Stock Picking in Disguise?

New Evidence that Hedge Fund Activism Adds Value

Benjamin S. Solarz Economics

April 6, 2009 Advisor: Dean Takahashi

Abstract Using a large hand-collected data set spanning from 2000 through 2008, I find substantial evidence that activist hedge funds do more than pick stocks; they improve them. Controlling for differences in target firms, activist investments outperform the same hedge funds passive investments. Over two months, they earn 3.8% greater returns; over two years, they earn 18.4% greater returns; and measured by financial statements, they improve margins and ROA, increase payout and leverage, reduce balance sheet assets, and sell their targets more frequently and for higher prices. Hedge funds follow five cohesive activist strategies: they (1) improve business strategies, (2) advise on mergers and acquisitions, (3) demand that target firms sell themselves, (4) optimize capital structures, and (5) fix corporate governance. These results imply that skilled hedge funds can add significant value as corporate partners.

Introduction If you ask CEOs about their greatest fears, they might start talking about activist hedge funds. If someone with a 203 area code calls you, call them back very carefully, said public relations maven Joele Frank at a 2008 panel on activist hedge funds.1,2 Apparently, Greenwich, CT can be very dangerous. As the hedge funds located there and across the country redesign corporate governance, they have kicked up a storm of controversy. In the past few years, Nelson Peltz purchased Wendys, Carl Icahn pushed Yahoo into Microsofts arms, and William Ackman encouraged Target to divest its real estate. To top it all off, activists have even targeted other hedge funds, as when Carl Icahn sued over Steel Partners plans to go public.3 While there is no shortage of bold claims, hard evidence is tougher to come by. Do activist hedge funds truly threaten corporate health? To the contrary, using hand-collected data on 718 activist investments, I find that activist hedge funds add substantial value to their socalled victimsmeasured either by stock market returns or accounting performance. I improve on the existing literature, which fails to distinguish activist investing from astute stock picking, to show that hedge funds actually catalyze their targets outperformance. In particular, I test whether activism adds value beyond stock picking by comparing activist investments to the same hedge funds passive investments. In addition, I describe the most common activist strategies. Judging by short-term returns, investors applaud an activist hedge funds decision to target a company. Moreover, investors prefer these activist investments to passive investments. Over 61 days, controlling for the differences in target firms, activist investments outperform passive investments by 3.8%, and investments seeking board seats outperform other activist
1

While the term activist may draw to mind images of Gandhi, I use it to refer to a better dressed, better paid, and much more self-interested cohort: aggressive investors who attempt to influence their targets. Throughout this thesis, I will use the terms activist and active interchangeably. 2 Discussing the Perils of Activist Investors, The ew York Times [New York], 3 April 2008. 3 Icahn and BofA Take Steel Partners to Court, The ew York Times [New York], 7 February 2007.

-1-

investments by 3.2%. In addition, investors react positively to activism even when it contains no stock picking signal. When activists change their intentions from passive to active, stock prices jump significantly. Except for activism seeking to fix corporate governance, investors reward all activist strategies. While I find evidence that activism yields significant excess returns, I also find that these returns have declined over time. Long-term returns also indicate that hedge fund activism enriches shareholders. Over two years, controlling for the differences in target firms, activist targets outperform passive targets by 18.4%, and investments seeking board seats outperform other activist investments by 18.6%. Again, except for corporate governance activism, all strategies generate substantial long-term returns. Also consistent with short-term results, returns have declined over time. Activism earns these excess returns by improving its targetsthey increase profitability, optimize their capital structures, divest unprofitable assets, and sell themselves more frequently and for higher prices. Controlling for the differences in target firm characteristics, activist targets increase their return on assets (ROA) by 3.0% more than passive targets, increase their margins by 2.9% more, increase their payout ratios by 2.5% more, and increase their leverage by 0.7% more. Activist targets also reduce their assets by 4.6% more than passive targetsmost likely by shedding unprofitable divisions. Finally, not only do activist targets sell themselves more frequently, but they also sell for higher prices. In improving target firms, activists follow five cohesive strategies: they (1) improve business strategies, (2) advise on mergers and acquisitions, (3) demand that target firms sell themselves, (4) optimize capital structures, and (5) fix corporate governance. Each strategy targets different firms and improves them in distinct ways. Further, except for corporate governance, every strategy earns significant excess returns.

-2-

The new evidence presented in this paper suggests that, far from economic hitmen, activist hedge funds add significant value to their targetsjudged by either stock prices or financial statements. Thus, instead of avoiding hedge funds calls, both management teams and shareholders alike should welcome their insights. This paper is organized as follows. Section I reviews the literature. Section II explains my methodology. Section III describes hedge fund activism; it details the firms that activists target and the strategies that they use. Turning to the analysis, Section IV shows that activists earn excess short-term returns. Section V presents similar evidence for long-term returns. Section VI describes the fundamental improvements in activist targets. Section VII concludes.

Section I: Background For hundreds of years, CEOs have eroded the wealth of nations. In 1776, Adam Smith warned, The directors of such [joint-stock] companies, however, being the managers rather of other peoples money than of their own, it cannot well be expected, that they should watch over it with the same anxious vigilance with which the partners in a private copartnery frequently watch over their own.4 In modern terms: CEOs often roll up their sleeves not for auditing spreadsheets but for dealing cards. In August 2007, while errant hedge funds brought his firm to its knees, Bear Stearns CEO Jimmy Cayne spent 10 days playing bridge out of state.5 Agency costs arise whenever management (the agent) takes actions which harm shareholders (the principal), generally because their incentives differ. In general, shareholders want management to invest as much as possible in projects whose internal rates of return exceed the firms weighted average cost of capital (WACC) (Koller et al. 2005). But CEOs paid based
For a more modern treatment, see Berle and Means (1932) and Jensen and Meckling (1976). The same November 17, 2007 Wall Street Journal Article, titled Wall Street Bosses Spending Too Much Time on the Golf Course, also reported that in June alone, Cayne squeezed in 13 rounds of golf.
5 4

-3-

on earnings growth might increase earnings by investing in projects with internal rates of return (IRRs) greater than zero but below the firms WACC. Similarly, CEOs paid based on ROA might forgo projects with attractive IRRs if those projects would drag down a firms overall returns. While some herald stock options as an incentive alignment panacea, managers only earn those options by meeting performance targets, which offer their own perverse incentives.6 Further, Jensen (1986) argues that managers may grow the company beyond its optimal size, indiscriminately purchasing assets. While doing so harms shareholders, it also increases managements compensation, pride, and job security. Gabaix and Landier (2008) find that from 1980 through 2003, the six-fold increase in large US firms market capitalization led directly to a six-fold increase in CEO pay. So while shareholders might want the firm to disgorge unproductive assets, management might harbor other intentions. And even if managers wanted to maximize shareholder value, they just might not know any better. At Prudential Bancorps February 9, 2007 annual meeting, for example, shareholder activist Joseph Stilwell publicly wagered $25,000 that CEO Thomas Vento could not define return on equity on a per share basis. Vento failed.7 For these reasons, Black (1992) and Pound (1992) argue that shareholder activism could improve corporate performance. With their voting rights, large shareholders might combat misguided management teams. But on the whole, the literature strongly disagrees. The first problem is that too few activists play the game. Due to the free rider problem, activists personal benefits do not justify the necessary costs.8 Bebchuk (2008) discusses the costs that activists must endure. Parrino et al. (2003) suggest that stockholders may simply prefer to walk away.

For the problems with performance targets, see Yermack (1995) and Bebchuk and Fried (2003). For the problems with options themselves, see also Hall and Murphy (2003). 7 Predential Bancorp, SEC Schedule 13D, Filed 2 November 2008. 8 See Rock (1992), Black (1998), Bainbridge (2006), and Kahan and Rock (2006).

-4-

Even when activists do take up arms, their wooden arrows make little dent in the corporate fortress. Most institutional investors suffer conflicts of interest, such as those mutual funds face; regulatory constraints, including diversification requirements and insider trading regulations; and political constraints, especially for public pension funds.9 The empirical record agrees; activists sometimes induce improvements in corporate governance, but they rarely, if ever, induce long-term changes in performance.10 In a literature review, Black (1998) concludes, A small number of American institutional investors spend a trivial amount of money on overt activism efforts. Institutions achieve the effects on firm performance that one might expect from this level of effortnamely, not much. While the early literature focuses on mutual funds, pension funds, and block purchasers, this thesis argues that hedge funds succeed where others have failed. First, drawing on prior legal scholarship, I argue why activist hedge funds might outperform their activist peers. Second, I document empirical evidence that activist hedge funds do improve corporate value. (A) Hedge Funds Are Different As many a pitchbook eagerly points out, hedge funds are different. They enjoy many structural advantages that distinguish them from prior activists. In particular, hedge funds have the incentives to engage in activism(1) they earn pay-for-performance bonuses, (2) they suffer fewer conflicts of interest, and (3) they can avoid ERISA regulationand they have the weapons to engage in activism(4) they can make concentrated bets, (5) they can take can take longterm, illiquid positions, and (6) they can use derivatives. Below, I elaborate on these points.

See Black (1990) and Romano (1993). For other critiques, see Lipton and Rosenblum (1991), Karpoff, Malatesta, and Walkling (1996), Carleton, Nelson, and Weisbach (1998), Davis and Kim (2008), and Del Gurcio, Wallis, and Woidtke (2008). 10 For corporate governance successes, see Bebchuck (2005, 2007). For other failures, see Wahal (1996), Karpoff, Malatesta, and Walkling (1996), Johnson and Shackell (1997), Black (1998), Gillan and Starks (2000), Karpoff (2001), Romano (2001), Barber (2007), Gillan and Starks (2007), and Del Guercio, Wallis, and Woidtke (2008).

-5-

First, hedge funds earn pay-for-performance bonuses, which incentivize them to wage successful activist campaigns. Hedge funds do not face the 1940 Investment Company Acts pay restrictions. Indeed, hedge fund managers often earn up to 20% or more of their firms annual return (Brav et al. 2008). In contrast, mutual fund managers rarely earn performance bonuses.11 Second, hedge funds suffer fewer conflicts of interest than other institutional investors. For example, mutual fund managers might placate a company to win its pension business (Davis and Kim 2007), and public pension fund managers might flatter a CEO to win contributions for a re-election campaign (Kahan and Rock 2006). Third, hedge funds can avoid ERISA or prudent man regulations, which restrict the risks other institutional investors can stomach.12 The threat of liability may scare these potential activists from taking the risk. Fourth, hedge funds can make concentrated bets to fight entrenched management teams. Similarly, activists might threaten to purchase the entire target company, often as a negotiating ploy. In contrast, restrained by subchapter M of the Internal Revenue Code and the 1940 Investment Company Act, mutual funds cannot hold more than 10% of any companys stock or concentrate more than 5% of their own portfolio in a given name (Klein and Zur 2008). Fifth, hedge funds can take illiquid positions to weather a long fight. Whereas hedge funds sign investors into lengthy lock-up agreements, mutual funds must maintain constant liquidity to satisfy redemption requests (Klein and Zur 2008). And finally, hedge funds use exotic strategies. Through stock lending and derivatives, hedge funds might acquire voting rights without taking on economic exposure.13 Or hedge funds

Golec (1992) and Deli (2002) find that only 6-7% of mutual fund companies offer their managers performance bonuses. Kahan and Rock (2006) report that 97% of mutual funds annual compensation ignores investment gains. 12 See Coffee (1991), Bhide (1993), and Aragon (2007). 13 See Christoffersen, et. al. (2007) and Hu and Black (2007).

11

-6-

might use these derivatives expressly to gain economic exposure; by increasing activisms potential gains, they justify its requisite costs. In contrast, mutual funds must follow the Investment Company Acts short-selling and leveraging restrictions (Klein and Zur 2008). (B) Similar Studies Do ot Distinguish Activism From Stock Picking While activists proclaim their successes in myriad press releases, it is not so easy to assess their claims. Simply because an activist earns abnormal returns, for instance, does not mean it actually used activism to earn those returns. And simply because an activist target improves its margins does not mean the activist did anything to catalyze the change. Rather than activism, these findings might simply mark great passive investingchoosing companies primed to improve beyond the markets expectations. Thus, to test whether activism adds value beyond stock picking, we must compare activist investments to passive investments. To date, nonetheless, the literature continues to compare apples to pairshedge funds activist investments to a non-hedge fund benchmark. Several recent papers focus on hedge fund activism.14 Using a case study approach, Becht et al. find that the Hermes U.K. Focus Funds activism earns large, positive excess returns. More similar to this study, Puustinen (2007), Boyson and Moradian (2008), Brav et al. (2008), and Klein and Zur (2008) study large crosssections of hedge fund activism. They find strong evidence for short-term abnormal returns and mild evidence for long-term improvement. But none of these studies distinguishes activism from passive investing in disguise; they provide no evidence that activists have done anything more than simply pick undervalued companies poised to outperform expectations. To mitigate this problem, Clifford (2008) compares activist investments to passive investments. But this study offers a flawed dichotomy. First, Clifford fails to control for

Becht et al. (2006), Puustinen (2007), Boyson and Mooradian (2008), Brav et al. (2008), Clifford (2008), Greenwood and Schor (2008), and Klein and Zur (2008) study hedge fund activism.

14

-7-

differences in target firms. Without doing so, activism might outperform simply by targeting more attractive firms. Second, Clifford compares Schedule 13D filings, which he calls active, to Schedule 13G filings, which he calls passive. But these filings differ across other crucial dimensions. Most importantly, hedge funds must file Schedule 13Ds within 10 days of building their position but need not file Schedule 13Gs for an entire year. The larger reaction around Schedule 13D filings, therefore, might simply represent the market preferring fresh ideas to pickled ones. Clifford also defines activism too broadly. In particular, he calls all Schedule 13D filings activist investments, even though many explicitly renounce activist intentions. In my sample of 942 Schedule 13Ds, for instance, 48.1% simply hold the stock for investment purposes. Instead of capturing the active-passive distinction, Cliffords study might measure some other confounding variable.15 This paper eliminates these problems. In addition to ignoring stock picking, prior studies barely discus heterogeneity within activism. While certain activist strategies, like selling the target firm, might generate large returns, others, like improving corporate governance, may not. Further, these strategies might affect companies in conflicting ways that cancel each other out, leaving little aggregate effect. For example, business strategy activism might reduce leverage, while capital structure activism might increase leverage. Without testing for such heterogeneity, previous studies improperly conclude that fundamentals change very little. In contrast, I analyze activists strategies in detail. Greenwood and Schor (2008) present some evidence on heterogeneity. In particular, they disagree that activists improve their targets; instead, they propose that activists simply put firms

15

There are several possible confounds. First, even passive investors holding at least 20% in a company must file a Schedule 13D, so Cliffords results might reflect investors preference for these concentrated bets. Second, an investor might file a Schedule 13D to convey information to other investors. Well-known passive investors, for example, might benefit from leaking their best ideas to the public. Sweren (2008) documents abnormal returns when hedge funds leak their best ideas. And third, astute investors might prefer to file Schedule 13Ds to protect themselves from potential litigation. These legally savvy investors might also have the best ideas.

-8-

in play. The authors show that, excluding sales, activist hedge funds earn negligible long-term excess returns. But by asymmetrically removing successful sales, and not failed sales, Grenwood and Schor have imprecisely measured activists non-sale returns, biasing their results downward. In Section V, I correct for this bias. Additionally, Greenwood and Schor (2008) argue that hedge funds poor performance during the 2008 Credit Crunch proves that they rely on selling their targets. But this overlooks the crisiss main symptom; levered institutions like hedge funds unwound dramatically. This structural pressure, and not activists reliance on selling their targets, might explain the results. In summary, while the literature documents some activist outperformance, (1) it fails to distinguish this outperformance from stock picking, and (2) it largely ignores heterogeneity.

Section II: Experimental Design (A) Contributions to the Literature To argue that activism adds value, I improve on the literature in several ways. This is the first paper to: (1) use a comprehensive sample of experienced hedge funds, (2) compare activist Schedule 13D investments to passive Schedule 13D investments, (3) control for differences in target firms, (4) measure short-term abnormal returns around events which occur after the hedge fund has disclosed its position, (5) analyze board seats, and (6) fully document cross-sectional variation in activists strategies. This section elaborates on these frontiers. First, in contrast to the existing literature, I study hedge funds that repeatedly pursue activism. Prior scholars call activist any hedge fund which has ever filed an activist Schedule 13D. But this includes inexperienced hedge funds, which bring down the averages. Expanding the activist sample to include these novices resembles defining an athlete as anyone who has ever

-9-

owned a baseball glove; after rounding up two NY Yankees and five hundred physical education teachers, prior scholars conclude that no one can consistently throw a 90 mph fastball. By selecting only hedge funds known for activism, I resolve this bias. Second, to disentangle activism from stock picking ability, I compare hedge funds active Schedule 13D investments to the same hedge funds passive Schedule 13D investments. I expand on this method below. While Clifford (2008) also compares activist returns to passive returns, I improve his methodology by (A) limiting my focus to Schedule 13Ds and (B) recognizing that many Schedule 13Ds espouse passive intentions. I also expand on his study by more thoroughly examining changes in target firms accounting performance. Third, I control for variation in target firm characteristics like undervaluation and underperformance. I expand on this method below. This improves on Cliffords (2008) study by dismissing the alternative hypothesis that activists target more attractive firms. Fourth, I document short-term abnormal returns around events that take place after the initial Schedule 13D filing. By following each investments SEC trail, I find subsequent filings in which a hedge fund changes its intentionsfor example, the first time it expresses activist intentions. Because the hedge fund has already disclosed a passive interest in the target firm, this method more cleanly isolates the activist signal from the stock picking signal. Fifth, I distinguish activist instruments, like seeking board seats, from activist strategies, like selling the firm; an activist uses instruments to implement its strategies. I consider five instruments: seeking board seats, launching proxy fights, offering to buy the target, providing private financing, or suing the target. Because 80% of all investments using an instrument seek board seats, I consider them most thoroughly. Board seats distinguish activism from stock picking by providing tangible proof that the hedge fund seeks to influence its target.

- 10 -

Sixth, I identify five cohesive activist strategies: business strategy, M&A, sale, capital structure, and corporate governance. While Puustinen (2007) and Brav et al. (2008) analyze short-term returns by strategy, I add evidence that strategies differ over the long-run. I also investigate heterogeneity in fundamental improvements. This is important for two reasons. First, prior studies have observed little change in target firms fundamental accounting performance, perhaps because different strategies cancel each other out. And second, only by understanding their strategies can we replicate activist hedge funds ability to improve firms. (B) Testing For Activism To prove that hedge fund activism adds value, we must demonstrate that (1) shareholders enjoy greater returns than they otherwise would and (2) target firms improve their accounting performance more than they otherwise would. But unlike the Beatles, these findings cannot stand alone. If activism earned excess returns without improving target firms, then hedge fund activism would simply be a penname for astute stock picking. On the other hand, if activist targets became better companies without earning excess returns, then the market must have already expected those improvementsand the activist probably did not cause the change. By comparing activist targets to passive targets, I come closer to measuring how these activist targets otherwise would perform than does any prior paper. Both activist and passive investments benefit from hedge funds superior stock picking abilities, but only activist targets benefit from hedge funds activist abilities. Thus, to the extent activist investments outperform passive investments, activism must add value.16 There is, however, a second layer of stock picking to peel back. Hedge funds might pick more attractive firms for activist investments than for passive investments. To eliminate this bias,

16

Thus, activist outperformance is a sufficient condition to show that activism adds value. It is probably a necessary condition, too. Because activism costs more than passive investing, hedge funds should demand greater returns.

