Rev Account Stud (2016) 21:808–858 DOI 10.1007/s11142-016-9362-2
What are the consequences of board destaggering? Weili Ge1 • Lloyd Tanlu2 • Jenny Li Zhang3
Published online: 28 June 2016 Ó Springer Science+Business Media New York 2016
Abstract In this paper, we examine the consequences of the decision to destagger the election of directors using a sample of firms that switched from a staggered to a destaggered board structure from 2002 through 2010. We find that the likelihood of destaggering increases in shareholder activism, firm size, and poor prior accounting performance. Furthermore, we find that firms that destagger tend to have larger boards and a lower entrenchment index prior to destaggering. We then use our determinants model to identify a sample of control firms that maintained a staggered board structure. Employing a difference-in-differences research design, we find that, relative to our control firms, firms that destaggered experience declines in Tobin’s q and accounting performance, measured by ROA. In addition, the negative effect on Tobin’s q is most pronounced in firms with greater advisory needs, consistent with the notion that destaggering results in worse performance when the advisory role of boards is more important. Contrary to claims made by proponents of destaggered boards, we find no evidence that CEOs are less entrenched after destaggering. We also provide some evidence suggesting that investment in R&D falls in the post-destaggering period, consistent with the view that after destaggering board members have shortened & Jenny Li Zhang
[email protected] Weili Ge
[email protected] Lloyd Tanlu
[email protected] 1
Michael G. Foster School of Business, University of Washington, Paccar Hall, Box 353226, Seattle, WA 98195, USA
2
D’Amore-McKim School of Business, Northeastern University, 360 Huntington Ave., Boston, MA 02115, USA
3
Sauder School of Business, The University of British Columbia, 2053 Main Mall, Vancouver, BC V6T 1Z2, Canada
123
What are the consequences of board destaggering?
809
incentive horizons. Taken together, our evidence is contrary to the earlier studies that claim that destaggered boards are generally optimal and value-increasing. Keywords Staggered board Classified board Consequences of board destaggering JEL Classifications G34 G30 M29
1 Introduction Directors of staggered boards are divided into several classes, with only one class of directors elected in a given annual shareholders’ meeting.1 The stated benefit of staggered boards is continuity, as generally two-thirds of the board will stay in place in any given year (Koppes et al. 1999). Staggered boards have been widely criticized for looking out for themselves and management instead of shareholders; their opponents claim that the annual election of directors makes directors more accountable to shareholders and improves firm performance (Bebchuk and Cohen 2005; Bebchuk et al. 2002). Guided by this view, efforts led by the Harvard Law School Shareholder Rights Project and proxy advisory firms like Institutional Shareholder Services have led to the recent wave of destaggering boards.2 We examine whether board destaggering affects firm performance and the mechanisms through which board destaggering affects firm performance. It has been well documented in the literature that staggered boards are associated with lower firm value and director effectiveness (Bebchuk and Cohen 2005; Faleye 2007; Cohen and Wang 2013). This evidence is consistent with the argument that staggered boards are captive and entrenched, insulating managers from the market for corporate control and allowing managers to potentially engage in opportunistic behaviors (such as shirking, empire-building, and other actions that extract value from shareholders). These studies suggest that the presence of a staggered board is a symptom of corporate governance failure and therefore board destaggering would lead to an improvement in shareholder value. In contrast, a thought piece by Koppes et al. (1999) argues that staggered boards bring value to shareholders by promoting the continuity and stability of the corporation’s leadership and preventing drastic changes in corporate strategies. A 3-year term likely offers directors stronger incentives to invest in firm-specific human capital, which results in better director performance in terms of both monitoring and advising management. On the other hand, annual elections might reduce directors’ incentive horizons, which could result in myopia; this view has recently been highlighted by Subramanian (2015). In fact, the boards of companies 1
Staggered boards of directors are also known as classified boards. Typically directors of staggered boards are divided into three classes, and each election is for a term of 3 years.
2
The Harvard Law School Shareholder Rights Project reports that its work resulted in a total of 121 S&P 500 and Fortune 500 companies adopting (or committing to adopt) staggered board structures between 2012 and 2014. In addition, Tonello and Aguilar (2012) show that, among shareholder proposals that target changing corporate governance practices, the largest proportion relates to destaggering boards.
123
810
W. Ge et al.
such as Netflix, Boeing, and Safeway recommended that shareholders vote against the proposals to destagger their respective boards for the aforementioned reasons. Therefore it may be premature to conclude that a staggered board is unconditionally a poor governance practice for most firms and that destaggering can lead to an improvement in shareholder value. Given the mixed views on destaggering as well as the increasing rate at which public firms are moving toward destaggered board structures, we delve into the consequences when firms destagger their boards. We start by examining performance consequences, if any, of destaggering and then dig deeper into whether destaggering alleviates CEO entrenchment (consistent with the view of proponents of destaggering) or reduces incentive horizons of directors, causing directors to be more myopic (consistent with the view of opponents). While prior studies have largely taken the staggered nature of boards as given and have focused on the association between staggered boards and firm value, Adams et al. (2010) argue that governance structures are chosen endogenously in response to firm-specific governance issues. Therefore we begin our analyses by investigating the determinants of a firm’s decision to destagger. We compile a sample of 188 firms that destaggered from 2002 through 2010. We find that firms under more pressure from activist shareholders are more likely to destagger their boards; the extent of shareholder activism is evident in the increased likelihood of having a shareholder proposal that advocates having a destaggered board. This finding is consistent with the increasing trend of shareholder empowerment after the Sarbanes–Oxley Act of 2002 (Ferri 2012). We also find that destaggering firms tend to be larger and have poor prior accounting performance (measured as accounting-based return on assets for the prior 2 years), consistent with the idea of Koppes et al. (1999) that shareholders blame poor performance on firms’ board structure and demand changes to the board. In terms of governance attributes, firms more prone to entrenchment, as measured by the entrenchment index, or E-index, proposed by Bebchuk et al. (2009), are less likely to destagger their boards, consistent with the claim that directors of such firms prefer to keep the staggered board structure to deter takeovers. Using our determinants model, we generate a control sample by matching each destaggering firm to a firm that has always had a staggered board during our sample period. We use this matched sample and perform a difference-in-differences analysis to examine whether board destaggering alleviates the alleged problems that accompany director and manager entrenchment (i.e., those claimed by much of the literature opposing staggered boards). Generally, we find that firm performance deteriorates in terms of Tobin’s q and return on assets, contrary to their claims. In our supplemental tests, we also find evidence that the negative effect on performance (Tobin’s q) is most pronounced in firms with greater advisory needs from the board, consistent with the notion that destaggering results in worse performance when the advisory role of boards is more important. We attempt to explain our performance results by directly investigating the mechanisms through which destaggering affects performance, as suggested in the literature by both proponents and opponents of destaggering. We first examine the effect of destaggering on management entrenchment and board monitoring by
123
What are the consequences of board destaggering?
811
focusing on director and CEO turnover, as well as CEO pay-for-performance sensitivity. We do not find any significant changes in destaggered boards’ CEO or director turnover relative to those of matched control firms, even after conditioning on firm performance. We also find evidence of reduced CEO pay-for-performance sensitivity following destaggering, indicating the potential for the reduced effectiveness of board monitoring. Taken together, our results contradict the argument of proponents that a destaggered board results in better governance. Not only does overall performance decline after destaggering, but there is also a lack of evidence that managers are penalized for poor performance. Secondly, we explore whether destaggering differentially affects the incentive horizons of directors. Opponents claim that shortening a director’s term from 3 to 1 year may lead directors to focus on short-term performance instead of long-term value creation. We find evidence that destaggering firms experience a decline in their research and development expenditure, consistent with directors having shorter horizons and investing less in potentially longer-term projects such as R&D. Finally, we examine how destaggering influences board characteristics and director compensation. On average, there is no statistically significant increase in director turnover after destaggering relative to control firms. We also examine whether there is a relative change in the characteristics of directors actually elected to destaggered boards; we do not find significant changes in board size or the percentage of independent directors on destaggered boards relative to control firms. In terms of newly elected director characteristics (e.g., age, experience, achievements), we also do not find any significant differences between directors elected after destaggering and directors elected beforehand or those elected in the control sample. Therefore our performance and R&D results do not seem to be driven by changes in the types of newly elected directors. In addition, while we find a marginally significant decline in the level of total director compensation after destaggering relative to control firms, we do not observe a significant change in the proportion of equity-based compensation, suggesting that directors are not awarded more equity-based compensation to mitigate potential shorter horizon concerns resulting from annual elections. Overall, our findings suggest that, contrary to the implications of the research on staggered boards and claims made by activist investors, destaggering does not appear to lead to improved firm performance on average. On the contrary, our evidence suggests that destaggering could lead to managerial short-termism and less effective board monitoring. We note that, if firms make destaggering decisions ‘‘optimally,’’ we would not expect to observe an association between board destaggering and future firm performance or investment decisions, especially given that our difference-indifferences design already accounts for the likelihood that destaggering firms differ systematically from firms that choose not to destagger their boards. Our findings of declining future performance are consistent with the conjecture that some firms are not making optimal decisions with respect to moving to staggered boards. However, to the extent that our performance variables explain the destaggering decision but do not fully capture expected future performance at the time of destaggering, our findings of declining performance may indicate that firms with anticipated poor
123
812
W. Ge et al.
future performance are those that choose to destagger their boards. In other words, board destaggering would only be a cosmetic governance change. Nevertheless, the latter explanation would still be inconsistent with the assertion that destaggering would lead to an improvement in shareholder value. Our study contributes to the literature in several ways. First, our findings inform the current debate on destaggering boards. While proxy advisors have unconditionally called for the destaggering of all public boards (Davidoff 2012), there has been limited evidence on the determinants and consequences of the recent trend of board destaggering. We build on previous research that examines determinants of destaggering (Ganor 2008; Guo et al. 2008) and use a more recent and carefully culled sample of destaggering firms to tease out the determinants of the decision. Although activists for shareholders’ rights argue that staggered boards are valuedestroying for shareholders, our findings suggest that shareholder activism plays a significant role in a board’s destaggering decision but this one-size-fits-all approach for board structure does not always benefit shareholders. Our findings are consistent with the implications of the work of Larcker et al. (2015), who find that proxy advisors’ recommendations do not result in increased firm value. Furthermore, our results corroborate the arguments made in the thought piece by Koppes et al. (1999) that whether the board is staggered is not necessarily critical in ensuring that the board provides good monitoring and advising. Our paper also extends and builds on the literature on staggered boards by speaking directly to the effects of board destaggering on firm performance. The earlier papers on staggered boards, which are mostly cross-sectional studies, suggest that these boards are a suboptimal governance mechanism due to entrenchment effects (Bebchuk and Cohen 2005; Faleye 2007). Moreover, our results complement the findings in a concurrent working paper by Cremers et al. (2015). Like our work, theirs questions the common view in the literature that staggered boards are valuedecreasing. They question the research methodology of Bebchuk and Cohen (2005) by replicating the study and subjecting the data to a series of additional tests. They find that staggered boards are associated with increased firm value, contrary to the findings in Bebchuk and Cohen (2005). While our results are consistent with those of Cremers et al. (2015), the two studies differ on at least two dimensions. First, as previously mentioned, we use a difference-in-differences approach, which we believe addresses the issue of the endogenous choice of board structure suggested by Adams et al. (2010) and acknowledged by Cremers et al. (2015). Second, in addition to Tobin’s q, we examine both accounting performance and the mechanisms by which board destaggering may help or hurt firm value and accounting performance (e.g., changes in R&D). We find that firms that destagger tend to have poorer accounting performance. We then examine the reasons behind the poorer performance, including board monitoring, manager entrenchment, investment decisions, and board characteristics. The remainder of the paper is organized as follows. Section 2 discusses the background and our predictions. Section 3 describes our sample and research design. Section 4 presents our empirical results, and Section 5 concludes.
123
What are the consequences of board destaggering?
813
2 Motivation and predictions 2.1 Background and research motivation The role of boards of directors is an issue of fundamental importance in the literature on corporate governance. Boards are generally involved in the hiring and firing of CEOs, setting strategy, and selecting major projects (Adams et al. 2010). A board’s actions could be influenced by various board attributes, such as the percentage of independent directors, board size, and CEO-chairman duality. An attribute that has received much attention from activist investors and researchers relates to the timing of the election of the directors within a board. Boards are either staggered, in which only a fraction of directors stands for election in any given year, or destaggered, in which all directors stand for election at the same time—usually annually.3 For firms with staggered boards, usually only a third of the directors stands for election in a given year at the annual shareholders’ meeting to serve 3-year terms. Therefore, for a staggered board, it would take at least 2 years to acquire a majority of seats on the board and 3 years to completely turn over the board. Efforts by shareholder rights activists have led to the proliferation of board destaggering over the last decade; however, there is limited evidence on whether destaggering achieves the goals claimed by these activists (e.g., improving firm performance). To shed light on this issue, we use more recent hand-collected data and a difference-in-differences research design to examine whether board destaggering benefits shareholders by reducing the extent of management entrenchment. Specifically, we investigate how destaggering affects firm performance, director and management turnover, and investment strategies. Below, we first develop our predictions following the arguments against staggered boards. Then we discuss the views held by proponents for staggered boards and the corresponding predictions. 2.2 Predictions based on the view held by opponents of staggered boards Recent research has criticized staggered boards for entrenching and protecting ineffective managers and directors. According to the literature, the staggering of elections renders boards ineffective and unable to quickly fire and replace poorly performing managers. Opponents argue that the use of a staggered board, particularly in combination with mechanisms such as a poison pill, inhibits the replacement of incumbent directors and consequently allows for the potential entrenchment of a poorly performing CEO. Empirical evidence to date seems to be consistent with the above view of the entrenchment effect of staggered boards. For instance, staggered boards are associated with lower firm valuation, lower sensitivity of CEO compensation to performance, and lower sensitivity of CEO turnover to firm performance (Faleye 2007; Bebchuk and Cohen 2005; Frakes 3
Staggered boards are also known as ‘‘classified’’ boards and categorize directors into several groups or classes. Every year, only one group or class of directors is up for election; the others maintain their seats and stand for election at a future date. Destaggered boards are also known as declassified, single-class, or unitary boards.
