Active Ownership

Abstract

We analyze an extensive proprietary database of corporate social responsibility engagements with U.S. public companies from 1999–2009. Engagements address environmental, social, and governance concerns. Successful (unsuccessful) engagements are followed by positive (zero) abnormal returns. Companies with inferior governance and socially conscious institutional investors are more likely to be engaged. Success in engagements is more probable if the engaged firm has reputational concerns and higher capacity to implement changes. Collaboration among activists is instrumental in increasing the success rate of environmental/social engagements. After successful engagements, particularly on environmental/social issues, companies experience improved accounting performance and governance and increased institutional ownership.

Major institutional investors are often termed “universal owners” because of their diversified and ultra-long-term holdings with substantial ownerships (Monks and Minow 1995; Hawley and Williams 2000a, 2000b; Mattison, Trevitt, and van Ast 2011; Dimson et al. 2013). Since these investors own most of the equities in the market, their portfolios are exposed to risks from corporate externalities, and it is in their interest to minimize the potential costs and maximize the potential benefits of those externalities by influencing investee firms' businesses. At the same time, socially responsible investing (SRI), which seeks to deliver social as well as financial benefits, has attracted increasing attention.¹ There are 8,346 companies in 161 countries that now commit to responsible and sustainable corporate practices under the UN Global Compact (2015). The Principles for Responsible Investment (2015) lists 1,387 signatories with over $59 trillion in assets under management, and the Global Sustainable Investment Alliance (2015) estimates that $21.4 trillion of professionally managed assets worldwide incorporate environmental, social, and governance (ESG) concerns into their decisions. A growing proportion of pension funds, insurance companies, endowments, sovereign funds, and other investors is engaging with public firms on these issues (Goldstein 2011); the world's largest asset owners are addressing social and environmental concerns (Skancke et al. 2014); more social issue resolutions are being filed (Glac 2010; Carroll et al. 2012; Katz and McIntosh 2015); and the extent of engagement between corporations and investors is now said to be at an all-time high (Goldstein 2014). In brief, reflecting their ESG concerns, business owners are increasingly exercising their rights to influence the way businesses are managed.

Active engagement by universal owners on ESG issues (hereafter “ESG activism” or “active ownership”) differs in motivation from traditional shareholder activism by institutions, such as pension funds and mutual funds. It also differs from hedge fund activism and, more generally, from entrepreneurial activism.² Traditional shareholder activism and hedge fund activism typically focus on issues related to the interests of shareholders only, whereas ESG activism focuses on issues related to the interests of a broader range of stakeholders, including employees, customers, and creditors. Universal owners have multiple roles, for example, as shareholders or creditors, and long-term fiduciary responsibilities to their customers, beneficiaries, and the wider community (Hawley and Williams 2000a, 2000b): this explains their focus on broader stakeholders' interests. Consistent with this view, there is an emerging literature that emphasizes the potential positive role of nonshareholder stakeholders in companies' values and corporate governance systems, including Zingales (2000), Jensen (2001), Acharya, Myers, and Rajan (2011), and Allen, Carletti, and Marquez (2015).

Despite the growing prevalence of active ownership, data limitations have left unanswered even the most basic questions about ESG activism: Which firms do active owners engage, and how are these engagements executed? Do active owners compete or collaborate with other shareholders, and with what effect? How do engaged firms respond? What determines the success of these engagements? How does the market react to engagements? Do active owners succeed in implementing their objectives? And, more fundamentally, how do ESG activities affect firm performance? In this paper, drawing on a proprietary dataset of environmental, social, and governance engagements and outcomes, we address the above questions.

Our dataset is unusual in being a point-in-time record of active engagements. It has been provided by a large institutional investor with a major commitment to responsible investment. During the period spanned by our data, the firm was ranked between 80th and 100th largest in the world by assets under management (P&I 2014). The organization's heritage of responsible investing extends back to its first ethical fund, launched in 1984, and it uses its influence as one of the world's major shareholders to promote the adoption of good ESG practices. It actively engages in dialogues with target companies (4,186 of them in 2014) via letters, emails, telephone conversations, and direct conversations with senior management. It exercises ownership rights at shareholders' meetings on behalf of both its internal and external clients (e.g., voting globally on 60,076 resolutions in 2014), in addition to screening out irresponsible companies from its investment portfolios. In a typical year, the asset manager achieves the change it seeks in several hundred cases (244 in 2014). The asset manager also intensifies its efforts through active partnerships with other investors, such as SRI, pension, and religious funds (we refer to this as “hard collaboration”), and by working with bodies such as the Extractive Industries Transparency Initiative and the Coalition for Environmentally Responsible Economies (“soft collaboration”). Engagements are compiled as a detailed electronic file. Although the asset manager engages worldwide, this paper focuses on engagements with U.S. public companies.

We examine highly intensive engagements on environmental, social, and governance areas, each of which is further divided into different themes and issues. Given the relative lack of research on environmentally and socially themed engagements, we emphasize the environmental and social (ES) engagements throughout the paper and use the corporate governance (CG) engagements as a basis for comparison.³ Our primary sample consists of 2,152 engagement sequences (1,252 ES-, and 900 CG-based sequences) for 613 public firms between 1999 and 2009. The success rate for engagements in our sample is 18%, and, on average, it takes a sequence of 2–3 engagements before success can be recorded. The elapsed time from initial engagement to success averages nearly one-and-a-half years; the median time is one year. In comparison to CG themes, the chance of achieving success for ES themes is lower (13% vs. 24%), and the number of engagements per sequence is higher (3.7 vs. 2.2).

Compared to a matched sample of companies, firms are more likely to be engaged if they are large, mature, and performing poorly. The likelihood of being engaged is further increased if the asset manager and other socially conscious institutional investors (such as pension activists and SRI funds) have high shareholdings in the firm. Engagement is also more likely if reputation is important for the target company and if the company has inferior governance. The asset manager's ownership plays a less important role in relation to ES engagements than to CG engagements. On the other hand, reputational concerns are a more important determinant of engagement with firms on ES themes. These last two results indicate the importance of potential collaborations with other stakeholders and of customer opinion and loyalty, notably in consumer-facing industries, for the active ownership.

Conditional on being engaged, which firms are more likely to implement the asset manager's proposed changes? We refer to cases in which changes are implemented as successful engagements. Success is more likely if the target firm has reputational concerns, a capacity to implement change, economies of scale, and headroom for improvement. For the ES engagements, we find reputational concerns and a capacity to change play a more important role in achieving success. This may be attributed to the costly nature of improvements in areas related to ES dimensions.

Analyzing the engagement features and tactics, we find that successful prior engagement experience with the same target firm increases the likelihood of subsequent engagements being successful. In addition, we find collaborations among the asset manager and other active investors and/or stakeholders to contribute positively to the success of engagements, particularly for the ES engagements. This suggests that it requires more coordinated effort to convince an engaged company's management regarding the ES issues, in comparison to CG issues.

How does the market react to ESG activism? We find that ESG engagements generate a cumulative size-adjusted abnormal return of +2.3% over the year following the initial engagement. Cumulative abnormal returns are much higher for successful engagements (+7.1%) and gradually flatten out after a year, when the objective is accomplished for the median firm in our sample. We do not find any market reaction to unsuccessful engagements. The abnormal return patterns and magnitudes are similar for the subsamples of CG and ES engagements.⁴ This suggests the existence of a threshold for success to be pursued and achieved for both types of engagements. We then examine the cross-section of abnormal returns (controlling for industry and year fixed effects) and find that the positive market reaction to successful engagements is most pronounced for the themes of corporate governance and climate change. For these themes, the cumulative abnormal return of an additional successful engagement over a year after the initial engagement averages +8.6% and +10.3%, respectively.

To investigate the sources of the positive market reaction to successful engagements, we take a difference-in-differences approach and examine the subsequent changes in target firms' operating performance, profitability, efficiency, institutional ownership, stock volatility, and governance after successful engagements relative to after unsuccessful engagements. We observe significant improvements in all these measures (i.e., an increase in firm performance, investor base, and governance, and a decrease in stock return volatility) following successful engagements, as compared to the unsuccessful ones. Particularly focusing on the ES and CG subsamples, we first find that the return on assets and the ratio of sales to the number of employees improve significantly one year after successful ES engagements, as compared to the unsuccessful ones; but such improvements are less pronounced for successful CG engagements. These findings support the view that successful ES initiatives enhance customer and employee loyalty. Second, we observe an increase in shareholdings by the asset manager, pension activists, and SRI funds one year after successful ES engagements; but such an increase is not apparent for successful CG engagements. These results support the view that ES initiatives generate a clientele effect among shareholders. Third, we find improvements in the corporate governance structure of targeted firms, as measured by the Bebchuk, Cohen, and Ferrell (2009) entrenchment index, two years after successful engagements on all ESG issues. This suggests that good ESG practices signal improving governance quality.

We conclude that environmental, social, and governance activism of the type that we study improves social welfare to the extent that it increases stakeholder value when engagements are successful and does not destroy firm value even when engagements are unsuccessful. We note that, after successful engagements (particularly on ES issues), firms with inferior governance subsequently improve their governance and performance. Our interpretation is that active ownership attenuates managerial myopia and hence helps to minimize intertemporal losses of profits and negative externalities (see Benabou and Tirole 2010). This approach is differentiated from other styles of shareholder action, particularly hedge fund activism. Responsible investment initiatives are less confrontational, more collaborative, and more sensitive to public perceptions; yet they achieve success.

Our paper makes new contributions on four dimensions, which we outline here and expand on in the next section. To our knowledge, this study is the first to examine the impact of shareholder activism on environmental and social issues. We are the first to document the nature and impact of collaboration among these institutional activists. We contribute to the growing literature on corporate social responsibility (CSR) that mainly focuses on the link between responsible investing and firm performance, but suffers from data limitation and methodological criticisms. Finally, our paper complements the extensive literature on shareholder activism and corporate governance.

1. Literature and Hypotheses

In this section, we locate our paper in the context of an extensive body of research that embraces both corporate governance and corporate social responsibility. We start with a necessarily brief overview of the relevant literature and explain how our paper fits within this area. Based on our summary of the literature, we then outline hypotheses that are testable using our database. The data are described in Section 2.

1.1 Literature and contributions of this paper

Corporate governance has been studied extensively, and good practices have been shown to be important for shareholder value (see Shleifer and Vishny 1997; Gompers, Ishii, and Metrick 2003; Bebchuk, Cohen, and Ferrell 2009, among others). Traditional shareholder activism, whether through engaging with investee companies or through responding to shareholder proposals, emphasizes corporate governance. This activity is judged by Smith (1996), Karpoff, Malatesta, and Walkling (1996), and Gillan and Starks (2000) to provide, at best, negligible benefits to shareholders. Hedge fund activism generates considerable abnormal stock returns (+7% to +10%), but typically through engagements on issues such as business strategy, takeover decisions/activities, obtaining board seats, or financial engineering (see Brav et al. 2008; Greenwood and Schor 2009; Klein and Zur 2009; Brav, Jiang, and Kim 2015). In a long-only activist fund, Becht et al. (2009) report the highest positive market reactions when the stated objectives are associated with restructuring activities. Klein and Zur (2009) find that other entrepreneurial activists generate the highest performance when they intend to buy more stocks of the target or to become active investors. Bebchuk, Brav, and Jiang (2015) find no evidence of adverse effect of hedge fund activism on long-term interests of firms and its shareholders, whereas Klein and Zur (2011) document that hedge fund activism reduces bondholders' wealth.

Our paper introduces and examines the impact of shareholder activism on issues such as environmental and social questions. This is a form of owner behavior that differs in objectives, tactics, and outcomes from traditional shareholder activism. Yet it also differs from the approach of hedge funds and other entrepreneurial activists. Over the last decade, the number of shareholder proposals on environmental and social issues filed with the SEC has increased, and approval rates for these proposals have risen (Glac 2010; Welsh and Smith 2011; Allen et al. 2011; Flammer 2015; Sullivan & Cromwell 2014; Katz and McIntosh 2015). Given the increasing importance and prevalence of shareholder resolutions on environmental and social issues, our study provides timely feedback on this recent phenomenon and fills in a gap in the literature. Our engagement data also includes corporate governance issues, which in part goes beyond standard corporate governance activities, and addresses issues such as diversity, CSR disclosure, stock-option expensing, and “say on pay.” Ferri and Sandino (2009) and Ertimur, Ferri, and Muslu (2011) discuss the increased activities of shareholders on voluntarily expensing stock options and “say on pay” after the Enron scandal and the 2002 Sarbanes-Oxley Act.

The literature on CSR and SRI includes many studies of the link between responsible investing and firm performance. In a survey of all the CSR studies published in the management field over 1972–2007, Margolis, Elfenbein, and Walsh (2009) find that most studies report a nonsignificant relation, a small proportion document a negative relation, and only a quarter find a positive relation. They conclude that the overall impact of CSR on firm performance is positive but small, a finding that is confirmed, by Renneboog, ter Horst, and Zhang (2008). Moreover, most studies are subject to methodological criticisms, such as endogeneity. Indeed, Edmans (2012), who reviews a large number of studies in both management and finance, concludes that all prior work fails to identify a CSR dimension that improves risk-adjusted stock returns over the long haul.

As noted by Ferrell, Liang, and Renneboog (2014) and others, a drawback of most previous research is that it has relied on static and delimited measures for CSR performance, such as the scores of firms' social responsibility produced by KLD (now MSCI). In contrast, this paper benefits from the dynamic and incremental nature of our dataset, enabling us to conduct event-study analyses and to link subsequent changes in firm performance to prior ESG activities. This offers an improved prospect of discerning causality, rather than simply noting measures of association. In addition, instead of “the convenient yet difficult to validate measures such as the Fortune ratings of admired companies and company insiders’ self-reported impressions” (Margolis, Elfenbein, and Walsh 2009), our data are objective and quantified.

