues. For example, ideological firms with strong liberal values were more likely to accept the formation of LGBTQ employee groups within their organizations, and these groups often used their organizations to collectively sponsor political events such as annual Gay Rights parades and to fund nonprofit groups that advocate for gay and lesbian rights (Briscoe and Safford, 2010; Briscoe and Gupta, 2016). 3.2.2. Spillover effects of ideological firms’ contributions The following question is, would ideological firms’ proactive support of advocacy issues encourage other firms to follow suit? Because most studies in the corporate ideology literature focus on investigating the nature of the relationship between a focal firm’s ideology and its own social actions, we are yet to study if and how advocacy by ideological firms affect the issue-related behaviors by other firms in the population. Yet, this is a serious oversight, given that how a firm’s activism impacts both it and society 107 would critically depend on how other firms would react to it (Hunt, Benford and Snow, 1994; Briscoe and Safford, 2008; Briscoe et al., 2015). In particular, the key question of interest is if an ideological firm’s action can encourage participation of others that would create broader issue support, as often assumed in the media accounts that celebrate corporate activism as a catalyst for social change (Aiello, 2017; Bersin, 2018; Murray, 2019). Existing scholarly accounts of diffusion and institutionalization of general corporate social responsibility (CSR) practices do suggest that such spillover effects may arise. Scholars have found that firms tend to give to the social and environmental issues that other firms gave to, showing a surprisingly high level of similarity in the portfolio of giving across competitors in the same industry (Useem and Kutner, 1986; Raffaelli and Glynn, 2014, Marquis and Tilcsik, 2016). For instance, housing characterizes the focus of corporate giving in the finance industry, whereas corporate giving in the electric utilities sector has long focused on the issue of technology education (Kirchberg, 1995). This pattern of homogeneity, the argument goes, arises from the uncertainty that many firms face in their decisions around how much to give, and where to give (Marquis et al., 2007; Marquis and Tilcsik, 2016; Peterson and Su, 2017). In face of such high uncertainty, firms closely observe and follow the pattern of giving by other firms that face a similar stakeholder environment—notably, industry competitors—as a valuable cue that can guide their own social actions. Rivals’ activism may signal public interest for the social issue (Yue et al., 2013; Briscoe et al., 2015). Also, it may endogenously create audience who increasingly expect other firms to speak up as well (Lipsky, 1968; Andrews, 2001), 108 thus putting an issue into play and creating rare windows of sociopolitical openness for institutional change (Cress and Snow, 2000; Meyer, 2004). In doing so, firms can essentially act as issue entrepreneurs—the role traditionally assumed to be that of extra- institutional actors such as social movement organizations (Brayden and Mae?, xx)—that create and amplify the salience of the issue they care about (McCarthy and Zald, 1977). Overall, this argument suggests that ideological firms’ activism may indeed encourage activism from other competitors, potentially creating broader support for the issue that they hoped to champion. 3.3. Possibility of Bifurcated Response from Followers However, this simple extension of existing accounts of CSR diffusion fails to consider a potential variation in the way that other competitors can react to an initiating ideological firm’s activism; specifically, potential followers can vary significantly with respect to not just whether they will follow the initiating firms’ issue engagement, but also how they will engage with the issue. Specifically, while some followers may expand issue support by advocating for the issue position of the original contributors, others may choose to promote different or even opposing positions in their issue engagement. This varied response is especially likely when firms have reasons to believe that not everyone agrees with original givers’ issue stand, as would be in the case of corporate activism that revolves around advocacy issues of socio-politically contentious nature. Existing literature highlighting mimetic influence in CSR tends to implicitly assume a uniform consensus that emerge on what constitutes normatively appropriate behavior. However, unlike so-called service issues whose need elicit broad agreement among most 109 stakeholders (e.g., providing food and shelter to people in need)—which has been the focus of most CSR studies—, recent studies linking insight from political psychology and movement literature to CSR have started to highlight that a substantial number of people take a stand both for and against advocacy issues. Contention around these issues often follows ideological fault lines, with conservatives and liberals holding different and often competing views on many advocacy issues such as LGBTQ rights and immigration (Skitka and Tetlock, 1993; Jost, 2006). Initiating firms’ activism to promote the issue stand aligned with their own ideological preferences is thus unlikely to garner universal support and fail to emerge as a dominant standard. While it may be celebrated by people of similar ideological leaning, it may alienate those of different ideological leaning who prefer a different issue position. Broader scholarship on diffusion highlight such absence of a clear norm as a source of a wide variation in the way firms adopt and implement organizational practices, instead of homogeneity that has long been the focus of the traditional scholarship examining how unitary practices of corporate action spread throughout the field. To uncover this potential variation behind what may look like smooth and irrational mimicry on the surface, it is thus crucial to go beyond existing focus on an aggregate increase in the issue engagement by all other peers and understand the distribution of response among different types of firms in the population of potential followers (i.e., which firms follow). Existing scholarship points to the across-firm heterogeneity in the population of followers as a source of a variation that emerges in the way firms adopt a given practice in the diffusion process, showing how followers often customize the practice to fit their 110 own unique organizational characteristics and interests (Ansari, Fiss and Zajac, 2010; Fiss, Kennedy, and Davis, 2012; Litrico and David, 2017). In this paper, I extend this insight by bringing attention to the overlooked role of the heterogeneity in the followers’ ideological identity: that is, how followers may vary significantly in the ideological preferences of their own employees, who act as key stakeholders that shape the nature of these followers’ response. As noted, emerging literature on corporate political ideology document how a sorting process at the organizational level allows firms to develop a distinctive ideology from each other—some firms lean ‘red’ (i.e., conservatives) and others lean ‘blue’ (i.e,. liberals)—, with such variation often emerging even among firms in the same industry. Ideologically-charged nature of advocacy issues make these firms at both ends of the ideological spectrum viable and willing participants in the public discourse around those issues (Hambrick and Wowak, 2019; McDonnell and Darnell, 2020). Therefore, the population of potential followers may include not just firms whose employees hold similar ideological preferences to the original contributors (i.e., ideological allies), but also those whose employees hold different and even opposing preferences (i.e., ideological opponents). Importantly, this heterogeneity in the followers’ ideological disposition among the workforce may lead to the bifurcated response from these two groups in terms of how they would engage with the issue. On the one hand, response from original contributors’ ideological allies is likely to add support for their issue position. In the eyes of these allies’ employees, initiating ideological firms’ issue stand is closely aligned with their own ideological values and seen as an action to be emulated by their own firm. To win 111 favor from their employees who see the issue stand of the initiating firm as normatively desirable, ideological allies thus have an incentive to add further support for that position. On the other hand, ideological opponents of initiating firms have strategic incentives to behave differently. Given their contrasting ideological views from the original contributors, employees of ideological opponents may view initiating ideological firms’ issue stand as normatively undesirable and be inclined to reward the firm for taking a different or even an opposing stand. Indeed, perceptions of threats by opposing ideological interests is understood to be one of the main triggers for the rise of movement-countermovement dynamics (Useem, 1980; McCarthy and Wolfson, 1992; Meyer and Staggenborg, 1996). In this situation, ideological opponents have incentives to join in the debate around the given issue, but espouse a different issue position by funding alternative sets of advocacy nonprofits and think tanks that promote an alternative position aligned with their own ideological preferences, rather than that of original contributors (Desimone and Popoff, 2000; Bonardi and Keim, 2005). This posited dynamic is consistent with the extant understanding in the social movement literature that as the movement grows, an issue draws in diverse supporters that promote frames different from those employed by original contributors (Haines, 1988; Meyer and Staggenborg, 1996; Carlos et al., 2018). For example, Staggenborg (1991) notes that pro- life movement by conservative ideological interests has grown primarily in response to success of pro-choice movement led by groups with liberal ideologies. 112 3.4. Asymmetry in the Response between Ideological Allies vs. Opponents The following question is, which firms are likely to follow more than others? Depending on the relative magnitude of reaction from each group, the net social implications of initiating ideological firms’ actions may thus vary widely; in particular, when the latter response from ideological opponents dominates the response from ideological allies, not only may an initial action of ideological firms fail to create widespread issue support, it may even create damaging effects to the cause that they hope to champion by creating multiple and conflicting opinions around the issue. Indeed, proliferation of conflicting and fragmented views on advocacy issues is known to be the cause of the demise of many change initiatives (Levy, 1997; Bonardi and Keum, 2005), as it often prevents audiences’ opinions from coalescing around the final decision and promotes dissension, resulting in gridlock. Intuitively, initiating ideological firms would prefer to encourage a stronger reaction from their ideological allies who would support their issue positions, rather than from their ideological opponents with competing issue-related claims. Yet, my conjecture is that the strategic incentives for the subsequent action for ideological opponents will be stronger than the incentives for ideological allies’ action, so that, on average, the competitive response from ideological opponents would dominate the response from ideological allies. First and foremost, this claim rests on the logic that—as the classical theory of competitive differentiation posits (Porter, 1997; Baum and Haveman, 1997; Zhao et al., 113 2017)—the firm’s strategic gain and competitive advantage from its response would be maximized when it can differentiate itself from the original contributors: specifically, differentiation in terms of its issue position. While initial activism of the original contributors creates strategic incentives for both groups of their ideological allies and opponents to engage with the issue, one main strategic advantage that is unique to ideological opponents is that they can sufficiently distance themselves from the original contributors by taking a different stand (Mohliver, Crilly, and Kaul, 2019); this additional differentiation advantage is likely to be valued by their workforce and result in significant strategic gains such as stronger commitment and loyalty26. In contrast, ideological allies of initial contributors face relatively limited scope for strategic differentiation, as they are likely to stand close in their issue stance to original contributors, given their similarity in organizational members’ ideological preferences. Since they are not the first to take a given stand on the issue, these ideological allies’ actions are likely to be seen as less novel and unique—and even taken for granted—by their employees and thus less appreciated, which can limit strategic rewards that firms can expect in return for their action. The best that these ideological allies may hope to achieve by following original contributors’ actions is to do as well as them (Barnett and Woywode, 2004), or spend significantly greater resources to differentiate themselves from the initiating firms, all of which would limit the expected strategic returns from their issue investment. While this notion that for-profit corporations can use ideological contention around divisive social 26 Positioning decisions on sociopolitical issues involve different considerations from positioning on traditional product market attributes. Given the externalities, while stakeholders wouldn’t be offended by a firm’s decision to sell products in a color that they do not like, they can be offended by its decision to take a issue position that they do not agree with. 114 issues as the basis for their strategic differentiation and competitive advantage is relatively new to our management and CSR literature, it harkens back to the classic model of competition between opposing parties in political science, where parties with varying ideological positions differentiate themselves by espousing different positions on advocacy issues of divisive nature (Cox, 1990; Kahn et al., 1999; Carrillo and Castanheira, 2008; De Sio and Weber, 2014). Ideological opponents’ action can also produce more strategic benefits compared to the action of ideological allies, as employees may see their decision to take a different stand from others as a more authentic representation of firms’ values. According to CSR scholars, perceived authenticity of corporate social actions is one of the key determinants of if and how much firms can benefit from those actions (Hawn and Ioannou, 2016; Clark et al., 2018). Corporate social action seen as driven by external pressures or incentives (e.g., impression management) is often discredited as inauthentic and less valued by stakeholders, which produce limited strategic benefits or even backfire (Moulard, Raggio and Folse, 2016; Joo, Miller and Fink, 2019). Unlike non-profit entities such as social movement organizations, for-profit firms usually do not operate in the realm of civic discourse—especially that revolve around contentious issues—and thus lack in public trust and moral authenticity that legitimize their participation, which makes stakeholders especially wary of potential ulterior motives behind firms’ activism (Selznick 1949; Duffy et al. 2010, Walker and Rea, 2014; Yue 2015; McDonnell, 2016; Korschun et al., 115 2018)27. Importantly, an ideological ally’s support of initiating firms’ issue stand can be especially prone to this authenticity-related concern; employees may see the firm’s decision to echo what other firms have already said as a forced and symbolic attempt to deal with external pressure from their peers and increasing public attention on the issue (Kim and Lyon, 2014; Marquis, Toffel and Zhou, 2016), rather than a sincere expression of its intrinsic issue-related interests. In comparison, ideological opponents’ issue counter-positioning would be less prone to such concern, given that such action would be at odds with external cue from their industry peers; this may lead employees to favorably see the firm’s action as an internally driven action based on organizational values and thus form positive attitudes and commitment toward the organization. Indeed, Cuypers, Koh and Wang (2016) find that stakeholders value a novel and unique practices of corporate giving—which differentiates the firm from other corporate givers—more than conventional forms of giving, because the former is believed to require a greater internal deliberation and involvement on the part of firms and thus reflect their sincere and substantive commitment of firms to social causes. These reasons suggest that, all else being equal, initial actions of ideological firms can ironically create disproportionately larger economic incentives for their ideological opponents’ subsequent action, compared to their ideological allies’ action. That is, 27 As elite actors within the field, firms “lack one of the few (but central) advantages that insurgents traditionally enjoy: the appearance of speaking for the authentic interests of ‘the people’ in a fashion not motivated primarily by gaining economic resources or political power” (Walker 2015, p. 3). 116 ideological rivals may simply have more to gain from responding to initiating ideological firms’ actions than ideological allies. I therefore predict: H1: A positive relationship between a focal firm’s giving and its ideological opponents’ giving is stronger than the relationship between a focal firm’s giving and its ideological allies’ giving. 