NBER WORKING PAPER SERIES THE EFFECT OF LIQUIDITY ON GOVERNANCE Alex Edmans Vivian W. Fang Emanuel Zur Working Paper 17567 http://www.nber.org/papers/w17567 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 November 2011 We thank Charles Trzcinka for data and Niki Boyson, Chris Clifford, Slava Fos, Robin Greenwood, Cliff Holderness, Oguzhan Karakas, Robert Kieschnick, Alexander Ljungqvist, Ernst Maug, Greg Nini, Ed Rock, and seminar participants at Boston College, Oregon, UT Dallas, and Wharton for helpful comments. Vivian also gratefully acknowledges the financial support from Rutgers Business School Research Resources Committee (RRC). The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peer- reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications. © 2011 by Alex Edmans, Vivian W. Fang, and Emanuel Zur. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. The Effect of Liquidity on Governance Alex Edmans, Vivian W. Fang, and Emanuel Zur NBER Working Paper No. 17567 November 2011 JEL No. G12,G23,G34,G38 ABSTRACT This paper studies the effect of stock liquidity on blockholders’ choice of governance mechanisms. We focus on hedge funds as they are unconstrained by legal restrictions and business ties, and thus have all governance channels at their disposal. Since the threat of governance, not just actual governance, can discipline managers, we use Section 13 filings to measure governance intent rather than only studying instances of actual governance. We find that liquidity increases the likelihood that a hedge fund acquires a block in a firm. Conditional upon acquiring a stake, liquidity reduces the likelihood that a blockholder governs through voice (intervention) – as evidenced by the greater propensity to file Schedule 13Gs (passive investment) rather than 13Ds (active investment). Liquidity is more likely to lead to a 13G filing if the manager’s wealth is sensitive to the stock price, consistent with governance through exit (trading). A 13G filing leads to positive announcement returns, especially in liquid firms. These two results suggest that liquidity does not dissuade blockholders from governing altogether, but instead encourages them to govern through exit rather than voice. We use decimalization as an exogenous shock to liquidity to identify causal effects. Alex Edmans Emanuel Zur The Wharton School Zicklin School of Business, Baruch College University of Pennsylvania City University of New York 2318 Steinberg Hall - Dietrich Hall New York, NY 10010 3620 Locust Walk [email protected] Philadelphia, PA 19104 and NBER [email protected] Vivian W. Fang Rutgers Business School Rutgers University Newark, NJ 07102 [email protected] This paper studies the effect of stock liquidity on an investor’s decision to acquire a block in a firm, and her choice of how to govern the firm thereafter. The theoretical literature yields conflicting predictions regarding the effect of liquidity on governance. The traditional view is that investors govern through intervening in a firm’s operations (also known as “voice”), for example by firing a shirking manager or blocking a pet project. Under this view, liquidity weakens governance because it provides the blockholder with the option of selling her stake in a troubled firm rather than bearing the cost of intervening to fix it (Coffee (1991); Bhide (1993)), or trading on inside information rather than monitoring (Maug (2002)). However, this view has been recently challenged along two fronts. First, even when considering voice as the only governance mechanism, Maug (1998) shows that liquidity encourages blockholders to intervene as they can buy additional shares at a price that does not incorporate the gains from intervention. Kyle and Vila (1991), Maug (1998), and Kahn and Winton (1998) demonstrate that liquidity facilitates block formation in the first place. Faure-Grimaud and Gromb (2004) show that liquidity encourages intervention as it increases stock price informativeness. Thus, if the activist is forced to sell prematurely due to a liquidity shock, the price she receives will partially reflect the gains from intervention. Second, Admati and Pfleiderer (2009), Edmans (2009), and Edmans and Manso (2011) demonstrate that the act of selling one’s shares (also known as engaging in “exit”, following the “Wall Street Rule”, or taking the “Wall Street Walk”) can be a governance mechanism in itself.1 By gathering and trading on private information, blockholders cause the stock price to more closely reflect the firm’s fundamental value. If the manager is compensated according to the stock price, the threat of exit induces him to maximize fundamental value – for example, by exerting effort and investing efficiently. Liquidity increases the threat of exit in two 1 While it is the threat of exit that induces superior managerial actions ex ante, we follow the literature by referring to this governance mechanism as “exit” for brevity. 