- 11 -

I control for differences in target firmsin particular, differences in undervaluation and in underperformance. If activist investments still outperform passive investments, we can conclude that activism adds value. Additionally, I analyze board seats to better distinguish activist investments from passive investments. Unlike the mere rhetoric accompanying many filings, seeking board seats helps activists influence the firms decisions. Even if it never wins a seat, the hedge fund has opened a dialogue with management. At best, the activist might get its way. At worst, the activist might threaten a proxy fight as a bargaining chip. Thus, whereas a fine line separates activism from passive investing, a much thicker line separates board seat investments from other investments. As before, controlling for target firm variation, if we observe that board seat investments outperform other activist investments, board seats must add value. (C) Collecting Data Unlike previous studies, which include hedge funds that have made very few activist investments, I study only hedge funds which focus on activism. 17 In particular, I consider the 48 hedge funds which the shareholder activism advisory firm 13D Monitor defines as most active. 18 13D Monitor sought to include every fund that repeatedly uses activism; that is, rather than assembling a random sample of a broad universe, they compile a comprehensive list that fits a narrow criteria. Thus, my sample generalizes not to every ex-Goldman Sachs trader who files a Schedule 13D, but rather to experienced activists. Because it is comprehensive, the list suffers from little if any bias in describing the historical record. I find no evidence of survivorship bias
17

Puustinen (2007), Boyson and Mooradian (2008), Brav et al. (2008), Clifford (2008), and Klein and Zur (2008) all use news searches to form an initial hedge fund list. Thus, even hedge funds with only one prior activist investment enter the sample. Brav et al. (2008) limit their sample to hedge funds with at least two activist investments. 18 13D Monitor maintains a subscription-based database of Schedule 13D filings at http://www.13dmonitor.com/. I am completely indebted to Ken Squire and Robyn Rosenblatt for allowing me access to this database and for putting up with my annoying questions. To the extent I had time to emerge from my dorm room at any point during the month of January, Ken and Robyn deserve full credit.

- 12 -

when comparing my sample to those in similar papers.19 As evidence against backfill bias, several hedge funds in my sample performed quite poorly.20 And finally, even if the sample were biased towards hedge funds whose activist investments perform quite well, it would not necessarily be biased towards hedge funds whose activist investments outperform their passive investments. In fact, if activism does simply represent disguised passive investing, these hedge funds may be the biggest culprits. Thus, 13D Monitors sample provides a fitting laboratory. Using 13D Monitors online database, I collected information on all 942 investments that these hedge funds made. To further dissect these investments, I turned to the Edgar database, which houses SEC filings dating back to 1994. Section 13(d) of the 1934 Exchange Act requires investors to file with the SEC when they amass large stakes with activist intentions. In particular, hedge funds must file a Schedule 13D within 10 days of acquiring at least 5% in any public company.21 In Item 4, filers must declare why they acquired the shares. By reading over 6,000 individual SEC filings, I collected the entire activist history for each investment. From these 942 investments, I narrowed the sample. First, to focus on the current activist climate, I excluded filings before January 1, 2000. Second, I excluded all firms which declared bankruptcy before the hedge fund invested. This narrowed my sample to 718 investments, of which I classified 393 as active and 325 as passive. I classify an investment as activist if the investor announces any activist instrumentboard seats, proxy fights, buyout offers, private financings, or lawsuitsor any activist strategybusiness strategy, M&A, sale, capital structure, or corporate governance. Of the 393 activist investments, 213 sought board seats.

I compared my sample to two sources: Puustinen (2007)s sample and hedge fund consulting firm Hedge Fund Solutionss database. Without exception, the most frequent funds in each source overlapped with my sample. 20 Of the 46 firms with long-term return data, 23.9% exhibit negative average activist abnormal returns. 21 While activists rarely pierce the 5% threshold in large firms, this sample paints an accurate picture; activists tend to focus on targets small enough that they can establish large positions.

19

- 13 -

Because this is the first paper to suggest that many Schedule 13D filings might harbor passive intentions, some elaboration is in order. Passive institutional investors who acquire at least 5%, but less than 20%, of a companys stock may file a Schedule 13G with the SEC.22 But sometimes these firms will instead file a Schedule 13D. The strongest evidence for this proposition comes from the hedge funds themselves: many filings announce that they are holding the stock for investment purposes only. Should we believe them? First, the SEC and target firms use the courts to enforce truthful disclosures.23 Second, citing conversations with hedge fund managers, Clifford (2008) argues that hedge funds exercise extreme caution to avoid filing any misleading disclosures. Third, each hedge fund in my sample announces either passive or activist intentions at least once, which implies that these funds do differentiate passive from activist investments. Fourth, I find that after 32.4% of passive Schedule 13Ds, hedge funds subsequently announce activist intentions. This devotion to updating their goals indicates that hedge funds take care to convey honest information. And finally, in searching Bloomberg news for certain randomly chosen passive investments, I found no evidence that the hedge funds engaged any attempts to change the target firms. So why might a hedge fund file a Schedule 13D rather than a Schedule 13G? First, using the more rigorous Schedule 13D safeguards the hedge fund against liability. Second, funds might take advantage of Schedule 13Ds immediate disclosure. They may wish to communicate their investment thesis so that other investors, upon realizing the firms true value, bid up its price.24 After collecting information on these investments, I pulled share price histories from the Center for Research in Security Prices (CRSP). I calculate the target firms return in excess of

22

Passive investors acquiring at least 20%, however, must file a Schedule 13D. While this might bias my Schedule 13D sample towards large passive firms, I find almost no such instances of 20%+ passive positions. 23 See, e.g., SEC vs. Montgomery Medical Ventures, LP (1996) and Ronson Corp v. Steel Partners II (2005). 24 For example, Sweren (2008) documents excess returns after hedge funds file Schedule 13Fs.

- 14 -

the market model measured against the CRSP value-weighted NYSE/Amex/Nasdaq index. I describe the event study method more fully in Appendix A. While unreported, all results are robust to the use of the Fama French factors. The analysis excludes firms for which CRSP lacks datamost likely small, unlisted firms. This might bias my findings. I use two types of data on target fundamentals. First, I use accounting data pulled from Bloomberg. Second, I calculate industry-adjusted statistics to better analyze a target firms idiosyncratic performance. For example, following Ang and Chen (2006) and Puustinen (2007), I use abnormal q to proxy for undervaluation. I calculate the firm-specific component (denoted abnormal q) as the percentage difference between the firms Tobins q and the median of its industry peers, using the GICS industry classification. I also calculate abnormal margins, abnormal return on equity (ROE), and abnormal leverage, where the firm-specific component (denoted abnormal) is the difference between the firms value and the median value of its industry peers. I calculate these variables by hand using firm-level data from Compustat. Unfortunately, I could only obtain Compustat data dating back to January 1, 2004. Both Bloomberg and Compustat were missing data for several firms, presumably smaller firms.

Section III: Hedge Fund Activism Before turning to the results, this section elaborates on the differences between activism and passive investing. First, I describe the firms which activist hedge funds target. In general, they target undervalued firms, with low Tobins q.25 While hedge funds target relatively profitable firms overall, their activist efforts target less profitable firms than do their passive

Tobins q equals the firms market value divided by its replacement valuemeasured as the market value of debt plus equity divided by the book value of debt plus equity. A low Tobins q implies undervaluation.

25

- 15 -

Table I Target Firm Characteristics

The table reports the characteristics of hedge funds' targets. Column 1 compares activist investments to passive investments. Column 2 compares activist investments seeking board seats to other activist investments. All variables refer to the trailing twelve months prior to the Schedule 13D. All data is winsorized at the 10% level. *, **, and *** refer to statistical significance at the 10%, 5%, and 1% levels.

(1) Activism Activist Market Cap Tobin's q Abnormal q EBITDA Margin Abnormal Margin Return on Assets Return on Equity Abnormal ROE Leverage Abnormal Leverage Payout Ratio Cash/Assets N= 757.1 1.6 -16.6 10.9 3.0 8.5 20.0 0.7 22.2 15.1 9.1 16.3 303 Passive 654.4 1.7 12.3 12.1 3.2 9.4 22.3 2.4 22.8 11.6 9.8 14.5 492 Diff 102.7 -0.1 -29.0 -1.1 -0.2 -1.0 -2.4 -1.7 -0.6 3.5 -0.7 1.9 * * *** * ** * ** Board 636.0 1.5 -22.9 8.4 0.7 7.4 17.0 0.2 19.7 12.8 10.0 16.6 157

(2) Board Seats Other 891.4 1.7 -11.1 13.5 5.0 9.5 23.0 1.2 24.7 17.1 8.3 16.1 146 Diff -255.4 -0.2 -11.7 -5.1 -4.3 -2.1 -6.0 -1.0 -5.0 -4.3 1.7 0.5 *** *** *** *** *** *** **

efforts. Second, I discuss heterogeneity among activists strategies. I describe five main categories: business strategy, M&A, sale, capital structure, and corporate governance. (A) Activists Target Undervalued Firms Before we can measure activist investments against a passive benchmark, we should first describe the ex-ante differences in the two samples. Table I reports summary statistics for active investments, passive investments, active investments seeking board seats, and active investments not seeking board seats. In Appendix D, Table D-1 adds more information, including definitions. Column 1 shows that hedge funds go activist in undervalued and underperforming firms. Measured by Abnormal q, relative to the median firm in its industry, the average activist target is 16.6% more undervalued. Relative to passive targets, activist targets are 29.0% more undervalued (t = -3.68). While hedge funds target overall profitable firms, they go active in the

- 16 -

least profitable ones. The average activist target has 3.0% higher margins and 0.7% higher ROE than the median firm in its industry. Relative to passive targets, however, they have 1.1% lower margins (t = 1.34), 1.0% lower ROA (t = 1.71), and 2.4% lower ROE (t = 1.49).26 Activist targets also hoard assets. Relative to passive targets, they have 1.9% more cash per dollar in assets (t = 1.49) and 0.7% lower payout ratios (t = 0.42).27 Overall, the table provides moderate evidence that hedge funds try to unlock value in underperforming target firms. While these results resemble most prior studies, Clifford (2008) finds that active targets are more profitable than passive targets.28 This might be explained by his methodology; because he considers all Schedule 13Ds to be active, he includes some passive investments. Consistent with the literature, these passive investments might lift the observed profitability. Column 2 shows that within activist investments, hedge funds seek board seats in the most troubled companies. Compared to other activist investments, those seeking board seats are even more undervalued, with 0.2 lower Tobins q (t = 2.47). They are also less profitable, with 5.1% lower margins (t = 4.13), 2.1% lower ROA (t = 2.49), and 6.0% lower ROE (t = 2.59). This evidence refutes the alternate hypothesis that activists seek board seats to associate themselves with successful firms. Instead, it seems they seek board seats to improve underperforming firms. (B) Activists Use Different Strategies Activists tailor their strategy to each target. From the hedge funds SEC filings, I categorized the investments into five broad strategies: (1) Business Strategy activism, which attempts to improve the firms competitive strategy and capital allocation, (2) M&A activism, which seeks to maximize value through extraordinary transactions, like acquisitions or spinoffs,

26

I calculate margins as EBITDA divided by sales. I calculate ROA as EBITDA divided by book value of assets. I calculate ROE as EBITDA divided by the book value of equity. 27 I calculate payout ratio as common dividends divided by net income. 28 See Boyson and Mooradian (2008), Brav et al. (2008), and Klein and Zur (2008).

- 17 -

(3) Sale activism, which demands that the firm sell itself, (4) Capital Structure activism, which tries to influence leverage or payout policies, and (5) Corporate Governance activism, which addresses the firms governance policies. I improve on Brav et al. (2008) by separating what they call business strategy into business strategy and M&A. This way, I distinguish between activism which improves companies through capital allocation and activism which improves companies through corporate transactions. In Appendix B, I elaborate on these categories by excerpting from SEC filings. In my sample, hedge funds most frequently use M&A and Sale activism (130 times each) and least frequently use business strategy activism (75 times). While correlations among the strategies vary from 15% to 37%, business strategy and sale activism correlate the least, at 7%. This supports the hypothesis that business strategy and sale activism seek mutually exclusive goalswhere one builds long-term businesses, the other quickly sells them. If activist strategies truly differ, we should expect each strategy to target firms with particular weaknesses. Table II shows a logit regression that uses firm characteristics to predict the hedge funds strategy. Business strategy activism targets poorly performing companies with lower sales growth (t = -1.33) and lower margins (t = -1.22). Targets are also larger (t = 3.13),
Table II

Logit Analysis By Strategy

Using a logit regression, the table reports the liklihood that activists target firms with various characteristics. The dependent variable is a dummy equal to one if the hedge fund targets the firm with a particular activist strategy. The sample population includes all hedge fund investments, active or passive. All data is winsorized at the 10% level. *, **, and *** represent statistical significance at the 10%, 5%, and 1% levels.

Bus Strat t -stat Coeff. Market Cap Tobin's q Growth Margin Leverage Cash/Assets Constant No. and Pseudo-R 2 0.06 9.19 -1.43 -2.41 0.17 1.02 -3.14 547 3.13 0.42 * -1.33 -1.22 0.19 * 1.60 *** -6.78 0.06 ***

M&A t -stat Coeff. 0.04 *** 2.78 -15.72 -0.89 0.62 0.76 -0.11 -0.07 0.05 0.07 1.43 ** 1.78 -2.21 *** -5.75 547 0.04

Sale Coeff. -0.03 * -21.02 -0.20 0.85 -0.33 1.04 ** -1.76 *** 547 t -stat -1.51 -1.04 -0.22 0.51 -0.42 1.66 -4.71 0.03

Cap Struc t -stat Coeff. 0.03 -56.34 1.21 4.61 -1.30 1.57 -2.25 547 * 1.29 *** -2.35 1.24 *** 2.61 * -1.41 ** 2.06 *** -5.14 0.06

Corp Gov t -stat Coeff. 0.00 0.13 10.14 0.50 -0.06 -0.07 -0.78 -0.43 0.31 0.39 0.79 * 1.43 -2.67 *** -6.46 547 0.01

- 18 -

Table III Board Seats By Strategy

The table present the logistic regression odds ratios that each strategy (1) seeks board seats and (2) wins board seats. The dependent variable is a dummy set equal to 1 if the activist sought or won a board seat. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

(1) Board Logit Odds Bus Strategy M&A Advice Sale Cap Structure Corp Gov No. and Pseudo-R 2 2.1 *** 1.2 1.9 ** 1.3 1.7 * 316 z -stat 2.59 0.89 2.40 1.01 1.99 0.04 Odds 1.5 1.1 0.9 1.3 1.3 155

(2) Win Logit z -stat 0.92 0.33 -0.29 0.60 0.61 0.02

perhaps because large companies struggle to allocate capital among their many divisions. M&A and sale activism target undervalued companies (t = -0.89 and t = -1.04). M&A activism targets large companies (t = 2.78), perhaps because larger firms have underperforming divisions rife for spinoffs. Sale activism targets smaller companies (t = -1.51), perhaps because they are easier to sell in whole. Capital structure activism targets healthy firms, which can increase their payout. Targets enjoy very high margins (t = 2.61) with lots of cash on hand (t = 2.06). They also have low existing leverage (t = -1.41).29 Corporate governance activism targets firms with excess cash (t = 1.43), likely following Jensens (1986) hypothesis that misaligned managers hoard assets. Table III describes the frequency with which each strategy seeks board seats. Column 1 and Column 2 present odds ratios from logistic regressions that predict which strategies seek board seats and which strategies win board seats. Business strategy activism nominates directors most frequently (z = 2.59)probably to nurse businesses back to profitabilitywhile M&A activism nominates directors least frequently (z = 0.89)probably because the hedge funds anticipate exiting after quick flips. Similarly, business strategy activism wins board seats most
29

I measure leverage as long-term debt divided by total assets.

- 19 -

often (z = 0.92), while M&A and sale activism succeed least often (z = 0.33 and z = -0.29). This makes sense; targets prefer long-term partners, like hedge funds advising on business strategy, to short-term speculators, like hedge funds seeking to sell the business. In summary, activists target undervalued firms with room to improve. Depending on the strategy they want to use, they choose different targets. Having described activists methods, I now turn to the first research question. Does this activism increase returns?

Section IV: Short-Term Stock Returns Everyday in the stock market, shareholders vote with their wallets; with good news, they bid up prices, and with bad news, they drive them back down. Particularly in efficient markets, these price swings presage actual changes in firms values. Thus, if activism adds value, once shareholders discover hedge funds intentions, we should observe immediate excess returns. Absent efficient markets, however, these short-term returns do not themselves prove that activism adds value. They may reflect hedge funds solving the adverse selection problem, not the moral hazard problemthat is, the filings may signal target firms already good prospects without actually signaling that hedge funds will in fact improve those prospects. Investors may believe that hedge funds can access better information or conduct better analysis without adding any activist value. In fact, Sweren (2008) documents short-term abnormal returns around hedge fund Schedule 13F filings, which convey no explicit activist intentions. In this section, using three tests, I argue that activism does add value beyond simple stock picking. First, I compare activist filings to passive filings. I find strong evidence that, in the short-term, the market favors activist investmentsand in particular, activist investments which seek board seats. Second, I use a dummy variable regression to compare activist filings to

- 20 -

passive filings, controlling for target firm undervaluation and underperformance. Again, activist investments outperformparticularly when the hedge fund seeks board seats. Third, I identify events which isolate the activist signal from the passive investing signal. I find significant excess returns. Thus, the stock market responds favorably to activism, not simply to stock picking. I also investigate heterogeneity across these returns. In particular, I investigate when activists add the most valuewhat target firm characteristics, activist instruments, or activist strategies does the market judge most effective? Undervalued, underperforming target firms experience the largest excess returns. I find evidence that the market favors investments which take active steps, like seeking board seats. Except for corporate governance activism, all strategies enjoy abnormal returns. Finally, I note that these returns have declined over time. (A) Activism Exceeds Stock Picking Table IV reports the markets reaction to activism. In Appendix D, Table D-2 shows more details. I calculate excess returns over 61- and 5-day windows relative to the market model
Table IV Short-Term Returns by Investment Type
The table presents average daily returns around the schedule 13D filing in excess of the market model using the CRSP value-weighted index. I estimate parameters over the 200-day window prior to the Schedule 13D filing. Panel A compares activist investments to passive investments. Panel B compares active investments seeking board seats to active investments that do not. I report p -values for medians using the Wilcoxon Signed Rank Test. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active vs. Passive Active (-30,+30) Median Average % Positive N= 11.8% *** 12.9% *** 80.4% *** 194 (-2,+2) 4.4% *** 5.7% *** 79.9% *** 194 Passive (-30,+30) 3.6% *** 4.6% *** 57.6% *** 342 (-2,+2) 1.9% *** 2.7% *** 67.4% *** 340 Difference (-30,+30) 8.2% *** 8.3% *** (-2,+2) 2.5% *** 3.0% ***

Panel B: Board Seats vs. Other Active Investments Board Seats (-30,+30) Median Average % Positive N= 12.5% *** 17.7% *** 88.9% *** 45 (-2,+2) 7.8% *** 8.6% *** 77.8% *** 45 No Board Seats (-30,+30) 10.8% *** 12.6% *** 77.9% *** 149 (-2,+2) 4.1% *** 5.7% *** 80.5% *** 149 Difference (-30,+30) 1.7% * 5.1% * (-2,+2) 3.7% * 2.9% *

- 21 -

In the Short-Run, Active Investments Outperform Passive Over a 61-Day Event Window
20% Board 15% Active NoBoard

10%

5%

Passive

0%

-5%

Fi lin g

D ay s

ay s

ay s D 25 30

ay s

ay s

ay s

ay s

-5

10

15

-3 0

-2 5

-2 0

-1 0

-1 5

Figure 1. For active (blue) and passive (red) hedge fund investments, the chart plots the average excess return around the schedule 13D filing, from 30 days prior to the filing until 30 days afterward. The dashed lines divide active positions into those in which the hedge fund sought board seats (long dashes) and those in which the hedge fund did not (short dashes).

using the CRSP value-weighted index. Both active and passive investments experience positive and significant excess returns surrounding the filing date. Over 61 days, 80.4% of active positions and 57.6% of passive positions show positive returns (z = 8.91 and z = 3.37). Figure 1 plots the average excess return, from 30 days prior to the Schedule 13D filing date to 30 days afterward. Before the filing, between 10 days to 1 day prior, active positions jump by 4.6% and passive positions by 2.1%. On filing day and the following day, active positions jump by 4.5% and passive positions by 1.4%. Afterward, the abnormal returns trend up to 13.9% and 4.6% over 30 days. The results are consistent with Boyson and Mooradian (2008), Brav et al. (2008), Clifford (2008), and Klein and Zur (2008). To test whether activism differs from stock picking, Panel A in Table IV compares activist filings to passive filings. Over 61 days, activism outperforms by 8.3% (t = 3.96). This - 22 -

13 D

20

ay s

ay s

ay s

ay s

ay s

In the Short-Run, Active Investments Outperform Passive Controlling For Undervaluation and Underperformance
10.0% 8.0% 6.0% 4.0% 2.0% 0.0%
Active Passive Board No Board

Figure 2. The chart shows the average 5-day returns to active, passive, board, and no board investments, controlling for Tobin's q and margins, measured by a dummy variable regression.

supports Clifford (2008)s results. Within active investments, those seeking board seats earn significantly higher returns. Over 61-days, these investments outperform by 5.1% (t = 1.33). Overall, it seems activism add value. Still, far from a controlled experiment, these comparisons do not prove that activism adds value. Activist targets might differ from passive targets in a way that makes them more likely to outperform. In other words, hedge funds might simply call their best ideas activist. I use a dummy variable regression to compare activist to passive investments, controlling for the possibility that activist investments might be more attractive ex-ante.30 Figure 2 shows the excess returns to an average firm which a hedge fund targets for an active, passive, board, or no board investment. I regresses the 5-day abnormal return on activism and board seat dummy variables, controlling for the target firms undervaluation and underperformance. The regression uses Tobins q to proxy for undervaluation and margins to

30

In Appendix A, I elaborate on these dummy variable regressions.