3.5. Methodology 3.5.1. Research Context and Data I use the same database used in the second essay of this dissertation for the records of corporate advocacy giving by US public companies28. In addition to the data on corporate advocacy giving, this paper’s inquiry also requires the dataset on firms’ ideological characteristics. I continue to follow the approach of identifying organizational ideologies of firms based on campaign contributions of individuals within a firm (Gupta et al., 2017; Bermiss and McDonald, 2018; Carnahan and Greenwood, 2018; Gupta and Briscoe, 2019; McDonnell and Cobb, 2019). The data on individuals’ campaign contributions are obtained from Bonica’s 2013 DIME database29. Building on the insight of recent papers (Gupta et al., 2017; Gupta and Briscoe, 2019) that firms’ CSR practices are often a reflection of the ideological characteristics exhibited by the entire body politic, not just only those of corporate elites, I collect political contributions of all individuals that report their main employer as a given firm in my sample to calculate the 28 In Appendix Table 3.3, I run the main results in this paper using the records of corporate giving that goes to service giving, and find that the same pattern of results is not replicated, such that there exists no pattern of asymmetry in the effect of giving by ideological opponents vs. giving by ideological allies. 29 Data.stanford.edu/dime. 117 organization-level ideological characteristics30. Finally, a variety of other sources I draw on include Compustat, KLD Research & Analytics, lobbyview.org (Kim, 2018)31, and Labor Force statistics from the Current Population Survey, which provide additional firm-level financial and operational information. My final sample frame is 433 U.S. public firms that are identified with at least one record of advocacy giving and whose employees made campaign contributions at least once during my study period32, and their 25,440 firm-issue-year observations. In this dataset, for each corporation-advocacy issue pair in each year, I can observe whether or not the corporation gave to the issue in that year, and if so, how much. 3.5.2. Variables Dependent Variable The dependent variable of this study is the Choice of Advocacy Giving, which is measured as a dummy variable indicating whether a focal firm made donation to a given issue area in a given year, This measure proxies for a firm’s decision to engage in a given “market” of advocacy issue in each year. This builds on the understanding each advocacy issue area (e.g., environment, health, education, civil rights, etc.) can be understood as a market different from others that involve its own stakeholder base (e.g., customers, 30 In the first mechanism analysis (reported in Table 3.4), I separate out the effect of organizational ideological intensity into that of ideological intensity by senior executives vs. ideological intensity by lower-level employees. 31 This is the firm-level lobbying database (Kim, 2018), which is originally generated from lobbying reports filed between 1999 and 2017 and recorded at the Senate Office of Public Records. 32 Most of the US public firms with affiliated foundations as reported in the FDO dataset do have employees who spent positive amount of campaign contributions at least once during my study period. Yet, there do exist firms in total whose employees cannot be matched to any records of campaign contributions, which are excluded from my analyses. 118 employees, and activists that care about the issue) and its own set of participants (e.g., advocacy nonprofits and grantmakers, including corporations). Accordingly, the market condition and opportunities of social action vary significantly across different issue areas over time (Meyer, 2004). The definition of cause areas follows, again, NTEE-CC classification system; thus, according to the construction of my data, in every year, each firm is assumed to make decisions on whether or not they would contribute to each one of 22 issue areas, which would be reflected in the variation observed in this dependent variable. In the robustness check, I confirm the stability of my findings in the design using the logged amount of a firm’s advocacy giving (i.e., how much a focal firm gave to a given issue) as an alternative dependent variable. Independent Variables The main baseline regressor in this paper is Giving by Industry Peers. It is measured as the dummy variable indicating whether or not a focal firm’s industry peers made donation to a given issue in a given year, consistent with how advocacy giving of a focal firm (dependent variable) is measured. Building on the aforementioned studies that highlight industry peers as one of the most crucial reference sets for understanding the pattern of firms’ philanthropic contributions (Useem and Kutner, 1986; Bertels and Peloza, 2008; Marquis and Tilcsik, 2016)33, I define firms’ peers as those with common industry membership (NAICS 2 digit). 33 According to Marquis and Tilcsik (2016), peer group can be further specified not just by their industry, but also by their geographic location, based on which firms can be categorized into four groups: 1) same industry x same state, 2) same industry x different state, 3) different industry x same state, 4) different industry x different state. While their paper posits and shows that the first group of so-called “institutional 119 To test the main theoretical conjecture in this paper, I break down the variable of Giving by Industry Peers into two components based on the ideological position of givers: Giving by Ideological Allies and Giving by Ideological Opponents. To this end, I measure an ideological position of each firm, by calculating the average ideological score of the firm’s individual employees that made campaign contributions during my study period, with each individual’s ideological scores weighted by the frequency of their contributions. Ideological scores of each firms’ individuals used to calculate this measure are derived from the common-space CFScore in DIME dataset (Bonica, 2013). Using the intuition that individuals with similar ideologies donate to similar political candidates, Bonica’s CFScore places individuals’ ideology on a scale of –2 (highly liberal) to +2 (highly conservative) based on their political contribution patterns. Corporate ideological position, which is calculated by aggregating these individuals’ positions at the firm-level, varies significantly across the entire ideological spectrum, which is shown in Figure 3.1. Firms whose ideological position takes the positive values are categorized as conservative firms (i.e., red firms), and firms with negative values are categorized as liberal firms (i.e., blue firms). Based on this categorization, Giving by Ideological Allies is defined as other blue industry peers’ previous advocacy giving for a focal blue firm, equivalents” has the strongest mimetic effect on a focal firm’s pattern of giving, I find no such pattern in my data of advocacy giving. If anything, only the second group of firms that are in the same industry but are located in different states exerts a significant effect on a focal firm’s giving. There can be two reasons behind this finding. Theoretically, unlike service giving, not a lot of advocacy giving corresponds to local giving (i.e., giving to nonprofits in firms’ headquartered states). This can be one reason why giving by firms in the same area do not have strong normative or informational values on a focal firm’s giving. Empirically, because advocacy giving is rarer than service giving, I can expect (and indeed find that) not many firms have institutional equivalents. That is, not many firms have peers that are advocacy givers and are operating in the same industry and same state. In fact, this observation is consistent with their finding that, for firms without any institutional equivalents, firms use giving by industry peers or giving by community peers as alternative reference points. 120 and other red industry peers’ previous giving for a focal red firm; Giving by Ideological Opponents is defined as red industry peers’ previous giving for a focal blue firm vs. blue industry peers’ previous giving for a focal red firm. Control Variables First, I control for the dummy variable of Giving by Other Firms, which indicates whether or not firms outside the boundary of a focal firm’s industry gave to a given issue in a given year. This serves as an empirical counterpart compared to which we can evaluate the effect of industry peers’ advocacy giving decisions on a focal firm’s giving. Second, I control for firm-level financial and operational attributes. They include a firm’s R&D-to-Sales Ratio, Advertising-to-Sales Ratio, and Capex-to-Sales Ratio, measured by its expenditure in each area scaled by total sales, all of which have been shown to affect the firm’s philanthropic activities in previous studies (Surroca, Tribó and Waddock, 2010; Zhao and Murrell, 2016). I also control for a firm’s Lobbying-to-Sales Ratio, to account for the extent to which its advocacy giving is politically motivated and used as a complementary tactic for lobbying (Bertrand et al., 2020). Additional controls include a firm’s Size, which is a logged measure of its total assets, and a firm’s financial performance, measured by return on assets (ROA). I include the index of CSR rating to control for the potential role of firms’ other areas of social engagement on their advocacy giving decisions, which is measured as the firm’s net KLD score, i.e., its number of strengths minus number of concerns in four major dimensions (environment, diversity, employment, community) (Mattingly and Berman, 2006). I also control for Female 121 Board Ratio, which is calculated by the number of a firm’s female board members scaled by its total board size, given the prior evidence that documents its link to firms’ social actions (Marquis and Lee, 2013). The third set of controls is to account for the potential influence of external market conditions on firms’ advocacy giving decisions. Specifically, I account for the product-market competitive intensity in the firm’s main market by including the ratio of Market Concentration, measured by the Herfindahl index of market share distribution across firms in the same NAICS 2 digit code as the focal firm. In light of previous studies that highlight the importance of a firm’s geographic community in its CSR and corporate philanthropic practices (Marquis et al., 2007), I also include the size of Headquarter Population (in millions) and the Headquarter Poverty Ratio in the firm’s headquartered state. Table 3.1 reports summary statistics and correlations of variables. Unsurprisingly, the measures of Giving by Ideological Allies and Giving by Ideological Opponents show a relatively high level of correlation with each other (31%), which raises concerns about potential multicollinearity. While my main analyses use raw measures of each variables for ease of interpretation of the findings, I find that the findings are consistent with the use of orthogonalized measures of two variables and also in the specifications that separately estimate the effect of each variables. ***Insert Table 3.1 about here*** 122 3.5.3. Specification My main specification estimates the following regression that examines the relationship between a focal firm’s giving and its ideological opponents’ and allies’ giving in the previous year. 𝐺𝑖𝑣𝑖𝑛𝑔 𝑏𝑦 𝑎 𝑓𝑜𝑐𝑎𝑙 𝑓𝑖𝑟𝑚𝑖𝑗𝑡+1 = 𝑓(𝐺𝑖𝑣𝑖𝑛𝑔 𝑏𝑦 𝑖𝑑𝑒𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝑜𝑝𝑝𝑜𝑛𝑒𝑛𝑡𝑠𝑖𝑗𝑡 , 𝐺𝑖𝑣𝑖𝑛𝑔 𝑏𝑦 𝑖𝑑𝑒𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝑎𝑙𝑙𝑖𝑒𝑠𝑖𝑗𝑡 , 𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑖𝑡, 𝐹𝐸𝑖𝑗 , 𝐹𝐸𝑡) where 𝐺𝑖𝑣𝑖𝑛𝑔 𝑏𝑦 𝑎 𝑓𝑜𝑐𝑎𝑙 𝑓𝑖𝑟𝑚𝑖𝑗𝑡+1 is the dummy variable of advocacy giving by firm 𝑖 in issue 𝑗 at year 𝑡 + 1, 𝐺𝑖𝑣𝑖𝑛𝑔 𝑏𝑦 𝑖𝑑𝑒𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝑜𝑝𝑝𝑜𝑛𝑒𝑛𝑡𝑠𝑖𝑗𝑡 is the dummy variable of advocacy giving made by firm 𝑖’s ideological opponents in issue 𝑗 at year 𝑡, and 𝐺𝑖𝑣𝑖𝑛𝑔 𝑏𝑦 𝑖𝑑𝑒𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝑎𝑙𝑙𝑖𝑒𝑠𝑖𝑗𝑡 is the dummy variable of advocacy giving made by firm 𝑖’s ideological allies in issue 𝑗 at year 𝑡. In addition, I include a vector of time-varying firm and industry level characteristics 𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑖𝑡 at year 𝑡. I further include year fixed effects 𝐹𝐸𝑡 in all specifications, and also control for issue and firm fixed effects. My preferred specification, as shown in the equation above, includes firm-issue pair fixed effects 𝐹𝐸𝑖𝑗. One of the most salient concerns in any empirical analysis that link different firms’ behaviors to identify interdependencies is the possibility of omitted third variables, which may cause behaviors of firm A and firm B in the same industry to vary in the same direction. For one, it is possible that the unobserved industry-issue specific factors may create strategic incentives for firms in the same industry to co-invest to a given issue than others (e.g., oil firms giving to environmental issue). Or, within a given industry, there may exist a sub-group of firms with shared organizational capabilities and 123 resources, which make them pay more attention to a set of issues than others. Firm-issue level fixed effects mitigate these concerns by controlling for all time-invariant firm-issue specific heterogeneity, with industry-issue specific influences also captured in these effects. I use a linear probability model in my main models reported below, though results are robust to a variety of other specifications including fixed effects logit or standard OLS regressions using the logged amount of giving by a focal firm and the logged amount of giving by its ideological opponents and allies as alternative dependent and independent variables. All regressions use robust standard errors clustered at the firm- issue pair level. 3.6. Results 3.6.1. Main Results Before presenting the main findings, it is useful to confirm that the pattern of institutional isomorphism found in previous studies does arise in this context of advocacy giving, which would serve as a baseline for my analysis: that is, if firms are more likely to give to the issue areas that their industry peers gave to. Model 1 to 3 in Table 3.2 reports this baseline result, where I examine the effect of Giving by Industry Peers of a focal firm on its subsequent choice of advocacy giving in a given issue area, with a full set of controls. M1 employs year and issue fixed effects, M2 employs year, issue, and firm fixed effects, and finally, M3 employs year and firm-issue fixed effects (the default and most preferred structure as described above). Across all three specifications with 124 different FE structure, I observe a highly significant and positive effect of Giving by Industry Peers. Such consistent and strong positive effect that peers’ giving has on a focal firm’s giving at the issue-level is in contrast with the insignificant effect of Giving by Other Firms, which is an important control against which we can gauge how the pattern of industry peers’ previous giving, but not those of firms outside the boundary of the industry, triggers a strong mimetic response by a focal firm. These results thus strongly suggest that firms do closely follow their industry peers’ issue area of investment in their advocacy giving, replicating the findings of previous papers that demonstrate largely the same patterns of conformity and learning between industry peers (e.g., Marquis and Tilcsik, 2016). Yet, as noted, this result does not account for the presence of heterogeneous types of actors within a given institutional environment. In particular, my theory focuses on the unexplored possibility that ideological firms’ giving might invite response not just from their ideological allies, but also from their ideological opponents, and how the response from the latter may be stronger than the response from the former. To test these possibilities, I divide the variable of Giving by Industry Peers into Giving by Ideological Opponents and Giving by Ideological Allies, and re-run the regression. The result of this analysis is presented in the Model 4 to 6 in Table 3.2, with each specification controlling for increasingly more demanding sets of fixed effects. According to the baseline result in M4 that only control for issue and year fixed effects but not for firm fixed effects (thus, between-firm estimation), both variables—not just Giving by Ideological Allies but also Giving by Ideological Opponents—exert a strong and positive effect on the focal firm’s 125 giving, with a Chi-square test reporting no statistically significant difference between the two coefficients. Yet, this baseline result does not account for the confounding effects of the unobserved firm heterogeneity that may act as omitted third variables, which may cause firm A and firm B to co-invest to a given issue due to their similar organizational characteristics. To account for this effect, I add firm fixed effects in M5, or firm-issue pair fixed effects in M6 (most preferred specification), which mitigates these concerns above by controlling for all time-invariant firm-issue specific heterogeneity. The results in M5 show that the observed effect of Giving by Ideological Opponents (𝛽=0.0377, p=0.0000) stay to be highly significant and positive, whereas the effect of Giving by Ideological Allies (𝛽=0.0110, p=0.1906) completely loses its significance in this model. A Chi-square test strongly confirms the significant difference between the two coefficients (p=0.0169) as well. The following M6 that control for more demanding firm- issue pair level fixed effects reports largely the same pattern of results; the coefficient of Giving by Ideological Opponents (𝛽=0.0343, p=0.0001) is shown to be significantly bigger and stronger than that of Giving by Ideological Allies (𝛽=0.0139, p=0.1355), which stays insignificant at conventional levels. These findings lend strong initial support for Hypothesis 1, suggesting that firms react more strongly to the giving by their ideological opponents in the same industry, compared to the giving by their ideological allies. According to the estimates in M6, when a focal firm’s ideological opponents give to a given issue, the probability of its own giving to the same issue area increases by 126 3.4% in the subsequent year, which corresponds to a 13.6% increase over its baseline giving probability of 25.9%. This effect size is substantially bigger than the corresponding increase in a focal firm’s subsequent probability of giving associated with its ideological allies’ giving, which is 1.3% (i.e., a 5.5% increase over its mean value). In other words, the results confirm the posited asymmetry in the pattern of response to ideological opponents’ issue-related actions vs. allies’ actions, showing that ideological firms’ giving invites a disproportionately stronger reaction from ideological rivals whose issue position is likely to be opposed to their own. ***Insert Table 3.2 about here*** 3.6.2. Robustness Analyses So far, the main findings provide support for the posited spillover effects of ideological firms’ giving on other ideological firms in the same industry: in particular, on their ideological opponents. Yet, there may still remain a concern about the confounding effects of the omitted third variables. Even though the main specification in my analyses use a demanding set of firm-issue and year fixed effects to mitigate this concern, one may still be concerned about the potential common shock of time-varying nature, which may drive adjacent giving behaviors of firms in the same direction. These common shocks could be exogenous events that raise the saliency of an issue over others (e.g., recent Supreme Court ruling on LGBTQ discrimination in the workplace), which may lead different firms to simultaneously increase their spending on the issue. 