123
814
W. Ge et al.
2007). Firms with staggered boards have also been found to have more valuedecreasing acquisitions (Masulis et al. 2007). Furthermore, takeover targets with staggered boards are associated with lower returns to shareholders (Bebchuk et al. 2002). The above-mentioned evidence has motivated the recent wave of investor activism to push for the destaggering of boards. For instance, the Harvard Law School Shareholder Rights Project (SRP) considers the annual election of directors as best practice and has been advocating that shareholders submit proposals to destagger boards.4 The aforementioned studies suggest that directors and CEOs are held more accountable to firm performance when director elections are annual rather than staggered. Therefore, to the extent that the threat of removal from office (either from the board or as CEO of the firm) as a result of board destaggering strengthens both directors’ and CEO’s accountability, we expect that board destaggering would lead to improvements in accounting and stock market-based performance for firms that destagger their boards relative to similar firms who choose to maintain a staggered board structure. Furthermore, if board destaggering truly makes directors and CEOs more accountable for firm performance, we expect that board destaggering is associated with higher director and CEO turnover in the presence of poor firm performance. We also expect CEO compensation to be more sensitive to firm performance under a regime of more effective board monitoring. Finally, given that both directors and CEOs of firms with destaggered boards are held more accountable year to year, the incentives to compromise firm value for selfenrichment are reduced. Thus, in line with these arguments, destaggering is expected to lead to higher investment efficiency as well. 2.3 Predictions based on the view held by proponents of staggered boards There is also a contrary argument that staggered boards are valuable for shareholders because they promote continuity and stability of leadership and allow the board to focus on long-term strategies (Koppes et al. 1999). Typically, the board of directors is responsible for developing and implementing long-term strategies, which often take years. The staggered election of the board directors prevents drastic changes in corporate strategy that might happen if the entire board were replaced each year. Destaggering and moving toward annual elections increase the possibility of directors being removed after 1 year and thus likely shortens directors’ incentive horizons; as a result, the board members might be less inclined to invest their human capital in understanding and investing in projects that would take longer to implement, regardless of the potential profitability of these projects. Because destaggering causes directors’ terms to be reduced from 3 to 1 year, it may lead these directors to focus on short-term performance rather than long-term value creation. In this scenario, where directors are expected to be more myopic, we expect destaggering to be associated with poorer subsequent firm performance. In 4
See http://srp.law.harvard.edu/.
123
What are the consequences of board destaggering?
815
addition, Dechow and Sloan (1991) show that CEOs spend less on R&D in their final years of office, suggesting an effect of the CEO horizon on corporate investments. Following the same logic, myopia of both directors and CEOs may also lead to a decline in long-term investments, reflected for instance in R&D expenditures.5 We note, however, that having directors actually turn over and exit the board is not a necessary condition for strategy formulation or implementation to change. The threat of removal from the board that directors face every year (as opposed to every three in a staggered board) changes their incentive horizons and may make them more short-term focused. Annual election provides shareholders more opportunities to express their dissatisfaction with companies by withholding votes for the directors. Grundfest (2004) notes that even symbolic votes can create negative publicity and embarrassment. The change in incentives can therefore affect the behavior of directors, regardless of whether a board’s composition and structure changes drastically. Moreover, research on director elections documents that shareholders’ ability to have a voice in electing or removing the board members is limited, due to the fact that the vast majority of director elections are uncontested. To the extent that destaggering increases the likelihood of a proxy contest, we would expect the incumbent directors to face a more meaningful risk of being replaced, resulting in greater turnover after destaggering. Additionally, directors’ terms (and subsequently their horizon incentives) may be reduced, as voluntary turnover likely increases if it is costly for directors to continuously face election (Lipton and Savitt 2007).
3 Sample selection and research design 3.1 Sample selection We begin our sample construction with the governance data in the RiskMetrics Database, available through Institutional Shareholder Services (ISS) Governance Services, which provides the classification status of the board for S&P 1500 companies. Table 1 Panel A presents the sample selection process. We identify 390 firms that destaggered their boards from 2002 through 2010. We start our sample period in 2002 because the Sarbanes–Oxley Act, which became effective in 2002, may be correlated with many significant changes in the corporate governance environment (e.g., shareholder activism) and thus the determinants of the decision to destagger boards. Our primary interest is to understand the consequences of the
5
Destaggering can also affect takeover risk and thereby R&D spending by changing the incentives of the managers. Chemmanur and Tian (2013) find that firms with more anti-takeover provisions (including staggered boards) are more innovative, as measured by the number of patents, which is the output of R&D activities. The idea is that these anti-takeover provisions insulate managers from short-term pressures arising from the equity market and thus allow them to focus on long run value-creation. Consistent with this finding, removing the anti-takeover provision (i.e., staggered board) increases management pressure to focus on the shorter term instead of the longer term, resulting in lower R&D.
123
816
W. Ge et al.
Table 1 Sample description Panel A: Sample selection Destaggering firm from RiskMetrics (2002–2010)
390
Less: RiskMetrics data errors Firms that never destaggered
(8)
Firms that always have had destaggered boards
(41)
Firms that actually destaggered boards before 2002 (Risk Metrics’ dates are wrong)
(36)
Less: duplicates or destaggered as a result of a merger or acquisition
(30)
Less: proxy statements cannot be found
(12)
Less: no Compustat match
(41)
Less: REITS
(10)
Less: dual-class firms
(15)
Less: missing SIC codes
(9)
Total
188
Panel B: Frequency of destaggering firms by year Year
Number of destaggering firms
Percentage of total destaggering firms (%)
2002
13
2003
16
6.91 8.51
2004
19
10.11
2005
32
17.02
2006
28
14.89
2007
39
20.74
2008
21
11.17
2009
18
9.57
2010
2
1.06
Total
188
100.00
Panel C: Frequency of destaggering firms by industry Industry
Destaggering firms (number)
Destaggering firms (%)
Compustat population (%)
Consumer nondurables
7
3.72
4.42
Consumer durables
3
1.60
2.09
Manufacturing Energy Chemicals Business equipment
17
9.04
8.38
7
3.72
4.88
7
3.72
1.90
21
11.17
15.73
Telecom
3
1.60
3.41
Utilities
25
13.30
2.97
Wholesale, retail, and some services
26
13.83
8.09
Health
15
7.98
8.54
Finance
38
20.21
24.82
Other
19
10.11
14.77
Total
188
100.00
100.00
123
What are the consequences of board destaggering?
817
recent wave of destaggering due to increasing shareholder activism; therefore we focus on the sample period starting from 2002.6 The RiskMetrics Database has two limitations. First, it lacks annual data before 2006. It only provides data on various corporate governance provisions for each included company in 2002, 2004, and 2006 and annual data after that. Second, the year that is identified by RiskMetrics as the year of destaggering is unclear and inconsistent, particularly if the destaggering phased in over several years. The data on RiskMetrics span three different possible destaggering events: (1) when the board chose to adopt a destaggered structure, regardless of when the destaggering actually happened; (2) when all new directors elected during a year have 1-year terms (while the other directors finish out their respective terms); and (3) when all directors have 1-year terms. Thus, based on these 390 firms, we hand-collect data from the destaggering firms’ proxy statements and identify the three different dates mentioned. As a result of our hand-collection, we eliminate 85 destaggering cases that appear to be data errors in RiskMetrics, resulting in 305 firms that destaggered their boards from 2002 through 2010. Among these, we remove 30 that involve duplicates or being destaggered as a result of a merger or acquisition. In addition, we cannot identify 12 firms’ proxy statements and cannot find Compustat matches for 41 firms. Finally, following Bebchuk and Cohen (2005), we remove 10 REITS, 15 dual-class firms, and nine firms without SIC codes, resulting in a final sample of 188 destaggering firms. For our analyses, we consider all available firm-years before the board’s decision to adopt a destaggered structure (the first date mentioned above) as the pre-destaggering period; we consider all firm-year observations after new directors are elected to 1-year terms (which is the second date mentioned above) as part of the post-destaggering period.7 Table 1 Panel B reports the distribution of destaggering firms by year, and Panel C presents the distribution of destaggering firms across industries. We follow the Fama and French 12 industry classification scheme. It appears that the destaggering firms in our sample are generally distributed across industries in a similar fashion to the Compustat population. The only exception is the utilities industry; 13.30 % of our destaggering firms are in that industry, while only 2.97 % of Compustat firms are. 3.2 Research design We perform a difference-in-differences analysis that takes endogeneity into consideration to examine the consequences of destaggering. Destaggering firms likely differ systematically from firms that did not destagger their boards as our determinants analysis demonstrates; therefore the difference in the consequences of destaggering could be driven by omitted variables that are correlated with both destaggering and the consequence variables. We perform a difference-in-differences 6
We focus on larger firms covered by RiskMetrics to ensure that we can gather all the variables for the determinants model and that we have a higher likelihood of finding a comparable control sample to be used in our consequences tests. Our conclusions thus might not be generalizable to firms outside of the coverage of RiskMetrics.
7
In unreported tests, we also consider the periods after all directors are elected to 1-year terms for robustness and find similar results to those reported in this paper.
123
818
W. Ge et al.
analysis to mitigate the omitted variable concern driven by time-invariant differences between destaggering firms and control firms as well as the same temporal trends for both groups of firms. We generate a matched control sample using propensity-score matching (Lemmon and Roberts 2010). We discuss our matching in greater detail shortly below. We analyze the determinants of the destaggering decision using the following logistic regression specification: DESTAGGERit ¼a0 þa1 PROPOSALit1;t2 þa2 %INSOWNit1 þa3 TOBINQit1 þa4 TOBINQit2 þa5 ROAit1 þa6 ROAit2 þa7 RETURNit1 þa8 RETURNit2 þa9 EINDEXit1 þa10 DELAWAREit1 þa11 BOARDSIZEit1 þa12 %INDDIRECTit1 þa13 CEOTENUREit1 þa14 CEOCHAIRMANit1 þa15 PCTOWNit1 þa16 INTANGit1 þa17 RDit1 þa18 CAPEXit1 þa19 FCFit1 þa20 SIZEit1 þa21 AGEit1 þa22 NBSEGit1 þa23 NGSEGit1 þa24 LEVERAGEit1 þ
48 X q¼1
INDUSTRY q þ
10 X
YEARp þet
p¼1
ð1Þ We estimate the above regression using all the RiskMetrics firms that have staggered boards and either (a) continued to have staggered boards or (b) chose to destagger their boards at any point within our sample period. The indicator variable DESTAGGER equals zero for firms that did not destagger. For destaggering firms, DESTAGGER equals zero in the years before destaggering and one during the year of destaggering (i.e., when the destaggering decision was made, as discussed in Sect. 3.1); we exclude the destaggering firms’ post-destaggering observations from estimating Eq. (1). We measure all the independent variables before the destaggering year to capture the information that was available to shareholders when the destaggering decision was made. We regress DESTAGGER on our measures of investor activism, financial performance, governance, and the board’s importance in terms of both its monitoring and advisory roles. We start with two measures for shareholder activism: (1) an indicator variable equal to one if there is at least one shareholder proposal to destagger the board in the previous two proxy-voting seasons and zero otherwise (PROPOSAL) and (2) the percentage of institutional ownership obtained from the Thomson Financial database (%INSOWN). Next, we measure prior firm performance using Tobin’s q (TOBINQ), return on assets (ROA), and market-adjusted stock returns over the fiscal year (RETURN). We measure each of these performance variables for each of the 2 years before the decision to destagger to capture the trend in firms’ financial performance. We conjecture that poor performance increases the demand to reform governance practices within the firm. Hence we expect negative associations between past performance and the decision to destagger.
123
What are the consequences of board destaggering?
819
We then include measures of governance in the model: the entrenchment index (EINDEX), constructed following Bebchuk et al. (2009); whether a firm is incorporated in Delaware (DELAWARE); board size (BOARDSIZE); the percentage of independent directors on board (%INDDIRECT); CEO tenure (the number of years the manager has been in the CEO position, CEOTENURE); whether the CEO is also chairman of the board (CEOCHAIRMAN); and the percentage of shares held by the top five executives at the end of the year (PCTOWN). Finally, we include several measures for monitoring and advisory needs (i.e., greater need suggests more importance of the board as a governance mechanism): asset intangibility (INTANG, measured as the ratio of intangible assets to total assets); research and development expenditure (RD), defined as research and development expenditure deflated by average total assets; capital expenditure (CAPEX), defined as capital expenditure deflated by average total assets; free cash flows (FCF), defined as the sum of net income and depreciation minus capital expenditure scaled by average total assets; firm size (SIZE); firm age (AGE); complexity (measured by the number of business and geographic segments, NBSEG and NGSEG); and financial leverage (LEVERAGE).8 We also include year fixed effects and Fama–French 48 industry fixed effects, to control for time effects and industry characteristics. We cluster our standard errors by firm. All variables are defined in ‘‘Appendix 1’’. In ‘‘Appendix 2’’, we discuss the reasons and predictions for the determinants in Eq. (1). Using our determinants analysis, we form a propensity score for all firm-year observations included in our determinants analysis. From the list of firms that have always had a staggered board and have never destaggered during our sample period, we then select the firm that has the closest predicted probability of destaggering (i.e., propensity score) to each of our treatment firms during the year immediately before the destaggering decision. We use this set of control firms for all subsequent tests. In our multivariate analyses of the effects of destaggering, we continue to include many explanatory variables from Eq. (1), to control for any remaining differences between the treatment and control groups. We employ the following model to test the consequences of destaggering decision: CONSEQUENCEit ¼ b0 þ b1 DESTAGGERi þ b2 AFTERit þ b3 DESTAGGERi AFTERit X X X þ bk CONTROLSit þ rp YEARp þ hq INDUSTRYq þ eit
ð2Þ DESTAGGER is an indicator variable, equal to one for destaggering firms. CONSEQUENCE refers to the consequences related to performance, manager and director turnover, and investment strategies discussed in Sect. 2. Specifically, we examine three measures of financial performance: Tobin’s q (which has been used conventionally in the literature), annual return on assets (ROA), and value-weighted market-adjusted annual returns (RETURN). With respect to turnover, we examine both management turnover (an indicator variable equal to one if the CEO from the previous 8
Firms that need more monitoring (and therefore might benefit from having a destaggered board) are also likely to have strong advisory needs (and therefore might also benefit from having a staggered board). Therefore it is very difficult to empirically separate one effect from the other.