- 23 -

Table V Abnormal Returns After 13D Filing Date

The table presents average returns around events that occur after the initial Schedule 13D filing. I calculate returns in excess of the market model using the CRSP value-weighted index. I estimate parameters over the 200-day window prior to the event. I report p -values for medians using the Wilcoxon Signed Rank Test. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

(1) First Action (-30,+30) Median Average % Positive N= 7.5% *** 8.3% *** 73.6% *** 53 (-2,+2) 1.2% *** 4.4% *** 66.0% *** 53

(2) First Board Seat (-30,+30) 6.3% *** 6.9% *** 60.4% ** 48 (-2,+2) 2.5% *** 3.4% *** 77.1% *** 48

(3) First Board Win (-30,+30) 5.2% *** 6.2% *** 66.4% *** 110 (-2,+2) 0.3% 0.8% * 53.2% 109

proxy for underperformance.31 The results are robust to other models. Activist investments outperform passive investments by 3.8% (t = 4.78), and board investments outperform other activist investments by 3.2% (t = 1.71). Thus, activism adds value beyond stock picking. (B) Purely Active Filings Show Excess Returns In addition to comparing returns around Schedule 13D filings, I isolate the activist signal in subsequent filings. Because the activist has already disclosed a 5% stake in the target firm, the only new information that these subsequent filings convey is the activists intention to improve the target firms.32 In addition, because activist hedge funds cannot fully control whether they win board seats after they decide to seek them, filings in which the activist wins seats convey little information about hedge funds ability to pick stocksonly their ability to add value as activists. To isolate these events, I read each investments entire SEC trail. I also conducted basic news searches. I calculate abnormal returns around three events: (1) having previously filed a passive 13D, the first time a hedge fund announces activist intentions, (2) having previously filed an activist 13D, the first time a hedge fund seeks board seats, and (3) having already announced its intention to seek board seats, the first time a hedge fund wins board seats.
31 32

In Appendix D, Table D-3 presents the regression results. If a hedge fund increased its exposure in any subsequent filing, however, it might convey some additional passive signal. But given that the hedge fund has already accumulated 5% of the company, this passive signal is of a small magnitude. Further, this critique does not apply to incidents where activist hedge funds win board seats.

- 24 -

Purely Active Filings Outperform After Initial Schedule 13D

10% 8% 6% 4% 2% 0% -2%
ay s ay s ay s ay s ay s Fi lin g ay s ay s D ay s D ay s D ay s D ay s D ay s 30 D D D D D D D

FirstActivist SeekBoard WinBoard

-3 0

-2 5

-2 0

-1 5

-1 0

-5

15

20

Figure 3. The table shows abnormal returns around events after the initial Schedule 13D. Having previously filed a passive 13D, the blue line shows the hedge fund's first activist intention; having previously filed an activist 13D, the red line shows the hedge fund's first time seeking a board seat; and having previously filed board seat 13D, the green line shows the hedge fund's first time winning a seat.

Table V reports the 61- and 5-day abnormal returns surrounding these events.33 Figure 3 shows the returns over time. Target funds earn positive and significant excess returns surrounding each event date. Over 61-days, 73.6%, 60.4%, and 66.4% of target firms enjoy positive returns around each event (z = 3.76, z = 1.70, and z = 3.82), respectively. On average, the firms outperform by 8.3%, 6.9%, and 6.2% (z = 3.38, z = 2.69 and z = 3.80), respectively. Thus, the market reacts favorably to the activist signal, not just the stock picking signal. (C) Returns Differ By Strategy Just as important as average abnormal returns is the cross-sectional variation of those returns. Using a regression approach, Table VI analyzes (1) what characteristics make a firm most susceptible to improvement, (2) what instruments make an activist most effective, and (3) what strategies add the most value. The dependent variable is the 5-day abnormal return
33

In Appendix D, Table D-4 reports more details.

13 D

- 25 -

10

25

Table VI Cross-Sectional Variation of Short-Term Returns

The dependent variable is the 5-day abnormal return around the Schedule 13D date. For all activist investments, Column 1 regresses abnormal returns on target firm fundamentals. For all investments using activist instruments, Column 2 regresses abnormal returns on dummy variables for those instruments. And for all investments using activist strategies, Column 3 regresses abnormal returns on dummy variables for those strategies. All accounting variables are winsorized at the 10% level. All non-dummy covariates are expressed as the deviation from the sample average values. In columns 2 and 3, intercepts are suppressed . *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

(1) Dependent Variable: Abnormal Returns Constant ln(MV) Abnormal q Abnormal Margin Leverage BoardWin Board Buyout Financing Proxy Bus Strategy M&A Sale Cap Structure Corp Governance Other No. and R 2 Target Fundamentals Coeff. t -stat 6.49 *** 6.98 -0.53 -0.77 -1.93 ** -1.87 -0.23 ** -2.22 0.00 -0.08 120 0.11 Coeff. -6.22 -0.64 6.69 4.30 17.87 3.56 0.26 32

(2) Activist Instruments t -stat -2.47 -2.46 0.99 1.75 4.30 0.65 0.07 0.75 Coeff. -1.58 -0.25 2.68 1.63 7.41 2.92 -1.60 3.34 112

(3) Activist Strategies t -stat -1.51 -2.17 1.31 1.26 4.73 1.53 -0.85 0.55 0.38

*** ***

* **

** ***

* *** *

surrounding the schedule 13D filing. I include as regressors (1) the firms abnormal q, the logarithm of market capitalization, the firms industry-adjusted margins, and the firms debt-toassets ratio; (2) dummy variables for various activist instruments, including winning board seats, seeking board seats, offering to buy the target firm, providing private financing, or launching an additional proxy fight; and (3) dummy variables for the various activist strategies, including business strategy, M&A, sale, capital structure, and corporate governance. For interpretation purposes, I adapt the methodology in Brav et al. (2008). I express all non-dummy covariates as the deviation from the sample mean, and in Column 2 and Column 3, I suppress the intercept because of the dummy variables full span. Thus, we can interpret each dummy coefficient as the

- 26 -

partial return for an average target firm when the hedge fund uses that dummy variable. Both Column 2 and Column 3 include dummy variables which are not mutually exclusive. For example, if a hedge fund uses both Business Strategy and M&A activism, we expect the total abnormal return to be 4.3% (= 2.7% + 1.6%).34 Turning first to target firm fundamentals, the market reaction indicates activists can improve underperforming firms. In particular, activists earn the highest returns when targeting firms with low abnormal q (t = -1.87) and with low industry-adjusted margins (t = -2.22). This is what we should expect; if activists do add value, theyre more likely to succeed at firms with more room to improve. Thus, not only do activists target undervalued and underperforming firms, as documented in Section III, but they earn the highest returns while doing so, too. I also find that activist instruments yield excess returns. Consistent with my earlier findings, seeking board seats generates average abnormal returns of 4.3% (t = 1.75). Hedge funds that win board seats generate 6.7% additional returns (t = 0.99), for 11.0% total excess returns. Thus, the market seems to think that, through board seats, hedge funds can add value to target firms. The market also responds very positively to hedge funds that offer to buy the entire firm, driving target prices up 17.9% (t = 4.30). Whereas prior literature has noted hedge funds tendency to drive returns by putting firms in play, the findings presented here identify an overlooked subtlety: rather than simply agitating for change, hedge funds put their own money on the line.35 Often these buyouts often push targets into another acquirers armsmaking them a formidable weapon in the activists arsenal. But when the hedge fund does purchase the target, stock market returns no longer proxy for the hedge funds return, which depends on what happens after the buyout. Still, these buyout offers generate large returns for shareholders.

34 35

In Appendix A, I describe more fully this dummy variable regression. See Greenwood and Schor (2007) and Brav et al. (2008).

- 27 -

In the Short-Run, Selling the Firm Generates Highest Returns Controlling for Undervaluation and Underperformance

10%

5%

0%

-5% Business Strategy M&A Sale Capital Structure Corporate Governance

Figure 4. The chart shows the average 5-day abnormal return for each strategy, defined as the regression coefficients. The bars show the 90% confidence interval.

Figure 4 shows the average return to each activist strategy. These strategies generate statistically different returns (F = 4.03). Sale activism generates 7.4% excess returns (t = 4.73). Business strategy and capital structure activism also generate significant abnormal returns of 2.7% (t = 1.31) and 2.9% (t = 1.53), which implies that activism can improve its targets. While consistent with Brav et al. (2008), these results contradict Boyson and Mooradian (2008), who find that corporate governance activism generates large returns. This is due to their longer time frame. Indeed, when I include investments from 2000 through 2003 by substituting regular data for the industry-adjusted data, I find that capital structure activism generates positive returns. Perhaps due to improving governance, this strategy has become less successful over time. Additionally, because both Column 2 and Column 3 control for abnormal q and abnormal margin, the coefficients reflect activisms effects, not just target firms fundamentals. This thoughtful variation provides further evidence that shareholders expect activism to generate

- 28 -

Short-Term Returns Decline Over Time As Investments Increase in umber

25% 20% 15% 10% 5% 0% -5%
20 04 20 00 20 01 20 02 20 03 20 05 20 06 20 07 20 08

180 150 120 90 Active Passive 60 30 0

Active Filings

Passive Filings

Figure 5. The line graphs (left axis) show show the average 5-day returns by 13D year, as measured by the coefficients in a dummy variable regression that controls for Tobin's q and margins. The blue line shows the average activist return, and the red line shows the average passive return. The blue bar (right axis) shows the number of active Schedule 13Ds and the red bar (right axis) shows the number of passive Schedule 13Ds filed in my sample each year.

value; if shareholders were purchasing simply on expectations of superior passive investing, they would not parse a hedge funds press releases to discern its intentions. (D) Returns Have Decreased Over Time Finally, I measure how these returns have changed over time. I regress the 5-day abnormal returns against dummy variables for each filing year, controlling for Tobins q and margins. The results are robust to other models. All non-dummy covariates are expressed as the deviation from the sample mean, and constants are suppressed. I reject the hypothesis that annual returns equal each other (F = 2.14). To the contrary, returns have declined; in 2001, the average firm earned 23.1%, but in 2008, the average firm earned only 5.9%. In Figure 5, the blue lines plots the average activist returns to Schedule 13Ds filed each year (left axis). The bar graph shows the number of Schedule 13Ds filed each year (right axis). Consistent with Puustinen (2007), I find that short-term activist returns have declined over time.

- 29 -

This pattern suggests three interpretations. First, as activism becomes more common, shareholders might begin to expect it. Thus, when hedge funds actually do intervene, investors are less surprised. Second, as companies have learned to accept activism, hedge funds have shed their most aggressive tactics. Thus, outside observers might confuse true activism for passive investing. As evidence against this hypothesis, however, passive returns (red line) have not demonstrably improved; in fact, I cannot reject the null hypothesis that the coefficients all equal each other (F = 0.48). Finally, activists may simply have lost their edge, putting money to work in increasingly less attractive situations. In Section VI, however, I find no evidence that activists have lost their ability to improve target firms. In summary, I find substantial evidence that activism adds value. Activist investments outperform passive investments, adjusting for undervaluation and underperformance. In addition, investors applaud hedge funds activist intentions, even if the hedge funds have already disclosed passive investments in the same targets. Activism aimed at undervalued firms and attempts to sell target firms earn the highest returns. Recently, though, these returns have declined.

Section V: Long-Term Stock Returns In the Intelligent Investor, Benjamin Graham (1973) writes, In the short run, the market is a voting machine. In the long run it is a weighing machine. Using what Graham might consider an exit poll, Section IV argued that investors believe activism adds value. This section builds on that work to examine whether the weighing machine agrees. After all, if activism fails to generate long-term excess returns, neither corporate shareholders nor hedge funds clients benefit; activism represents wasted energy at best and value destruction at worst.

- 30 -

If hedge fund activism improves its targets beyond the markets pre-targeting expectations, then target firms should outperform their risk-adjusted benchmarks. In addition, we should expect these investments to continue to outperform as the activism persists. At every point in time, the market should bid stock prices to a level reflecting the expected probability that the activist will succeed (Brav et al. 2008). Because success is never certain, as the activist continues to wage a campaign, the market should continue to reward the effort. If this were not the caseif the market price ever reached a level where further activism would not generate further abnormal returnsthen the hedge fund would exit its position. As in the short-term, however, the efficient markets hypothesis proves crucial to the argument. Instead, if hedge funds could persistently pick winning stocks without improving them, then abnormal returns might simply mark great passive investing, not activist ability. In this section, using two tests, I conclude that activists add value in the long-run. First, I compare activist filings to passive filings. I find strong evidence that activist investments, particularly those which seek board seats, outperform passive investments. Even excluding the 61-days surrounding the filing, activist investments earn significantly higher returns. Second, I compare activist filings to passive filings, controlling for undervaluation and underperformance. Again, activist investments outperform passive investmentsparticularly when the hedge fund seeks board seats. Thus, activism appears to add long-term value. I also investigate heterogeneity across these returns. In particular, I investigate when activists add the most valuewhat target firm characteristics, activist instruments, or activist strategies prove most effective? As in the short-run, undervalued, underperforming target firms earn the highest long-term returns, likely because activism has rejuvenated them. Hedge funds

- 31 -

earn excess returns when they sell target firms, improve targets business strategies, or optimize targets capital structures. Finally, I observe that these returns have declined over time. Before turning to the analysis, however, it is worth elaborating on the methods I will use. The major obstacle to analyzing long-term returns is picking a time period. Brav et al. (2008) analyze hedge fund holding period returnsbuying when hedge funds buy and selling when hedge funds sellbut this methodology (1) introduces error because we cannot perfectly measure these holding periods and (2) conflates hedge funds market timing ability with their activist ability. Instead, to better measure the reward to corporate shareholders (at the expense of the returns to hedge fund investors), I analyze long-term returns along arbitrary windows. To the extent activist hedge funds adopt different time frames for different investments, this method should bias observed abnormal returns downward. In particular, once target firms realize their potential, their abnormal returns should trend towards zero. Thus, any observed effect, whether positive or negative, is likely understated. Further, because this downward bias affects passive investments less than it affects activist investments, any observed difference between active and passive investments is similarly understated.36 Even in spite of this, I still find that active investments both generate positive abnormal returns and outperform passive investments. Long time horizons also complicate how we categorize activism. Over 5 or even 61 days, we can isolate a hedge funds intentions, but over 2 years, hedge funds revise their strategies numerous times. I find that hedge funds significantly change strategies in 79.5% of investments. To most purely match the hedge funds behavior to the relevant investment period, I distill my sample as follows: passive investments include all instances where a hedge fund never amends its passive Schedule 13D, activist investments include all instances where a hedge fund announces an activist instrument or an activist strategy within 180 days of its Schedule 13D, no
36

Unlike activist investments, passive investments have no catalyst. Thus, they have no holding period sensitivity.