127 To remove confounding effects attributable to these common shocks, I add issue- year fixed effects to the current fixed effect structure and re-run the analysis. The design in this set of robustness analyses thus correspond to my most stringent specification with firm-issue and issue-year level fixed effects, which absorb all time-invariant firm-issue specific heterogeneity and all issue-specific time trends. As shown in M7 of Table 3.3, Giving by Ideological Allies stays to be insignificant in this model, with its effect size getting sufficiently smaller than that observed in previous models; this finding suggests that the positive correlation between adjacent advocacy giving decisions among ideological allies observed in the previous models may in large part reflect their response to common issue trends in the environment. In contrast, the effect of Giving by Ideological Opponents stays highly significant and positive in this demanding specification (𝛽=0.0310, p=0.0006), with a Chi-square test conforming the statistical difference between two coefficients (p=0.0349). Of course, there may remain concerns about the possibility of certain geography- specific time trends around the issue areas, which the general issue-level time trends may fail to capture. For example, the effect of certain external shocks that may change the salience of given issues over others may be geographically bounded (e.g., mass shooting, oil spill, demographic profile change). To address this concern, the following M8 replaces main variables of giving by ideological opponents and allies with giving made by each group of firms that are headquartered in the different states to a focal firm. The intuition for these analyses is that, unlike giving by ideological firms that are located in the same geographic area to the focal firm, a variation in giving by industry peers located 128 in different areas would not be subject to the possibility of common geography-specific shock that may confound the results. The results using these alternative measures produce consistent findings, with a Chi-square test still conforming the statistical difference between two coefficients (p=0.0226). Another main issue in my analysis is the potential reverse causality—the direction of the influence from ideological firms to others in the industry can flow in the opposite way—as well as potential autocorrelation concerns. While my main specification uses a lagged structure to alleviate this concern, the possibility of potential autocorrelation in my baseline model remains. In M9, I re-estimate my results using a system dynamic panel model (Arellano and Bover, 1995; Blundell and Bond, 1998). I find the consistent results; the coefficient of Giving by Ideological Allies is shown to be insignificant and even negative, whereas Giving by Ideological Opponents is highly significant and positive (𝛽=0.0515, p=0.0000), with the statistical difference between the two coefficients strongly supported (p=0.0003). Controlling for the recent past of a focal firm’s giving, the effect size of giving by ideological opponents grows substantially bigger than that shown in the previous models; based on the estimates of M9, when a focal firm’s ideological opponents give to an issue, its own probability of entry into the same issue increases by 20.4% over its mean value, whereas the corresponding change in ideological allies’ giving is associated with -1.7% decrease in a focal firm’s subsequent probability of giving. In M10 and M11, I re-run my default model (M6 in Table 3.2 with firm-issue and year fixed effects) by using the logged amount of giving as an alternative independent 129 variable, dependent variable, or both. Specifically, in M10, I replace my main regressors of Giving by Ideological Opponents and Giving by Ideological Allies from dummy variables to the logged amount of giving (i.e., how much each group of ideological opponents and allies gave to a given issue area). In the following M11, I further replace my dependent variable of a focal firm’s Choice of Giving with the logged amount of its giving (i.e., how much a focal firm gave to a given issue) and re-run the OLS FE regression. While the effect of Giving by Ideological Allies gains its marginal significance in these models using the amount of giving as alternative regressors, both models continue to report a highly consistent pattern of asymmetry between the two regressors, which report that the effect of Giving by Ideological Opponents is significantly bigger and stronger than that of Giving by Ideological Allies. In the untabulated results, I check the robustness of the results in the design using the orthogonalized measures of Giving by Ideological Opponents and Giving by Ideological Allies, to address the concerns about the potential instability of the results that may arise from the high correlation between the two variables. The pattern of results also stays largely the same with the use of fixed effect logit model as an alternative specification. ***Insert Table 3.3 about here*** 3.6.3. Mechanism analyses (1): Role of Employees These overall results lend strong support for my Hypothesis 1, showing that ideological firms’ giving invites a strong reaction from their ideological opponents, and 130 such response from opponents is stronger than that shown from their allies. Having demonstrated support for my main conjectures, I now turn to consider the effect of several contingent factors (summarized in Figure 3.2); these analyses would not only help to better understand the mechanisms underlying my observed main findings, but also to add credibility to the results. Given the consistency between my main result and the estimation in the robustness check, I use firm-issue, and year level fixed effects for the analyses below, though the results adding issue-year level time trends are largely the same. ***Insert Figure 3.2 about here*** Mechanism analyses (1.1): Executive ideology vs. Non-executive employee ideology According to my theory, selective response to ideological opponents’ giving is internally driven by the preferences and interests of their organizational members around given advocacy issues. After all, my empirical design based on which I identify such response categorizes firms into ideological allies vs. ideological opponents according to their organizational members’ ideological preferences, demonstrating the role of internal employees as key stakeholders in firms’ advocacy giving decisions. Yet, even within this scenario, there can be two explanations of different nature as to why firms do this. The first scenario, which is most consistent with my theory, is that firms respond to their ideological opponents because their organizational members care about a given issue, and this response is strategically beneficial as their members who appreciate such counter- response would reward the firm by showing increased commitment and loyalty (Burbano, 131 2016; Flammer and Luo, 2017). Alternatively, this selective response to ideological opponents might simply reflect agency behavior of top executives (Friedman, 1970; Ang, Cole and Lin, 2000; Tan and Tang, 2016); according to this second scenario, CEOs and other top executive members may be using advocacy giving to pursue their private ideological preferences separate from other employees’ preferences, which may not necessarily be aligned with the firm’s strategic interests. To evaluate the validity of the first scenario compared to the second, I re-run the estimation by splitting the sample into four cases based on the separate measures of ideological intensity among the firm’s top executives vs. among its non-executives in Table 3.4: 1) high executive ideological intensity & high non-executive ideological intensity (M12a); 2) high executive intensity & low non-executive intensity (M12b); 3) low executive intensity & high non-executive intensity (M12c); 4) low executive intensity & low non-executive intensity (M12d). Ideological intensity is measured by how much each group of individuals spent on campaign contributions during my study period, which would capture which group is relatively more ideological than the other (Hambrick and Wowak, 2019). According to my findings, only the first group of firms (M12a)—where both top executives and lower-level employees are highly ideologically motivated—selectively respond to the previous giving of their ideological opponents; in contrast, such selective response does not arise in any other cases, including the sample where only the senior executives, but not the lower-level employees, exhibit strong ideological preferences (M12b). These findings are consistent with the aforementioned first scenario, suggesting that the observed response to ideological opponents is being 132 driven by ideological preferences of the entire body politic of the firm—not just senior executives, but also other organizational members of the firm (Gupta et al., 2017; Gupta et al., 2018)—, and is inconsistent with the second interpretation of such response as the agentic behavior based on top executives’ personal preferences (Gupta et al., 2017; Gupta et al., 2018)34. Mechanism analyses (1.2): Ideologically homogeneous vs. heterogeneous firms The second ideological characteristic of firms that may further shed light on the validity of the first scenario is the level of ideological heterogeneity within the firm: i.e., whether the firm’s organizational members hold similar ideological positions to each other or not. According to my theory, ideological firms give more when their opponents gave, because of their organizational members’ ideological interests to defend their preferred issue positions. Importantly, such role of internal preferences of employees would be pronounced in the group of ideologically homogeneous firms whose employees think similarly to each other, which gives rise to a prevailing system of beliefs that would enable firms to reach internal consensus in their advocacy giving decisions as to which issues to support and in what ways (Schneider, 1987; Gupta et al., 2017). Moreover, such counter-response to ideological rivals’ giving is likely to translate into a firm’s strategic advantages only when the firm faces a sufficiently low level of within-firm ideological diversity, whereas counter-response by a firm with a high level of ideological 34 In the untabulated supplementary analysis, I run the analysis that tests how the pattern of response varies by the level of firm governance, which is proxied by the number of independent board members scaled by total board size. Inconsistent with the possibility that the asymmetric response to rivals’ giving may reflect agency costs, I find that such response is actually stronger among firms with strong governance, rather than those with weak governance. 133 heterogeneity may backfire by alienating a significant portion of their employee base who disagrees with the stated position and may show reduced engagement and motivation (Burbano, 2020). To explore this conjecture, I run a split sample analysis based on the median value of within-firm ideological heterogeneity, which is measured by the standard deviation of ideological positions among firm employees that made campaign contributions during my study period. As shown in Table 3.4, the results presented in M13a and M13b are consistent with my expectation, showing that it is the group of ideologically homogeneous firms (with a low level of within-firm ideological diversity) that respond selectively to their ideological opponents’ giving, but not the group of ideologically heterogenous firms. Mechanism analyses (1.3): Labor market competition Finally, I look at how the pattern of response to ideological rivals varies by the level of labor market competition. If it is the case that firms’ counter-reaction to their opponents is their strategic decision to make their ideologically motivated employees happy, it should be concentrated among firms that face ‘talent wars’, having to compete intensely to hire and retain personnel (Coff, 1997; Gardner, 2005; Somaya and Williamson, 2008; Tan and Rider, 2017). I test this conjecture by running a split sample analysis based on the median value of unemployment rate in a focal firm’s industry (NAICS 2 digit), which is used as the inverse proxy for the level of labor market competition that the firm faces. The results in Table 3.4 show that selective response to ideological opponents arises among the firms that face higher labor market competition (M14a), but not among firms that face lower labor market competition (M14b), 134 confirming the role of internal employees as the key stakeholder that drives firms’ counter-response to their ideological rivals. In the untabulated analysis using the proxy of labor intensity to capture firm sensitivity to employees’ preferences (measured by the average industry wage at NAICS 2 digit), I also observe similar results, finding that asymmetric response to ideological opponents is stronger among firms with higher labor intensity. In comparison, the supplementary analyses (the details shown in Appendix Table 3.2) show that the role of external stakeholders (e.g., consumers) is rather limited in the firms’ decisions for counter-response to their ideological opponents. In the analysis where I examine the moderating role of product-market competition (using the inverse measure of market concentration as a proxy), I find that stronger response to opponents’ giving arises only among firms that face a low level of competition, not a high level of competition. I also find that the exclusive response to ideological opponents is coming from the group of firms that exhibit lower consumer sensitivity, which is proxied by the level of advertising spending ratio. Both results are inconsistent with the pattern of response that we would expect if the observed selective response to ideological rivals is driven by the preferences of external stakeholders such as consumers (Wang and Qian, 2011; Flammer, 2015). Rather, stronger response to ideological rivals seems to be more pronounced among the group of firms in a market condition where the voice of internal employees can be maximized. ***Insert Table 3.4 about here*** 135 3.6.4. Mechanism analyses (2): Ideological competition According to my theory, this observed selective response to ideological opponents reflects firms’ competitive attempts to appeal to their own organizational members who are alienated by the issue-related actions of initiating ideological firms with opposing ideologies and want to defend their preferred issue position against such actions. Mechanism analyses (2.1): Swing vs. non-swing states If this is the case, we can expect that the pattern of response would vary by the level of ideological competition in the environment. Specifically, I expect that the selective response to rivals would get stronger with the intensifying ideological competition in the firms’ stakeholder environment, with such pattern being concentrated in so-called swing or battleground states. In swing states characterized by high ideological heterogeneity among stakeholders, there is no solid status quo, with stakeholders disagreeing over how to approach a given issue. Such uncertainty in the stakeholder environment make swing states “sweet spots” for institutional change (Gamson and Meyer, 1996), where initiating ideological