123
820
W. Ge et al.
year is no longer the CEO during the current year and the CEO age is below 659) and director turnover (the percentage of directors from the previous year who are no longer on board during the current year). We also examine investment strategies using research and development expenditures (R&D) and net expenditures for total investment (ITOTAL, which Richardson (2006) defines as the sum of all outlays on capital expenditures, acquisitions, and R&D, less receipts from the sale of property, plant, and equipment). We scale both investment variables by average total assets. In Eq. (2), AFTER is an indicator variable, equal to one for the period following the destaggering decision. For control firms, AFTER is equal to one for the same years as AFTER for the corresponding matched destaggering firms. The coefficient of interest, b3, represents the difference in the dependent variable between destaggering firms and control firms from the pre-destaggering period to the postdestaggering period. All variables are defined in ‘‘Appendix 1’’. Furthermore, we include year and industry fixed effects to control for macroeconomic and industryspecific factors and cluster standard errors by year and industry. For each consequence variable, we include all years following the destaggering decision as part of the analysis. Specifically, our analysis starts on the fiscal year immediately after the shareholders’ meeting when new directors are elected to 1-year terms.10
4 Empirical results 4.1 Determinants of board destaggering Table 2 Panel A reports the logistic regression results of estimating Eq. (1). Our results indicate that, while the coefficient on institutional ownership (%INSOWNt-1) is insignificant, firms having shareholder proposals of board declassification (PROPOSALt-1,t-2) are more likely to destagger their boards.11,12 Our results are consistent with recent evidence suggesting that low-cost shareholder activism in the 9
To capture CEO turnover due to poor performance as opposed to retirements or health concerns, we have excluded CEOs above 65 years of age at the time of CEO turnover. Guo and Masulis (2015) use the same measure to measure involuntary CEO turnover.
10 The effects of destaggering may not be immediately evident; we do not necessarily expect drastic changes to happen instantly upon destaggering. Thus, in addition to the reported results for the overall sample, we also run the same set of analyses excluding the first year in the post-destaggering period. We find similar results to those reported when omitting the first year after destaggering. These untabulated results imply that our results are generally persistent and are not just driven by the first year after destaggering. 11 Given the significant role of shareholder proposals in determining firms’ destaggering decision, we also examine the characteristics of firms that receive shareholder proposals to destagger their boards. In the untabulated logistic analysis, we find that the firms that receive shareholder proposals to destagger tend to be larger and older. They also exhibit poorer market performance (measured in terms of Tobin’s q and prior period returns). 12 We further break down institutional ownership into dedicated investors, quasi-indexers, and transient investors (per Bushee 1998) to determine whether destaggering is demanded by only the short-term (transient) investors or by truly activist longer-term investors. We do not find a significant coefficient on either type of institutional investor, suggesting that board destaggering is not demanded by any particular type of institutional investor.
123
– ? ? ? ? ? ? ? ? ? ?
RETURNi,t-1
EINDEXi,t-1
DELAWAREi,t-1
BOARDSIZEi,t-1
%INDDIRECTi,t
CEOTENUREi,t-1
CEOCHAIRMANi,t-1
PCTOWNi,t-1
INTANGi,t-1
RDi,t-1
CAPEXi,t-1
?
0.200
–
RETURNi,t-1
AGEi,t-1
-1.625
–
ROAi,t-2
?
-0.137***
–
ROAi,t-1
SIZEi,t-1
-0.187
–
?
-0.060
–
TOBINQi,t-1
TOBINQi,t-2
FCFi,t-1
0.104
?
%INSOWNi,t-1
0.004
0.153***
4.958***
2.852
-0.324
0.087
-0.004
-0.001
0.042*
0.068
-0.095
-1.901***
-4.514***
0.124
1.223***
?
-5.808***
?
Coefficients
Intercept
Pred. sign
Dependent variable: DESTAGGERit
PROPOSALi,t-1,t-2
Panel A: Logistic regression
Table 2 The determinants of the decision to destagger the board: logistic regression
0.257
0.001
0.002
0.133
0.859
0.659
0.146
0.379
0.593
0.844
0.070
0.495
0.008
0.277
0.544
0.002
0.003
0.149
0.523
0.763
0.000
0.000
p value
What are the consequences of board destaggering? 821
123
123 35.78 % 531.71
Pseudo R2
Max-Rescaled R2
Likelihood ratio
0.055 0.057
1.929
2.000
0.058
0.062
0.085
TOBINQi,t-1
TOBINQi,t-2
ROAi,t-1
ROAi,t-2
RETURNi,t-1 0.015
1.525
1.494
0.730
0.703
0.000
0.053
%INSOWNi,t-1
0.408 0.742 1.647 1.737 0.044 0.041 0.015
\ \ [ [ [ [ [
Mean (C)
Median (B)
Mean (A)
-0.026
0.038
0.038
1.333
1.303
0.798
0.000
Median (D)
Destaggering firms (N = 169)
Predicted
(N = 6960)
Other staggered firms in the RiskMetrics population
PROPOSALi,t-1,t-2
Variables
0.784
0.523
0.514
p value
-0.069***
-0.021***
-0.014**
-0.263***
-0.282***
0.039**
0.355***
Difference in Mean (C)–(A)
Panel B: Descriptive statistics of destaggering firms during the year of destaggering versus other (staggered) firm-years in the RiskMetrics population
7129 7.19 %
N
Yes
-0.095 Yes
?
LEVERAGEi,t-1
0.017
0.016
Industry fixed effects
?
Coefficients
Year fixed effects
?
NGSEGi,t-1
Pred. sign
Dependent variable: DESTAGGERit
NBSEGi,t-1
Panel A: Logistic regression
Table 2 continued
822 W. Ge et al.
0.609
0.062
0.717
0.028
0.054
0.035
7.407
20.435
3.060
2.888
0.213
CEOCHRMNi,t-1
PCTOWNi,t-1
INTANGi,t-1
RDi,t-1
CAPEXi,t-1
FCFi,t-1
SIZEi,t-1
AGEi,t-1
NBSEGi,t-1
NGSEGi,t-1
LEVERAGEi,t-1 0.207
2.000
2.000
17.000
7.273
0.040
0.039
0.000
0.783
0.036
1.000
4.753
6.875
CEOTENUREi,t-1 (unlogged)
?
?
?
?
?
?
?
?
?
?
?
?
?
?
9.000 75.000
9.319
?
?
[
1.000
71.781
0.594
DELAWAREi,t-1
2.000
0.016
BOARDSIZEi,t-1
2.253
%INDDIRECTi,t-1
0.108
EINDEXi,t-1
0.246
2.373
2.858
26.746
8.746
0.031
0.046
0.019
0.709
0.031
0.663
5.630
77.966
10.763
0.550
2.435
0.014
Mean (C)
Median (B)
Mean (A)
0.223
1.000
1.000
25.000
8.946
0.030
0.036
0.000
0.777
0.019
1.000
4.304
80.000
11.000
1.000
2.000
-0.032
Median (D)
Destaggering firms (N = 169)
Predicted
(N = 6960)
Other staggered firms in the RiskMetrics population
RETURNi,t-2
Variables
0.033**
-0.503***
-0.202
6.310***
1.336***
-0.003
-0.008**
-0.009***
-0.008
-0.031***
0.054
-1.245***
6.202***
1.426***
-0.044
0.182*
-0.094***
Difference in Mean (C)–(A)
Panel B: Descriptive statistics of destaggering firms during the year of destaggering versus other (staggered) firm-years in the RiskMetrics population
Table 2 continued
What are the consequences of board destaggering? 823
123
123
0.014
0.046
0.040
8.668
CAPEXi,t-1
FCFi,t-1
SIZEi,t-1
77.575
%INDDIRECTi,t-1
RDi,t-1
10.822
BOARDSIZEi,t-1
0.690
0.521
DELAWAREi,t-1
0.032
2.266
EINDEXi,t-1
PCTOWNi,t-1
0.004
RETURNi,t-2
INTANGi,t-1
0.078
RETURNi,t-1
4.926
11.000
0.050
ROAi,t-2
0.633
1.000
0.054
ROAi,t-1
CEOCHRMNi,t-1
1.348
1.617
TOBINQi,t-2
CEOTENUREi,t-1 (unlogged)
1.359
1.596
TOBINQi,t-1
8.845
0.035
0.042
0.000
0.736
0.019
1.000
3.685
81.818
2.000
0.001
0.061
0.046
0.054
0.754
0.734
0.000
0.377
%INSOWNi,t-1
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
8.746
0.031
0.046
0.019
0.709
0.031
0.663
5.630
77.966
10.763
0.550
2.435
0.014
0.015
0.041
0.044
1.737
1.647
0.742
0.408
Mean (C)
Median (B)
Mean (A)
Destaggering firms (N = 169)
Predicted
(N = 169)
Matched control firms
PROPOSALi,t-1,t-2
Variables
8.946
0.030
0.036
0.000
0.777
0.019
1.000
4.304
80.000
11.000
1.000
2.000
-0.032
-0.026
0.038
0.038
1.333
1.303
0.798
0.000
Median (D)
Panel C: Descriptive statistics of destaggering firms during the year of destaggering versus matched control (staggered) firm-years
Table 2 continued
0.078
-0.009
0.000
0.005
0.019
-0.001
0.030
0.704
0.391
-0.059
0.029
0.169
0.010
-0.063
-0.009
-0.010
0.120
0.051
0.008
0.031
Difference in mean (C)–(A)
824 W. Ge et al.
2.308
0.228
LEVERAGEi,t-1
0.018
0.065
0.201
157
N
157
0.425
-0.181 %
0.205 % 157
0.660
157
0.202
-0.536 %
Median
Panel D reports the mean and median 3 day (5 day) value-weighted market-adjusted return around the destaggering announcement. All p values are two tailed
Panels B and C report summary statistics of the determinants of destaggering. ***, **, * indicate that the destaggering firm-years are significantly different from the nondestaggering firm-years at the 1, 5, and 10 % levels, respectively, based on a two-tailed t-test for the mean. All the variables are defined in ‘‘Appendix 1’’. Each of the continuous variables is winsorized at 1 and 99 % to mitigate outliers
Panel A reports the logistic regression results for the decision to destagger for a sample of 169 firms that destaggered and the control group of firms that continue to have a staggered board throughout the sample period. The dependent variable is an indicator variable that equals one for the sample firms that destaggered classified boards in the destagger year. Coefficient estimates and p values are reported. *, **, and *** denote two-tailed statistical significance at 10, 5, and 1 %, respectively. Standard errors are clustered by firm. All the variables are defined in ‘‘Appendix 1’’
0.074 %
0.842
Market-adjusted return
p value
Mean
0.223
1.000
Mean
0.246
2.373
-1.355
5 day return (-2, ?2) Median
=
=
1.000
25.000
Median (D)
Difference in mean (C)–(A)
3 day return (-1, ?1)
Panel D: Short-window stock market reaction to destaggering decision
0.210
1.000
2.858
26.746
NGSEGi,t-1
=
2.657
=
33.000
28.101
AGEi,t-1
NBSEGi,t-1 1.000
Mean (C)
Median (B)
Destaggering firms
Mean (A)
Predicted (N = 169)
Matched control firms
(N = 169)
Variables
Panel C: Descriptive statistics of destaggering firms during the year of destaggering versus matched control (staggered) firm-years
Table 2 continued
What are the consequences of board destaggering? 825
123
826
W. Ge et al.
form of shareholder proposals and votes have helped push through shareholderinitiated governance changes at target firms (Ferri 2012; Ertimur et al. 2010b). Additionally, anecdotal evidence suggests that pressure from shareholder activists has played a significant role in convincing companies to destagger. Proxy advisory firms (e.g., Institutional Shareholder Services) have also been working with institutional investors to destagger boards. The push to destagger seems to coincide with the dramatic increasing trend of shareholder empowerment after the Sarbanes– Oxley Act of 2002 (SOX). After the public disclosures of accounting fraud cases such as Enron, WorldCom, and Tyco International, activist shareholders have increased their demands for the accountability of management and directors, and the effectiveness of such shareholder pressure has also increased. In terms of the economic magnitude, we find that firms with shareholder proposals to destagger are 2.83 % (untabulated) more likely to destagger their boards than other firms. This economic magnitude is quite large, as the mean likelihood for destaggering in our sample is 2.37 %.13 This indicates the effectiveness of shareholder activism.14 We also find that firms with poor prior accounting performance (ROAt-1, ROAt-2) are more likely to destagger their boards. This result is consistent with research suggesting that poor firm performance increases demand for reforming the governance practices within the firm (Finegold et al. 2007). Research has shown that poor prior-year performance is associated with increasing the number of outside directors (Pearce and Zahra 1992) and replacing the CEO (Bhagat et al. 1999). Similarly, we conjecture that poor prior-year performance likely increases the demand for potential changes to the board and positively influences the decision to destagger. Thus we expect firm performance to influence the pressure from shareholders to destagger boards. Firms with poor performance are likely to get more attention from shareholders, and shareholders might blame poor performance on firms’ board structure (Koppes et al. 1999).15 We also find that firms with high entrenchment index (EINDEX) are significantly less likely to destagger their boards, consistent with the notion that incumbent directors of more entrenched boards may want to keep mechanisms (such as poison pills and staggered boards) that prevent them from easily being removed or
13
Within firms with shareholder proposals, we also find that the voting percentage for declassifying the board is significantly and positively associated with the decision to destagger (not tabulated). 14 In addition to shareholder proposals to destagger, we also examine the effect of the total number of shareholder proposals unrelated to destaggering. We find that it is highly correlated with the number of shareholder proposals to destagger (Pearson correlation = 0.302) and significantly associated with the likelihood of board destaggering. It appears that investors who are active in pursuing board destaggering are also active in pursuing other changes. We do not include this variable in the determinant regression due to multicollinearity concerns. 15
Ganor (2008) and Guo et al. (2008) do not find firm performance to be a significant factor in the decision to destagger. The difference in the results might be due to the fact that we use a more recent sample and investor activism is a more recent phenomenon (Ferri 2012).