A critical aspect of the activism is the collaboration among activists during their engagements. Despite the importance of collaborations in engagements, the absence of data has limited the empirical research in this area. Using survey evidence, McCahery, Sautner, and Starks (2015) report that 59% of institutional investors would consider coordinating their actions with respect to shareholder activism. Gillan and Starks (2007) show that voting support in favor of shareholder proposals has increased over time, citing “more concerted action by institutional investors,” together with “the existence of proxy voting advisory firms” and “public disclosure of mutual fund proxy votes,” as explanatory factors. Artiga González and Calluzzo (2014) perceive coordination as offering the possibility of superior outcomes from target firms.

We expect the coordination and partnership among institutions and stakeholders to be particularly prevalent for environmental and social engagements. This is because these changes are less standard and often more costly to implement, and it is hence single-handedly more difficult to convince the management to make changes. We hand-collect the information on the asset manager's collaboration with other investors and stakeholders and analyze the impact of these collaboration activities on targeting and success of engagements. To our knowledge, this is the first paper to present direct evidence on collaboration among ES activists.

1.2 Testable hypotheses

The theoretical literature on corporate responsibility is relatively new and developing. Benabou and Tirole (2010) summarize this literature, offering three views with different predictions for the impact of CSR on firm value. One view is that CSR practices allow management to take a long-term perspective and maximize intertemporal profits, consistent with the interests of universal owners. This assessment accords with recent evidence, such as the study by Kim, Park, and Wier (2012), that more truthful firms, as judged by their aversion to earnings management, tend to be more active on CSR issues. Benabou and Tirole articulate a second view that socially responsible businesses act as an efficient channel to express personal values on behalf of their stakeholders, which may be regarded as a form of delegated philanthropy. While these two views would imply a positive impact of corporate responsibility on firm value, a third is that CSR reveals insider-initiated corporate philanthropy or a managerial agency problem. Benabou and Tirole note that, in this scenario, CSR activities would most likely be value destroying.⁵

For the enhancement of firm value through ESG activism, the literature highlights four channels. First, more socially conscious consumers have greater customer loyalty, and increased product differentiation supports premium pricing (Besley and Ghatak 2007; Albuquerque, Durnev, and Koskinen 2014). Second, firms with high employee satisfaction tend to outperform the market (Edmans 2011, 2012). Third, more virtuous companies attract a broader clientele than “sinful” companies (Grossman and Sharpe 1986; Hong and Kacperczyk 2009; Dimson, Marsh, and Staunton 2015), and political leanings, which attract particular stockholder clienteles, also influence corporate behavior (Di Giuli and Kostovetsky 2014). Fourth, successful investor interventions signal future governance improvements (Gompers, Ishii, and Metrick 2003; Brav et al. 2008; Klein and Zur 2009). In addition, engaged firms may be induced to look for improvements in other areas.

Based on the literature on CSR and activism, we summarize a number of predictions that are prime candidates for our empirical setting. First, following the discussion above, companies are more likely to be targeted for engagement when they are sensitive to perceptions regarding their reputation. Second, engagements are more likely to achieve success when target firms operate in consumer-oriented businesses with high customer awareness and loyalty, as discussed by Servaes and Tamayo (2013) and Albuquerque, Durnev, and Koskinen (2014) (also see Lev, Petrovits, and Radhakrishnan 2010). Third, companies benefiting from economies of scale and having the capacity and headroom for improvements are more likely to respond to engagement. Fourth, we expect a positive association between investor collaboration and the success rate of ESG activism, based on McCahery, Sautner, and Starks (2015) and Gillan and Starks (2007), among others. Last, following successful engagements, we expect profitability to improve, stock prices to react positively, and stock volatility to decrease, provided ESG practices maximize intertemporal profits and/or act as an effective channel of delegated philanthropy, as articulated by Benabou and Tirole (2010).

In general, we expect the role of reputational concerns and investor collaboration to be more pronounced for ES engagements, in comparison to CG engagements. This is because, ES engagements are prone to be more costly and more challenging to implement when presented to reluctant management, and need a longer period to realize their benefits. Throughout the paper, we test our hypotheses on the whole sample first and then provide further evidence by comparing ES engagements with CG engagements.

2. Data

Our data provider uses its influence as a major shareholder to promote the adoption of sound ESG practices. We believe the detailed electronic file of the firm's engagements is the most complete point-in-time dataset that is currently available for research of this type. We use only the data for U.S.-listed public companies, and our sample covers 613 companies from 1999–2009. Focusing on the U.S. sample offers two advantages: the availability of comprehensive and reliable financial, ownership, and governance datasets (see Appendix B), as well as the opportunity to compare our findings with those in the relevant activism and CSR literature that typically investigates U.S. firms. The total market capitalization of our sample firms is about 26% of the aggregate stock market capitalization reported by Center for Research in Security Prices (CRSP) during our sample period.

2.1 Engagement data

The data used in this paper include detailed information about the different engagement actions taken by the asset manager. Target companies are often identified by using ESG screening metrics and are chosen from the asset manager's current and prospective holdings. In addition, the asset manager also engages with companies beyond its investment portfolio and on behalf of at least twenty consulting clients. Engagements with target companies involve two types of actions: Raising Awareness and Request for Change. When the data provider records an engagement as Raising Awareness, it is aiming to inform and warn the target companies about certain ESG issues. In contrast, a Request for Change is usually a more stringent step, in which the asset manager asks for specific changes in the target company to address its unsatisfactory ESG practice. Accompanying the engagement data is a record of the improvements that the target company achieves in its ESG practices (if there are any), which are recorded as Milestones. On average, milestones are achieved nearly one-and-a-half years after the initial engagement. The original engagement dataset includes 2,482 unique engagement sequences, consisting of 2,462 Raising Awareness, 2,149 Request for Change, and 405 Milestones over 1999–2009. After requiring the target firm to have minimum company-level identifier data from Compustat and removing engagements on sector specific issues unrelated to ESG topics (115 Raising Awareness, 73 Request for Change, and 23 Milestones), our sample consists of 2,152 unique engagement sequences, of which 382 are successful.

In Appendix A, we present three illustrations of the engagements in our dataset. The first is a sequence of interactions with a well-known technology firm on environmental issues. The target was engaged three times before a milestone was recorded. A search on Factiva reveals that the initial engagement was triggered by a series of public events, such as prior demands by Greenpeace, a sustainability group, that the target be more environmentally friendly. After a take-back and recycling plan was announced and approved by shareholders, the outcome was recorded as a milestone. The second and third examples deal with social and governance issues. Unlike the first example, Factiva did not carry any news articles discussing these issues around the engagement dates, and we conclude that these engagements were unlikely to have been initiated by public events. Communication is probably through private channels, but we do not expect such engagements to be less effective than those triggered by public events, especially since Becht et al. (2009) show that shareholder activism can successfully and effectively be undertaken through private communications.

As mentioned above, many engagements are triggered by public events. To get a better idea of the frequency of these cases, we obtain information on public news coverage of our target firms up to seven calendar days prior to the engagement dates from the Capital IQ Key Development database. We find that 46.6% of ESG engagements in our sample are preceded by public news, some of which relates to the engagement in question.⁶ Milestones are public events when they coincide with shareholder meetings, at which requested changes are approved. We obtain information on shareholder meeting dates from the ISS, Capital IQ Key Development, and SEC's Edgar databases. We find that 32.5% of milestones in our sample occurred within two weeks of the date of target companies' shareholder meetings. Finally, the data confirm that the process for recording engagements and milestones is not based on ex post criteria: there are no indications that entries are backdated after observing the target firms' stock price movements.

2.2 Firm-level data

We obtain our data for firm characteristics from several sources. We download corporate accounting data from Compustat North America Fundamentals Annual, stock return data from Compustat North America Security Monthly, abnormal return data from the CRSP monthly stock file, analyst coverage data from I/B/E/S, institutional ownership data from Thomson Reuters 13F, corporate governance measures from RiskMetrics, legal lawsuits data from AuditAnalytics, and data to calculate liquidity measures from the CRSP daily stock file. Data from different sources are merged together using company identifiers, such as CUSIP, Gvkey, Permno, CIK, and firm name. Definitions and descriptions of each variable and of the data sources are provided in Appendix B.

3. Active Ownership Themes

Based on the stated objectives, engagements are divided by our data provider into nine themes belonging to three major areas: governance, environmental, and social. Table 1 lists the detailed description of different issues within each theme and reports the number of engagement sequences. An engagement sequence is defined as a series of interactions, including Raising Awareness (RA) or Request for Change (RC) or both, dealing with the same issue. The most commonly engaged theme is corporate governance, followed by labor standards, environmental management, and business ethics.

Table 1

Breakdown of engagement areas, themes, and issues

*Our data provider lists the “Business ethics” and “Sustainability management & reporting” themes under the broader area of “Governance”.

This table enumerates the different engagement areas, themes, and issues in the sample. Each area contains themes, and each theme comprises several issues. For example, the governance area comprises a corporate governance theme (labeled CG) plus business ethics and sustainability management and reporting themes. We distinguish between the CG theme and all non-CG themes. The non-CG themes cover environmental and social aspects and are labeled ES. The final column lists the number of engagement sequences under each theme.

In our analyses, we first examine the entire sample of engagements, including corporate governance, as well as environmental and social themes. This is for several reasons. First, ESG issues are often dealt with together by responsible investors and ESG activists. Therefore, the approach of including the corporate governance theme in our main sample is consistent with current practice. Second, corporate governance issues are often consolidated with social and environmental issues, the latter being a manifestation of the former. For example, an environmental setback could be a reflection of poor corporate governance. Third, issues related to CSR disclosure are also classified by our data provider under the corporate governance theme (see example A.3 at Appendix A).

There is an overlap, albeit partially, of governance engagements with the existing literature on shareholder activism. Therefore, throughout the paper, we emphasize the lesser known environmental and socially themed engagements and use the corporate governance themed engagements mostly in a comparative framework. As well as themes under the environmental and social areas, the ES classification in our study incorporates business ethics and the sustainability management and reporting themes (which are originally classified by our data provider under the governance area). This approach follows the proxy voting guidelines developed by ISS Social Advisory Service for socially responsible investment clients. ISS (2012) categorizes board of directors, ratification of auditors, takeover defense/shareholder rights, capital structure, executive and director compensation, shareholder rights, and mergers and corporate restructuring as governance proposals and others as environmental and social proposals.

3.1 Analysis by year and industry

The upper panel of Table 2 reports the number of engagement sequences by calendar year, classified by the date of the initial engagement (the first one in a sequence), and the success rate, defined as the number of successful engagement sequences divided by total number of engagement sequences. An engagement sequence is defined as “successful” if a milestone is achieved at the end of the sequence and recorded in the database. There are relatively few observations in the early years due to narrow coverage within the database. The apparent decline in the success rate from 2007 onward, especially for ES themes, is probably due to the fact that when our data stop at mid-2009, some engagements are still works in progress and milestones have not yet been achieved. Classifying not-yet-successful engagements as unsuccessful biases us against finding differences between successful and unsuccessful engagements in our analyses.

Table 2

Summary of engagements by year and industry

	Whole sample			CG subsample		ES subsample
	Num. of sequences	% Sample	% Success	Num. of sequences	% Success	Num. of sequences	% Success
Engagement year
1999	8	0.4%	25.0%	–	–	8	25.0%
2000	27	1.3%	37.0%	7	100.0%	20	15.0%
2001	77	3.6%	29.9%	11	72.7%	66	22.7%
2002	103	4.8%	47.6%	32	46.9%	71	47.9%
2003	158	7.3%	34.2%	50	52.0%	108	25.9%
2004	419	19.5%	27.0%	332	25.9%	87	31.0%
2005	207	9.6%	25.1%	71	32.4%	136	21.3%
2006	200	9.3%	16.0%	95	21.1%	105	11.4%
2007	207	9.6%	4.3%	76	10.5%	131	0.8%
2008	434	20.2%	7.1%	143	14.7%	291	3.4%
2009	312	14.5%	2.2%	83	4.8%	229	1.3%
Total/Average	2,152	100.0%	17.8%	900	24.2%	1,252	13.1%
Industry division
Agriculture	10	0.5%	10.0%	5	20.0%	5	0.0%
Mining	103	4.8%	7.8%	32	3.1%	71	9.9%
Construction	12	0.6%	25.0%	6	50.0%	6	0.0%
Manufacturing	963	44.7%	19.3%	378	27.0%	585	14.4%
Transportation	169	7.9%	17.8%	94	22.3%	75	12.0%
Wholesale trade	30	1.4%	13.3%	13	23.1%	17	5.9%
Retail trade	203	9.4%	19.2%	83	28.9%	120	12.5%
Financial	437	20.3%	15.6%	185	21.6%	252	11.1%
Services	166	7.7%	20.5%	83	24.1%	83	16.9%
Public admin.	28	1.3%	32.1%	9	33.3%	19	31.6%
Missing	31	1.4%	–	12	–	19	–
Total/Average	2,152	100.0%	17.8%	900	24.2%	1,252	13.1%

This table shows how engagements are distributed over time and across industry sectors. The upper panel reports the number of engagement sequences and the success rates of engagements by calendar year for the whole sample, for the CG subsample, and for the ES subsample. Engagement sequences are classified into calendar years according to the initial engagement date. The lower panel reports the number of engagement sequences categorized by the industry of the target firm. Agriculture denotes Agriculture, Forestry, Fishing; Transportation denotes Transportation, Communications, Electric, Gas, and Sanitary Services; Financial denotes Finance, Insurance, Real Estate; Missing denotes Missing Industry Identification.