firms’ actions have the potential to move the needle and thus present threats to the firms with opposing ideologies. In contrast, in solidly red and blue states with low stakeholder heterogeneity, political opportunity structures for institutional change are relatively closed (Schneiberg and Lounsbury, 2008; York and Lenox, 2014; McDonnell et al., 2015), where ideological firms’ issue engagement is less likely to be perceived as salient threats for their ideological opponents. Accordingly, ideological firms’ efforts to undermine any issue- 136 related change made by their ideological rivals would be strategically targeted to these battleground states (Gamson and Meyer, 1996; Briscoe et al., 2014). To test this prediction, in Table 3.5, I re-estimate my main analysis by using more granular measures of dependent variable: a focal firm’s advocacy giving in swing states (M15a) vs. advocacy giving in non-swing states (M15b). I measure the House election margin between Democratic votes and Republican votes in each election cycle during my study period, and categorize states with narrower-than-median margin as swing states, and the rest as non-swing states. Each grant of a focal firm is assigned to swing or non- swing states based on its nonprofit recipients’ location information. Consistent with my expectation, the results show that selective response to ideological firms’ opponents arise only in swing states, but not in non-swing states. This pattern is consistent with the empirical regularity that competition between political parties, reflected in the distribution of campaign contributions and political ad spending, is concentrated in a few swing states. Mechanism analyses (2.2): Ideological radicals vs. Ideological moderates In addition to the level of ideological competition in the external environment, corporations’ own ideological attributes may also act as a key contingency that moderates their competitive response to ideological opponents. Specifically, I expect that radicalness of a focal firm’s ideological position would moderate its counter-response to ideological opponents’ giving. Within the group of firms that hold ideologically aligned positions to each other, there is a range in the level of radicalness in their ideological 137 positions, such that some firms are placed on the extreme end of the ideological spectrum (i.e., ideological radicals), whereas others are placed more closely towards the middle (i.e., ideological moderates). I expect this heterogeneity to shape the pattern of ideological firms’ response to their ideological opponents’ giving, such that selective response to ideological opponents would be stronger for the former group of ideological radicals, if it truly reflects competitive attempts to defend the issue position held by a focal firm’s organizational members. The more extreme a focal firm’s ideological position is, the distance between its organizational members’ preferred issue position and that held by its ideological opponents is likely to be greater. Accordingly, ideological radicals are more likely to perceive their ideological opponents’ issue-related actions as salient threats, which leads to a greater propensity for counter-response to defend their issue position (May, 1973; Fisher, 1999). For those with extreme ideologies, such struggle with alternative perspectives may even be seen as their reason for being (Selznick, 1960; Barnett and Woywode, 2004). To test this prediction, I split the sample into the group of ideological radicals (M16a) and ideological moderates (M16b), based on the median value of ideological position in each ideological camp of conservatives and radicals. Consistent with my expectation, the results show that the selective response to ideological opponents’ giving is primarily coming from the group of firms with radical ideological positions, but not those with moderate positions. This finding lends support for my theory, suggesting that the observed stronger response to ideological rivals may indeed reflect the dynamics of 138 within-issue competition between firms at polar ends of the ideological spectrum, which is likely to contribute to increasing polarization and conflict around the issue. Mechanism analyses (2.3): Ideological liberals vs. Ideological conservatives In addition, the tendency of stronger response to ideological opponents may vary also by the focal firm’s ideological position: in particular, whether a focal firm holds a liberal ideology (i.e., blue firm) vs. conservative ideology (i.e., red firm). While my main argument has treated the counter-mobilization by blue firms and red firms as symmetrical possibilities, the extensive literature on political psychology (Jost, 2006; Jost et al., 2007) finds that liberals and conservatives tend to differ fundamentally in their preference of stability and preserving status quo, suggesting the potential asymmetry in the role played by two groups (Bermiss and McDonald, 2018). Specifically, scholars have found that, while liberals are relatively open to the prospect of change, conservatives put more emphasis on stability, exhibiting greater psychological needs to reduce the perception of uncertainty and threats (Skitka and Tetlock, 1993; Jost et al., 2003; Christensen et al., 2015). This inherent psychological difference between individuals at two ideological poles suggests that, if the firm’s response to ideological rivals reflects the attempts of counter-positioning to defend the issue position preferred by their organizational members, such response should be stronger among the group of red firms with conservative workforce than blue firms with liberal workforce. Conservative organizational members in red firms would perceive any issue-related changes— especially the changes made by their ideological opponents that would be undesirable in their eyes—as bigger threats than liberal members in blue firms, and thus are more likely 139 to mobilize to undermine such changes. Indeed, this is consistent with the extant notion that “rightist movements tend to be known for what they are against, not for what they support” (Blee and Creasap, 2010), as shown in the cases of anti-immigrant movements fueled by success of pro-immigrant movements, anti-LGBTQ movements mobilized by change in public perceptions around LGBTQ rights, etc. I test this conjecture by splitting the sample into ideologically liberal firms (M17a) and conservative firms (M17b) based on the average ideological position of their organizational members. The results support the posited asymmetry between the two groups, showing that stronger response to ideological opponents is coming from the group of ideological conservatives, whereas the response by ideological liberals does not vary by the ideological identity of original contributors. According to the estimates, ideologically liberal firms’ giving is associated with a 27.4% increase in the subsequent giving probability by ideologically conservative firms, whereas the corresponding change in giving by their ideological allies (i.e., conservatives) is only associated with a 3.8% increase. ***Insert Table 3.5 about here*** 3.6.5. Mechanism Analyses (3): Recipient choices within each issue area Lastly, I run the supplementary analyses to examine whether firms that belong to two different ideological camps tend to support different issue positions within each issue, which is an implicit assumption in my interpretation of the observed response to ideological opponents as the competitive attempts of counter-positioning. While the ideal test is directly observing the issue positions held by advocacy nonprofits that each firm 140 supports, an indirect approach to test this dynamic using my dataset is to examine if the pattern of ideological clustering arises in firms’ recipient choices within each issue area. Specifically, within each issue, I can observe if ideologically adjacent firms tend to fund the similar set of nonprofits, such that ideological conservative firms fund the nonprofits that other conservative firms gave to, and ideological liberal firms fund the nonprofits that other liberal firms gave to. Empirically, this means that, when it comes to the choices of nonprofit recipients within each issue, we would observe the opposite pattern of asymmetry in the effects of giving by ideological allies vs. ideological opponents on a focal firm’s giving, with the effect of the former having a far stronger and positive effect on a focal firm’s own choice of nonprofit recipient than the latter. To examine this, I construct my dataset at a more granular, nonprofit recipient-firm dyad level to observe firms’ recipient choices within each issue area. The dependent variable in this analysis is whether or not a focal firm gives to the nonprofit in each year, and the main regressors are whether its ideological allies and opponents gave to the given nonprofit in the previous year. Consistent with my main analyses, I run a linear probability model, with year, firm and recipient fixed effects (issue-specific influences captured in recipient FE). The results shown in Table 3.6 are consistent with my expectation. As shown in M18, while both variables are significant and positive, the effect of giving by ideological allies (𝛽=0.0801, p=0.0000) is not only stronger but also significantly bigger in its effect size than the effect of giving by ideological opponents (𝛽=0.0258, p=0.0158), with the statistical difference between the two coefficients being strongly supported (p=0.0000). This result suggests that, within each issue, firms are more likely to give to the nonprofits 141 that their ideological allies gave to, compared to the nonprofits that their ideological rivals gave to, which demonstrates the posited dynamics of ideological clustering in firms’ recipient choices. The same pattern of results holds in M19, where I use the logged amount of giving by a focal firm as an alternative dependent variable and the logged amount of giving by its ideological allies and opponents as alternative regressors; in this model, giving by ideological opponents is shown to have only a marginally significant and positive effect, whereas the effect of giving by ideological opponents is still highly significant at p-value of 0.0000. In the following M20 and M21, I test the robustness of this finding by estimating the main effect of the variable of Ideological Gap on a focal firm’s giving, which I measure by calculating the absolute difference between the average ideological position of corporate donors that gave to the nonprofit in a previous year and a focal firm’s ideological position35. My expectation is that this variable would have a negative effect on a focal firm’s giving to the given nonprofit, which would suggest that a firm prefers to give to the nonprofits that other ideologically adjacent firms gave to. Consistent with this expectation, I find that the coefficient of Ideological Gap is highly significant and negative in both models that use the alternative construction of independent and dependent variables as dummy of giving (M20) and the logged amount of giving (M21), respectively. In other words, the higher the ideological gap between a focal firm and the 35 Note that the construction of this analysis excludes firm-recipient-year observations in which no other firms made a giving to the given recipient in the previous year, significantly reducing N of this model. I also find the same result when I use the weighted average ideological score of previous corporate donors to the nonprofit (by amount) to calculate the variable of Ideological Gap, and also when I use the two year window as an alternative time frame to calculate the average ideology of previous donors. 142 nonprofit’s previous corporate funders, the less likely the firm is likely to fund the given nonprofit. Overall, these collective results strongly confirm that, when giving to an issue, a firm tend to fund nonprofits whose previous funders are ideologically similar to itself. ***Insert Table 3.6 about here*** 3.7. Discussion The results of my empirical analyses strongly support my theoretical arguments about how ideological firms’ advocacy giving opens the door for not just their ideological allies, but also for their ideological rivals, which would lead to within-issue competition between firms on the opposite ends of the ideological spectrum. As hypothesized, I find that firms respond to the previous giving not only by their ideological allies but also by their ideological opponents, with the response to opponents dominating the response to allies. This result is robust across a variety of econometric specifications, with the most stringent specification with firm-issue and issue-year level fixed effects, which absorb all time-invariant firm-issue specific heterogeneity and all issue-specific time trends. Moreover, the results from my supplementary analyses validate the posited mechanisms underlying my main findings. I find that selective response to ideological opponents’ giving is driven by firms whose entire body politic tend to be highly ideologically motivated and think similarly to each other, especially when they face a higher level of labor market competition. These findings are highly consistent with my theory that the observed counter-response to ideological opponents reflects ideological 143 firms’ competitive attempts to defend the issue positions held by internal organizational members. Moreover, selective response to ideological opponents is stronger in swing states where initiating actions of ideological firms would be perceived as salient threats for the firms with opposing ideologies, and stronger among firms whose organizational members hold radical and conservative preferences. Finally, I find that, when supporting a given issue, ideological firms of different positions tend to fund different nonprofits, using a within-issue, nonprofit recipient-level variation to detect the pattern of ideological clustering in recipient choices. My findings make contributions to multiple streams of research. My study contributes to a rapidly growing body of literature on corporate engagement in socio- political issues and, in particular, the line of literature that highlights the role of corporate organizational ideology as the main driver for such action. Recently, many practitioners and scholars called for attention to the mounting evidence that corporate actors are increasingly engaging in social issues of contentious and political nature, which have long been understood to lie outside the traditional boundary of firms’ participation. In light of this, scholars have started to explore the drivers of these activities, with a substantial body of research pointing to the role of political ideology among managers (Gupta, Nadkarni, and Mariam, 2019; Hambrick and Wowak, 2019) and employees as key internal drivers of such actions (Gupta et al., 2017; Gupta and Briscoe, 2019; McKean and King, 2019). However, the literature remains largely silent on the consequences of such corporate activism: in particular, how advocacy engagement by ideological firms affects the behaviors of other firms in the population. 