123
What are the consequences of board destaggering?
827
replaced.16 Finally, we find that firms with more free cash flows (FCF) and larger firms (SIZE) are more likely to destagger their boards. The overall fitness of the determinants model seems reasonable given the likelihood ratio of 531.7. 4.2 Propensity-score matching Next, we generate a matched control sample by matching each destaggering firm to a firm that has always had a staggered board and has the closest predicted probability to destagger in the year before destaggering. The probabilities of destaggering are calculated as the fitted values from estimating Eq. (1). We also require that both destaggering firms and their matched control firms have at least one valid observation before the decision to destagger and at least one valid observation after the year that destaggering has actually occurred (i.e., the year when new board members are elected to 1-year terms).17 In Table 2 Panel B, we compare firm characteristics (i.e., determinant variables) between destaggering firm-year observations and all other firm-year observations before the matching procedure. The results suggest that destaggering firm-years differ systematically from other firm-year observations in the direction similar to the multivariate results reported in Panel A of Table 2. For example, the percentage of shareholder proposals to destagger boards is much higher for destaggering firm-year observations (40.8 %) than for other firm-years (5.3 %). Table 2 Panel C shows that none of the determinant variables differ significantly between destaggering firms and their propensity-score-matched control firms. In addition, the mean predicted probability of destaggering boards is 18.86 % for destaggering firms and 18.27 % for their control firms (not tabulated); the difference is insignificant (p value = 0.75). Overall, the above results suggest that our matched control sample provides high quality matches of the destaggering sample. Finally, in Panel D of Table 2, we examine the short-window stock market reaction to the announcement of destaggering. We hand-collect the following event dates and examine the combined 3-day market-adjusted stock returns around these dates: (1) the day on which the management proposal to destagger the board was 16 For example, since a board controls the implementation of poison pills (which is one component of the entrenchment index) without the necessity of having shareholder approval, the combination of a staggered board and a poison pill creates a veritable fortress for the incumbent board for deterring hostile takeovers. Thus, to the extent that incumbent directors have control over the presence of these mechanisms, we expect them to push back against their removal. 17 Note that, because our consequence analysis requires that all firms have at least one valid observation before and one observation after the decision to destagger, this research design potentially excludes destaggering firms that may have received takeover attempts (if destaggering invites takeovers) within one year. To the extent that being taken over increases shareholders’ value, we may have excluded the observations of which firm value would have increased following destaggering. Thus, we examine whether our sample of destaggering firms were subject to being taken over. Specifically, we use mergers and acquisitions data from the SDC Platinum database and find that 13 destaggering firms and 22 control firms were taken over after the destaggering year. However, these acquisitions were all friendly takeovers; that is, there were no hostile takeovers of our destaggering or control firms. Among these firms, only one destaggering firm and four control firms were acquired through friendly takeovers within one year following the destaggering decision. Therefore the explanation that destaggering invites hostile takeovers that increase value for shareholders is unlikely to be the case for our sample.
123
828
W. Ge et al.
announced, (2) the day on which the destaggering decision was announced in the press, and (3) the release date of the proxy statement containing a proposal to destagger the board. We note that not every firm has all three dates. As reported in Panel D, the mean (median) 3-day value-weighted market-adjusted return is 0.074 % (-0.181 %) and is insignificantly different from zero. The average 5-day value-weighted market-adjusted return is also insignificantly different from zero.18 This result suggests that shareholders do not view board destaggering as beneficial for them. This finding is consistent with the conjecture of Larcker et al. (2011) that investors do not always view destaggering the board as value-enhancing for shareholders.19,20 To the extent that shareholders correctly responded to news related to board destaggering, this result indicates that board destaggering does not lead to an improvement in future firm performance. In addition, we examine whether annual market-adjusted stock returns following destaggering are different between firms that destagger versus those that do not. In untabulated results, we find that, in the 1-year return window following destaggering, destaggering firms have marginally significantly lower stock returns (mean = 1.65 %) than matched control firms (mean = 8.26 %) and the difference is significant at the 10 % level. This result provides preliminary support for the conjecture that destaggering does not benefit shareholders. 4.3 Consequences of the destaggering decision We present our analysis of the consequences of the decision to destagger the board in the following order. First, we examine whether destaggering has performance consequences. Second, we examine whether management entrenchment and the board’s effectiveness in monitoring have changed after board destaggering. Third, to illuminate the effect of destaggering on the horizon over which directors and management make corporate decisions, we investigate how investments change after boards are destaggered for destaggering firms relative to firms that do not destagger. 4.3.1 The effect of destaggering on firm performance We first examine three consequence variables related to financial performance: Tobin’s q (which has been used conventionally in the literature), annual return on assets (ROA), and value-weighted market-adjusted annual returns. Table 3 presents the regression results based on Model (2) described in Sect. 3.2. Columns (1), (3), and (5) exclude all the control variables (except for year and industry fixed effects). 18 For comparison, we also measure the short-window stock returns for the control sample and find that the results are not significantly different from those of the firms that destaggered. 19 Our results differ from those of Guo et al. (2008), who find a positive and significant return on the announcement of firms’ intention to destagger boards when the implementation of the annual election of directors is immediate. This might be due to the sample difference: we use a more recent sample (2002–2010), while their sample period is 1987–2004. 20 Note that Larcker et al. (2011) examine short-window stock returns to corporate governance regulation, and one proposed regulatory change they study involves the Shareholder Bill of Rights Act, which, if passed, would have required destaggered boards across all publicly listed firms.
123
CEOTENUREit
%INDDIRECTit
BOARDSIZEit
DELAWAREit
EINDEXit
ROAit
%INSOWNit
DESTAGGERi 9 AFTERit
AFTERit
DESTAGGERi
Intercept
0.000 (0.898)
0.003 (0.476)
(0.001)
(0.002)
(0.821)
-0.000***
-0.001***
-0.004
(0.056)
(0.001)
-0.004*
-0.007***
(0.158)
(0.684)
(0.161) -0.108
-0.000
-0.039
(0.003)
0.003***
(0.969)
0.000
(0.763)
0.001
(0.564)
-0.010
(0.012)
0.016**
1.168** (0.014)
(0.229)
-0.081
(0.130)
0.032
(0.870)
-0.005
(0.077)
-0.026*
(0.170)
-0.179
(6)
(0.000)
(0.892)
0.003
(0.531)
0.015
(0.800)
0.003
(0.001)
-0.150***
(5)
Returnsit
8.222***
0.003 (0.603)
(0.049)
-0.005**
(0.529)
0.001
(0.114)
0.003
(0.339)
-0.018
(4)
(0.009)
(0.047)
-0.015**
(0.089)
0.016*
(0.049)
0.014**
(0.000)
0.146***
(3)
ROAit
-0.386***
(0.030)
(0.003)
(0.005) -0.211**
(0.022)
-0.341***
0.210***
(0.016)
(0.034)
0.275**
0.216**
0.307**
3.074*** (0.000)
3.239***
(2)
(0.000)
(1)
Tobin Qit
Table 3 The effects of destaggering on firm performance
What are the consequences of board destaggering? 829
123
123
LEVERAGEit
NGSEGit
NBSEGit
AGEit
SIZEit
FCFit
CAPEXit
RDit
INTANGit
PCTOWNit
CEOCHAIRMANit
Table 3 continued
(1)
Tobin Qit
-0.018* (0.059)
-0.288 (0.109)
(0.090)
(0.514) -0.001*
-0.000
0.003 (0.754) 0.094***
(0.674)
(0.084)
(0.001)
-0.000
-0.006*
0.004*** (0.000)
0.023
(0.000)
(0.389) (0.365)
0.944***
-0.965
0.905*** (0.000)
3.441**
(0.140)
(0.000) (0.021)
-0.079
7.989***
0.047*** (0.000)
0.253
(0.745)
(0.071) (0.398)
0.013
-1.237*
0.001 (0.383)
(4)
0.093**
(3)
(0.037)
(2)
ROAit (5)
Returnsit
(0.379)
0.039
(0.314)
0.011
(0.158)
0.004
(0.247)
-0.001
(0.443)
-0.007
(0.512)
0.261
(0.271)
-0.497
(0.188)
0.375
(0.778)
-0.025
(0.137)
0.201
(0.061)
-0.025*
(6)
830 W. Ge et al.
Yes
3892
0.378
N observations
Adj. R squared
Tobin Qit
0.648
3240
Yes
Yes
(2)
0.194
3693
Yes
Yes
(3)
ROAit
0.879
3240
Yes
Yes
(4)
0.126
3898
Yes
Yes
(5)
Returnsit
0.186
3240
Yes
Yes
(6)
This table reports the results of the consequence regressions regarding future firm performance. Dependent variables are specified in column headings. All p values are two-tailed. Standard errors are clustered by industry and year. All the variables are defined in ‘‘Appendix 1’’
Yes
Industry FE
(1)
Year FE
Table 3 continued
What are the consequences of board destaggering? 831
123
832
W. Ge et al.
We include industry fixed effects to control for macroeconomic and other factors affecting the industry that each firm operates in. This particular specification assumes that both the treatment and control firms have been matched well on all the determinants. In these tests, the coefficient estimates on DESTAGGER 9 AFTER are negative and statistically significant for Tobin’s q (coefficient is -0.341, p \ 0.01) and ROA (coefficient = -0.015, p \ 0.05), suggesting that both firm value and accounting performance decrease after the destaggering decision for our sample firms relative to the firm value and accounting performance, respectively, of control firms. We find no statistically significant results for our regression on annual stock returns. Overall, these results are inconsistent with prior research that links staggered boards with reduced firm value (Bebchuk et al. 2002; Bebchuk and Cohen 2005) and the belief that board declassification would lead to improving performance and increasing firm value. We then relax the assumption that the matching produces perfect matches for our treatment firms and add in control variables that may affect performance. In Columns (2), (4), and (6), we include time-varying control variables as well as industry fixed effects and year fixed effects. Consistent with our first set of reported results, the coefficient estimates on DESTAGGER 9 AFTER are negative and statistically significant (p \ 0.05) for our Tobin’s q regressions in Column (2) of Table 3, suggesting that firm value decreases after the destaggering decision for our sample firms relative to the control firms—and these results are unaffected by the inclusion of the additional control variables. Our results for return on assets (ROA) suggest a similar pattern. The coefficient on DESTAGGER 9 AFTER in Column (4) is significantly negative (-0.005, two-tailed p value \0.05). Once more, there are neither economically or statistically significant results for our returns tests.21 Taken together, these findings suggest that destaggering does not result in improved firm value, as argued by the Harvard Law School Shareholder Rights Project; on the contrary, our evidence is more consistent with the view that destaggering eventually leads to reduced firm performance. 4.3.2 The effect of destaggering on board monitoring As discussed in Sect. 2, opponents of staggered boards argue that destaggering would reduce management entrenchment by making directors and CEOs more accountable for firm performance. Therefore we examine how destaggering influences the effectiveness of board monitoring of CEOs. Although monitoring is difficult to measure, following prior literature, we focus on CEO turnover, director turnover, and the sensitivity of CEO compensation to firm performance (Hwang and Kim 2009). Opponents of staggered boards argue that destaggering is associated with higher director turnover and higher CEO turnover in the presence of poor firm performance as well as higher pay-for-performance sensitivity for CEO compensation. 21 We also replicate all our tests (untabulated) where we exclude the first year as well as the first two years following destaggering, to see whether the effects take more time to manifest, and we find similar results to what we report in the paper.
123
What are the consequences of board destaggering?
833
We begin with investigating CEO turnover (an indicator variable equal to one if the CEO from the previous year is no longer the CEO during the current year and the CEO age is below 65) and director turnover (the percentage of directors from the previous year who are no longer on board during the current year). The regression results are reported in Table 4 Panel A. Across all specifications [Columns (1) through (4)], we find no statistically significant results on either management or director turnover. Given our earlier finding that destaggering firms have poorer performance following destaggering relative to the matched control firms, the lack of management turnover and director turnover appears to be inconsistent with the argument made by the opponents of staggered boards. We do note, however, that, even if we find no evidence of increased board turnover, the incentives of the directors (e.g., incentive horizon) may still have changed as a result of destaggering, which would affect firm performance (as examined in Sect. 4.3.1). We examine the potential effects of destaggering on directors’ incentive horizons as reflected in firms’ investment strategies in Sect. 4.3.3. We next estimate the following model to examine the sensitivity of management and director turnover to firm performance—in particular, poor performance. TURNOVERit ¼ b0 þ b1 AFTERit þ b2 LOW ROAi;t;t1 þ b3 LOW RETURNi;t;t1 þ b4 AFTERit LOW ROAi;t;t1 þ b5 AFTERit LOW RETURNi;t;t1 X X X þ bk CONTROLSit þ rp YEARp þ hq INDUSTRYq þ eit ð3Þ We use two indicator variables—LOW_ROAi,t,t-1 and LOW_ RETURNi,t,t-1— which take on values of 1 if the firm-year observation falls below the median ROA and market-adjusted stock returns, respectively, over the preceding 2 years. All the other variables are as described previously and defined in ‘‘Appendix 1’’. Our coefficients of interest are those on the interaction terms between AFTER and each of the performance variables (i.e., b4 and b5), which capture the change in turnoverperformance sensitivity following destaggering. Positive coefficients on the interactions indicate that turnover is associated with poorer performance following destaggering. We estimate Eq. (3) separately for destaggering firms and matched control firms and then perform an F-test for the difference in the coefficient estimates (i.e., b4, and b5) between these two groups. The results for CEO turnover and director turnover are reported in Panels B and C, respectively, of Table 4. The results in Panel B show that, although there seems to be an improvement in CEO turnover-to-performance sensitivity [as evidenced by the positive and significant coefficient on AFTERit 9 LOW_ROAi,t,t-1 in Columns (1) and (3)], these coefficients are not statistically significantly different from the coefficients on the interaction terms for the control firms (p values are above 10 % for both AFTERit 9 LOW_ROAi,t,t-1 and AFTERit 9 LOW_RETURNi,t,t-1). Turning to director turnover (Panel C), it appears that, in the post-destaggering period, director turnover becomes less sensitive to market performance for control firms firms. (b5 is negative and at least marginally significant under Columns 2 and and 4.) The difference in the coefficient estimates between destaggering and control firms