- 32 -

Table VII Long-Term Returns by Investment Type

The table presents average monthly returns around the Schedule 13D filing in excess of the market model using the CRSP value-weighted index. I estimate parameters over the 36 months prior to the Schedule 13D. Panel A compares activist investments to passive investments. Panel B compares activist investments seeking board seats to activist investments that do not. I report p -values for medians using the Wilcoxon Signed Rank Test. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active vs. Passive Active (-1,+24) Median Average % Positive 20.1% *** 21.3% *** 66.8% *** (+1,+24) 12.1% *** 13.5% *** 61.1% *** (-1,+24) 10.4% * 3.7% 55.1% ** Passive (+1,+24) 5.3% 1.5% 53.7% *** Difference (-1,+24) 9.7% 17.6% ** (+1,+24) 6.8% * 11.9% *

Panel B: Board Seats vs. Other Active Investments Board Seats (-1,+24) Median Average % Positive 28.2% *** 31.2% *** 75.2% *** (+1,+24) 18.3% *** 21.6% *** 66.4% *** No Board Seats (-1,+24) 13.0% *** 14.3% *** 60.9% *** (+1,+24) 7.6% 7.8% *** 58.0% ** Difference (-1,+24) 15.2% ** 16.9% * (+1,+24) 10.8% ** 13.8% *

board seat investments include all activist investments where a hedge fund never seeks board seats, and board seat investments include all activist investments where a hedge fund seeks a board seat within 180 days of its Schedule 13D. (A) Activism Exceeds Stock Picking Table VII measures long-term returns. I calculate excess returns (1) from one month before the Schedule 13D until 24 months after and (2) from one month after the Schedule 13D until 24 months after. In Appendix D, Table D-5 shows more details. Consistent with short-term results, both activist and passive investments earn positive and significant excess returns. Over the entire 25-month period, 66.8% of activist investments and 55.1% of passive investments show positive returns (z = 5.70 and z = 2.08). Figure 6 plots the excess returns over time. From one month before the Schedule 13D until one month after, activist investments earn 10.3% excess returns and passive investments earn 2.8%. After the short-term reaction subsides, from one month after the Schedule 13D until 12 months after, activist investments flourish while passive investments falter: activist investments increase an additional 8.5% (z = 3.87) while - 33 -

In the Long-Run, Active Investments Outperform Passive Over 25-month Event Window
40% Board 30% 20% 10% Passive 0% -10%
-1 13 M D on F t 1 ilinh M g on th 3 M on th s 6 M on th s 9 M on th s 12 M on th s 15 M on th s 18 M on th s 21 M on th s 24 M on th s

Active NoBoard

Figure 6. For active (blue) and passive (red) hedge fund investments, the chart plots the average excess return around the schedule 13D filing, from 1 month prior to the filing until 24 months afterward. The dashed lines divide active positions into those in which hedge funds sought board seats (long dashes) and those in which hedge funds did not (short dashes).

passive investments fall 0.4% (z = -0.29). Over a longer window, from one month after the Schedule 13D until 24 months after, activist investments earn 13.5% excess returns (z = 5.55) while passive investments earn 1.5% (z = 0.16). These results are consistent with Boyson and Mooradian (2008), Brav et al. (2008), and Klein and Zur (2008). The increasing returns imply that over two years, the market continues to receive positive information. This makes sense; to the extent activists do improve their targets, these changes play out slowly over timeand so should returns. In contrast, passive investors subsequent actions matter very little. As passive investing reveals itself all at once, so should returns. Its also worth noting that this activist momentum is not a free lunch. First, the strategy is not investible because certain investments reveal their activism after the event date. Additionally, activism might carry more risk than passive investing. By estimating target firms factor loadings

- 34 -

In the Long-Run, Active Investments Outperform Passive Controlling For Undervaluation and Underperformance
40.0%

30.0%

20.0%

10.0%

0.0%
Active Passive Board No Board

Figure 7. The chart shows the average 25-month returns to active, passive, board, and no board investments, controlling for Tobin's q and margins, measured by a dummy variable regression.

before the hedge funds launch their activism, I underestimate the target firms risk and thus their benchmark returns. What looks like excess risk-adjusted returns relative to the stocks prior risk might actually represent nothing more than benchmark returns relative to the stocks current risk. To test whether activist hedge funds do more than pick stocks, I compare activist returns to passive returns. Panel A in Table VII shows that over 25 months, activism outperforms passive investing by 17.6% (t = 2.28). Of this outperformance, 11.9% comes after the short-term reaction (t = 1.61). Panel B shows that board seat activism outperforms other activism by 16.9% over the full period (t = 1.57) and by 13.8% after the short-term reaction (t = 1.31). Overall, activism seems to add value. To more cleanly distinguish activism from stock picking, I use a dummy variable regression to control for the possibility that hedge funds might go activist in more attractive targets. I regress the 25-month return on dummy variables for activism and board seat activism,

- 35 -

Table VIII Cross-Sectional Variation of Long-Term Returns

The dependent variable is the 25-month abnormal return around the Schedule 13D date. For all activist investments, Column 1 regresses abnormal returns on target firm fundamentals. For all investments using activist instruments, Column 2 regresses abnormal returns on dummy variables for those instruments. And for all investments using activist strategies, Column 3 regresses abnormal returns on dummy variables for those strategies. All accounting variables are winsorized at the 10% level. All non-dummy covariates are expressed as the deviation from the sample average values. In columns 2 and 3, intercepts are suppressed . *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

(1) Dependent Variable: Abnormal Returns Constant ln(MV) Abnormal q Abnormal Margin Leverage BoardWin Board Buyout Financing Proxy Law Bus Strategy M&A Sale Cap Structure Corp Governance Other No. and R 2 Target Fundamentals Coeff. 8.72 * 1.91 -10.11 ** -1.14 ** 0.04 243 t -stat 1.60 0.47 -1.65 -1.88 0.14 0.03 Coeff. -23.63 -0.14 -20.68 39.86 -15.93 -8.82 -14.87 -48.01 90

(2) Activist Instruments t -stat *** -2.45 -0.16 -1.31 3.12 -0.88 -0.31 -1.16 -1.91 0.21

(3) Activist Strategies Coeff. -16.10 * -0.48 13.90 4.70 0.62 12.94 -8.84 -51.29 101 t -stat -1.56 -0.42 0.67 0.38 0.04 0.71 -0.46 -0.93 0.08

* ***

**

controlling for Tobins q and margins.37 The results are robust to other models. Activism outperforms passive investing by 18.4% (t = 2.17), and board investments outperform other activism by 18.6% (t = 1.59). Thus, activism adds value beyond stock picking. (B) Returns Differ By Strategy As with short-term returns, I investigate the conditions that make activism most effective. In Table VIII, I regress 25-month abnormal returns on (1) target firm fundamentals, (2) activist instruments, and (3) activist strategies. Turning first to target firm fundamentals, activists earn larger returns from undervalued targets and underperforming targets. Consistent with short-term
37

In Appendix D, Table D-6 shows the underlying regression.

- 36 -

returns, activists perform best when firms have low abnormal q (t = -1.65) and low abnormal margins (t = -1.88). The evidence supports the hypothesis that activists do their best work turning around unloved, unprofitable target firms. With respect to activist instruments, board seats continue to drive value in the long-run. Over 25-months, investments that seek board seats earn 39.9% excess returns (t = 3.12), far exceeding any other activist instrument. Consistent with short-run expectations, this finding strongly supports the hypothesis that activists can add value; by seeking board seats, they ensure that management hears their message. Proxy fights and lawsuits generate moderately significant negative returns of -14.9% (t = -1.16) and -48.0% (t = -1.91), respectively. In these hostile situations, activist hedge funds might distract management from running the firm. Alternatively, the hedge fund might be running into a brick walla stubborn management team that refuses to cooperate with value-increasing propositions. Contrary to expectations, within investments that seek board seats, those which win board seats actually underperform those which do not win board seats by 20.7% (t = -1.31). This result contradicts short-term returnswhere shareholders applaud the news that a hedge fund has won a board contest. But this contradiction is not very robust. First, this coefficient is biased because the regression excludes investments initiated from 2000 through 2003. When I increase the sample by substituting Tobins q for abnormal q and substituting margin for abnormal margin, the effect becomes positive; winning adds 6.4% (t = 0.40) to investment returns.38 Second, this coefficient is biased by hedge funds that withdraw their requests after management cooperates, making the investment a success. In fact, of all failed board contests in my sample, in 28.3% of cases the hedge fund voluntarily withdraws its request. After removing withdrawn board seats
38

Part of this increase is due to reduced precision in the covariates. To corroborate that sample size plays a role, in unreported results I run the new regression for the 2004-2008 period and for the 2000-2008 period. As expected, the longer period shows a larger coefficient on winning board seats.

- 37 -

Many Strategies Generate High Returns Controlling for Undervaluation and Underperformance

40%

20%

0%

-20%

-40% Business Strategy M&A Sale Capital Structure Corporate Governance

Figure 8. The chart shows the average 25-day abnormal return for each strategy, defined as the regression coefficients. The bars show the 90% confidence interval.

from the original model, the effect of winning becomes positive. Winning board seats increases returns by 7.2% (t = 0.28). Thus, it appears activists use board seats as a bargaining chip, dropping the board contest once management acquiesces to their demands. In addition to activist instruments, multiple activist strategies earn positive returns. In fact, different strategies returns are statistically indistinguishable from one another (F = 0.20). This does not itself disprove the hypothesis that activists follow different strategies; if strategies differ, they should differ in fundamental performance, not necessarily returns. Figure 8 shows the average return per strategy. As in the short-run, investments which change the targets business strategy and capital structure perform very well, earning 13.9% excess returns (t = 0.67) and 12.9% excess returns (t = 0.71), respectively. Also consistent with short-run expectations, corporate governance activism underperforms the market by 8.9% (t = -0.46). When broadening the sample by substituting Tobins q for abnormal q and margin for abnormal margin, however, this effect disappears. Instead, corporate governance activism outperforms by 9.5% (t = 0.57).

- 38 -

Activism Adds Value Beyond Selling Companies But Selling Generates Very High Returns
40% Sold Firms All Firms 30%

20% All Firms ex. Sale 10%

0%
-1 13 M D on F th 1 ilin M g on th 3 M on th s M on th s M on th s on th s on th s on th s on th s M 21 24 on th s M

12

15

Figure 9. The chart plots the average excess return around the schedule 13D filing, from 1 month prior to the filing until 24 months afterward. The chart plots active investments only. The solid line includes all active investments. The long dashes show only investments which were subsequently sold. The short dashes excludes any investment which was sold as well as any investment in which the activist sought a sale but failed.

Most shockingly, sale activism, which generated the highest short-term returns, generates no long-term returns. This result is most likely due to failed sales. Especially if activists try to sell wounded companies, those remaining unsold will dramatically underperform. To counter this effect, in unreported results, I run the same regression with a dummy variable set to one if the company was sold. As expected, successful sales earn positive returns (9.3% with t = 0.48). To test Greenwood and Schors (2008) hypothesis that activists add value exclusively by putting firms in play, I re-run the event study excluding those firms which the hedge funds try to sell. Unlike Greenwood and Schor, who asymmetrically remove only successful sales, I remove both successful and failed sales. By eliminating the bias from failed sales, this better measures

- 39 -

18

Long-Term Returns Decline Over Time As Investments Increase in umber

175% 125% 75% 25% -25%
20 04 20 06 20 01 20 00 20 02 20 03 20 05 20 07 20 08

180 135 90 45 0

Active Passive

Active Filings

Passive Filings

Figure 10. The line graphs (left axis) show show the average 25-month return by year, as measured by the coefficients in a dummy variable regression controlling for Tobin's q and margins. The blue line plots the average active return, and the red line plots the average passive return. The blue bar (right axis) shows the number of active Schedule 13Ds and the red bar (right axis) shows the number of passive Schedule 13Ds filed in my sample each year.

hedge funds ability to add value beyond selling the target. In stark contrast to Greenwood and Schor, I find that activism still generates significant excess returns. Figure 9 shows that while they underperform firms which were sold, these targets still generate 22.3% returns over 25 months (t = 5.79). This result differs from Greenwood and Schor both because I exclude failed sales and because my sample includes more experienced hedge funds. (C) Returns Have Decreased Over Time Finally, I measure how these returns have changed over time. I regress the 25-month abnormal returns against dummy variables for each filing year, controlling for Tobins q and margins. The results are robust to other models. Figure 10 plots the results over time. I reject the hypothesis that annual returns equal each other (F = 2.56). To the contrary, returns have declined over time; in 2001, the average investment earned 23.1%, and in 2007 the average investment earned only 7.6%. This pattern suggests the same three interpretations: (1) investors might

- 40 -

increasingly expect activism, (2) activism might begin to look more like passive investing, and (3) activists might have lost their edge. As evidence against the second hypothesis, however, I cannot reject the hypothesis that passive returns equal each other in every year (F = 0.27). In summary, I find further evidence that activism adds value. Over the long-run, activist investments outperform passive investments, even adjusting for undervaluation and underperformance. Activism aimed at undervalued and underperforming firms earns the greatest excess returns. Recently, however, these returns have declined.

Section VI: Target Firm Fundamentals Having observed that activism earns greater returns than stock picking might imply, we now turn our attention to target firm fundamentals. By adding further evidence that activism adds value, this section strengthens earlier results. Perhaps more importantly, it also describes the levers activists pull to earn those returns. I document four non-mutually exclusive ways that activist hedge funds improve target firms. First, activists improve profitability. Both relative to passive targets and in absolute terms, activist targets increase their margins, return on assets, and return on equity. Second, activists optimize the capital structure, often by disgorging cash. Activist targets increase their leverage and payout ratios. Third, activists reduce firm size. While passive target firms actually grow, activist target firms shrink dramatically. Fourth, activists sell target firms. Relative to passive targets, activist targets sell themselves more frequently and for higher premiums. Throughout, I also find evidence that activists use board seats to build long-term valueconsistent with the evidence presented in Section III that business strategy activism most frequently seeks board seats while M&A activism least frequently seeks board seats.

- 41 -

Before turning to the data, a cautionary note is in order. Identifying firms which outperform by accounting standards presents a more difficult task than identifying those which outperform stock market indices. The beauty of financial markets lies in their ability to correctly price target firms fundamentals, not in their ability to equalize those fundamentals; no efficient markets hypothesis governs accounting performance. Thus, there is no reason that even a passive hedge fund should not be able to consistently pick stocks whose profitability increases more than does the aggregate market. To the contrary, by simply investing in stocks with high P/E ratios, for instance, a hedge funds portfolio might regularly outpace its peers in revenue growth. Thus, we must distinguish activist hedge funds ability to improve their targets performance from activist hedge funds ability to pick stocks that everyone expects will increase profitability. Three factors mitigate this stock picking hypothesis. First, activists target wellperforming firms.39 To the extent certain companies can improve more than others, it shouldnt be well-performing firms but rather underperforming firms with more room to improve.40 Second, activists target undervalued companies. The companies expected to outperform should not be undervalued but rather overvalued to account for yet-unrealized profits. And third, excess returns increase over two years. If the market had predicted that target firms would improve, excess returns should not rise over time but rather should equal zero. Thus, hedge funds dont seem to simply pick stocks that everyone expects will increase profitability. In addition to arguments about target firms characteristics, I further debunk the stock picking hypothesis by comparing hedge funds active targets to the same hedge funds passive targets. If active and passive targets come from the same underlying pools, then hedge fund activism must cause any outperformance by the activist group. As discussed earlier, though, the
39

As noted in Section III, relative to passive targets, activist targets are poor performers, but relative to the median firms in their industries, activist targets are good performers. 40 I find significantly negative correlations between initial levels and subsequent improvements.

- 42 -

Table IX Profitability Improvement By Investment Type

The table shows the 1- and 2-year increases, from the date of the first Schedule 13D filing, in target firms' trailing twelve months EBITDA divided by sales (Margin) , EBITDA divided by total assets (ROA) , and EBITDA divided by book value of equity (ROE) . All data from Bloomberg. All data is winsorized at the 10% level. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active vs. Passive Active 1 yr Margin ROA ROE -0.44 0.27 2.48 2 yr 0.73 2.28 8.29 1 yr -1.20 -1.84 -4.73 Passive 2 yr -3.38 -2.87 -8.70 Difference 1 yr 2 yr 0.76 2.11 *** 7.21 *** 4.11 *** 5.15 *** 16.99 ***

Panel B: Board Seats vs. Other Active Investments Board Seats 1 yr 2 yr Margin ROA ROE 0.21 0.54 0.95 0.97 2.41 6.13 No Board Seats 1 yr 2 yr -1.59 -0.33 2.92 -0.08 1.02 8.71 Difference 1 yr 2 yr 1.80 ** 0.86 -1.97 1.05 1.39 -2.58

pools differ. Most noticeably, hedge funds go active in more undervalued companies. For the reasons discussed above, this should handicap activist targets ability to outperform. Because I nonetheless observe outperformance, this does not appear to be a problem. Additionally, hedge funds go active in worse performing companies. While this might give activist investments more room to improve, after controlling for this effect, I still observe that activist targets outperform. (A) Activism Improves Profitability By improving target firms capital allocation and strategic positioning, hedge funds increase cash flowand thus, firm value. Management might fail to do so for many reasons. For example, CEOs paid based on earnings growth might increase earnings by investing in projects with IRRs greater than zero but less than the firms WACC. Just as bad, CEOs paid based on ROA might forgo projects with attractive IRRs if they drag down a firms overall returns.

- 43 -

Active Investments Become More Profitable Controlling For Initial Levels Return on Assets
3.0% 2.0% 1.0% 0.0% -1.0% -2.0% Active Passive Board No Board 2.0% 1.0% 0.0% -1.0% -2.0% -3.0% Active Passive Board No Board

EBITDA Margins

Figure 11. The chart shows the average 2-year improvement in ROA and in margins for active, passive, board, and no board investments, controlling for Tobin's q and for the initial levels of each variable.

Table IX measures increased profitability in three ways: EBITDA margins, ROA, and ROE. The former two measures are particularly appropriate because they isolate core operational performance.41 Panel A shows that, while passive investments falter, activist targets improve their profitability in absolute terms. Further, over two years, activist targets improve their margins by 4.1% more than passive targets (t = 4.26), their ROA by 5.2% more (t = 4.49), and their ROE by 17.0% more (t = 4.07). Panel B shows similar, but less significant results for board seats; they improve margins and ROA. While they have little effect on ROE, this is because board seats decrease target firms leverage. Figure 11 dismisses the alternate hypothesis that this outperformance comes from stock picking. I regress the 2-year increase in ROA and in margins on dummy variables set to one for activist investments or board seat investments. I control for the Tobins q and for the initial value of each accounting metric. I express all non-dummy covariates as the deviation from the sample
41

ROE includes capital structure decisions. It equals ROA divided by one minus the debt-to-assets ratio.

- 44 -

mean. We interpret the constant term as the profitability increase to an average passive target, and we interpret the dummy coefficient as the difference for an average active target.42 While the average passive firm becomes less profitable, with ROA falling by 1.2% (t = -1.70) and margins falling by 2.5% (t = -3.71), the average active firm becomes more profitable. The average active target increases its ROA by 3.0% more than the average passive target (t = 3.05), and the average active target increases its margins by 2.9% more (t = 2.76). Board seats show but less significant similar results. Overall, activists increase firms profitability. (B) Activism Optimizes Capital Structure By optimizing a firms capital structure, activists increase the cash flow to shareholders. First, by increasing leverage, and thus interest expense, activists can reduce taxes. Second, by disgorging cash otherwise earning at best middling returns, activists can increase the present discounted value of the firms equity cash flows. Management might fail to optimize the capital structure for many reasons. For example, because CEOs depend on the firms for their jobs, they might take more caution than shareholders might like before leveraging the firm. Table X tracks changes in target firms capital structures. Leverage refers to the debt-toassets ratio and payout measures dividends as a percent of net income. In addition to the average change in each ratio, the table also shows the percent of firms which increased these ratios. Panel A shows that activist targets increase their leverage and payout ratios more than passive targets. The results are much stronger over one-year horizons than two-year horizons. This makes sense; unlike profitability measures, which ideally increase over the years, these capital structure ratios fluctuate. Leverage naturally declines over time as firms pay down their debt, and payout ratios might bounce around with special dividends. Over one year, activist targets increase their leverage by 1.1% more than passive targets (t = 1.65) and increase their payout ratio by 4.5%
42

In Appendix D, Table D-7 contains the regression results.