There was a transitory surge in CG engagements in 2004 (increasing from fifty observations in 2003 to 332 in 2004 and dropping back to 71 in 2005). This is mainly driven by engagements on issues of voluntary employee stock option (ESO) expensing, which was a heavily debated accounting topic in the early 2000s. The Financial Accounting Standards Board (FASB) released the final version of FAS no. 123R, which requires all U.S. companies to expense ESOs, effective from June 15, 2005 onward. In 2004, while the final FASB rule was still under debate, the asset manager sent a letter to the CEOs of a large group of target firms, asking them to voluntarily expense ESOs. Ferri and Sandino (2009) provide a detailed discussion about shareholder proposals on voluntary expensing ESO issues during that period. To make sure that our results are not sensitive to this specific event, we repeat our analysis throughout the paper after excluding engagements on this topic and find qualitatively similar results (data unreported). The distribution of ES engagements is relatively stable across years.

The lower panel of Table 2 reports the number of engagement sequences by industry, based on single-digit SIC codes of the target companies. Engaged companies are from all the major industries, with observations concentrated in manufacturing and finance.

3.2 Analysis by area and theme

Table 3 reports the summary of engagement sequences by different engagement areas and themes (e.g., number of sequences, percentages of successful engagements).⁷ This table also reports comparison statistics between CG and ES subsamples. Columns (1) and (2) report the number of successful engagements and the percentage success rate under each theme. Column (6) reports the number of unsuccessful engagements. As can be seen in Column (2), engagements on corporate governance, environmental management, and labor standards themes are most likely to be successful, with success rates of 24.2%, 17.6%, and 16.9%, respectively. Engagements on public health, sustainability management & reporting, and human rights themes are least likely to be resolved, with success rates below 10%.

Table 3

Summary of engagements by area and theme

	Successful						Unsuccessful
Engagement areas and Themes	Num. of sequences	% Success	Num. of RA & RC	Horizon (days)		Num. of collab'tors	Num. of sequences	Num. of RA & RC	Num. of collab'tors
	(1)	(2)	(3)	(4i)	(4ii)	(5)	(6)	(7)	(8)
1. Governance
Corporate gov'ce (CG)	218	24.2%	2.2	525	[369]	1.0	682	1.6	0.3
2. Environment
Climate change	16	10.3%	3.9	521	[524]	4.7	140	1.9	1.0
Ecosystem services	8	10.4%	3.0	512	[123]	0.9	69	2.1	0.6
Environmental mgt	39	17.6%	3.2	386	[246]	2.8	182	1.8	0.6
3. Social
Public health	2	6.5%	3.5	622	[622]	3.5	29	1.6	0.3
Human rights	18	9.9%	4.7	591	[472]	3.9	164	3.1	1.3
Labor standards	38	16.9%	2.8	410	[165]	3.6	187	1.6	0.5
Business ethics	29	13.7%	4.8	647	[539]	4.0	182	2.2	0.9
Sustainability mgt & rptg	14	9.4%	3.8	284	[77]	3.3	135	1.8	0.5
Total/Average (ES)	164	13.1%	3.7	474	[265]	3.5	1,088	2.1	0.8
t-stat [z-stat] (ES-CG)		−6.47	5.06	−1.04	[−1.93]	8.68		5.05	7.35
Total/Average	382	17.75%	2.90	503	[349]	2.03	1,770	1.90	0.58

This table decomposes engagement sequences by area and theme and includes an analysis of success rates for ES, as compared to CG, engagements. Engagement sequences comprise a series of raising awareness (RA) plus requests for change (RC) engagements, dealing with the same issue. Columns (1)–(5) report on the engagement sequences that were rewarded with success; Columns (6)–(8) report on the sequences that did not experience success by the end of our sample period. For successful sequences, Columns (1) and (2) report the number in each category and the percentage among all sequences that were successful; for unsuccessful sequences, Column (6) reports the number in each category. Columns (3) and (7) report the number of RA and RC interactions with target companies. Column (4) presents the average [median] number of days between the initial engagement and the achievement of the milestone. Columns (5) and (8) report the average number of collaborating investors within each sequence. The prepenultimate row reports overall statistics for ES engagements, which may be compared with CG engagements in the first row. The penultimate row provides chi-square tests of the difference between ES and CG engagements. The last row provides summary statistics for the entire sample, comprising both CG and ES engagements.

The ES subsample as a whole has an average success rate of 13.1%, which is significantly lower than the 24.2% success rate of the CG subsample (⁠$t$-statistic $=-6.47$⁠) and much below that of hedge funds (40.6% in Brav et al. 2008 and 60% in Klein and Zur 2009) or other entrepreneurial activists (65% in both Klein and Zur 2009 and Becht et al. 2009). We posit two explanations for this lower success rate. First, there is difficulty in convincing management or other shareholders to accept projects that are costly but potentially beneficial to other stakeholders, such as employees, suppliers, local community, and/or consumers. Note, however, that this rate is consistent with the approval rate of below 20% for shareholder proposals on environmental and social issues during proxy seasons before 2011, as documented by Allen et al. (2011). Second is the lesser influence on the target firm of ESG engagement strategies that are less aggressive than those of hedge funds activists and other entrepreneurial activists. However, the success rate of our CG subsample (24.2%) is comparable to traditional shareholder activism via shareholder proposals. For example, Gillan and Starks (2000) document an average success rate of 23% for shareholder proposals on corporate governance issues. Using a more recent sample, Aggarwal, Erel, and Starks (2014) report an average success rate of 29% for shareholder proposals.

Columns (3) and (7) report the average number of Raising Awareness and Request for Change for successful and unsuccessful engagement sequences, respectively. We find the engagements on human rights and business ethics themes to have the largest number of engagements per sequence, despite their low success rates. This suggests that it might be particularly difficult to persuade target companies to resolve issues in these areas. Columns (4i) and (4ii) report the mean and median number of days between the initial engagement date and the milestone date for successful engagement sequences under each theme, respectively. For the whole sample, the mean (median) horizon is 503 (349) days, an elapsed time that is consistent with the shareholder activism literature: Becht et al. (2009) find that the median duration of investment is 469 days for collaborative engagements and 1,284 days for confrontational ones, whereas Brav et al. (2008) find that the median holding period of their hedge fund sample is 369 days.

Columns (5) and (8) report the average number of activist partners and/or initiatives that the asset manager collaborates with for successful and unsuccessful engagement sequences, respectively. We hand-collect this information by reading and coding the original engagement records that were provided by our data contributor (see Section 3.3 for more details). We note that on average successful engagements have more collaborators than unsuccessful ones (2.0 vs. 0.6). We also observe that successful engagements in ES themes have significantly more collaborators than those in corporate governance theme (3.5 vs. 1.0). This finding provides evidence that our asset manager adopts a set of engagement strategies that differ between ES and CG themes.

3.3 Collaboration and other tactics

How are the engagements carried through? To address this question, we analyze the engagement tactics in this section.

As noted in Section 3.2, collaboration by the asset manager with other activist investors and/or initiatives plays an important part in the asset manager's engagement strategies. There is limited empirical evidence on collaboration among investors in the shareholder activism literature, and we take a first step to further investigate the identity and number of collaborators in each engagement. We classify collaborations into two major categories based on the identity of those involved: hard collaborations and soft collaborations. Hard collaborations include the partnership of the asset manager with activist investors, such as SRI funds, pension funds, asset managers, financial institutions, religious funds, foundations, endowments, aid agencies, individuals, and union funds. Soft collaborations refer to asset managers who benefit from the ESG principles and initiatives established by investment bodies, nonprofit organizations, and/or the industry in which the firm operates.

The asset manager may prefer cooperating with single, multiple, or no collaborators during each of its engagement sequence. The upper panel of Table 4 reports the number of engagement sequences and success rate by collaborator types. Note that an engagement sequence may show up multiple times in this table if the asset manager cooperates with multiple types of collaborators. In the univariate analysis using a chi-square test, we observe that cooperation with hard collaborators, compared with soft collaborators (48.5% vs. 39.1%, with a $p$-value of 0.00), leads to a higher success rate, as the former are activist investors, whereas the latter are passive principals. We also observe that this difference exists for ES engagements, but not for CG engagements.

Table 4

Collaboration and success rates

		Whole sample		CG subsample		ES subsample
Type of engagement	Detail on engagement	Num. of sequences	% Success	Num. of sequences	% Success	Num. of sequences	% Success
Collaborator types and success rates
Hard collab'tors	SRI Fund	415	51.6%	100	57.0%	315	49.8%
	Pension Fund	137	32.1%	65	33.8%	72	30.6%
	Asset Manager	62	48.4%	28	67.9%	34	32.4%
	Financial Institution	27	63.0%	5	60.0%	22	63.6%
	Religious Fund	23	47.8%	6	83.3%	17	35.3%
	Foundation, Endowment	17	35.3%	11	27.3%	6	50.0%
	Aid Agency, CSR Activist, Individual	16	93.8%	1	100.0%	15	93.3%
	Union Fund	15	53.3%	9	22.2%	6	100.0%
	All hard collaborators	712	48.5%	225	49.8%	487	47.8%
Soft collab'tors	Code, Principle, Initiative, Standard, Bill	540	31.9%	95	37.9%	445	30.6%
	Forum, Network, Coalition	231	36.4%	32	59.4%	199	32.7%
	Peer, Media	92	85.9%	23	87.0%	69	85.5%
	Institute, University, ThinkTank, Center	66	30.3%	3	100.0%	63	27.0%
	Index, Research, Proxy Advising, Audit	62	45.2%	28	53.6%	34	38.2%
	Special Interest	58	25.9%	–	–	58	25.9%
	Government, Regulatory	41	68.3%	4	75.0%	37	67.6%
	Other	6	83.3%	2	100.0%	4	75.0%
	All soft collaborators	1,090	39.1%	185	51.9%	905	36.5%
Hard vs. soft	Chi square [p-value]	15.45	[0.00]	0.18	[0.67]	17.02	[0.00]
Engagement features and success rates
Aggressiveness	Non-aggressive engagements	921	13.7%	462	19.9%	459	7.4%
	Aggressive engagements	1,231	20.8%	438	28.8%	793	16.4%
Intensity	One-time engagements	1,320	11.7%	567	17.3%	753	7.6%
	Repeated engagements	832	27.3%	333	36.0%	499	21.4%
Experience	Without success history	1,304	13.5%	610	20.2%	694	7.6%
	With success history	848	24.3%	290	32.8%	558	19.9%
Focus	Multiple engagements	917	10.7%	207	17.9%	710	8.6%
	Focused engagements	1,235	23.0%	693	26.1%	542	19.0%
Collaboration	Without collaborator	1,485	11.0%	739	19.2%	746	2.8%
	With collaborator(s)	667	32.8%	161	47.2%	506	28.3%
	Without soft collaborator	1,570	12.6%	783	20.8%	787	4.3%
	With soft collaborator(s)	582	31.8%	117	47.0%	465	28.0%
	Without hard collaborator	1,865	13.2%	830	21.7%	1,035	6.4%
	With hard collaborator(s)	287	47.4%	70	54.3%	217	45.2%

The upper panel of this table lists the collaborator types and the success rates of engagements for the whole sample and the CG and ES subsamples. The lower panel provides corresponding analysis for other engagement features. Chi-square tests compare hard and soft collaboration for the whole sample and the CG and ES subsamples. Variables are defined in Appendix B.

In the lower panel of Table 4, we analyze the asset manager's engagement tactics by comparing the success rate across subsamples with five different engagement features: aggressiveness, intensity, successful experience, focus, and collaboration. First, we posit Request for Change to be more aggressive than Raising Awareness and define engagement sequences as aggressive if they contain at least one Request for Change. Second, we consider repeated engagements in a sequence to be more intensive than one-time engagements. Third, if the asset manager has a history of successful engagements with the same target firm, we interpret it as a positive factor. Fourth, we define an engagement sequence as focused if, at the initial engagement, the asset manager exclusively addresses issue(s) under the same theme. Fifth, we consider an engagement sequence as collaborative if the asset manager cooperates with at least one hard collaborator or soft collaborator at any point during the whole engagement sequence.

In our univariate analysis using a chi-square test, we find aggressive, intensive, experienced, focused, and collaborative (both hard and soft) engagement sequences are associated with higher success rates (with two-tailed $p$-values lower than 0.01 for all cases and hence are untabulated for brevity). We present multivariate probit analysis of engagement tactics as the determinants of success in Section 5.

4. Characteristics of Target Companies Prior to Engagement

Which types of companies are targeted for active ownership? To address this question, we examine the characteristics of the target firms in the year before the initial engagement and compare them with a matched sample of firms. To construct the matched sample, we first create a matching pool using all companies from Compustat North America and follow the Brav et al. (2008) matching rule. We remove all the target companies from the pool and require both the target and the matching firms to have data on industry, firm size, and the market-to-book ratio. The matched firms for each target company are assigned from the same year, industry (3-digit SIC), and 10 × 10 size- and market-to-book-sorted portfolios. If the above rule does not yield any match, we relax the industry to two-digit SIC and the size-/market-to-book to 5 × 5-sorted portfolios. Each target firm is therefore matched to a portfolio of control firms. We further average firm characteristics across the matched portfolio for each target firm to construct a (pseudo) control firm corresponding one-to-one to the target firm. In tests of robustness (unreported), we adopt another matching rule, where we relax the industry to twelve Fama-French categories and directly use 5 × 5 portfolios sorted by size and the market-to-book ratio. Then, among all the matched firms, we keep only the one with size closest to the target company. Using this alternative rule with fewer restrictions, we are able to find matches for more engagement sequences and the size difference between the target company and the matched firm is smaller, but our test results remain similar.

