Quasi-Insider Shareholder Activism: Corporate Governance at the Periphery of Control

Abstract

We analyze shareholder activism by “quasi-insiders”: founders, former executives, and other individuals tangentially connected to a firm. These individuals seek control in their campaigns, use aggressive tactics, and target smaller, poorly performing firms. Their campaigns are associated with positive announcement returns. Former CEOs who engage in campaigns often depart from the target under negative circumstances before launching their campaigns soon afterward. Our results suggest that insiders at the periphery of control may intervene in firms too small for traditional activists to target. (JEL: G34, G32)

Activist shareholders attempt to induce change in firms through a combination of persuasion, proxy contests, and other formal campaign tactics. The archetypal shareholder activist is an outsider, such as a hedge fund, whose only connection to a firm it targets is through its ownership of an equity stake, often acquired in conjunction with the launch of the campaign (Cohn, Gillan, and Hartzell 2016). However, a number of activism campaigns in recent years involve shareholders who are not pure outsiders but rather have a past or current connection with the firms they target. For example, Steven Vestergaard, founder of Destiny Media Technologies, was fired as CEO of the company in 2017 but continued to own a significant stake in the company. In 2019, he launched a proxy contest, nominating five directors, including himself, to the company’s five-person board in opposition to management’s nominees, arguing that he was wrongfully terminated and that “the company has gone to cutting expenses to show short term profits at the expense of long term innovation and revenue growth.”¹

We classify shareholders who are founders, former top executives, former directors, and/or current directors of a firm as “quasi-insiders.” While these shareholders have little or no formal control of a firm to which they are connected, they have knowledge of the firm’s inner workings and relationships with insiders and long-time shareholders that may give them an advantage in activism campaigns. They may also often own significant equity stakes. These shareholders may be motivated to become activist by a perception, based on their knowledge of the firm, that the firm is following a suboptimal strategy. They also may be motivated by concerns about their legacies, a desire to reassert control, and/or ego. How frequently do quasi-insiders become activist? What types of firms do they target? What are their objectives and tactics? How often and in what circumstances do they succeed in achieving their objectives? How are their campaigns perceived by other investors? What are the long-term consequences of their campaigns? What types of firms have quasi-insiders who could potentially become activist? This paper seeks to shed light on the answers to these and related questions.

Using a combination of shareholder activism data from FactSet and a manual search through 13D SEC filings, we identify 280 public campaigns launched by quasi-insiders between 1995 and 2021. Collectively, these campaigns involve 327 quasi-insiders. Of these, 37.6% are former CEOs, 29.4% founders, 28.8% former board chairs, 21.1% former directors, 16.5% former non-CEO executives, and 33.6% current directors.² In some cases, quasi-insiders cooperate with traditional activist investors, such as hedge funds, in campaigns. For example, Pershing Square, a hedge fund, launched a campaign at J.C. Penney in 2013 to bring back former CEO Allen Questrom as CEO and Board Chair. While the total number of public quasi-insider campaigns is modest, these observable campaigns likely represent the tip of the iceberg in terms of interventions by quasi-insiders. Because of their connections within the firm, quasi-insiders may be better-positioned than true arm’s-length shareholders to induce changes, without the need for an expensive public campaign.

One unique feature of quasi-insider campaigns is that the match between the activist and target firm is effectively predetermined. While traditional activists, such as hedge funds, choose which firms to target, a quasi-insider is, by definition, only linked to a firm with which she has a current or prior relationship. Thus, a quasi-insider does not choose which firm to target in a campaign but rather whether to initiate a campaign at the specific firm with which she already has a connection. Firms at which quasi-insiders launch activism campaigns tend to be smaller than the average Compustat firm in the same industry. This tendency is consistent with a greater cost of initiating a campaign at a larger firm (Brav et al. 2008). These firms also tend to have low valuations, as measured by Tobin’s q, and poor recent performance, as measured by either return-on-assets or stock returns, relative to other firms in the same industry. In addition, they are disproportionately in struggling industries. Quasi-insider activist campaigns, then, tend to target firms where at least the perceived scope for a potential turnaround is high.

An activist shareholder in general may seek a specific one-time action, such as payment of a dividend or divestiture of assets, generic improvements in value maximization through unspecified means, or some degree of ongoing control through the appointment of activist-affiliated directors to the firm’s board or a hostile acquisition. Quasi-insiders typically seek control in campaigns rather than specific actions or generic value maximization. They seek at least some board representation 62.1% of the time and full control of the board 31.8% of the time. In addition, they seek a sale to themselves another 4.3% of the time. The fact that they seek control in so many campaigns suggests that they often see themselves as better able to set the firm’s direction than incumbent management. Quasi-insider activists also frequently use aggressive tactics, such as writing public letters to the board or to shareholders and, in some cases, filing lawsuits and calling for special shareholder meetings as part of their campaigns.

Quasi-insider campaigns often succeed in achieving at least some of their objectives. The success rate among the 280 campaigns in our sample is 43.6%. Campaigns seeking board control have the highest success rate, achieving their objectives 51.1% of the time. The likelihood of success increases with the activist’s ownership stake. This is not surprising, as a larger stake gives the quasi-insider more voting rights and may also make a campaign more credible in the eyes of management and other shareholders. The likelihood of success decreases with the target firm’s stock return over the year prior to the initiation of the campaign. Thus, it appears that shareholders are more willing to side with an activist quasi-insider when recent performance raises doubts about the competence of current management.

To contextualize these findings further, we compare campaigns initiated by quasi-insiders to those initiated by hedge funds. Quasi-insider activists are more likely to seek board representation than hedge fund activists and much more likely to seek board control (32.1% of quasi-insider campaigns versus 7.3% of hedge fund campaigns). In contrast, hedge funds are much more likely to seek general shareholder value maximization as an objective. Quasi-insider campaigns are also more aggressive. Quasi-insiders are more likely to file lawsuits, call for special shareholder meetings, and send public letters to shareholders as part of their campaigns than hedge fund activists are. The firms that quasi-insiders target in campaigns tend to be smaller and have had weaker recent performance than those that hedge funds target. The size difference suggests that quasi-insiders target firms activist hedge funds may avoid because of the cost of accumulating a stake in a smaller, less liquid firm (Kahn and Winton 1998; Maug 1998) - a cost that a quasi-insider who already owns a stake can avoid. The difference in recent performance could indicate that quasi-insiders face a higher cost of launching a campaign and therefore only launch a campaign when performance has deteriorated precipitously.

After analyzing the characteristics of quasi-insider campaigns and the firms involved, we next analyze the financial implications of these campaigns. The mean cumulative abnormal return (CAR) from 10 days prior to a campaign announcement to the day after announcement is a statistically significant 3.9%. As with campaigns initiated by other activists (Boyson and Mooradian 2011), targets in the small number of campaigns where the quasi-insider activist’s objective involves forcing a sale of the firm experience the largest CARs, though CARs are positive and statistically significant in other campaigns as well. CARs are smaller when insiders in the firm own a larger stake, which may reflect greater difficulty in achieving campaign objectives that current insiders oppose. CARs show little correlation with other observables, including recent firm performance and the size of the activist’s stake in the firm.

The positive announcement returns suggest that the market may anticipate increases in cash flow subsequent to quasi-insider campaigns, perhaps as a result of improvements in operating performance. Next, we examine changes in operating profits (EBITDA/Total assets) from the year before to the 2 years after campaigns, relative to firms matched on industry, size, and precampaign performance. On average, operating profits decrease by 0.2 percentage points from the year prior to the campaign to the first year after and increase by about 0.5 percentage points from the year prior to the second year after, relative to matched firms. However, operating performance is extremely noisy, and the changes are not statistically significant. The increases in the first and second years after a campaign are larger and positive for campaigns with positive CARs, whereas they are negative for campaigns with nonpositive CARs, but, again, the changes are not statistically significant. In the end, because the standard deviation of changes in operating profits is so large, we are unable to discern much about the long-run consequences of quasi-insider activism campaigns.

We conduct two additional forms of analysis using the subsample of quasi-insiders who were previously CEOs. The advantage of focusing on former CEOs is that we can observe information about their employment stints, including their departure dates. First, we examine the circumstances in which former CEOs who initiate activism campaigns departed the target firm. Among former CEOs who subsequently launch activism campaigns, the fraction who departed involuntarily is more than three times the fraction of CEO departures in general that are involuntary as documented by Parrino (1997). Moreover, recent stock returns and operating performance in the year prior to departure are substantially worse for former CEOs who subsequently launch campaigns than market and industry benchmarks. These findings suggest that individuals who initiate quasi-insider campaigns are not stellar performers.

Finally, we take a step back and examine the prevalence of quasi-insiders who could potentially become activist in the future and which firms they tend to target in campaigns, focusing again on former CEOs. We focus more specifically on former CEOs who own at least 5% of their former employer’s stock since ownership implies at least some ongoing connection to the firm. We identify 687 former CEOs in 621 firms who own at least 5% of their former employer’s stock at some point during our sample period. We find that these former CEOs tend to hold stakes in larger firms with good performance, but, conditional on having a stake, tend to target smaller firms with poor performance. These findings suggest that the tendency of quasi-insider campaigns to involve primarily smaller, poorly performing firms is a function of selective targeting rather than the types of firms in which quasi-insiders are present.

Our paper adds to the literature on shareholder activism (see Denes, Karpoff, and McWilliams 2017 for a recent survey). Shareholder activism has become an increasingly influential force in corporate governance. Most of the literature on activism focuses on activism campaigns initiated by hedge funds (Brav, Jiang, and Kim 2010; Brav et al. 2008; Klein and Zur 2009).³ Our results suggest that individuals who are not in positions of control in a firm but have a prior or current connection with the firm sometimes engage in activism as well and often do so aggressively. However, we do not see evidence that this activism meaningfully improves firm performance, at least in the short run.

