Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Three essays on board structure
(USC Thesis Other)
Three essays on board structure
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THREE ESSAYS ON BOARD STRUCTURE
by
Derek Horstmeyer
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)
August 2012
Copyright 2012 Derek Horstmeyer
ii
Acknowledgements
I would like to thank my dissertation chairs, Kevin Murphy and Oguzhan Ozbas, for their
guidance and encouragement throughout my PhD. I deeply appreciate their time and patience
with me throughout the entire process. I also thank my dissertation committee members: Guofu
Tan, John Matsusaka, and Pedro Matos, for offering valuable insights throughout the entire
process.
I thank Jerchern Lin, Breno Schmidt, Kara Wells, John Zhu, Fernando Zapatero, Wayne
Ferson, Chris Jones, Tom Chang, Harry DeAngelo and seminar participants in the 2010 FMA
Annual Meeting, University of Southern California, Simon Fraser University, and George Mason
University for helpful comments.
I would also like to thank my classmates who have made my years at USC a great
experience and my family for their support throughout my academic studies.
iii
Table of Contents
Acknowledgements .......................................................................................................................... ii
List of Tables ................................................................................................................................... v
Abstract .......................................................................................................................................... vii
Chapter 1: Beyond Independence: CEO Influence and the Internal Operations of the Board ...... 1
1.1 Introduction ........................................................................................................................ 1
1.2 Background Information, Data, and Summary Statistics .................................................... 9
1.2.1 Post-SOX Regulatory Environment ......................................................................... 9
1.2.2 Background Literature and Committees ................................................................ 11
1.2.3 Development of Cross-sectional Hypotheses.......................................................... 13
1.2.4 Data Construction and Variable Specification ....................................................... 17
1.2.5 Specification of Primary Measures ........................................................................ 23
1.2.6 Summary Statistics .................................................................................................. 25
1.3 Empirical Design .............................................................................................................. 31
1.3.1 Cross-Sectional Determinants of the Operational Form of the Board .................... 31
1.3.2 Determinants of Internal Monitoring Structure....................................................... 36
1.3.3 Determinants of Investment/Policy Control ........................................................... 44
1.3.4 Changes in the Internal Operations of the Board .................................................... 48
1.4 Conclusion ....................................................................................................................... 59
Chapter 2: Monitoring the Monitors ............................................................................................. 63
2.1 Introduction ...................................................................................................................... 63
2.2 Data, Variable Specification and Summary Statistics ...................................................... 68
2.2.1 Regulatory Environment and Data Construction ................................................... 68
2.2.2 Summary Statistics ................................................................................................. 72
2.3 Empirical Design ............................................................................................................. 77
2.3.1 Outside Director Turnover and Nominating Committee Structure ........................ 78
2.3.2 Board Decisions and Committee Structure ............................................................ 90
2.4 Conclusion ....................................................................................................................... 97
Chapter 3: The Internal Governance of the Board ...................................................................... 99
3.1 Introduction ..................................................................................................................... 99
3.2 Role of the Nominating Committee & Development of Hypotheses.............................. 103
3.2.1 Regulatory Environment and Background Literature ........................................... 103
iv
3.2.2 Nominating Committee Formation ....................................................................... 106
3.3 Empirics - Data, Variable Specification and Summary Statistics ................................... 109
3.3.1 Data Set Construction and Variable Specification ................................................ 110
3.3.2 Summary Statistics ................................................................................................ 112
3.4 Empirical Design ............................................................................................................ 117
3.4.1 Director Characteristics and the Formation of the Nominating Committee ......... 117
3.4.2 Determinants of Nominating Committee Structure at the Firm Level .................. 125
3.4.3 Director Compensation and Nominating Committee ............................................ 132
3.5 Conclusion ...................................................................................................................... 139
Bibliography ................................................................................................................................ 141
v
List of Tables
Table 1.1: Summary Statistics .................................................................................................... 27
Table 1.2: Operational Form and CEO Power ............................................................................ 33
Table 1.3: Determinants of Internal Monitoring Structure ......................................................... 39
Table 1.4: Determinants of Internal Investment Control ............................................................ 46
Table 1.5: Changes in Board Structure Over Time ..................................................................... 51
Table 1.6: Changes in Firm, CEO and Board Characteristics ..................................................... 56
Table 1.7: Compliant v. Non-Compliant Firms .......................................................................... 58
Table 2.1: Summary Statistics .................................................................................................... 73
Table 2.2: Director Turnover and Firm-Level Committee Structure .......................................... 80
Table 2.3: Compensation Director Turnover and Firm-Level Committee Structure .................. 86
Table 2.4: Audit Director Turnover and Firm-Level Committee Structure ................................ 89
Table 2.5: CEO Compensation and Board Committee Structure ............................................... 91
Table 2.6: Probability of Restatement and Board Committee Structure ..................................... 93
Table 2.7: Changes in Corporate Governance and Board Committee Structure ........................ 95
Table 3.1: Summary Statistics .................................................................................................. 114
Table 3.2: Univariate Analysis of Nominating Committee Positions. ....................................... 120
Table 3.3: Determinants of Nominating Committee Holdings ................................................. 124
Table 3.4: Determinants of Nominating Committee Structure at the Firm-Level .................... 126
vi
Table 3.5: Matched Sample Differences in Director Compensation ........................................ 134
Table 3.6: Director Compensation and Firm-Level Nominating Committee Holdings ............ 137
vii
Abstract
This dissertation is comprised of three essays.
The first essay, titled “Beyond Independence: CEO Influence and the Internal Operations
of the Board”, serves to extend our understanding of board control beyond traditional measures of
independence. Using a detailed, hand-collected dataset on board sub-structure, I document that
the board of directors for the average firm underwent a significant transformation between 1999
and 2005, not in terms of its size or composition (independence), but in terms of the CEO’s
participation in the internal decision-making processes of the board. Over this time period, (i) the
CEO presided over (had a voting stake in) fewer board meetings; (ii) the fraction of meetings held
in independent (outside director controlled) monitoring committees increased over 80 percent;
and (iii) the fraction of board-time spent by the CEO in the executive committee decreased 40
percent. Together, these findings suggest that the principal governance reform following the
corporate scandals of 2000-2002 was through an alternative channel of ‘independence’ on the
board - the reduction in the CEO’s ability to influence board oversight and control the internal
operations of the board. Next, consistent with the view that the drastic shift in board operations
between 1999 and 2005 was contrary to the preferences of the CEO, I find that CEO power is
negatively associated with the fraction of meetings held outside of the CEO’s presence in
independent monitoring committees, and positively associated with the fraction of meetings held
in the executive committee. In total, the results extend the literature on board structure by
viii
highlighting that the governance standards of the modern board are not only defined by the
composition and size of the board, but also by its internal operating form.
In the second essay, titled “Monitoring the Monitors”, I address whether the sub-structure
of the board has a material impact on the policies and internal governance of the board. Namely, I
investigate how the size of the nominating committee and the cross-membership of directors
between the nominating committee and other monitoring committees affect the effort level of
directors and the self-assessment of the board. Through an examination of compensation, audit,
and governance decisions, large nominating committees are at best weakly associated with lower
effort/quality of decisions. More significantly, both the size of the nominating committee and the
cross-membership between the nominating committee and other committees are negatively
related to outside director turnover. This finding highlights the self-evaluation problem faced by
large nominating committees and the benefits of such a board feature to directors - outside
directors primarily structure large nominating committees to secure their board positions and
insulate themselves from turnover.
In the third essay, title “The Internal Governance of the Board”, I detail how the
formation and composition of the nominating/governance committee relate to board member
characteristics and firm determinants in a manner that is in accordance with the contracting
relationship between shareholders and board members. For a sample of NYSE firms over the
2005 to 2009 period, I find that directors who are in high demand and directors who have
performed well are more likely to hold nominating committee positions. In addition, firms
associated with higher risk levels, and boards structured with a greater number of these high
demand and strong performing directors create larger nominating committees to attract and retain
their directors’ services. The results serve to address the apparent conflict between the prevalence
of large nominating committees and the insulating nature of such a board feature.
1
Chapter 1
Beyond Independence: CEO Influence
and the Internal Operations of the Board
1.1 Introduction
Beginning in the early 1980s, shareholder advocacy groups, academics, and institutions devoted
to corporate governance issues started calling for increased outside director representation on
U.S. boards.
1
Accompanying these demands for board reform, the composition of the average
board changed in a material manner over the latter half of the 20th century, with the fraction of
outside directors serving on the board roughly increasing from 50% to 80%.
2
Moreover, by the
time that NASDAQ/NYSE first formally imposed board independence requirements in 1999 for
firms listed on their exchanges, the vast majority of firms already had ‘outsider-dominated’
boards, and further, many firms had the CEO sitting as the sole inside director on the board.
Yet, despite this level of board independence, a series of corporate malfeasance,
accounting, and backdating scandals occurred at the turn of the 21st century.
3
Thus, if these
1
See the American Law Institute (1982) and The Business Roundtable (1997) for evidence pertaining to governance
advocacy groups demands for greater board independence. For academic research on the issue of board independence,
see Weisbach (1988) for evidence on outside directors’ ability to remove poor performing CEOs, and Fama and Jensen
(1983) for additional supporting conjectures on the value of independent boards.
2
Lehn et al. (2005) study the evolution of 81 firms over time and note that independence increased from 50% to 83%
during the second half of the 20th century. See Coles et al. (2008) for additional supporting evidence.
3
See Heron and Lie (2007) for a time-line of option backdating events to hit U.S. public firms.
2
corporate failures were at least partially attributable to lax oversight on the part of the board, this
suggests that the mere presence of many outside directors on the board might not be sufficient. If
CEOs have the ability to alter the monitoring operations of the board through alternative
channels, such as presiding over the decisions made on the board, or implementing policies
removed from the oversight of the full board (via the executive committee), then the high level of
board independence witnessed at the time of the scandals does not necessarily imply that CEOs
had forfeited their control/influence over board proceedings. Nor does it mean that the
overwhelming number of ‘outsider-dominated’ boards which existed at this point in time were, in
fact, ‘dominated’ by outside directors. Therefore, to have a clearer picture of both control within
the boardroom, and how CEO influence over the board has truly changed over time, a deeper and
more robust understanding of the internal workings of the board seems imperative.
In this paper, I explore just such an alternative dimension of ‘independence’ on the board:
the extent to which outside directors are able to handle their board responsibilities removed from
the influence of the CEO, or conversely, the CEO’s ability to control the internal monitoring and
investment decision-making processes of the board. Using a detailed, hand-collected dataset on
board committee and meeting structure, I proxy for outside directors’ control over board
proceedings using the fraction of meetings which outside directors hold in independent
monitoring committees (audit, compensation, nominating) removed from the CEO’s voting
influence. This measure, often noted as ‘the fraction of meetings held in independent monitoring
committees’ or ‘the fraction of board work controlled by outside directors’, is constructed as the
ratio of the number of meetings which outside directors hold in independent committees to the
number of meetings which the CEO presides over (has a voting stake in).
4
Implicit in the
4
Specifically, this primary measure of the operational control on the board is constructed as the number of meetings
held in a particular independent monitoring committee divided by the sum of full board meetings, executive committee
meetings, and the number of meetings in the particular monitoring committee. Alternative constructs to this measure,
and issues associated with the measure are discussed in full in Section 1.2.
3
construction of this measure is the notion that having directors perform their board
responsibilities in the presence of management can alter board oversight, and further, may benefit
the CEO. Past research and anecdotal evidence on the inner-workings of boards provide support
for this contention. Charles Elson notes that directors face significant pressure when speaking in
front of the executive officers of the firm: ‘In a boardroom, there is nothing more difficult to do
than to talk about the CEO while the CEO is present.’
5
Moreover, Mace (1986) details a case
study where an independent director was removed from the firm’s proxy statement after openly
disagreeing with management during a board meeting.
6
Hence, if CEOs desire to control board
oversight, then this measure appropriately functions to capture the operational control over
monitoring decisions on the board and the degree to which outside directors handle their duties
free from CEO interference.
With this measure of internal board control in mind, I empirically investigate how CEO
influence over the operations of the board has changed over time. Following the various
corporate malfeasance scandals of 2000-2002, the Sarbanes-Oxley Act of 2002 (SOX) and the
NYSE/NASDAQ listing requirement changes in 2003 were enacted with the intention of being a
comprehensive solution to the governance problems which brought about the scandals. However,
since most firms were already in compliance with the board independence mandates, did these
events have any real impact on the CEO’s control over the board? Examining board changes for
a sample of 586 NYSE firms, I document that the structure of the board underwent a significant
transformation between 1999 and 2005, not in terms of size or independence, but in terms of the
internal decision-making processes on the board. While board independence increased a
marginal 5% between 1999 and 2005, the structural form of the average board transitioned from
5
See ‘Emerging Trends in Corporate Governance’, a supplement to Corporate Board Member, 2001.