4.1 Univariate analysis of targeting

Summary statistics for the target firms' characteristics, as of one year before the initial engagement, are provided in Columns (1)–(6) of Table 5. The detailed variable definitions and data sources are included in Appendix B. Columns (7)–(12) report the difference between target companies and matched firms averaged across the target sample. As in Brav et al. (2008), the difference between a sample firm $i$ and its matched firms is calculated as follows:

$$Dif{f}_i=X_i-\frac{1}{m}\sum _{j=1}^m{X}_j,$$

where $X_i$ is defined as a characteristic variable and firms $j=1,\dots ,m$ are from the matching group. To test whether the differences are statistically different from zero, we report the $t$-statistics in Columns (8), (10), and (12). Wilcoxon signed-rank statistics, which test the median difference between two samples, yield similar results (data unreported). The number of observations as reported in Columns (2), (4), and (6) varies due to the availability of data to calculate companies' characteristics.

Table 5

Characteristics of target companies

	Summary Statistics						Relative to Matched Firms
	All		CG		ES		All		CG		ES
Firm characteristics	Mean	Obs	Mean	Obs	Mean	Obs	Mean	t-stat	Mean	t-stat	Mean	t-stat
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Firm size	53.54	1,746	39.78	749	63.88	997	49.07	27.76	35.54	15.35	59.24	23.56
Market-to-book	4.05	1,746	3.84	749	4.20	997	−0.29	−2.70	−0.12	−0.87	−0.42	−2.67
Tobin's q	2.98	1,739	2.80	745	3.11	994	−0.14	−2.21	−0.10	−1.24	−0.17	−1.83
Firm age	33.07	1,746	32.19	749	33.72	997	12.14	27.42	11.87	18.19	12.34	20.54
Sales growth	0.12	1,742	0.11	747	0.13	995	−0.15	−10.52	−0.12	−6.65	−0.17	−8.22
Stock return	0.11	1,692	0.14	724	0.08	968	−0.14	−9.76	−0.14	−6.51	−0.13	−7.26
Stock return volatility	0.09	1,711	0.09	734	0.09	977	−0.02	−16.25	−0.02	−8.64	−0.03	−13.94
Return on assets	0.15	1,596	0.14	704	0.16	892	0.00	−0.24	0.00	−0.98	0.00	0.47
Asset turnover	0.85	1,746	0.81	749	0.88	997	−0.09	−5.76	−0.08	−3.65	−0.10	−4.47
Sales over employees	0.68	1,717	0.55	736	0.78	981	−0.47	−7.16	−0.25	−3.66	−0.64	−6.20
Cash flow	0.10	1,596	0.09	704	0.11	892	0.00	0.63	0.00	−0.28	0.00	1.02
Leverage	0.37	1,739	0.37	745	0.37	994	0.04	6.76	0.05	5.30	0.03	4.37
Cash holding	0.09	1,708	0.09	738	0.08	970	−0.02	−8.21	−0.01	−3.70	−0.02	−7.79
Dividend yield	0.02	1,746	0.02	749	0.02	997	0.00	3.19	0.00	1.51	0.00	2.89
Dividend payout	0.32	1,746	0.29	749	0.34	997	0.06	2.52	0.10	3.06	0.03	0.84
R&D expenditure	0.03	1,746	0.03	749	0.03	997	0.00	−2.53	0.00	−1.32	0.00	−2.18
Capital expenditure	0.05	1,700	0.04	727	0.05	973	0.00	−3.65	−0.01	−4.92	0.00	−1.18
Advertising expenditure	0.01	1,746	0.01	749	0.01	997	0.00	4.06	0.00	2.69	0.00	3.05
Industry Herfindahl index	0.34	1,657	0.32	707	0.35	950	0.02	3.63	0.01	1.56	0.02	3.38
Industry advertising intensity	0.45	1,660	0.42	712	0.48	948	0.11	4.34	0.06	1.60	0.15	4.33
Shareholding of asset manager	0.06	1,746	0.06	749	0.06	997	0.03	9.21	0.04	8.18	0.02	4.79
Shareholding of pension activists	2.17	1,746	2.24	749	2.13	997	0.28	8.34	0.39	7.32	0.21	4.66
Shareholding of SRI funds	0.21	1,746	0.23	749	0.20	997	0.11	11.02	0.12	7.21	0.10	8.38
Amihud illiquidity	0.01	1,702	0.02	725	0.01	977	−0.02	−34.92	−0.02	−22.44	−0.02	−26.75
Governance index	9.15	1,199	9.36	574	8.95	625	0.22	2.36	0.41	3.08	0.04	0.34
Entrenchment index	2.00	1,432	2.13	639	1.89	793	−0.34	−7.98	−0.19	−2.99	−0.47	−7.98
Number of analysts	16.63	1,746	16.26	749	16.91	997	6.89	35.73	6.62	22.00	7.09	28.28
Number of pension activists	11.09	1,746	11.34	749	10.91	997	2.29	17.90	2.73	13.86	1.97	11.70
Tangibility	0.24	1,582	0.23	703	0.26	879	0.00	0.04	−0.01	−1.24	0.01	1.12
H&P product similarity score	0.06	1,674	0.07	719	0.05	955	−0.01	−8.24	−0.01	−3.53	−0.02	−8.03
Market share (segment)	0.11	1,545	0.10	670	0.12	875	0.06	14.21	0.05	8.61	0.07	11.36

This table reports the characteristics of target companies and comparisons with a set of matched companies. Firm characteristics are measured at the year before the initial engagement date. Only the initial engagement is kept for each sequence. The first two columns report the mean and number of observations in our sample, and columns (3) and (4) and (5) and (6) provide these statistics for the CG and ES subsamples. Columns (7)–(12) report the average difference between the sample firms and the industry/size/market-to-book matched firms and $t$-statistics for the average differences. Variables are defined in Appendix B. Shareholdings of asset manager, pension activists, and SRI funds are multiplied by 100, for ease of readability. All variables are winsorized at the 1st and 99th percentile levels.

4.1.1 Size and maturity

Unlike activist hedge funds or other entrepreneurial activists that need considerable voting power for intervention, and hence target medium- or small-sized companies in which they can acquire a sizeable ownership block (Brav et al. 2008; Klein and Zur 2009), our data provider engages with large and mature firms; they have higher firm size and firm age and lower sales growth compared with the matched group. This suggests that the asset manager aims to achieve its goals by relying more on the economies of scale and benefiting from the accompanying reputational concerns faced by large-sized target companies. The focus on large firms is consistent with that documented in traditional shareholder activism (e.g., Smith 1996; Karpoff, Malatesta, and Walkling 1996). Our data provider's relatively less confrontational strategy may explain the lower success rates reported in Table 3, in comparison to the success rates of hedge fund activists. Note that the voting power is exploited as a mechanism to publicize a position in support of, or in opposition to, the firm's decisions. The market-to-book ratio and Tobin's$q$ are only modestly different from the control firms', especially for sample medians (data not reported), suggesting effective matching by the market-to-book ratio. Because of their large size, our target firms also have higher liquidity (lower Amihud illiquidity), a higher number of analysts covering the firm, and higher market share. Higher liquidity might make the (threat of) exit more credible (see Edmans, Fang, and Zur 2013).

4.1.2 Institutional ownership

Target firms appear to attract more socially conscious investors, characterized with higher number of pension activists holding and higher shareholding of the asset manager, pension activists, and SRI funds, although these institutions' shareholding percentages are relatively low (0.06%, 2.17%, and 0.21%, respectively), due to the large firm size. As we demonstrate in the rest of the paper, despite this relatively low ownership, engagements are effective in influencing the companies' business. Thus, this finding complements the existing literature on the impact of large shareholders, with at least 5% ownership, on large firms (e.g., Becker, Cronqvist, and Fahlenbrach 2011; Clifford and Lindsey 2013). The finding that target companies have a larger shareholding from the asset manager and SRI funds, which are potentially collaborators in the engagements, is in line with the argument that benefits related to improvement accrue in proportion to the size of the shareholding, and ownership is related to voting power.

4.1.3 Performance

In contrast to hedge funds targeting more profitable firms (Brav et al. 2008; Klein and Zur 2009), our active owner targets relatively less profitable ones. Stock return is the buy-and-hold return, including reinvested dividends, from the previous year, and it is significantly lower for target firms compared with that of control firms. The strategy of targeting poorly performing firms is consistent with that of traditional shareholder activism (Smith 1996; Karpoff, Malatesta, and Walkling 1996) and other entrepreneurial activism (Klein and Zur 2009; Becht et al. 2009). In addition, targets are less efficient firms, with a lower asset-turnover ratio and a lower sales-over-employees ratio.

4.1.4 Discretionary spending

Whereas hedge funds target firms paying lower dividends (Brav et al. 2008), our sample emphasizes those paying more. They have a higher dividend yield and a higher dividend-payout ratio. In addition, engaged companies have lower research and development (R&D) expenditure and have a lower capital expenditure. The latter finding suggests that the asset manager targets firms with more scope to incur additional expenses that may be necessary for ESG improvements. Klein and Zur (2009) find no difference in discretionary spending between their hedge fund targets and the control group in their univariate analysis.

4.1.5 Capital structure

Target firms have higher leverage and lower cash holding, similar to those targeted by active hedge funds in Brav et al. (2008). In contrast, although Klein and Zur (2009) do not find cash holding to be different between their hedge fund targets and the control group, they find that relative to targets of other entrepreneurial activists, hedge fund targets have higher cash holding.

4.1.6 Corporate governance

We measure target firms' corporate governance mechanisms using the Gompers, Ishii, and Metrick (2003)^,governance index and the Bebchuk, Cohen, and Ferrell (2009)^,entrenchment index. These two indexes measure the extent to which management is entrenched (see Bebchuk, Cohen, and Wang 2013).⁸ We find mixed results with univariate analysis, as indicated by the higher governance index and lower entrenchment index of target firms related to benchmark firms. The results are more uniform in favor of lower governance of target firms in the multivariate analyses below, and we defer further discussion to the next section.

4.1.7 Customer awareness and loyalty

Firms in our sample face high customer awareness and loyalty, characterized by higher advertising expenditure, higher industry advertising intensity, and higher product differentiation (lower H&P product similarity score). This suggests that the asset manager targets firms in consumer-facing industries with high reputational concerns. This is in line with Fisman, Heal, and Nair (2005) and Servaes and Tamayo (2013), who find that CSR is more prevalent and beneficial in advertising-intensive industries and firms, respectively. It is also consistent with Eccles, Ioannou, and Serafeim (2014), who note superior performance from ESG-focused firms in consumer-facing, brand-driven, and natural resource sectors. This finding resonates with the theoretical argument in Albuquerque, Durnev, and Koskinen (2014) that ESG activities reduce firm risk by establishing customer loyalty—the effect being stronger for firms producing differentiated goods for which substitution is hard.

In general, we observe similar patterns for the overall sample as for the CG and ES subsamples. We highlight the differences of target firm characteristics between the CG and ES subsamples, using multivariate analysis, in Section 4.2.

4.2 Multivariate analysis of targeting

Table 6 presents a probit regression model of targeting. In this multivariate model, we see in the left half of the table the marginal effect of each firm characteristic on the likelihood of being targeted for the whole sample (see columns titled “All”). The results are largely consistent with the previous table. In these models, we control for year fixed effects, and standard errors are clustered at the firm level. Consistent with the asset manager targeting large and mature firms and firms with poor performance, we observe that target firms have larger size, older age, lower sales growth, higher liquidity, more analyst following, larger market share in the industry, and lower return on assets, relative to the benchmark firms. Additionally, consistent with the asset manager targeting firms with high reputational concerns among customers, we observe that target firms have higher advertising expenditure and product differentiation. We also find that target firms have higher shareholdings from the asset manager, pension activists, and SRI funds, who represent socially conscious investors and potential collaborators.

Table 6

Probit analysis on targeting

Dependent variable:	Whole sample						CG and ES subsamples
1 if targeted, 0 otherwise*	All		All		All		CG		ES		ES – CG
Firm	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat
characteristics	Effect		Effect		Effect		Effect		Effect		Effect
Firm size	0.01	7.25	0.01	6.32	0.01	5.74	0.02	5.50	0.00	4.72	0.00	1.05
Tobin's q	−0.01	−0.91	0.00	−1.06	0.00	−0.10	0.01	0.71	0.00	−0.65	0.00	−0.03
Firm age	0.00	2.49	0.00	3.01	0.00	1.69	0.00	3.56	0.00	−0.19	0.00	−0.12
Sales growth	−0.06	−2.82	−0.07	−2.20	−0.06	−2.06	−0.26	−2.93	−0.02	−1.26	0.02	0.14
Stock return	−0.02	−1.08	−0.01	−0.59	0.00	−0.37	−0.05	−1.20	0.00	0.40	0.04	0.66
Return on assets	−0.25	−2.68	−0.07	−0.99	−0.13	−1.79	−0.35	−1.58	−0.04	−1.14	0.33	0.96
Sales over employees	−0.02	−1.32	−0.01	−1.12	−0.01	−0.78	−0.01	−0.18	0.00	−0.80	0.07	2.69
Cash holding	0.17	1.72	−0.01	−0.17	−0.02	−0.29	0.14	0.74	−0.04	−1.11	−0.32	−1.35
Leverage	−0.01	−0.15	0.00	−0.16	0.04	1.60	0.13	1.66	0.01	0.87	0.03	0.36
Dividend yield	0.27	0.52	−0.21	−0.64	−0.04	−0.11	−1.00	−1.02	0.11	0.64	0.96	0.58
Capital expenditure	0.06	0.38	−0.13	−1.26	−0.12	−1.13	−0.48	−1.47	−0.04	−0.72	0.55	1.05
R&D expenditure	−0.29	−1.42	−0.10	−0.70	0.10	0.71	0.13	0.30	0.03	0.48	0.26	0.59
Advertising expenditure	0.56	2.02	0.43	2.09	0.36	1.78	0.81	1.49	0.19	1.93	1.36	1.93
Shareholding of asset manager	36.09	5.68	13.98	2.86	17.26	3.49	75.81	4.90	1.84	0.57	−48.94	−2.57
Shareholding of pension activists	−0.54	−0.79	0.70	1.68	0.98	2.21	4.25	3.23	0.26	1.12	−2.36	−1.48
Shareholding of SRI funds	6.49	4.16	4.25	4.09	3.95	3.50	8.48	2.62	1.72	2.67	−1.51	−0.52
Amihud illiquidity	−0.56	−1.70	−0.81	−2.39	−0.70	−2.08	−1.59	−1.46	−0.38	−1.92	−2.98	−1.58
Number of analysts	0.01	3.24	0.00	2.91	0.00	3.98	0.01	3.92	0.00	2.85	0.00	−0.42
Entrenchment index			0.01	2.33	0.01	2.01	0.02	1.74	0.00	1.61	−0.01	−0.93
H&P product similarity score					−0.55	−3.79	−1.11	−2.36	−0.26	−4.07	−0.77	−1.83
Market share (segment)					0.07	1.77	0.10	0.96	0.03	1.77	0.26	2.25
Year fixed effect	Yes		Yes		Yes		Yes		Yes		Yes
Obs	2,908		2,558		2,356		1,048		1,308		1,178
Pseudo ${\text{R}}^2$	0.48		0.54		0.56		0.51		0.63		0.18

* For final regression, 1 if targeted in ES, 0 if targeted in CG.