Our paper also adds to the large literature on blockholder governance (for surveys, see Edmans, 2014; Edmans and Holderness, 2017). Cronqvist and Fahlenbrach (2008) document significant heterogeneity in the importance of different blockholders in explaining differences in firm policies and performance. Among other factors, they find that blockholders with a larger block size, board seats, and direct management involvement are more influential. Becker, Cronqvist, and Fahlenbrach (2011) find that blockholders influence firms, using geographic variation in blockholder location to separate selection from treatment effects. Agrawal (2012) finds that union-affiliated blockholders may reduce firm value (see also Ertimur, Ferri, and Muslu 2010). Our paper specifically identifies former insiders as potentially important blockholders. Existing corporate governance research often explicitly excludes these agents when studying the role of external governance providers (Clifford and Lindsey 2016; von Lilienfield-Toal and Schnitzler 2015). Hadlock and Schwartz-Ziv (2019) find that blockholders tend to crowd each other out, which may make quasi-insider blockholders, who typically hold ownership stakes because of their prior involvement with the firm, especially influential. Our evidence suggests that these blockholders often play an active role in the firms in which they are present.

Finally, our paper adds to the literature examining the role of former CEOs specifically in corporate control. Fahlenbrach, Minton, and Pan (2011) find that firms with former CEOs on their boards experience better accounting performance. In contrast, Evans, Harry, Nagarajan, and Schloetzer (2010) find lower long-run stock price performance after an outgoing nonfounder CEO ascends to the board. Andres, Fernau, and Theissen (2014) find that German firms whose former CEO serves on the supervisory board pay their current CEO more, though they also find a positive announcement return when a retiring CEO transitions to the supervisory board. These papers study board membership, an internal source of governance. Our paper adds to this literature by examining a broader set of former insiders and focusing on activism, an external source of governance. Our conclusions are mixed, with evidence of a positive stock price response to activism campaigns launched by these individuals but inconclusive evidence of improvements in profitability, at least in the short run.

1. Data and Sample

Our empirical analysis of quasi-insider intervention takes two forms. We first analyze shareholder activism campaigns and then study the consequences of having a former CEO as a blockholder. To implement this analysis, we construct two samples. The first sample consists of shareholder activism campaigns initiated by quasi-insiders. The second sample takes the form of a panel of firm-years, within which we identify firm-years in which a firm has a former CEO who owns a substantial block of the firm’s shares.

1.1 Quasi-insider activism campaigns

We define a quasi-insider as an individual who is not a current executive or board chair but is a founder, former top executive, former chair, former director, or current director. We build a sample of quasi-insider-initiated activism campaigns. We identify campaigns involving quasi-insiders primarily using FactSet’s SharkWatch corporate activism database. This database contains 11,940 shareholder activism campaigns as of February 1, 2021 and has been used as a basis for other recent studies of shareholder activism (e.g., Appel, Gormley, and Keim 2019; Francis, Hasan, Shen, and Wu 2012). FactSet identifies activism campaigns through a combination of SEC filings and news sources. Thus, the activism campaigns we analyze are those that reach the level of being public and do not include those that take place behind the scenes.

Other studies of shareholder activism use 13D filings to identify campaigns (e.g., Brav, Jiang, and Kim 2010; Brav et al. 2008). We use FactSet rather than 13D filings to identify quasi-insider campaigns for two reasons. First, many quasi-insiders hold less than 5% of the target firm’s stock and are thus not required to file a 13D filing. Second, many quasi-insiders who file 13Ds file their original 13D while they are still insiders.⁴ The result is tens of thousands of individual 13D filers, only a small fraction of whom are likely to engage in activism in a traditional sense, and a time that likely does not correspond to the original 13D filing in many cases. As a result, classifying any 13D filing by a quasi-insider as an activism campaign would result in a large number of false positives. Nevertheless, we use 13D filings to augment the SharkWatch database, as we will describe shortly.

FactSet provides a detailed synopsis for each campaign in its database. We read the synopsis for each campaign, look for associated 13D filings, and conduct extensive Google searches to determine whether an individual meeting our definition of a quasi-insider was involved in the campaign. Altogether, this process yields 265 unique campaigns, of which 247 were launched by a quasi-insider and 18 were launched by a hedge fund but involved a quasi-insider.⁵

While the SharkWatch campaign data appears fairly comprehensive, we nevertheless supplement this data by using 13D filings in the SEC’s EDGAR database to identify quasi-insider activism campaigns not in the SharkWatch database. We start with all 277,315 13D filings in EDGAR with filing dates between January 1, 2000, and December 31, 2020. Within this set, we identify filings potentially made by individuals by dropping any filing for which the primary filer name field contains terms, such as “Capital,” “LLC,” “LP,” “Partners,” or “Trust” that suggest that the filer is not an individual.⁶ This filtering leaves 98,642 13D filings potentially made by individuals. We then use Capital IQ’s People Intelligence database (Capital IQ hereafter), which contains start and end dates for each executive and director role that an individual has had in a given firm, to identify primary filers who are current directors, former CEOs, former officers, former board chairs, or founders.⁷ We link firms that are subject to the 13D filings with Capital IQ using their CIKs. We link the individuals filing 13Ds to individuals in Capital IQ using a fuzzy merge based on their names, which we hand-check for accuracy. This process yields 4,502 unique individual-company pairs with a 13D filing.

For each of these individual-company pairs, we retain all 13D filings, including amendments, from EDGAR until the end of 2020. This filtering yields a list of 31,741 filings. As noted, only a small fraction of individual filers is likely to ever become activist. We therefore read Item 4 (Purpose of Transaction) for all filings made by quasi-insiders and identify 13D filings that with activist requests.⁸ Campaigns identified using this approach fit the definition of campaigns in our sample, which involve at least some publicly observed indication of an activist role. This process yields an additional 15 campaigns initiated by quasi-insiders that are not included in FactSet, bringing our total sample of quasi-insider campaigns to 280, involving 327 separate quasi-insiders. That this process results in so few additional campaigns attests to the comprehensiveness of the SharkWatch database.

For each of the campaigns in our sample, we collect information about the role of the quasi-insider from FactSet, 13D filings, Capital IQ, and Google searches. We collect information about the campaign type (proxy fight, exempt solicitation, or other stockholder campaign) and the objective of the campaign based on the objective categories provided by FactSet. For the 265 campaigns in FactSet, we collect this information directly from FactSet. For the 15 campaigns not in FactSet, we collect type information from the 13D filing and assign the objective based on Item 4 of the 13D. We also collect information from FactSet about the tactics that the activist uses for the 265 campaigns in FactSet. In addition, we collect information about the success of each campaign in achieving its stated objectives from the FactSet campaign synopsis and news articles about the outcome of the campaign. We classify a campaign as successful if the firm implemented at least one of the activist’s stated objectives. Based on this definition, 43.6% of campaigns achieve success.

FactSect provides a CUSIP for each target firm in the SharkWatch data, and 13D filings provide a CIK for the firm to which the filing is related. We are able to match 255 of the 280 firms subject to quasi-insider campaigns in our sample to Compustat based on CUSIP and CIK. We use the Compustat-CRSP link file to match each firm in our sample to CRSP, from which we obtain stock return data. We match each firm based on CUSIP to 13F holdings data from Thomson Reuters to obtain information about institutional ownership, correcting for known errors in the holdings data.⁹ We obtain information about the activist’s ownership from FactSet, which provides a campaign text synopsis that frequently includes this information, and from 13D filings. We hand-collect information about insider ownership for each firm from the most recent 10-K filing prior to the campaign. Finally, for all quasi-insiders in our sample who are former CEOs, we attempt to identify the CEO’s departure date in Capital IQ. We then attempt to determine whether the departure was voluntary or forced using the FactSet campaign synopsis, where available, and Google searches. Table A1 in the appendix defines all variables.

1.2 Former CEO blockholders

To construct our second sample, we start with 11,718 13D filings and amendments made by former CEOs out of the 31,741 quasi-insider 13D filings identified above. We then identify 13G filings and amendments made by former CEOs, match these to Capital IQ, and retain only those filed by former CEOs between the time that they become quasi-insiders and December 31, 2020. This process yields a sample of 10,919 13G filings. We add these to our sample of 13D filings to create a sample of 22,637 13D and G filings and amendments made by former CEOs. We focus on former CEOs because we need comprehensive end dates for the individuals, and end dates in Capital IQ for individuals with other prior roles (e.g., non-CEO executives) are frequently missing.

Since our objective is to build a panel in which we can identify firm-years with a former CEO blockholder, we need to determine whether a former CEO is a blockholder in each individual year. Determining whether an individual is a 5% blockholder at a specific point in time is challenging. A shareholder is required to file either an initial form 13D or 13G with the SEC after obtaining a holding of 5% or more of a publicly listed company’s stock. The shareholder is then required to file an amended 13D or 13G when there is a change in either the ownership level of greater than 1% relative to the most recent filing or when the shareholder’s intentions change. In theory, shareholders are also required to file a final 13D/G amendment when their ownership stake falls below 5%. However, anecdotal evidence and discussions with regulators suggest that filers often neglect to file a terminal 13D/G amendment, making it difficult to determine when a blockholder ceases to be a blockholder.¹⁰

As a conservative approach, we identify a former CEO as a blockholder in a given year if two criteria are satisfied: (1) Capital IQ reports a CEO role for the individual with an end date prior to the year in question and (2) the individual files a 13D or 13G (or amendment) in the year of or any year subsequent to the year in question. For each firm-year from 2000 through 2020, we define an indicator variable QIBlockholder, which equals one if the firm has a former CEO blockholder based on our definition in that year and zero otherwise. This approach yields 2,221 firm-years in which a former CEO is a blockholder (i.e., QIBlockholder = 1), with 687 former CEOs in 672 unique firms. Because our approach is conservative, we set QIBlockholder to zero for likely many firm-years in which the firm has a former CEO who is, in fact, a blockholder. Of the 122 campaigns involving former CEOs in the quasi-insider campaign sample, 101 are initiated in firm-years for which QIBlockholder = 1 in our firm-year panel. The remainder are campaigns where the former CEO’s holding is below the 5% threshold for filing a 13D/G or that occur after the last 13D/G filing.