123
834
W. Ge et al.
Table 4 The effects of destaggering on CEO turnover and director turnover Panel A: CEO turnover and director turnover Director turnoverit
CEO turnoverit (1) Intercept
0.079 (0.486)
DESTAGGERi
-0.011***
AFTERit
-0.034
(0.000) (0.228) DESTAGGERi 9 AFTERit
(2)
(3)
0.185**
10.687***
(0.043) -0.003 (0.478) -0.022 (0.506)
0.025
0.015
(0.456)
(0.634)
%INSOWNi,t
-0.014
TOBINQi,t
-0.018
(0.783) (0.365) TOBINQi,t-1
(0.000) -1.007** (0.034) -0.393 (0.755) -0.629 (0.483)
(4) 17.022*** (0.001) -0.927 (0.223) 0.038 (0.976) -0.812 (0.377) -2.218*** (0.002) -0.113 (0.840)
0.035
0.815*
(0.150)
(0.065)
ROAi,t
-0.388 (0.446)
(0.007)
ROAi,t-1
-0.336*
-10.618*
(0.052) RETURNi,t
0.023 (0.544)
RETURNi,t-1
-0.011 (0.590)
EINDEXi,t
0.011 (0.107)
DELAWAREi,t
-0.003 (0.846)
BOARDSIZEi,t
-0.001 (0.778)
%INDDIRECTi,t
-0.000 (0.870)
CEOTENUREi,t
-0.022***
CEOCHAIRMANi,t
-0.125***
(0.000) (0.000) PCTOWNi,t
0.127 (0.366)
INTANGi,t
0.008 (0.870)
123
-34.121***
(0.091) -1.472 (0.257) -0.900* (0.070) 1.223*** (0.000) 0.247 (0.606) -1.033*** (0.000) 0.018 (0.587) -0.117*** (0.000) -0.513 (0.232) -6.487 (0.251) -0.651 (0.594)
What are the consequences of board destaggering?
835
Table 4 continued Panel A: CEO turnover and director turnover Director turnoverit
CEO turnoverit (1)
(2)
RNDi,t
(3)
-0.494
CAPEXi,t FCFi,t SIZEi,t
-23.949***
(0.171)
(0.000)
0.504
22.571*
(0.254)
(0.079)
0.258
26.154**
(0.529)
(0.012)
0.010**
0.829***
(0.016) AGEi,t NBSEGi,t NGSEGi,t LEVERAGEi,t
(4)
(0.001)
0.000
0.029
(0.640)
(0.282)
-0.003*
0.049
(0.087)
(0.689)
0.003
0.115
(0.581)
(0.559)
0.019
-2.339
(0.805)
(0.398)
Industry FE
Yes
Yes
Year FE
Yes
Yes
N observations
3738
3011
3574
3020
Adj. R squared
0.016
0.172
0.049
0.112
Panel B: CEO turnover conditional on performance
Intercept
Destagger firms (1)
Control firms (2)
0.007
-0.103
(0.966) AFTERit LOW_ROAi,t,t-1
0.186
0.413*
(0.385)
(0.096)
0.035
0.013
0.066
0.059
(0.762)
(0.130)
(0.168)
0.026 0.030 (0.121)
AFTERit 9 LOW_ROAi,t,t-1
-0.009
AFTERit 9 LOW_RETURNi,t,t-1
-0.034
(0.796) (0.357) %INSOWNi,t
Control firms (4)
(0.433) (0.237) LOW_RETURNi,t,t-1
(0.596)
Destagger firms (3)
0.073*** (0.001) -0.017 (0.396) -0.063* (0.091) -0.071* (0.058)
0.001 (0.966) 0.037 (0.203) -0.053 (0.148) -0.034 (0.348) -0.010 (0.866)
0.087*** (0.003) -0.047 (0.121) -0.111*** (0.003) -0.075** (0.048) -0.029 (0.603)
123
836
W. Ge et al.
Table 4 continued Panel B: CEO turnover conditional on performance Destagger firms (1) TOBINQi,t
Control firms (2)
Destagger firms (3)
Control firms (4)
-0.007
-0.040
(0.705) TOBINQi,t-1
0.047*** (0.008)
ROAi,t
-0.180 (0.633)
ROAi,t-1
-0.438** (0.016)
RETURNi,t
0.041 (0.256)
RETURNi,t-1
-0.004 (0.898)
EINDEXi,t DLWi,t
%INDDIRECTi,t
0.004
-0.021 (0.445) 0.001
0.000
PCTOWNi,t
-0.057
(0.000) (0.824) -0.116 (0.229) -0.388 (0.334) -0.148 (0.755) -0.049 (0.891) 0.012 (0.171) -0.000 (0.763)
123
(0.201)
(0.870)
(0.000)
AGEi,t
-0.044
(0.893)
-0.127***
SIZEi,t
(0.736)
0.001
CEOCHAIRMANi,t
FCFi,t
(0.551) -0.014
0.013
-0.020***
CAPEXi,t
(0.429) -0.127
(0.200)
CEOTENUREi,t
RDi,t
-0.429
(0.495)
(0.725)
INTANGi,t
0.023 (0.385)
0.007
(0.842) BSIZEi,t
(0.122)
-0.000 (0.579) -0.026*** (0.000) -0.143*** (0.000) 0.070 (0.755) 0.131* (0.077) -0.164 (0.783) 1.331** (0.014) 0.413 (0.418) -0.007 (0.535) -0.000 (0.932)
What are the consequences of board destaggering?
837
Table 4 continued Panel B: CEO turnover conditional on performance Destagger firms (1)
Control firms (2)
NBSEGi,t
Destagger firms (3)
Control firms (4)
-0.002
-0.002
(0.611) NGSEGi,t
(0.696)
-0.004 (0.489)
LEVERAGEi,t
0.015** (0.016)
-0.016
0.055
(0.835)
(0.545) Yes
Industry FE
Yes
Yes
Yes
Year FE
Yes
Yes
Yes
Yes
N Observations
1873
1865
1478
1533
Adj. R squared
0.020
0.046
0.177
0.220
Tests of differences in coefficients between destagger versus control firms AFTERit * LOW_ROAi,t,t-1
v2 = 1.73
v2 = 2.06
p = 0.19
p = 0.15
AFTERit * LOW_RETURNi,t,t-1
v2 = 0.29
v2 = 0.34
p = 0.59
p = 0.56
Panel C: Director turnover conditional on performance Destagger firms (1) Intercept AFTERit LOW_ROAi,t,t-1 LOW_RETURNi,t,t-1
AFTERit 9 LOW_RETURNi,t,t-1
Destagger firms (3)
0.311
5.198
14.128**
(0.958)
(0.257)
(0.032)
0.338
-1.239
-0.777
(0.810)
(0.342)
(0.590)
1.124
0.818
-1.773*
(0.109)
(0.212)
(0.064)
1.021 (0.109)
AFTERit 9 LOW_ROAi,t,t-1
Control firms (2)
2.633*** (0.000)
-0.687 (0.475)
1.637
1.852*
2.225*
(0.156)
(0.097)
(0.068)
0.321
-2.821**
1.550
(0.784)
(0.013)
(0.202)
%INSOWNi,t
-4.029**
TOBINQi,t
-0.647
(0.034) (0.287) TOBINQi,t-1
Control firms (4) 24.144*** (0.004) -0.837 (0.524) -0.362 (0.686) 2.432*** (0.008) 2.499** (0.031) -4.782*** (0.000) -1.265 (0.452) 1.164 (0.141)
0.789
0.777
(0.184)
(0.332)
123
838
W. Ge et al.
Table 4 continued Panel C: Director turnover conditional on performance Destagger firms (1) ROAi,t
Control firms (2)
Destagger firms (3) -30.803** (0.015)
ROAi,t-1 RETURNi,t
0.269 (0.967)
-0.800 -0.027 1.057*** (0.001)
DLWi,t
-0.219 (0.751)
BSIZEi,t
-1.093*** (0.000)
%INDDIRECTi,t
-0.006 (0.788)
CEOTENUREi,t
-0.124** (0.037)
CEOCHAIRMANi,t
-0.896
PCTOWNi,t
-7.746
(0.188) (0.357) INTANGi,t
2.944 (0.358)
RDi,t CAPEXi,t FCFi,t SIZEi,t
LEVERAGEi,t
1.665*** (0.000) 0.695 (0.407) -1.355*** (0.000) 0.032 (0.162) -0.074 (0.218) -1.135* (0.080) -3.216 (0.637) -1.032 (0.648) 4.905
21.342
-2.430
(0.176)
(0.886)
17.596
27.147*
(0.143)
(0.089)
1.339*** -0.017 0.091
0.918*** (0.005) 0.084*** (0.007) -0.024 (0.876)
0.029
0.352*
(0.888)
(0.055)
0.037 (0.989)
123
(0.016)
(0.793)
(0.541) NGSEGi,t
(0.059) -2.514**
(0.000)
(0.549) NBSEGi,t
-2.358*
-50.332***
(0.000) AGEi,t
(0.079)
(0.016)
(0.979) EINDEXi,t
-29.295*
-14.684**
(0.504) RETURNi,t-1
Control firms (4)
-4.296 (0.124)
What are the consequences of board destaggering?
839
Table 4 continued Panel C: Director turnover conditional on performance Destagger firms (1)
Control firms (2)
Destagger firms (3)
Control firms (4)
Industry FE
Yes
Yes
Yes
Yes
Year FE
Yes
Yes
Yes
Yes
N Observations
1781
1793
1481
1539
Adj. R squared
0.045
0.116
0.113
0.210
Tests of differences in coefficients between destagger versus control firms AFTERit * LOW_ROAi,t,t-1
v2 = 0.01 p = 0.93
p = 0.92
AFTERit * LOW_RETURNi,t,t-1
v2 = 3.60
v2 = 14.51
p = 0.06*
v2 = 0.01
p = 0.00***
Panel A reports the results of the consequence regressions regarding CEO and director turnover. Dependent variables are specified in column headings. Panels B and C report the results of consequence regressions regarding CEO and director turnover, respectively, conditional on performance. All p values are two-tailed. Standard errors are clustered by industry and year. All the variables are defined in ‘‘Appendix 1’’
is significant at the 5 % level in our first specification without additional controls (p value = 0.02) and at the 1 % level (p value = 0.01) in our full model. We find no differences between destaggering and control firms in the relative sensitivity to accounting performance. These results provide some evidence that directors’ terms are less affected by performance for firms that continue to have staggered boards. However, the turnover in directors does not seem to result in higher turnover of poorly performing entrenched CEOs, as the proponents of destaggering suggest. Taken all together, our analysis of CEO turnover suggests that there is no statistically significant difference in the rate of CEO turnover after destaggering, even conditional on poor performance. Proponents of destaggering propose that entrenchment will be reduced after a board destaggers, but we do not find any empirical evidence supporting this claim. We next examine the sensitivity of CEO compensation to firm performance, which has been used in prior research as indicating the effectiveness of board monitoring (Hwang and Kim 2009). We expect that, under a regime of more effective board monitoring, CEO compensation is more sensitive to firm performance. We test CEO pay-for-performance sensitivity using the following equation: CEOCOMPit ¼ b0 þ b1 AFTERit þ b2 ROAi;t;t1 þ b3 RETURNi;t;t1 þ b4 AFTERit X ROAi;t;t1 þ b5 AFTERit RETURNi;t;t1 þ bk CONTROLSit X X þ rp YEARp þ hq INDUSTRYq þ eit ð4Þ Similar to Eq. (3), we estimate Eq. (4) separately for destaggering firms and matched control firms. For this specification, we define our dependent variable
123
840
W. Ge et al.
(CEOCOMP) as the logarithm of the sum of salary, bonus, long-term incentive plan payouts, the value of restricted stock grants, the proceeds from options exercised during the year, and any other annual pay for the CEO during the year. The other variables are as described previously. Our coefficients of interest are those on the interaction terms between AFTER and each of the performance variables (i.e., b4 and b5). Generally, we expect that CEO compensation to increase with performance. If destaggering results in improved governance (as argued by the proponents of destaggering), we expect b4 and b5 to be more positive for firms that destagger, relative to our control sample. Decreases in pay-for-performance sensitivity after destaggering relative to the control sample would indicate relatively less effective monitoring after destaggering. Table 5 presents the results. The results for firms that destagger are presented in Columns (1) (3), while the results for control firms are presented in Columns (2) and (4). First, across all columns, CEOCOMP is significantly and positively associated with the performance variables (RETURN and ROA). Next, turning to the interaction terms, the coefficient on AFTER 9 ROA is insignificant for destaggering firms [Column (1)] and significantly positive for control firms [Column (2)]. After including control variables, the coefficient on AFTER 9 ROA is significantly negative (p \ 0.05) for destaggering firms as shown in Column (3), while the coefficient is statistically insignificant for control firms in Column (4). F-tests indicate that the difference in the coefficients on AFTER 9 ROA are indeed significantly different (p value = 0.069) between the two groups, while the coefficients on AFTER 9 RETURN do not differ between the destaggering sample and control group. Overall, our findings suggest that the change in pay-forperformance sensitivity (particularly to accounting earnings) following destaggering is lower for destaggering firms than the control firms, consistent with relatively less effective monitoring after destaggering. Taking the results reported in Tables 4 and 5 together, we do not find any evidence suggesting that board destaggering strengthens the effectiveness of board monitoring, particularly in terms of management turnover-performance sensitivity and pay-for-performance sensitivity. The lack of evidence on any improvement in board monitoring contradicts the argument by the opponents of staggered boards that destaggering reduces management entrenchment. 4.3.3 The effect of destaggering on investment strategies Given that there is no significant effect of board destaggering on management entrenchment, it is unclear why the destaggering decision is negatively associated with firm performance. Therefore we next investigate whether destaggering results in any change in investment strategies. As discussed in Sect. 2, to the extent that destaggering leads directors to focus on short-term performance rather than longterm value creation, we expect to observe a decrease in investments following destaggering for firms that destagger their boards relative to matched control firms. Specifically, we examine research and development expenditures and net expenditures for total investment (ITOTAL), which Richardson (2006) defines as the sum of all outlays on capital expenditures, acquisitions and R&D, less receipts from the
123
What are the consequences of board destaggering?