This table reports the marginal effects of characteristics of being targeted. The dependent variable is a dummy variable equal to one if the company is targeted during the following year, and zero for a control firm-year. Firm characteristics are measured at the year before the initial engagement date. Only the initial engagement is kept for each sequence. Engagement year fixed effects are included in all regressions. Standard errors are clustered at the firm level. Independent variables are defined in Appendix B. All variables are winsorized at the 1st and 99th percentile levels. Coefficients in bold are statistically significant at the 10% level.

Finally, we find that engaged firms have weaker corporate governance, evidenced by the positive coefficient on the entrenchment index. We find qualitatively similar results by replacing the entrenchment index with the governance index (data unreported). This is consistent with the evidence in Table 2 that corporate governance is the theme that is most frequently associated with action. The fact that targeted firms have a weaker governance structure suggests that firms with scope for improvement are more likely to be engaged. Because we are controlling for other firm characteristics, this result is more meaningful than the univariate analysis, in which the entrenchment index has the opposite sign.

In the right half of Table 6, we partition the sample into CG and ES engagements and conduct the probit regressions separately for these two subsamples (see the columns titled “CG” and “ES”). The coefficients on return on assets become insignificant for both subsamples, potentially due to the smaller sample sizes. The coefficients on firm age and sales growth stay significant only for the CG subsample. The coefficients on advertising expenditure, liquidity, and market share are insignificant for the subsample with a CG theme, but are significant for the subsample with ES themes. These findings suggest that the asset manager adopts different targeting strategies for CG and ES engagements, with the former focusing more on mature firms with poor corporate governance and the latter focusing more on large firms with reputational concerns among customers.

For the subsample with ES themes, we also find the coefficient on the shareholding of asset manager and pension activists to be insignificant. That is, the asset manager does not necessarily rely on its shareholding as a determinant for engaging on environmental and social issues. Particularly for these issues, it is not uncommon for the asset manager to draw attention to the reputational concerns of the target company and/or to collaborate with other investors and stakeholders in order to achieve its goals (see Tables 4 and 7). For example, the asset manager sometimes engages with investee companies by sending representatives alongside other sustainable investment analysts from KLD for a meeting with managers from the company in question, by participating in a multistakeholder working group, or by supporting shareholder proposals sent by other stockholders. In the engagement example described in Section 2.1, the asset manager was acting in response to the demand by Greenpeace, a nongovernmental environmental organization. This finding reflects the relatively nonconfrontational strategy adopted by the asset manager.

Table 7

Probit analysis on success

Dependent variable:	Whole sample						CG and SEE subsamples
1 if successful, 0 otherwise*	All		All		All		CG		ES		ES – CG
Firm	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat	Marg.	t-stat
characteristics	Effect		Effect		Effect		Effect		Effect		Effect
Firm size	0.00	3.65	0.00	3.12	0.00	2.91	0.00	1.66	0.00	2.09	0.00	1.63
Tobin's q	0.00	−0.36	0.00	−0.38	0.00	−0.23	0.00	0.01	0.00	−0.79	−0.02	−0.93
Firm age	0.00	1.35	0.00	1.50	0.00	1.67	0.00	2.40	0.00	0.40	0.00	−0.84
Sales growth	0.03	0.67	0.08	1.40	0.09	1.44	0.11	0.86	0.06	1.20	0.15	0.70
Stock return	−0.04	−1.14	−0.06	−1.40	−0.06	−1.45	−0.05	−0.89	−0.04	−0.79	−0.04	−0.33
Return on assets	−0.36	−2.05	−0.53	−2.79	−0.58	−2.88	−0.52	−1.45	−0.50	−2.70	0.37	0.45
Sales over employees	−0.01	−0.84	−0.02	−1.36	−0.02	−1.25	−0.02	−0.42	−0.01	−0.55	0.10	1.32
Cash holding	0.29	1.96	0.41	2.47	0.35	2.25	0.42	1.63	0.32	2.28	0.20	0.34
Leverage	−0.08	−1.44	−0.10	−1.45	−0.11	−1.48	0.02	0.18	−0.18	−2.41	−0.21	−0.98
Dividend yield	−0.50	−0.61	−1.07	−1.11	−1.37	−1.23	−1.20	−0.75	−0.65	−0.59	−0.63	−0.18
Capital expenditure	−0.77	−2.49	−0.65	−1.83	−0.72	−2.02	−0.13	−0.18	−0.75	−2.44	−0.87	−0.66
R&D expenditure	−1.00	−3.73	−1.18	−3.69	−1.33	−3.28	−0.37	−0.57	−1.78	−4.68	−4.31	−3.45
Advertising expenditure	0.71	1.62	1.05	2.08	1.05	2.06	1.16	1.31	1.15	2.60	4.12	2.58
Shareholding of asset manager	16.93	1.27	13.94	0.97	8.86	0.62	1.55	0.08	6.31	0.39	−94.96	−1.78
Shareholding of pension activists	−0.32	−0.33	−0.65	−0.58	−0.07	−0.06	−0.15	−0.06	1.01	0.94	2.88	0.77
Shareholding of SRI funds	−2.11	−0.92	−1.14	−0.48	−0.78	−0.32	0.12	0.03	−6.32	−1.45	−38.20	−1.84
Amihud illiquidity	−3.67	−1.93	−6.74	−2.45	−7.18	−2.45	−11.45	−2.24	−3.35	−1.46	−5.27	−0.65
Number of analysts	0.01	4.44	0.01	3.85	0.01	3.54	0.01	2.39	0.01	3.38	0.01	1.42
Entrenchment index			0.01	1.06	0.01	1.06	0.02	0.85	0.00	0.17	−0.01	−0.25
H&P product similarity score					−0.33	−0.87	−1.51	−1.90	0.29	1.00	1.14	0.77
Market share (segment)					−0.22	−2.25	−0.07	−0.40	−0.24	−3.16	−0.03	−0.10
Year fixed effect	Yes		Yes		Yes		Yes		Yes		Yes
Obs	1,454		1,279		1,178		524		654		262
Pseudo ${\text{R}}^2$	0.22		0.22		0.22		0.26		0.33		0.21
Aggressive engagement	0.06	2.00	0.06	1.66	0.05	1.42	0.05	0.74	0.03	1.21	0.12	1.06
Intensive engagement	−0.03	−1.39	−0.03	−1.00	−0.03	−1.21	−0.01	−0.29	−0.07	−3.37	−0.33	−3.55
Successful experience	0.10	3.56	0.09	3.15	0.10	3.29	0.21	2.73	0.04	2.20	0.01	0.04
Focused engagement	0.06	2.86	0.07	2.92	0.08	3.13	0.11	2.15	0.02	1.39	−0.23	−2.13
Hard collaboration	0.14	3.97	0.16	4.15	0.18	4.34	−0.05	−0.78	0.19	4.87	0.43	4.80
Soft collaboration	0.05	2.07	0.05	1.61	0.04	1.27	0.11	1.38	0.07	3.05	0.54	6.18
Firm controls	Yes		Yes		Yes		Yes		Yes		Yes
Year fixed effect	Yes		Yes		Yes		Yes		Yes		Yes
Obs	1,454		1,279		1,178		524		654		262
Pseudo ${\text{R}}^2$	0.27		0.26		0.27		0.28		0.45		0.46

*For final regression, 1 if successful in ES, 0 if successful in CG.

The upper panel of this table reports the marginal effects of characteristics of being successful. The dependent variable is a dummy variable equal to one if an engagement sequence is successful and zero for unsuccessful engagements. Firm characteristics are measured at the year before the initial engagement date. Only the initial engagement is kept for each sequence. The bottom panel reports the marginal effects of engagement features of being successful, after controlling for firm characteristics used in the upper panel (coefficients on firm controls omitted for brevity). Engagement year fixed effects are included in all regressions. Standard errors are clustered at the firm level. Independent variables are defined in Appendix B. All variables are winsorized at the 1st and 99th percentile levels. Coefficients in bold are statistically significant at the 10% level.

The above analysis compares each target firm with their counterparts in the same industry. In the final pair of columns in Table 6, we also compare the target firm characteristics between CG and ES engagements (see the column titled “ES-CG”). In this specification, we remove control firms from our sample and examine the likelihood of engaging with the target firm on an ES issue relative to the likelihood of engaging on a CG issue. We observe that, relative to CG targets, ES targets are characterized with higher market share, higher employee efficiency (sales over employees), lower shareholdings from the asset manager, and, more importantly, higher reputational concerns (higher advertising expenditure and lower product similarity). These findings further confirm and highlight the difference between traditional activism on corporate governance issues and activism on ES issues.⁹

5. Determinants of Successful Engagements

With which types of target firms are engagements more likely to be successful? To answer this question, we examine the firm characteristics of the successful engagements in the year before the initial engagement and compare them with those of the unsuccessful ones. The upper panel of Table 7 reports the marginal effects of our probit regression models. In these models, we continue to control for year fixed effects, and standard errors are clustered at the firm level. To have a consistent sample of firms, we keep the same set of target firms as those used in the previous table, that is, those with valid matches available.

We report the results for the whole sample in the upper left panel of Table 7. Compared with the results reported in Table 6, coefficients on size, return on assets, advertising intensity, illiquidity, and analyst coverage continue to be significant with the same signs, indicating that target firms with poorer performance and higher reputational concerns can benefit most from ESG activities. Moreover, as in Table 6, the positive coefficient on size also indicates that the potential benefits are scalable and the fixed costs of the desired changes are more affordable for large firms. On the other hand, coefficients on the entrenchment index and product similarity score lose their significance, indicating that managerial entrenchment and product differentiation are not determining factors for success. We find that the coefficient on market share becomes negative and significant, suggesting that engagement with industry leaders is less likely to be successful, probably because they have less headroom for improvement.

We further discuss our findings focusing on the upper right panel, which reports the results using CG and ES subsamples separately. For the CG subsample results, coefficients on firm size, firm age, and analyst coverage continue to be significant with the same signs as those reported in Table 6, suggesting that the asset manager tends to achieve greater success in engagements on corporate governance issues, with larger and longer established firms. On the other hand, coefficients on the asset manager, pension activists, and SRI funds' shareholdings lose their significance, suggesting that the success of CG engagements does not rely on the owner's voting rights. This finding is similar to that documented by Smith (1996) in connection with activism by pension fund CalPERS, but contrasts with the positive association between voting outcome and institutional ownership in Gillan and Starks (2000) and Gordon and Pound (1993). This is again consistent with the relatively active, but generally less confrontational engagement strategy that the asset manager employs. The coefficients on the cash holding, capital expenditure, R&D expenditure, and advertising expenditure become insignificant for the CG subsample. This suggests that the success of engagements on corporate governance issues neither requires heavy spending nor does it rely on the reputational concerns of the engaged companies.

For the ES subsample results, the positive coefficient on firm size indicates that the potential benefits are scalable and the fixed costs of the desired changes are more affordable for large firms. The significant coefficients on advertising intensity and analyst coverage suggest that reputational concerns play an important role for the success in ES issues. The positive coefficients on cash holding and negative coefficients on capital expenditure and R&D expenditure suggest that improvements in relation to ES issues are potentially costly. The negative coefficients on return on assets and market share suggest that the scope for improvement is a potential indicator of success for ES engagements. We also find that the asset manager tends to achieve greater success with targets with greater financial slack (leverage plays a negative role in determining success). Similar to the CG subsample results, we do not find shareholdings of the asset manager or other investors or entrenchment index to play a role in determining success of ES engagements.

Hong, Kubik, and Scheinkman (2012) show that CSR is costly, and hence it is more prevalent in less financially constrained firms. Our finding is consistent with the impact of being less financially constrained, since we observe that ES engagements with firms that have lower capital expenditure, R&D expenditure, and leverage and larger cash holdings are more likely to succeed. Overall, target firms that have the potential and means to benefit most from active ownership initiatives are most likely to adopt the changes that have been proposed to them, although the experience of other activists could of course be different (c.f. Smith (1996)).