6
See Lorsch and MacIver (1989) for further evidence on the pressures associated with speaking out in the boardroom,
and Sonnenfeld (2002) for a discussion and examples of boardroom conformity.
4
one where the CEO was present for and had a voting stake in the majority of board meetings, to a
structural form where the vast majority of board meetings were held by outside directors in
independent committees, removed from the CEO’s voting influence. In particular, in 1999 the
average CEO presided over 9.10 meetings a year, while outside directors held a total of 8.56
meetings in the independent monitoring committees (3.48 audit meetings, 3.98 compensation
meetings, and 1.11 nominating meetings).
7
By 2005, the average CEO presided over 8.59
meetings a year, while outside directors held a total of 18.42 meetings in the independent
monitoring committees (9.15 audit meetings, 5.45 compensation meetings, and 3.82 nominating
meetings). This implies that 48% of board meetings were held in independent committees in
1999, and 68% of board meetings were held in independent committees in 2005. Or,
alternatively, the average CEO participated as a member in 52% of meetings held in 1999 and
only 32% of meetings in 2005.
8
Further, on an individual committee basis this entails that over
this time period the fraction of meetings held outside of the CEO’s voting influence in the
independent audit, compensation, and nominating committees rose 80, 30, and 200 percent,
respectively. Hence, the fact that the CEO presided over fewer board meetings, in conjunction
with the significant increase in work allocation to independent monitoring committees suggest
that the CEO’s influence over the monitoring decision-making processes on the board decreased
over this period.
While these results highlight a shift in board oversight control, equally important to our
understanding of how board structure changed during this period are the investment operations of
7
Meetings which the CEO presides over (has a voting stake in) is constructed as the number of full board meetings
plus the number of committee meetings which the CEO serves as a member on. Further, throughout this investigation,
‘independent monitoring committees’ denote committees where outside directors are the sole voting members.
8
The frequency of meetings held on ‘miscellaneous monitoring committees’ other than the audit, nominating, and
compensation committees are sparse (under 5% of committee meetings held in both 1999 and 2005). The inclusion of
these meetings does not alter results, though is further detailed in later sections.
5
the board. If the executive committee functions as an environment where the CEO may
implement policy decisions (i.e. dividend and capital structure changes) with far fewer outside
directors scrutinizing such decisions, then how did the operations of this committee change
surrounding the regulatory events of 2002-2003? I document that in 1999 32% of firms held one
or more meetings in the executive committee, while by 2005 only 19% of firms held one or more
executive committee meetings. In a similar manner, examining the average number of meetings
held in the committee, the fraction of board-time spent by the CEO in the executive committee
decreased by 40% over this time period.
9
Together, these findings support the contention that the
CEO’s ability to side-step the oversight of the full board and implement policy/investment
decisions through the executive committee were significantly curtailed between 1999 and 2005.
While past empirical research has demonstrated that SOX had a strong mechanical
impact on director workloads and the risks associated with holding board positions (Linck et al.
(2009)), the results presented here extend these findings by providing strong supporting evidence
that these regulatory events not only affected board work levels, but also the CEO’s involvement
in the decision-making processes of the board. Though the finding that independent monitoring
committee meetings increased post-SOX is consistent with the notion that more work had to be
done on the board, the fact that the number of meetings which the CEO presided over and the
number of meetings held in the executive committee both significantly decreased suggest that
CEO influence over board oversight/operations was abated during this time period.
10
Hence,
given that the average board already exhibited a high degree of ‘nominal independence’ by 1999
9
The fraction of board-time spent on the executive committee is constructed as the number of meetings held in the
executive committee divided by the sum of full board meetings and executive committee meetings. In addition, the
number meetings held in ‘miscellaneous investment committees’ are few and remained relatively constant over the
time period. Such results are presented in greater detail in later sections.
10
Further, any ‘scaling’ argument with regard to board operations would necessitate that not only should the fraction of
meetings in monitoring committees increase, but also the fraction of meetings in the executive committee (under the
condition that CEO influence remained constant).
6
(80% outsider representation), the documented changes in board structure indicate that
shareholder demands for greater board scrutiny following the corporate malfeasance scandals of
2000-2002 were primarily satisfied via an alternative channel of ‘independence’ on the board -
the removal of the CEO as a participating member in the board’s internal operations.
Next, to lend further clarity on how operational control within the board relates to firm
determinants and the balance of power between outside directors and the CEO, I investigate the
cross-sectional variation in the operational form of the board over the 2005-2006 time period.
Several authors have formulated and tested numerous theories pertaining to the relationship
between board composition and firm-level determinants (Raheja (2005); Lehn et al. (2005);
Boone et al. (2007); Harris and Raviv (2008); Duchin et al. (2010)). Collectively these works
demonstrate that board independence and size are a product of a firm’s business environment,
information environment, and various contracting costs. Following this line of literature, I
categorize these theories pertaining to board structure into three primary hypotheses: the scope of
operations hypothesis, the monitoring hypothesis, and the negotiation hypothesis. To further our
understanding of what determines the internal structure of the board, I extend each of these
hypotheses by examining how control over the monitoring/investment decision-making processes
of the board is associated with the bargaining position of the CEO, information costs for outside
directors, and other firm-level factors.
First, in accordance with the scope of operations hypothesis, I demonstrate that the
fraction of board meetings held in committees is positively related to firm size. This finding is
consistent with the notion advanced by Fama and Jensen (1983) that complex firms develop more
hierarchical organizations. Following this, I investigate how information costs faced by outside
directors and firm-level managerial private benefits relate to the operational form of the board
(monitoring hypothesis). I ultimately find weak supporting evidence that monitoring costs are
7
negatively associated with the fraction of meetings held by outside directors in monitoring
committees, and that managerial private benefits are positively associated with the fraction of
board meetings held in monitoring committees. Yet, this lack of conclusive evidence regarding
the monitoring hypothesis is not entirely surprising given some of the indeterminate empirical
results documented in the prior literature (Boone et al. (2007); Coles et al. (2008); Linck et al.
(2008)).
11
Next, I examine how CEO influence affects operational control over board proceedings.
If board structure follows from a negotiation process between the CEO and the outside directors
on the board (Hermalin and Weisbach (1998)), then in what manner do high power CEOs bargain
for lower levels of board oversight? While prior empirical studies have demonstrated that the
proportion of independent directors on the board is negatively related to measures of CEO
influence (Boone et al. (2007); Linck et al. (2008)), I extend the negotiation hypothesis by
detailing that high power CEOs (high ownership, high tenure) are associated with a lower fraction
of meetings held in independent monitoring committees (higher fraction of meetings held in the
CEO’s presence). Further, CEO power is also negatively related to the fraction of board-time
spent in non-management executive sessions.
12
Together, these findings suggest that CEOs who
have the ability to alter board structure will pull the monitoring operations of the board away
from independent committees and back toward an environment where they may preside over
monitoring discussions and influence board oversight.
11
Boone et al. (2007) and Coles et al. (2008) find ultimately inconclusive results (not in accordance with the
monitoring hypothesis) regarding the association between R&D expenditures (monitoring costs) and board
independence. Similarly, Linck et al. (2008) document an insignificant relation between board size and R&D. With
respect to the effect of managerial private benefits on board composition, Boone et al. (2007) find an insignificant
association between independence and free cash flow.
12
Vafeas (1999) provides evidence that CEO bargaining power is negatively related to the number of board meetings.
Hence, these results extend such a finding by demonstrating that high bargaining position CEOs prefer to have outside
directors handle their monitoring operations in their presence as opposed to independent committees.
8
Following this investigation into the relationship between CEO influence and the
monitoring structure of the board, I also extend the negotiation hypothesis through an
examination of board-level control over policy decisions. In particular, I find that CEO
bargaining power is positively associated with the fraction of board meetings handled in the
executive committee. CEOs with the capability to affect the structure of board operations spend
less time convening the full board for approval on investment decisions, and instead bypass the
oversight of outside directors by enacting such decisions through the executive committee. In
total, these results highlight the mechanism by which powerful CEOs, despite being subject to
boards with 80+% outsider representation, still control board-level investment and monitoring
decisions in the modern boardroom.
13
Moreover, the results associated with the negotiation hypothesis offer support for the
notion that the drastic shift in the operational form of the board between 1999 and 2005 was
contrary to the preferences of the CEO. Since CEOs who have a greater ability to influence board
structure allocate a lower fraction of meetings to be held outside of their presence in independent
the documented changes in board operations over the 1999 to 2005 time frame (more board-time
spent in outsider controlled committees, and fewer meetings held in the executive committee)
were against the desires of the CEO.
Overall, the results presented here extend our understanding of board structure in two
ways. First, the significant transformation in the internal operations of the board between 1999
and 2005 lends support to the idea that the second wave of board governance reform to occur in
the past 60 years was one in which CEOs were removed from the decision-making processes of
the board, diminishing their ability to influence and control board oversight/investment
13
Adams et al. (2005) demonstrate that powerful CEOs are associated with higher variability in firm performance and
decisions. Further, Core et al. (1999) document that CEOs who hold the board chair position demand higher cash-
based and total compensation.
9
operations. Second, the results pertaining to the cross-sectional determinants of board operating
form demonstrate that the internal structure of the board is an important feature to consider when
discussing issues related to board control and governance in the modern board. Together, these
findings shed light on a previously unexplored area of board structure, and highlight that the
internal operations of the board may offer a more complete and robust understanding of
‘independence’ on the board.
This paper proceeds as follows. Section 1.2 develops the boards hypotheses and details
the construction of the data. Section 1.3 presents the cross-sectional determinants of board
operational form and the time trends. Section 1.4 concludes the paper.
1.2 Background Information, Data, and Summary Statistics
In this section, I first summarize how recent governance and listing requirement changes relate to
this empirical investigation. Next, I highlight the existing research on the role of the board and
the responsibilities of committees. Following this, I describe and extend three hypotheses in the
boards literature. Finally, I detail the construction of the dataset used in this investigation and
provide summary statistics.
1.2.1 Post-SOX Regulatory Environment
The regulatory changes of the late 1990s and early 2000s constituted a significant shift in the
governance standards applied to U.S. public firms. Beginning in 1999, NYSE and NASDAQ
implemented listing rules requiring the complete independence of audit committees. Following
the Enron, Tyco, and WorldCom corporate and accounting scandals, the Sarbanes-Oxley Act of
2002 (SOX) was enacted with the intention of being a thorough solution to the governance
deficiencies which engendered the scandals. Included in the regulation were rules to formally
declare audit committee independence, improve the quality of financial statements, and
10
strengthen the enforcement of securities law.
14
In 2003, NYSE and NASDAQ both took
measures to further strengthen the SOX regulatory requirements by mandating that publicly listed
firms have a majority of independent directors on their boards. In addition, both exchanges
refined the SOX definition of ‘independent’ board member, set rules for the composition of board
committees, and required that audit committees have financially literate members.
While the two exchanges set similar listing requirements regarding audit committee
composition, their rules regarding other committees differed slightly. NYSE required that all
firms establish audit, nominating/governance, and compensation committees comprised entirely
of independent directors and that such independent directors were to meet separately from inside
board members in non-management executive sessions on a regular basis. NASDAQ took similar
measures regarding committee formation, yet allowed more flexibility in the composition of these
committees. NASDAQ did not explicitly require that firms have nominating or compensation
committees, but compensation payable to the CEO and other officers had to be approved either by
a majority of the independent directors on the board or a compensation committee of independent
directors. Similarly, for these companies, nominations had to be approved either by a majority of
the independent directors on the board or a nominating committee of independent directors. Also,
if a company elected to establish compensation and nominating committees of at least three
members, then one director who is not independent under NASDAQ’s rules may hold a position.
In addition, both exchanges granted certain entities (e.g. controlled companies, limited
partnerships, foreign private issuers, and other passive organizations) exemptions to a number of
these rules regarding committee and board independence. Both exchanges instituted timetables
14
See The Practitioner’s Guide to Sarbanes-Oxley Act, Volume 1, The American Bar Association (2004) for more
information on SOX requirements.
11
by which firms had to comply with the rulings. In general, firms had to meet the listing
requirements by late 2004, with extended time (late 2005) given to firms with staggered boards.
Given NYSE’s more definitive rulings regarding committee independence, the sample of
firms used in this empirical analysis is based on a set of post-SOX NYSE firms. NYSE’s 2003
mandate to require the complete independence of the monitoring committees (audit,
compensation, nominating/governance) enables the collection of a distinct and ‘clean’ dataset.
Since inside directors are prohibited from serving on such committees, this yields an environment
where the distinction between the influence of independent directors and inside directors (or the
CEO) is easily observable. Further, this delineation between the board and its monitoring
committees also allows for the development and extension of several hypotheses in the boards
literature.
1.2.2 Background Literature and Committees
The board of directors’ responsibilities extend far beyond that of monitoring the CEO’s
performance and replacing the CEO, should the situation warrant it. The Business Roundtable
(1990) details the five primary functions of the board: (1) review and approve the major plans and
strategies of the corporation; (2) advise executive officers on corporate issues; (3) evaluate, and if
necessary, replace executive officers, and set compensation practices; (4) evaluate board
performance and provide shareholders a slate of candidates for the board of directors; (5)
formulate and review systems for corporate legal and regulation compliance.