1 ways: it induces blockholders to gather information (Edmans (2009); Edmans and Manso (2011)), and it encourages investors to acquire a larger block to begin with (Edmans (2009)). Thus, even if liquidity discourages intervention, it may not reduce governance overall but instead cause the blockholder to govern through exit rather than voice. Despite the rich theoretical literature analyzing the effect of stock liquidity on governance choices, there are very few papers that address this debate empirically. This likely results from a number of empirical challenges. First, many blockholders do not have the governance mechanism of voice at their disposal to begin with. Mutual funds have diversification requirements that prevent them from acquiring the large positions necessary to exercise control,2 and pension funds are subject to “prudent man” rules that hinder them from acquiring stakes in troubled firms that may be particularly in need of intervention (Del Guercio (1996)). Even if not legally restricted, certain blockholders may choose not to engage in activism due to conflicts of interest: suppliers and customers may risk losing supply-chain relationships, and mutual funds may have business ties through managing a firm’s pension plan.3 Del Guercio and Hawkins (1999) find that pension fund activism has little effect on stock or accounting performance; more generally, the survey by Yermack (2010) concludes that “the success of institutional investor activism to date appears limited.” Thus, liquidity may not affect the choice between exit and voice, since many blockholders do not engage in voice to begin with. Moreover, even if they do use voice, certain blockholders may have little motivation to make the optimal governance 2 Most mutual funds prefer to designate themselves as “diversified”. Under the Investment Company Act of 1940, such a fund, with respect to 75% of its portfolio, can have no more than 5% of it invested in any one security and can own no more than 10% of the voting rights in one company. 3 Davis and Kim (2007) show that mutual funds with more business ties in aggregate are more likely to vote with management in general, although at the individual firm level they are no more likely to vote with management of client firms than of non-clients. 2 choice in response to liquidity (e.g. due to weak financial incentives), or pursue objectives other than shareholder value maximization (Agrawal (2011)). Second, while many existing papers study actual exit (e.g., Parrino, Sias, and Starks (2003)) or actual voice (e.g., Norli, Ostergaard and Schindele (2009)), the threat of exit or threat of voice also exerts governance. The absence of instances of exit or voice does not mean the blockholder is failing to govern – on the contrary, it may suggest that the blockholder is governing effectively, inducing the manager to maximize firm value and so actual exit or voice is not needed (cf. Becht et al. (2009), Klein and Zur (2009), and Fos (2011)). Third, liquidity and governance may be jointly determined by a firm’s unobservable characteristics, or the causality may run from governance to liquidity.4 This paper aims to study the effect of liquidity on governance while addressing the three above challenges. We address the first challenge by focusing on activist hedge funds. Hedge funds have few business ties or regulatory constraints that would hinder voice: they have no diversification requirement, and they have few disclosure needs, allowing them to act with greater secrecy and flexibility (Yermack (2010)). McCahery, Sautner, and Starks (2011) find that hedge funds are more willing to engage in activism than other institutions. Hedge funds thus have the full “menu” of governance mechanisms to choose from, and liquidity may drive their selection from this menu. Indeed, over half (69 out of 101) of the funds in our sample engage in both passive and active monitoring. They also have high performance-based fees which induce them to make the optimal choice from this menu. Clifford and Lindsey (2011) show that 4 Chung, Elder, and Kim (2010) show that superior governance improves liquidity, potentially due to improved transparency and thus reduced informational asymmetries. Gallagher, Gardner and Swan (2011) demonstrate that governance through exit enhances stock liquidity. By contrast, Cohen (2011) shows that a Schedule 13 filing by hedge funds, particularly those close to the target company, leads to a decrease in liquidity, potentially because investors fear trading against an informed investor. 