The above analysis compares successful engagements with unsuccessful ones under the same themes. In the last two columns of Table 7, we compare successful engagements across CG and ES issues (i.e., only successful engagements are kept in this analysis). We observe that relative to successful CG engagements, successful ES engagements have marginally larger firm size, significantly lower R&D expenditure, higher advertising expenditure, and somewhat lower shareholdings from the asset manager and SRI funds. This is consistent with the argument that ES engagements are costly and the success of these engagements relies on target firms' reputational concerns and the asset manager's potential collaborations with other investors. This finding further highlights the differences between traditional activism on corporate governance issues and ESG activism.

In the bottom panel of Table 7, we repeat the analysis from the upper panel after additionally including variables for engagement features, as discussed in Section 3.3.¹⁰ The results indicate that successful experience with the same target firm is an important determinant for the success of both CG and ES engagements. We also find that focused CG engagements and collaborative (both hard and soft) ES engagements are more likely to be successful. It is interesting that more intensive engagements (i.e., repeated ones) are less productive for the success of ES engagements. In the final pair of columns, we compare the engagement features of successful ES engagements with successful CG engagements and find that the former are more collaborative but are less intensive and less focused.

To examine whether lawsuits would contribute to the success of engagements, in unreported analysis, we include the number of lawsuits as an independent variable in the probit regression of success. We find that the number of lawsuits is positively associated with the probability of success, for both the whole sample and the subsample of ES engagements, whereas the coefficient on the number of lawsuits is not significantly different from zero for the subsample of CG engagements. Comparing successful ES engagements with successful CG engagements, we find that the number of lawsuits is a more important determinant for the former. This suggests that target firms facing potential legal pressure are more likely to adopt changes in environmental and social issues suggested by the ESG activist. Our finding is in line with Glac (2010), who states that shareholder activism on ESG challenges the existing legal boundaries and initiates a shift in legislation and the interpretation of regulations. This shift allows broader increases in public awareness, especially through increased engagements through proxy process. As a response to shareholder and public demands, firms embrace ESG as a strategic opportunity.

6. Stock Market Responses to Engagements

How does the stock market respond to active ownership? To answer this question, we look at cumulative abnormal returns around initial engagements, over various investment horizons.

6.1 Cumulative abnormal returns around initial engagements

In our analysis, stock returns are measured by calendar month and the month of the initial engagement date is defined as month 0. We use monthly stock returns rather than daily returns because some of the engagements are private, and hence it might take time for the market price to reflect information that is not initially in the public domain. Using monthly returns is also prudent in cases of information leakage or engagements being triggered by earlier public events.

We use stock return data from the CRSP monthly files. We calculate monthly abnormal returns in two ways. First, since our target firms are large in market capitalization, we use size-adjusted returns, calculated as the monthly stock return minus size-decile matched portfolio return. This follows the event study methodology developed by Dimson and Marsh (1986) and implemented by CRSP; it is also an approach used by Klein and Zur (2009). Decile size portfolio returns and size breakpoints are downloaded from Kenneth French's website. Second, we use market-adjusted returns, calculated as the monthly stock return minus the value-weighted market return from CRSP.¹¹ We tabulate results based on size-adjusted returns. For completeness, we discuss in the text results using market-adjusted returns when appropriate, though we find consistent results using both measures.

The cumulative abnormal returns of target companies around the initial engagement dates are plotted in the upper panel of Figure 1. For each event month, we calculate the average abnormal return from holding an equal-weighted portfolio of all target firms that initiated engagements in month 0. All abnormal returns are winsorized at the 1st and 99th percentiles before calculating sample means for each event window. We set the base value for month −1 as zero and cumulate the average abnormal returns from months −1 through +18. The bold line for the whole sample trends upward, indicating that engagements increase shareholders' value on average. This line portrays a +2.3% cumulative abnormal return over a postengagement horizon of one year. If we use market-adjusted abnormal return, the one-year CAR is +1.8%.¹²

Figure 1

Cumulative abnormal returns around initial engagements

This figure plots the cumulative monthly abnormal returns (CARs) around the initial engagements from one month prior to the engagement month to 18 months afterward. The upper chart examines the entire sample and the lower one looks separately at the CG and ES subsamples. Each CAR is decomposed into the CAR for successful engagements (i.e., those that achieved milestones) and the CAR for unsuccessful engagements. For each event month, we calculate the average abnormal return from holding an equally weighted portfolio of all target firms that initiated engagements in month 0. The stock returns are adjusted for Fama-French decile size-matched returns. The stock returns are winsorized at the 1st and 99th percentile levels before calculating the average CARs.

We further split the sample into successful (the light line) and unsuccessful engagements (the dotted line). To remove duplications, for each sample, we keep only one engagement for each firm in a calendar month (our conclusions are not impacted by this empirical choice). The successful engagements generate a cumulative abnormal return of around +7.1% over the year following engagement. The corresponding CAR using market-adjusted return is +4.9%. The chart shows that the cumulative abnormal return on successful engagements is much higher than that of the unsuccessful ones and the difference becomes larger with time. The difference reaches a peak of +6.3% at month +12 and of +7.5% at month +16. These two horizons correspond roughly to the times at which the median and average target firm in our sample achieves its milestone (see Table 3). The concave curve for successful engagements reveals efficiency in the market's response to engagements, insofar as significant improvements are usually made before milestones are recognized and recorded.

Although the +2.3% (+1.8%) annual size-adjusted (market-adjusted) cumulative abnormal return to activism for the whole sample is lower than the +7% to +10% abnormal returns generated by activist hedge funds, as documented in Brav et al. (2008) and Klein and Zur (2009), it is much higher than the negligible abnormal returns generated by traditional shareholder activism, as discussed in detail by Becht et al. (2009). Thus, in terms of its impact on stock market values, the active ownership we examine here lies between traditional shareholder activism and hedge fund activism. The +7.1% (+4.9%) annual size-adjusted (market-adjusted) abnormal return associated with successful engagements broadly matches the annual abnormal return of +4.9% generated by the Hermes U.K. Focus Fund, the strategy of which is midway between a traditional shareholder activist and an activist hedge fund (see Becht et al. 2009). It is also comparable with +5.1% abnormal returns generated by other entrepreneurial activism documented in Klein and Zur (2009).

In the lower panel of Figure 1, we split the sample into the ES theme (bold solid line and bold dashed line) and the CG theme (light solid line and light dotted line). We observe similar patterns and magnitudes across these two subsamples. For example, we observe a one-year size-adjusted abnormal return of +7.2% for the ES subsample and +7.1% for the CG subsample. This suggests the existence of a threshold for success to be pursued and achieved for both types of engagements.¹³ It also provides some reassurance that the outperformance arising from engagement success is robust to the type of interaction with the target company.

Note that to construct the figures, we first calculate the portfolio abnormal return of each event month by averaging the abnormal returns of all engagements and then calculate CARs by cumulating the portfolio abnormal returns. This methodology does not allow us to evaluate the statistical significance of the portfolio CARs or whether CARs are statistically different for the successful and unsuccessful subsamples. To have an idea of the statistical significance of CARs at different event windows, we adopt two different approaches (results untabulated for brevity). The first approach is to calculate the CAR for a specific window for all the engagements in the sample and then to compute the sample mean and $p$-value (two-tailed) to evaluate whether the average CAR is significantly different from zero. For CAR (0, +12), the whole sample and the successful subsample have a positive mean of +2.9% (⁠$p$-value = 0.00) and +7.2% (⁠$p$-value = 0.00) for size-adjusted CARs, respectively, and +2.4% (⁠$p$-value = 0.00) and +5.1% (⁠$p$-value = 0.00) for market-adjusted CARs, respectively. In both size-adjusted and market-adjusted cases, unsuccessful subsamples have CARs that are statistically indistinguishable from zero. The difference between successful and unsuccessful subsamples is also statistically significant for both size-adjusted and market-adjusted CARs. We observe similar patterns for the CG and ES subsamples.

The second approach is to compute buy-and-hold returns. We calculate the return of a portfolio that buys the stock of the target company at the month of the initial engagement and sells it at the month when the milestone is recorded. For unsuccessful engagements, since there is no milestone date, we form the portfolio using the median horizon of the successful engagements (12 months) as the holding period. With this buy-and-hold analysis, we confirm our main findings that both the size-adjusted and market-adjusted returns to successful engagements are significantly positive and higher than the zero returns to unsuccessful engagements. For example, we find that successful engagements generate an annualized size-adjusted buy-and-hold return of +10.4% (⁠$p$-value = 0.00) and market-adjusted buy-and-hold return of +6.8% (⁠$p$-value = 0.00), while the annualized size-adjusted or market-adjusted return of the unsuccessful sample is not statistically different from zero. We also observe similar patterns for the CG and ES subsamples. An advantage of using the buy-and-hold approach is that holding periods correspond to the horizons of successful engagements in different themes (see Table 3).

To sum up, this section indicates that active-ownership engagements increase shareholders' value on average and the positive returns are apparent for engagements on both CG and ES themes. This accords with Aktas, de Bodt, and Couisin (2011) who find a positive market reaction for acquirers investing in target firms with good social and environmental risk management practices. While it is conceivable that the better stock performance of engaged companies is solely attributable to extraordinary stockpicking skills by the asset manager, the fact that we document differential abnormal returns for successful and unsuccessful engagement subsamples mitigates this potential concern.

6.2 Cross-sectional variation of abnormal returns

Table 8 reports the cross-sectional analysis of the cumulative abnormal returns over different event windows. For some firms there are multiple engagements in a month on the same or different issues. To disentangle market reactions to different engagements, we aggregate the information at a monthly frequency. We count the numbers of successful and unsuccessful engagements under different themes for each engagement month and regress cumulative abnormal returns over three different windows (event month, months 0 to +6, and months 0 to +12) on these counting variables. We include industry and year fixed effects to control for unobserved heterogeneity between industries and years. We also experiment with using a dummy in place of each counting variable, if its value is positive, which yields analogous results (data unreported). To be consistent with the previous section, we continue reporting results using size-adjusted abnormal returns. In unreported analysis, we also use market-adjusted abnormal return to calculate CARs and obtain similar results.¹⁴

Table 8

Cross-sectional variation on abnormal returns

	Size-adjusted abnormal returns (decile match)
Dependent variable:	Event CAR		CAR(0,+6)		CAR(0,+12)
	Coef.	t-stat	Coef.	t-stat	Coef.	t-stat
Successful engagements
Successful corporate governance	−0.006	−0.90	0.060	3.10	0.086	3.34
Successful business ethics	−0.001	−0.08	0.001	0.03	−0.006	−0.10
Successful sustanability management	−0.009	−0.80	0.033	0.75	−0.012	−0.23
Successful climate change	−0.012	−0.94	0.062	2.55	0.103	4.16
Successful ecosystem services	−0.027	−1.10	0.042	1.02	0.049	0.42
Successful environmental management	−0.014	−0.85	0.013	0.44	0.020	0.40
Successful public health	−0.021	−0.48	0.092	1.31	−0.008	−0.19
Successful human rights	0.060	2.36	0.146	2.42	0.004	0.05
Successful labor standards	0.013	0.72	0.068	1.95	0.063	1.02
Unsuccessful engagements
Unsuccessful corporate governance	−0.002	−0.43	0.027	1.95	0.025	1.50
Unsuccessful business ethics	−0.012	−1.83	0.017	0.87	0.007	0.32
Unsuccessful sustanability management	0.003	0.34	0.013	0.47	0.021	0.70
Unsuccessful climate change	−0.003	−0.39	−0.024	−0.99	−0.035	−1.30
Unsuccessful ecosystem services	−0.005	−0.49	−0.027	−0.94	−0.037	−1.02
Unsuccessful environmental management	−0.008	−1.47	0.008	0.48	0.004	0.16
Unsuccessful public health	0.012	0.55	0.060	1.48	0.042	0.80
Unsuccessful human rights	−0.001	−0.10	0.043	1.63	0.022	0.67
Unsuccessful labor standards	−0.009	−1.23	−0.016	−0.66	−0.016	−0.56
Industry fixed effect	Yes		Yes		Yes
Year fixed effect	Yes		Yes		Yes
Obs	1,415		1,408		1,384
${\text{R}}^2$	0.09		0.10		0.11

This table reports the results of regressions of cumulative abnormal returns (CARs) around the initial engagements for nine engagement themes, as defined in Table 1. CARs are measured relative to the investment returns of similar sized companies. The independent variables are counting variables indicating the number of successful and unsuccessful engagement sequences under each theme initiated during the event month. Event CAR is the monthly abnormal return for the event calendar month. CAR(0,+6) is the sum of monthly abnormal returns over month 0 to month +6. CAR(0,+12) is the sum of monthly abnormal returns over month 0 to month +12. Intercepts are suppressed. Engagement year and industry fixed effects are included in all regressions. Standard errors are clustered at the firm level. All dependent variables are winsorized at the 1st and 99th percentile levels. Coefficients in bold are statistically significant at the 10% level.

Over the long run, we observe different and statistically significant market reactions to various types of engagement. For example, the cumulative abnormal return over window (0, +6) is +6.0% for one additional engagement in the corporate governance theme, extending to +8.6% over window (0, +12). The cumulative abnormal return over window (0, +6) is +6.2% for one additional engagement in the climate change theme, extending to +10.3% over window (0, +12).¹⁵

These results confirm that activism on responsible investing is different from hedge fund activism. According to Brav et al. (2008) and Klein and Zur (2009), the largest market reactions from hedge fund activism come from engagements on issues of mergers and acquisitions, demanding a board seat, or suggesting an increase in a stake with the intention of buying it in. This is also distinct from other entrepreneurial activism, which generates the highest market reaction when the activist intends to buy more stocks in the firm or expects to become more active (Klein and Zur 2009) or to be involved with restructuring activities (Becht et al. 2009).