- 45 -

Table X Capital Structure Changes By Investment Type

The table shows the 1- and 2-year increases, from the date of the first Schedule 13D filing, in trailing twelve months debtto-assets (Leverage) and dividends divided by net income (Payout) . The chart shows both the mean difference in the ratio as well as the percent of firms which increased their ratios. All data from Bloomberg. All data is winsorized at the 10% level. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active Investments Lever Capital Structure More Than Passive Investments Active 1 yr Leverage % Increase Payout % Increase 0.64 43.6% 5.37 32.6% 2 yr 1.51 45.3% 0.53 18.5% 1 yr -0.45 35.4% 0.91 12.7% Passive 2 yr 1.56 42.2% 10.00 16.7% Difference 1 yr 2 yr 1.09 ** 8.2% 4.46 ** 19.9% *** -0.05 3.1% -9.47 ** 1.9%

Panel B: Board Investments Lever Capital Structure Less Than Non Board Investments Board Seats 1 yr 2 yr Leverage % Increase Payout % Increase -0.75 33.8% 7.36 26.7% 0.72 40.0% 0.76 17.6% No Board Seats 1 yr 2 yr 2.08 58.1% 3.58 40.0% 2.86 55.0% -0.75 20.0% Difference 1 yr 2 yr -2.83 ** -2.15 -24.2% ** -15.0% 3.78 -13.3% 1.51 -2.4%

more (t = 2.08). In addition, 8.2% more active firms increased their leverage than did passive firms (t = 1.25), and 19.9% more active firms increased their payout ratios than did passive firms (t = 2.38). Thus, activist targets lever up more than their passive peers. Panel B shows that investments seeking board seats take on 2.8% less leverage (t = 2.21) than other activist investments. Two mutually exclusive reasons might account for this observation. Most consistent with results elsewhere, activists might seek board seats to provide long-term strategic advice. In Section III, I showed that activists most frequently take board seats to advance business strategy goals and least frequently to advance M&A goals. These activists might focus their efforts on building value before taking on additional financial risk. Alternatively, activists might seek board seats when they anticipate management will otherwise dismiss their proposals. These stubborn, entrenched management teams might resist leverage.

- 46 -

Active Investments Increase Payout Controlling For Initial Levels and Initial Cash Payout Ratio
3.0% 2.0%

Leverage

2.0%

1.0%

1.0%

0.0%

0.0% Active Passive Board No Board

-1.0% Active Passive Board No Board

Figure 12. The chart shows the average 1-year change in payout ratio and in the debt-to-assets ratio for active, passive, board, and no board investments, controlling for Tobin's q , initial levels of each variable, and initial cash holdings.

Figure 12 shows that, once again, these results are due to activism, not to stock picking. I control for Tobins q, the initial value of each accounting metric, and the initial level of cash as a percentage of total assets.43 While the average passive firm marginally increases its payout by 2.3% (t = 1.79) and decreases its leverage by 0.4% (t = -0.83), the average active firm increases both payout and leverage. Payout increases by 2.5% more than passive firms (t = 1.26), and leverage increases by 0.7% more than passive firms (t = 0.98). Thus, it appears activism earns some of its gains by disgorging cash and leveraging the capital structure. (C) Activism Reduces Firm Size Activist shareholders also add value by counteracting managements harmful tendency to build empires. For CEOs, larger firms lead to greater compensation and greater job security. To grow the firm, they might purchase assets which yield suboptimal returns. Thus, by divesting unprofitable assets activists can improve future returns.
43

In Appendix D, Table D-8 shows more detail.

- 47 -

Table XI Agency Cost Reductions By Investment Type

The table shows the 1- and 2-year increases, from the date of the first Schedule 13D filing, in inflation-adjusted firm size, measured by assets, sales, and cash. Each variable is expressed as a percentage of its value in the year prior to the Schedule 13D filing. All data from Bloomberg. All data is winsorized at the 10% level. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active Investments Shrink More Than Passive Investments Active 1 yr Assets Sales Cash -4.8% -2.4% 22.1% 2 yr -6.5% 0.0% 35.2% 1 yr 0.0% 1.7% 30.5% Passive 2 yr 0.1% 1.4% 24.2% Difference 1 yr 2 yr -4.8% *** -6.6% *** -4.2% *** -1.4% -8.3% 11.0%

Panel B: Board Investments Shrink More Than Non-Board Investments Board Seats 1 yr 2 yr Assets Sales Cash -7.0% -4.6% 26.0% -9.6% -3.1% 27.9% No Board Seats 1 yr 2 yr -1.4% 0.3% 20.6% 0.5% 5.3% 50.0% Difference 1 yr 2 yr -5.6% *** -10.1% *** -5.0% *** -8.5% ** 5.4% -22.1%

Table XI examines this hypothesis. It shows assets, sales, and cash, adjusted for core inflation, expressed as a percent change from their values in the year prior to activism. Panel A shows that while passive target firms grow, active target firms actually shrink. On average, active targets reduce their assets by 4.8% more than passive targets over one year (t = -3.23) and by 6.6% more over two years (t = -2.36). Similarly, activist targets reduce their sales by 4.2% more than passive firms (t = -2.89) over one year and by 1.4% more (t = 0.58) over two years. While active targets also appear to reduce their cash by 8.3% more than passive firms over one year (t = 1.04), they appear to increase their cash by 11.0% more over two years (t = 1.11). The inconsistent evidence likely reflects two conflicting forces: on one hand, activist target firms are shrinking, but on the other hand, they are also increasing profitability. Panel B shows that board seat investments shrink the most. Investments where activists seek board seats shrink their assets by 5.6% more than other activist investments (t = -2.80) over

- 48 -

Active Investments Reduce Assets Controlling For Initial Assets Asset Growth
4.0% 2.0% 0.0% -2.0% -4.0% Active Passive Board No Board 6.0% 4.5% 3.0% 1.5% 0.0% Active Passive Board No Board

Sales Growth

Figure 13. The chart shows the average 2-year change in assets and sales, adjusted for inflation, for active, passive, board, and no board investments, controlling for Tobin's q and for the initial firm size.

one year and by 10.1% more (t = -2.55) over two years. They also shrink their sales by 5.0% more (t = 1.04) over one year and by 8.5% more (t = 1.97) over two years. To isolate activism from stock picking, Figure 13 controls for differences Tobins q and in firm size.44 While passive targets actually increase assets by 2.8% (t = 1.81), active investments decrease assets by 1.8%, shrinking 4.6% more (t = -2.12). Similarly, active investments without board seats actually increase assets by 0.8% (t = 0.34), but board seat investments decrease assets by 3.7%, shrinking 4.4% more (t = -1.49). The changes in sales show similar patterns; while all investments grow their sales, the activist investments seeking board seats grow the least. Taken together with the earlier evidence that activist firms increase profitability, it appears that activist investments purposefully pare down, rather than unintentionally shrink into irrelevance. Thus, activist investments, particularly those with board seats, effectively reduce firm size. As Jensen (1986) argues, this decreases agency costs.
44

In Appendix D, Table D-9 shows more details.

- 49 -

Table XII Target Firm Sales By Activism Type

The table shows the percent of all firms which sold themselves and the average premiums they received. Average premium is calculated from one month before the hedge fund files a Schedule 13D until the date on which the sale closes. Panel A separates active investments from passive investments. Panel B separates active investments seeking baord seats from those which do not. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active Investments Sold More Frequently Active 6 mo % Sold Avg. Premium umber Sold 9.2% 20.4% 24 24 mo 17.9% 39.8% 47 Passive 6 mo 0.0% 0 24 mo 4.2% 21.5% 9 Difference 6 mo 9.2% *** 24 mo 13.7% *** 18.3% *

Panel B: Investments With Board Seats Sold Less Frequently, But For Higher Prices Board Seats 6 mo % Sold Avg. Premium umber Sold 2.2% 44.0% 3 24 mo 10.4% 57.9% 14 No Board Seats 6 mo 19.3% 16.8% 21 24 mo 30.3% 31.9% 33 Difference 6 mo -17.0% *** 27.1% *** 24 mo -19.8% *** 26.1% **

(D) Activism Sells Target Firms Finally, activists add value by selling their targets to acquirers who values them more highly. Executive managers might fail to sell the firm, even if it benefits shareholders, because doing so imperils their job security. Table XII measures how well activists sell their targets. First, it measures the percent of targets which sell themselves, and second, it measures the average premium those firms fetched. To capture the premium attributable to the activist, I measure the total buy-and-hold gross return from one month before the Schedule 13D until the sale closes. Panel A shows that, relative to their passive investments, hedge funds activist targets sell themselves more frequently and for higher prices. Within the first 6 months, activists sell 9.2% of their targets, while passive investors sell none. Over two years, activist investors sell 13.7% more targets (t = 4.93) and on average sell them for 18.3% higher premiums (t = 1.64). Thus, activists both put their targets into play and market them better than do passive investors. Panel B shows that when activists seek

- 50 -

board seats, they sell firms less frequently but for much higher prices. Again, this result is consistent with the interpretation that activists use board seats to build long-term businesses. When doing so, theyre less likely to sellin particular, 19.8% less likely over two years (t = 3.77). But when they do sell, having built a strong business, they demand a healthy premium on average, 26.1% larger (t = 2.10). In total, relative to passive investments, activism unlocks substantial value by selling the target firm. (E) Improvements Differ by Strategy If hedge fund activism, rather than sheer luck, catalyzes these observed changes, we should expect activists professed strategies to predict target firms improvements with reasonable accuracy. This section tests that hypothesis. In particular, I describe the differences among the various activist strategies. To describe these strategies, I run cross-sectional regressions like those in Section IV and Section V. For all investments which use activist strategies, I regress the accounting statistics described earlier on dummy variables for each strategy. When applicable, as described earlier in this section, I control for the dependent variables initial value. To facilitate interpretation, all non-dummy covariates are expressed as the deviation from the sample mean, and all constants are suppressed. Thus, we interpret the coefficients on each dummy variable as that strategys partial effect on a firm with average characteristics. In general, I find strong evidence that different strategies work in different ways. Figure 14 shows the key results.45 Business strategy activism builds profitable businesses. Over two years, ROA increases 6.9% (t = 2.71) and margins increase 2.3% (t = 0.91). Perhaps to remove distractions from building these businesses, activists also reduce target firms payout ratios by 10.0% (t = -1.78)

45

In Appendix D, Table D-10, Table D-11, Table D-12, and Table D-13 show more detail.

- 51 -

Panel A: Business Strategy Activism Improves ROA

12%

F = 2.74 6%

0%

-6% Business Strategy M&A Sale Capital Structure Corporate Governance

Panel B: Capital Structure Activism Increases Payout

20% F = 1.39 10%

0%

-10%

-20% Business Strategy M&A Sale Capital Structure Corporate Governance

Panel C: Corp Gov & Sale Activism Decrease Assets

10% F = 1.11 5%

0%

-5%

-10% Business Strategy M&A Sale Capital Structure Corporate Governance

- 52 -

Panel D: M&A Activism Sells Firms Most Frequently Corporate Governance Activism Earns Highest Premiums
250% 50%

200% Sale Odds Ratio Sale Odds (left axis) 150% Sale Premium (right axis)

40% Averarge Premium

30%

100%

20%

50%

10%

0% Bus Strat M&A Sale CapStruc CorpGov

0%

Figure 14. These charts plot the results of cross-sectional regressions on various accounting metrics: ROA (Panel A), payout ratios (Panel B), and asset growth (Panel C). I also show cross-sectional variation of firm sale statistics (Panel D). For all activist investments which use activist strategies, these dependent variables are regressed on dummy variables for each strategy. Where applicable, regressions control for the dependent variable's initial level. All non-dummy covariates are expressed as the deviation from the sample mean, and all constants are suppressed. Panel A, Panel B, and Panel C show 90% confidence intervals and F statistics on the null hypothesis that all strategies generate equal improvements. In Panel D, the blue bars (left axis) show the sale odds ratio from a logistic regression that regresses a dummy variable equal to one if the firm sold itself on strategy dummies. The red outlines (right axis) show a linear regression that regresses the average premium on the strategy dummies.

and leverage by 1.2% (t = -0.73). Presumably due to increased profitability and lower payout ratios, target firms assets increase 3.2% (t = 0.77) and sales increase 4.6% (t = 1.14). M&A and sale activism effectively push target firms to sell themselves; M&A activism sells firms more frequently, and sale activism sells them for higher prices. Within investments that use activist strategies, those in which the hedge fund uses a M&A-related strategy are 2.3 times more likely to sell themselves as those which do not (t = 1.74). When a hedge fund demands that a target firm sell itself, it earns on average a 24.6% premium (t = 2.04). When hedge funds fail to sell these firms, they often seek divestitures, reducing total assets by 4.5% over two years (t = -1.43).

- 53 -

Capital structure activism increases cash to investors. Target firms increase their payout ratio by 10.4% (t = 2.21). Interestingly, leverage actually falls by 1.8% (t = -1.8), perhaps reflecting heterogeneity even within the strategywhile some hedge funds want the firm to leverage itself, others want the firm to de-lever a precarious balance sheet. Corporate governance activism reduces agency costs. Target firms reduce their assets by 4.0% over two years (t = -1.10). As Jensen (1986) argues, management teams which ignore shareholder interests may seek to grow assets at the expense of profitability. Activist hedge funds seem to take particular interest in reducing firm size when they distrust management. (F) Activist Ability Has ot Declined In unreported results, I find no evidence that activists abilities have declined over time. I regress the changes in activist targets accounting measures on dummy variables for the Schedule 13D filing year, controlling for Tobins q and for the accounting variables initial levels. I conduct joint hypothesis tests on the null hypothesis that the yearly coefficients all equal one another. While I find that activists ability to reduce assets has changed over time (F = 2.71), it hasnt decreased. Early in the decade, activists were very effective at reducing firm size, towards the middle of the decade, activists struggled, and towards the end of the decade, activists regained their effectiveness. No other accounting measures provide any significant results. Nor do I find evidence that passive returns have changed over time. Thus, I find no support for the hypothesis that activist skill has declined. To summarize, I find manifold evidence that activism adds value. Target firms improve their profitability, optimize their capital structure, reduce their potential agency costs, and sell themselves more often and for richer valuations. In addition, activism adds value in the pattern we would expect given hedge funds stated strategies. I find much more significant evidence that

- 54 -

activist hedge funds improve target firms than do previous studies.46 While this is in part due to my unique sample, it is also because I dissect these improvements by strategy. In contrast, prior studies looked at the results in aggregateso different strategies canceled each other out.

Section VII: Conclusion This paper argues that activist outperformance represents more than stock picking in disguise. By comparing activist investments to passive investments while controlling for ex-ante differences in target firms, I isolate hedge fund activism from hedge fund investing. Activism earns substantial short-term and long-term returns by unlocking shareholder value; in particular, activists revive struggling firms. They improve profitability, increase payout, spin off assets, and sell target firms. Activists also tailor their approach to particular firms needs, focusing on improving business strategies, advising on M&A transactions, selling target firms, optimizing capital structures, or fixing corporate governance. The new evidence presented in this paper expands the existing literature on activism. In particular, while prior studies have shown that activist hedge funds earn excess returns, I present strong evidence they earn these returns by doing more than picking stocks. I use several unique tests to distinguish activism from passive investingI compare activist investments to passive investments, I control for ex-ante differences in target firms, I study activist instruments, and I identify short-term events which contain only activist signals. Each test strongly supports the hypothesis that activism adds value. This paper has focused on the returns to corporate shareholders, sometimes at the expense of measuring the returns to hedge fund investors. Most notably, I have made no adjustments for

46

See, e.g., Boyson and Mooradian (2008), Brav et al. (2008), and Klein and Zur (2008).

- 55 -

hedge fund fees. Thus, while I have shown that activism benefits shareholders, further research might study whether it similarly benefits hedge funds limited partners. In a broader context, this paper presents several lessons that might call corporate boards to activism themselves. Given hedge funds ability to improve shareholders returns, perhaps corporate boards should encourage hedge funds to invest. I also present evidence that expert advice can turn around struggling firms. Perhaps undervalued companies should more frequently hire outside consultants to revisit their policies on capital allocation, capital structure, M&A, and corporate governance. Further research should investigate whether activist hedge funds follow any predictable patterns that corporate executives might mimic. Additionally, I show that agency costs detract from a firms value. In particular, why can hedge funds unlock value beyond managements ability even though hedge funds invest in many companies, while managers devote their whole lives to just one firm? Most likely, this is due to agency costs: whereas the hedge fund activists benefit from stock returns, the CEO benefits from other factorslike compensation, pride, and job security. Perhaps corporate boards might increase their firms values by compensating their CEOs less like CEOs and more like hedge fund managers. Further research might investigate whether activists earn greater returns in firms where managers interests are most poorly aligned with shareholders interests.

- 56 -

Appendix A: Statistical Methods (A) Event Studies This essay uses the event study methodology. To determine how certain events affect security returns, I compare observed returns to what they would have been without that event. I calculate these counterfactual returns using two popular models: the market model (Fama 1970) and the Fama French three factor model (Fama and French 1992). Both models assume that investments with greater risk offer greater reward. First, over a measurement window, I calculate the firms risk. Then, over the event window, I assume the firm performed like other firms with similar risk levels. The market model assumes we can measure risk by a firms correlation with the overall market. It also assumes asset returns follow a normal distribution (MacKinlay 1997). Rit = i + i Rmt + it E ( it ) = 0 Var ( it ) = i
2

In this model, Rit is the return to security i, and Rmt is return to the valued-weighted market portfolio in period t. i measures the securitys correlation with the overall market. i represents the idiosyncratic risk associated with a specific company. The error term it has an expected mean of zero, and the variance of it is assumed to be homoskedastic with a normal distribution. Fama and French (1992) recommend augmenting the market model to include more possible risksin particular, the premiums and discounts awarded to firms based on their market capitalization and ratio of book value of equity to market value of equity (B/M). Small firms tend to be more risky and participate more frequently and for longer periods of time in earnings downturns. Companies with high B/M ratios often suffer low earnings yields on their assets. To

A1

incorporate these factors into the market model, Fama and French suggest running time-series regressions to determine the premiums and discounts associated with market capitalizations and B/M ratios. They propose a three factor model: Rit = i + i Rmt + si SMBt + hi HMLt + it E ( it ) = 0 Var ( it ) = i
2

As before, Rit is the return on security i and Rmt is the return on the value-weighted index at time t. i is the intercept term. The two new terms represent the Fama French factors, where SMB represents the difference in returns on portfolios of small and big market capitalization equities, and HML is the difference between returns of portfolios separated by B/M ratios (Fama and French 1992). Although these models are similar, most activism studies use the simpler market model.1 Thus, I report results using the market model. To test for robustness, I check my results against the Fama French three factor model. To estimate the parameters, I follow two steps. First, I run times-series regressions over an estimation window prior to the event. In this study, for daily returns I run regressions over the 200 day window ending 10 days before the event, and for monthly returns I run regressions over the 36 month window ending one month before the event. In the diagram below, T-2 and T-1 represent the start and end of the estimation period, respectively. T0 represents the event day. T1 represents the end of the event window.

T-2

T-1

T0

T1

Estimation Window

Event Window

See Puustinen (2007), Boyson and Mooradian (2008), Brav et al. (2008), and Klein and Zur (2008).

A2

Second, I measure abnormal returns. I use the coefficient estimates from the time-series regression to estimate counterfactual returns over the event window. Abnormal returns simply equal actual returns minus expected returns . Under the null hypothesis, abnormal returns have conditional mean zero and conditional variance

, where 200 represents estimation periods length. To test returns over more than one day or one month, I calculate abnormal returns (CARs) over the event window. For any two times t1 , t 2 , CAR( t1 , t 2 ) = .

This paper tests whether the mean CARs among a portfolio of activist investments differ significantly from zero, where and .