2. Results

2.1 Quasi-insider relationships

Table 1 reports the nature of the quasi-insiders involved in campaigns in our sample. Categories of quasi-insiders are founder, former CEO, former president, former other executive, former board chair, former (nonchair) director, and current director. Note that these categories are not mutually exclusive. Some campaigns involve multiple quasi-insiders, and some individuals fit in multiple categories.

Table 1

Quasi-insider activists’ relationships with target firms

A. Campaign level
	N	% of quasi-insider campaigns
Founder	90	32.1
Former CEO	122	43.6
Former president	61	21.8
Former other executive	50	17.9
Former chair	94	33.6
Former director	60	21.4
Current director	91	32.5
Total	280

B. Quasi-insider level
	N	% of Quasi-Insider Individuals
Founder	96	29.4
Former CEO	123	37.6
Former president	61	18.7
Former other executive	54	16.5
Former chair	94	28.8
Former director	69	21.1
Current director	110	33.6
Total	327

This table summarizes the relationships of quasi-insider activists with the target firms. The sample consists of activist campaigns obtained from FactSet SharkWatch and 13D filings for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. There are 327 quasi-insiders who participate in 280 quasi-insider activist campaigns. Information on the activists’ relationships to target firms is obtained from FactSet campaign synopses, Capital IQ, SEC 13D and proxy filings, and web searches. Panel A reports the relationship breakdown at the campaign level, and panel B reports the relationship breakdown at the quasi-insider level. The relationship classifications are not mutually exclusive because quasi-insiders may have multiple relationships with a firm and a campaign may include multiple quasi-insiders.

Panel A reports the breakdown by campaign across the 280 campaigns in our sample. Most of the campaigns involve individuals who once held substantial direct control over the target company but no longer do. Of the campaigns, 43.6% involve former CEOs, 33.6% former board chairs, and 32.1% founders. These individuals are likely to at least perceive themselves to be well-informed about factors affecting the target firm’s optimal strategic direction. They are also likely to be well-connected to executives within the firm, members of the board of directors, and long-time institutional shareholders, and to own stakes in the firm. In addition, they may be concerned about their legacies, which may prompt them to act when they perceive current management to be making decisions they believe to be suboptimal.

Panel B reports the breakdown by the 327 individual quasi-insiders in our sample. Patterns here are similar to those in panel A, with former CEOs, former board chairs, and founders representing 37.6%, 28.8%, and 29.4% of the quasi-insiders in our sample, respectively. Noteworthy is the fact that 33.4% of quasi-insiders are current directors. Of the 110 quasi-insiders who are current directors, 24.6% are former CEOs, 16.4% are former board chairs, and 28.2% are founders. The combination of prior direct control and continued presence on the board seems likely to make an individual feel especially well-positioned to reassert control if they perceive current management to be making suboptimal decisions.

2.2 Quasi-insider campaign objectives

Table 2 reports the breakdown of campaign types, objectives, and tactics. Panel A reports the breakdown of campaign types. The majority (57.9%) of quasi-insider campaigns are categorized as proxy fights, in which the dissident shareholder nominates directors to run against directors nominated by management and engages in proxy solicitation, soliciting all shareholder to vote for the dissident’s nominees. Another 3.2% are categorized as exempt solicitations. These campaigns also involve the nomination of dissident directors, but the dissident in these campaigns solicits 10 or fewer shareholders, making it exempt from the SEC’s proxy solicitation rules. Note that not every campaign involving a proxy contest ends in a vote on competing slates of directors, since the dissident may withdraw the nominations prior to shareholder vote. Withdrawal sometimes occurs because the firm agrees to grant board seats or other concessions to the dissident as a form of settlement.

Table 2

Frequency of quasi-insider campaign type, objectives, and tactics

A. Campaign type
	Quasi-insiders
	N	%
Proxy fight	162	57.9%
Exempt solicitation	9	3.2%
Other stockholder campaign	109	38.9%

B. Campaign objectives
Maximize shareholder value	43	15.4%
Board representation	84	30.0%
Board control	90	32.1%
Hostile/unsolicited acquisition	12	4.3%
Other specific requests
Enhance corporate governance	10	3.6%
Remove director(s)	4	1.4%
Remove officer(s)	6	2.1%
Support dissident group in proxy fight	3	1.1%
Vote against a management proposal	9	3.2%
Vote for a stockholder proposal	8	2.9%
Vote/activism against a merger	11	3.9%
Total	280

C. Tactics
	N	%
Binding proposal	34	12.8%
Board letter	117	44.2%
Call meeting	22	8.3%
Lawsuit	42	15.8%
Stockholder letter	114	43.0%
Written consent	27	10.2%

This table summarizes the type of campaigns launched by quasi-insider activists (panel A), the objectives of the activists (panel B), and the tactics employed (panel C). The sample consists of 280 activist campaigns obtained from FactSet and 13D filings for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. The type of campaign is classified by FactSet. Campaign objectives are classified based on FactSet primary campaign objectives and Item 4 of SEC 13D filings. The first column indicates the category of the objective and the second column indicates the specific objective. Campaigns with more than two main objectives are classified as General Value. Campaign tactics are classified by FactSet.

Panel B reports the breakdown of campaigns by primary objective. We rely here on the categories of objectives defined by FactSet. These categories are general value maximization, board representation, board control, sale-related, and a number of categories that involve requests for specific actions, such as an increase in leverage or the spin-off of an business unit, that we lump together into an “other specific requests” category. The second, third, and fourth categories all involve the quasi-insider seeking some degree of ongoing control, either through the board of directors or through ownership of the firm. Campaigns seeking general value maximization involve an attempt neither to gain direct control nor to induce specific actions.

A substantial majority of quasi-insider campaigns seek at least some degree of ongoing control, with 30.0% seeking board representation (but not full control), 32.1% seeking board control, and 4.3% seeking sale of the target firm. Cases in which the quasi-insider activist seeks specific actions (18.2%) or general value maximization (15.4%) are less common. The fact that most quasi-insider campaigns involve efforts to gain at least some degree of ongoing control over the target is consistent with the nature of quasi-insiders. These activist shareholders are more likely than true outside dissident shareholders to at least believe that they have the target firm-specific expertise necessary to make better strategic and operating decisions than current management.

Panel C reports a breakdown of the tactics that quasi-insider activists use in their campaigns. Quasi-insider activists employee a broad variety of aggressive tactics. They frequently send public letters to the board of directors (44.2% of campaigns) or to shareholders directly (43.0% of campaigns). Activists typically send such letters to put pressure on the board to adopt proposed changes or to garner shareholder support for campaigns. Quasi-insider activists also sometimes file lawsuits (15.8% of campaigns), call special shareholder meetings (8.3% of campaigns), and request that shareholders be able to vote via written consent (10.2% of campaigns).

2.3 Characteristics of quasi-insider campaign targets

Of the 280 firms targeted in quasi-insider campaigns, 255 have nonmissing Compustat total assets as of the fiscal year-end prior to the initiation of the campaign. Table 3 reports the mean, median, and standard deviation of various characteristics for the year prior to the campaign for these 255 firms. We winsorize all variables at the 1st and 99th percentiles to address concerns about possible outliers. The table also reports these values for the median Compustat firm in each target’s three-digit SIC code in the same year for the sake of comparison.¹¹

Table 3

Quasi-insider campaign target summary statistics

	N	Mean	Median	SD	SIC3 Median	median diff
Total assets	255	3,338	138	14,385	180	−42**
log(Total assets)	255	5.012	4.927	2.480	5.193	−0.266**
Market cap	249	1,697	84	6,147	140	−56***
log(Market cap)	249	4.775	4.441	2.139	4.954	−0.513***
Tobin’s q	248	2.694	1.286	7.405	1.545	−0.259***
Market-to-book equity	248	1.733	1.326	8.366	1.609	−0.283**
Cash	255	0.216	0.129	0.234	0.122	0.007
R&D	255	0.064	0.000	0.164	0.000	0.000
Capital expenditures	255	0.041	0.022	0.052	0.021	0.001
Dividend yield	255	0.014	0.000	0.027	0.000	0.000***
Debt	255	0.253	0.126	0.477	0.138	−0.012
ROA	255	−0.327	−0.028	1.083	0.002	−0.030***
Stock return	196	−0.102	−0.160	0.480	0.002	−0.162***
Institutional ownership	211	0.430	0.372	0.530	0.496	−0.124***
Activist ownership	237	0.160	0.112	0.138
Insider ownership	234	0.187	0.137	0.171

This table reports summary statistics of characteristics of firms targeted by quasi-insider activists as well as industry comparisons. The sample consists of activist campaigns obtained from FactSet SharkWatch and 13D filings for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. The sample is restricted to 255 campaigns for which data on firm characteristics are available in Compustat in the fiscal year prior to the campaign. The table also and reports the median of the distributions for the median of each characteristic in the same year for all firms in the same three-digit SIC code as each quasi-insider target. Table 13 defines all variables.

*

p <.1;

**

p <.05;

***

p <.01 (for Wilcoxon signed-rank tests that compare medians of firm-years with former CEO blockholders with and without former CEO activist campaigns).