3 blockholders with greater incentive pay, such as hedge funds, govern more effectively; those without incentive pay are unlikely to choose voice to begin with.5 While Yermack (2010) concludes that activism in general leads to little improvement in performance, Brav et al. (2008), Klein and Zur (2009), Clifford and Lindsey (2011), and Boyson and Mooradian (2011a) document significant gains to hedge fund activism, and so hedge fund activism is particularly important from a policy perspective. Kahan and Rock (2007) argue that “hedge fund activism is strategic and ex ante”, whereas “mutual fund and public pension fund activism, if it occurs, tends to be incidental and ex post.” Thus, hedge funds’ decision to acquire a block and their choice of filing is more likely to be driven by governance (cf. Brav et al., 2008), whereas mutual funds typically acquire a block for other reasons, e.g., undervaluation. We note that, while all hedge funds have the option of engaging in voice, several never do so – for example, some funds focus entirely on trading as this is their core skill. We thus focus on activist hedge funds as they both have the ability and willingness to engage in intervention. We identify activist hedge funds as those who have ever engaged in activism with any of their investments; however, such funds may choose to govern through trading for their other holdings. We find that activist hedge funds are more likely to acquire a block (a stake of at least 5%) in firms that exhibit high stock liquidity, measured using the proxy of either Amihud (2002) or Fong, Holden, and Trzcinka (2011). This result supports the voice theories of Kyle and Vila (1991) and Kahn and Winton (1998), and the exit model of Edmans (2009). Consistent with the exit mechanism in particular, the effect of liquidity is stronger in firms with high managerial sensitivity to the stock price. 5They, like us, posit a link between liquidity and block formation by using liquidity as an instrument for blockholdings. They similarly find that liquidity encourages block formation, but reduces the likelihood of activism conditional upon block formation. 4 Having established that liquidity stimulates the entry of hedge fund blockholders, we next examine how it affects their choice of monitoring strategy once they have decided to acquire a block. We address the second challenge – that the threat of governance also matters in addition to actual governance – by using the blockholder’s choice of filing to measure her intent, rather than studying only instances of actual exit and voice. Blockholders who intend to engage in activism are required to file Schedule 13D upon acquiring a block in a public firm and state their activist intentions.6 Blockholders who intend to remain passive are able to file 13G. They will take advantage of this option due to the costs of filing a 13D, described later in Section 1. A separate advantage of using 13D filings is that they are not limited to a specific type of monitoring variable. Norli, Ostergaard and Schindele (2009) examine voice in the form of contested proxy solicitations and shareholder proposals. While these are important instances of activism, relying on two specific vehicles could potentially omit other channels of voice. We find, among the targeted firms, a strong negative relation between stock liquidity and the likelihood of a hedge fund blockholder filing a 13D. This finding is consistent with the view that liquidity weakens governance as it deters the blockholder from engaging in voice (Coffee (1991); Bhide (1993); Maug (2002)). However, it is also consistent with the argument that trading is itself a governance mechanism (Admati and Pfleiderer (2009); Edmans (2009); Edmans and Manso (2011)) and so liquidity merely causes a blockholder to employ a different governance channel, i.e., move from “voice” to “exit” rather than from “voice” to “no governance”. To support the “exit” view over the “no governance” view, we undertake two additional tests. First, we show that liquidity has a particularly large effect in inducing a 6 Examples of activists’ stated intentions filed in a 13D include: change the CEO or board, pursue strategic alternatives, oppose or induce a merger, increase the dividend, induce a buyback, and change the firm’s corporate governance. 5 blockholder to file 13G rather than 13D for firms with high managerial sensitivity to the stock price. Second, we find a significantly positive market reaction to a 13G filing, and that this reaction is particularly strong for firms with above-median liquidity. The above results show that, while liquidity increases the likelihood of a block acquisition, it decreases the likelihood that a 13D is filed conditional upon block acquisition. We show that the first effect outweighs the second, i.e. that liquidity increases the unconditional probability of a 13D being filed. Since liquidity increases the incidence of voice as well as exit, it has an overall positive effect on governance. Finally, we address the third empirical challenge – that liquidity is endogenous – in two ways. First, since we study a governance event (a Section 13 filing) rather than governance characteristics, a regression of a Section 13 filing on lagged liquidity is unlikely to be driven by reverse causality from the future event to past liquidity. Second, we