To test for statistical significance, the paper will calculate the variance of the mean CAR , where N is the number of events, and . This equation holds when the estimation period length T is large, such that

approaches zero (Dasgupta, Laplante, and Mamingi 1997). A3

(B) Dummy Variable Regressions This paper also makes extensive use of dummy variable regressions to analyze binary variables like an activists strategy. First, I use dummy variable regressions to analyze whether investments which I consider active enjoy greater returns or accounting outperformance than other investments. To do so, I use a regression of the form

where Y i is the return to investment i, and the X ij s represent other covariates. I express all nondummy covariates as the deviation from the mean. For a passive target firm whose covariates equal sample averages, returns equal averages earns . An activist target firm whose covariates equal sample

^ greater returns. If is significantly greater than zero, I conclude that activism

generates significantly greater returns than passive investing. Second, I use dummy variable regressions to measure the average partial returns on various binary variablesmost frequently, binary variables for various activist strategies. I use a regression of the form

where Y i and the X ij s have the same interpretations. I express all non-dummy covariates as the deviation from the mean. Because I have suppressed the constant term, I interpret as the

average partial return to a target firm whose covariates equal the sample averages in which the activist hedge fund uses strategy j. For example, if firm is covariates equal the sample averages,

A4

and if a hedge fund targets firm i with strategy 2 and strategy 3, then all the all the s equal zero, except and

s equal zero, and .

equal one, so we expect the return to be

A5

Appendix B: Excerpted Hedge Fund Letters To elaborate on the distinctions among activist strategies, this section excerpts various hedge fund disclosures. The first section below presents a typical letter announcing that a hedge fund will seek board seats. The next five sections describe business strategy activism, M&A activism, sale activism, capital structure activism, and corporate governance activism, respectively.

Investments Seeking Board Seats In a letter to Motorola, Carl Icahn wrote1: You stated during todays conference call, we discussed Board Nominees with Carl Icahn and we proposed two nominees and he declined. Again this is only partially true. It is true that Sandy Warner, head of the Nominating Committee called me and offered seats to two of my Nominees if I would drop the proxy fight. However, you failed to mention in your conference call that I told Mr. Warner that I would gladly accept this offer if the Board would also accept Keith Meister. Mr. Warner replied summarily to this offer that Meister did not qualify. I asked Mr. Warner what does one have to do to qualify lose $37 billion dollars? Mr. Warner then replied that the Board did not know Meister. My answer was that Meister would fly anywhere at any time to meet the Board so they could know him (I did mention that the situation at Motorola is too serious for the Board to remain a country club). My offer to Motorola stills stands. You have stated to the press that our request for information about what steps the Board actually took to correct the problem at Motorola is an unnecessary distraction. We disagree. In a political election when constituents believe their representatives performance was inadequate, they are certainly not denied information as to whether their representative acted in a grossly negligent fashion. Why should it be different in Corporate America? I do however agree with you that this proxy fight is a distraction that Motorola at this junction can ill afford. If as you have stated, we all want to benefit the stockholders of Motorola, then what possible reason is there for not putting Keith Meister on the Board. After all, how much can he eat at the Board meetings? On a positive side, having a highly intelligent, energetic individual like Keith, who has 145 million reasons to spend his time working toward the spin-off being accomplished, may well make this promise come true in a timely fashion. We ask
1

Motorola Inc., SEC Schedule 13D, filed 27 March 2008.

B1

the Board meet with Meister, put egos aside and lets get on with the urgent business at hand.

Business Strategy Activism In a letter to Pier 1 Imports, Elliot Associates wrote2: Elliott is puzzled by the pace of cost-cutting and restructuring actions taken by the Company over the past two years. Whereas we believe a sense of urgency is appropriate, the Company has reacted at a glacial pace to the intensified competitive landscape that has developed in the US home furnishing retail industry over the past five to ten years. It is imperative that the Company immediately focus on executing cost-cutting and restructuring initiatives in order to stem its current cash bleed. We suggest Pier 1 close down a significant number of stores, right-size the corporate and divisional SG&A, lease part or all of the corporate headquarters, and add new independent members to the Board of Directors. Elliott believes that these initiatives, if executed properly, will yield immense value to shareholders by creating a leaner, profitable company fit to navigate the intense competition of the future.

M&A Activism In a letter to Brinks Co., MMI Investments wrote3: We understand that BAX operations might be further improved, but conclude that this is insufficient grounds to retain them. The depressed market capitalization of BCO far outweighs any benefit of holding BAX. Furthermore, we believe that by waiting until 2006 to divest BAX in the hope it will earn its cost of capital would be making a decision to trade the certainty of the currently robust M&A and freight markets for the uncertainty of BAXs hypothetical ability to earn its cost of capital. We believe BCO is currently in a brief cyclical window of robustness in world freight markets and strong earnings for BAX, coupled with low-interest rates and a vibrant M&A market. Any concerns about losing key people during a sale process can be addressed with meaningful divestiture incentive bonuses. This happens all the time and we would support such extra compensation for BAX executives in the event of a successful exit, whether that results in one or more transactions or a shut-down/liquidation. Most divestitures today, we believe, are achieved via an organized process, which includes a qualified investment banker with a board-approved mandate to approach qualified prospective buyers. Waiting for the buyers to come will likely result in no transaction.

2 3

Pier 1 Imports Inc., SEC Schedule 13D, filed 9 April 2007. Brinks Co., SEC Schedule 13D, filed 20 April 2005.

B2

It is rare in life or business that the sun, moon and stars all align. We believe that BCO has such an occasion at this moment: a win-win transaction for exiting an undesirable subsidiary and funding a legacy liability, and a highly favorable environment in which to do so. We conclude that it is crucial for the future of BCO that a buyer is sought promptly. We would welcome the opportunity to discuss this or present further details to the Board.

Sale Activism: In a letter to Gehl Co., Newcastle Partners wrote4: We were surprised to learn by press release of the rejection by Gehl Company (the Company) of the offer made in a letter to you dated December 22, 2000 from Newcastle Partners, L.P. and CIC Equity Partners, Ltd. (collectively, the "Interested Parties") to acquire the Company, as described in such letter, for $18.00 per share in cash. Common sense aside, given that the offer represented a 67% premium to the Company's then current market price of $10.75 per share, we believe that the offer warranted, at minimum, a phone call to the Interested Parties to address any questions or concerns of the Company's Board of Directors. Instead, we were disappointed to see that you failed to return any of the phone calls made to you by the Interested Parties in connection with such offer, or in any way seek any additional information about the offer. The Board's rejection of the offer, which rejection was announced in a press release only a few hours after it was delivered to you, clearly indicates to us that the offer was not given serious consideration by the Board. The offer was summarily rejected with no indication that the Board consulted with an investment banking firm or relied on a "fairness opinion" to justify the rejection. Instead of giving any consideration to the offer, it appears that the Board spent their time preparing a press release attempting to belittle the ALL CASH offer. If there are concerns with any of the standard conditions contained in the offer, we invite the Board to contact either of the undersigned to directly address such concerns in a productive manner. The Board's rejection of the offer demonstrates the Board's lack of dedication to pursuing the best interests of the stockholders of the Company and maximizing stockholder value.

Capital Structure Activism In a letter to Ikon Office Solutions, Steel Partners wrote5: We commend management for its efforts that have resulted in improved operating margins and a simplified capital structure. We also recognize and commend the
4 5

Gehl Co., SEC Schedule 13D, filed 20 March 2001. Ikon Office Solutions, Inc., SEC Schedule 13D, filed 29 June 2007.

B3

Company for expanding the Company's existing share repurchase program. However, we believe that the Company's common stock continues to trade below its intrinsic value and your balance sheet remains highly inefficient. We therefore recommend that the Company capitalize on its operating momentum and utilize the strength of its balance sheet to pursue a public recapitalization at $17.50 a share. This represents a 20% premium to the 20-day closing average of the shares. With the Company's excess cash and a conservative debt financing package, the Company should be in a position to self tender for at least $850 million of common stock, or approximately 39% of its outstanding shares at the aforementioned price per share. Based on our discussions with several leading investment banks, we believe this refinancing is highly achievable in a short period of time. The proposed recapitalization would provide immediate liquidity at a meaningful premium to the current share price for your shareholders who elect to tender. We believe it will be well received by the shareholders who desire liquidity. It would also be accretive immediately, and more so in the long term. As a long term shareholder, Steel would commit not to tender into this offer. Importantly, in addition to lowering the Company's cost of capital and enhancing its EPS growth, this recapitalization would result in a strong and flexible balance sheet that would support the Company's long term strategic vision.

Corporate Governance Activism In a letter to Steel Gas Partners, Third Point Capital wrote6: Sadly, your ineptitude is not limited to your failure to communicate with bond and unit holders. A review of your record reveals years of value destruction and strategic blunders which have led us to dub you one of the most dangerous and incompetent executives in America. (I was amused to learn, in the course of our investigation, that at Cornell University there is an "Irik Sevin Scholarship." One can only pity the poor student who suffers the indignity of attaching your name to his academic record.) Irik, at this point, the junior subordinated units that you hold are completely out of the money and hold little potential for receiving any future value. It seems that Star Gas can only serve as your personal "honey pot" from which to extract salary for yourself and family members, fees for your cronies and to insulate you from the numerous lawsuits that you personally face due to your prior alleged fabrications, misstatements and broken promises. I have known you personally for many years and thus what I am about to say may seem harsh, but is said with some authority. It is time for you to step down from your role as CEO and director so that you can do what you do best: retreat to your waterfront mansion in the Hamptons where you can play tennis and hobnob with
6

Star Gas Partners LP, SEC Schedule 13D, filed 14 February 2005.

B4

your fellow socialites. The matter of repairing the mess you have created should be left to professional management and those that have an economic stake in the outcome.

B5

Appendix C: Hedge Fund Heterogeneity Because it requires substantial skill to improve a firm beyond the management teams ability (or willingness), we might expect substantial variation hedge funds activist success. Especially because 13D Monitor selectively chose the hedge funds for my sample, it is worth investigating whether certain outlier funds unduly influence the aggregate outcomes. This does not appear to be the case. In this section, I analyze hedge fund heterogeneity in (1) short-term returns, (2) long-term returns, and (3) target firm improvements. I find no evidence that outliers have distorted my results. Turning first to short-term returns, I look for outliers in any particular funds activist or passive investments. For all hedge funds with at least three activist observations and three passive observations, Figure C1 plots active returns against passive returns. For all funds above
In Short-Run, Most Funds' Active Investments Exceed Passive Investments Results ot Driven By Outliers
40%
Cannell

30% A ctive Exces s Return

Ramius Icahn Elliot Shamrock

FEF

Stilwell Third Ave Riley SAC Pershing Southeastern

20%

Newcastle Sandell JANA Third Point Harbinger

10%
Greenlight Discovery DE Shaw Wynnefield

Barington

0%

-10%

Nierenberg

-20% -20%

-10%

0%

10% 20% Passive Excess Return

30%

40%

Figure C1. For each hedge fund with at least three activist investments and three passive investments, the chart plots the fund's activist return aginst its passive return, from 30 days prior to each investment until 30 days after. For funds above the 45 line, activist investments outperformed.

C1

In Long-Run, Most Funds' Active Investments Exceed Passive Investments Results ot Driven By Outliers
120%
SCSF

100% 80% A ctive Excess Return

Steel

SAC

60% 40% 20% 0% -20% -40% -40%

Barington

Nierenberg Southeastern Wynnefield FEF JANA Ramius Elliot Riley

Icahn Cannell ValueAct Third Point Discovery

-20%

0%

20% 40% 60% Passive Excess Return

80%

100%

120%

Figure C2. For each hedge fund with at least three activist investments and three passive investments, the chart plots the fund's activist return aginst its passive return, from one month before each investment until 24 months after. For funds above the 45 line, activist investments outperformed.

the 45 line, active investments outperform passive investments. The greater the distance from the 45 line, the more the activist investments outperformed (or underperformed). For 73.9% of the 23 funds, active investments outperformed passive investments. Further, I find few outliers. Coupled with the strong median outperformance, this adds comfort that my prior results were not skewed by outliers. I use the same method to next analyze long-term returns. Figure C2 plots hedge funds active excess returns against passive excess return, from one month prior to the investment until 24 months after. If above the 45 line, a funds activist investments outperformed its passive investments. Interestingly, the short-term outliers have disappeared; Cannell Capitals high flying short-run returns have fallen back to earth, and both Nierenberg and SAC migrate above the 45 line as their activist investments outperform their passive investments over two years. In

C2

fact, for 76.5% of the 17 funds, activist investments outperformed passive investments over the period. This median outperformance adds more comfort that my prior results were not skewed by outliers. While SCSF appears to be a positive outlier, its 109.4% activist return is driven by only four investments. Thus, few outliers coupled with strong median outperformance give no reason to question the earlier results. Finally, I turn to heterogeneity in target firm improvements. Figure A3 repeats the above analysis for changes in ROA (Panel A), leverage (Panel B), and assets (Panel C). Due to a scarcity of observations, I analyze the change in ROA for any fund with at least two activist and two passive observations, but I analyze the change in leverage and the change in assets for any fund with at least three activist and three passive observations. Turning first to profitability, as expected, in 63.6% of hedge funds, their activist targets improve ROA, and in 90.9% of investments, their passive targets reduce ROA. For 81.8% of hedge funds, activist target firms ROA increase more than passive target firms ROA. While firms above the line show much more variance than those below the line, the distribution gives no reason to suspect that outlier funds have skewed the results, especially given the dramatic median outperformance. Leverage shows a similar pattern. As expected, in 57.1% of hedge funds, their activist targets increase leverage, and in 57.1% of hedge funds, their passive targets reduce leverage. For 71.4% of hedge funds, activist target firms leverage increases more than passive target firms leverage. Both due to the strong median performance, and because there appear to be no clear outliers, I find no reason to reject the long-term results discussed earlier. Finally, analyzing the change in assets lends further strength to my earlier conclusions. As expected, activist investments grow more slowly than passive investments; for 69.6% of hedge funds, activist targets reduce total assets, and for 65.2% of hedge funds, their passive

C3

Panel A: 82% of Hedge Funds Improve ROA

4%
Steel

2%
JANA

ValueAct Third Point

Ramius Elliot

Southeastern

Active ROA Increase

0%
Greenlight

Wynnefield

-2%

Icahn Farallon

-4%

-6%

-8% -8%

-6%

-4%

-2% 0% Passive ROA Increase

2%

4%

Panel B: 71% of Hedge Funds Increase Leverage

4%
Wynnefield Steel ValueAct

3% Active Leverage Increase

2%
Ramius

1%
Icahn

0%

Ramius

JANA

-1%

-2% -2%

-1%

0% 1% 2% Passive Leverage Increase

3%

4%

C4

Panel C: 74% of Hedge Funds Reduce Assets

20%

Greenlight

10% Active Asset Growth

Nierenberg SAC ValueAct Third Point

0%

Cannell Barington SCSF Ramius Wynnefield

JANA

Breeden Icahn

Stilwell Southeastern Highland Riley Discovery

FEF

Harbinger

Elliot

-10%

Steel Farallon

-20% -20%

-10%

0% Passive Asset Growth

10%

20%

Figure C3. For various hedge funds, the chart plots changes in ROA, leverage, and assets for activist and passive investments in Panel A, Panel B, and Panel C, respectively. Panel A includes all hedge funds with at least two activist and passive observations. Panel B and Panel C, which had more observations, include all hedge funds with at least three activist and passive observations.

targets increase total assets. In a sample of 23 hedge funds, in 73.9% of funds, activist targets shrink more than passive targets. Thus, I find strong evidence to support the earlier conclusions that hedge funds reduce assets in their activist investments. To give more details about the funds comprising my sample, Table C-1 describes each funds investments by strategy and by instrument. To summarize, despite using a sample comprised of hand-picked hedge funds, I find no evidence that any particular hedge funds have skewed my results. To the contrary, examining returns and improvements on a fund-by-fund level has only strengthened my earlier conclusions.

C5

Table C-1 Hedge Fund Sample

This table describes the hedge funds in my sample. For each fund, the table shows how many filings each fund made over the time period, what percentage of that fund's investments fall into each strategy category, and what percentage of that fund's investments fall into each instrument category.

Total Filings Steel Partners ValueAct Capital Ramius Capital Carl Icahn Third Point Farallon Wynnefield Capital Harbinger Capital Riley Invetment Mgmt Financial Edge Fund Blum Capital Barington Capital Canell Capital JANA Partners SCSF Equities Shamrock Activist Fund Discovery Group Elliot Associates Joseph Stilwell Newcastle Partners Southeastern Asset Mgmt MMI Investments SAC Capital Highland Capital Nierenberg Investment Third Avenue Mgmt DE Shaw Sandell Asset Mgmt Clinton Group Costa Brava David M Knott Greenlight Capital Cresecendo Breeden Capital Pershing Square Blue Harbor New Mountain Vantage Relational Investors Total 51 49 39 37 37 35 29 24 24 23 22 20 20 18 18 17 15 15 15 14 14 13 13 12 12 12 11 10 9 9 9 9 8 7 7 6 5 5 718

Business Strategy 3.9% 4.1% 17.9% 2.7% 13.5% 2.9% 10.3% 12.5% 16.7% 8.7% 4.5% 30.0% 0.0% 11.1% 5.6% 29.4% 0.0% 13.3% 6.7% 0.0% 21.4% 0.0% 0.0% 0.0% 16.7% 8.3% 0.0% 30.0% 11.1% 0.0% 0.0% 22.2% 25.0% 14.3% 14.3% 16.7% 20.0% 80.0% 10.4%

Activist Strategies Capital Corp Sale Structure Governance 19.6% 10.2% 15.4% 37.8% 24.3% 5.7% 10.3% 12.5% 45.8% 21.7% 9.1% 30.0% 15.0% 22.2% 22.2% 5.9% 26.7% 33.3% 33.3% 21.4% 7.1% 0.0% 15.4% 25.0% 0.0% 16.7% 9.1% 40.0% 33.3% 11.1% 0.0% 0.0% 0.0% 0.0% 14.3% 0.0% 20.0% 0.0% 18.1% 13.7% 4.1% 20.5% 5.4% 16.2% 0.0% 0.0% 12.5% 20.8% 13.0% 4.5% 45.0% 5.0% 22.2% 5.6% 52.9% 6.7% 0.0% 26.7% 0.0% 7.1% 7.7% 0.0% 8.3% 25.0% 16.7% 18.2% 40.0% 33.3% 11.1% 0.0% 0.0% 12.5% 14.3% 14.3% 16.7% 20.0% 40.0% 13.6% 11.8% 4.1% 20.5% 2.7% 27.0% 0.0% 13.8% 12.5% 25.0% 17.4% 4.5% 35.0% 0.0% 22.2% 0.0% 76.5% 20.0% 13.3% 6.7% 7.1% 21.4% 0.0% 7.7% 16.7% 33.3% 25.0% 9.1% 20.0% 22.2% 0.0% 11.1% 0.0% 12.5% 14.3% 0.0% 0.0% 0.0% 60.0% 14.8%

Any Strategy 33.3% 18.4% 59.0% 56.8% 56.8% 11.4% 37.9% 29.2% 70.8% 47.8% 13.6% 60.0% 25.0% 55.6% 27.8% 88.2% 46.7% 66.7% 40.0% 35.7% 42.9% 38.5% 53.8% 33.3% 41.7% 66.7% 27.3% 70.0% 77.8% 55.6% 33.3% 33.3% 37.5% 28.6% 71.4% 16.7% 20.0% 80.0% 44.0%