Firms targeted in quasi-insider campaigns tend to be significantly smaller than the average firm in the same industry. The median targeted firm has total assets of $138M, while the median firm in the same-industry comparison group has total assets of $180M. This difference is statistically significant based on a Wilcoxon signed-rank test. Similarly, firms subject to quasi-insider activism campaigns have lower median logged assets. These differences are consistent with a greater cost of initiating a campaign at a larger firm (Brav et al. 2008).

Targeted firms also tend to exhibit relatively poor recent performance as measured by return-on-assets over the fiscal year prior to the campaign and stock returns over the calendar year prior to the campaign. Median ROA for targeted firms is negative, and is 0.03-percentage-points lower than the median for firms in the same industry. Median stock return in targeted firms over the year prior to the campaign is −16.0%, 16.2 percentage points less than the median for firms in the same industry. Targeted firms also have a lower median Tobin’s q, suggesting lower valuations.

It is worth noting that the industries of firms targeted by quasi-insiders tend to exhibit relatively poor recent performance themselves. Industry median ROA is barely positive, at 0.002. Similarly, industry median stock return over the year prior to the campaign is 0.2%. By comparison, the annual return on the S&P 500 over the period 1994–2019 is 11.5%.¹² This industry-level weakness suggests that the industries in which quasi-insiders become activist are experiencing dislocations. Quasi-insiders may at least perceive that their experience is especially valuable for firms in industries experiencing such dislocations.

2.4 Quasi-insider campaign success

Next, we examine the factors that predict the success of quasi-insider activism campaigns. We first examine differences in the probability of success by campaign objective. Table 4 reports these probabilities. 43.6% of all quasi-insider campaigns achieve success. The success rate is higher in campaigns in which the activist seeks board representation, at 45.2%, and is highest in campaigns in which the activist seeks full board control, at 51.1%. Campaigns seeking specific actions have the lowest success rate, at 35.3%.

Table 4

Frequency of quasi-insider campaign success

	N	No. successful	% successful
All	280	122	43.6%
By objective
General value	43	15	34.9%
Board representation	84	38	45.2%
Board control	90	46	51.1%
Sale related	12	5	41.7%
Other specific requests	51	18	35.3%

This table reports data on the success of activist campaigns for quasi-insider activists. The sample consists of activist campaigns obtained from FactSet SharkWatch and 13D filings for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. A campaign is classified as being successful if the activist achieves its stated objectives, according to information in the FactSet synopses and press reports. Success rates are reported for all campaigns as well as separately by objective. Campaign objectives are classified using information from FactSet campaign synopses and SEC 13D and proxy filings (see panel B of Table 2).

Next, we estimate a linear probability model where the dependent variable is an indicator equal to one if a campaign is successful and zero otherwise. The dependent variables are various campaign, firm, and ownership characteristics. Table 5 presents the results of these regressions.

Table 5

Quasi-insider campaign success regressions

	Success	Success	Success	Success
	(1)	(2)	(3)	(4)
General value	0.047			−0.080
	(0.43)			(−0.56)
Sale related	0.113			0.087
	(0.64)			(0.43)
Board control	0.164			0.066
	(1.62)			(0.45)
Board representation	0.109			0.087
	(1.12)			(0.73)
log(Total assets)		−0.012		−0.016
		(−0.55)		(−0.55)
Stock return		−0.208**		−0.238**
		(−2.26)		(−2.09)
Tobin’s q		0.001		0.016
		(0.03)		(0.48)
Cash		−0.107		−0.105
		(−0.51)		(−0.39)
ROA		−0.181		−0.091
		(−1.00)		(−0.45)
Dividend yield		−1.567		−1.228
		(−1.10)		(−0.80)
Debt		−0.115		−0.169
		(−0.85)		(−1.09)
Activist ownership			0.614*	0.655*
			(1.92)	(1.85)
Insider ownership			−0.159	−0.129
			(−0.55)	(−0.40)
Institutional ownership			−0.059	−0.056
			(−0.66)	(−0.55)
Year fixed effects	Yes	Yes	Yes	Yes
Observations	252	196	185	167
Adjusted R-squared	.096	.167	.112	.210

This table reports results from a linear probability model where the dependent variable is equal to one if a campaign was successful and zero otherwise. A campaign is classified as being successful if the activist achieves its stated objectives, according to information in the FactSet synopses and press reports. The explanatory variables in column 1 are campaign objectives, with Other Specific Requests as the omitted category. The explanatory variables in column 2 are firm characteristics. The explanatory variables in column 3 are ownership variables. Column 4 includes all variables. All specifications include year fixed effects. Table 13 defines all variables. Standard errors are reported in parentheses.

*

p <.1;

**

p <.05;

***

p <.01.

The explanatory variables in column 1 are indicator variables for each campaign objective. The omitted objective is Other Specific Requests. The positive coefficients in column 1 indicate that success is more likely for all campaign objectives than for Other Specific Requests. The success rate is highest when the campaign objective is Board Control, for which success is 16.4 percentage points more probable than for Other Specific Requests. This difference is large, considering that the unconditional probability of success is 43.6%. However, none of the coefficients in column 1 are statistically significant, though the Board Control coefficient is almost significant at the 10% level (t-stat of 1.62). So, while we cannot draw strong conclusions, it appears that shareholders may be more likely to support a quasi-insider activist when the activist seeks outright control of the target firm.

The explanatory variables in column 2 are firm characteristics. Among the seven firm characteristics included in column 2, only stock return over the past year has explanatory power over campaign success probability at a statistically significant level. The coefficient for the stock return of -0.208 implies that a one-standard-deviation higher stock return over the year prior to the campaign (48.0%) is associated with a 9.6-percentage-point lower probability of campaign success. Campaign success probability also decreases with ROA, though not at a statistically significant level. Overall, it appears that shareholders are significantly more likely to support a quasi-insider campaign when the target firm is struggling.

The explanatory variables in column 3 are ownership characteristics, including the quasi-insider activist’s ownership percentage, insiders’ ownership, and institutional ownership. The relationship between campaign success probability and the quasi-insider activist’s ownership is positive and statistically significant at the 10% level. The 0.614 coefficient for activist ownership implies that a one standard deviation higher level of activist ownership (13.8%) is associated with an 8.5-percentage-point higher probability of campaign success. A larger ownership stake increases the quasi-insider activist’s voting power in a proxy contest or other shareholder vote. It also likely gives the activist more leverage with management and credibility with other shareholders. The relationships between success probability and insider and institutional ownership are statistically insignificant.

Finally, column 4 includes all of the explanatory variables from columns 1 through 3. Campaign success probability continues to be negatively related to stock return over the past year and positively related to activist ownership. That so few variables predict campaign success probability and that the 14 variables in column 4 only explain 21.0% of total variation in success probability suggest that most of the factors affecting success probability are unobservable. These factors likely include the nature of behind-the-scenes interactions between the activist and management and between the activist and other shareholders, the reputations of the activist and management, and the nature of the shareholder base more generally.

2.5 Hedge fund comparison

To provide further context for the activities of quasi-insider activists and the firms involved, we compare quasi-insider campaigns with 2,969 activist campaigns that FactSet flags as initiated by hedge funds between 1995 and February 1, 2021, that do not include quasi-insiders. Table 6 presents this analysis. Panel A compares activism type; panel B campaign objectives; panel C campaign tactics; and panel D target characteristics.

Table 6

Quasi-insider and hedge fund campaign comparisons

A. Activism type	QI %	HF %	Difference	t-stat
Exempt solicitation	3.2	1.8	1.4%*	(1.67)
Other stockholder campaign	38.9	74.0	−35.1%***	(−12.66)
Proxy fight	57.9	24.2	33.7%***	(12.38)

B. Campaign type	QI %	HF %	Difference	t-stat
Board control	32.1	7.3	24.8%***	(13.94)
Board representation	30.0	34.8	−4.8%	(−1.61)
Enhance corporate governance	3.6	2.9	0.7%	(0.67)
Hostile/unsolicited acquisition	4.3	1.3	3.0%***	(3.83)
Maximize shareholder value	15.4	33.5	−18.1%***	(−6.26)
Public short position	0.0	3.8	−3.8%***	(−3.33)
Remove director(s)	1.4	0.7	0.7%	(1.23)
Remove officer(s)	2.1	0.5	1.6%***	(3.27)
Seats granted, no pub actvsm	0.0	0.0	0.0%	(−0.31)
Support diss grp in proxy ft	1.1	2.0	−0.9%	(−1.07)
Vote against mgmt proposal	3.2	2.4	0.8%	(0.85)
Vote for mgmt proposal	0.0	0.9	−0.9%	(−1.63)
Vote for stockholder proposal	2.9	2.3	0.6%	(0.60)
Vote/activism against merger	3.9	7.5	−3.6%**	(−2.20)

C. Tactics	QI %	HF %	Difference	t-stat
Binding proposal	12.8	2.3	10.5%***	(9.70)
Board letter	44.2	39.9	4.3%	(1.37)
Call meeting	8.3	1.8	6.5%***	(7.00)
Lawsuit	15.8	4.4	11.5%***	(8.17)
Stockholder letter	43.0%	12.1%	30.9%***	(14.25)
Written consent	10.2	1.2	9.0%***	(10.83)

D. Corporate	QI mean	QI median	HF median	Median diff
Total assets	3,338	138	554	−416***
log(Total assets)	5.012	4.927	6.317	−1.39***
Market cap	1,697	84	321	−237***
log(Market cap)	4.775	4.441	5.771	−1.33***
Tobin’s q	2.694	1.286	1.239	0.047
Market-to-book equity	1.733	1.326	1.379	−0.053
Cash	0.216	0.129	0.097	0.032*
R&D	0.064	0.000	0.000	0.000
Capital expenditures	0.041	0.022	0.023	−0.001
Dividend yield	0.014	0.000	0.000	0.000
Debt	0.253	0.126	0.195	−0.069**
ROA	−0.327	−0.028	0.005	−0.033***
Stock return	−0.104	−0.162	−0.033	−0.129***
Institutional ownership	0.406	0.393	0.693	−0.300***
Activist ownership	0.160	0.113	0.079	0.034***

This table reports summary statistics of characteristics of firms targeted by quasi-insider and hedge fund activists. The sample consists of targets of activist campaigns obtained from FactSet SharkWatch for the period 1995–2021. Panel A details the activism type. Panel B details the primary campaign objective as detailed by FactSet. Panel C details the tactics employed according to FactSet. Panel D details summary statistics for stock returns, Compustat variables, and ownership variables. Table 13 defines all variables.