841
Table 5 The effects of destaggering on CEO pay-for-performance relation Dependent variable: Log(CEOCOMP) Destagger firms (1) Intercept
Control firms (2)
9.327***
8.161***
(0.000) AFTERit
0.197*
RETURNi,t,t-1 AFTERit 9 ROAi,t,t-1
0.388***
0.157
2.931***
AFTERit 9 RETURNi,t,t-1
-0.194
-0.046
(0.000)
0.292***
0.283***
(0.000)
(0.000)
-1.697**
0.364
(0.012)
(0.593)
(0.000)
(0.245)
1.973***
(0.000)
(0.000)
(0.851)
(0.661)
3.074***
(0.002)
0.360***
-0.037
(0.832)
1.476***
(0.000)
(0.000)
0.019
(0.850)
(0.000)
4.636***
(0.000)
-0.019
4.313***
Control firms (4)
3.365***
(0.000)
(0.078) ROAi,t,t-1
Destagger firms (3)
-0.002
(0.785)
-0.009
(0.990)
CEOTENUREi,t
(0.949)
0.014***
0.020***
(0.000) SIZEi,t-1
(0.000)
0.480***
0.429***
(0.000)
(0.000)
0.028
SP500i,t
-0.050
(0.594) BTMi,t-1
(0.351)
-0.861***
-0.598***
(0.000)
(0.000)
Industry FE
Yes
Yes
Yes
Yes
Year FE
Yes
Yes
Yes
Yes
N observations
1806
1792
1805
1790
Adj. R squared
0.323
0.404
0.562
0.563
Tests of differences in coefficients between destagger versus control firms AFTERit 9 ROAi,t,t-1
v2 = 3.69
AFTERit 9 RETURNi,t,t-1
v2 = 0.38
v2 = 0.00
p = 0.54
p = 0.97
p = 0.06*
v2 = 3.06 p = 0.08*
This table reports the results of the consequence regressions regarding CEO pay-performance relation. Dependent variables are specified in column headings. All p values are two-tailed. Standard errors are clustered by industry and year. All the variables are defined in ‘‘Appendix 1’’
sale of property, plant, and equipment.22 We scale both investment variables by average total assets. The regression results are reported in Table 6. The marginally significant and negative coefficients on DESTAGGER 9 AFTER (p value \10 %) in Columns (1) 22 Financial institutions typically have low levels of R&D expenditure. We already include industry fixed effects in our consequence analyses (Eq. 2); nevertheless, we examine the robustness of our results to excluding financial firms from our sample (untabulated). We obtain inferentially similar results.
123
842
W. Ge et al.
Table 6 The effects of destaggering on firm investment ITOTALit
R&Dit (1) Intercept
0.008*** (0.006)
DESTAGGERi
0.006** (0.044)
AFTERit DESTAGGERi 9 AFTERit
-0.005
-0.002 (0.923) 0.005** (0.048) -0.004*
(0.106)
(0.056)
-0.004*
-0.003*
(0.093) %INSOWNi,t
(2)
(0.072) -0.006 (0.220)
TOBINQi,t
0.008*** (0.001)
ROAi,t
-0.127*** (0.009)
RETURNi,t
-0.003** (0.037)
EINDEXi,t
0.000 (0.662)
DELAWAREi,t
-0.000 (0.995)
BOARDSIZEi,t %INDDIRECTi,t CEOTENUREi,t CEOCHAIRMANi,t
CAPEXi,t
0.087*** (0.002)
0.004
0.003
(0.590)
(0.527)
-0.019*** (0.009) -0.003 (0.429)
-0.013* (0.074) -0.000 (0.959) -0.002 (0.849) 0.010*** (0.004) 0.682*** (0.000) -0.011*** (0.002) -0.001 (0.509) 0.017*** (0.000) 0.001
(0.027)
(0.251)
0.000
0.000
(0.337)
(0.885)
0.000
0.000
(0.670)
(0.782)
-0.001 -0.022 (0.353)
INTANGi,t
0.202*** (0.000)
(4)
-0.001**
(0.600) PCTOWNi,t
(3)
0.008** (0.048) -0.052 (0.324)
0.010
0.001
(0.203)
(0.956)
0.050 (0.427)
FCFi,t
0.047 (0.336)
SIZEi,t
-0.000
AGEi,t
-0.000
(0.939) (0.385)
123
-0.779*** (0.000) -0.009*** (0.000) -0.000** (0.022)
What are the consequences of board destaggering?
843
Table 6 continued ITOTALit
R&Dit (1) NBSEGi,t
(2)
(3)
-0.001*
-0.001
(0.089)
(0.405)
-0.000
NGSEGi,t
0.001
(0.420) LEVERAGEi,t
(4)
(0.447)
-0.026***
0.069***
(0.005)
(0.001)
Industry FE
Yes
Yes
Yes
Yes
Year FE
Yes
Yes
Yes
Yes
N observations
3693
3240
3377
2942
Adj. R squared
0.620
0.681
0.316
0.503
This table reports the results of the consequence regressions regarding investment decisions. Dependent variables are specified in column headings. All p values are two-tailed. Standard errors are clustered by industry and by year. All the variables are defined in ‘‘Appendix 1’’
and (2) provide evidence suggesting that destaggering firms have decreased R&D spending following destaggering. These results are consistent with the idea that board destaggering shortens the director incentive horizon and leads to manager short-termism.23 In addition, the coefficients on DESTAGGER 9 AFTER in Columns (3)–(4), where ITOTAL is the dependent variable, are not statistically different from zero. Prior literature has suggested that R&D investment is generally more risky than capital expenditure (e.g., Coles et al. 2006), which might explain why destaggering influences the relative level of R&D investment but not capital expenditure.24 4.4 Cross-sectional variation in the effects of destaggering Following the analyses of the consequences of destaggering, a natural subsequent analysis is to examine the types of firms for which destaggering would either be value-enhancing or value-decreasing. We explore whether the performance consequences of destaggering are influenced by the importance of the board as a governance mechanism. Ahn and Shrestha (2013) suggest that, when firms have more advisory needs, destaggering 23 We expect to find opposite results for firms that switched from a destaggered to a staggered board structure. However, empirically, the frequency of this happening is rather low (i.e., 70 cases between 1991 and 2010). This is consistent with the finding by Cremers et al. (2015) that the frequency of staggering-up was highest during the 1978–1989 period. We have available data for only 19 firms that moved to staggered boards. We find that for these firms, capital expenditures and R&D spending increased after they move to a staggered board structure. 24 In an untabulated robustness test, we examine whether our results are driven by having the financial crisis period in the post-destaggering sample period. We did three additional analyses: (1) excluding board destaggering that occurred during the financial crisis (2008–2009); (2) only using the pre-crisis sample period; (3) excluding financial institutions from our sample. The results of each robustness test are similar to what is reported in this paper, suggesting that the financial crisis does not drive our results.
123
844
W. Ge et al.
might not be beneficial. We use free cash flows (FCF) to proxy for the importance of the board, particularly in terms of the board’s advisory role.25 Then, we also investigate whether the presence of mechanisms that allow for manager entrenchment affects the performance consequences of board destaggering by using the entrenchment index constructed by Bebchuk et al. (2009) (EINDEX). To the extent that firms with lower manager entrenchment have greater advisory (relative to monitoring) needs from the board, we expect that destaggering is associated with more negative future performance when free cash flows are high and managers are less entrenched. Operationally, we create two indicator variables based on the medians of firmlevel average FCF and EINDEX, equal to one when the value is greater than the median value and zero otherwise. We then estimate the following regression for both the firms above and below the median: CONSEQUENCEit ¼ b0 þ b1 DESTAGGERi þ b2 AFTERit þ b3 DESTAGGERi X X AFTERit þ bk CONTROLSit þ rp YEARp X þ hq INDUSTRYq þ eit ð5Þ We use the same control variables as in our consequence analyses in Eq. (2) and Table 4. For parsimony, we focus on Tobin’s q and ROA as consequence variables. We compare the coefficient of interest on the interaction term AFTER 9 DESTAGGER (i.e., b3) for the two different groups. Table 7 Panel A presents the results for Tobin’s q, and Table 7 Panel B presents the results for ROA. Overall, we find that, after destaggering, Tobin’s q decreases for sample firms with higher levels of free cash flows (Column 1) and lower levels of entrenchment (Column 4). In contrast, there are no changes in Tobin’s q for sample firms with lower levels of free cash flows and higher levels of entrenchment (Columns 2–3). The differences in the coefficient estimates of DESTAGGER 9 AFTER between the two groups are significant at the 1 % level for both partitioning variables. Turning to Panel B, it appears that ROA also declines more for sample firms with lower levels of entrenchment, although the difference is not statistically significant. Taken together, to the extent that firms that have higher free cash flows or low entrenchment would benefit more from strong boards’ advisory function, our results are consistent with the expectation that destaggering weakens the advisory function of our sample firms’ boards, resulting in weaker subsequent performance. 4.5 Additional analysis 4.5.1 The effect of destaggering on board characteristics and director compensation The fact that destaggering appears to hurt firm performance, combined with our finding that R&D investment is lower for destaggering firms relative to matched 25
The inferences remain similar when we use intangibles (INTANG) to proxy for board importance.
123
What are the consequences of board destaggering?
845
Table 7 Cross-sectional analysis of destaggering on firm performance: board importance Panel A: Tobin’s q Partitioning variable
FCF High (1)
DESTAGGER
EINDEX Low (2)
0.322***
0.102***
(0.000) AFTER
(0.006)
0.307***
-0.001
(0.000) DESTAGGER 9 AFTER
(0.990)
-0.388***
-0.006
High (3)
-0.038 (0.209) 0.068 (0.252) 0.036
Low (4)
0.692*** (0.000) 0.363*** (0.000) -0.599***
(0.000)
(0.448)
(0.512)
(0.000)
No. of observations
1799
1441
1800
1440
Adj. R squared
0.710
0.541
0.684
0.687
Tests of differences in coefficients for each partitioning variable: v2 = 7.59
DESTAGGER*AFTER
v2 = 11.20
p = 0.01***
p = 0.00***
Panel B: ROA Partitioning variable
FCF High (1)
DESTAGGER AFTER
EINDEX Low (2)
High (3)
0.001
0.003*
0.001
(0.319)
(0.073)
(0.331)
0.001
-0.002
Low (4)
0.007*** (0.000)
-0.001
(0.447)
-0.004**
-0.006**
(0.043)
(0.027)
(0.138)
(0.000)
No. of observations
1730
1510
1800
1440
Adj. R squared
0.888
0.869
0.868
0.915
DESTAGGER 9 AFTER
(0.638)
0.004
(0.503)
-0.004
(0.119) -0.009***
Tests of differences in coefficients for each partitioning variable: DESTAGGER 9 AFTER
v2 = 0.09
v2 = 1.08
p = 0.76
p = 0.30
This table reports the results of the cross-sectional regressions testing how board importance affects the relationship between destaggering and performance, by partitioning the sample into two pairs of subsamples: (1) ‘‘High FCF’’ sub-sample vs. ‘‘Low FCF’’ sub-sample; and (2) ‘‘High EINDEX’’ sub-sample versus ‘‘Low EINDEX’ sub-sample. Dependent variables are Tobin’s q in Panel A and ROA in Panel B, respectively. Control variables have been suppressed. All p values are two-tailed. Year fixed effects are included. Standard errors are clustered by firm. All the variables are defined in ‘‘Appendix 1’’
control firms, suggests that destaggering does not necessarily always benefit shareholders. Our findings with respect to management and director turnover and CEO pay-for-performance sensitivity also suggest that the effectiveness of board monitoring does not appear to improve following destaggering. In this section, to
123
846
W. Ge et al.
provide further evidence on the effect of destaggering on board monitoring, we examine how destaggering influences board characteristics and director compensation. Prior research has shown that board size is negatively associated with the variability of firm performance (Cheng 2008) and the percentage of independent directors is positively associated with the effectiveness of board monitoring (Klein 2002). Therefore, using the directors data in RiskMetrics, we investigate and report our analyses on two board characteristics: board size (the number of directors on the board) and the percentage of independent directors on the board. The regression results based on Model (2) are reported in Table 8. There appears to be no effect of board destaggering on board size or the percentage of independent directors. The coefficient on DESTAGGER 9 AFTER is statistically insignificant across Columns (1)–(4). We also investigate whether director compensation serves as a mechanism that can mitigate the effect of board destaggering on director incentives. We first examine whether directors’ total compensation increases on average to compensate for the increased risk of losing the seat in one year. Then we test the conjecture that directors are rewarded with more equity-based compensation to mitigate potentially shorter horizon concerns resulting from annual elections. As reported in Columns (5) and (6), we do not observe any significant increases in total director compensation in terms of median director compensation within each board.26 In fact, the results reported in Column (6) indicate that there is a marginal decline in destaggering firms’ total compensation relative to matched control firms. The results in Columns (7) and (8) of Table 8 suggest that the proportion of equity-based compensation for destaggering firms does not change significantly after destaggering relative to control firms, inconsistent with the idea that directors are compensated with more equity-based compensation for the increased risk of losing the seat in any 1 year. Overall, we find that after destaggering, board size and the relative proportion of independent directors do not change for destaggered boards. In addition, we find some evidence of a decline in director compensation after destaggering. 4.5.2 New (replacement) director characteristics Although we do not find statistically or economically significant results for director turnover and board size, we select those destaggering firms with directors who have exited the board to examine the characteristics of directors who were elected to the board in the post-destaggering period. Using director-level data from the BoardEx database, we test whether directors elected after destaggering differ from (a) those elected beforehand or (b) those elected by control firms. We find that directors elected after destaggering to be part of destaggered boards do not significantly differ from directors elected before destaggering, nor from those elected in the control sample, in terms of age, years of work experience, concurrent CEO-level experience at other firms, number of directorships, and number of awards/achievements. Therefore it does not appear that our previously documented performance and R&D results are driven by changes in the types of newly elected directors. 26
The results are similar for the compensation of the highest-paid director within each board.