144 On the one hand, ideological actors’ active participation in advocacy issues and its cascade effect on others can be socially beneficial as it increases the total supply level of corporate action around key socio-political issues from which many firms would have otherwise stayed away. In a sense, initiating ideological actors thus act as “critical mass” that solve collective action problem in the provision of social goods, as they are willing to pay the initial start-up costs of institutionalizing a given issue as a legitimate area of firms’ participation. On the other hand, ideological firms’ advocacy efforts may backfire, by inviting ideological opposition from their rivals and thus increasing conflict and polarization around the issue. This paper focuses on this tension, by showing that those initial actions by ideological firms do create an opening not just for the firms whose ideology is aligned to their own, but also for their ideological rivals who may undermine any issue-related changes that initiating ideological firms were trying to achieve. In doing so, I present a more nuanced and less optimistic take on the role of corporate activism in society, raising important questions about the appropriate limits on powerful business actors’ participation in civic discourse, and the use of tax-exempt philanthropic giving to those ends. Also, I bring into corporate activism and broader CSR literature a relatively new idea that growing socio-political contention around many issues in many societies can be leveraged by firms as a newly emerging source of strategic differentiation and competitive advantage. That is, firms can compete against each other not just based on traditional product or service attributes but also their stand on salient issues that stakeholders care about. This is a notably different take from the conceptualization in 145 existing studies on corporate activism that view such contention primarily as a source of strategic liability that expose firms to risks of alienating a portion of a firm’s key stakeholders who disagree (Burbano, 2020; Hambrick and Wowak, 2020). Yet, the imperatives of strategic differentiation suggests that such heterogeneous stakeholder preferences are not necessarily a bad thing from a strategic angle; in my paper, I highlight that ideological differences across different people can be used as a basis of firms’ novel differentiation strategy and value creation, as they espouse different positions to appeal to stakeholders with opposing preferences and thus maximize their competitive advantage. Second, my findings extend our understanding of how firms influence others’ social actions. Most studies on this topic tend to build on classic insight of the normative pressures and homogeneity in the institutional theory (DiMaggio and Powell, 1983; Haunschild and Miner, 1997) and show how unitary practice spreads throughout the field as actors in similar stakeholder environment develop a uniform consensus on the appropriate level and focus of corporate social action (Marquis et al., 2007; Marquis and Tilcsik, 2016), Yet, recent work that started to connect political psychology literature to CSR highlights how CSR can be contested, pointing to heterogeneous and often polarizing stakeholder preferences around social issues (Rao, 2009; King and Pearce, 2010). Importantly, this contested nature of social issues suggests that the diffusion process of CSR may not be as smooth as what one may expect based on existing studies, which is also the insight shared by broader diffusion and movement scholarship that emphasizes the fragile nature of institutionalization (Oliver, 1991; Fiss and Zajac, 2004; Lounsbury, 2008; Briscoe and Safford, 2008; Ansari et al., 2010; Fiss et al., 2012; Litrico 146 and David, 2017). My theory and findings extend this line of literature by offering a strategic differentiation-based account of corporate social counter-positioning (Porter, 1997; Zhao et al., 2017), which shows that the firm’s initial issue-related action creates strong economic incentives for its rivals to take an opposing issue position. In so doing, this paper also uncovers an underexplored source of contention in the diffusion process of firms’ social practices: the across-firm heterogeneity in the ideological preferences of the labor force. This study also brings new insight into the social movement literature, which usually try to find the source of contention around corporate social practices from the outside of the institution: for example, contention between social movement organizations called ‘extra-institutional entrepreneurs’ (King and Soule, 2007; King and Pearce, 2010) or issue entrepreneurs (McCarthy and Zald, 1977). In these accounts, firms are portrayed as relatively passive actors around the market of social and environmental issues whose market-related activities are affected as a result of exogeneous competition between these movement organizations that attempt proactive changes to the issue priority structure of a society to their own preferences—especially between ideological contest between organizations with opposing ideologies (e.g., rightist vs. leftist movement). Even in few papers that do recognize how firms can participate in such contention, their focus goes to the dynamics of competition between firms and other activist groups, where firms are considered to be unitary actors with homogeneous interests fighting to defend existing quo against activists’ effort to change the system. Yet, this paper’s focus on varying ideological preferences of a firm’s organizational 147 members suggests not only that some firms may play a proactive role of issue entrepreneurs to promote their organizational members’ unique interests to change the system, but also that competition between different ideological interests may even arise within the population of firms themselves. In doing so, this paper broadens our understanding on the role of for-profit actors in the market of advocacy issues. Figure 3.1. Firm-level distribution of ideological positions 148 Figure 3.2. Summary of Moderating Analyses Key mechanism in question Moderator Logic Prediction 1. Is the selective response to ideological opponents internally driven by employee preferences? 1) Executive vs. non- executive ideological intensity If the response to opponents reflects the internal ideological preferences of organizational members, both executive and non- executive ideology should matter Response to opponents will be strongest when both executive and non- executive ideological intensity are high 2) Within-firm ideological heterogeneity The more homogeneous the firm’s organizational members are in their ideology, the greater the returns from the response to ideological opponents Response to opponents will be stronger from firms with lower ideological heterogeneity 3) Labor market competition The more competitive the labor market is, the greater incentives for firms to give in ways that make employees happy Response to opponents will be stronger from firms that face a higher level of labor market competition 2. Is the selective response to ideological opponents the competitive attempts of counterpositioning? 1) Giving to swing vs. non- swing states Opponents’ giving would be perceived as bigger threats in swing states with a higher level of ideological competition Response to opponents will be concentrated in swing states than in non-swing states 2) Ideological radicals vs. ideological moderates Firms with more extreme ideological preferences would perceive the giving by opponents as bigger threats Response to opponents will be stronger among firms with radical preferences than among those with moderate preferences 3) Ideological conservatives vs. liberals Firms with conservative ideology would perceive opponents’ giving as bigger threats due to their greater aversion to uncertainty Response to opponents will be stronger among red firms with conservative ideology than among blue firms with liberal ideology 149 Table 3.1. Summary statistics and correlation table Mean SD Min Max 1 Giving by a focal firm 0.259 0.438 0.000 1.000 2 Giving by industry peers 0.787 0.410 0.000 1.000 3 Giving by other firms 0.997 0.053 0.000 1.000 4 Giving by ideological opponents 0.529 0.499 0.000 1.000 5 Giving by ideological allies 0.700 0.458 0.000 1.000 6 Firm size 9.679 2.710 0.000 14.936 7 ROA 0.046 0.072 -2.283 0.953 8 R&D-to-sales ratio 0.025 0.075 0.000 3.351 9 Advertising-to-sales ratio 0.013 0.030 0.000 0.313 10 Capex-to-sales ratio 0.060 0.100 0.000 2.955 11 Board female ratio 0.134 0.110 0.000 0.556 12 Lobbying-to-sales ratio 0.010 0.022 0.000 0.551 13 CSR rating 1.098 2.428 -7.000 13.000 14 Market concentration 0.031 0.043 0.011 0.321 15 Headquarter population 13.266 9.905 0.567 38.994 16 Headquarter poverty rate 13.378 2.868 6.800 25.800 150 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 1.000 2 0.069 1.000 3 0.009 0.043 1.000 4 0.068 0.552 0.040 1.000 5 0.080 0.796 0.040 0.315 1.000 6 0.188 0.051 -0.026 -0.002 0.068 1.000 7 -0.034 -0.088 -0.001 -0.059 -0.083 -0.017 1.000 8 -0.026 0.073 -0.015 0.083 0.056 0.003 0.038 1.000 9 0.006 -0.017 0.006 -0.069 -0.004 -0.034 0.203 0.092 1.000 10 -0.022 -0.100 0.009 -0.038 -0.105 0.035 -0.042 0.173 -0.048 1.000 11 0.137 -0.027 -0.013 -0.067 -0.009 0.302 0.104 0.036 0.141 -0.008 1.000 12 -0.003 -0.009 -0.012 0.004 -0.014 0.047 0.068 0.353 -0.010 0.163 0.050 1.000 13 0.078 0.061 -0.028 -0.040 0.104 0.230 0.089 0.191 0.163 -0.128 0.363 0.057 1.000 14 0.048 -0.398 0.000 -0.290 -0.322 0.057 0.090 -0.047 0.027 0.022 0.165 -0.070 -0.003 1.000 15 0.005 0.047 -0.026 -0.035 0.071 0.127 0.060 0.189 0.119 0.048 0.052 0.083 0.233 -0.127 1.000 16 0.019 -0.074 -0.015 -0.069 -0.074 0.177 -0.027 -0.034 0.050 0.113 -0.006 0.028 -0.002 0.088 0.388 1.000 Notes: Number of firm-issue-year observations for all variables is 25,440. Missing values of all variables are recoded as zero. 151 Table 3.2. Baseline Results M1 M2 M3 M4 M5 M6 Giving by industry peers 0.0364*** 0.0234* 0.0277** (0.0086) (0.0095) (0.0106) Giving by other firms 0.0021 0.0064 0.0022 0.0030 0.0071 0.0036 (0.0525) (0.0524) (0.0549) (0.0527) (0.0526) (0.0550) Giving by ideological opponents 0.0319*** 0.0379*** 0.0343*** (0.0072) (0.0079) (0.0086) Giving by ideological allies 0.0242** 0.0110 0.0139 (0.0076) (0.0084) (0.0093) Firm size 0.0226*** 0.0107*** 0.0107*** 0.0223*** 0.0107*** 0.0107*** (0.0020) (0.0024) (0.0025) (0.0021) (0.0025) (0.0025) ROA -0.2412*** -0.2136*** -0.2138*** -0.2423*** -0.2165*** -0.2164*** (0.0530) (0.0608) (0.0629) (0.0529) (0.0608) (0.0630) R&D-to-sales ratio -0.1297*** -0.1603** -0.1602** -0.1377*** -0.1627** -0.1622** (0.0373) (0.0508) (0.0525) (0.0379) (0.0506) (0.0524) Advertising-to-sales ratio -0.0785 -0.8327* -0.8333* -0.0571 -0.8364* -0.8358* (0.1824) (0.3429) (0.3551) (0.1820) (0.3436) (0.3556) Capex-to-sales ratio -0.0427 0.1007 0.1005 -0.0402 0.0990 0.0988 (0.0401) (0.0778) (0.0805) (0.0400) (0.0768) (0.0796) Board female ratio 0.2181*** 0.0413 0.0411 0.2184*** 0.0420 0.0418 (0.0397) (0.0526) (0.0545) (0.0397) (0.0527) (0.0546) Lobbying-to-sales ratio 0.4810** 0.6882** 0.6875* 0.4899** 0.6921** 0.6925* (0.1832) (0.2659) (0.2752) (0.1837) (0.2650) (0.2744) CSR rating 0.0019 -0.0065** -0.0065** 0.0019 -0.0065** -0.0065** (0.0019) (0.0022) (0.0023) (0.0019) (0.0022) (0.0023) Market concentration 0.5880*** 1.6427** 1.6406** 0.6322*** 1.6666** 1.6616** (0.1332) (0.5283) (0.5468) (0.1337) (0.5283) (0.5469) Population in HQ state -0.0004 0.0010 0.0010 -0.0003 0.0010 0.0010 (0.0005) (0.0062) (0.0064) (0.0005) (0.0062) (0.0064) Poverty rate in HQ state 0.0041* 0.0096*** 0.0096*** 0.0041* 0.0094*** 0.0094*** (0.0018) (0.0027) (0.0028) (0.0018) (0.0027) (0.0028) Constant -0.0108 -0.4055* -0.2053 -0.0177 -0.4022* -0.2101 (0.0646) (0.1761) (0.1815) (0.0651) (0.1757) (0.1814) Year FE Y Y Y Y Y Y Issue FE Y Y Y Y Firm FE Y Y Firm-Issue FE Y Y R2 0.0759 0.1940 0.3325 0.0763 0.1951 0.3329 N 25440 25440 25440 25440 25440 25440 Notes: Dependent variable is the dummy variable of advocacy giving of a focal firm in a given issue area in each year. Coefficients of OLS panel regressions. Robust standard errors clustered by firm-issue level are used. Standard errors in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 152 Table 3.3. Robustness analyses M7 M8 M9 M10 M11 Issue-level time trends as common shock Dynamic panel Alternative construction of variables General issue- level trend Geography- specific trends IV as amount of giving IV and DV as amount of giving Giving by ideological opponents 0.0310*** 0.0515*** (0.0091) (0.0104) Giving by ideological allies 0.0040 -0.0044 (0.0095) (0.0116) Giving by ideological opponents in different states 0.0331*** (0.0090) Giving by ideological allies in different states 0.0041 (0.0093) Amount of giving by ideological opponents’ giving 0.0037*** 0.0360*** (0.0009) (0.0087) Amount of giving by ideological allies’ giving 0.0016+ 0.0177* (0.0009) (0.0089) Giving by other firms 0.0153 0.0152 -0.0152 (0.0847) (0.0846) (0.0639) Amount of giving by other firms 0.0002 0.0019 (0.0037) (0.0369) Firm size 0.0110*** 0.0109*** 0.0092* 0.0107*** 0.1077*** (0.0025) (0.0025) (0.0046) (0.0025) (0.0251) ROA -0.2224*** -0.2225*** -0.1900** -0.2178*** -2.1227*** (0.0586) (0.0587) (0.0661) (0.0629) (0.5804) R&D-to-sales ratio -0.1658** -0.1658** -0.0783 -0.1609** -1.4543** (0.0521) (0.0521) (0.0809) (0.0524) (0.4794) Advertising-to-sales ratio -0.8731* -0.8721* -0.2581 -0.8298* -8.8828* (0.3583) (0.3584) (0.5188) (0.3557) (3.5933) Capex-to-sales ratio 0.0955 0.0953 -0.0814 0.0964 0.9446 (0.0780) (0.0779) (0.1035) (0.0794) (0.7405) Board female ratio 0.0455 0.0456 0.0385 0.0422 0.4513 (0.0543) (0.0543) (0.0752) (0.0546) (0.5460) Lobbying-to-sales ratio 0.7116** 0.7127** 0.1097 0.6934* 6.4344* (0.2753) (0.2752) (0.3411) (0.2740) (2.6755) CSR rating -0.0072** -0.0072** 0.0005 -0.0065** -0.0742** (0.0023) (0.0023) (0.0028) (0.0023) (0.0228) Market concentration 1.7041** 1.7017** -0.2303 1.6464** 15.3863** (0.5512) (0.5511) (0.8486) (0.5473) (5.3403) Population in HQ state 0.0018 0.0018 0.0061 0.0011 -0.0157 (0.0062) (0.0062) (0.0068) (0.0064) (0.0650) Poverty rate in HQ state 0.0091*** 0.0091*** 0.0070* 0.0095*** 0.1071*** (0.0027) (0.0027) (0.0033) (0.0028) (0.0273) Lag(amount of a focal firm’s advocacy giving) 0.2614*** (0.0095) constant -0.2831 -0.2846 -0.1308 -0.2131 -1.7293 (0.2485) (0.2485) (0.1351) (0.1839) (1.8654) Firm-issue FE Y Y Y Y Y Issue-year FE Y Y Year FE Y Y Y R2 0.3428 0.3429 0.3330 0.3964 N 25440 25440 25440 25440 25440 Notes: Dependent variable is the dummy variable of advocacy giving of a focal firm in an issue in each year, with an exception of M11 which use the logged amount of a focal firm’s advocacy giving in a given issue in each year as an alternative DV. M7, 8, 10 and 11 report the results of OLS panel regressions. M9 runs dynamic panel regression. Robust standard errors. Standard errors in parentheses. +p<0.10, *p<0.05, **p<0.01, ***p<0.001 153 Table 3.4. Mechanism analysis (1): Role of employees M12a M12b M12c M12d M13a M13b M14a M14b Executive vs. non-executive ideological intensity Ideological heterogeneity Labor market competition High exec & high non- exec intensity High exec & low non-exec intensity Low exec & high non- exec intensity Low exec & low non-exec intensity High Low High Low Giving by ideological opponents 0.0444** 0.0202 0.0277 0.0213 0.0356** 0.0378** 0.0331** 0.0449*** (0.0155) (0.0208) (0.0245) (0.0162) (0.0130) (0.0127) (0.0119) (0.0136) Giving by ideological allies 0.0027 0.0144 0.0340 0.0167 0.0252+ 0.0009 -0.0007 0.0272* (0.0198) (0.0222) (0.0237) (0.0155) (0.0132) (0.0153) (0.0148) (0.0133) Giving by other firms -0.0584 -0.1657 0.0327 0.1626** -0.0370 0.0124 0.0143 -0.0127 (0.0833) (0.1899) (0.0927) (0.0501) (0.0829) (0.0697) (0.0853) (0.0856) Firm size 0.0301 0.1476*** 0.1807*** 0.0030 0.0228*** 0.0128*** 0.0067* 0.0169*** (0.0201) (0.0400) (0.0282) (0.0041) (0.0051) (0.0035) (0.0029) (0.0044) ROA -0.3038* 0.0031 -0.2710* -0.2513** -0.1603+ -0.2697** -0.3806*** -0.1738* (0.1238) (0.1605) (0.1310) (0.0800) (0.0944) (0.0876) (0.1053) (0.0764) R&D-to-sales ratio -0.1207 -0.2019 0.4044 -0.1763* -0.2011* -0.1463 -0.1685 -0.1353* (0.1634) (0.2968) (0.3017) (0.0829) (0.0932) (0.1492) (0.1736) (0.0640) Advertising-to-sales ratio 0.1229 -0.0077 -0.1410 -1.0511* -1.8673** -0.5227 -0.6130 -1.2477** (0.8230) (1.5746) (0.8008) (0.5022) (0.5912) (0.4843) (0.5295) (0.4550) Capex-to-sales ratio 0.2960* 0.4909* -0.0594 0.1811 0.3164+ 0.0505 0.1050 -0.0406 (0.1259) (0.2243) (0.2049) (0.2129) (0.1715) (0.1009) (0.0869) (0.1053) Board female ratio 0.0598 -0.0865 0.0221 0.1947* 0.1146 0.0447 -0.0813 0.1666+ (0.1004) (0.1331) (0.2095) (0.0990) (0.0854) (0.0826) (0.0780) (0.0851) Lobbying-to-sales ratio 1.0385* -0.7789 -0.2000 1.9758* -0.2480 0.6356+ 0.9589* 0.7836* (0.4652) (1.0571) (0.4468) (0.8659) (0.4328) (0.3715) (0.3979) (0.3698) CSR rating -0.0112** 0.0112+ -0.0173* -0.0054 -0.0114*** 0.0026 -0.0036 -0.0104** (0.0041) (0.0057) (0.0072) (0.0046) (0.0032) (0.0038) (0.0033) (0.0034) Market concentration -0.5527 2.4659+ 2.9193 1.1077 2.3123*** 0.3728 -0.7000 0.3665 (1.0979) (1.4202) (1.9248) (0.9622) (0.6833) (1.0405) (1.2708) (0.7090) Population in HQ state 0.0110 -0.0023 0.0151 -0.0080 0.0020 -0.0226 0.0206+ -0.0105 (0.0165) (0.0362) (0.0238) (0.0076) (0.0082) (0.0154) (0.0115) (0.0073) Poverty rate in HQ state 0.0119* 0.0018 0.0079 -0.0000 0.0012 0.0143** 0.0162*** 0.0066 (0.0055) (0.0072) (0.0077) (0.0045) (0.0041) (0.0044) (0.0039) (0.0047) Constant -0.6368 -1.1119 -1.0609*** -0.1300 -0.2325 -0.2713* -0.4015 0.0387 (0.4755) (1.4706) (0.3038) (0.1011) (0.2453) (0.1157) (0.3060) (0.2324) R2 0.4198 0.4664 0.4490 0.3747 0.4210 0.3927 0.3882 0.3218 N 8222 4470 4435 8313 12288 12364 13883 10849 Notes: Dependent variable is the dummy variable of advocacy giving of a focal firm in a given issue area in each year. Coefficients of OLS panel regressions 154 Table 3.5. Mechanism analysis (2): Ideological competition M15a M15b M16a M16b M17a M17b Swing vs. non-swing states Ideological radicals vs. moderates Ideological liberals vs. conservatives DV: Advocacy giving in swing states DV: Advocacy giving in non- swing states Ideological radicals Ideological moderates Ideological liberals Ideological conservatives Giving by ideological opponents 0.0165* 0.0164* 0.0314** 0.0356** 0.0258** 0.0596** (0.0068) (0.0071) (0.0122) (0.0123) (0.0096) (0.0199) Giving by ideological allies 0.0015 0.0183** 0.0004 0.0249+ 0.0205+ 0.0083 (0.0073) (0.0068) (0.0127) (0.0135) (0.0113) (0.0165) Giving by other firms -0.0273 0.0048 0.0098 0.0045 0.0450 -0.3166 (0.0377) (0.0486) (0.0713) (0.0848) (0.0484) (0.2149) Firm size 0.0044* 0.0041* 0.0136*** 0.0113 0.0107*** 0.0155* (0.0021) (0.0020) (0.0029) (0.0072) (0.0027) (0.0072) ROA -0.1385** -0.1169** -0.2694*** -0.1356 -0.2946*** 0.0516 (0.0457) (0.0376) (0.0682) (0.0937) (0.0831) (0.1013) R&D-to-sales ratio -0.0385 -0.1164** -0.1652+ -0.1471* -0.1304* -0.4528* (0.0364) (0.0430) (0.0942) (0.0613) (0.0614) (0.1782) Advertising-to-sales ratio -0.4849+ -0.3791 -1.0380** 0.4469 -0.8206* -1.3865 (0.2553) (0.2723) (0.3495) (0.7125) (0.3650) (1.5507) Capex-to-sales ratio 0.0228 0.0600 0.0791 0.1562 0.0138 0.1803+ (0.0532) (0.0563) (0.0934) (0.1319) (0.1094) (0.1040) Board female ratio 0.0336 0.0437 -0.0548 0.1733* 0.0381 0.0166 (0.0419) (0.0432) (0.0790) (0.0809) (0.0595) (0.1353) Lobbying-to-sales ratio 0.2490 0.4120* 0.5363 0.7692* 0.7559* 1.1558+ (0.1947) (0.2031) (0.3695) (0.3834) (0.2984) (0.6189) CSR rating -0.0055** -0.0043* 0.0021 -0.0139*** -0.0043+ -0.0138** (0.0018) (0.0018) (0.0034) (0.0031) (0.0026) (0.0047) Market concentration 0.6896+ 1.7343*** 2.2071** 1.1207 2.0930*** -0.5653 (0.3977) (0.4531) (0.8320) (0.7342) (0.5787) (1.5061) Population in HQ state -0.0113* 0.0127* 0.0234* -0.0068 0.0046 -0.0187 (0.0048) (0.0062) (0.0108) (0.0075) (0.0073) (0.0163) Poverty rate in HQ state 0.0075*** 0.0041+ 0.0089* 0.0093* 0.0082* 0.0110* (0.0022) (0.0021) (0.0040) (0.0039) (0.0033) (0.0052) Constant 0.2445+ -0.4812** -0.2272 -0.0348 -0.3415+ 0.8867 (0.1370) (0.1732) (0.1399) (0.2310) (0.2016) (0.6664) R2 0.3516 0.3332 0.3190 0.3524 0.3436 0.3017 N 25440 25440 12439 13001 19464 