*

p <.1;

**

p <.05;

***

p <.01 (for Wilcoxon signed-rank tests that compare medians of firm-years with former CEO blockholders with and without former CEO activist campaigns).

Several differences between the two samples are worth noting. First, quasi-insider activists are far more likely to seek at least some board representation than hedge fund activists (62.1% of quasi-insider campaigns versus 42.1% of hedge fund campaigns) and especially more likely to seek full board control (32.1% of quasi-insider campaigns vs. 7.3% of hedge fund campaigns). In contrast, hedge funds are much more likely to seek general shareholder value maximization. Quasi-insider campaigns are also much more likely to involve formal proxy fights.

Second, in addition to seeking more direct control in their campaigns, quasi-insider activists tend to employ more aggressive tactics. Quasi-insiders file lawsuits in 15.8% of their campaigns, while hedge funds file lawsuits in only 4.4% of their campaigns. Quasi-insiders are also more likely than hedge funds to call for a special shareholder meeting (8.3% versus 1.8%), send public letters to shareholders (43.0% versus 12.1%), and request written consent for votes (10.2% versus 1.2%). One explanation for the relatively aggressive tactics of quasi-insider activists is that they are only likely to own a large stake in the firm to which they are connected and therefore do not need worry as much about their public reputations as hedge funds do. Another is that they seek more control than hedge funds do in their campaigns and therefore may need to use more aggressive tactics to support their objectives. A third possibility is that they are more emotionally invested in their campaigns, since these campaigns involve firms with which they already have relationships.

Third, quasi-insiders tend to target different types of firms than hedge funds target. The median quasi-insider campaign target is less than one-fourth of the size of the median hedge fund target. These differences suggest that quasi-insiders play an active role in firms that may be too small for hedge fund activists to bother targeting. Quasi-insider campaign targets also exhibit weaker recent performance in terms of both ROA and stock return relative to hedge fund targets. Thus, it appears that quasi-insiders wait until a firm’s condition has deteriorated to a greater degree before attempting to intervene. This difference in thresholds is consistent with quasi-insiders, who do not regularly engage in activism campaigns, facing higher costs of intervening and therefore waiting until performance is worse before doing so.

Fourth, quasi-insiders tend to own a larger fraction of the shares of firms they target in activism campaigns than hedge funds do. This difference is not surprising, since many quasi-insider activists are founders and early employees. A larger stake presumably allows a quasi-insider to absorb more of the fixed costs associated with an activism campaign, which might otherwise make a campaign at a smaller firm cost prohibitive. It is worth noting that, because quasi-insiders tend to target smaller firms, they tend to have smaller stakes in the target firm in dollar terms than hedge funds do when they launch campaigns (untabulated).

Fifth, while hedge fund activists tend to target firms with high levels of institutional ownership relative to other firms, quasi-insiders do not. Existing research suggests that hedge funds prefer to target firms with high levels of institutional ownership because they rely on these institutional owners to support their campaigns (Brav et al. 2008). Because of their inside connections, quasi-insiders may not need to rely as much on institutional investor support to achieve their objectives. Alternatively, institutional investors’ mandates may prevent them from investing in the types of smaller firms that quasi-insiders target. Quasi-insiders potentially make up for less institutional support through their larger ownership stakes. The lack of institutional ownership also may be partly mechanical, since quasi-insiders tend to own larger stakes in the firms they target, crowding out ownership by others.

2.6 Quasi-insider financial performance

Next, we examine the financial outcomes of quasi-insider campaigns. We begin by examining abnormal announcement returns around campaigns to assess the market’s reaction to these campaigns. Figure 1 plots average cumulative abnormal returns (CARs) over the $(-10,+10)$ window around the campaign announcement date.¹³  Figure 1, panel A, plots CARs for all campaigns in the sample. It shows that a firm’s stock experiences statistically significant abnormal returns of 3% to 4% around the announcement of a quasi-insider activism campaign. A large fraction of this abnormal return occurs in the run-up to the campaign announcement, suggesting leakage of information about the pending campaign.

Figure 1

Quasi-insider campaign announcement CARs and abnormal turnover

This figure plots the cumulative abnormal returns (CARs) and abnormal turnover around the announcement of quasi-insider activist campaigns, starting 10 days before and ending 10 days after the announcement date. The sample consists of activist campaigns obtained from FactSet SharkWatch and 13D filings for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. The sample is restricted to campaigns for which data on target total assets is available in Compustat in the fiscal year prior to the campaign. The sample is further restricted to firms for which data are available on returns in CRSP (see Table 13 for definitions), resulting in 184 quasi-insider activist campaigns. CARs are computed following standard event study methodology using the market model (see Table 13). Abnormal daily turnover in the event period is measured relative to the average daily turnover (calculated as daily trading volume divided by shares outstanding) for the same firm during the (−100,−40) period relative to the event date. CARs and abnormal turnover are winsorized at the 5th and 95th percentiles. Panel A plots the CAR and turnover data for all quasi-insider campaigns; panel B plots CARs for quasi-insider campaigns separated by Objective; and panel C plots CARs for sale-related and non-sale-related quasi-insider campaigns separately.

Figure 1, panel A, also plots average abnormal daily turnover of firms (daily trading volume divided by shares outstanding) in the event period, computed relative to the average daily turnover for each firm during the (−100,−40) period relative to the campaign announcement date. Trading volume appears to be abnormally large around the time of quasi-insider campaigns. The high volume right before a campaign provides further evidence of information leakage. The high volume after the campaign announcement is consistent with investors with strong views about the campaign selling and buying shares in expectation of the outcome.

Figure 1, panel B, plots CARs for campaigns with different objectives. It shows that campaigns attempting to induce a sale of the firm exhibit the highest abnormal returns, in excess of 15%. The difference is consistent with findings from the hedge fund activism literature that much of the value increase around activism campaign announcements in general is driven by the possibility of a takeover (Mulherin and Poulsen 1998; Boyson and Mooradian 2011). However, there are only 12 such campaigns, and we can measure CARs for only 10 of these. Campaigns where the objective is general value maximization or board control also exhibit large abnormal returns. Campaigns where the objective is Other Specific Actions exhibit the smallest abnormal returns, suggesting that the market responds more positively to campaigns in which quasi-insider activists seek to reassert a degree of control rather than just force specific one-time actions.

Given the large CARs associated with the 10 sale-related campaigns, it is possible that the statistically significant average CAR for the full sample is driven by these 10 campaigns. To assess this possibility, Figure 1, panel C, plots CARs for campaigns where the objective is to force a sale of the firm and all other campaigns separately. It shows that non-sale-related campaigns exhibit statistically significant abnormal returns. The mean CAR for non-sale-related campaigns is approximately 3%. These campaigns appear to exhibit more information leakage, with most of the CAR occurring prior to the announcement date.

To more formally assess the announcement returns around quasi-insider campaigns, we compute and report CARs over the $(-10,+1)$ window around quasi-insider campaign announcement dates. Table 7 reports these CARs. Panel A reports CARs for all campaigns and for campaigns with different objectives. The average CAR for the full sample is 3.9%, which is statistically significant at the 1% level based on a two-tailed t-test. Announcement CARs are positive around campaigns with each different objective, though they are only statistically significant for sale-related campaigns. As is apparent in Figure 1, panel B, by far the largest announcement CARs occur around these campaigns. CARs around the announcement of these campaigns are 18.1%, on average.

Table 7

Quasi-insider campaign announcement CARs

A. All campaigns & campaigns by objective
		CAR(−10,+1)
	N	Mean	p-value
All	184	0.039	0.00
General Value	31	0.054	0.21
Sale-related	10	0.181	0.01
Board Control	40	0.048	0.11
Board Representation	67	0.017	0.24
Other Specific Requests	36	0.017	0.44
Non-sale-related	174	0.031	0.01
Sale-related—Non-sale-related difference		0.150	0.01

B. Campaigns by ownership characteristics
		CAR(−10,+1)
	N	Mean	p-value
$\ge$Median Activist Ownership	84	0.066	0.00
<Median Activist Ownership	84	0.024	0.22
Difference		0.042	0.11
$\ge$Median Insider Ownership	89	0.040	0.00
<Median Insider Ownership	88	0.035	0.02
Difference		0.005	0.84
$\ge$Median Institutional Ownership	91	0.030	0.09
<Median Institutional Ownership	90	0.050	0.01
Difference		−0.019	0.44

C. Campaigns by other characteristics
		CAR(−10,+1)
	N	Mean	p-value
Founder	65	0.067	0.00
No Founder	119	0.024	0.14
Difference		0.044	0.09
Forced CEO Departure	33	0.012	0.66
Voluntary CEO Departure	51	0.057	0.06
Difference		−0.045	0.28
Successful Campaign	82	0.047	0.01
Unsuccesful Campaign	102	0.032	0.06
Difference		0.015	0.55

This table reports mean cumulative abnormal returns for the (-10,+1) window around the date of the campaign announcement (see Table 13). The sample consists of firms that are targets of activist campaigns obtained from FactSet SharkWatch and 13D filings for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. The sample is restricted to 184 campaigns for which data are available on returns in CRSP. Panel A reports mean CARs for quasi-insider activist campaigns overall, by objective, and depending on whether the campaign is sale related. Panel B reports mean CARs for quasi-insider campaigns split by median activist, insider, and institutional ownership. Panel C reports mean CARs for quasi-insider campaigns split by whether the campaign includes a founder, whether the former CEO was forced out (for campaigns involving former CEOs), and whether the campaign was successful. p-values for CARs are based on t-tests comparing means to zero. p-values for differences in CARs are based on t-tests comparing means to each other.