Each of these responsibilities of the board may be handled by the full board, where all
members discuss such issues, or may be delegated to committees, where a select few individuals
focus on particular tasks. Kesner (1988) argues that the primary monitoring decisions of the
board originate at the committee level. Further, Vance (1983) notes that corporate decisions are
primarily influenced by four board committees: the audit, executive, compensation, and
12
nominating committees. Hence, the structure of committees within the board may be an
important determinant of board performance if the board is primarily operating through its
committees.
This notion that board oversight is a function of not only the composition of the board as
a whole, but also of the structure of the board’s committees has recently been advanced by
various empirical works. Klein (1998) finds little association between overall board structure and
performance, but does find a positive relation between the percentage of insiders on investment
committees and firm performance. Xie et al. (2003) provide evidence that committee
composition affects the likelihood of earnings management. The authors highlight that audit
committees comprised of members with financial or corporate backgrounds are associated with
smaller discretionary accruals at the firm-level. In similar fashion, Klein (2002) also documents a
negative relation between audit committee independence and abnormal accruals. Shivdasani and
Yermack (1999) investigate the CEO’s influence over the director nominating process. The
authors find that when the CEO sits on the nominating committee, or no such committee exists,
fewer independent directors are chosen for available board seats.
For the purposes of this study, to understand how the internal structure of the board
relates to firm determinants, it is important to first summarize how various committees operate
within the board. I provide a detailed look at the tasks and responsibilities of the four committees
of most importance in this study.
15
What follows below is a conglomerate description taken from
firm proxy statements on the operating functions of the audit, compensation, nominating, and
executive committees.
The audit committee’s primary responsibilities are to oversee the financial reporting of
the firm, the disclosure process, the appointment of independent auditors, and to monitor the
15
Hayes et al. (2004) provide a similar look at all the functions of committees in their sample. See the authors’ work
for a detailed look at the functions of less frequent committees (e.g. technology, pension plan, corporate responsibility).
13
performance of the auditors. The committee also monitors the internal control process, consulting
auditors to discuss these matters, and monitors the choice of accounting policies. In addition, the
committee may also be tasked with discussing risk management practices, compliance with laws
and regulations, and reviewing safety and environmental audit functions.
The compensation committee’s primary tasks are to review and recommend to the full
board the CEO’s and officers’ compensation - including salary, benefits, and long-term incentive
plans. The committee may also establish and monitor performance guidelines for the CEO and
evaluate such performance. In addition, it can make recommendations concerning director
compensation and oversee the appointment of consultants to help with such compensation issues.
The nominating/governance committee is responsible for reviewing, assessing, and
nominating members of the board of directors. It also reviews criteria for new directors, deals
with consultants to find appropriate new members, and recommends committee assignments
within the board. The committee is also responsible for developing corporate governance
principles, shaping the governance standards of the company, and is often tasked with overseeing
the company’s CEO succession planning process.
The executive committee is responsible for exercising the powers of the board and the
affairs of the firm when the board is not in session. The committee primarily deals with dividend
and capital structure decisions, and has the right to alter or change such practices (including the
issuance of equity). Limitations to the powers of the executive committee are set by firm by-
laws. One near universal restriction on the powers of the executive committee is that it cannot
change by-laws or amend the firm’s articles of incorporation.
1.2.3 Development of Cross-sectional Hypotheses
Past empirical and theoretical studies on board structure provide evidence that firm and market
determinants affect the size and composition of the board. Here, I detail how these determinants
14
relate to three primary hypotheses in the boards literature, and in turn, how these hypotheses
apply to this investigation into the internal allocation of work on the board.
Fama and Jensen (1983) conjecture that the manner in which a firm is organized stems
from the complexity of its operations. Large firms, or firms with more detailed and complex
processes, will function in a more hierarchical manner. This notion, often referred to as the scope
of operations hypothesis, has served as a basis for investigations into the relation between firm
complexity and board structure.
16
Through studies into the size and composition of the board, the
scope of operations hypothesis has also been empirically supported by numerous authors. If
outside directors bring valuable expertise and oversight to the board, then firms with disparate
business lines and larger structures should benefit from larger and more independent boards.
Boone et al. (2007) document such a finding - firms construct more independent and larger
boards as they grow in size and complexity over time. Coles et al. (2008) find that diversified
firms have more independent directors sitting on the board to monitor and advise the vast set of
operations of the firm.
17
Knyazeva et al. (2009) and Anderson et al. (2009) extend these ideas
and provide evidence that complex firms form more heterogeneous boards in terms of director
expertise and director occupation.
Thus, with respect to this empirical investigation, the scope of operations hypothesis
would imply a positive association between firm complexity and work allocation on the board. If
firm complexity fosters a more rigid hierarchical firm form (Fama and Jensen (1983)), then it
stands to reason that the same should apply to the form of the board. Large and diverse firms
should tend to structure the monitoring and investment aspects of the board as distinct units, with
16
See Raheja (2005) for similar predictions regarding the association between firm size and board size. The author
argues that the trade-off between the extra monitoring ability of additional members and the free-riding (moral hazard)
problems associated with additional members defines board size.
17
See Denis and Sarin (1999), Lehn et al. (2005) and Linck et al. (2008) for further evidence validating the positive
relation between board size and firm size.
15
the board spending more time in separate committees as compared to time spent making decisions
as a full board. In other words, if large firms construct boards with more outside directors and
greater levels of expertise heterogeneity, then such firms should be more inclined to partition the
tasks of the board, and allocate a greater fraction of board work to be handled in committees.
Consistent with the past literature, to proxy for firm complexity, I use firm size, firm age, and the
number of business segments.
A second hypothesis in the boards literature is that the form of the board should reflect
the costs of monitoring and the managerial private benefits present at the firm-level. This two-
fold hypothesis is often denoted as the monitoring hypothesis. First, Adams and Ferreira (2007)
model the structure of the board and suggest that the number of outsiders sitting on the board
should decrease with the costs of monitoring. Extending this idea empirically, Coles et al. (2008)
provide some evidence that since outside directors are ineffective at monitoring firms with high
growth potential, the fraction of insiders on the board will be positively related to R&D
expenditures. Further, Linck et al. (2008) and Boone et al. (2007) make similar arguments that
firms should create smaller and less independent boards the greater the level of asymmetric
information between the firm and outside directors.
If inside director and CEO knowledge is an important feature to a well functioning board
in high asymmetric information environments, then outside directors in such boards should stand
to benefit from a discussion (transfer of knowledge) with inside members before making their
board-level decisions. Since outside directors must serve by themselves on the primary
monitoring committees in this post-SOX period, firms in high monitoring cost environments
should be more inclined to pull the operations of the board away from monitoring committees
(where inside directors have no say) and structure board operations so that oversight decisions
are discussed at full board meetings. In such, the monitoring hypothesis would predict that the
16
fraction of board-time spent by outside directors in monitoring committees (removed from the
full board) is negatively related to monitoring costs. Consistent with the prior literature, R&D
intensity is used to proxy for the importance of firm-specific knowledge (monitoring costs).
In addition to the costs of monitoring, board composition should also be related to the
level of managerial private benefits. This second facet of the monitoring hypothesis has been
theoretically modeled by various authors (Raheja (2005); Adams and Ferreira (2007); Harris and
Raviv (2008)). Simply, these authors demonstrate that as firm-level managerial private benefits
increase a more independent board is optimally constructed to properly constrain the actions of
management. In the context of this empirical investigation, managerial private benefits should be
positively associated with monitoring work allocation to committees. If higher levels of board
oversight are needed to constrain management as private benefits increase, it follows that a
greater fraction of the internal monitoring operations of the board should be handled outside the
influence of the CEO. Following the existing literature on the issue of managerial private
benefits, I implement two measures to proxy for private benefits: free cash flow and antitakeover
provisions (E-Index).
18
In total, the monitoring hypothesis predicts that the fraction of board-time
spent on the independent monitoring committees (apart from the full board) should be positively
associated with managerial private benefits (free cash flow, antitakeover provisions) and
negatively associated with information costs (R&D expenditures).
A third primary hypothesis in the boards literature is the negotiation hypothesis. The
predictions of this hypothesis generally follow from the idea that CEOs bargain with shareholders
for certain board features that suit their interests. Hermalin and Weisbach (1998) formalize this
hypothesis in a model where CEOs use their influence (via surplus production) to negotiate for
18
Jensen (1986) argues that free cash flow is associated with agency conflicts since management have strong
incentives to waste it on pet projects instead of making investment decisions in the interests of shareholders. Hence,
free cash flow serves as an appropriate measure of managerial private benefits. For evidence pertaining to the value
destroying nature of takeover protection see Scharfstein (1988), Gompers et al. (2003), and Bebchuk et al. (2009).
17
insiders, or affiliated directors, to be placed on open board seats. The model suggests that as a
CEO’s bargaining position increases, board independence will fall.
19
Support for this theory
comes from a number of recent empirical investigations (Baker and Gompers (2003); Boone et al.
(2007); Linck et al. (2008)).
If CEOs dislike the monitoring role played by outside directors and derive private
benefits from control over the operations of board, then CEOs with considerable influence over
the firm will mandate that the internal processes of the board be handled in their presence. As
CEO bargaining power increases, it stands to reason that the monitoring and investment functions
of the board will be controlled by the CEO and not by outside directors. Hence, the fraction of
board work performed by outside directors removed from the CEO’s presence (fraction of board-
time spent in independent monitoring committees) will be negatively related to CEO bargaining
power. Further, if CEOs desire to control dividend and capital structure decisions with minimal
interference from outside directors, the negotiation hypothesis also predicts that the fraction of
work handled by the CEO in the executive committee will be positively associated with CEO
bargaining position. Consistent with the literature, to proxy for CEO influence I consider two
primary measures: CEO ownership and CEO tenure.
1.2.4 Dataset Construction and Variable Specification
The sample of firms used in the empirical analysis to test the three boards hypotheses is based on
a set of post-SOX NYSE firms from 2005-2006. NYSE’s 2003 listing requirement changes to
mandate the complete independence of the primary monitoring committees (audit, compensation,
nominating/governance) yields an environment where the construction of a ‘clean’ dataset to
extend and test the hypotheses is feasible. To construct my sample of NYSE firms, I start by
19
Other predictions of Hermalin and Weisbach (1998) are that more outsiders will be added to the board following
poor performance, insiders are added to the board as the CEO reaches retirement, and that independence will decrease
with CEO tenure. In addition, see Raheja (2005) for an alternative set of predictions regarding board composition and
CEO influence.
18
accessing Compustat for the following firm-specific information: total assets, firm age, number of
business segments, book leverage, R&D intensity (R&D/Sales), free cash flow (FCF),
acquisitions, ROA, market-to-book (ratio of the market value to book value of assets).
20
In
addition to information regarding firm characteristics, the CRSP monthly files are used to define
all firm prices and returns. To ensure that outliers do not have an impact on the results, variables
are winsorized at the 1% level.
To obtain information on firm-level institutional ownership and charter provisions, I
access the Thomson Financial Institutional Ownership database and the IRRC database,
respectively. In conjunction, these two databases serve to provide the necessary information
needed to construct various proxies for the governance standards of the firm. First, the IRRC
database provides annual data on firm-level provisions regarding staggered boards, poison pills
and other charter/bylaws for approximately 1,500 firms, primarily from the S&P 500 and other
large corporations. Bebchuk et al. (2009) construct a takeover defense index, the entrenchment
index (E-Index), based on charter amendments, supermajority requirements, golden parachutes,
poison pills, limits to shareholder bylaw amendments, and staggered boards. The E-Index used in
subsequent sections, a proxy for firm-specific shareholder rights, follows from the authors’
construction. Next, for information regarding institutional holdings at the firm-level, I access
Form 13-F statements from the Thomson Financial Institutional Ownership database. This
database provides the required information needed to construct a measure of aggregate
institutional ownership (sum of shares held by all institutional investors).
20
Specifically, return on assets (ROA) is operating income before depreciation over assets. Market-to-book is the book
value of assets minus book value of equity plus the market value of equity normalized by the book value of assets.
Book leverage is the ratio of debt (long term total debt plus debt in current liabilities) to shareholders equity. Free cash
flow (FCF) is income before extraordinary items plus depreciation and amortization less total dividends paid
normalized by total assets.