5976 Notes: Dependent variable is the dummy variable of advocacy giving of a focal firm in a given issue area in each year. 155 Table 3.6. Recipient choices within each issue area M18 M19 M20 M21 Giving by ideological opponents vs. ideological allies Effect of ideological gap Dummy of giving Amount of giving Dummy of giving Amount of giving Giving by ideological opponents 0.0258* 0.0183+ (0.0107) (0.0109) Giving by ideological allies 0.0801*** 0.0801*** (0.0058) (0.0060) Ideological gap -0.0780*** -0.7538*** (0.0118) (0.1148) Giving by industry peers -0.0290** -0.0344*** (0.0099) (0.0102) Giving by other firms 0.0748*** 0.0749*** -0.0614*** -0.0720*** (0.0043) (0.0043) (0.0104) (0.0107) Firm size 0.0629*** 0.6374*** 0.0681*** 0.6797*** (0.0061) (0.0590) (0.0149) (0.1462) ROA -0.1891*** -1.6841*** -0.1357+ -1.1579+ (0.0388) (0.3566) (0.0789) (0.7034) R&D-to-sales ratio -0.2699*** -2.4126*** -0.5048* -4.5949+ (0.0707) (0.6557) (0.2422) (2.5008) Advertising-to-sales ratio -0.2200 -2.1526 -0.4932 -4.3183 (0.1802) (1.7938) (0.4019) (4.1265) Capex-to-sales ratio 0.1196** 1.1937** 0.2306* 1.9727* (0.0419) (0.3941) (0.0986) (0.9347) Board female ratio -0.0332 -0.2581 -0.0389 -0.2870 (0.0260) (0.2503) (0.0632) (0.6281) Lobbying-to-sales ratio 0.8986*** 8.1259*** 1.0080* 9.3552* (0.1553) (1.4970) (0.4050) (3.9988) CSR rating -0.0053*** -0.0502*** -0.0072** -0.0678** (0.0010) (0.0100) (0.0024) (0.0239) Market concentration 1.1644*** 10.0005*** 0.6532 3.7550 (0.2381) (2.4059) (0.4989) (5.1246) Headquarter population 0.0056+ 0.0250 0.0014 -0.0115 (0.0032) (0.0318) (0.0067) (0.0706) Headquarter poverty rate 0.0086*** 0.0871*** 0.0094** 0.0888** (0.0013) (0.0126) (0.0031) (0.0302) Constant -0.8631*** -7.6531*** -1.0973*** -9.4710*** (0.0982) (0.9821) (0.2316) (2.3403) Year FE Y Y Y Y Firm FE Y Y Y Y Recipient FE (& Issue FE) Y Y Y Y R2 0.0910 0.0993 0.2089 0.2112 N 93192 93192 21993 21993 Notes: Dependent variable is the dummy variable of advocacy giving of a focal firm in a given nonprofit in each year in M18 and M20, and the logged amount of advocacy giving of a focal firm in a given nonprofit in each year in M19 and M21. Coefficients of OLS panel regressions. Robust standard errors clustered by firm-recipient dyad level are used. Standard errors in parentheses. +p < 0.10, *p < 0.05, **p < 0.01, ***p < 0.001 156 GENERAL CONCLUSION AND DISCUSSION By the existing CSR accounts, corporate philanthropy is typically conceptualized as a form of delegated philanthropy, whereby firms give back to society in expectation of rewards from external stakeholders by funding of service nonprofits (e.g., soup kitchens, homeless shelters) that deliver social goods and services to people in need. While this view of corporate giving as a means of voluntary redistribution of corporate wealth to serve external social needs applies well to the traditional form of service giving, it does not hold in the case of corporate advocacy giving. The pattern of findings in the three chapters of my dissertation suggest that corporate advocacy giving may represent an overlooked yet important source of ideological influence of many firms’ internal organizational members, who often hold homogeneous and extreme preferences. Also, I found that ideological preferences of the firm’s organizational members are especially likely to translate into its advocacy giving activities, when such activities allow the firm to distance themselves from their competitors in terms of ideological issue positions and thus gain strategic benefits of competitive advantage. Academically, these findings contribute to nonmarket and corporate philanthropy literature by challenging and extending its scholarly conceptions of what corporate giving is. My findings suggest that firms give to advocacy not to meet exist external social needs and demands, but to cater to their internal organizational members’ private, ideological preferences to give to certain advocacy issues to promote their preferred issue positions. In doing so, I extend our view not only on why firms give but also raise concerns about the potential welfare implications of CSR and corporate philanthropic activities that firms 157 may use for their strategic benefits. These findings contribute not only to business literature examining the impact of corporate giving on both firms and society, but also to broader inter-disciplinary literature straddling economics, law, and political science that investigates the prevalence of money from private and ideological interests in socio- political discourse. In particular, this means that we need to move beyond our focus on campaign finance and lobbying in the existing literature (on which fine-grained data is readily available), and consider less visible avenues of influence such as corporate donation to fully understand the role of private, ideological interests in society. Practically, the findings from my dissertation point to multiple potential ways in which tax-exempt philanthropic activities of corporations may often fail to help society and, in some cases, even create social costs, thus raising the need for effective social monitoring on corporate donation and foundations. My findings for firm employees as a key driver of advocacy giving suggest that it may amplify their ideological voice in a way that is difficult for the public to discern. Given that these organizational members’ ideological preferences and issue-related views may significantly diverge from those of the broader society, it raises concerns not only about the welfare consequences of these giving activities, but also a fundamental question about if such use of corporate giving is democratic. After all, there exist a variety of regulations in place (e.g., campaign finance laws) that are designed to protect equal voice among individuals and to avoid giving greater voice to those with more power or financial resources. Yet, my dissertation suggests that corporate donation may serve as a less visible channel that may amplify the 158 impact of certain individuals who are often members of prestigious and financially resourceful organizations to begin with, and often hold extreme preferences. Moreover, my findings for the dynamics of ideological competition between firms with opposing ideologies suggest that corporate advocacy giving may also add to further polarization and gridlock around key issues in US society, which is already acknowledged as an issue of growing importance. Thus, my research proposes the need for effective social monitoring on corporate donation to ensure that it is allocated in ways that truly help society. Monitoring can be achieved through state regulation by enforcing stricter public disclosure requirements, and/or through collective efforts of private actors in society (e.g., the media, activists, consumers) to closely observe and evaluate how philanthropic dollars from firms are actually spent. Without such vigilance, my research shows, corporate philanthropy may fail to contribute to social benefits, despite the huge tax breaks it enjoys. 159 REFERENCES Abramowitz, Alan I. 2010. The disappearing center: Engaged citizens, polarization, and American democracy. New Haven: Yale University Press. Aiello, 2018. Social change is part of business, Ben & Jerry’s CEO says. CNBC (Auguest 17): https://www.cnbc.com/2018/08/17/ben--jerrys-ceo--social-changeis-part-of- business.html Ang, J. S., Cole, R. A., & Lin, J. W. (2000). Agency costs and ownership structure. Journal of Finance, 55(1), 81-106. Ansari, S. M., Fiss, P. C., & Zajac, E. J. (2010). Made to fit: How practices vary as they diffuse. Academy of management review, 35(1), 67-92. Ansolabehere, S., De Figueiredo, J. M., & Snyder Jr, J. M. (2003). Why is there so little money in US politics?. Journal of Economic perspectives, 17(1), 105-130. Barnett, M. L. (2019). The business case for corporate social responsibility: A critique and an indirect path forward. Business & Society, 58(1), 167-190. Barnett, W. P., & Woywode, M. (2004). From Red Vienna to the Anschluss: Ideological competition among Viennese newspapers during the rise of national socialism. American journal of sociology, 109(6), 1452-1499. Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99-120. Baron, D. P. (2001). Private politics, corporate social responsibility, and integrated strategy. Journal of Economics & Management Strategy, 10(1), 7-45. Battilana, J., Leca, B., & Boxenbaum, E. (2009). How actors change institutions: towards a theory of institutional entrepreneurship. Academy of Management annals, 3(1), 65-107. Baum, J. A., Li, S. X., & Usher, J. M. (2000). Making the next move: How experiential and vicarious learning shape the locations of chains' acquisitions. Administrative Science Quarterly, 45(4), 766-801. Baum, J.A.C., & Haveman, H.A. 1997. Love thy neighbour? Differentiation and agglomeration in the Manhattan hotel industry, 1898-1990. Administrative Science Quarterly, 42: 304-338. Bénabou, R., & Tirole, J. (2010). Individual and corporate social responsibility. Economica, 77(305), 1-19. Bermiss, Y. S., & McDonald, R. (2018). Ideological misfit? Political affiliation and employee departure in the private-equity industry. Academy of Management Journal, 61(6), 2182-2209. 160 Berry, J. M. (2015). Lobbying for the people: The political behavior of public interest groups. Princeton University Press. Bershin, 2018. Yes, CEOs, You Do Need To Speak Up On Social Issues. Fortune (Sep 5): https://www.forbes.com/sites/joshbersin/2018/09/05/yes-ceos-you-do-need-to- speak-up-on-social-issues/#2047224d5898 Bertels, S., & Peloza, J. (2008). Running just to stand still? Managing CSR reputation in an era of ratcheting expectations. Corporate Reputation Review, 11(1), 56-72. Bertrand, M., Bombardini, M., Fisman, R., & Trebbi, F. (2020). Tax-exempt lobbying: Corporate philanthropy as a tool for political influence. American Economic Review, 110(7), 2065-2102. Bertrand, M., Bombardini, M., Fisman, R., Hackinen, B., & Trebbi, F. (2018). Hall of Mirrors: Corporate Philanthropy and Strategic Advocacy. National Bureau of Economic Research. Available at: https://www.nber.org/papers/w25329 Bettis, R., Gambardella, A., Helfat, C., & Mitchell, W. (2014). Quantitative empirical analysis in strategic management. Strategic Management Journal, 949-953. Blee, K. M., & Creasap, K. A. (2010). Conservative and right-wing movements. Annual review of sociology, 36, 269-286. Bode, C., & Singh, J. (2018). Taking a hit to save the world? Employee participation in a corporate social initiative. Strategic Management Journal, 39(4), 1003-1030. Bode, C., Singh, J., & Rogan, M. (2015). Corporate social initiatives and employee retention. Organization Science, 26(6), 1702-1720. Bonardi, J. P., & Keim, G. D. (2005). Corporate political strategies for widely salient issues. Academy of Management Review, 30(3), 555-576. Bonica, A. (2013). Database on Ideology, Money in Politics, and Elections: Public version 1.0 [computer file]. Stanford, CA: Stanford University Libraries. Retrieved October, 1, 2013. Bonica, A. (2016). Avenues of influence: On the political expenditures of corporations and their directors and executives. Business and Politics, 18(4), 367-394. Briscoe, F., & Gupta, A. (2016). Social activism in and around organizations. The Academy of Management Annals, 10(1), 671-727. Briscoe, F., & Safford, S. (2008). The Nixon-in-China effect: Activism, imitation, and the institutionalization of contentious practices. Administrative science quarterly, 53(3), 460-491. Briscoe, F., & Safford, S. (2010). Employee affinity groups: Their evolution from social movement vehicles to employer strategies. 161 Briscoe, F., Chin, M. K., & Hambrick, D. C. (2014). CEO ideology as an element of the corporate opportunity structure for social activists. Academy of Management Journal, 57(6), 1786-1809. Briscoe, F., Gupta, A., & Anner, M. S. (2015). Social activism and practice diffusion: How activist tactics affect non-targeted organizations. Administrative Science Quarterly, 60(2), 300-332. Bromiley, P. (2010). Looking at prospect theory. Strategic Management Journal, 31(12), 1357-1370. Burbano, V. (2016). Social responsibility messages and worker wage requirements: Field experimental evidence from online labor marketplaces. Organization Science, 27(4), 1010-1028. Burbano, V. (2019). The Demotivating Effects of Communicating a Social-Political Stance: Field Experimental Evidence from an Online Labor Market Platform. Management Science, Forthcoming. Burden, Barry C. 2004. “Candidate Positioning in US Congressional Elections.” British Journal of Political Science 34(2): 211–227. Cable DM, DeRue DS (2002) The convergent and discriminant validity of subjective fit perceptions. Carlos, W. C., Sine, W. D., Lee, B. H., & Haveman, H. A. (2018). Gone with the wind: The evolving influence of social movements and counter movements on entrepreneurial activity in the US wind industry. Emerald Publishing Limited. Advances in Strategic Management. 38, 339-364 Carnahan, S., & Greenwood, B. N. (2018). Managers’ political beliefs and gender inequality among subordinates: Does his ideology matter more than hers? Administrative Science Quarterly, 63(2), 287-322. Carnahan, S., Kryscynski, D., & Olson, D. (2017). When does corporate social responsibility reduce employee turnover? Evidence from attorneys before and after 9/11. Academy of Management Journal, 60(5), 1932-1962. Carrillo, J. D., & Castanheira, M. (2008). Information and strategic political polarisation. The Economic Journal, 118(530), 845-874. Chatman, J. A. (1989). Improving interactional organizational research: A model of person-organization fit. Academy of management Review, 14(3), 333-349. Chatterji, A. K., & Toffel, M. W. (2019). Assessing the impact of CEO activism. Organization & Environment, 32(2), 159-185. Chin, M. K., & Semadeni, M. (2017). CEO political ideologies and pay egalitarianism within top management teams. Strategic Management Journal, 38(8), 1608-1625. 162 Chin, M. K., Hambrick, D. C., & Treviño, L. K. (2013). Political ideologies of CEOs: The influence of executives’ values on corporate social responsibility. Administrative science quarterly, 58(2), 197-232. Choi, S. J., Jia, N., & Lu, J. (2015). The structure of political institutions and effectiveness of corporate political lobbying. Organization Science, 26(1), 158- 179. Christensen, D. M., Dhaliwal, D. S., Boivie, S., & Graffin, S. D. (2015). Top management conservatism and corporate risk strategies: Evidence from managers' personal political orientation and corporate tax avoidance. Strategic Management Journal, 36(12), 1918-1938. Coff, R. W. (1997). Human assets and management dilemmas: Coping with hazards on the road to resource-based theory. Academy of management review, 22(2), 374- 402. Coke, T. E., Nielsen, S., Ramos, H. A., Seward, S., & Smith, B. K. Social Justice Grantmaking II: An Update on US Foundation Trends, ed. Steven Lawrence (New York: Foundation Center, 2009). Foundation Center. Cox, G. W. (1990). Centripetal and centrifugal incentives in electoral systems. American Journal of Political Science, 903-935. Cress, D. M., & Snow, D. A. (2000). The outcomes of homeless mobilization: The influence of organization, disruption, political mediation, and framing. American Journal of Sociology, 105(4), 1063-1104. De Figueiredo, J. M., & Richter, B. K. (2014). Advancing the empirical research on lobbying. Annual Review of political science, 17, 163-185. De Figueiredo, J. M., & Silverman, B. S. (2006). Academic earmarks and the returns to lobbying. The Journal of Law and Economics, 49(2), 597-625. De Figueiredo, J. M., & Tiller, E. H. (2001). The structure and conduct of corporate lobbying: How firms lobby the Federal Communications Commission. Journal of Economics & Management Strategy, 10(1), 91-122. De Sio, L., & Weber, T. (2014). Issue yield: A model of party strategy in multidimensional space. American Political Science Review, 108(4), 870-885. Delmas, M., Lim, J., & Nairn-Birch, N. (2016). Corporate environmental performance and lobbying. Academy of Management Discoveries, 2(2), 175-197. DeSimone, L. D., & Popoff, F. (2000). Eco-efficiency: the business link to sustainable development. MIT press. 163 DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American sociological review, 147-160. Dineen, B. R., & Noe, R. A. (2009). Effects of customization on application decisions and applicant pool characteristics in a web-based recruitment context. Journal of Applied psychology, 94(1), 224. Dorobantu, S., Henisz, W. J., & Nartey, L. (2017). Not all sparks light a fire: Stakeholder and shareholder reactions to critical events in contested markets. Administrative Science Quarterly, 62(3), 561-597. Dyke, N., Soule, S. A., & Taylor, V. A. (2004). The targets of social movements: Beyond a focus on the state. In Authority in contention (pp. 27-51). Emerald Group Publishing Limited. Earl, J., Martin, A., McCarthy, J. D., & Soule, S. A. (2004). The use of newspaper data in the study of collective action. Annual Review of Sociology, 30, 65-80. Elfenbein, D. W., & McManus, B. (2010). A greater price for a greater good? Evidence that consumers pay more for charity-linked products. American Economic Journal: Economic Policy, 2(2), 28-60. Fischbacher, U., & Gachter, S. (2010). Social preferences, beliefs, and the dynamics of free riding in public goods experiments. American economic review, 100(1), 541- 56. Fisher, J. (1999). Modelling the decision to donate by individual party members: the case of British parties. Party politics, 5(1), 19-38. Fiss, P. C., & Zajac, E. J. (2004). The diffusion of ideas over contested terrain: The (non) adoption of a shareholder value orientation among German firms. Administrative science quarterly, 49(4), 501-534. Fiss, P. C., Kennedy, M. T., & Davis, G. F. (2012). How golden parachutes unfolded: Diffusion and variation of a controversial practice. Organization Science, 23(4), 1077-1099. Flammer, C. (2015). Does product market competition foster corporate social responsibility? Evidence from trade liberalization. Strategic Management Journal, 36(10), 1469-1485. Flammer, C., & Luo, J. (2017). Corporate social responsibility as an employee governance tool: Evidence from a quasi‐experiment. Strategic Management Journal, 38(2), 163-183. Fremeth, A. R., & Richter, B. K. (2011). Profiting from environmental regulatory uncertainty: Integrated strategies for competitive advantage. California Management Review, 54(1), 145-165. 