Panel B reports announcement CARs for campaigns with different ownership characteristics. We divide the sample into campaigns with above and below median activist ownership (11.5%), insider ownership (14.5%), and institutional ownership (37.4%).¹⁴ Announcement CARs are 6.6% when activist ownership is above the median, compared to 2.4% when activist ownership is below median. The difference between these two average CARs is nearly statistically significant, with a p-value of.11. This difference suggests that campaigns in which the activist has a larger ownership stake may be perceived as more credible and therefore more likely to lead to outcomes that benefit shareholders. Differences in announcement CARs between campaigns with above and below median insider ownership and institutional ownership are small and statistically insignificant.

Finally, panel C reports announcement CARs for campaigns with differences in other characteristics, including whether the activist is a founder, whether the activist is a CEO who departed involuntarily or voluntarily (for campaigns where the activist is a former CEO), and whether the campaign is successful in achieving its objectives. Founder-initiated campaigns earn significantly higher announcement CARs than non-founder-initiated campaigns. This difference suggests that campaigns initiated by founders, who likely have a closer connection with the firm than other quasi-insiders, such as former executives, are perceived as more credible. Announcement CARs are higher for campaigns initiated by former CEOs who departed voluntarily are higher than those initiated by former CEOs who departed involuntarily. However, these differences, with the exception of founder-initiated campaigns, are statistically insignificant. While announcement CARs are higher around successful campaigns than unsuccessful campaigns, the difference is small and statistically insignificant. Note that investors do not know the success of the campaign at the time it is announced, though they may have some ability to forecast campaign outcomes.

To dig further into the incremental importance of various factors affecting announcement CARs, we regress $(-10,+1)$ announcement CARs on campaign, firm, and ownership characteristics. Table 8 presents the results. The explanatory variables in column 1 are indicator variables for each campaign objective. The omitted objective is Other Specific Requests. The positive coefficients in column 1 indicate that announcement returns are higher for all other campaign objectives than for Other Specific Requests. However, the difference is statistically significant for sale-related campaigns only, which experience the highest announcement CARs, consistent with Figure 1, panel B.

Table 8

Campaign announcement CAR regressions

	(1)	(2)	(3)	(4)
	CAR(−10,+1)	CAR(−10,+1)	CAR(−10,+1)	CAR(−10,+1)
General value	0.041			−0.010
	(0.050)			(0.059)
Sale related	0.167**			0.171**
	(0.066)			(0.073)
Board control	0.043			0.007
	(0.040)			(0.050)
Board representation	0.008			−0.002
	(0.030)			(0.035)
log(Total assets)		−0.016**		−0.020*
		(0.008)		(0.010)
Stock return		−0.027		−0.051
		(0.032)		(0.034)
Tobin’s q		0.004		0.011
		(0.008)		(0.010)
Cash		−0.049		−0.034
		(0.086)		(0.112)
R&D		−0.147		−0.132
		(0.146)		(0.161)
ROA		0.086		0.133**
		(0.057)		(0.063)
Dividend yield		−0.121		0.146
		(0.501)		(0.548)
Debt		0.018		0.013
		(0.042)		(0.049)
Activist ownership			0.003*	0.002
			(0.001)	(0.001)
Insider ownership			−0.002*	−0.002*
			(0.001)	(0.001)
Institutional ownership			−0.054	−0.044
			(0.035)	(0.029)
Year FE	Yes	Yes	Yes	Yes
Observations	185	182	161	160
Adjusted R-squared	.022	−.041	−.025	−.004

This table reports coefficients from ordinary least squares (OLS) regressions where the dependent variable is equal to the cumulative abnormal return in the (-10,+1) window around the date of campaign announcement. The sample consists of firms that are targets of activist campaigns obtained from FactSet SharkWatch and SEC 13D filings for the period for the period 1995 through February 1, 2021, initiated by a founder, former top executive, former director, or current director who is not a current executive or board chair. The sample is restricted to 184 campaigns for which data on returns in CRSP is available. The explanatory variables in column 1 are campaign objectives, with Other Specific Requests as the omitted category. The explanatory variables in column 2 are firm characteristics. The explanatory variables in column 3 are ownership variables. Column 4 includes all variables. All specifications include year fixed effects. Campaign objectives are classified using information from FactSet campaign synopses and SEC 13D and proxy filings (see panel B of Table 2). All other variables are defined in Table 13. t-statistics are reported in parentheses.

*

p <.1;

**

p <.05;

***

p <.01.

The explanatory variables in column 2 are firm characteristics. Among the seven firm characteristics included in column 2, only firm size has explanatory power over announcement CARs, with smaller firms earning larger CARs (significant at the 5% level). This sensitivity could reflect the fact that changes are more difficult to implement in larger firms.

The explanatory variables in column 3 are ownership characteristics. Consistent with the univariate comparisons in Table 7, panel B, announcement CARs increase with activist ownership. Announcement CARs also decrease with insider ownership. Note that this result does not stem from a lower campaign success rate in firms with more insider ownership, as we find no evidence of a relationship between campaign success and insider ownership (Table 5). However, a campaign could be more likely to induce meaningful changes when insiders hold a smaller stake and hence are less able to resist, regardless of whether the campaign is successful in achieving its stated objectives.

Finally, column 4 includes all of the explanatory variables from columns 1 through 3. Announcement CARs continue to be larger in campaigns with sale-related objectives and to decrease with firm size and insider ownership. The relationship between announcement CARs and activist ownership continues to be positive but ceases to be statistically significant when we include all of the characteristics in the regression. Interestingly, the positive relationship between CARs and ROA becomes statistically significant, at the 5% level, when we include all of the characteristics in the regression. This result is somewhat counterintuitive but may suggest that meaningful changes are easier to implement in healthier firms, even if the scope for improvements is larger in less healthy firms. Alternatively, the market may anticipate a campaign at a poorly performing firm with higher probability and therefore may already price in expected value gains associated with a campaign more in these firms.

2.7 Firm operating performance

The positive abnormal returns around campaign announcements suggest that investors view quasi-insider campaigns as increasing future cash flows to shareholders. To further assess the consequences of quasi-insider campaigns, we next analyze changes in measures of operating performance over the years around quasi-insider campaigns. Following an approach similar to Brav et al. (2008), we analyze changes in EBITDA/Assets over the period from the year prior to a campaign (t—1) to up to two years after a campaign (t + 2), relative to a matched sample of observably similar firms. We construct the matched sample by selecting a firm for each targeted firm from the same two-digit SIC industry that has the closest propensity score computed as the fitted value from a probit regression of an indicator for a quasi-insider campaign on total assets and operating performance in year (t—2). We only include firms that have data on operating performance available for years (t—2) through to (t + 2), and we exclude financial firms since operating performance measures for financial firms are difficult to compare to those for nonfinancial firms. The distribution of changes in EBITDA/Assets exhibits substantial noise and several potential outliers. We therefore winsorize the change in EBITDA/Assets at the 5th and 95th percentiles tails. Table 9 reports the difference between the mean performance of the quasi-insider targets and the matched firms, with p-values comparing the differences.

Table 9

Changes in operating performance around quasi-insider campaigns

A. All campaigns (N = 148)
	Diff w/match	p-value
(t + 1)-(t-1)	−0.024	.316
(t + 2)-(t-1)	−0.004	.881

B. Campaigns with CARs (N = 114)
	Diff w/match	p-value
(t + 1)-(t-1)	−0.002	.929
(t + 2)-(t-1)	0.005	.878
Campaigns with positive CARs (N = 69):
(t + 1)-(t-1)	0.005	0.853
(t + 2)-(t-1)	0.022	0.556
Campaigns with nonpositive CARs (N = 45):
(t + 1)-(t-1)	−0.012	0.765
(t + 2)-(t-1)	−0.023	0.664

This table reports mean change in operating performance for targets of quasi-insider activist campaigns in excess of the mean change in performance of a matched sample in years before and after being targeted by quasi-insider activists. Operating performance is measured as EBITDA divided by Total Assets. The table reports mean changes in performance between years (t−1) and (t + 1) or (t + 2) relative to the year of the campaign. The sample is restricted to firms with data available in years t-2 through to t + 2 relative to the year of the campaign. Financial firms are excluded. The sample in panel A consists of all quasi-insider activist target firms that meet these criteria. The sample in panel B consist only of targets for which data on cumulative abnormal returns (CARs) around the announcement of the campaigns in the (−10,+1) window is available. A matched firm for each campaign target is selected from the same two-digit SIC industry, and is closest in a propensity score from a probit regression on total assets and operating performance in year t-2 to the campaign target. Changes in operating performance are winsorized at the 5th and 95th percentiles. p-values are reported for t-tests comparing the mean changes in operating performance of the quasi-insider campaign targets and the matched sample.

Panel A shows the evolution of the operating performance measures for all campaigns for which we can obtain this data for the years t—2 through t + 2. Operating profits decrease, on average, from the year before a quasi-insider campaign to both the first and second years after. The estimated decreases in EBITDA/Assets are large, at 2.4 percentage points to the first year after and 0.4 percentage points to the second year after a campaign. However, neither of these changes is statistically significant. The standard deviation of the change in EBITDA/Assets is so large that identifying statistically significant changes in operating performance would require much larger average changes.