19
Next, as a basis for the necessary board-level data, I use the Corporate Library for
information on director characteristics and board membership. The Corporate Library provides
data on board size, director affiliation, director tenure, director ownership, and committee
structure. In particular, from this database, board independence is constructed as the fraction of
non-employee directors on the board: the number of outside directors divided by the total number
of directors, where affiliated outside directors are denoted as outsiders. While not in direct
accordance with the NYSE definition of independence, this measure is consistent with the prior
literature (Coles et al. (2008), Huson et al. (2001)). Further, this measure of independence is
preferred to the alternative construct, where affiliated directors are treated as insiders, due to the
fact that the definition of ‘affiliated director’ has changed over time (i.e. different specifications
by the exchanges in 1999 and 2005). Moreover, the Corporate Library’s categorization of
affiliated director often does not match NYSE definitions. Hence, this treatment gives the
cleanest and most consistent measure of board composition over time. Following this, for
information pertaining to CEO characteristics, I access the ExecuComp database. ExecuComp
provides CEO and officer data, including compensation, CEO age, CEO ownership, and CEO
tenure.
To supplement the board-level data provided by the Corporate Library, I hand collect
detailed board operations information from firm proxy statements (DEF 14A) over the 2005 to
2006 fiscal years (2006 to 2007 reporting years), available from the SEC’s EDGAR reporting
system. Should pertinent information be unavailable in these proxy statements, firm 10-K
statements (annual reports) are used to provide supplemental information. To avoid
complications with changes in board behavior which may have occurred following the financial
20
crisis of 2007, the 2005 and 2006 fiscal years serve as the central time frame in this study.
21
To
limit the size of the pre-collection dataset, I require that necessary firm-level data be available
from Compustat (total assets, market-to-book), and that the firm be present in the IRRC (E-
Index), Thomson (institutional ownership), Corporate Library (board membership), and
ExecuComp (CEO ownership) databases. Further, since the empirical analysis requires previous
year observations as controls, each firm must have available information for two consecutive
years. All regulated entities (utilities and financials) and firms that are not in compliance with the
2003 NYSE rulings (e.g. foreign private issuers, controlled companies, firms in bankruptcy and
other passive organizations) are also removed from the dataset.
22
These necessary conditions
result in 1,356 firm-year observations over the 2005 to 2006 period.
From firm proxy statements (DEF 14A), I record detailed information on each firm’s
committee structure - which standing committees exist within the board, the composition of each
committee, and the number of meetings held by each committee in the fiscal year.
23
In addition, I
cross check the board-level information (board members and affiliation) provided by the
Corporate Library and change any differences in accordance with the proxy statement
information. Consistent with Adams (2003), board committees are classified by their three
primary functions: monitoring, investment (advising), and stakeholder interest. The three
monitoring committees of foremost concern in this investigation are the compensation,
21
Unobservable firm-level exposure to certain factors may have caused large shifts in the operations of the board after
2007, and for that reason a post-financial crisis time period is avoided in this study.
22
The inclusion of these firms does not alter subsequent results and in fact strengthens the results regarding the primary
hypotheses.
23
Schedule 14A of the Securities Exchange Act of 1934 requires firms to disclose the functions performed by their
committees, the names of committee members, and the number of committee meetings during the last fiscal year.
Anecdotal evidence suggests that board meetings held via teleconference are a fraction of the length of in-person board
meetings (full meetings). Hence, such meetings are treated as half-meetings in this investigation. Results throughout
hold in a qualitatively identical fashion whether teleconference meetings are treated as regular meetings (full meetings)
or completely omitted.
21
nominating/governance, and audit committees. As mandated by NYSE’s listing requirements,
each firm has such a committee and discloses the operations of each of these monitoring
committees in its proxy statements. Predominately, the operations of these monitoring
committees are handled by outside directors apart from managerial input. The independent
chairman of each committee sets the agenda for all meetings and reserves the right to call other
officers of the firm to their committee meetings to assist with decisions, yet the language of the
disclosure statements suggest that in general a vast majority of meetings are handled in isolation
from inside director influence.
24
The one exception to this rule is the audit committee. The audit
committee frequently meets with external auditors and the CFO of the firm to prepare and review
financial statements. Considering the nominating and compensation committees, 20 out of the
1,356 firm-year proxy statements explicitly note, or imply through the language of the document,
that the CEO attended a majority of the meetings. The inclusion or exclusion of these
observations has no material impact on results throughout the paper. Hence, the monitoring
committee meeting in the post-SOX board represents an environment where not only does the
CEO/insider have no voting stake, but the CEO/insider has also relinquished all control over
decisions to the outside directors on the board (i.e. cannot dictate agenda/policies, and attends
the meeting only if called there by an independent director).
The primary investment committee of greatest concern in this investigation is the
executive committee. The executive committee operates in the board’s stead when the full board
is not in session and may make decisions on behalf of the board should the full board not be able
to convene. The board may also delegate to the executive committee the authority to make
certain policy decisions, limited by the articles of incorporation. Following this, committees
24
Further, a typical proxy ‘Report of the Compensation Committee’ or ‘Report of the Nominating Committee’ will
state that non-management executive sessions were held following any committee meeting in which an executive was
called to attend.
22
organized to represent the stakeholders’ interests constitute the smallest fraction of committees in
the sample. Committees dealing with public image issues (e.g. contributions, human resources,
environment, diversity, corporate responsibility, public issues) are all classified as stakeholder
committees.
Next, although small in total numbers, many firms have other miscellaneous committees
operating within the board. Committees organized to deal with safety, retirement/pension,
options, and succession are denoted as ‘miscellaneous monitoring committees’. Committees
dealing with technology, strategy, and acquisition issues are recorded as ‘miscellaneous
investment committees’.
25
The final committee not classified into any particular category is the
finance committee. The finance committee may function as a monitoring committee, scrutinizing
the capital structure decisions of the CEO, yet may also serve an advisory role to the executives
of the firm (Klein (1998)).
26
Given its dual functions, I do not allocate the finance committee to
either the ‘miscellaneous monitoring committee’ group or the ‘miscellaneous investment
committee’ group. Throughout the empirical analysis, allocation of this committee to either of
the two groups does not alter results.
Following the assignment of committees, NYSE’s 2003 listing rules also required boards
to hold regularly scheduled outside executive sessions, where independent directors meet
amongst themselves, separate from the CEO and any other current employee directors.
27
Since
outside executive sessions constitute an NYSE mandate, and not a specific committee, the
disclosure of the number of such meetings is not explicitly required. Nevertheless, firms often
25
All together, miscellaneous monitoring committee meetings, miscellaneous investment committee meetings, and
stakeholder meetings account for a small fraction of the total committee meetings in the sample (8%).
26
19% of firms have a standing finance committee in the sample.
27
See SEC Release No. 34-48745 (November 4, 2003) at http://www.sec.gov/rules/sro/34-48745.htm for more details
on the issue.
23
report the number of outside executive sessions in proxy statements. In fact, only 14% of firms
make no mention of the issue, and 21% of firms state that they are in compliance with the NYSE
listing requirements or that ‘executive sessions of outside directors were regularly held’. In the
data, firms appear to reveal the number of outside executive sessions held in a given fiscal year
with a lower bound of one-quarter the level of full board meetings (e.g. 8 full board meetings and
2 outside executive sessions in a given year). In accordance with this finding, missing
observations, or firm observations which simply state compliance with the NYSE outside
executive session mandate, are recorded as having one-quarter the number of outside executive
sessions as full board meetings. In later sections, I address robustness checks to this assumption.
For the second part of the empirical analysis in this paper, detailed information on
committee and board operations is also needed from the pre-SOX period. Taking the original set
of 2005 to 2006 firm-year observations, I create a matched sample to the year 1999 where
inclusion is conditional on being present in the 2005-2006 dataset. I use the IRRC database to
provide supplementary information for the 1999 set of firm observations. Identical committee
and board operations variables are collected for this earlier set of data, and identical cross-checks
regarding board affiliation are made as well. This construction yields 586 firm observations with
available board, ownership and financial data for the 1999 fiscal year.
28
1.2.5 Specification of Primary Measures
Given the construction of the 1999 and 2005-2006 datasets, I now address the exact specification
of the primary measures used in this study to proxy for the allocation of and control over internal
board work. First, the fraction of board meetings handled by independent directors in a particular
monitoring committee (Frac Monitoring) is constructed as the number of meetings held in the
28
The loss of seventy observations follows generally from insufficient information (lack of coverage) in Compustat
and ExecuComp for the 1999 sample.
24
particular monitoring committee divided by the sum of full board meetings, executive committee
meetings, and the number of meetings in the particular monitoring committee. This measure
functions to capture the fraction of board monitoring work controlled by outside directors in the
committee. The denominator includes the sum of full board meetings and executive committee
meetings since executive committee meetings serve as a substitute to full meetings for the CEO
(i.e. the CEO may call executive committee meetings in lieu of full board meetings). If the CEO
was constrained by directors who would not allow executive committee meetings to take place,
such meetings would otherwise be full board meetings. In essence, the denominator of the
measure operates to proxy for the amount of work which the CEO controls (has a voting stake in),
while the numerator operates to proxy for the amount of monitoring work which the independent
directors control.
Next, the fraction of board work handled in the executive (investment) committee is
constructed in a similar manner: the number of meetings held in the executive (investment)
committee divided by the sum of full board meetings and executive (investment) committee
meetings. This measure, denoted as Frac Exec (Frac Exec/Inv), serves to proxy for the CEO’s
ability to control policy/investment decisions within the board (i.e. avoid the oversight of the full
board and make decisions through the executive committee).
Moreover, it is critical to note that the forthcoming cross-sectional results associated with
these measures are robust to alternative constructions. If the fraction of board meetings handled
by a particular monitoring committee (Frac Monitoring) is altered to include the number of
‘miscellaneous investment committee’ meetings in which the CEO holds a position (in the
denominator of the measure), the primary results detailed in this investigation hold in a
25
qualitatively similar manner.
29
Similarly, if the fraction of board work handled in the executive
committee (Frac Exec) is altered to exclude executive committee meetings from the denominator,
the significance of the primary findings do not change. Additional alternative constructions to the
primary measures are also detailed throughout the subsequent analysis.
Further, it is also important to discuss the limitations of these measures before
proceeding. First, each constructed measure includes only a ‘count’ of committee and full board
meetings, and hence the observed length of time spent in each meeting, and the effort intensity of
each meeting is unobservable. The assumption throughout is that each meeting is equivalent to a
unit of board-time and that across firms and within the board itself, such a measure of board-time
captures a relatively consistent fraction of the work/effort devoted to a task. Next, though Frac
Monitoring is constructed to capture the degree to which outside directors handle their monitoring
operations in isolation from the CEO’s influence, it is true that it is ultimately indeterminate the
degree to which CEOs participate in committee meetings as non-voting members. Though the
reports filed in firm proxy statements suggest that CEOs generally do not attend meetings,
whether the CEO is physically present for some or many meetings is again inconclusive. Yet, as
previously discussed, the monitoring committee still represents an environment where the CEO
has relinquished ultimate voting control and procedural control to the outside directors. With the
central measures of board operational control well defined, I next address sample descriptive
statistics.
1.2.6 Summary Statistics
Table 1.1 provides summary statistics for the 1,356 firm-year observations over the 2005-2006
period. Panel A includes the mean, median, standard deviation, 25th percentile, and 75th
29
Since the percent of committee meetings held in ‘miscellaneous investment committees’ amounts to 2%, and the
vast majority of these committees do not have the CEO as a voting member, such an alteration to the measure has no
material impact.
26
percentile for various firm financial and governance measures. The mean (median) value of total
assets is 10331 $MM (2733 $MM) in the sample. The average firm in this study is larger
compared to firms in previous boards studies (Boone et al. (2007); Linck et al. (2008)), yet this
follows as a natural consequence of the stringent sample requirements previously detailed. The
average firm in the sample has free cash flow at 8.5% of total assets, book leverage of 35%, is 29
years old, and has an R&D intensity of 0.021. The mean ROA and market-to-book of a firm in
the sample is 0.148 and 1.87, respectively. The median ROA and market-to-book of a firm in the
sample is 0.140 and 1.61, respectively. The governance summary statistics are also consistent
with past studies. The mean E-Index and institutional ownership for the sample firm is 2.44 and
81%, respectively. The median E-Index and institutional ownership is 2 and 84%, respectively.
Compensation statistics are also comparable with previous studies. The mean level of CEO
ownership in the sample is 1.30%. The median level of CEO ownership in the sample is 0.26%.
Together, these two points suggest that a few CEOs hold considerable and sizeable stakes in their
firm, while most CEOs in the sample hold low levels of firm equity. In addition, the average
CEO has held the executive position for 6.40 years. Mean director ownership (the average
percent of shares held by outside directors, by firm) is about one-fifth the level of mean CEO
ownership (0.21% v. 1.30%), and mean tenure levels are comparable between CEOs and outside
directors (6.40 v. 6.98 years).