Seek Board 37.3% 16.3% 51.3% 29.7% 29.7% 2.9% 24.1% 16.7% 62.5% 34.8% 4.5% 50.0% 0.0% 44.4% 22.2% 29.4% 20.0% 0.0% 46.7% 71.4% 7.1% 38.5% 7.7% 16.7% 8.3% 0.0% 18.2% 70.0% 33.3% 55.6% 22.2% 22.2% 100.0% 42.9% 14.3% 0.0% 40.0% 60.0% 29.7%

Win Board 23.5% 14.3% 38.5% 18.9% 16.2% 2.9% 20.7% 8.3% 50.0% 17.4% 4.5% 50.0% 0.0% 33.3% 22.2% 17.6% 0.0% 0.0% 33.3% 57.1% 7.1% 23.1% 0.0% 0.0% 8.3% 0.0% 18.2% 50.0% 11.1% 44.4% 11.1% 22.2% 100.0% 42.9% 14.3% 0.0% 40.0% 40.0% 21.7%

Activist Instruments Proxy Buyout Fight 9.8% 10.2% 5.1% 24.3% 2.7% 5.7% 3.4% 0.0% 20.8% 0.0% 0.0% 5.0% 0.0% 11.1% 16.7% 0.0% 0.0% 26.7% 0.0% 21.4% 0.0% 0.0% 7.7% 8.3% 0.0% 0.0% 0.0% 0.0% 0.0% 11.1% 0.0% 0.0% 0.0% 0.0% 14.3% 0.0% 0.0% 0.0% 6.7% 13.7% 2.0% 17.9% 0.0% 8.1% 0.0% 3.4% 4.2% 25.0% 13.0% 0.0% 5.0% 0.0% 16.7% 0.0% 17.6% 20.0% 6.7% 13.3% 14.3% 14.3% 15.4% 15.4% 8.3% 0.0% 0.0% 9.1% 30.0% 0.0% 0.0% 0.0% 11.1% 25.0% 0.0% 14.3% 0.0% 0.0% 0.0% 8.8%

Lawsuit 3.9% 2.0% 5.1% 5.4% 8.1% 0.0% 0.0% 8.3% 8.3% 13.0% 0.0% 10.0% 0.0% 16.7% 0.0% 5.9% 0.0% 0.0% 20.0% 7.1% 0.0% 0.0% 0.0% 16.7% 0.0% 0.0% 9.1% 10.0% 0.0% 44.4% 11.1% 11.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 5.4%

Any Instrument 37.3% 22.4% 56.4% 45.9% 35.1% 11.4% 24.1% 54.2% 66.7% 39.1% 4.5% 55.0% 0.0% 50.0% 33.3% 47.1% 33.3% 33.3% 53.3% 71.4% 21.4% 38.5% 23.1% 33.3% 16.7% 0.0% 27.3% 70.0% 33.3% 66.7% 22.2% 33.3% 100.0% 42.9% 28.6% 0.0% 40.0% 60.0% 37.0%

Appendix D: Unreported Regressions and Additional Analysis

Table D-1 Characteristics of Target Companies

The table reports the characteristics of hedge funds' targets. Column 1, Column 2, Column 4, and Column 5 report means, medians, and standard deviations. Column 3 and Column 6 report the differences in means and associated t -statistics. Panel A separates activist investments from passive investments. Panel B separates activist investments which seek board seats from activist investments which do not. All variables refer to the trailing twelve months prior to the Schedule 13D and are from Bloomberg or Compustat. Market Cap is market capitalization, Tobin's q is the ratio of market value of equity plus long-term debt over book-value of equity plus long-term debt, Abnormal q is the target firm's Tobin's q as a percent of its industry's median Tobin's q , P/E Ratio is the firm's trailing twelve months price-to-earnings ratio, Growth is the firm's one-year sales growth, Margin is the firm's EBITDA divided by sales, Abnormal Margin is the firm's margin less its industry median margin, Return on Assets is the firm's EBITDA divided by its assets, Return on Equity is the firm's EBITDA divided by the book value of equity, Abnormal ROE is the firm's ROE less its industry median ROE, Leverage is the firm's debt-to-assets ratio, Abnormal Leverage is the firm's leverage less its industry median, Payout Ratio is the firm's common dividends divided by net income, Dividend Yield is the firm's total dividends divided by market capitalizatoin, Cash/Assets is the firm's cash and cash equivalents divded by total assets, Capex/Assets is the firm's capital expenditures divided by total assets, and R&D/Assets is the firm's R&D expense divided by total assets. All data is winsorized at the 10% level. *, **, and *** refer to statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active vs. Passive Targets (1) Active Mean Market Cap Tobin's q Abnormal q P/E Ratio Growth Margin Abnormal Margin Return on Assets Return on Equity Abnormal ROE Leverage Abnormal Leverage Payout Ratio Dividend Yield Cash/Assets Capex/Assets R&D/Assets 757.1 1.6 -16.6 15.8 3.0 10.9 3.0 8.5 20.0 0.7 22.2 15.1 9.1 0.5 16.3 -0.8 2.2 Median 329.9 1.3 -34.0 16.2 2.3 8.7 0.9 8.8 16.8 -0.7 20.5 4.1 0.0 0.0 7.8 -0.5 1.9 StDev 888.6 0.8 74.9 18.1 16.4 9.9 8.0 6.8 18.5 7.3 19.2 34.9 16.1 1.0 17.0 0.8 1.6 Mean 654.4 1.7 12.3 13.3 4.1 12.1 3.2 9.4 22.3 2.4 22.8 11.6 9.8 0.6 14.5 -0.9 1.9 (2) Passive Median 270.0 1.4 -4.2 15.3 3.9 10.3 1.8 9.3 19.1 0.8 19.8 0.0 0.0 0.0 7.8 -0.7 1.5 StDev 814.8 0.9 81.4 18.5 16.4 9.9 8.0 6.9 19.7 8.1 20.6 34.4 16.9 1.1 15.5 0.8 1.7 Mean 102.7 -0.1 -29.0 2.5 -1.2 -1.1 -0.2 -1.0 -2.4 -1.7 -0.6 3.5 -0.7 -0.1 1.9 0.1 0.3 * * *** * * ** * ** (3) Difference t -stat 1.55 1.47 3.68 1.36 0.89 1.34 0.23 1.71 1.49 2.22 0.40 1.00 0.42 0.63 1.49 0.84 1.30

* *

Panel B: Active Investments Seeking Board Seats vs. Active Investments Which Do Not (4) Board Seats Mean Market Cap Tobin's q Abnormal q P/E Ratio Growth Margin Abnormal Margin Return on Assets Return on Equity Abnormal ROE Leverage Abnormal Leverage Payout Ratio Dividend Yield Cash/Assets Capex/Assets R&D/Assets 636.0 1.5 -22.9 14.3 0.5 8.4 0.7 7.4 17.0 0.2 19.7 12.8 10.0 0.5 16.6 -0.8 2.0 Median 253.3 1.2 -42.6 16.2 1.0 6.9 -0.8 6.5 13.9 -1.6 16.4 4.0 0.0 0.0 9.0 -0.5 1.8 StDev 837.4 0.7 72.6 18.8 16.6 8.8 7.2 6.9 17.9 7.2 19.4 33.5 17.2 1.0 17.0 0.8 1.6 Mean 891.4 1.7 -11.1 17.1 5.5 13.5 5.0 9.5 23.0 1.2 24.7 17.1 8.3 0.6 16.1 -0.8 2.3 (5) No Board Seats Median 472.9 1.5 -29.8 16.6 5.8 11.5 3.7 9.3 20.3 0.0 25.1 4.7 0.0 0.0 6.7 -0.5 2.0 StDev 926.8 0.8 76.7 17.6 15.9 10.4 8.1 6.5 18.6 7.4 18.8 36.3 15.2 1.0 17.1 0.8 1.6 Mean -255.4 -0.2 -11.7 -2.9 -5.0 -5.1 -4.3 -2.1 -6.0 -1.0 -5.0 -4.3 1.7 0.0 0.5 0.0 -0.3 *** *** (6) Difference t -stat 2.44 2.47 1.10 1.07 2.54 4.13 3.54 2.49 2.59 0.93 2.21 0.85 0.66 0.10 0.25 0.35 0.95

*** *** *** *** *** **

Table D-2 Short-Term Returns by Investment Type

The table presents average daily excess returns around the Schedule 13D filing for various event windows. I measure returns in excess of the market model using the value-weighted CRSP NYSE/Amex/Nasdaq index, estimating parameters over the 200 day window prior to the Schedule 13D. Panel A compares active investments to passive investments. Panel B further splits active investments into those which seek board seats and those which do not. I report p -values for medians using the Wilcoxon Signed Rank Test. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active Investments Perform Better Than Passive Investments Active (-30,+30) 5% 25% Median 75% 95% Average % Positive Patel z-statistic Non par. z-statistic N= -21.5% 1.6% 11.8% *** 25.4% 53.5% 12.9% *** 80.4% *** 10.77 8.91 194 (-2,+2) -7.5% 0.8% 4.4% *** 10.8% 25.9% 5.7% *** 79.9% *** 16.50 8.76 194 (+0,+2) -3.7% 0.1% 3.0% *** 9.7% 21.8% 4.6% *** 75.8% *** 17.16 7.61 194 (+0,+30) -12.2% -2.6% 5.6% *** 16.7% 38.8% 7.5% *** 69.1% *** 8.89 5.75 194 (-30,+30) -33.3% -10.7% 3.6% *** 19.2% 44.9% 4.6% *** 57.6% *** 4.04 3.37 342 (-2,+2) -7.0% -0.9% 1.9% *** 6.0% 16.7% 2.7% *** 67.4% *** 9.27 6.95 340 Passive (+0,+2) -5.2% -0.7% 1.2% *** 4.4% 10.9% 1.8% *** 64.1% *** 7.82 5.76 340 (+0,+30) -24.2% -5.1% 2.0% *** 11.5% 30.3% 2.7% *** 57.4% *** 3.00 3.26 340 (-30,+30) (-2,+2) Difference (+0,+2) (+0,+30)

8.2% ***

2.5% ***

1.9% ***

3.6% ***

8.3% ***

3.0% ***

2.8% ***

4.8% ***

Panel B: Active Investments Seeking Board Seats Perform Better Than Active Investments That Do Not Board Seats (-30,+30) 5% 25% Median 75% 95% Average % Positive Patel z-statistic Non par. z-statistic N= -13.8% 6.1% 12.5% *** 29.8% 59.6% 17.7% *** 88.9% *** 5.90 5.48 45 (-2,+2) -7.8% 0.7% 7.8% *** 16.0% 29.2% 8.6% *** 77.8% *** 9.41 3.98 45 (+0,+2) -5.0% -0.8% 5.0% *** 14.1% 31.3% 7.6% *** 71.1% *** 10.16 3.09 45 (+0,+30) -8.7% -1.3% 9.3% *** 23.6% 40.6% 11.7% *** 73.3% *** 4.98 3.39 45 (-30,+30) -23.3% 0.9% 10.8% *** 24.4% 51.2% 12.6% *** 77.9% *** 9.04 7.15 149 No Board Seats (-2,+2) -6.1% 0.9% 4.1% *** 8.6% 21.6% 5.7% *** 80.5% *** 13.66 7.81 149 (+0,+2) -3.5% 0.1% 2.9% *** 7.6% 19.4% 4.6% *** 77.2% *** 14.00 6.99 149 (+0,+30) -13.6% -3.6% 5.0% *** 15.3% 35.1% 7.5% *** 67.8% *** 7.40 4.70 149 (-30,+30) (-2,+2) Difference (+0,+2) (+0,+30)

1.7% *

3.7% *

2.1%

4.3%

5.1% *

2.9% *

3.1% **

4.2% *

Table D-3 Short-Term Returns By Type, Controlling For Undervaluatoin

The dependent variable is the abnormal return during the 5 days around the Schedule 13D filing date. Panel A controls for abnormal q and abnormal margin and Panel B controls for Tobin's q and margin. Column 1 and Column 3 regress all abnormal returns on Active , a dummy set to one for activist investments. Column 2 and Column 4 regress abnormal returns from exclusively active investments on Board , a dummy set to one for board investments. All dummies refer only to information contained in the initial schedule 13D filing. All nondummy variables are winsorized at the 10% level and are expressed as the deviation from the sample average values. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Industry-Adjusted Variables (1) Dependent Variable: Abnormal Returns Constant Abnormal q Abnormal Margin Active Board 2 No. and R Coeff. 3.07 -1.14 -0.09 3.15 290 Activism t -stat *** 5.09 ** -1.97 * -1.56 *** 3.37 0.07 (2) Board Seats Coeff. t -stat 5.65 *** 6.69 -2.07 ** -2.15 -0.23 ** -2.23 2.45 1.24 122 0.11

Panel B: Non-Adjusted Variables (3) Dependent Variable: Abnormal Returns Constant Tobin's q Margin Active Board 2 No. and R Activism Coeff. t -stat 2.45 *** 5.13 -0.93 ** -1.96 -0.02 -0.58 3.75 *** 4.78 441 0.06 (4) Board Seats Coeff. t -stat 5.55 *** 6.45 -2.38 *** -2.46 -0.03 -0.41 3.15 ** 1.71 162 0.07

Table D-4 Abnormal Returns After 13D Filing Date

The table presents average returns after the initial Schedule 13D filing, in excess of the value-weighted market, for various event windows. Column 1 analyzes schedule 13D/A filings in which, having already filed a passive 13D, a hedge fund first announces activist intentions. Column 2 analyzes schedule 13D/A filings or news stories in which, having already announced activist intentions, a hedge fund first seeks board seats. Column 3 analyzes schedule 13D/A filings or news stories in which, having already announced its intention to seek board seats, the hedge fund first wins board seats. I measure returns in excess of market model using the value-weighted CRSP NYSE/Amex/Nasdaq index, estimating parameters over the 200 day window prior to the event. I report p -values for medians using the Wilcoxon Signed Rank Test. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

(1) First Action (-30,+30) 5% 25% Median 75% 95% Average % Positive Patel z-statistic Non par. z-statistic N= -30.2% -2.3% 7.5% *** 20.0% 46.3% 8.3% *** 73.6% *** 3.38 3.76 53 (-2,+2) -9.5% -2.2% 1.2% *** 10.7% 21.7% 4.4% *** 66.0% *** 5.83 2.66 53 (+0,+2) -3.1% -0.3% 2.8% *** 9.3% 20.5% 5.3% *** 71.7% *** 8.81 3.49 53 (+0,+30) -12.5% -1.9% 5.6% *** 17.0% 26.9% 7.1% *** 69.8% *** 4.34 3.21 53 (-30,+30) -23.4% -4.9% 6.3% *** 18.4% 41.3% 6.9% *** 60.4% ** 2.69 1.70 48

(2) First Board Seat (-2,+2) -3.3% 0.2% 2.5% *** 6.7% 13.6% 3.4% *** 77.1% *** 4.95 4.01 48 (+0,+2) -2.8% 0.5% 2.6% *** 4.6% 8.3% 2.5% *** 77.1% *** 4.71 4.01 48 (+0,+30) -13.7% -3.5% 2.5% *** 11.1% 34.1% 5.3% *** 66.7% *** 2.75 2.56 48 (-30,+30) -29.1% -6.2% 5.2% *** 20.7% 35.7% 6.2% *** 66.4% *** 3.80 3.82 110

(3) First Board Win (-2,+2) -7.9% -2.2% 0.3% 4.1% 10.2% 0.8% * 53.2% 1.63 1.05 109 (+0,+2) -5.4% -1.0% 0.7% *** 3.3% 6.6% 0.8% ** 61.5% *** 2.21 2.78 109 (+0,+30) -18.2% -5.6% 0.6% * 10.7% 25.8% 2.5% *** 51.8% 2.41 0.76 110

Table D-5 Long-Term Returns by Investment Type

The table presents average monthly excess returns around the Schedule 13D filing for various event windows. I measure returns in excess of the market model using the value-weighted CRSP NYSE/Amex/Nasdaq index, estimating parameters over the 36 month window prior to the Schedule 13D. Panel A compares active investments to passive investments. Panel B further splits active investments into those which seek board seats and those which do not. I report p -values for medians using the Wilcoxon Signed Rank Test. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Active Investments Perform Better Than Paassive Investments Active (-1,+24) 5% 25% Median 75% 95% Average % Positive Patel z-statistic Non par. z-statistic N= -132.5% -8.4% 20.1% *** 57.3% 172.3% 21.3% *** 66.8% *** 8.43 5.70 244 (+1,+6) -48.2% -12.6% 4.2% *** 19.9% 59.6% 4.9% *** 56.3% *** 2.88 2.39 238 (+1,+12) (+1,+24) (-1,+24) -153.3% -31.3% 10.4% * 49.2% 125.2% 3.7% 55.1% ** 1.08 2.08 227 (+1,+6) -58.8% -20.1% 0.1% 20.3% 54.2% -0.6% 50.0% -0.57 0.54 216 Passive (+1,+12) -89.4% -24.5% -0.1% 30.5% 75.2% -0.4% 49.8% -0.29 0.48 217 (+1,+24) -132.5% -30.3% 5.3% 45.9% 114.8% 1.5% 53.7% *** 0.16 16.29 218 (-1,+24) Difference (+1,+6) (+1,+12) (+1,+24)

-77.7% -127.6% -19.2% -18.0% 7.4% *** 12.1% *** 38.2% 47.5% 111.3% 159.4% 8.5% *** 56.9% *** 3.87 2.58 239 13.5% *** 61.1% *** 5.55 3.87 239

9.7%

4.1%

7.5%

6.8% *

17.6% **

5.4% **

8.9% **

11.9% *

Panel B: Active Investments Seeking Board Seats Perform Better Than Active Investments That Do Not Board Seats (-1,+24) 5% 25% Median 75% 95% Average % Positive Patel z-statistic Non par. z-statistic N= -91.9% 0.2% 28.2% *** 60.7% 150.7% 31.2% *** 75.2% *** 7.48 6.03 121 (+1,+6) -37.0% -10.8% 6.1% *** 22.4% 58.3% 8.3% *** 58.5% ** 2.67 2.32 118 (+1,+12) -53.5% -11.6% 11.3% *** 38.6% 95.7% 15.5% *** 62.2% *** 4.19 3.14 119 (+1,+24) -75.9% -11.6% 18.3% *** 53.1% 149.2% 21.6% *** 66.4% *** 5.15 4.06 119 (-1,+24) -154.0% -8.1% 13.0% *** 56.5% 177.8% 14.3% *** 60.9% *** 4.97 2.45 115 No Board Seats (+1,+6) -55.8% -13.0% 2.0% 19.1% 64.6% 2.8% ** 55.4% * 1.89 1.30 112 (+1,+12) -106.8% -24.1% 2.2% * 36.4% 116.2% 3.8% ** 54.5% 1.84 1.11 112 (+1,+24) -144.7% -20.7% 7.6% 40.6% 175.3% 7.8% *** 58.0% ** 3.19 1.87 112 (-1,+24) Difference (+1,+6) (+1,+12) (+1,+24)

15.2% **

4.1%

9.2% *

10.8% **

16.9% *

5.5%

11.7% *

13.8% *

Table D-6 Long Term Returns By Type, Controlling For Undervaluatoin

The dependent variable is the abnormal return during the 25 months around the Schedule 13D filing date. Panel A controls for abnormal q and abnormal margin and Panel B controls for Tobin's q and margin. Column 1 and Column 3 regress all abnormal returns on Active , a dummy set to one for activist investments. Column 2 and Column 4 regress abnormal returns from exclusively active investments on Board , a dummy set to one for board investments. All nondummy variables are winsorized at the 10% level and are expressed as the deviation from the sample average values. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Industry-Adjusted Variables (1) Dependent Variable: Abnormal Returns Constant Abnormal q Abnormal Margin Active Board 2 No. and R Activism Coeff. 6.67 -9.21 * -1.01 ** 4.32 251 t -stat 0.90 -1.59 -1.80 0.45 0.03 Coeff. 7.81 -14.40 ** -0.01 2.45 150 (2) Board Seats t -stat 1.00 -1.87 -0.01 1.24 0.03