123
RETURNi,t-1
RETURNi,t
ROAi,t-1
ROAi,t
TOBINQi,t-2
TOBINQi,t-1
%INSOWNi,t
DESTAGGERi 9 AFTERit
AFTERit
DESTAGGERi
Intercept
0.198 (0.791)
0.117 (0.393)
(0.616)
(0.246)
(0.907)
-0.529
-9.286
-0.104 0.074
(0.028)
(0.655)
-42.122**
(0.003)
(0.041)
(0.828)
-6.038***
-1.120**
(0.574)
(0.667)
-0.027
-0.486
-0.045
(0.000)
(0.903)
-0.264
(0.000)
(0.875)
-0.328
(0.870)
-0.240
(0.391)
1.384
(0.000)
61.858***
(4)
13.774***
(0.400)
(0.431)
(0.855)
0.297
(0.423)
1.420
(0.000)
53.778***
(3)
% Independent directors
-1.974***
-0.247
-0.212
(0.735)
(0.063)
(0.350)
-0.076
(0.829)
0.442*
0.227
0.063
4.967*** (0.000)
8.135***
(2)
(0.000)
(1)
Board size
Table 8 The effects of destaggering on board characteristics
(0.303)
-13.541
(0.058)
11.850*
(0.946)
-0.355
(0.000)
205.182***
(5)
(0.115)
13.776
(0.855)
-1.988
(0.487)
50.706
(0.378)
-60.676
(0.273)
6.577
(0.148)
8.135
(0.135)
29.205
(0.079)
-19.971*
(0.860)
-0.625
(0.914)
1.164
(0.226)
-94.714
(6)
Director compensation
(0.773)
-0.017
(0.317)
0.053
(0.758)
-0.014
(0.000)
0.457***
(7)
(0.316)
0.026
(0.530)
-0.025
(0.301)
0.231
(0.203)
-1.274
(0.079)
0.081*
(0.086)
-0.070*
(0.438)
0.049
(0.568)
-0.027
(0.361)
0.048
(0.806)
-0.008
(0.644)
-0.075
(8)
% Equity in director comp
What are the consequences of board destaggering? 847
123
123
FCFi,t
CAPEXi,t
RDi,t
INTANGi,t
PCTOWNi,t
CEOCHAIRMANi,t
CEOTENUREi,t
%INDDIRECTi,t
BOARDSIZEi,t
DELAWAREi,t
EINDEXi,t
Table 8 continued
(1)
47.284*** (0.007)
7.796***
(0.009)
(0.000) (0.003)
62.716***
12.201***
(0.287)
(0.011)
(0.123) 23.919
(0.174)
-9.263**
7.722
-0.683
(0.001)
(0.234) (0.123)
4.861** (0.014)
-0.222 -42.493***
(0.099)
-1.970
-0.168*
0.007 (0.607)
(0.896)
10.077
(0.808)
33.819
(0.094)
167.438*
(0.297)
-12.124
(0.006)
236.499***
(0.009)
-12.164***
(0.006)
-1.090***
0.154 (0.630)
(0.048)
(0.265)
(0.047)
-0.013**
(0.949) -1.391
-0.613
(0.937)
-0.316
(6)
(0.186)
(0.029)
(5)
Director compensation
-0.612**
1.582
-0.444**
1.475**
(4)
(0.027)
(3)
0.164*
(2)
% Independent directors
(0.085)
Board size (7)
(0.187)
0.756
(0.200)
1.368
(0.799)
0.132
(0.441)
0.073
(0.834)
0.057
(0.661)
-0.012
(0.841)
0.000
(0.094)
0.001*
(0.008)
0.011***
(0.039)
-0.053**
(0.995)
-0.000
(8)
% Equity in director comp
848 W. Ge et al.
Yes
3898
0.177
N observations
Adj. R squared
0.438
3065
Yes
Yes
Yes
0.318
3898
Yes 0.458
3065
Yes
Yes
5.928 (0.384)
(0.846)
(0.047)
-0.148
-0.570**
(0.067)
(0.371) 0.029
0.297*
0.032
(0.617)
0.140** (0.049)
0.028*** (0.007)
1.022
(4)
(0.232)
(3)
0.748***
(2)
% Independent directors
(0.000)
Board size
0.414
2430
Yes
Yes
(5)
0.526
2097
Yes
Yes
(0.845)
-5.660
(0.527)
-1.423
(0.123)
-1.878
(0.117)
-0.447
(0.000)
33.168***
(6)
Director compensation
0.131
2430
Yes
Yes
(7)
0.227
2097
Yes
Yes
(0.105)
0.107
(0.700)
0.003
(0.254)
0.008
(0.190)
-0.001
(0.021)
0.026**
(8)
% Equity in director comp
This table reports the results of the consequence regressions regarding board characteristics. Dependent variables are specified in column headings. All p values are twotailed. Standard errors are clustered by firm and by year. All the variables are defined in ‘‘Appendix 1’’
Yes
Year FE
(1)
Industry FE
LEVERAGEi,t
NGSEGi,t
NBSEGi,t
AGEi,t
SIZEi,t
Table 8 continued
What are the consequences of board destaggering? 849
123
850
W. Ge et al.
4.5.3 Controlling for firm fixed effects All our reported tests include industry fixed effects to control for factors that affect specific industries in which firms operate. To control for any time-invariant firmspecific factors, we have also run our tests using an alternate specification using firm fixed effects. For our difference-in-differences design, we alternatively include firm and year fixed effects and exclude DESTAGGER from the regression (but keep DESTAGGER 9 AFTER).27 In untabulated results, we find that all the results remain similar to those reported here, except that our results on R&D expenditures, while still negative, have become statistically insignificant. We continue to find that firms that destaggered experience declines in Tobin’s q and ROA, and the negative effect on Tobin’s q is most pronounced in firms with greater advisory needs, consistent with the notion that destaggering results in worse performance when boards are more important in their advisory role. 4.5.4 Shareholder activism and destaggering Our earlier analyses suggest that the destaggering decision is significantly influenced by investor activism (as proxied by having shareholder proposals) and that board destaggering appears to have harmed firm performance. A natural question is why would investors push for board destaggering if it does not benefit them? In this section, we investigate whether investors are rational in pressuring firms to destagger. We partition the sample based on whether the board of each firm received a shareholder proposal to destagger before destaggering to determine whether shareholder pressure boosts or hurts firm performance in the post-destaggering period. If shareholders know that destaggering would be good for the firm, then we would expect the firms that exhibit greater shareholder pressure to have better performance after destaggering than firms without shareholder pressure. On the other hand, if shareholders are ‘‘irrational’’ and blindly push for destaggering, then we would expect to see firms’ performance deteriorate in the post-destaggering period for those firms with shareholder proposals relative to firms that did not receive shareholder proposals. Overall, we find that Tobin’s q decreases for our sample firms relative to control firms regardless of whether these firms received a shareholder proposal to destagger in the pre-destaggering period. However, the magnitude of the decrease in Tobin’s q for firms that received shareholder proposals to destagger is not significantly different from the decrease in Tobin’s q for firms without shareholder proposals. Since the difference in performance between firms with and without shareholder proposals is not statistically significant, we cannot draw any conclusions regarding shareholder activism and its effect on performance. Furthermore, it is reasonable to conjecture that firms that choose to destagger, in spite of not receiving any shareholder proposals, may be pressured to do so from other sources, such as proxy advisory firms and other large shareholder blocks. Shareholder proposals are only one way by which shareholders can exert pressure on the board. We cannot conclude nor test this conjecture, however, and leave it to future research. 27
The variable DESTAGGER drops out as it is perfectly correlated with firm fixed effects.
123
What are the consequences of board destaggering?
851
5 Conclusions This study examines the consequences of board destaggering. We begin by analyzing determinants of destaggering; we find that the likelihood of destaggering increases with shareholder activism, firm size, and poor prior accounting performance. Furthermore, we find that firms that destagger tend to have larger boards and are less likely to be entrenched. Using the determinants analysis combined with propensity-score matching, we construct a matched control sample of firms that opted to continue having a staggered board structure and use a difference-in-differences research design to examine the consequences of destaggering. We first examine performance consequences of the destaggering decision and find that both Tobin’s q and accounting performance, measured by ROA, are lower after destaggering. We then investigate whether, as the proponents of destaggering suggest, CEO entrenchment is reduced after boards destagger and find little evidence of this. Destaggering does not appear to lead to an overall increase in CEO and director turnover. Conditional on performance, director turnover increases with poor performance; however, this effect does not carry over to CEO turnover. We also show that CEO pay-forperformance sensitivity declines after destaggering, suggesting that monitoring quality does not improve post-destaggering. Overall, the lack of evidence seems to run contrary to the conjectures of destaggering proponents. Next, we try to determine whether boards behave differently after destaggering. We find evidence that investment in R&D decreases after the decision to destagger the board, consistent with directors having shorter horizons and therefore investing less in long-term projects such as R&D. We then examine how destaggering affects board structure. We do not find any significant board structure changes after the destaggering decision. Supplemental cross-sectional analyses show that destaggering firms that would likely benefit from strong advisory boards exhibit significant declines in Tobin’s q, while this effect is not present among firms with fewer advisory needs. This evidence suggests that destaggering may be more value-destroying for some firms than others. The evidence provided in this study suggests that destaggering hurts shareholders. Furthermore, our cross-sectional results are consistent with the conjecture that destaggering is not necessarily the one-size-fits-all board structure that is being endorsed by proxy advisory firms. These findings contradict some of the earlier cross-sectional studies on destaggered boards and support the view held by proponents of keeping the previously more popular staggered board structure. The study is timely, given the wave of destaggering over the past decade. We note, however, that our results may be limited in terms of generalizability to all destaggered boards. In our research design, we focus only on firms that switched to a destaggered board from a staggered board. Thus our results may not be generalizable to firms that have always had a destaggered board. We leave it to future research to examine firms that recently have gone public with either staggered or destaggered board structures.
123
852
W. Ge et al.
Acknowledgments We thank Ge Bai, Xia Chen, Mei Feng, Allison Germain, Russ Lundholm, Xiumin Martin, Ryan McDonough, Sarah McVay, Jed Neilson, Gaizka Ormazabal (discussant), Lorenzo Patelli, Ray Pfeiffer, Phil Shane, Terry Shevlin, Lakshmanan Shivakumar (editor), D. Shores, Mary Stanford, Christina Synn, Jake Thornock, our anonymous reviewer, workshop participants at Bentley University, College of William and Mary, Northeastern University, Texas Christian University, University of Toronto, University of Washington, Virginia Tech, and Washington University in St. Louis, and participants at the 2016 RAST Conference at London Business School, 2013 HKUST Accounting Symposium, the 2014 AAA Annual Meeting and the 2014 AAA MAS Midyear Meeting for helpful comments and suggestions. We thank Allison Germain for research assistance. Earlier versions of this paper were circulated under the title ‘‘Board Destaggering: Corporate Governance Out of Focus?’’ Weili Ge thanks the Moss Adams Professorship for financial support. Jenny Li Zhang acknowledges the financial support from the Social Sciences and Humanities Research Council of Canada (SSHRC).
Appendix 1: Variable definitions
Variable
Definition
Determinants of the destaggering decision DESTAGGER
An indicator variable for firms which destaggered classified boards during 2002–2010 sample period
Shareholder activism PROPOSAL
An indicator variable for shareholder proposal to destagger; equal to one if there is a shareholder proposal to de-stagger in the previous two fiscal year and zero otherwise
%INSOWN
Percentage of institutional ownership
Firm performance TOBINQ
The ratio of the market value of assets to book value of total assets The market value of assets is obtained as total assets - common equity - deferred taxes ? market value of equity: (PRCC_F 9 CSHO ? AT-CEQTXDB)/AT
ROA
Return on assets, defined as the ratio of income before extraordinary items to beginning total assets: IB/LAG(AT)
RETURN
Value-weighted market-adjusted annual returns
Governance attributes EINDEX
Entrenchment index constructed following Bebchuk et al. (2009). It is the number of the following provisions: staggered board, limitation on amending bylaws, limitation on amending the charter, supermajority to approve a merger, golden parachute, and poison pill. We took out staggered board from the e-index, so the e-index ranges from 0 to 5. We use the RiskMetrics database to compile the entrenchment index
DELAWARE
Indicator variable equal to one if the firm is incorporated in Delaware and zero otherwise
BOARDSIZE
The number of directors on the board for the year
%INDDIRECT
Percentage of independent board of directors
CEOTENURE
The logarithm of the number of years the manager has been in the CEO position
CEOCHAIRMAN
An indicator variable if the CEO is concurrently the board chairman
PCTOWN
The percentage of shares held by the top five executives at the end of year
123
What are the consequences of board destaggering?