164 Friedman, M. (1970). A Friedman doctrine: The social responsibility of business is to increase its profits. The New York Times Magazine, 13(1970), 32-33. Gamson, W., & Meyer, D. (1996). Framing Political Opportunity in Comparative Perspectives on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings edited by McAdam D., McCarthy J., Zald M. Gardner, T. M. (2005). Interfirm competition for human resources: Evidence from the software industry. Academy of Management Journal, 48(2), 237-256. Garud, R., Jain, S., & Kumaraswamy, A. (2002). Institutional entrepreneurship in the sponsorship of common technological standards: The case of Sun Microsystems and Java. Academy of management journal, 45(1), 196-214. Gavetti, G., Helfat, C. E., & Marengo, L. (2017). Searching, shaping, and the quest for superior performance. Strategy Science, 2(3), 194-209. Gehringer, T. (2021). Corporate foundations as hybrid organizations: A systematic review of literature. Voluntas: International Journal of Voluntary and Nonprofit Organizations, 1-19. Godfrey, P. C. (2005). The relationship between corporate philanthropy and shareholder wealth: A risk management perspective. Academy of Management Review, 30(4), 777-798. Granovetter, M. (1978). Threshold models of collective behavior. American journal of sociology, 83(6), 1420-1443. Green, D. P., Kahneman, D., & Kunreuther, H. (1994). How the scope and method of public funding affect willingness to pay for public goods. Public Opinion Quarterly, 58(1), 49-67. Greene, W. H. (2012). Econometric Analysis, 7th Edition. Stern School of Business, New York University. Greve, H. R. (2000). Marketing niche entry decisions: Competition, learning, and strategy in Tokyo banking, 1894–1936. Academy of Management Journal, 43(5), 816-836. Gupta, A., & Briscoe, F. (2019). Organizational Political Ideology and Corporate Openness to Social Activism. Administrative Science Quarterly, Forthcoming. Gupta, A., Briscoe, F., & Hambrick, D. C. (2017). Red, blue, and purple firms: Organizational political ideology and corporate social responsibility. Strategic Management Journal, 38(5), 1018-1040. Haidt, J. (2012). The righteous mind: Why good people are divided by politics and religion. Vintage. 165 Haidt, J., Graham, J., & Joseph, C. (2009). Above and below left–right: Ideological narratives and moral foundations. Psychological Inquiry, 20(2-3), 110-119. Haines H. (1988). Black Radicals and the Civil Rights Mainstream, 1954-1970. University of Tennessee Press: Knoxville. Hall, Andrew B. 2015. “What happens when extremists win primaries?” American Political Science Review 109(1): 18-42. Hambrick, D. C., & Wowak, A. (2019). CEO Sociopolitical Activism: A Stakeholder Alignment Model. Academy of Management Review, Forthcoming. Hardin, R. (1982). Collective action. Resources for the Future. Haunschild, P. R., & Miner, A. S. (1997). Modes of interorganizational imitation: The effects of outcome salience and uncertainty. Administrative science quarterly, 472-500. Hawn, O., & Ioannou, I. (2016). Mind the gap: The interplay between external and internal actions in the case of corporate social responsibility. Strategic Management Journal, 37(13), 2569-2588. Helfat, C. E. (2007). Stylized facts, empirical research and theory development in management. Strategic Organization, 5: 185–192. Hiatt, S. R., & Carlos, W. C. (2019). From farms to fuel tanks: Stakeholder framing contests and entrepreneurship in the emergent US biodiesel market. Strategic Management Journal, 40(6), 865-893. Hiatt, S. R., Sine, W. D., & Tolbert, P. S. (2009). From Pabst to Pepsi: The deinstitutionalization of social practices and the creation of entrepreneurial opportunities. Administrative science quarterly, 54(4), 635-667. Hillman, A. J., Zardkoohi, A., & Bierman, L. (1999). Corporate political strategies and firm performance: indications of firm‐specific benefits from personal service in the US government. Strategic Management Journal, 20(1), 67-81. Hogan, L. 2019. The Rise of Corporate CEO Activism in the Age of Information. Kennedy School Review (Septermber 25): https://ksr.hkspublications.org/2019/2009/2025/the-rise-of-corporate-ceo- activism-in-the-age-of-information/. Hunt, S. A., Benford, R. D., & Snow, D. A. (1994). Identity fields: Framing processes and the social construction of movement identities. New social movements: From ideology to identity, 185, 208. Ingram, P., Yue, L. Q., & Rao, H. (2010). Trouble in store: Probes, protests, and store openings by Wal-Mart, 1998–2007. American Journal of Sociology, 116(1), 53- 92. 166 James, S. D., & Shaver, J. M. (2016). Motivations for voluntary public R&D disclosures. Academy of Management Discoveries, 2(3), 290-312. Jenkins, J. C., & Halcli, A. (1999). Grassrooting the system? The development and impact of social movement philanthropy, 1953–1990. Philanthropic foundations: New scholarship, new possibilities, 229-56. Jia, N. (2014). Are collective political actions and private political actions substitutes or complements? Empirical evidence from China's private sector. Strategic Management Journal, 35(2), 292-315. Jost, J. T. (2006). The end of the end of ideology. American psychologist, 61(7), 651. Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism as motivated social cognition. Psychological bulletin, 129(3), 339. Jost, J. T., Napier, J. L., Thorisdottir, H., Gosling, S. D., Palfai, T. P., & Ostafin, B. (2007). Are needs to manage uncertainty and threat associated with political conservatism or ideological extremity?. Personality and social psychology bulletin, 33(7), 989-1007. Kahn, K. F., Fridkin, K., Kenney, P. J., & Kenney, P. (1999). The spectacle of US Senate campaigns. Princeton University Press. Kahneman, D., & Lovallo, D. (1993). Timid choices and bold forecasts: A cognitive perspective on risk taking. Management science, 39(1), 17-31. Kaul, A., & Luo, J. (2018). An Economic Case for CSR: The Comparative Efficiency of For-Profit Firms in Meeting Consumer Demand for Social Goods. Strategic Management Journal, 39: 1650-1677. Kim, (2018). "LobbyView: Firm-level Lobbying & Congressional Bills Database." Working paper. Kim, In Song (2018). “LobbyView: Firm-level Lobbying & Congressional Bills Database.” Working paper available from http://web.mit.edu/insong/www/pdf/lobbyview.pdf Kimberlin, S. E. (2010). Advocacy by nonprofits: Roles and practices of core advocacy organizations and direct service agencies. Journal of Policy Practice, 9(3-4), 164- 182. King, B. G. (2008). A political mediation model of corporate response to social movement activism. Administrative Science Quarterly, 53(3), 395-421. King, B. G., & Pearce, N. A. (2010). The contentiousness of markets: Politics, social movements, and institutional change in markets. Annual review of sociology, 36, 249-267. 167 King, B. G., & Soule, S. A. (2007). Social movements as extra-institutional entrepreneurs: The effect of protests on stock price returns. Administrative Science Quarterly, 52(3), 413-442. Kirchberg, V. (1995). Arts Sponsorship and the State of the City. Journal of Cultural Economics, 19(4), 305-320. Koh, P. S., Qian, C., & Wang, H. (2014). Firm litigation risk and the insurance value of corporate social performance. Strategic Management Journal, 35(10), 1464– 1482. Kollman, K. (1998). Outside lobbying: Public opinion and interest group strategies. Princeton University Press. Korschun, D., Rafieian, H., Aggrawal, A., & Swain, S. 2018. Taking a stand: Consumer responses when companies get (or don’t get) political. SSRN Scholarly Paper no. ID 2806476, Rochester, NY. Kutcher, E. J., Bragger, J. D., & Masco, J. L. (2013). How interviewees consider content and context cues to person–organization fit. International Journal of Selection and Assessment, 21(3), 294-308. Lev, B., Petrovits, C., & Radhakrishnan, S. (2010). Is doing good good for you? How corporate charitable contributions enhance revenue growth. Strategic Management Journal, 31(2), 182–200. Liebeskind, J. P., Oliver, A. L., Zucker, L., & Brewer, M. (1996). Social networks, learning, and flexibility: Sourcing scientific knowledge in new biotechnology firms. Organization science, 7(4), 428-443. Lipsky, M. (1968). Protest as a political resource. The American Political Science Review, 62(4), 1144-1158. Litrico, J. B., & David, R. J. (2017). The evolution of issue interpretation within organizational fields: Actor positions, framing trajectories, and field settlement. Academy of Management Journal, 60(3), 986-1015. Lohmann, S. (1994). The dynamics of informational cascades: The Monday demonstrations in Leipzig, East Germany, 1989-91. World Politics., 47, 42. Lounsbury, M. (2007). A tale of two cities: Competing logics and practice variation in the professionalizing of mutual funds. Academy of management journal, 50(2), 289-307. Lounsbury, M. (2008). Institutional rationality and practice variation: New directions in the institutional analysis of practice. Accounting, organizations and society, 33(4- 5), 349-361. 168 Luo, J., Kaul, A., & Seo, H. (2018). Winning us with trifles: Adverse selection in the use of philanthropy as insurance. Strategic Management Journal, 39(10), 2591-2617. Lyon, T. P., Delmas, M. A., Maxwell, J. W., Bansal, P., Chiroleu-Assouline, M., et al. (2018). CSR needs CPR: Corporate sustainability and politics. California Management Review, 60(4), 5-24. Mackey, A., Mackey, T. B., & Barney, J. B. (2007). Corporate social responsibility and firm performance: Investor preferences and corporate strategies. Academy of Management Review, 32(3), 817-835. Marquis, C., & Lee, M. (2013). Who is governing whom? Executives, governance, and the structure of generosity in large US firms. Strategic Management Journal, 34(4), 483-497. Marquis, C., & Tilcsik, A. (2016). Institutional equivalence: How industry and community peers influence corporate philanthropy. Organization Science, 27(5), 1325-1341. Marquis, C., Glynn, M. A., & Davis, G. F. (2007). Community isomorphism and corporate social action. Academy of management review, 32(3), 925-945. Mason, Liliana. 2018. Uncivil agreement: How politics became our identity. Chicago: University of Chicago Press. Mattingly, J. E., & Berman, S. L. (2006). Measurement of corporate social action: Discovering taxonomy in the Kinder Lydenburg Domini ratings data. Business & Society, 45(1), 20-46. May, J. D. (1973). Opinion structure of political parties: the special law of curvilinear disparity. Political studies, 21(2), 135-151. McCarthy, J. D., & Wolfson, M. (1992). Consensus movements, conflict movements, and the cooptation of civic and state infrastructures. Frontiers in social movement theory, 273-297. McCarthy, J. D., & Zald, M. N. (1977). Resource mobilization and social movements: A partial theory. American journal of sociology, 82(6), 1212-1241. McCright, A. M., & Dunlap, R. E. (2000). Challenging global warming as a social problem: An analysis of the conservative movement's counter-claims. Social problems, 47(4), 499-522. McDonnell, M. H. (2016). Radical repertoires: The incidence and impact of corporate- sponsored social activism. Organization Science, 27(1), 53-71. McDonnell, M. H., & Cobb, J. A. (2019). Take a Stand or Keep Your Seat: Board Turnover after Social Movement Boycotts. Academy of Management Journal, Forthcoming. 169 McDonnell, M. H., & King, B. (2013). Keeping up appearances: Reputational threat and impression management after social movement boycotts. Administrative Science Quarterly, 58(3), 387-419. McDonnell, M. H., & King, B. G. (2018). Order in the court: How firm status and reputation shape the outcomes of employment discrimination suits. American Sociological Review, 83(1), 61-87. McDonnell, M. H., King, B. G., & Soule, S. A. (2015). A dynamic process model of private politics: Activist targeting and corporate receptivity to social challenges. American Sociological Review, 80(3), 654-678. McDonnell, M., & Darnell., S. (2020). Profiting from Protest: A Contingency Model of the Effects of Anti-Corporate Activism, Working paper. McKean, A. E., & King, B. G. (2019, July). Signing to Signal: Examining Factors Behind Progressive Corporate Activism. Academy of Management Proceedings (Vol. 2019, No. 1, p. 13454). Meyer, D. S. (2004). Protest and political opportunities. Annual Review of Sociology, 30, 125-145. Meyer, D. S., & Staggenborg, S. (1996). Movements, countermovements, and the structure of political opportunity. American journal of sociology, 101(6), 1628- 1660. Minkoff, D., & Agnone, J. (2010). Consolidating social change: the consequences of foundation funding for developing social movement infrastructures. American foundations: roles and contributions. Brookings Press, Washington, 347-367. Moeen, M., & Agarwal, R. (2017). Incubation of an industry: Heterogeneous knowledge bases and modes of value capture. Strategic Management Journal, 38(3), 566- 587. Mohliver, A., Crilly, D., & Kaul, A. (2019). In dubious battle: Issue salience, polarization, and corporate social counterpositioning. Polarization, and Corporate Social Counterpositioning (January 7, 2019). Murray, 2019. America’s CEOs Seek a New Purpose for the Corporation. Fortune (August 19): https://fortune.com/longform/business-roundtable-ceos- corporations-purpose/ O’Neil, M. (2018). What Business Wants. The Chronicle of Philanthropy, September 6. Odziemkowska, K., & McDonnell, M. H. (2019). Ripple Effects: How Firm-Activist Collaborations Reduce Movement Contention. Available at: https://ssrn.com/abstract=3428050. 170 Oelberger, C. R. (2018). Cui bono? Public and private goals in Nonprofit Organizations. Administration & Society, 50(7), 973-1014. Oliver, C., & Holzinger, I. (2008). The effectiveness of strategic political management: A dynamic capabilities framework. Academy of Management Review, 33(2), 496- 520. Oliver, P., Marwell, G., & Teixeira, R. (1985). A theory of the critical mass. I. Interdependence, group heterogeneity, and the production of collective action. American journal of Sociology, 91(3), 522-556. Olson, M. (1965). Collective action. In The invisible hand (pp. 61-69). Palgrave Macmillan, London. Oreskes, N., & Conway, E. M. (2011). Merchants of doubt: How a handful of scientists obscured the truth on issues from tobacco smoke to global warming. Bloomsbury Publishing USA. Oxley, J. E., Rivkin, J. W., Ryall, M. D., & Strategy Research Initiative. (2010). The strategy research initiative: Recognizing and encouraging high-quality research in strategy. Strategic Organization, 8: 377–386. Poole, K. T., & Rosenthal, H. (1984). The polarization of American politics. The journal of politics, 46(4), 1061-1079. Porter, M. E. (1997). Competitive strategy. Measuring business excellence. Porter, M. E., & Kramer, M. R. (2002). The competitive advantage of corporate philanthropy. Harvard Business Review, 80(12), 5–16. Raffaelli, R., & Glynn, M. A. (2014). Turnkey or tailored? Relational pluralism, institutional complexity, and the organizational adoption of more or less customized practices. Academy of Management Journal, 57(2), 541-562. Ramarajan, L., Bezrukova, K., Jehn, K. A., & Euwema, M. (2011). From the outside in: The negative spillover effects of boundary spanners' relations with members of other organizations. Journal of Organizational Behavior, 32(6), 886-905. Rich, A. (2005). Think tanks, public policy, and the politics of expertise. Cambridge University Press. Richter, B. K., & Werner, T. (2017). Campaign contributions from corporate executives in lieu of political action committees. The Journal of Law, Economics, and Organization, 33(3), 443-474. Rothman, S. M., Raveis, V. H., Friedman, A., & Rothman, D. J. (2011). Health advocacy organizations and the pharmaceutical industry: an analysis of disclosure practices. American Journal of Public Health, 101(4), 602-609. 