Panel B presents summary statics for board, committee, and meeting structure. The
median board size in the sample is 9 members, while the median level of independence (fraction
of non-employee directors on the board) is 87.5%. Altering this definition and treating affiliated
directors as inside directors decreases independence by approximately 8% for the sample. Given
that the median firm in this sample has a board size of 9, this indicates that the most common
board structure by 2005 is one in which the CEO serves as the single insider (employee director)
27
Table 1.1: Summary Statistics
This table reports summary statistics for the sample of 1356 firm-year observations from 2005 and 2006. The firm
policy descriptive statistics in Panel A include: assets ($MM), book leverage, R&D intensity (R&D/Sales), acquisition
ratio (total value of acquisitions over market equity), free cash flow, business segments, firm age, ROA, and the ratio of
the market value to book value of assets (market-to-book). The governance and compensation descriptive statistics in
Panel A include: the Bebchuk, Cohen, and Ferrell (2009) entrenchment index (E-Index), institutional ownership
(aggregate), block (top blockholder), CEO salary, CEO total compensation, CEO percent ownership, CEO equity
compensation (equity compensation over total compensation), CEO tenure, CEO age, mean director ownership (mean
holdings of independent directors by firm), and director tenure by firm. Panel B presents summary statistics for the
sample board structure. The board descriptive statistics include: board size, the ratio of outsiders to board size
(independence), the fraction of board members holding three or more board seats (fraction busy), the fraction of family
boards, the size of various committees (audit, compensation, nominating/governance, executive), the independence of
the executive committee, and the number of committee positions held per director. In addition, the summary statistics
for the meeting structure of the board are also presented. These statistics include: the number of full board meetings,
the number of committee meetings (audit, compensation, nominating/governance, executive, miscellaneous investment,
outside executive sessions), the number of monitoring meetings (excluding audit, but including miscellaneous
monitoring meetings and stakeholder meetings), and the fraction of meetings held in each committee.
Panel A: Firm Statistics Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Financial & Investment Policies
Assets 10331.88 36763.02 1158.36 2733.50 7362.19
Book Leverage 0.346 0.226 0.191 0.333 0.475
R&D Intensity 0.021 0.041 0 0 0.022
Acq Ratio 0.036 0.094 0 0.002 0.025
Free Cash Flow 0.085 0.082 0.055 0.086 0.123
Segments 3.40 1.98 1 3 5
Firm Age 29.24 14.33 15 34 44
ROA 0.148 0.089 0.102 0.140 0.191
Market-to-Book 1.87 0.948 1.27 1.61 2.155
Governance & Compensation
E-Index 2.44 1.16 2 2 3
Institutional Holdings 0.811 0.146 0.74 0.841 0.941
Block 0.104 0.044 0.074 0.097 0.126
CEO Salary 878.49 356.01 645 847 1026
CEO Total Comp 6747.89 7665.97 2366 4540.97 8494
CEO Equity Comp 0.451 0.252 0.287 0.489 0.646
CEO Ownership (%) 1.30 3.50 0.095 0.265 0.765
CEO Tenure 6.40 6.18 2 5 8
CEO Age 55.77 6.61 51 56 60
Mean Director Ownership 0.214 0.748 0.008 0.026 0.087
Director Tenure 6.98 3.82 4 6 9
28
Table 1.1 (Continued)
Panel B: Board Statistics Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Board & Committee Structure
Board Size 9.67 2.12 8 9 11
Independence 0.841 0.082 0.80 0.875 0.90
Fraction Busy 0.330 0.219 0.142 0.333 0.50
Family Board 0.096 0.290 0 0 0
Audit Committee Size 3.96 1.01 3 4 5
Nom/Gov Committee Size 4.06 1.44 3 4 5
Compensation Committee Size 3.85 1.09 3 4 4
Executive Committee Size 3.89 1.43 3 4 5
Executive Committee Indep 0.641 0.231 0.60 0.667 0.80
Committee Positions Held 1.46 0.836 1 1 2
Meeting Structure
Full Board Meetings 7.98 3.46 6 7 9
Audit Committee Meetings 9.01 3.31 7 9 11
Nom/Gov Committee Meetings 3.81 1.81 3 4 5
Comp Committee Meetings 5.46 2.66 4 5 7
Monitoring Meetings 10.00 4.48 7 9 12
Executive Committee Meetings 0.66 1.96 0 0 0
Misc Inv Committee Meetings 0.24 1.14 0 0 0
Standing Executive Committee 0.398 0.49 0 0 1
Fraction Audit 0.514 0.108 0.440 0.526 0.601
Fraction Nom/Gov 0.310 0.113 0.235 0.315 0.40
Fraction Comp 0.389 0.110 0.311 0.400 0.465
Fraction Monitoring 0.531 0.117 0.461 0.542 0.616
Fraction Outside Exec Sessions 0.327 0.137 0.20 0.307 0.50
29
on the board. Next, family boards are defined as boards where two or more individuals on the
board are direct family members and one of the board members is an insider (exclusively this
amounts to one of the family members holding the CEO position). This construction includes
father-son, brother-brother, husband-wife, sister-sister, etc., boards, but excludes boards where
two or more cousins sit on the board. Family boards constitute 9% of the sample. Following this,
the average audit, nominating, compensation, and executive committees have 3.96, 4.06, 3.85,
3.89 members, respectively. This is in accordance with the past empirical work of Hayes et al.
(2004) who find that most committees have a median size of four directors. In addition, although
not detailed in the table, the frequency of lesser committees (‘miscellaneous monitoring
committees’, and ‘miscellaneous investment committees’) follows in a similar manner to Hayes
et al. (2004) and Adams (2003).
Panel B also details the meeting structure of the board. The mean (median) number of full
board meetings over this time period is 7.98 (7). This is nearly identical to the much earlier study
of Vafeas (1999) who reports that the mean (median) number of full board meetings over the
1990 to 1994 time frame was 7.45 (7). The nominating and compensation committees meet an
average of 3.81 and 5.46 times a year, respectively. Next, given the post-SOX increased scrutiny
on firm financial statements, it is not surprising that the audit committee spends more time on
their work alone (9 meetings a year) than the number of full board meetings (7 meetings a year).
Combining the nominating, compensation, stakeholder, and miscellaneous monitoring
committees (excluding the audit committee), the average firm holds 10 monitoring committee
meetings a year. Moreover, adding the number of monitoring committee meetings (10.00) to the
number of audit committee meetings (9.01) highlights that the 2005-2006 board holds over two
times the number of meetings in outsider committees as compared to full board meetings (19.01
meetings per year v. 7.98 meetings per year). Hence, since the median director holds one
30
committee position on the board (considering the four primary committees), this implies that the
average 2005-2006 board operates in a highly fragmented manner. Outside directors spend a
considerable fraction of their board-time interacting only with other members within committee
(as compared to interacting with all board members in full board meetings).
Turning to the issue of the executive committee, 40% of firms have a standing executive
committee in the sample. Yet, looking at the number of meetings held in the executive
committee, it appears to be highly skewed. The 75th percentile of executive committee meetings
held is still 0, yet the mean is 0.66 meetings a year. In fact, only 19% of firms held one or more
executive committee meetings in a given year. Though, these firms which do have executive
committee meetings hold a significant number of meetings in the executive committee, with an
average of over 3 meetings a year being held in the committee. Further, the CEO holds a position
on the executive committee in 91% of the cases where an executive committee operates within
the board. Situations where the CEO does not hold a position are almost exclusively firm-year
observations where a CEO turnover event has occurred.
Following this, it is important to summarize the fraction of meetings held in various
committees since these measures serve as central variables in this investigation. First, the mean
(median) fraction of meetings in the audit committee, Frac Audit, is 0.51 (0.52). Given the small
number of firms actually holding executive committee meetings, a similar statistic persists when
altering the construction of Frac Audit to be the number of audit committee meetings normalized
by the sum of the number of board meetings and audit committee meetings (mean of 0.53).
Similarly, the mean fraction of meetings held in the nominating committee is 0.31. The mean
fraction of meetings held in the compensation committee is 0.39. The mean fraction of meetings
held in the monitoring committees (excluding meetings held in the audit committee) is 0.53. In
addition, the mean (median) fraction of outside executive sessions is 0.33 (0.31). The 25th and
31
75th percentile for this statistic are 0.20 and 0.50, respectively. On the lower end, this indicates
that over 25% of firms are merely stating that they are in compliance with the NYSE’s
requirement regarding outside executive sessions. In fact, 35% of firms are simply stating
compliance with the ruling (or not mentioning the issue in proxy statements). On the upper end,
over 25% of firms are stating that the number of outside executive sessions matches the number
of full board meetings. With the summary statistics delineated for the sample, I next turn to the
empirical tests of the three board hypotheses.
1.3 Empirical Design
In this section, I address the cross-sectional determinants of the operational form of the board and
changes to the operational form of the board over time. First, I investigate how the three board
hypotheses relate to the internal structure of board operations in a post-SOX environment.
Following this, I detail how the operating structure of the board has changed over the pre- and
post-SOX time frame. To conclude, I summarize how board compliance in the pre-SOX
environment relates to operational changes.
1.3.1 Cross-Sectional Determinants of the Operational Form of the Board
Before explicitly testing the primary board hypotheses, I isolate one particular hypothesis, the
negotiation hypothesis, and detail in a univariate sense how it is associated with the internal
operations of the board. The negotiation hypothesis argues that the form which the board takes
follows from a bargaining process between the CEO and outside directors representing the
interests of shareholders (Hermalin and Weisbach (1998)). Past empirical research has
demonstrated that CEOs with considerable power use their position to bargain for more inside
directors on the board (Boone et al. (2007); Linck et al. (2008)). In the context of this
investigation, I conjecture that the influence of the CEO also has a material impact on the internal
32
structure of the board. If CEOs dislike the monitoring role of outside directors or desire to
control overall board decisions, then they will bargain to pull the operations of the board out of
the hands of outside directors. This entails performing a greater fraction of the board tasks in
their presence and not in independent committees where outside directors control the decision-
making processes.
Table 1.2 provides univariate support for the negotiation hypothesis. First, to mitigate the
effect that CEO turnover may have on the operations of the board, 353 CEO turnover
observations are removed from the table, leaving 1003 firm-year observations. Next, in the table,
I consider three measures of CEO power: high tenure (10 or more years as the CEO), high
ownership (greater than 1% ownership of the common shares outstanding), and family board (two
or more family members sitting on the board).
30
Before detailing the differences between high
and low power CEOs, consider the example of Maryjo Cohen, the CEO of National Presto
Industries for over 10 years and the firm’s largest blockholder. In fiscal year 2005, the firm’s
board met 6 times, the audit committee met 5 times, and the compensation and nominating
committees met once. In addition, the firm made no mention of outside executive sessions. In
essence, this serves as a prototypical example of how a high power CEO structures the oversight
operations of the board. All nominating and compensation issues are discussed in the presence of
the CEO and outside directors are in all likelihood meeting infrequently in outside executive
sessions.
First, the difference in mean board size between high tenure (ownership) and low tenure
(ownership) CEOs is -0.81 (-0.99). This difference in means (along with differences in medians
according to the Wilcoxon signed-rank test) is significant at the 5% level for both measures of
CEO power. Next, high tenure CEOs are associated with mean independence of 82%, while low
30
With the CEO turnover events removed from the sample, high tenure CEOs and high ownership CEOs both
constitute one-third of the remaining sample.
33
Table 1.2: Operational Form and CEO Power
This table reports board structure statistics for the 2005-2006 sample, partitioned by various measures of CEO power.
The three measures of CEO power are CEO tenure, CEO ownership and family boards. High CEO tenure denotes
CEOs that have held the executive position for ten or more years. High Own denotes CEOs who hold greater than one
percent of the common shares outstanding. Family board denotes boards where two or more family members sit on the
board. In addition, 353 turnover event-years are excluded from the analysis (observations where the CEO departs in
the current or prior year), leaving 1003 firm-year observations. The table presents differences for high and low CEO
power across the following variables: board size, independence, full board meetings, compensation meetings,
nominating/governance meetings, the fraction of meetings held in the compensation committee, the fraction of
meetings held in the nominating/governance committee, outsider executive sessions, the fraction of meetings held in
the executive committee, and work in exec (the fraction of firms who hold twenty-five percent of their board meetings
in the executive committee). Differences in means (and medians in brackets) denoted in bold represent statistical
significance at the 95% confidence level.