Variable
853
Definition
Monitoring and advisory roles of the board INTANG
Intangibility, calculated as one minus the ratio of net property, plant and equipment, to total assets: 1-PPENT/AT
RD
Research and development expenditure, deflated by average total assets. It is set to zero when this variable is missing
CAPEX
Capital expenditure deflated by average total assets
FCF
Free cash flow to assets, calculated as the ratio of free cash flow to total assets. Free cash flow is defined as the sum of net income and depreciation minus capital expenditure: (IB ? DP-CAPX)/AT
SIZE
The logarithm of the total assets at the end of the fiscal year: log(AT)
AGE
Firm age, measured as the number of years since a firm’s first appearance in the CRSP monthly stock return files
NBSEG
The logarithm of the number of business segments from the Compustat segment files at the end of a fiscal year
NGSEG
The logarithm of the number of geographic segments from the Compustat segment files at the end of a fiscal year
LEVERAGE
Leverage ratio, measured as total debt (the sum of debt in current liabilities and longterm debt) to total assets:(DLC ? DLTT)/AT
Consequence and additional control variables AFTER
An indicator variable for years after the year of destaggering
CEO_TURN
An indicator variable for CEO turnover for the year
DIR_TURN
The percentage of board members from the previous year who leave the board
CEOCOMP
Total payout to CEO. It is the sum of salary, bonus, long-term incentive plan payouts, the value of restricted stock grants, the proceeds from options exercised during the year, and any other annual pay for the CEO in the year
DIRCOMP
The median of total annual compensation for the board members
ITOTAL
Total investment, defined as the sum of all outlays on capital expenditure, acquisitions, and research and development less receipts from the sale of property, plant, and equipment, deflated by total assets
BTM
Book-to-market ratio, calculated as the book value of assets deflated by the sum of book value of liabilities and market value of equity at the end of the year
SP500
An indicator variable for S&P 500. It is set to one if the firm is in the S&P 500 at the end of year and zero otherwise
Appendix 2: Determinants of the destaggering decision We examine the determinants of the decision of the board to opt for a destaggered structure to find control firms for our difference-in-differences analysis. Based on the literature and anecdotal evidence, we posit that several factors influence the destaggering decision, which we discuss below. Shareholder activism We expect that firms under more pressure from activist shareholders are more likely to destagger their boards. In the years after the Sarbanes–Oxley Act of 2002 (SOX), evidence has surfaced suggesting that low-cost shareholder activism in the form of
123
854
W. Ge et al.
shareholder proposals and shareholder votes have helped push through shareholderinitiated governance changes at target firms (Ertimur et al. 2010a; Ferri 2012). Moreover, in the past decade, there has also been a remarkable increase in monitoring from activist hedge funds. Brav et al. (2008) have documented that hedge funds can influence management and board decisions. Anecdotal evidence also suggests that pressure from shareholder rights activists has played a significant role in convincing companies to destagger. Proxy advisory firms (e.g., Institutional Shareholder Services) have also been working with institutional investors to destagger boards. The push to destagger boards seems to coincide with the dramatic upward trend of shareholder empowerment after SOX. After the public disclosures of accounting fraud cases such as Enron, WorldCom, and Tyco International, activist shareholders have increased their demand for accountability of management and directors, and the effectiveness of such shareholder pressure has also improved. Therefore, to the extent that activist shareholders favor destaggering boards, we expect to observe a positive association between firm-specific shareholder activism level and the likelihood of destaggering boards. As discussed above, low cost shareholder activism in the form of shareholder proposals has become more effective in pushing through governance changes following SOX. Thus our first measure for shareholder activism is PROPOSAL, an indicator variable equal to one if there is at least one shareholder proposal to destagger boards in the previous two fiscal years and zero otherwise. Our second measure is the percentage of institutional ownership obtained from the Thomson Financial database (%INSOWN). Crane et al. (2014) have shown that an increase in institutional ownership is associated with a significant increase in proxy voting. Given their investment, institutional investors probably have strong incentives to actively monitor the firms they invest in. Guo et al. (2014) also find that the pressure from activist shareholders is associated with a higher likelihood and speedier process of destaggering. Thus we expect both PROPOSAL and %INSOWN to be positively associated with the destaggering decision. Firm performance Finegold et al. (2007) suggest that poor firm performance (in terms of accounting and stock market measures), which itself is a function of current corporate governance mechanisms, increases demand for reforming the governance practices within the firm. Indeed, research has shown that poor prior year performance is associated with increasing the number of outside directors (Pearce and Zahra 1992) and replacing the CEO (Bhagat et al. 1999). Similarly, we conjecture that poor prior year performance likely increases the demand for potential changes to the board and positively influences the decision to destagger. Thus we expect firm performance to influence the pressure from shareholders to destagger boards. Firms with poor performance are likely to get more attention from shareholders, and shareholders might blame poor performance on firms’ board structure (Koppes et al. 1999).
123
What are the consequences of board destaggering?
855
Governance attributes In terms of governance mechanisms, we believe that firms that have stronger board and management protection, and thus a higher likelihood of management and director entrenchment, face two countervailing forces regarding destaggering. On one hand, some boards are protected by mechanisms that allow directors (and managers) to be entrenched, such as poison pills and insider directors, and these mechanisms are enhanced by a staggered board structure. These boards are likely to be pushed by investors to destagger. This is exacerbated by the increasingly urgent call by several organizations, including the Harvard Law School Shareholder Rights Project and proxy advisor firms (Davidoff 2012), who claim that boards should be destaggered. When shareholders are concerned about poor corporate governance practices, the credibility of management, or the transparency of corporate disclosures, managers are likely to face greater pressure to destagger boards. On the other hand, however, incumbent directors may want to keep mechanisms (such as poison pills and staggered boards) that prevent them from easily being removed or replaced. For example, since a board controls the implementation of poison pills without the necessity of having shareholder approval, the combination of a staggered board and a poison pill creates a veritable fortress for the incumbent board. Thus, to the extent that incumbent directors have control over the presence of these mechanisms, we expect directors to push back on their removal. Overall, it is not clear which forces are stronger between the demand to destagger versus the reluctance of boards to do so. We therefore do not make any directional predictions on how governance attributes influence the choice to destagger. Importance of board’s role in governance: monitoring and advisory roles of the board Boards have both a monitoring and an advisory function when dealing with managers.28 Ahn and Shrestha (2013) argue that the decision to opt for a destaggered board is contingent on the tradeoff between two general roles that the board takes on. On one hand, there are firms whose shareholders benefit from high monitoring from the board, such as firms with opaque operations or complex structures (Harris and Raviv 2008; Hermalin and Weisbach 1998; Raheja 2005). In these cases, any mechanism that allows for managerial entrenchment, such as staggered boards, increases the ability of managers to reap private benefits and thus potentially increases monitoring costs. Therefore firms requiring more monitoring from the board are likely to adopt a destaggered board structure to reduce any entrenchment effect that increases monitoring costs. On the other hand, there are firms that have greater advisory needs from the board, such as firms with higher free cash flows, more intangible assets, and substantial R&D projects that may involve multiple years to implement and see payoffs (Boone et al. 2007; Coles et al. 2008). In these particular firms, a staggered 28 The management literature analogously refers to the monitoring and advisory functions as the control and service roles, respectively, of the board. See Johnson et al. (1996) for a review.
123
856
W. Ge et al.
structure promotes continuity within the board, strengthening the ability of the board to provide advisory services. We contend that a staggered board, which is more stable, can better help in the implementation of long-term strategies than a destaggered board. We therefore expect that firms that would benefit from a strong advisory board are less likely to adopt a destaggered board structure. However, firms do not necessarily have to trade off the board’s monitoring function and its advisory role; a firm in need of more monitoring (which therefore would benefit from having a destaggered board) might also have strong advisory needs (and therefore would also benefit from having a staggered board). Empirically, it is very difficult to separate one effect from the other. In addition, Kim et al. (2014) find that outside director tenure is positively associated with both advising performance (acquisition and investments) and monitoring performance (CEO compensation); to the extent that a staggered board is associated with longer director tenure, even firms with strong monitoring needs might not always desire a destaggered board. Thus, rather than examining the monitoring versus advisory tradeoff, we examine how the importance of the board’s role (both as advisor to and monitor of management) impacts destaggering.
References Adams, R., Hermalin, B. E., & Weisbach, M. S. (2010). The role of boards of directors in corporate governance: A conceptual framework and survey. Journal of Economic Literature, 48(1), 58–107. Ahn, S., & Shrestha, K. (2013). The differential effects of classified boards on firm value. Journal of Banking & Finance, 37(11), 3993–4013. Bebchuk, L. A., & Cohen, A. (2005). The costs of entrenched boards. Journal of Financial Economics, 78, 409–433. Bebchuk, L. A., Coates, J. C, I. V., & Subramanian, G. (2002). The powerful antitakeover force of staggered boards. Stanford Law Review, 54, 887–951. Bebchuk, L. A., Cohen, A., & Ferrell, A. (2009). What matters in corporate governance? Review of Financial Studies, 22(2), 783–827. Bhagat, S., Carey, D., & Elson, C. (1999). Director ownership, corporate performance, and management turnover. Business Lawyer, 54(3), 885–919. Boone, A., Field, L., Karpoff, J., & Raheja, C. (2007). The determinants of corporate board size and composition: An empirical analysis. Journal of Financial Economics, 85, 66–101. Brav, A., Jiang, W., Partnoy, F., & Thomas, R. (2008). Hedge fund activism, corporate governance, and firm performance. Journal of Finance, 63(4), 1729–1775. Bushee, B. (1998). The influence of institutional investors on myopic R&D investment behavior’’. The Accounting Review, 73(3), 305–333. Chemmanur, T., & Tian, X. (2013). Do anti-takeover provisions spur corporate innovation. Working paper, Boston College and Indiana University. Cheng, S. (2008). Board size and the variability of corporate performance. Journal of Financial Economics, 87(1), 157–176. Cohen, A., & Wang, C. Y. (2013). How do staggered boards affect shareholder value? Evidence from a natural experiment. Journal of Financial Economics, 110, 627–641. Coles, J., Daniel, N., & Naveen, L. (2006). Managerial incentives and risk-taking. Journal of Financial Economics, 79, 431–468. Coles, J., Daniel, N., & Naveen, L. (2008). Boards, does one size fit all? Journal of Financial Economics, 87, 329–356. Crane, A. D., Michenaud, S., & Weston, J. P. (2014). The effect of institutional ownership on payout policy: Evidence from index thresholds. Working Paper: Rice University.
123
What are the consequences of board destaggering?
857
Cremers, M., Litov, L., & Sepe, S. (2015). Staggered boards and firm value, revisited. Working paper, University of Notre Dame. Davidoff, S. M. (2012). The case against staggered boards. New York Times website. 3/20/2012. http:// dealbook.nytimes.com/2012/03/20/the-case-against-staggered-boards/. Dechow, P. M., & Sloan, R. G. (1991). Executive incentives and the horizon problem. Journal of Accounting and Economics, 14, 51–89. Ertimur, Y., Ferri, F., & Muslu, V. (2010a). Shareholder activism and CEO pay. Review of Financial Studies, 24(2), 535–592. Ertimur, Y., Ferri, F., & Stubben, S. R. (2010b). Board of directors’ responsiveness to shareholders: Evidence from shareholder proposals. Journal of Corporate Finance, 16, 53–72. Faleye, O. (2007). Classified boards, firm value, and managerial entrenchment. Journal of Financial Economics, 83, 501–529. Ferri, F. (2012). ‘Low-cost’ shareholder activism: A review of the evidence. In C. A. Hill & B. H. McDonnell (Eds.), Research Handbook on the Economics of Corporate Law. Cheltenham: Edward Elgar Publishing. Finegold, D., Benson, G., & Hecht, D. (2007). Corporate boards and company performance: Review of research in light of company reforms. Corporate Governance: An International Review, 15(5), 865–878. Frakes, M. (2007). Classified boards and firm value. Delaware Journal of Corporate Law, 32(1), 113–158. Ganor, M. (2008). Why do managers dismantle staggered boards? Delaware Journal of Corporate Law, 33, 149–198. Grundfest, J. (2004). Advice and consent: An alternative mechanism for shareholder participation in the nomination and election of corporate directors. In L. Bebchuk (Ed.), Shareholder Access to the Corporate Ballot. Boston, MA: Harvard University Press. Guo, L., & Masulis, R. W. (2015). Board structure and monitoring: New evidence from CEO turnovers. Review of Financial Studies, 28(10), 2770–2811. Guo, R., Kruse, T. A., & Nohel, T. (2008). Undoing the powerful anti-takeover force of staggered boards. Journal of Corporate Finance, 14, 274–288. Guo, R., Kruse, T. A., & Nohel, T. (2014). Activism and the shift to annual director elections. Journal of Accounting and Finance, 14, 83–103. Harris, M., & Raviv, A. (2008). A theory of board control and size. Review of Financial Studies, 21, 1797–1832. Hermalin, B., & Weisbach, M. (1998). Endogenously chosen board of directors and their monitoring of the CEO. American Economic Review, 88, 96–118. Hwang, B., & Kim, S. (2009). It pays to have friends. Journal of Financial Economics, 93, 138–158. Johnson, J. L., Daily, C. M., & Ellstrand, A. E. (1996). Boards and directors: A review and research agenda. Journal of Management, 22(3), 409–438. Kim, K., Mauldin, E., & Patro, S. (2014). Outside directors and board advising and monitoring performance. Journal of Accounting and Economics, 57, 110–131. Klein, A. (2002). Audit committee, board of director characteristics and earnings management. Journal of Accounting and Economics, 33, 375–400. Koppes, R. H., Ganske, L. G., & Haag, C. T. (1999). Corporate governance out of focus: The debate over classified boards. The Business Lawyer, 54, 1023–1055. Larcker, D. F., Ormazabal, G., & Taylor, D. J. (2011). The market reaction to corporate governance regulation. Journal of Financial Economics, 101(2), 431–448. Larcker, D. F., McCall, A. L., & Ormazabal, G. (2015). Outsourcing shareholder voting to proxy advisory firms. Journal of Law and Economics, 58, 173–204. Lemmon, M., & Roberts, M. R. (2010). The response of corporate financing and investment to changes in the supply of credit. Journal of Financial and Quantitative Analysis, 45(3), 555–587. Lipton, M., & Savitt, W. (2007). The many myths of Lucian Bebchuk. Virginia Law Review, 93, 733–758. Masulis, R. W., Wang, C., & Xie, F. (2007). Corporate governance and acquirer returns. Journal of Finance, 62, 1851–1889. Pearce, J., & Zahra, S. (1992). Board composition from a strategic contingency perspective. Journal of Management Studies, 29(4), 411–438. Raheja, C. (2005). Determinants of board size and composition: A theory of corporate boards. Journal of Financial and Quantitative Analysis, 40(2), 283–306. Richardson, S. (2006). Overinvestment of free cash flows. Review of Accounting Studies, 11, 159–189.
123
858
W. Ge et al.
Subramanian, G. (2015). Corporate governance 2.0. Harvard Business Review, 93, 96–105. Tonello, M., & Aguilar, M. (2012). Shareholder proposals: Trends from recent proxy seasons (2007– 2011). Working paper, The Conference Board, Inc.
123