The positive abnormal return on successful climate change engagements indicates that investors expect the changes on environmental issues to increase firm value. Consistent with this result, Bauer and Hann (2014) show that firms with proactive environmental engagements, particularly on climate change issues, have lower cost of debt; similarly, Chava (2014) finds that firms with environmental concerns have higher cost of capital.

7. Post-Engagement Changes in Performance

Finally, we examine the mechanisms through which successful engagements could be associated with a favorable stock market response. The existing literature highlights four potential sources: first, attracting more socially conscious consumers (Besley and Ghatak 2007); second, increasing the loyalty of consumers and employees, thereby enhancing operating performance and efficiency (Baron 2008; Portney 2008; Benabou and Tirole 2010); third, attracting more socially conscious shareholders and thereby improving stock market performance (Baron 2008; Benabou and Tirole 2010); and fourth, signaling future governance improvements that enhance the value of the engaged company (Brav et al. 2008; Klein and Zur 2009).

To test the above mechanisms, we employ a difference-in-differences method by comparing subsequent changes in firm performance, investor base, and corporate governance following successful engagements with those following unsuccessful ones. For each engagement sequence we obtain information one year before and one year after the calendar year of the initial engagement date.¹⁶Table 9 reports the regression results on nine dependent variables of interest for the whole sample and the CG and ES subsamples, respectively. The independent variables include an indicator for post-engagement period (Post), an indicator for successful engagements (Success), the interaction of these two (Post × Success), and control variables, such as firm size, market-to-book ratio, industry performance (industry-year median of the dependent variable), and firm and engagement year fixed effects. As a firm can have multiple engagements in our sample, we cluster standard errors by firm. For brevity, we only report coefficients on Post × Success, which captures the difference-in-differences effect.

Table 9

Performance, ownership, and governance after engagements

	Whole		CG		ES
	Sample		subsample		subsample
Specification	Post x	Obs	Post x	Obs	Post x	Obs
	Success	${\text{R}}^2$	Success	${\text{R}}^2$	Success	${\text{R}}^2$
	(1)	(2)	(3)	(4)	(5)	(6)
Post = 1 if Window = +1, Post = 0 if Window = −1
Change in Firm Performance
1 Return on assets	0.010	3,614	0.002	1,555	0.014	2,059
	(2.24)	0.86	(0.42)	0.87	(2.44)	0.86
2 Profit margin	0.015	3,877	0.019	1,627	0.008	2,250
	(1.85)	0.88	(1.57)	0.82	(0.81)	0.91
3 Asset turnover	0.021	3,880	0.000	1,632	0.029	2,248
	(1.73)	0.98	(0.03)	0.98	(1.57)	0.97
4 Sales over employees	0.088	3,847	0.053	1,614	0.114	2,233
	(2.49)	0.93	(1.56)	0.94	(2.42)	0.92
Change in Shareholdings
5 Shareholding of asset manager	0.020	4,098	0.006	1,715	0.028	2,383
	(1.98)	0.64	(0.48)	0.76	(2.24)	0.55
6 Shareholding of pension activists	0.178	4,098	0.090	1,715	0.313	2,383
	(2.86)	0.83	(1.37)	0.84	(2.93)	0.84
7 Shareholding of SRI funds	0.047	4,098	0.024	1,715	0.079	2,383
	(2.20)	0.64	(0.63)	0.67	(3.82)	0.65
8 Stock return volatility	−0.013	3,730	−0.012	1,601	−0.013	2,129
	(−2.87)	0.61	(−1.86)	0.62	(−1.93)	0.65
Change in Corporate Governance
9 Entrenchment index	−0.091	3,706	−0.115	1,590	−0.095	2,116
	(−1.07)	0.73	(−1.06)	0.79	(−0.84)	0.71
Post = 1 if Window = +2, Post = 0 if Window = −1
10 Entrenchment index	−0.256	3,430	−0.235	1,506	−0.269	1,924
	(−3.06)	0.80	(−2.01)	0.81	(−2.13)	0.81

This table reports difference-in-differences regression results on all engagement sequences (successful and unsuccessful). The calendar year of the initial engagement date is defined as window 0. The year before (after) the initial engagement date is defined as window −1 (+1). The upper and middle panels and the first row of the final panel include observations from window −1 and window +1. The last row includes observations from window −1 and window +2 (the second year after the initial engagement date). The dependent variables are corresponding measures in firm performance, shareholdings, and corporate governance, all of which are defined in Appendix B. All regressions include Post, Success, and Post × Success. Post is a dummy variable indicating that the observation is from the period after the initial engagement date, and zero otherwise. Success is a dummy variable defined as one for successful engagement sequences, and zero otherwise. For brevity, we report only the coefficients on Post × Success and its $t$-statistics (in parentheses). Firm and engagement year fixed effects are included in all regressions. Regressions in the upper and middle panels also include firm controls, that is, firm size and market-to-book ratio. Regressions in the upper panel additionally include industry controls, that is, industry median of the corresponding dependent variable in a certain year. Standard errors are clustered at the firm level. All variables are winsorized at the 1st and 99th percentile levels. Shareholdings of asset manager, pension activists, and SRI funds are multiplied by 100, for ease of readability. Coefficients in bold are statistically significant at the 10% level.

Rows 1 to 4 of Table 9 report results on performance measures. We observe significant increases in return on assets and sales over employees after successful engagements for both the whole sample and the ES subsample, but not the CG subsample.¹⁷ This suggests improved operating performance, sales, and/or employee efficiency following successful engagements on environmental and social issues, consistent with the argument that active ownership initiatives can expand the customer base and enhance customer and employee loyalty (the first two mechanisms). Not observing significant improvements in operating performance for the CG subsample is largely consistent with research on the impact of traditional and hedge fund / entrepreneurial activism (see Smith 1996; Karpoff, Malatesta, and Walkling 1996; Becht et al. 2009; Klein and Zur 2009).

Rows 5 to 8 of Table 9 report the results on institutional ownership and stock volatility. We observe significant increases in the shareholdings of the asset manager, pension activists, and SRI funds after successful engagements for the whole sample and the ES subsample, but not the CG subsample. Active ownership appears to attract investments from socially conscious shareholders (the third mechanism), consistent with the findings of Dhaliwal et al. (2011) that firms disclosing superior CSR performance attract more institutional investors. We also observe a decrease in volatility for successful engagements on all ESG issues, consistent with Albuquerque, Durnev, and Koskinen (2014) that ESG activities reduce firm risk. Rows 9 and 10 report results on the entrenchment index that support the view that successful engagements lead to governance improvements (the fourth mechanism). This is a strong finding given that the entrenchment index is “sticky” as it is updated every three years. In particular, note that we observe a significant result for the entrenchment index using window +2 (two years after the initiate engagement date). We find similar results by using the governance index (data unreported). Together with the findings from targeting and success analyses (Tables 6 and 7), this contrasts with Cheng, Hong, and Shue (2013) who argue that CSR reflects managerial agency problems.

Results in this section suggest that the positive market reaction to successful engagements documented in Section 6 could be, at least partially, explained by post-engagement improvements in firm performance, shareholdings, and governance. This evidence also portrays active ownership as more effective than traditional interventions in facilitating changes in target firms (Smith 1996; Karpoff, Malatesta, and Walkling 1996). The above-discussed differences in results subsequent to successful CG and ES engagements suggest different mechanisms through which ESG activism can potentially improve firm value.

8. Discussion and Conclusion

A question that permeates the activism literature is whether one can infer a causal link between engagements and subsequent corporate performance. We consider several ways in which the favorable performance of successfully engaged companies might be an illusion. First, performance improvements could result from filtering by engaged companies, which accept value-enhancing proposals and reject value-destructive proposals. This would constitute good governance, so we introduce the governance and entrenchment indexes into the regressions in Table 8. These variables do not have any significant coefficients (data untabulated), so the observed performance improvement is unlikely to be attributable entirely to management filtering.

Second, we report in Section 6 a positive cumulative abnormal return for successful engagements and a zero return for unsuccessful ones, concluding that (expected) ESG improvements increase the market value of engaged companies. An alternative explanation is that target firms wait, and adopt the requested changes if their stock prices increase or firms signal their anticipated positive future performance through ESG activities. In this “reverse causality” explanation, it is positive stock market performance that triggers ESG changes in the target firms, rather than the other way round. We therefore include the CARs defined in Section 6.2 as an additional predictive variable for the models in Table 7 after controlling for other factors. We use a variety of windows for CARs but none of them has a coefficient significantly different from zero, and the target firm's stock return during the engagement period does not appear to be a determinant of success. This indicates that it is unlikely that ESG improvements in the target firms are a consequence of their superior future performance.

A third possibility is that milestones are recorded retrospectively after a positive stock market reaction. However, one third of the milestones coincide with the dates of shareholder meetings, and we have verified that they are recorded correctly. We repeat the analysis in Section 6.1, partitioning the successful engagement subsample into two based on the existence of a shareholder meeting. If the positive market reaction were an artifact of recording milestones after a price rise, the favorable performance would not be apparent when milestones coincide with shareholder meetings. Instead, we find very similar results across these two subsamples.

Given that engagement activities improve shareholder value, a final question is why firms might not voluntarily pursue such a strategy. It is possible for a firm to improve its ESG policies in the absence of intervention, but that is unlikely to happen to the fullest extent for a couple of reasons. On the one hand, Table 6 reveals that target firms have poorer corporate governance than control firms, indicating more serious agency issues and a greater likelihood of deviating from shareholder value maximisation, which would impede adoption even of value-enhancing ESG projects. On the other hand, the active owner provides directional guidance to target companies and, absent external input, some engaged companies will inevitably lack the ability to identify and respond appropriately to ESG opportunities. In other words, the asset manager (in partnership with other active investors and stakeholders) can establish appropriate standards and create the necessary ambience for engaged firms to shift to a new equilibrium with higher firm value and lower ESG risks.

To conclude, based on a proprietary dataset on responsible investment strategies, we document positive market reactions to active-ownership engagements in U.S. public firms over 1999–2009. On average, these ESG activities give rise to a positive size-adjusted abnormal return of +2.3% over the year after initial engagement. The average one-year size-adjusted abnormal return after initial engagement is +7.1% for successful engagements, but there is no adverse reaction to unsuccessful ones. The positive abnormal returns are most pronounced for engagements on the themes of corporate governance and climate change. Compared to matched firms, companies with poorer performance, inferior governance structure, greater reputational concerns and higher shareholding from the asset manager are more likely to be targeted. Engagements are more likely to be successful in achieving the activist's objectives if the target firm is more concerned about its reputation, and has higher capacity to implement change and larger headroom for improvement, especially for those on ES issues. In contrast with CG engagements, collaboration between the asset manager and other activist investors and stakeholders significantly increases the success rate of ES engagements. Consistent with arguments that ESG activities attract socially conscious customers and investors, we find that, after successful engagements, particularly for those on ES issues, engaged companies experience improvements in their operating performance, profitability, efficiency, shareholding, and governance.

What are the limitations of our work? We have studied a single and, in hindsight, successful example of active ownership. This calls for caution in interpreting our findings. The shareholder value that could be generated from activism by an under-skilled or under-resourced team would doubtless be lower. In addition, the abnormal return could also be specific to the time period studied in our sample. As investors learn to appreciate the difference between more and less responsibly managed firms, the reward from ESG engagement may evaporate. This would be consistent with the patterns, reported by Bebchuk, Cohen, and Wang (2013) and Borgers et al. (2013), for the abnormal returns associated with good corporate governance and stakeholder relations to disappear.

Our study provides the first detailed evidence on the impact of responsible investing. Future research might usefully focus on the precise mechanisms that determine the price reaction to activist engagements, and on examining whether the models developed here for the United States have validity in other markets around the world.

Appendix A. Examples of ESG Engagements

A.1 Environmental Engagements, Apple Inc. (Environmental Management)

Apple Inc. is a U.S. company producing consumer electronics, computers, and computer software. On December 14, 2006, in collaboration with a dialogue led by As You Sow (a foundation promoting environmental and social corporate responsibility through multiple initiatives), the asset manager sent a letter to Apple, copied to T.M., Senior Manager for Apple's Supplier Responsibility, and various board members. The letter highlighted various environmental issues that the company was facing, including its progress on product take-back and recycling efforts, assigning board responsibility for CSR, and publishing a comprehensive sustainability report. On January 22, 2007, the asset manager had a phone conversation with T.M. as a follow-up to the previous letter and reiterated the need for Apple to demonstrate its commitment to ESG. On February 12, 2007, the asset manager signed on a group letter, coordinated by As You Sow, asking Apple for specific commitments to addressing its environmental issues, to which Apple responded positively. The asset manager recorded these three engagements as Request for Change. On May 9, 2007, Apple announced new environmental commitments in advance of its 2007 annual general meeting, which was scheduled to include two environmental shareholder proposals: one on take-back and one on the use of toxics. Apple's commitments included eliminating the use of toxics in company products and the expansion of take-back and recycling efforts. The asset manager recorded this event as Milestone.

A.2 Social Engagements, Yahoo! Inc. (Human Rights)

Yahoo! Inc. is a U.S. Internet company owning a frequently visited Web portal and search facility. On August 25, 2006, the asset manager had a conference call with M.S., Vice President and Deputy General Counsel and founder of Yahoo!'s Business & Human Rights program, and M.N., Vice President for Yahoo!'s Investor Relations. They discussed the human rights issues on Access, Security, and Privacy (ASP). The meeting followed the media attention received by Yahoo! after the firm delivered user information to Chinese security officials, who then incarcerated one of their customers in 2005. The asset manager records this engagement as Raising Awareness. On November 26, 2006, M.S. attended the asset manager's seminar on ASP in the Global Digital Economy during which the asset manager issued best practice recommendations on how to manage ASP risks. The manager recorded this engagement as Request for Change. On June 12, 2007, at firm's annual meeting, J.Y., the founder of Yahoo!, announced the firm's commitment to human rights and freedom of expression online. The asset manager recorded this event as Milestone.