Panel B: Non-Adjusted Variables (3) Dependent Variable: Abnormal Returns Constant Tobin's q Margin Active Board 2 No. and R Activism Coeff. 4.60 -7.07 * -0.90 ** 18.35 ** 388 t -stat 0.75 -1.36 -2.11 2.17 0.04 Coeff. 15.33 ** -11.50 * -0.79 18.63 * 197 (4) Board Seats t -stat 1.86 -1.52 -1.25 1.59 0.04

Table D-7 Profitability Improvements, Controlling For Stock Picking

The dependent variable is the two-year increase in ROA (Panel A) and in EBITDA margins (Panel B). All regressions control for Tobin's q and for the initial ROA (Panel A) or EBITDA margins (Panel B). Column 1 and Column 3 regress improvements for all firms on Active , a dummy equal to one for activist investments. Column 2 and Column 4 regress improvements for activist targets on Board , a dummy equal to one for board seat investments. All nondummy variables are winsorized at the 10% level and are expressed as the deviation from the sample average values. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Return On Assets (1) Dependent Variable: Abnormal Returns Constant Tobin's q Initial Level Active Board 2 No. and R Activism Coeff. t -stat -1.15 ** -1.70 1.61 2.51 -0.31 *** -4.16 3.03 *** 3.05 108 0.26 Panel B: Margins (3) Dependent Variable: Abnormal Returns Constant Tobin's q Initial Level Active Board 2 No. and R Activism Coeff. t -stat -2.53 *** -3.71 0.16 0.24 -0.12 ** -2.32 2.87 *** 2.76 105 0.13 (4) Board Seats Coeff. t -stat -0.06 -0.05 0.23 0.24 -0.20 *** -2.34 0.23 0.16 38 0.16 Coeff. 1.20 1.73 -0.24 ** 0.82 40 (2) Board Seats t -stat 0.89 1.55 -1.76 0.46 0.10

Table D-8 Capital Structure Changes, Controlling For Stock Picking

The dependent variable is the two-year increase in payout ratio (Panel A) and in the debt-to-assets ratio (Panel B). All regressions control for Tobin's q, the initial payout ratio (Panel A) or leverage (Panel B), and the initial cash divded by initial assets. Column 1 and Column 3 regress improvements for all firms on Active , a dummy equal to one for activist investments. Column 2 and Column 4 regress improvements for activist targets on Board , a dummy equal to one for board seat investments. All nondummy variables are winsorized at the 10% level and are expressed as the deviation from the sample average values. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Payout (1) Dependent Variable: Abnormal Returns Constant Tobin's q Initial Level Cash Active Board 2 No. and R Activism Coeff. 2.25 ** -1.36 0.12 ** 10.45 * 2.48 102 t -stat 1.79 -0.99 2.02 1.50 1.26 0.07 Panel B: Leverage (3) Dependent Variable: Abnormal Returns Constant Tobin's q Initial Level Cash Active Board 2 No. and R Activism Coeff. t -stat -0.37 -0.83 0.05 0.11 -0.05 *** -2.68 -3.54 * -1.53 0.65 0.98 222 0.04 Coeff. 1.43 * -0.08 -0.05 ** -4.07 -2.08 ** 92 (4) Board Seats t -stat 1.59 -0.11 -1.78 -1.20 -1.86 0.07 Coeff. 4.24 * -0.56 0.16 * 10.83 0.75 42 (2) Board Seats t -stat 1.30 -0.21 1.39 0.70 0.19 0.06

Table D-9 Firm Size Reductions, Controlling For Undervaluatoin

The dependent variable is the two-year increase in assets (Panel A) and in sales (Panel B). All regressions control for Tobin's q and for initial assets (Panel A) or initial sales (Panel B). Column 1 and Column 3 regress improvements for all firms on Active , a dummy equal to one for activist investments. Column 2 and Column 4 regress improvements for activist targets on Board , a dummy equal to one for board seat investments. All nondummy variables are winsorized at the 10% level and are expressed as the deviation from the sample average values. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Panel A: Assets (1) Dependent Variable: Abnormal Returns Constant Tobin's q Initial Level Active Board 2 No. and R Activism Coeff. t -stat 2.77 ** 1.81 8.75 *** 6.66 0.00 -1.03 -4.56 ** -2.12 522 0.09 Panel B: Sales (3) Dependent Variable: Abnormal Returns Constant Tobin's q Initial Level Active Board 2 No. and R Activism Coeff. t -stat 4.95 *** 3.67 5.61 *** 4.85 0.00 0.23 -1.19 -0.63 494 0.05 (4) Board Seats Coeff. t -stat 4.15 ** 2.03 6.29 *** 3.61 0.00 0.79 -0.82 -0.30 244 0.06 (2) Board Seats Coeff. t -stat 0.75 0.34 5.90 *** 3.15 0.00 ** -1.89 -4.42 * -1.49 256 0.06

Table D-10 Profitability Improvement By Strategy

The table regresses the two-year change in (1) Return on Assets and (2) EBITDA margin on dummy variables for the various activist strategies, considering only the sample of investments which announce activist strategies. The regressions control for the initial ROA or the initial margin. The data is winsorized at the 10% level. Due to the full span of the dummy variables, the constants are supressed. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Return on Assets Coeff Initial Value Business Strategy M&A Advice Sale Capital Structure Corporate Governance Other No. and R 2 -0.4 *** 6.9 *** -0.7 -1.4 1.4 -0.3 0.5 29 t -stat -2.70 2.71 -0.42 -0.86 0.55 -0.12 0.22 0.50

EBITDA Margin Coeff -0.2 ** 2.3 -0.6 -3.4 ** 5.0 ** -0.7 2.0 26 t -stat -2.15 0.91 -0.39 -2.11 1.81 -0.32 0.87 0.46

Table D-11 Capital Structure Changes By Strategy

The table regresses the two-year change in (1) payout ratio and (2) leverage on dummy variables for the various activist strategies, considering only the sample of investments which announce activist strategies. The regressions control for the initial payout ratio or the leverage as well as the initial cash divided by assets. The data is winsorized at the 10% level. Due to the full span of the dummy variables, the constants are supressed. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Payout Ratio Coeff Initial Value Cash Business Strategy M&A Advice Sale Capital Structure Corporate Governance Other No. and R 2 0.0 21.8 * -10.0 ** 2.6 -2.8 10.4 ** 3.4 0.8 29 t -stat 0.11 1.35 -1.78 0.81 -0.81 2.21 0.62 0.13 0.29 Coeff

Leverage t -stat -1.58 -1.47 -0.73 1.11 0.90 -1.17 1.11 1.11 1.22

-0.1 * -6.1 * -1.2 1.1 1.0 -1.8 1.3 2.6 64

Table D-12 Firm Size Reductions By Strategy

The table regresses the two-year percentage change in (1) assets and (2) sales on dummy variables for the various activist strategies, considering only the sample of investments which announce activist strategies. The regressions control for the initial assets or sales. The data is winsorized at the 10% level. Due to the full span of the dummy variables, the constants are supressed. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

Assets Coeff Initial Value Business Strategy M&A Advice Sale Capital Structure Corporate Governance Other No. and R 2 0.0 ** 3.2 0.7 -4.5 * 2.6 -4.0 14.6 ** 185 t -stat -2.26 0.77 0.23 -1.43 0.68 -1.10 1.65 0.05 Coeff 0.0 4.6 3.6 -2.9 3.0 2.3 12.0 * 178

Sales t -stat -0.06 1.14 1.26 -0.93 0.83 0.65 1.32 0.06

Table D-13 Sale Efficiency By Activist Strategy

Column 1 shows a logit model that predicts whether a target firm will sell itself, and Column 2 shows a regression that predicts the average premium those firms receive. In Column 1, the dependent variable is a dummy set to 1 if the firm sold itself . In Column 2 the dependent variable is the average premium for each such sale. I measure premium as the total return to shareholders from one month before the activist first files a Schedule 13D until the sale closes. Both regressions include data only from investments which use activist strategies. Column 1 presents R 2 , and Column 2 presents Pseudo-R 2 . *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels.

(1) Odds of Sale Odds z -stat Bus Strategy M&A Sale Cap Struc Corp Gov Other No. and R 2 0.23 ** 2.27 ** 0.62 0.95 0.47 0.48 164 -2.23 1.74 -0.98 -0.11 -1.23 -0.62 0.10

(2) Avg Premium Coeff t -stat 7.3 13.7 24.6 15.8 40.7 74.8 32 0.33 1.65 2.04 1.14 2.26 2.21 0.57

** ** ** **

References Admati, Anat and Pfleiderer, Paul. 2009. The Wall Street Walk and Shareholder Activism: Exit as a Form of Voice. Stanford Graduate School of Business Research Paper No. 1918(R2). Ang, James and Yingmei Cheng. 2006. Direct Evidence on Market-Driven Acquisitions Theory. Journal of Financial Research 29: 199-216. Aragon, George. 2007. Share Restrictions and Asset Pricing: Evidence From the Hedge Fund Industry. Journal of Financial Economics 83: 3358. Bainbridge, Stephen. 2006. Directory Primacy and Shareholder Disempowerment. UCLA School of Law, LawEcon Research Paper No. 0525. Barber, Brad. 2007. Monitoring the Monitor: Evaluating CalPERS Shareholder Activism. Journal of Investing 59: 6680. Berle, Adolf and Gardiner Means. 1932. The Modern Corporation and Private Property. New York: Macmillan. Becht, Marco, Julian Franks, Colin Mayer, and Stefano Rossi. 2008. Returns to Shareholder Activism: Evidence From a Clinical Study of the Hermes UK Focus Fund. Working Paper. Available at SSRN: http://ssrn.com/abstract=934712. Bratton, William. 2006. Hedge Funds and Governance Targets. Working Paper. Available at SSRN: http://ssrn.com/abstract=928689. Bebchuck, Lucian. 2005. The Case for Increasing Shareholder Power. Harvard Law Review 118: 835914. Bebchuck, Lucian. 2007. The Myth of the Shareholder Franchise. Virginia Law Review 93: 675 732. Bebchuk, Lucian and Jesse Fried. 2003. Executive Compensation as an Agency Problem. Journal of Economic Perspectives 17: 1792. Bethel, Jennifer, Julia Liebeskind, and Tim Opler. 1998. Block share Purchases and Corporate Performance. The Journal of Finance 53: 60534. Bhide, Amar. 1993. The Hidden Costs of Stock Market Liquidity. Journal of Financial Economics 34: 3151. Black, Bernard. 1990. Shareholder Passivity Reexamined. Michigan Law Review 89: 520608.

Black, Bernard. 1992. Agents Watching Agents: The Promise of Institutional Investor Voice. UCLA Law Review 39: 811893. Black, Bernard. 1998. Shareholder Activism and Corporate Governance in the United States. The ew Palgrave Dictionary of Economics and the Law 3: 459465. Boyson, Nicole and Robert Mooradian. 2008. Hedge Funds as Shareholder Activists from 1994 2005. Working Paper. Available at SSRN: http://ssrn.com/abstract=992739. Bradley, Michael, Alon Brav, Itay Goldstein, and Wei Jiang. 2007. Shareholder Activism and Price Dynamics: Evidence From ClosedEnd Funds. Working Paper. Available: http://finance.wharton.upenn.edu/~itayg/Files/Research%20Statement%20 %20Website.pdf. Brav, Alon, Wei Jiang, Frank Partnoy, and Randall Thomas. 2008. Hedge Fund Activism, Corporate Governance, and Firm Performance. The Journal of Finance 63: 172975. Carleton, Willard, James Nelson, and Michael Weisbach. 1998. The Influence of Institutions of Corporate Governance through Private Negotiations: Evidence from TIAACREF. The Journal of Finance 53: 133562. Christoffersen, Susan, Chris Gczy, David Musto, and Adam Reed. 2006. Vote Trading and Information Aggregation. ECGI Finance Working Paper No. 141/2007. Clifford, Christopher. Value Creation or Destruction? Hedge Funds as Shareholder Activists. Journal of Corporate Finance 14: 32336. Coffee, John. 193. Liquidity Versus Control: The Institutional Investor As Corporate Monitor. Columbia Law Review 91: 12771368. Dasgupta, Susmita, Benoit Laplante, and Nlandu Mamingi. 1997. Capital Market Responses to Environment Performance in Developing Countries. The World Bank Development Research Group Working Paper No. 1909. Davis, Gerald and Han Kim. 2007. Business Ties and Proxy Voting by Mutual Funds. Journal of Financial Economics 85: 55270. Del Guercio, Diane and Jennifer Hawkins. 1999. The Motivation and Impact of Pension Fund Activism. Journal of Financial Economics 52: 293340. Del Guercio, Diane, Laura Wallis, and Tracie Woidtke. 2008. Do Board Members Pay Attention When Institutional Investors Just Vote No? CEO and Director Turnover Associated with Shareholder Activism. Working Paper. Available at SSRN: http://ssrn.com/abstract=575242.

Deli, Daniel. 2002. Mutual Fund Advisory Contracts: An Empirical Investigation. Journal of Finance 58: 10933. Fama, Eugene. 1970. Efficient Capital Markets: A Review of Theory and Empirical Work. The Journal of Finance 25: 383417. Fama, Eugene and Kenneth French. 1992. The Cross-Section of Expected Stock Returns. Journal of Financial Studies 47: 427-65. Fama, Eugene and Michael Jensen. 1983. Separation of Ownership and Control. Journal of Law and Economics 26: 30125. Fayele, Olubunmi. 2004. Cash and Corporate Control. Journal of Finance 59: 204160. Gabaix, Xavier and Augustin Landier. 2008. Why Has CEO Pay Increased So Much? Quarterly Journal of Economics 123: 49100. Gillan, Stuart and Laura Starks. 1998. A Survey of Shareholder Activism: Motivation and Empirical Evidence. Working Paper. Available at SSRN: http://ssrn.com/abstract=663523. Gillan, Stuart and Laura Starks. 2000. Corporate Governance Proposals and Shareholder Activism: The Role of Institutional Investors. Journal of Financial Economics 57: 275 305. Gillan, Stuart and Laura Starks. 2007. The Evolution of Shareholder Activism in the United States. Journal of Applied Corporate Finance 19: 5573. Golec, Joseph. 1992. Empirical Tests of a Principal Agent Model of the InvestorInvestment Advisor Relationship. Journal of Financial and Quantitative Analysis 27: 8195. Gordon, Lilli and John Pound. 1993. Information, Ownership Structure, and Shareholder Voting: Evidence From ShareholderSponsored Corporate Governance Proposals. Journal of Finance 47: 697718. Graham, Benjamin. 1973 [2003]. The Intelligent Investor, Revised Edition. New York: Harper Collins. Grossman, Sanford and Oliver Hart. 1980. Takeover Bids, the Free Rider Problem, and the Theory of the Corporation. Bell Journal of Economics 11: 4264. Greenwood, Robin and Michael Schor. 2008. Hedge Fund Investor Activism and Takeovers. Harvard Business School Working Paper. Hall, Brian and Kevin Murphy. The Trouble with Stock Options. Journal of Economic Perspectives 17: 4970.

Hu, Henry and Bernard Black. 2007. Hedge Funds, Insiders, and the Decoupling of Economic and Voting Ownership: Empty and Hidden (Morphable) Ownership. Journal of Corporate Finance 13: 34367. Jensen, Michael and William Meckling. 1976. Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics 3: 30560. Jensen, Michael. 1986. The Agency Cost of Free Cash Flow: Corporate Finance and Takeovers. American Economic Review 76: 32329. Johnson, Marilyn and Margaret Shackell. 1997. Shareholders Proposals on Executive Compensation. Working Paper. Available at SSRN: http://ssrn.com/sbstract=121. Kahan, Marcel and Edward Rock. 2006. Hedge Funds in Corporate Governance and Corporate Control. University of Pennsylvania, Institute for Law & Economics Research Paper No. 0616. Karpoff, Jonathan. 2001. The Impact of Shareholder Activism on Target Companies: A Survey of Empirical Findings. Working Paper. Available at SSRN: http://ssrn.com/abstract=885365. Karpoff, Jonathan, Paul Malatesta, and Ralph Walkling. 1996. Corporate Governance and Shareholder Initiatives: Empirical Evidence. Journal of Financial Economics 42: 36595. Koller, Tim, Marc Goedhart, and David Wessels. 2005. Valuation: Measuring and Managing the Value of Companies. Ne York: Wiley, John & Sons. Klein, April and Emauel Zur. 2009. Entrepreneurial Shareholder Activism: Hedge Funds and Other Private Investors. The Journal of Finance 64: 187229. Lipton, Martin and Steven Rosenblum. 1991. A New System of Corporate Governance: The Quinquennial Election of Directors. University of Chicago Law Review 58: 187253. Lynn, Matthew. 2007. Wall street bosses spending too much time on the golf course. Wall Street Journal. 17 November. MacKinlay, Craig. 1997. Event Studies in Economics and Finance. The Journal of Economic Literature 35: 1339. McConnell, John and Henri Servaes. 1990. Additional Evidence on Equity Ownership and Corporate Value. Journal of Financial Economics 27: 595612. Parrino, Robert, Richard Sias, and Laura Starks. 2003. Voting With Their Feet: Institutional Ownership Changes Around Forced CEO Turnover. Journal of Financial Economics 68: 346.

Pound, John. 1988. Proxy Contests and the Efficiency of Shareholder Oversight. Journal of Financial Economics 20: 237-65. Pound, John. 1992. Raiders, Targets, and Politics: The History and Future of American Corporate Control. Journal of Applied Corporate Finance 5: 618. Puustinen, Pekka. Determinants of Hedge Fund Activism and Shareholder Gains. Unpublished MA, Helsinki School of Economics, Helsinki. Rock, Edward. 1992. The Logic and (Uncertain) Significance of Institutional Shareholder Activism. Georgetown Law Review 79: 445406. Romano, Roberta. 1993. Public Pension Fund Activism in Corporate Governance Reconsidered. Columbia Law Review 79: 445506. Romano, Roberta. 2001. Less Is More: Making Shareholder Activism a Valued Mechanism of Corporate Governance. Yale Journal on Regulation 18: 174251. Smith, Adam. 1776 [1977]. An Inquiry Into the Chicago: University of Chicago Press. ature and Causes of the Wealth of ations.

Smith, Michael. 1996. Shareholder Activism by Institutional Investors: Evidence from CalPERS. Journal of Finance 51: 22752. Sweren, Joshua. 2008. The Wandering Eye: Do Investors Steal Ideas from 13Fs? Unpublished BA, Yale University, New Haven. Strickland, Deon, Kenneth Wiles, and March Zenner. 1996. A Requiem for the USA: Is Small Shareholder Monitoring Effective? Journal of Financial Economics 40: 31938. Wahal, Sunil. 1996. Pension Fund Activism and Firm Performance. Journal of Financial and Quantitative Analysis 31: 123. Yermack, David. 1995. Do Corporations Award CEO Stock Options Effectively? Journal of Financial Economics 39: 23769.