use decimalization as a natural experiment to provide an exogenous source of variation in liquidity. Between August 2000 and April 2001, the stock markets in the U.S. converted to the decimal-pricing system and reduced the minimum tick size from 1/16 dollar to one cent. Bid-ask spreads fell substantially across all market capitalization groups (Bessembinder (2003); Furfine (2003)). Thus, decimalization can instrument for liquidity as it led to an increase in liquidity, but was unlikely to affect a hedge fund’s governance strategy other than through liquidity. All of our results remain robust to using this instrument. We also show that decimalization has a stronger effect on governance in firms with low stock prices, for which a change in tick size has a greater impact on liquidity. This study contributes to two main literatures. First, we build on recent research studying the effect of liquidity on firm outcomes. Fang, Noe, and Tice (2009) identify a causal impact of 6 liquidity on firm performance.7 Bharath, Jayaraman, and Nagar (2010) show that the effect is stronger for firms with higher block ownership, which supports the exit governance mechanism, although they do not study the choice between exit and voice or the effect of liquidity on block acquisition. Like Bharath, Jayaraman, and Nagar (2010), our paper documents a potential mechanism to explain the positive effect of liquidity on firm value found by Fang, Noe, and Tice (2009) – the effect of liquidity on governance. Other papers study the effect of liquidity on voice, but do not consider exit. Heflin and Shaw (2000) document a negative correlation between block ownership and stock liquidity. This is consistent with liquidity hindering blockholder activism, although it could also suggest that block ownership reduces liquidity. In contrast, Norli, Ostergaard, and Schindele (2009) find a strong positive relation between stock liquidity and actual voice. Gerken (2009) finds that liquidity has no effect on governance choices, contrary to our findings that it causes blockholders to choose exit over voice. Our focus on hedge funds, which have both governance mechanisms at their disposal, may account for the difference in results. In addition, while we study the initial ex ante filing decision of a blockholder (13D or 13G), Gerken considers the subsample of 13G filers and investigates whether liquidity causes them to switch ex-post to a 13D. Such switches are much rarer (in our sample there are 42 switches from 13G to 13D out of the 1,112 initial 13G filings, and 31 out of 645 after adding controls), and this reduced power may account for the insignificant results. Kelly and Ljungqvist (2011) show that shocks to liquidity risk, caused by increases in information asymmetry resulting from exogenous broker closures, augment a stock’s required returns and thus reduce its price. 7 On the other hand, Fang, Tian, and Tice (2011) document a potential disadvantage of stock liquidity, that it may attract transient investors who impose short-term pressures on managers and impede firm innovation. In contrast, Kim and Kang show that liquidity leads to a less negative relationship between R&D and the probability of CEO firing, consistent with the idea that liquidity causes the benefits of R&D to be more closely incorporated into prices. 7 Second, the paper contributes to research on the role of hedge funds in corporate governance. Brav et al. (2008), Clifford (2008), Greenwood and Schor (2009), Klein and Zur (2009, 2011), and Boyson and Mooradian (2011a, 2011b) study the effect of hedge fund activism on firm outcomes. While the existing research typically focuses on activism alone, we examine the choice between exit and voice.8 The rest of the paper is organized as follows. Section 1 develops our hypotheses, Section 2 describes the data, Section 3 presents the results, and Section 4 concludes. 1. Hypothesis development and theoretical framework This section lays out our empirical hypotheses and the theoretical framework that underpins them. Our first hypothesis is as follows. H1: Liquidity increases the likelihood that a hedge fund acquires a block. This hypothesis is supported by both voice theories (Kyle and Vila (1991), Kahn and Winton (1998), and Maug (1998)) and exit theories (Edmans (2009)). In contrast, if block acquisition was motivated by undervaluation rather than governance concerns, liquidity should reduce its likelihood as liquidity increases efficiency (Chordia, Roll, and Subrahmanyam (2008)). The governance theories of voice and exit ascribe different roles to governance. Exit involves blockholder trades affecting the stock price, and thus requires the manager to be sensitive to the price to be effective. The effect of liquidity on block formation should therefore be stronger in 8 Clifford (2008) studies the value effect of a hedge fund filing a 13D versus a 13G, but not what determines this filing choice in the first place. 8