A.3 Governance Engagements, Illinois Tool Works Inc. (Corporate Governance)

Illinois Tool Works (ITW) is a U.S. company providing hardware, software, and services to consumers and businesses. On March 11, 2005, in a meeting with J.B., ITW's Vice President of Investor Relations, the asset manager asked whether ITW planned to produce a CSR report. ITW's response indicated that they were unaware of CSR and sustainability issues. On August 16, 2005, the asset manager sent a letter to J.W., General Counsel and Secretary of ITW, copied to J.B. The letter specifically asked for a CSR report together with a detailed description on what the manager would expect such a report to cover. The manager recorded these two engagements as Request for Change. On October 14, 2005, the asset manager had a follow-up call to J.B. and left a voice mail referencing the letter sent in August and inquiring about plans to issue a CSR report. The manager recorded this engagement as Raising Awareness. On November 16, 2005, the asset manager filed a shareholder proposal to be included on the 2006 proxy, calling on ITW to issue a sustainability report based on the Global Reporting Initiative guidelines. The proposal was co-filed with several other institutional investors, including Walden Asset Management, Trillium Asset Management, and Domini Social Investments. The manager recorded this engagement as Request for Change. On January 9, 2006, the asset manager received a phone call from J.W. regarding the shareholder proposal the asset manager filed. The conversation was investigative, nonconfrontational, and guiding with respect to the contextual requirements of the CSR report. The asset manager recorded this engagement as Raising Awareness. On March 9, 2006, the asset manager provided formal feedback to ITW's first interim CSR report, which it committed to publish on the company Web site within 60 days. The manager recorded this engagement as Request for Change. On May 19, 2006, ITW published “Response to Investors - Interim Sustainability Report.” As a result, the asset manager and co-filers withdrew the shareholder proposal calling for a sustainability report in 2006. The manager recorded this event as Milestone.

Appendix B. Variable Definitions

Variable name	Definition
Fundamental data (Source: Compustat NA)
Firm size	Market value of equity (in billion $)
Market-to-book	Market value of equity / Book value of equity
Tobin's q	(Market value of equity + Book value of debt) / (Book value of equity + Book value of debt)
Firm age	Firm age relative to the year when the firm initially appeared in Compustat
Sales growth	Annual sales growth rate
Return on assets	Earnings before interest, taxes, depreciation and amortisation (EBITDA) / Average total assets
Asset turnover	Sales / Average total assets
Sales over employees	Sales / Number of employees (in billion $)
Profit margin	Earnings before interest and taxes (EBIT) / Sales
Cash flow	(Net income before extraordinary items + Depreciation and amortisation) / Average total assets
Stock return	Buy-and-hold stock return of the fiscal year
Stock return volatility	Standard deviation of monthly stock returns during the fiscal year
Leverage	Book value of debt / (Book value of debt + Book value of equity)
Cash holding	Cash / Total assets
Dividend yield	Total dividends / (Market value of common equity + Book value of preferred equity)
Dividend payout	Total dividends / Net income before extraordinary items
R&D expenditure	R&D expenditures / Average total assets
Capital expenditure	Capital expenditures / Average total assets
Advertising expenditure	Advertising expenditures / Average total assets
Industry Herfindahl index	Herfindahl-Hirschman index computed using all firms within the same industry (four-digit SIC)
Fundamental data (Source: Compustat NA)
Industry advertising intensity	Industry (four-digit SIC) median of advertising intensity, (Advertising expenditures / Sales), multiplied by 100
Tangibility	Tangibility ratio (Net PP&E / Total assets)
Market share (segment)	Market share of sales calculated using segment data, calculated as firm's sales divided by industry (three-digit SIC) sales
Product similarity scores (Source: Hoberg and Phillips 2010)
H&P product similarity score	A firm-level variable that is inversely related to product differentiation
Shareholding data (Source: Thomson Reuters 13F)
Shareholding of asset manager	Number of shares held by the asset manager as a proportion of the number of shares outstanding
Shareholding of pension activists	Number of shares held by institutions defined as activist by Cremers and Nair (2005) and Larcker, Richardson, and Tuna (2007) as a proportion of the number of shares outstanding
Number of pension activists	Number of activist pension institutions
Shareholding of SRI funds	Number of shares held by a fund defined by Risk Metrics ISS shareholder proposal database as SRI fund or social fund as a proportion of the number of shares outstanding
Market data (Sources: CRSP, IBES, and RiskMetrics/IRRC)
Amihud illiquidity	Amihud (2002) illiquidity measure
Number of analysts	Number of analysts following the firm
Governance index	Gompers, Ishii, and Metrick (2003) governance index
Entrenchment index	Bebchuk, Cohen, and Ferrell (2009) entrenchment index
Engagement data (Source: Asset manager)
Successful (unsuccessful) theme	Number of successful (unsuccessful) engagements on that particular theme within the same calendar month (for Table 8)
Aggressive engagement	A dummy variable defined as one if there is at least one Request for Change in the engagement sequence, and zero otherwise
Intensive engagement	A dummy variable defined as one if there are more than one (repeated) engagements in the sequence, and zero otherwise
Successful experience	A dummy variable defined as one if the asset manager has had successful engagements with the same target firm before the initial engagement, and zero otherwise
Focused engagement	A dummy variable defined as one if the engagement sequence is focused, and zero otherwise. An engagement sequence is focused if the asset manager deals with issue(s) under a particular theme at the initial engagement. An engagement sequence is called multiple engagements if the asset manager deals with issues under different themes at the initial engagement
Collaboration data (Source: Engagement records)
Hard collaboration	A dummy variable defined as one if the engagement sequence has at least one hard collaborator, and zero otherwise
Soft collaboration	A dummy variable defined as one if the engagement sequence has at least one soft collaborator, and zero otherwise

We are grateful to the Berkeley-Haas Center for Responsible Business for awarding us the Moskowitz Prize for research in Socially Responsible Investing. We benefitted from comments by the Editor (Laura Starks), the Referee, Andrew Ang, Brad Barber, Sudipta Basu, Rob Bauer, Alon Brav, David Chambers, Alexis Cheang, Bhagwan Chowdhry, Gordon Clarke, Francesca Cornelli, Amil Dasgupta, Alberta Di Giuli, Alex Edmans, Otgo Erhemjamts, Julian Franks, William Goetzmann, John Griswold, Denis Gromb, Harrison Hong, Ioannis Ioannou, Robert Jenkins, Andrew Karolyi, Simi Kedia, April Klein, Idar Kreutzer, Philipp Krüger, Lloyd Kurtz, Rob Lake, Laura Lindsey, Karina Litvack, Paul Marsh, Jun Qian, Karthik Ramanna, Raghavendra Rau, Pietra Rivoli, Catherine Schrand, Rik Sen, Henri Servaes, Andrei Shleifer, Hege Sjo, Martin Skancke, Suraj Srinivasan, Mike Staunton, Avanidhar Subrahmanyam, İrem Tuna, Roger Urwin, Ross Watts, and workshop participants at Bentley University, Boston College, Cambridge University, Koç University, Norges Bank, Southwestern University of Finance and Economics at Chengdu, and Temple University. We also benefitted from comments by participants at the Asset Management Conference at London Business School, Charity Finance Group Conference in London, Conference on Inclusive Capitalism in London, Conference on the Impact of Responsible and Sustainable Investing at Hong Kong University of Science and Technology, ETHOS Centre for Responsible Enterprise Conference at Cass Business School, European Finance Association Annual Meeting in Cambridge, FMA European Conference in Istanbul, Grieg Investor Seminar on Responsible Investing in London, Harvard Business School Sustainability Conference, Inquire Europe & Inquire UK Spring Seminar in Budapest, JAE Corporate Accountability Conference at Harvard Business School, PRI–CBERN Conference on Responsible Investment in Toronto, Principles for Responsible Investment Signatories Conference in London, Society for Financial Studies Cavalcade at the University of Miami, SRI Conference in Connecticut, Summit on Responsible Finance at the University of Geneva, Summit on Responsible Investing at Cambridge Judge Business School, Sustainability and Finance Symposium at UC Davis, and World Investment Forum in Napa. We are grateful for the support of Forum pour l'Investissement Responsable, Inquire Europe, London Business School, the Newton Centre for Endowment Asset Management (Dimson), and the Young Research Scholars Seed Fund at the Fox School of Business, Temple University (Li). A preliminary version of this study was entitled “Activism on Corporate Social Responsibility.” We especially thank our data contributor for providing the dataset analyzed in this paper. Our research is independent, and we have no financial connection with organizations that stand to benefit from our findings. All errors are ours.

¹ In this paper we abbreviate five terms that we also define in the text: socially responsible investing (SRI), environmental, social, and governance (ESG), corporate social responsibility (CSR), corporate governance (CG), and environmental and social (ES).

² See, for example, Black (1998), Karpoff (2001), Romano (2001), Barber (2007), Carleton, Nelson, and Weisbach (1998), and Gillan and Starks (2007) for traditional shareholder activism; see Brav et al. (2008), Becht et al. (2009), Klein and Zur (2009), and Brav, Jiang, and Kim (2015) for hedge fund/entrepreneurial activism.

³ ES engagements include issues in the following themes: climate change, ecosystem services, environmental management, public health, human rights, labor standards, business ethics, and sustainability management and reporting. See Table 1 for further details on the engagement areas, themes, and issues.

⁴ Analyses with buy-and-hold returns generate similar results.

⁵ Beside suggesting differential impacts on firm value, these views also provide different predictions for the linkage between CSR activities and corporate governance. The first view predicts improvements in corporate governance following CSR activities. The second view does not imply any linkage between CSR and corporate governance since management still maximizes profits. The third view predicts that poor governance leads to CSR activities. Our results do not support the third view: we find no evidence that poor governance contributes to the success of engagements.

⁶ See Krüger (2015) for an empirical analysis of stock market reactions to public news regarding CSR events.

⁷ Within each theme, there is an issue type called “Other.” We read through the detailed records and find that these are engagements that are difficult to label as any of the listed issues. Our analyses are based on classifications at the engagement theme level. Additional data descriptions are provided by Bauer, Clark, and Viehs (2014).

⁸Bebchuk, Cohen, and Ferrell (2009) argue that six (out of 24) provisions in the governance index of Gompers, Ishii, and Metrick (2003) matter most for the corporate governance of the firm and hence construct the entrenchment index using these six provisions only.

⁹ To examine whether lawsuits could be a potential factor in the asset manager's targeting strategy, we obtain data on legal lawsuits from Audit Analytics and calculate the number of lawsuits each sample firm had in the year before being targeted. In an unreported analysis, we repeat all the regressions in Table 6 by additionally including the number of lawsuits as an independent variable. We do not find significant coefficients on the number of lawsuits in any of these regressions. However, when we further limit our sample to engagements under the climate change theme, we find a positive and significant coefficient on the number of lawsuits. It is possible that lawsuits may have attracted the CSR activist's attention for environmental issues.

¹⁰ We present the results without engagement features in the upper panel of Table 7 to be consistent with Table 6. In the bottom panel of Table 7, after additionally including engagement features, the estimation results for firm characteristics (serving as firm controls) are very similar to those reported in the upper panel and hence are untabulated for brevity.

¹¹ We compare the average firm characteristics, such as firm size, market-to-book ratio, and leverage, of our sample with those of the CRSP universe and find that our sample mean is much more comparable to the value-weighted average of the CRSP universe than to the equal-weighted average of the CRSP universe. The value-weighted market return from CRSP is therefore a more appropriate benchmark than an equal-weighted index.

¹²Klein and Zur (2009) also document higher size-adjusted abnormal returns in their event window, compared to market-adjusted returns. We find that the predicted probability of success from the probit model in Table 7 is positively associated with the CARs from months 0 to +12, with a coefficient of +0.08 and a $p$-value of 0.09 (data untabulated). Given that successful engagements lead to positive abnormal returns, this finding suggests that our success prediction model in Table 7 is well specified. This is also consistent with the observation in Klein and Zur (2009) that investors, on average, are able to differentiate, albeit with error, between successful and unsuccessful campaigns.

¹³ The Editor observed that the threshold is likely to be higher for ES engagements, compared to CG engagements. We find that post-engagement improvements in firm performance are stronger for the ES subsample (see the discussion in Section 7).

¹⁴ In regressions using market-adjusted return to calculate CARs, we include size, market-to-book ratio, and prior-year leverage in the regressions as controls for traditional risk factors. In addition, we also control for lagged stock return, calculated as the monthly stock return averaged over the same number of months prior to the event window (see Brav et al. 2008). The control variables are insignificant, except for size, which reinforces the case for using size-adjusted abnormal returns.

¹⁵ The positive coefficients on unsuccessful corporate governance engagements for CAR(0,+6) and CAR(0,+12) regressions could be due to the fact that not-yet-successful engagements are classified as unsuccessful in our sample. The coefficients on unsuccessful corporate governance engagements become negative and insignificant in both regressions if we limit our sample to engagements initiated before July 2008, that is, one year before our sample period ends (data unreported).

¹⁶ We choose a one-year window in pre- and post-engagement periods because a median successful engagement achieves the milestone after one year.

¹⁷ We also report results on profit margin and asset turnover, two components of return on assets, to illustrate the driving forces behind the increase in return on assets. We only observe significant increases in these two components for the whole sample.