171 Schneiberg, M., & Lounsbury, M. (2008). Social movements and institutional analysis. The handbook of organizational institutionalism, 648, 670. Schneider, B. (1987). The people make the place. Personnel psychology, 40(3), 437-453. Schneider, B., Goldstiein, H. W., & Smith, D. B. (1995). The ASA framework: An update. Personnel psychology, 48(4), 747-773. Scott, W. (2001). Institutions and Organizations, 2d ed. Thousand Oaks, CA: Sage. Selznick, P. (2014). The organizational weapon: A study of Bolshevik strategy and tactics. Glencoe, Ill.: Free Press. Seo, H., Luo, J., & Kaul, A. (2019). Giving a Little to Many or a Lot to a Few? The Returns to Variety in Corporate Philanthropy. Forthcoming. Strategic Management Journal. Available at: https://ssrn.com/abstract=3264902 Sharkey, A. J., & Bromley, P. (2015). Can ratings have indirect effects? Evidence from the organizational response to peers’ environmental ratings. American Sociological Review, 80(1), 63-91. Simons, T., & Ingram, P. (1997). Organization and ideology: Kibbutzim and hired labor, 1951-1965. Administrative Science Quarterly, 784-813. Skitka, L. J., & Tetlock, P. E. (1993). Providing public assistance: Cognitive and motivational processes underlying liberal and conservative policy preferences. Journal of Personality and Social Psychology, 65(6), 1205. Somaya, D., Williamson, I. O., & Lorinkova, N. (2008). Gone but not lost: The different performance impacts of employee mobility between cooperators versus competitors. Academy of Management Journal, 51(5), 936-953. Soule, S. A. (2009). Contention and corporate social responsibility. Cambridge University Press. Staggenborg, S. (1991). The pro-choice movement: Organization and activism in the abortion conflict. Oxford University Press. Su, W., & Tsang, E. W. K. (2015). Product diversification and financial performance: The moderating role of secondary stakeholders. Academy of Management Journal, 58(4), 1128–1148. Suárez, D. F. (2012). Grant making as advocacy: The emergence of social justice philanthropy. Nonprofit Management and Leadership, 22(3), 259-280. Surroca, J., Tribó, J. A., & Waddock, S. (2010). Corporate responsibility and financial performance: The role of intangible resources. Strategic management journal, 31(5), 463-490. 172 Tan, D., & Rider, C. I. (2017). Let them go? How losing employees to competitors can enhance firm status. Strategic Management Journal, 38(9), 1848-1874. Tan, J., & Tang, Y. (2016). Donate money, but whose? An empirical study of ultimate control rights, agency problems, and corporate philanthropy in China. Journal of Business Ethics, 134(4), 593-610. Thomsen, D. M. (2014). Ideological moderates won’t run: How party fit matters for partisan polarization in Congress. The Journal of Politics, 76(3), 786-797. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. science, 185(4157), 1124-1131. Useem, B. (1980). Solidarity model, breakdown model, and the Boston anti-busing movement. American Sociological Review, 357-369. Useem, M., & Kutner, S. I. (1986). Corporate contributions to culture and the arts: The organization of giving and the influence of the chief executive officer and of other firms on company contributions in Massachusetts. Nonprofit enterprise in the arts: Studies in mission and constraint, 93-112. Utych, S. M. (2020). Man bites blue dog: are moderates really more electable than ideologues?. The Journal of Politics, 82(1), 392-396. Van den Steen, E. (2010). On the origin of shared beliefs (and corporate culture). The RAND Journal of Economics, 41(4), 617-648. Chicago. Voegtlin, C., Crane, A., & Noval, L. (2019). When CEO sociopolitical activism attracts new talents: Exploring the conditions under which CEO activism increases job pursuit intentions. Available at SSRN 3410382. Walker, E. T. (2009). Privatizing participation: Civic change and the organizational dynamics of grassroots lobbying firms. American Sociological Review, 74(1), 83- 105. Walker, E. T. (2014). Grassroots for hire: Public affairs consultants in American democracy. Cambridge University Press. Walker, E. T., & Rea, C. M. (2014). The political mobilization of firms and industries. Annual Review of Sociology, 40, 281-304. Wang, H., & Qian, C. (2011). Corporate philanthropy and corporate financial performance: The roles of stakeholder response and political access. Academy of Management journal, 54(6), 1159-1181. Werner, T. (2015). Gaining access by doing good: The effect of sociopolitical reputation on firm participation in public policy making. Management Science, 61(8), 1989- 2011. 173 Werner, T. (2017). Investor reaction to covert corporate political activity. Strategic Management Journal, 38(12), 2424-2443. Werner, T., & Coleman, J. J. (2014). Citizens United, independent expenditures, and agency costs: Reexamining the political economy of state antitakeover statutes. The Journal of Law, Economics, & Organization, 31(1), 127-159. York, J. G., & Lenox, M. J. (2014). Exploring the sociocultural determinants of de novo versus de alio entry in emerging industries. Strategic Management Journal, 35(13), 1930-1951. Zhao, E.Y., Fisher, G., Lounsbury, M., & Miller, D. 2017. Optimal distinctiveness: Broadening the interface between institutional theory and strategic management. Strategic Management Journal, 38: 93-113. Zhao, X., & Murrell, A. J. (2016). Revisiting the corporate social performance‐financial performance link: A replication of Waddock and G raves. Strategic Management Journal, 37(11), 2378-2388 174 APPENDICES Appendix Figure 1.1 Notes: Baseline values represents the distribution of advocacy nonprofit presence across causes, as measured by % of nonprofit revenues in each area divided by aggregate revenue in all areas. Appendix Figure 1.2 Notes: % of general foundations’ advocacy giving to each area is calculated by dividing the total amount of general foundation giving to the area by the total amount of giving across all areas. 175 Appendix Table 1.1. List of full NTEE Code categorized as advocacy NTEE MAJOR CODE AND DESCRIPTION ADVOCACY CODE ADVOCACY CODE DESCRIPTION A. Arts, Culture, and Humanities A01 Alliances & Advocacy A05 Research Institutes & Public Policy Analysis B. Education B01 Alliances & Advocacy B05 Research Institutes & Public Policy Analysis C. Environmental Quality, Protection, and Beautification C01 Alliances & Advocacy C05 Research Institutes & Public Policy Analysis D. Animal-Related D01 Alliances & Advocacy D05 Research Institutes & Public Policy Analysis E. Health E01 Alliances & Advocacy E05 Research Institutes & Public Policy Analysis F. Mental Health, Crisis Intervention F01 Alliances & Advocacy F05 Research Institutes & Public Policy Analysis G. Diseases, Disorders, Medical Disciplines G01 Alliances & Advocacy G05 Research Institutes & Public Policy Analysis H. Medical Research H01 Alliances & Advocacy H05 Research Institutes & Public Policy Analysis I. Crime, Legal Related I01 Alliances & Advocacy I05 Research Institutes & Public Policy Analysis J. Employment, Job Related J01 Alliances & Advocacy J05 Research Institutes & Public Policy Analysis K. Food, Agriculture, and Nutrition K01 Alliances & Advocacy K05 Research Institutes & Public Policy Analysis L. Housing, Shelter L01 Alliances & Advocacy L05 Research Institutes & Public Policy Analysis M. Public Safety M01 Alliances & Advocacy M05 Research Institutes & Public Policy Analysis N. Recreation, Sports, Leisure, Athletics N01 Alliances & Advocacy N05 Research Institutes & Public Policy Analysis O. Youth Development O01 Alliances & Advocacy O05 Research Institutes & Public Policy Analysis P. Human Services - Multipurpose and Other P01 Alliances & Advocacy P05 Research Institutes & Public Policy Analysis Q. International, Foreign Affairs, and National Security Q01 Alliances & Advocacy 176 Q05 Research Institutes & Public Policy Analysis R. Civil Rights, Social Action, Advocacy R01 Alliances & Advocacy R02 Management & Technical Assistance R03 Professional Societies & Associations R05 Research Institutes & Public Policy Analysis R11 Single Organization Support R12 Fund Raising & Fund Distribution R19 Support N.E.C. R20 Civil Rights R21 Immigrants' Rights R22 Minority Rights R23 Disabled Persons Rights R24 Womens Rights R25 Seniors Rights R26 Lesbian & Gay Rights R27 Patients' Rights R28 Children's Rights R29 Employee & Workers' Rights(Draft Code R30 Intergroup & Race Relations R40 Voter Education & Registration R60 Civil Liberties R61 Reproductive Rights R62 Right to Life R63 Censorship, Freedom of Speech & Press R65 Freedom of Religion Issues R67 Right to Die & Euthanasia R99 Civil Rights, Social Action & Advocacy N.E.C S. Community Improvement, Capacity Building S01 Alliances & Advocacy S05 Research Institutes & Public Policy Analysis S21 Community Coalitions S22 Neighborhood & Block Associations U. Science and Technology U01 Alliances & Advocacy U05 Research Institutes & Public Policy Analysis V. Social Science Research Institutes, Services V01 Alliances & Advocacy V05 Research Institutes & Public Policy Analysis W. Public, Society Benefit W01 Alliances & Advocacy W05 Research Institutes & Public Policy Analysis W20 Government & Public Administration W22 Public Finance, Taxation & Monetary Policy W24 Citizen Participation W70 Leadership Development W90 Consumer Protection 177 Appendix Table 1.2. Summary statistics and correlation of variables Mean SD Min Max 1 Size 8.002 2.669 0.000 15.143 2 ROA 0.039 0.571 -16.118 46.455 3 R&D-to-Sales Ratio 0.018 0.055 0.000 1.555 4 Advertising-to-Sales Ratio 0.010 0.024 0.000 0.343 5 Capex-to-Sales Ratio 0.046 0.053 0.000 0.201 6 KLD Strength 1.116 1.769 0.000 14.000 7 KLD Concern 1.299 1.776 0.000 12.000 8 Market Concentration 0.033 0.042 0.000 0.427 9 Total Amount of Giving 2.873 5.653 0.000 18.806 1 2 3 4 5 6 7 8 9 1 1.000 2 0.007 1.000 3 0.084 -0.002 1.000 4 0.091 0.007 0.037 1.000 5 0.259 0.004 0.056 -0.008 1.000 6 0.345 0.014 0.197 0.128 0.079 1.000 7 0.315 0.005 0.026 -0.029 0.212 0.448 1.000 8 0.091 -0.003 -0.088 0.094 0.052 -0.009 0.056 1.000 9 0.264 0.008 0.057 0.070 0.060 0.351 0.240 0.002 1.000 Notes: Firm- and industry-level characteristics of Fortune 1,000 firms included as controls in regressions. Number of firm-year observations for all variables is 7,770. Missing values of all variables are recorded as zero. Given the high level of correlation between KLD Strength and KLD Concern, I use orthogonalized measures of each variables in all regressions. 178 Appendix Table 2.1. Replication of Results with Service Giving (AM1) (AM2a) (AM2b) (AM3a) (AM3b) Main Ideological heterogeneity Ideological radicalness High Low High Low Executive ideological intensity 0.0382 0.0691 0.0106 -0.0274 0.0679 (0.0669) (0.1274) (0.1354) (0.1115) (0.0851) Non-executive ideological intensity 0.0014+ 0.0017+ -0.0010 0.0004 0.0019* (0.0008) (0.0009) (0.0037) (0.0013) (0.0009) Giving by industry peers 0.0614 0.0791 0.0587 0.1093 0.0169 (0.0538) (0.0913) (0.0596) (0.0672) (0.0813) Giving by other firms -1.0637 1.3499 -0.0718 0.6765 -2.6733 (1.2265) (2.3466) (2.1774) (1.7431) (1.7527) Ideological position 0.2939 0.9232+ 0.6911+ 0.3119 0.3555 (0.1846) (0.4803) (0.3827) (0.2024) (0.3254) Size 0.0551 0.0382 -0.0035 0.1343 -0.0976 (0.0756) (0.1529) (0.1016) (0.0948) (0.1285) Total employees -0.0041 -0.0008 0.0044 0.0071 -0.0078 (0.0053) (0.0064) (0.0084) (0.0085) (0.0067) ROA -0.5087 -1.8339 0.0731 -0.5718 -0.7568 (0.8709) (1.1766) (2.2284) (1.2691) (1.1975) R&D-to-sales ratio -2.5007* -0.7645 -0.3711 -2.0413* -10.1899+ (1.0190) (3.2316) (1.9965) (0.9458) (5.6086) Advertising-to-sales ratio -14.6616 -8.9800 -1.3815 -23.7316+ 12.9267 (12.3496) (10.8028) (9.2230) (13.1103) (12.8901) Capex-to-sales ratio 2.1791+ 0.5261 0.4381 1.7531 3.3143 (1.2685) (3.8993) (1.7080) (1.2382) (3.3742) Board female ratio 0.4472 0.7550 1.8324 1.5115 -0.6883 (1.3187) (2.2862) (2.2214) (1.7660) (1.9939) Lobbying amount -0.0415 -0.0465 -0.0874+ -0.0435 -0.0413 (0.0261) (0.0442) (0.0467) (0.0415) (0.0317) CSR rating -0.0104 -0.0579 0.0574 -0.1315 0.0921 (0.0618) (0.0882) (0.1092) (0.1039) (0.0695) Market concentration 10.5610 33.8414 -7.8385 -0.8752 27.0461 (17.1295) (21.3908) (24.8122) (22.1079) (27.9149) Population in HQ state 0.1311 0.1022 -0.1279 0.1241 0.1069 (0.1368) (0.1342) (0.3518) (0.2486) (0.1562) Poverty rate in HQ state 0.1481+ 0.2646* 0.0317 0.0724 0.2507* (0.0768) (0.1200) (0.1285) (0.1065) (0.1093) Constant 18.9987 -23.8706 7.2150 -11.2547 47.6244 (21.3096) (40.9822) (37.2883) (30.2311) (30.5624) Year FE Y Y Y Y Y Firm FE Y Y Y Y Y R2 0.0473 0.0434 0.0384 0.0585 0.0477 N 5713 2475 2507 2866 2847 Notes: Dependent variable is the logged amount of service giving of a focal firm in a given each year. Coefficients of OLS panel regressions. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 179 Appendix Table 3.1. Economic magnitude of coefficients of Giving by Ideological Opponents and Ideological Allies Giving by ideological opponents Giving by ideological allies Table 3.2. Main results M4 Year, issue FE: Full model 12.6% 9.6% M5 Year, issue, firm FE: Full model 15.0% 4.4% M6 Year, firm-issue FE: Full model (*main model) 13.6% 5.5% Table 3.3. Robustness analysis M7 Firm-issue, issue FE: Full model 12.3% 1.6% M8 Firm-issue, issue FE: Using giving by firms in different states 13.1% 1.6% M9 Dynamic panel model 20.4% -1.7% Table 3.4. Mechanism analysis (1): Role of employees M12a High executive & High non-executive ideological intensity 15.0% 0.9% M12b High executive & Low non-executive ideological intensity 7.6% 5.4% M12c Low executive & High non-executive ideological intensity 11.4% 14.0% M12d Low executive & Low non-executive ideological intensity 9.8% 7.7% M13a Ideologically heterogeneous firms 13.3% 9.4% M13b Ideologically homogeneous firms 15.0% 0.4% M14a High labor market competition 12.3% -0.3% M14b Low labor market competition 20.2% 12.2% Table 3.5. Mechanism analysis (2): Ideological competition M15a Donation in swing states 11.5% 1.0% M15b Donation in non-swing states 12.3% 13.7% M16a Ideologically radical firms 13.1% 0.2% M16b Ideologically moderate firms 13.4% 9.4% M17a Ideologically liberal firms 9.8% 7.8% M17b Ideologically conservative firms 27.4% 3.8% Notes: Table shows the change in a focal firm’s subsequent probability of making advocacy giving to a given issue area over its mean value, which is associated with the change in the variables of Giving by Ideological Opponents and Giving by Ideological Allies from 0 to 1, based on their estimates in the corresponding model. 180 Appendix Table 3.2. Role of external stakeholders AM1a AM1b AM2a AM2b Product-market competition Consumer sensitivity High Low High Low Giving by ideological opponents 0.0168 0.0512*** 0.0153 0.0524*** (0.0120) (0.0132) (0.0122) (0.0121) Giving by ideological allies 0.0456** 0.0029 0.0261* 0.0032 (0.0154) (0.0120) (0.0128) (0.0135) Giving by other firms -0.0460 0.1019 -0.0401 0.0701 (0.0683) (0.0771) (0.0799) (0.0687) Firm size 0.0106*** 0.0064 0.0112* 0.0073* (0.0028) (0.0048) (0.0046) (0.0030) ROA -0.1229 -0.2203* -0.2618** -0.1905* (0.0782) (0.0929) (0.0828) (0.0875) R&D-to-sales ratio -0.1030 -0.1853*** -0.1092 -0.1684*** (0.1643) (0.0547) (0.1303) (0.0413) Advertising-to-sales ratio -1.1589+ -0.3454 (0.6425) (0.4751) Capex-to-sales ratio -0.0897 0.3131** -0.2163* 0.2519* (0.0981) (0.1079) (0.0852) (0.0993) Board female ratio -0.1893* 0.2075** 0.0566 0.0375 (0.0789) (0.0756) (0.0735) (0.0808) Lobbying-to-sales ratio 0.1669 1.5752*** 1.3998*** 0.4686 (0.3615) (0.4253) (0.3774) (0.4070) CSR rating -0.0049 -0.0101** -0.0010 -0.0137*** (0.0034) (0.0032) (0.0032) (0.0033) Population in HQ state 0.0002 0.0039 -0.0095 0.0252* (0.0078) (0.0109) (0.0070) (0.0117) Poverty rate in HQ state 0.0162*** 0.0041 0.0065 0.0133*** (0.0042) (0.0039) (0.0043) (0.0037) Market concentration 2.5124*** 0.2231 (0.6864) (0.8589) Constant -0.1080 -0.2136 0.1184 -0.2717 (0.2997) (0.3036) (0.2162) (0.2694) R2 0.3872 0.3219 0.3470 0.3236 N 12318 13122 12648 12792 Notes: Dependent variable is the dummy variable of advocacy giving of a focal firm in a given issue area in each year. Coefficients of OLS panel regressions with firm-issue and year fixed effects. Robust standard errors clustered by firm-issue level are used. Standard errors in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001 181 Appendix Table 3.3. Comparison of main results in service giving AM3 AM4 AM5 AM6 AM7 AM8 Giving by industry peers 0.0371*** 0.0277*** 0.0165* (0.0058) (0.0065) (0.0071) Giving by other firms -0.1069* -0.1109** 0.0551 -0.1022* -0.1132** 0.0469 (0.0433) (0.0412) (0.0447) (0.0433) (0.0412) (0.0447) Giving by ideological opponents 0.0273*** 0.0205*** 0.0172** (0.0052) (0.0057) (0.0063) Giving by ideological allies 0.0241*** 0.0245*** 0.0122+ (0.0053) (0.0059) (0.0064) Firm size 0.2202*** 0.0295 0.0299 0.2181*** 0.0288 0.0292 (0.0115) (0.0188) (0.0187) (0.0114) (0.0187) (0.0187) ROA 0.1380 -0.0477 -0.0369 0.1443 -0.0602 -0.0478 (0.2018) (0.2164) (0.2156) (0.2014) (0.2160) (0.2152) R&D-to-sales ratio 0.2692 -0.3209 -0.3168 0.1910 -0.3193 -0.3163 (0.3940) (0.4860) (0.4838) (0.3938) (0.4858) (0.4837) Advertising-to-sales ratio -1.1732 -4.6556* -4.6719* -1.0992 -4.5420* -4.5751* (1.1379) (2.2180) (2.2118) (1.1361) (2.2101) (2.2052) Capex-to-sales ratio -0.9607*** -0.1180 -0.1025 -0.9299*** -0.1743 -0.1430 (0.2344) (0.2872) (0.2861) (0.2345) (0.2874) (0.2863) Board female ratio 2.7573*** 1.4734*** 1.4852*** 2.7781*** 1.4653*** 1.4787*** (0.2490) (0.3155) (0.3145) (0.2491) (0.3152) (0.3144) Lobbying-to-sales ratio 9.5440*** 7.8099*** 7.8148*** 9.6692*** 7.8513*** 7.8456*** (1.0898) (1.3619) (1.3583) (1.0901) (1.3595) (1.3567) CSR rating 0.0569*** -0.1002*** -0.1001*** 0.0547*** -0.1006*** -0.1004*** (0.0141) (0.0176) (0.0175) (0.0141) (0.0176) (0.0175) Market concentration 7.7014*** 18.7688*** 18.8187*** 8.5269*** 17.9635*** 18.1415*** (1.2875) (3.1883) (3.1774) (1.3108) (3.1850) (3.1724) Headquarter population -0.0284*** 0.1201+ 0.1207+ -0.0284*** 0.1231* 0.1227* (0.0035) (0.0626) (0.0623) (0.0035) (0.0626) (0.0624) Headquarter poverty rate 0.0582*** 0.0582*** 0.0583*** 0.0601*** 0.0577*** 0.0579*** (0.0112) (0.0156) (0.0156) (0.0112) (0.0156) (0.0156) Constant 3.3775*** -0.6753 -2.0499+ 3.0885*** -0.7631 -2.0283+ (0.8686) (1.7213) (1.0674) (0.8682) (1.7165) (1.0644) Year FE Y Y Y Y Y Y Issue FE Y Y Y Y Firm FE Y Y Firm-Issue FE Y Y P-value in Chi-square test comparing giving by ideological opponents and allies 0.7066 0.6361 0.5917 R2 0.1670 0.4483 0.0050 0.1676 0.4485 0.0056 N 71368 71368 71368 71368 71368 71368 Notes: Dependent variable is the logged amount of service giving of a focal firm in a given issue area in each year. Coefficients of OLS panel regressions. Robust standard errors clustered by firm-issue level are used. Standard errors in parentheses. + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001