Board
Size
Indep Board
Meetings
Comp
Meetings
Nom
Meetings
Frac
Comp
Frac
Nom/Gov
Outside
Sessions
Frac
Exec
Work
in
Exec
High
Tenure
9.14 0.820 7.54 4.81 3.09 0.366 0.275 0.301 0.075 0.124
[9] [0.857] [7] [4] [3] [0.375] [0.285] [0.250] [0]
Low
Tenure
9.95 0.855 7.82 5.36 3.95 0.392 0.323 0.337 0.050 0.075
[10] [0.875] [7] [5] [4] [0.400] [0.333] [0.333] [0]
Difference -0.81 -0.035 -0.28 -0.55 -0.86 -0.026 -0.048 -0.036 0.025 0.049
[-1] [-0.018] [0] [-1] [-1] [-0.025] [-0.048] [-0.083] [0]
High Own 8.97 0.814 7.31 4.72 3.08 0.373 0.282 0.295 0.073 0.118
[9] [0.857] [6] [4] [3] [0.384] [0.294] [0.200] [0]
Low Own 9.96 0.856 7.90 5.38 3.92 0.389 0.320 0.337 0.052 0.079
[10] [0.875] [7] [5] [4] [0.400] [0.333] [0.333] [0]
Difference -0.99 -0.042 -0.59 -0.66 -0.84 -0.016 -0.038 -0.042 0.021 0.039
[-1] [-0.018] [-1] [-1] [-1] [-0.016] [-0.039] [-0.133] [0]
Family
Firm
9.77 0.771 6.71 4.61 2.94 0.379 0.288 0.286 0.074 0.105
[10] [0.777] [6] [4] [3] [0.400] [0.300] [0.200] [0]
No Family
Firm
9.72 0.857 7.89 5.29 3.81 0.386 0.314 0.332 0.054 0.085
[9] [0.875] [7] [5] [4] [0.400] [0.316] [0.333] [0]
Difference 0.05 -0.086 -1.18 -0.68 -0.87 -0.007 -0.026
-0.046 0.020 0.020
[1] [-0.098] [-1] [-1] [-1] [0] [-0.016] [-0.133] [0]
34
tenure CEOs are associated with mean independence of 85%, a difference significant at the 5%
level. Looking at the mean number of inside directors on each board (untabulated), high tenure
CEOs are associated with 1.62 insiders on the board (including themselves) and low tenure CEOs
are associated with 1.40 insiders on the board, a marginal difference.
Turning to the issue of board meetings, high power CEOs hold weakly fewer full board
meetings (significant at the 5% level when considering CEO ownership and family board as the
measures of CEO power). This is in accordance with the findings of Vafeas (1999) who
demonstrates that high power CEOs hold fewer full board meetings. Yet, when allocating
executive committee meetings to full board meetings, the difference between high and low power
CEOs becomes statistically insignificant. High power CEOs also hold far fewer compensation
and nominating committee meetings as well (significant across all measures of CEO power).
Most importantly, Table 1.2 demonstrates that high power CEOs have a lower fraction of
monitoring (compensation, nominating) meetings held outside of their presence on committees
(Frac Comp, Frac Nom/Gov). Specifically, high power CEOs are associated with a lower
fraction of meetings held in the compensation committee (significant at the 5% level when
considering CEO tenure, yet not significant at this level when considering CEO ownership or
family board as measures of CEO power). A similar result persists for the fraction of meetings
held in the nominating/governance committee (significant for all measures of CEO power which
are detailed in the table).
31
Economically, the fraction of meetings handled by outside directors
in the compensation (nominating) committees is 7% (17%) lower when considering high tenure
31
The allocation of audit committee meetings is not detailed in this table due to the assumption that in the modern
board auditing work is almost exclusively handled in committee. Further, it is difficult to reason that hypotheses such
as the negotiation hypothesis should apply to audit committee work allocation. While CEOs may strive to have greater
control over compensation issues or the election of directors, do high power CEOs really desire to control the
evaluation of financial statements given the corporate scandals of 2000-2001 and the severe penalties now associated
with financial misreporting? I operate under the assumption that auditing work will almost exclusively be dealt with in
committee and thus should have no relation to the negotiation hypothesis.
35
CEOs as compared to low tenure CEOs. Finally, high power CEOs are also associated with
fewer outside executive sessions. Across all measures of CEO power, the fraction of time spent
by independent directors in outside executive sessions is considerably lower when high power
CEOs sit on the board.
Next, if CEOs desire to control the investment decisions of the board, it should follow
that CEOs will implement policy decisions through the executive committee, side-stepping the
oversight of the full board, if given the opportunity to do so. Since, by construct, CEOs have a
greater voting stake in any decision implemented via the executive committee, this committee
functions as an alternative to full board meetings where CEOs may exercise investment decision
control and implement decisions without full outside director approval. In the final two columns
of Table 1.2, I address this conjecture. High tenure CEOs spend 33% more time in the executive
committee as compared to low tenure CEOs (0.075 v. 0.050 fraction of meetings held in the
executive committee). In addition to this measure of investment decision control, I also construct
an indicator variable which takes a value of one if a board spends over 25% of their board
meetings in the executive committee (Frac Exec > 25%). 12.4% of high tenure CEOs spend this
extreme amount of board time in the executive committee, while only 7.5% of low tenure CEOs
spend this level of time in the executive committee (significant at the 5%). Results are less
significant when considering family board as the measure of CEO power, but are significant at
the 5% level when considering CEO ownership as the measure of CEO power. Further, the
documented findings with respect to CEO power and board investment control are robust to
alternative thresholds for Frac Exec (including Frac Exec > 10%, and Frac Exec > 30%). In
addition, it is also important to note that results throughout Table 1.2 hold with the inclusion of
CEO turnover events as well.
36
Together, these findings provide supporting evidence pertaining to how high power
CEOs control board-level policies. First, Adams et al. (2005) demonstrate that high power CEOs
are associated with greater variability in firm performance. Thus, since Table 1.2 shows that high
power CEOs tend to call executive committee meetings in lieu of full board meetings, such a
result identifies the mechanism by which this variability in firm performance manifests. Through
the executive committee high power CEOs are able to enact quick and variable policy decisions,
avoiding the mediating constraint offered by outside directors. Second, Core et al. (1999) show
that CEOs who hold the board chair position demand higher cash-based and total compensation.
Hence, since high power CEOs appear to pull the compensation/monitoring operations of the
board away from committees and toward full board meetings (lower Frac Comp), this result again
identifies just how high power CEOs bargain for the form in which their compensation is granted
to them.
32
1.3.2 Determinants of Internal Monitoring Structure
While the univariate results presented in Table 1.2 lend support to the notion that high
power CEOs generally control the internal monitoring processes of the board by having outside
directors spend a greater fraction of their board-time in meetings in which the CEO has a voting
stake and not in independent committees, how do the other board hypotheses relate to the
monitoring structure of the board? In Table 1.3, I investigate this issue. A variety of monitoring
control measures are regressed on firm-level determinants used to capture the three board
hypotheses. In Columns (1) and (2) the dependent variable is Frac Comp (number of
compensation committee meetings normalized by the sum of full board meetings, executive
committee meetings, and compensation meetings). In Columns (3) and (4) the dependent
32
Though it is important to note that the regulations specify that the CEO is not allowed to be present in the boardroom
when deliberations regarding his pay are being made.
37
variable is Frac Nom/Gov (fraction of meetings held in the nominating/governance committee).
In Columns (5) and (6) the dependent variable is Frac Mon (fraction of meetings held in all
monitoring committees and stakeholder committees, excluding the audit committee). In Columns
(7) and (8) the dependent variable is Frac Sessions (fraction of board time spent in outside
executive sessions).
As discussed in previous sections, I use several firm-level controls to capture various
aspects of the three board hypotheses. Firm size, segments, and firm age are implemented as
proxies for the scope of operations hypothesis. This hypothesis predicts a positive relation
between work allocation to committees and firm complexity. Next, free cash flow (FCF), E-
Index, and R&D are used to serve as proxies for firm private benefits (FCF, E-Index) and the
costs of monitoring (R&D). The monitoring hypothesis predicts a positive association between
private benefits and the fraction of meetings held in monitoring committees, and a negative
association between the costs of monitoring and the fraction of meetings in committee. To
capture the level of CEO power, I focus on the two primary measures previously noted: CEO
ownership and CEO tenure.
33
The negotiation hypothesis predicts a negative relation between
CEO power and the fraction of time spent in monitoring committees. Finally, to control for other
factors which may influence internal board operations, I include the following variables: CEO
turnover (departure in the current or previous year), director turnover (departure in the current or
previous year), market-to-book, mean outside director ownership, industry-adjusted returns over
the prior year (adjusted by median returns in Fama-French 48 groupings), fraud/restatement
(indicator of one if fraud or a restatement was announced in the prior year), and high acq
(indicator of one if acquisitions normalized by market value were at the 75th percentile or higher
33
The log transform of CEO tenure and the log transform of CEO ownership yield qualitatively similar results
throughout the forthcoming tables.
38
in the previous year).
34
Firm-specific performance measures (industry-adjusted returns and
market-to-book) are implemented as controls to serve as proxies for CEO ability, leaving the
CEO power measures to capture the bargaining position of the CEO. In addition, all models
include time and industry fixed effects to control for underlying economic factors (either in a
given year, or specific to common market conditions) that may explain board operational
structure. Standard errors are computed using robust methods (heteroskedasticity-consistent with
clustering by firm) and p-values are denoted below coefficients in the table.
Columns (1) and (2) demonstrate a positive association between the fraction of meetings
held in the compensation committee and the firm complexity variables (firm size, segments, and
firm age), lending support to the scope of operations hypothesis. In particular, the coefficient on
firm size is positive and significant at the 1% level in the first column.
35
Moreover, considering
the coefficient on firm size, a shift from the 25th to the 75th percentile in firm total assets (within
sample) implies a 5% increase in the fraction of meetings in the compensation committee. Next,
in accordance with the monitoring hypothesis, free cash flow (FCF) is positively related to the
fraction of meetings held in the compensation committee, and R&D is negatively related to the
fraction of meetings in the committee, yet neither of the coefficients associated with these
variables are significant. In addition, both CEO tenure and CEO ownership are weakly negatively
related to the fraction of meetings in the compensation committee.
34
Data on financial restatements are taken from the U.S. Government Accountability Office (GAO). Data for the
revelation of fraud (charges brought against the firm regarding financial reporting violations) comes from a publicly
available repository of Accounting and Auditing Enforcement Releases (AAERs) issued by the SEC.
35
Note, this positive association between firm size and work allocation to committees holds despite an attenuation bias.
Separating firm size, segments, and firm age into different unique models serves to increase statistical significant for all
the variables, yet is omitted for brevity. Further, altering the dependent variable by removing executive committee
meetings from the denominator yields qualitatively identical results pertaining to significance of these measures
throughout the columns in Table 1.3.
39
Table 1.3: Determinants of Internal Monitoring Structure
The table reports results from regressing various measures of monitoring work allocation on firm-level determinants.
The sample includes 1356 firm-year observations from 2005-2006. The following firm-level variables are
implemented: firm size (log of total assets), segments (log of business segments), firm age, free cash flow, E-Index,
institutional ownership (aggregate), R&D (indicator of one if R&D expenditures over sales is at the seventy-fifth
percentile or higher), CEO tenure, CEO ownership, director ownership (average holdings of outside directors), market-
to-book, CEO turnover (departure of the CEO in the current or previous year), director turnover, industry-adjusted
returns over the prior year, fraud/restatement (indicator variable if there was an announcement of fraud or a restatement
in the current or prior year), and high acq (indicator of one if acquisitions over market equity is at the seventy-fifth
percentile or higher). The dependent variables presented are: the fraction of meetings held in the compensation
committee, the fraction of meetings held in the nominating/governance committee, the fraction of meetings held in all
monitoring committees (excluding the audit committee, but including miscellaneous monitoring committees and
stakeholder meetings), and the fraction of time spent in outside executive sessions. All regressions are estimated via
OLS, with the exception of outside executive sessions (Frac Sessions) which is estimated via tobit regressions.
Industry (Fama-French 48 classification) and year fixed effects are included in all regressions. Standard errors are
computed using robust methods (clustered by firm) and p-values are reported below coefficients in parentheses.