Panel B shows the same results for firms for which we are able to measure announcement CARs as well as firms with positive and negative announcement CARs separately. The increase in EBITDA/Assets from the year prior to a campaign to the year after is negative for the subsample of firms for which we can measure CARs, at 0.2 percentage points, and remains statistically insignificant. The change to the second year after a campaign becomes positive, at 0.5 percentage points, but is still statistically insignificant. The mean changes in EBITDA/Assets after campaigns are positive for campaigns with positive announcement CARs and negative for campaigns with nonpositive CARs, though even these changes are statistically insignificant. In the end, operating performance measures appear too noisy for our sample to allow for reliable inference about the longer run consequences of quasi-insider campaigns.¹⁵

2.8 Circumstances of former CEOs’ departures

We further characterize the quasi-insiders who initiate activism campaigns by examining the circumstances in which the 123 former CEOs engaging in campaigns ceased being CEO. We focus on former CEOs here because we can more readily identify the dates and reasons for departure for former CEOs than for other former executives. Table 10 presents this analysis.

Table 10

Characteristics of former CEO activists

Former CEOS (N = 123)
Departure	N	%
Forced departures	49	39.8%
Voluntary departures	74	60.2%
Market-adjusted stock return during 12-months prior to departure
	Mean	Median
All former CEOs	−12.8%	−20.3%
Forced departures	−14.5%	−31.8%
Voluntary departures	−11.7%	−7.2%
Industry median-adjusted ROA in year of departure
	Mean	Median
All former CEOs	−21.8%	−1.3%
Forced departures	−16.6%	−6.0%
Voluntary departures	−21.0%	0.1%
Days between departure and campaign announcement
	Mean	Median
All former CEOs	986	423
Forced departures	686	308
Voluntary departures	1,187	644

This table summarizes the characteristics of quasi-insider activists who are former CEOs. The sample consists of activist campaigns obtained from FactSet SharkWatch and 13D filings for the period 1995 through February 1, 2021, initiated by a former CEO who is not a current officer of the firm. There are 123 former CEOs who participate in 122 quasi-insider activist campaigns. Former CEO activists and their departure dates are identified and obtained from FactSet campaign synopses, Capital IQ, SEC 13D and proxy filings, and web searches. The nature of former CEO departure (i.e., forced or voluntary) is determined using the FactSet campaign synopsis, where available, and Google searches. Market-adjusted returns are computed from CRSP monthly returns and adjusted using the value-weighted CRSP index. Industry-adjusted ROA is computed as income before extraordinary items divided by total assets and is adjusted by the median of the firm’s four-digit SIC industry.

Some former CEOs (39.8%) who subsequently initiate activism campaigns at their former employer departed involuntarily. This fraction is more than three times as large as the 13.0% of overall CEO departures that Parrino (1997) finds to be involuntary, suggesting that fired CEOs are especially likely to attempt to reinvolve themselves in their former employers. The median time between a CEO departure and an activism campaign that the former CEO initiates is only 423 days. In addition to initiating more campaigns than CEOs who departed voluntarily, CEOs who departed involuntarily wait less time before launching campaigns, with a median lag of just 308 days.

The former CEOs who initiated activism campaigns in our sample do not appear to be star performers returning to resuscitate their former employers. The median market-adjusted stock return during the 12 months prior to departure is -20.3%, and the median industry-adjusted ROA the year prior to departure is -1.3%. Unsurprisingly, performance prior to departure is significantly worse for CEOs who departed involuntarily. However, even for those who departed voluntarily, the median market-adjusted stock return is -7.2% over the year prior to departure and the median industry-adjusted ROA is effectively zero.

This evidence suggests that former CEOs who launch activism campaigns at their former employer are unlikely to have been star managers. This conclusion may explain why firms subject to quasi-insider campaigns do not experience improvements in operating performance—and, in fact, experience declines—after these campaigns. It also helps to explain why the declines in performance are larger when the campaign is successful. In addition, this evidence may offer insight into the rationale for quasi-insider campaigns. It is possible that many of these campaigns are launched by former executives who feel that they were wrongfully terminated and are seeking to reassert themselves in the firm’s affairs, consistent with anecdote regarding Steven Vestergaard and Destiny Media Technologies that we described at the beginning of the paper.

2.9 Former CEO blockholders

We have thus far presented evidence documenting the initiation of activism campaigns by quasi-insiders. We now take a step back and use our second sample to analyze the prevalence and activism activities of former CEO blockholders. We first analyze the characteristics of firms that have these blockholders. We then analyze the factors that predict which former CEO blockholders initiate activism campaigns. This second form of analysis allows us to examine which former quasi-insiders launch campaigns from among a set of former quasi-insiders who could have launched campaigns.

Table 11 compares the characteristics of 2,221 firm-years in which a firm has a former CEO blockholder to firm-years in which a firm does not have a former CEO blockholder. Recall that our method for identifying blockholders is conservative, since we assume that an investor is no longer a blockholder in all years after the investor’s final 13D/G filing. The true number of firm-years with former CEO blockholders is likely considerably higher.

Table 11

Former CEO blockholders

	QI	QI	Non-QI	Non-QI	Median
	N	Median	N	Median	Diff
Total assets	2,221	291	205,334	230	61***
log(Total assets)	2,221	5.672	205,334	5.439	0.494***
Market cap	2,077	241	180,317	151	90***
log(Market cap)	2,077	5.484	180,317	5.014	0.470***
Tobin’s q	2,072	1.425	179,330	1.367	0.058**
Market-to-book equity	2,077	1.598	180,194	1.411	−0.570***
Cash	2,221	0.116	205,334	0.081	0.025***
R&D	2,221	0.000	205,334	0.000	0.000**
Capital expenditures	2,221	0.018	205,334	0.021	−0.003***
Dividend yield	2,221	0.000	205,334	0.000	0.000***
Debt	2,221	0.146	205,334	0.174	−0.028**
ROA	2,221	0.012	205,334	0.006	0.006***
Stock return	1,813	0.048	119,304	0.040	0.008
Institutional ownership	1,763	0.423	115,723	0.470	−0.047***

This table reports summary statistics of characteristics of firms with and without former CEO blockholders. Table 13 defines all variables.

*

p <.1;

**

p <.05;

***

p <.01 (for Wilcoxon signed-rank tests that compare medians of firm-years with former CEO blockholders with and without former CEO activist campaigns).

Compared to firms without former CEO blockholders, those with former CEO blockholders tend to be large and profitable. While speculative, one possible explanation for these differences is that CEOs are more likely to be fired from firms that are struggling to achieve profitability and growth, and terminated CEOs are more likely to liquidate their ownership stakes upon departure. Firms with former CEO blockholders also tend to have less institutional ownership. This difference is likely mechanical. By construction, these firms have a large CEO blockholder, which reduces the shares available for institutional shareholders to own.

Recall from Table 3 that the probability a firm is targeted in a quasi-insider campaign is negatively related to size and recent performance. There are two possible explanations for these relationships. They could reflect differences in the presence of quasi-insiders, who, by definition, can only target firms with which they have a quasi-insider relationship. Alternatively, they could reflect selection of targets on these characteristics, conditional on the presence of a quasi-insider. While we can only identify a subset of quasi-insiders who are clearly present by virtue of holding a large stake, the comparisons in Table 11 suggest that the distribution of quasi-insiders across different types of firms is unlikely to drive the results in Table 3.

Next, we further explore the argument that the relationship between the probability that a firm is targeted and both size and performance is driven by the selection of these targets, conditional on the presence of a quasi-insider. We do so by comparing the characteristics of firms that are and are not subject to a quasi-insider campaign in a given year, using only the subsample of firm-years that have former CEO blockholders. Table 12 presents the results.

Table 12

Former CEO blockholder targets and nontargets

	Active	Active	Nonactive	Nonactive	Median diff
	N	Median	N	Median	Diff
Total assets	101	171	2,120	299	−128*
log(Total assets)	101	5.140	2,120	5.702	−0.562**
Market cap	98	103	1,979	251	−148***
log(Market cap)	98	4.633	1,979	5.525	−0.892***
Tobin’s q	98	1.286	1,974	1.432	−0.146**
Market-to-book equity	98	1.307	1,979	1.612	−0.305**
Cash	101	0.129	2,120	0.116	0.013
R&D	101	0.000	2,120	0.000	0.000
Capital expenditures	101	0.024	2,120	0.018	0.004*
Dividend yield	101	0.000	2,120	0.017	−0.017*
Debt	101	0.137	2,120	0.148	−0.011
ROA	101	−0.011	2,120	0.013	−0.024***
Stock return	81	−0.218	1,732	0.056	−0.274***
Institutional ownership	80	0.487	1,683	0.420	0.067
Activist ownership	50	0.134	1,190	0.137	−0.003

This table reports summary statistics of characteristics of firm-years with and without former CEO activism conditional on having a former CEO blockholder. Table 13 defines all variables.

*

p <.1;

**

p <.05;

***

p <.01 (for Wilcoxon signed-rank tests that compare medians of firm-years with former CEO blockholders with and without former CEO activist campaigns).

The comparisons in this table reveal patterns similar to those in Table 3. Compared to firms with former CEO blockholders who do not launch campaigns, those where former CEO blockholders launch campaigns tend to be smaller and to exhibit poorer recent performance in terms of both return-on-assets and stock returns. Indeed, the stock price performance differences are more pronounced here than when we compare all quasi-insider targets to firms in the same industry in Table 3. That table shows that firms subject to quasi-insider campaigns in general have a 16.2-percentage-point lower stock return over the year prior to the campaign than the average firm in the same industry and year. The comparison in Table 12 shows that firms in which former CEO blockholders launch campaigns have a 27.4-percentage-point lower stock return over the prior year than firms in which former CEO blockholders are present but do not launch campaigns.