Frac
Comp
(1)
Frac
Comp
(2)
Frac
Nom/Gov
(3)
Frac
Nom/Gov
(4)
Frac
Mon
(5)
Frac
Mon
(6)
Frac
Sessions
(7)
Frac
Sessions
(8)
Firm Size 0.0071 0.0069 0.0092 0.0091 0.0148 0.0142 0.0074 0.0074
(0.01) (0.02) (0.00) (0.00) (0.00) (0.00) (0.07) (0.08)
Segments 0.0021 0.0023 0.0051 0.0074 0.0033 0.0048 -0.0029 -0.0030
(0.74) (0.62) (0.42) (0.26) (0.63) (0.50) (0.74) (0.73)
Firm Age 0.0004 0.0004 0.0005 0.0005 0.0007 0.0007 -0.0001 -0.0001
(0.16) (0.15) (0.08) (0.10) (0.03) (0.03) (0.81) (0.85)
FCF 0.0461 0.0393 0.0319 0.0501 0.0284 0.0307 0.0502 0.0350
(0.48) (0.55) (0.61) (0.45) (0.67) (0.65) (0.58) (0.71)
E-Index -0.0033 -0.0031 0.0019 0.0036 -0.0004 0.0005 -0.0019 0.0004
(0.37) (0.39) (0.57) (0.30) (0.91) (0.90) (0.69) (0.93)
R&D -0.0009 -0.0009 0.0012 0.0032 0.0014 0.0019 0.0582 0.0653
(0.94) (0.93) (0.91) (0.79) (0.90) (0.87) (0.00) (0.00)
CEO Tenure -0.0011 -0.0032 -0.0026 -0.0038
(0.15) (0.00) (0.00) (0.00)
CEO Own -0.0041 -0.0085 -0.0089 -0.0070
(0.08) (0.00) (0.00) (0.01)
CEO Turnover 0.0057 0.0103 -0.0193 -0.0053 -0.0097 0.0010 -0.0315 -0.0107
(0.50) (0.20) (0.05) (0.46) (0.28) (0.90) (0.03) (0.35)
Director Turnover -0.0038 -0.0041 0.0001 0.0025 -0.0038 -0.0023 0.0033 0.0085
(0.54) (0.51) (0.97) (0.69) (0.56) (0.71) (0.75) (0.41)
Director Own -0.0021 -0.0012 0.0005 0.0011 -0.0026 -0.0017 0.0022 0.0017
(0.59) (0.74) (0.88) (0.76) (0.56) (0.65) (0.72) (0.78)
Inst Own 0.0254 0.0166 0.0333 0.0150 0.0349 0.0167 0.0740 0.0611
(0.30) (0.50) (0.20) (0.57) (0.22) (0.55) (0.03) (0.07)
Market-to-Book 0.0011 0.0013 0.0020 0.0001 0.0019 0.0009 -0.0019 -0.0024
(0.80) (0.76) (0.59) (0.96) (0.67) (0.83) (0.79) (0.72)
Ind Adj Ret 0.0075 0.0101 0.0209 0.0238 0.0172 0.0217 -0.0118 -0.0106
(0.48) (0.35) (0.04) (0.02) (0.12) (0.07) (0.47) (0.52)
Fraud/Restatement 0.0071 0.0081 0.0043 0.0063 0.0098 0.0115 -0.0092 -0.0067
(0.50) (0.44) (0.65) (0.51) (0.35) (0.27) (0.51) (0.62)
High Acq 0.0065 0.0076 -0.0031 -0.0032 0.0014 0.0018 0.0010 0.0015
(0.39) (0.30) (0.67) (0.66) (0.84) (0.81) (0.93) (0.91)
N 1356 1356 1356 1356 1356 1356 1356 1356
R
2
0.0692 0.0719 0.1339 0.1392 0.1201 0.1331 0.0786 0.0776
40
Columns (3) and (4) present the results where the dependent variable is the fraction of
meetings held in the nominating committee. Similar associations persist throughout. Again, the
positive and significant coefficient on firm size implies that a shift from the 25th to the 75th
percentile in total assets yields a 6.5% increase in the fraction of time spent in the nominating
committee. While CEO tenure and CEO ownership were weakly associated with the allocation of
work to the compensation committee in Columns (1) and (2), in Columns (3) and (4) the
coefficients on these two measures of CEO power are significant at the 1% level. Following this,
in Columns (5) and (6) the dependent variable is the fraction of all monitoring meetings held in
committees (excluding audit committee meetings, but including stakeholder and miscellaneous
monitoring committee meetings).
36
The coefficients associated with firm size and firm age are
both positive and significant. A Wald test of the joint significance of the measures (all scope of
operations measures) is significant at the 1% level. Hence, firm complexity is positively related
to the allocation of monitoring tasks to committees. Next, the coefficients on CEO tenure and
CEO ownership are again negative and significant at the 1% level.
The dependent variable in Columns (7) and (8) is the fraction of time spent in outside
executive sessions. As previously noted, in the collected proxy statement data, firms appear to
report the actual number of outside executive sessions held in a given year with a lower bound of
one-fourth the number of full board meetings. In such, I have categorized those firms that do not
report, or simply state compliance with the NYSE mandate, as holding one-fourth the number of
outside executive sessions as full board meetings. This treatment creates a lower bound to the
distribution of observations. An upper bound to the distribution also exists due to the fact that
firms do not report holding more outside executive sessions than full board meetings. To control
36
Due to the relatively few times these committees meet, the removal of stakeholder and miscellaneous monitoring
committee meetings from this measure does not alter results.
41
for this issue, a tobit regression is implemented in Columns (7) and (8).
37
While the conjectures
associated with the three boards hypotheses primarily apply to the internal monitoring structure of
the board (Columns (1) - (6)), it is also of interest to see how they relate to the fraction of time
spent in outside executive sessions. Similar to previous results, firm size is positively related to
the fraction of meetings held in outside executive sessions. In addition, the coefficient on R&D is
positive and significant at the 1% level. While not directly in accordance with the monitoring
hypothesis, this result is not entirely surprising. If high R&D corresponds to high levels of inside
information, then this information which the CEO holds might also be associated with the ability
to manipulate boardroom discussions. Hence, since the CEO may have a greater opportunity to
guide board level proceedings and overstate future prospects, shareholders will demand that
independent directors meet apart from the CEO after most full board meetings to discuss whether
the CEO has in fact manipulated or falsified the performance outlook of the firm. Following this,
both coefficients on the CEO power measures are negative and significant as well.
Also of interest is the fact that, in general, control variables in Table 1.3 appear to be
insignificantly related to board monitoring structure. Turnover events (CEO or director) do not
appear to significantly alter the meeting structure on the board. Firms with high levels of
institutional ownership weakly structure the board so that a greater fraction of the meetings are
handled by outside directors in committees. Fraud accusations, restatements, and acquisitions
also seem to have little effect on the fraction of meetings held in monitoring committees. In
addition, while Vafeas (1999) demonstrates a strong negative association between full board
meetings held and performance (market-to-book), in this investigation the fraction of meetings in
monitoring committees does not appear to change over market-to-book states. In other words,
while the total number of meetings increases as firm performance falls, the number of meetings
37
All results in Columns (7) and (8) are qualitatively identical if an OLS regression is implemented as the alternative to
the tobit regression.
42
held in monitoring committees and the number of full board meetings scale together over
economic states.
In untabulated analysis, a variety of other methods are used to test the relationship
between monitoring meeting structure and the board hypotheses. First, to augment the results
presented in Columns (7) and (8), I construct two alternative measures to Frac Sessions: an
indicator variable taking a value of one if the firm has an equal number of full board meetings as
outside executive sessions, and an indicator variable taking a value of one if the firm does not
state the number of outside executive sessions. Running logit models with these two alternative
measures yields qualitatively similar results to those presented in Columns (7) and (8). In
particular, firm size is still positively associated with the fraction of time spent in outside
executive sessions, and the two measures of CEO power are both negatively associated with the
fraction of time spent in outside executive sessions (all significant at equivalent levels to results
presented in Table 1.3). Second, for Columns (1) - (6), I define threshold levels of monitoring
meeting allocation to committees using indicators at the 75th percentile for the sample. In
essence, these measures proxy for situations where outside directors control monitoring decisions.
Again, this specification does not alter the significance of the primary results presented in
Columns (1) - (6). Third, I also consider a committee-member weighted measure to determine
the fraction of meetings held in a particular committee. This amounts to scaling each number of
committee meetings held by the number of committee members in the particular committee and
the number of full board meetings by the number of members on the board. Since the number of
board members on each committee tends to move in accordance with board size (e.g. an 8 person
board has 4 members on a committee and a 10 person board has 5 members on a committee),
these alternative measures of board monitoring control function in nearly an identical fashion to
43
the measures used in Table 1.3. Results presented throughout hold in an equivalent manner if
these measures are implemented.
Following this, in regard to the monitoring hypothesis, a variety of additional models are
run to isolate the effect of managerial private benefits and monitoring costs on monitoring
structure. In these models, the firm complexity proxies and CEO power measures are removed,
leaving FCF, E-Index and R&D as the primary determinants. Equivalent weak supporting
evidence for the monitoring hypothesis is found in such models. Also, in accordance with other
past studies investigating the monitoring hypothesis, I use an alternative measure to capture firm-
level asymmetric information: the standard deviation of returns over the prior year (Linck et al.
(2008)). This measure of monitoring costs also yields qualitatively similar results to those
presented using R&D intensity (i.e. nothing of significance). Finally, alternative models where
board size and independence are implemented as control variables yield similar results throughout
Table 1.3, with the exception of the significance attached to the coefficient on firm size. Since
the single greatest determinant of board size is firm size or firm complexity (Boone et al. (2007)),
this correlation serves to diminish the significance associated with the variable. Yet, since
independence and board size are not fundamental firm-level determinants of monitoring meeting
structure on the board, and are in fact simultaneous choice variables which respond to similar
controls used in this study, their use is omitted in the preceding table.
In total, Table 1.3 demonstrates how the internal monitoring structure of the board relates
to the three board hypotheses. The results offer strong support for the scope of operations
hypothesis. Complex firms (large firms, older firms) are associated with a higher fraction of
meetings held in monitoring committees. Following this, Table 1.3 does not provide significant
support for the monitoring hypothesis. The conjecture that free cash flow and antitakeover
protection (E-Index) should be positively associated with monitoring task allocation, and that
44
R&D expenditures should be negatively related to the fraction of time spent in independent
committees does not appear to hold in any significant manner. Yet, this is not startling given
some of the inconclusive empirical results pertaining to this hypothesis in the prior literature
(Boone et al. (2007); Linck et al. (2008)).
38
Next, strong support is found for the negotiation
hypothesis. CEOs that wield considerable power appear to structure the board so that the
monitoring tasks are primarily done in their presence and not by outside directors in committees
(lower fraction of meetings held in monitoring committees and lower fraction of meetings held in
outside executive sessions).
1.3.3 Determinants of Investment/Policy Control
With the firm-level determinants of board monitoring control detailed, I now turn to the issue of
board investment control. In this section, I extend and test two of the three board hypotheses.
First, as previously discussed, if complex firms are associated with more hierarchical firm forms
(Fama and Jensen (1983)), then the fraction of meetings allocated to investment committees
should be positively associated with measures of complexity (scope of operations hypothesis).
Next, if CEOs desire to control the investment decisions of the board, then CEOs who have a
greater ability to affect the structure of their board’s operations should spend a greater amount of
time implementing decisions through the executive committee and possibly other investment
committees.
To test these conjectures, I follow a similar empirical procedure to that of the
methodology in the previous section. Table 1.4 presents a series of regressions where the
dependent variable is the fraction of meetings held in investment committees. In Columns (1)
and (2) the dependent variable is the fraction of board meetings in the executive committee
38
Boone et al. (2007) find that R&D is not negatively related to board independence (monitoring hypothesis) and is in
fact positively associated with independence. Linck et al. (2008) find an insignificant association between board size
and R&D.
45
(number of executive committee meetings normalized by the sum of full board meetings and
executive committee meetings). Since this measure is strongly skewed, for robustness I construct
an indicator variable which takes a value of one if the firm holds greater than 25% of its meetings
in the executive committee. This measure is implemented as the dependent variable in Columns
(3) and (4), and logit models are run to test its association with firm-level determinants.
39
Next,
in Columns (5) - (8) I run a similar set of tests with the single addition of miscellaneous
investment committee meetings to the dependent variable. Consequently, in Columns (5) and (6),
the fraction of meetings held in the executive/investment committees is constructed as the sum of
executive and miscellaneous investment committee meetings normalized by the sum of executive
committee meetings, miscellaneous investment committee meetings, and full board meetings. A
similar indicator variable to that used in Columns (3) and (4) follows in Columns (7) and (8). All
columns also include previously constructed control variables. In addition, although the
monitoring hypothesis is not associated with any definitive predictions in these models, I include
FCF, E-Index, and R&D as firm-level controls in all models.
In Columns (1) and (2), the firm complexity variables (firm size, segments, and age) are
all positively associated with the allocation of board meetings to the executive committee. In
particular, the coefficient on firm size is positive and significant at the 5% level (as detailed in
both of the columns). Next, the coefficients on CEO tenure and CEO ownership are also positive
and significant. The coefficient attached to CEO tenure is significant at the 5% level (as noted in
Column (1)) and the coefficient attached to CEO ownership is significant at the 10% level (as
noted in Column (2)). Also of interest is the observation that, in general, control variables are
insignificantly related to the fraction of meetings held in the executive committee. Though, higher
39
Note that in Columns (3) and (4) p-values denoted at the bottom of the columns follow from tests of model
significance (Model Chi-squared).
46
Table 1.4: Determinants of Internal Investment Control
The table reports results from regressing various measures of board/investment work allocation on firm-level
determinants. The sample includes 1356 firm-year observations from 2005-2006. The following firm-level variables
are implemented: firm size (log of total assets), segments (log of business segments), firm age, free cash flow, E-Index,
institutional ownership (aggregate), R&D (indicator of one if R&D expenditures over sales is at the seventy-fifth
percentile or higher), CEO tenure, CEO ownership, director ownership (average holdings of outside directors), market-
to-book, CEO turnover (departure of the CEO in the current or previous year), director turnover, industry-adjusted
returns over the prior year, fraud/restatement (indicator variable if there was an announcement of fraud or a restatement
in the current or prior year), and high acq (indicator of one if acquisitions over market equity is at the seventy-fifth
percentile or higher). The dependent variables presented are: the fraction of board meetings held by the CEO in
executive committee outside of the full board (Frac Exec), an indicator variable of one if the CEO holds greater than
twenty-five percent of board meetings in the executive committee (Work in Exec), the fraction of board meetings held
in the executive and investment committees (Frac Inv/Exec), and an indicator variable of one if the board holds greater
than twenty-five percent of board meetings in the executive/investment committees (Work in Inv/Exec). Regressions
for Frac Exec and Frac Inv/Exec are implemented via OLS, and regressions for Work in Exec and Work in Inv/Exec
are implemented via logit regressions. Industry (Fama-French 48 classification) and year fixed effects are included in
all regressions. Standard errors are computed using robust methods and p-values are reported below coefficients in
parentheses.