One factor that we were unable to consider when comparing all quasi-insider campaigns to industry averages in Table 3 is the ownership of the quasi-insider, since we do not observe this information for non-quasi-insider targets. Here, since we are conditioning on quasi-insider ownership, we are able to provide this comparison, though we only observe this information for about half of the firm-years in the former blockholder CEO sample because many campaigns occur prior to the former CEO’s first 13D/G filing in the data. Interestingly, ownership stake does not appear to predict whether a former CEO blockholder launches an activism campaign. It is plausible that two competing forces are at play here. On the one hand, a larger stake is likely to make an activism campaign more effective, encouraging such a campaign. The results in Table 5 support this argument. On the other hand, a large stake is also likely to allow a former CEO to influence a firm’s decisions without the need for a costly activism campaign.

3. Conclusion

Morgan Lewis, a prominent law firm, recently issued advice on how companies can make themselves less vulnerable to activism by investors who had a prior relationship with a company as insiders, such as founders and former CEOs. We examine the role of such investors, whom we term quasi-insiders, in the governance of firms. We document that they engage in shareholder activism campaigns just as activist institutional investors do but tend to target smaller companies that institutional investors are likely to ignore. These quasi-insiders appear to be relatively aggressive in their campaigns, seeking outright control rather than changes to specific corporate policies with greater frequency. This finding suggests that concerns about companies’ exposure to quasi-insider activism are well-founded.

While the market tends to respond positively to the announcement of a quasi-insider campaign, we do not find concrete evidence indicating that operating performance improves following campaigns. However, changes in operating performance measures are so noisy that tests of changes in operating performance likely have little statistical power. Given the apparent tendency of quasi-insiders to intervene in their former employers and, perhaps more importantly, the threat that they might do so, future research further exploring the long-term implications of quasi-insider campaigns would be useful. Such research would require more detailed data on the nature of specific actions taken in these firms, which is not generally publicly available. Future research shedding light on private interventions by quasi-insiders that do not result in public campaigns also would be useful since such interventions are likely to be even more common than campaigns. However, such interventions are, by definition, difficult to identify.

We are grateful to Andrew Ellul (Editor), an anonymous referee, Nick Gantchev, Bernadette Minton, Adam Pritchard, and Hong Zhao; conference participants at the 2016 Arizona Junior Finance Conference, the 2018 American Finance Association Meeting, the 2018 Conference for Empirical Legal Studies, and the 2019 Drexel Corporate Governance Conference; and seminar participants at Arizona State University, University of New South Wales, Wilfrid Laurier University, and Western University for helpful comments. We also thank Tim Brenden, Dennis Galinksy, Daniel Nikolic, Theodore Wolter, and Ye (Emma) Wang for research assistance.

Footnotes

1

Source: https://www.sec.gov/Archives/edgar/data/0001099369/000166357719000366/mainbody.htm

2

The percentages sum to more than 100% because some quasi-insiders had multiple roles.

3

Most research on shareholder activism studies U.S. firms. See Cziraki, Renneboog, and Szilagyi (2010) for a study of activism in Europe. See Appel, Gormley, and Keim (2016) for evidence that even passive institutions may play a role in governance.

4

Per conversations with Securities and Exchange Commission staff, executives, and other insiders who meet the 5% ownership threshold sometimes file 13Ds rather than 13Gs even though they never engage in any form of activism.

5

In a few cases the same former employee repeatedly launched campaigns over several years according to FactSet. For example, a former director of American Express unsuccessfully sought board representation at the company over 6 consecutive years. We do not view each of these campaigns as independent. To avoid giving undo weight to these cases, we consider these as a single campaign taking place when the activist targeted the firm for the first time.

6

The full list of strings we search for in the primary filer field is as follows: “LLC,” “L.L.C.,” “CORP,” “INC,” “LP,” “L.P.,” “LLP,” “L.L.P.,” “LTD,” “L.T.D,” “ASSOCIATE,” “FUND,” “PARTNERS,” “GROUP,” “TRUST,” “PLC,” “P.L.C,” “S.A.,” “S.P.A,” “INVESTMENT,” “ESTATE,” “BANK,” “CAPITAL,” “MUTUAL,” “PENSION,” “HOLDINGS,” “HOLDING,” “FOUNDATION,” “ASSOCIATION,” “INTERNATIONAL,” “DEVELOP,” “MANAGE,” “TECHNOLOG,” “LABORAT,” “RETIREMENT,” “COMMUNICATION,” “VENTURE,” “ENERGY,” “INVESTOR,” “COLLEGE,” “PHARMAC,” “ADVISER,” “EQUITY,” “ELECTRIC,” “SECURITY,” “CONSULTANT,” “COMMERCIAL,” “CREDIT,” “GOVERNMENT,” “SOCIETY,” “COMPANY,” “COMPANIES,” “CORPORATION,” “COOPERATIVE,” “CONSTRUCTION,” “CONCEPTS,” “GESELLSCHAFT,” “INDUSTR,” “SERVICE,” “SYSTEM,” “MORGAN STANLEY,” “RESOURCE,” “INSURANCE,” “AMERICA,” “BANCORP,” and “&.” We also search for primary filer names ending in “CO,” “AG,” and “SA.”

7

We identify founders by searching for the string “found” within three words of the company name in the individual’s Capital IQ biography.

8

As an example, Guy Cook made four 13D filings and amendments before departing as CEO of Bacterin International in April 2012. None of these filings indicated an activist role. Cook then filed a 13D in August 2013 including the following information in Item 4: “Mr. Cook is the founder of the Issuer and served as its chairman, chief executive officer and president until April 2013. Prior to the date of this report, the Reporting Persons acquired the shares of Common Stock reported herein solely for investment purposes, and not with any plans or proposals that relate to or would result in any of the transactions specified in clauses (a) though (j) of Item 4 of Schedule 13D. However, because the Reporting Persons now believe that the Issuer would be better able to realize its full value as a private entity, the Reporting Persons plan to engage legal and financial advisers to assist them in evaluating alternatives for taking the Issuer private.”

9

See Sias, Turtle, and Zykaj (2017), Blume and Keim (2011), and Gutierrez and Kelley (2009) for discussions of issues associated with the Thomson Reuters/WRDS 13(f) data.

10

Filers also frequently appear to file initial 13D/Gs when amendments are required and vice versa.

11

We focus on comparisons of medians because, even after winsorizing the data, a handful of firms report extreme observations that make a comparison of means difficult.

12

We use this time period as a comparison because we measure stock returns in the year prior to a campaign announcement.

13

We are able to measure CARs for 184 of the 255 campaigns in our sample for which Compustat data on total assets is also available. Eighty percent of the 71 campaigns that we are unable to match trade over the counter.

14

Note that these medians slightly differ from those reported in Table 3 as they are based on campaigns for which we can compute announcement CARs.

15

It is possible that impact of quasi-insider campaigns on operating performance takes longer than 2 years to take effect. Extending the horizon over which we examine changes in operating performance beyond two years post-campaign exacerbates sample attrition significantly. We lose 21% and 44% of our sample if we extend the horizon to 3 and 4 years post-campaign. Using these smaller samples, the change in EBITDA / Assets after campaigns remains statistically insignificant.

Appendix

Table A1

Variable definitions

Variable	Definition
Abnormal turnover	Daily turnover minus average daily turnover during the (-100,-40) period relative to the campaign announcement date, where daily turnover is daily trading volume divided by shares outstanding (Source: CRSP)
Board control	Indicator equal to one if the primary campaign objective is board control (Source: FactSet/Item 4 of 13D Filings)
Board representation	Indicator equal to one if the primary campaign objective is board representation (Source: FactSet/Item 4 of 13D Filings)
Capital expenditures	Target firm’s capital expenditures divided by total assets (Source: Compustat)
CAR(-i,+j)	Cumulative return from day -i to day +j relative to campaign announcement minus normal returns estimated using market model during estimation window of (-280,-30), with a minimum of 60 observations required (Source: CRSP)
Cash	Target firm’s cash and short-term investments divided by total assets (Source: Compustat)
Debt	Sum of target firm’s long-term debt and debt in current liabilities divided by total assets (Source: Compustat)
Dissident Ownership	Percentage of shares held by former CEOs in last 13D/G filed prior to that calendar year (Source: SEC EDGAR)
Dividend yield	Sum of target firm’s common and preferred dividends divided by sum of the market value of common equity and preferred equity (Source: Compustat)
General value	Indicator equal to one if primary campaign objective is to maximize shareholder value without specific requests (Source: FactSet/Item 4 of 13D Filings)
Institutional ownership	Percentage of shares held by institutions that file with a 13F (Source: Thompson Reuters)
Inside ownership	Percentage of shares held by insiders as reported in annual 10-K’s (Source: SEC EDGAR)
log(Market cap)	Natural logarithm of market capitalization in millions of dollars of target firm at end of fiscal year before campaign (Source: Compustat)
log(Total assets)	Natural logarithm of total assets at end of fiscal year before campaign (Source: Compustat)
Market-to-book equity	Ratio of target’s market value to book value of equity (Source: Compustat)
Other specific requests	Indicator equal to one if primary campaign objective is specific request not related to maximizing shareholder value, board representation, board control, or hostile/unsolicited acquisition (Source: FactSet/Item 4 of 13D Filings)
R&D	Target firm’s research and development expenses divided by total assets; set equal to zero when missing (Source: Compustat)
ROA	Target firm’s income before extraordinary items divided by total assets (Source: Compustat)
Sale related	Indicator equal to one if primary campaign objective is a Hostile/Unsolicited Acquisition (Source: FactSet/Item 4 of 13D Filings)
Stock return	Buy-and-hold return in year prior to campaign announcement in excess of value-weighted CRSP index return, computed using monthly return data (Source: CRSP)
Tobin’s q	Total assets minus book value of equity plus market value of equity, scaled by total assets (Source: Compustat)

This table contains the definitions and descriptions of the variables used in the paper.