Frac
Exec
(1)
Frac
Exec
(2)
Work
in Exec
(3)
Work in
Exec
(4)
Frac
Inv/Exec
(5)
Frac
Inv/Exec
(6)
Work in
Inv/Exec
(7)
Work in
Inv/Exec
(8)
Firm Size 0.0078 0.0092 0.1291 0.1457 0.0137 0.0148 0.2140 0.2245
(0.04) (0.02) (0.10) (0.07) (0.00) (0.00) (0.00) (0.00)
Segments 0.0002 0.0002 0.0365 0.0393 0.0022 0.0002 0.0950 0.0537
(0.97) (0.96) (0.83) (0.82) (0.81) (0.95) (0.52) (0.72)
Firm Age 0.0004 0.0004 0.0198 0.0186 0.0001 0.0001 0.0010 0.0008
(0.25) (0.33) (0.04) (0.04) (0.91) (0.98) (0.88) (0.91)
FCF -0.0076 -0.0223 -1.8653 -2.0428 -0.0131 -0.0386 -1.2169 -1.8467
(0.90) (0.73) (0.30) (0.29) (0.87) (0.64) (0.40) (0.25)
E-Index -0.0001 -0.0007 -0.0713 -0.0959 0.0009 0.0002 0.0274 0.0062
(0.97) (0.87) (0.44) (0.31) (0.84) (0.95) (0.73) (0.94)
R&D -0.0057 -0.0046 0.2312 0.2782 0.0210 0.0201 0.7572 0.7632
(0.59) (0.66) (0.57) (0.49) (0.17) (0.18) (0.02) (0.02)
CEO Tenure 0.0020 0.0460 0.0018 0.0305
(0.03) (0.01) (0.08) (0.07)
CEO Own 0.0054 0.1105 0.0053 0.0632
(0.07) (0.02) (0.12) (0.12)
CEO Turnover 0.0104 0.0022 0.2108 0.0159 0.0023 -0.0058 -0.0512 -0.2367
(0.30) (0.81) (0.43) (0.94) (0.83) (0.57) (0.82) (0.28)
Director Turnover -0.0007 -0.0004 -0.0174 -0.0259 -0.0028 -0.0018 -0.1201 -0.1188
(0.91) (0.94) (0.93) (0.90) (0.73) (0.82) (0.49) (0.50)
Director Own 0.0033 0.0033 0.1093 0.1133 0.0024 0.0026 0.0670 0.0720
(0.53) (0.52) (0.27) (0.26) (0.67) (0.64) (0.50) (0.47)
Inst Own -0.0680 -0.0544 -0.5156 -0.1930 -0.0646 -0.0490 -0.3113 -0.1403
(0.07) (0.15) (0.41) (0.77) (0.09) (0.21) (0.57) (0.79)
Market-to-Book -0.0019 -0.0006 -0.1873 -0.1285 0.0046 0.0058 0.1266 0.1603
(0.67) (0.88) (0.27) (0.45) (0.45) (0.33) (0.24) (0.15)
Ind Adj Ret -0.0039 -0.0065 -0.2064 -0.2285 -0.0032 -0.0049 -0.1116 -0.1084
(0.72) (0.57) (0.54) (0.50) (0.80) (0.71) (0.68) (0.70)
Fraud/Restatement -0.0039 -0.0036 -0.0473 -0.0691 0.0045 0.0052 0.1119 0.1277
(0.74) (0.76) (0.86) (0.80) (0.75) (0.72) (0.63) (0.58)
High Acq -0.0047 -0.0030 -0.0778 -0.0379 -0.0017 -0.0003 -0.0819 -0.0289
(0.59) (0.73) (0.76) (0.88) (0.86) (0.97) (0.70) (0.89)
N 1356 1356 1356 1356 1356 1356 1356 1356
R
2
/p-value 0.0828 0.0876 0.0001 0.0001 0.0820 0.0858 0.0001 0.0001
47
levels of institutional ownership appears to be weakly associated with a lower fraction of board
meetings handled in the executive committee. The results associated with the logit models in
Columns (3) and (4) provide similar results. Again, the findings support the notion that high
power CEOs are associated with greater control over investment decisions (higher fraction of
meetings in the executive committee).
Next, in Columns (5) - (8), the dependent variable is altered to include miscellaneous
investment committee meetings. First, firm size is positively related to the fraction of meetings in
the investment/executive committees. Next, the coefficients on CEO tenure and CEO ownership
are positive, though not at the same level of significance as the coefficients on these measures in
Columns (1) - (4). This result highlights that CEOs of high power desire to make investment
decisions for the full board through the executive committee, though are less likely to exercise
control through alternative investment committees (e.g. strategy, acquisitions, etc.).
40
Further,
results presented in Table 1.4 are qualitatively similar if finance committee meetings are also
included in ‘miscellaneous investment committee meetings’ and if alternative thresholds of work
allocation are used in Columns (3), (4), (7), and (8).
The results presented in Table 1.4 again lend support to the scope of operations and
negotiation hypotheses. Firm complexity (firm size) and CEO power are both positively related
to the fraction of meetings held in the executive committee. In particular, CEOs that have the
ability to dictate the operational form of the board (high tenure, high ownership CEOs), appear to
avoid the oversight provided by outside directors in full board meetings, and spend a greater
fraction of time making policy decisions in the executive committee. The results highlight how
40
This finding is not surprising if we consider the fact that CEOs hold such miscellaneous investment committee
positions in only one-quarter of the cases. Further models constructed to properly remove these miscellaneous
investment committee meetings in which the CEO does not hold a position on the committee from the numerator of the
measure serves to increase the statistical significance on the CEO power measures (coefficients positive and significant
at the 10% level).
48
powerful CEOs, despite being subject to outsider dominated boards, structure the internal
investment operations of the board in their favor.
1.3.4 Changes in the Internal Operations of the Board
Over the past 60 years, the composition of the board of directors has changed in a significant
manner. Hermalin and Weisbach (1988) highlight that outsider representation increased from
50% to 66% over the 1971 to 1983 time period for a sample of 142 NYSE firms. Lehn et al.
(2005) study the evolution of 81 firms over time and note that independence increased from 50%
in 1945 to 83% by 2000. In addition, Coles et al. (2008) highlight that the median fraction of
insiders on the board had decreased to 20% over the 1990s. If the vast majority of boards were
already ‘outsider-dominated’ by 1999, did the corporate scandals of 2000-2002 and the
subsequent regulatory changes which followed have any material impact on CEO influence
within the board? Further, if institutional investors did desire to push for greater board oversight
how did it manifest itself given that the average board already exhibited such a high degree of
‘nominal independence’ (80% outsider representation)? In this section, I detail just how the
board materially changed over this period.
41
To investigate the issue, I take the original set of board observations from 2005 and
match it to a sample of NYSE firms in 1999 with available board and proxy information, yielding
586 firm observations. Table 1.5 presents the differences between various board statistics over
the 1999-2005 time period. Included in Panel A of Table 1.5 are board size, independence, full
board meetings (including executive committee meetings), fraction of meetings held in the
compensation committee, fraction of meetings held in the nominating committee, fraction of
meetings held in miscellaneous monitoring committees, fraction of time spent in the executive
41
While the purpose of this investigation is not to go into great detail concerning the issue of changes in directors
characteristics on the board, consistent with past studies the primary change (between 1999 and 2005) in the type of
director who served on the board was the reduction in the number of CEOs from other firms serving as directors, and
the expansion of ‘financial experts’ serving as directors (Linck et al. (2009)).
49
committee, work in executive committee (indicator of one if the fraction of meetings held in the
executive committee is greater than 25%), and the fraction of firms holding one or more meetings
in the executive committee (Meet in Exec).
In Panel A, I first detail differences over the time period using an unadjusted matched
sample. Since NYSE firms in 1999 were not required to have independent monitoring
committees, CEOs and insiders could sit on such committees. In fact, for this matched sample,
20% of firms in 1999 had an insider or the CEO sitting on the nominating committee.
In addition, 34% of firms had no nominating committee in place. Hence, for 54% of
firms in the sample the CEO or an insider had a voting stake in the board composition discussion.
In addition, just over 5% of firms had an insider or the CEO sitting on the compensation
committee. For this unadjusted matched sample, I simply treat such situations as though no
insider was sitting on the monitoring committee.
The mean (median) board size in 2005 was 9.73 (9), while the mean (median) board size
in 1999 was 9.75 (9). This amounts to an insignificant difference of 0.02 (0). Independence
averaged 83.1% in 2005 and 79.0% in 1999. This yields a difference of 4.1% (significant at the
5% level). Altering the constructed definition of independence and treating affiliated directors as
non-independent members yields a similar 6% increase in the measure over the time period.
42
While this difference is significant, the overall economic magnitude is still quite small. Since
board size remained the same over the time period, the change in independence reflects the
removal of less than one-half of one insider from the board and the addition of less than one-half
of one outsider to fill the position. For the unadjusted matched sample, the number of board
meetings controlled by the CEO (full board meetings plus executive committee meetings) in 1999
42
Further, using this altered definition of independence, 87% of firms had a majority of independent directors on the
board in 1999. Using a much larger dataset, Duchin et al. (2010) show that 76% of firms had a majority of independent
directors on the board in 2000, and the level of independence increased by 10% between 2000 and 2005. Linck et al.
(2009) document a 7% increase in board independence over this time period.
50
was 8.54 and 8.59 in 2005. Next, although not directly noted in the table, the mean number of
executive committee meetings held in 1999 was 1.20, far above the mean level of 0.67 in 2005.
Following this, for the unadjusted matched sample, the fraction of meetings held outside
the CEO’s voting influence in the monitoring committees increased significantly between 1999
and 2005. The mean (median) fraction of time spent in the compensation committee was 0.321
(0.333) for the 1999 sample and 0.390 (0.400) for the 2005 sample (differences significant at the
5% level). The mean (median) fraction of meetings held in the nominating committee was 0.134
(0.111) for the 1999 sample and 0.313 (0.333) for the 2005 sample. This amounts to a difference
in means (medians) of 0.179 (0.222), significant at the 5% level. Similarly, the difference in
means (medians) for the time spent in the audit committee is 0.219 (0.233), significant at the 5%
level. In addition, the fraction of time spent in the executive committee was nearly cut in half
between 1999 and 2005. In 1999 the mean fraction of meetings held in the executive committee
was 0.093, and by 2005 it was 0.055. This constitutes a 40% drop in the measure between 1999
and 2005 (significant at the 5% level). A similar significant difference persists when considering
boards which do over 25% of their board work in the executive committee (Work in Exec
Committee).
To properly adjust for the presence of CEOs on committees in 1999, I create an adjusted
matched sample. If the CEO sits on a particular monitoring committee (compensation, audit or
nominating), any meetings that such a committee holds are now treated as ‘full board meetings’,
or in other words, meetings which the CEO presides over. It is important to note that this
adjustment is very conservative. Since, the CEO may exercise control over the committees in
alternative manners in 1999 such as placing inside directors or affiliated directors to committee
positions, this treatment is only isolating cases where the CEO definitely votes on decisions. Yet,
though constructed in a conservative manner, this adjusted match sample gives a more accurate
51
Table 1.5: Changes in Board Structure Over Time
This table reports board structure statistics for a matched sample of firms. The sample comprises 586 firms from the
original 2005-2006 data set which have available board and financial data for fiscal years 1999 and 2005. The full
matched sample (unadjusted) includes all matched firms regardless of CEO or insider positions on monitoring
committees. The unadjusted matched sample makes no alteration to firm-year observations where the CEO holds a
committee position. In addition to the unadjusted full matched sample, the table also presents board structure statistics
for an adjusted match sample. If a CEO holds a particular monitoring committee position and presides over meetings
in a given year, then such committee meeting observations are treated as full board meeting observations. The
variables presented below are the following: board size, independence, full board meetings (including executive
committee meetings), the fraction of compensation committee meetings (Frac Comp), the fraction of
nominating/governance committee meetings (Frac Nom/Gov), the fraction of audit committee meetings (Frac Audit),
the fraction of miscellaneous monitoring meetings (including stakeholder committee meetings), the fraction of
meetings in the executive committee (Frac Exec), Work in Exec Com (an indicator variable of one if the board holds
greater than twenty-five percent of board meetings in the executive committee), and the fraction of firms which hold
one or more executive committee meetings in a given year (Meet Exec Com). Board meetings include executive
committee meetings. Panel B presents board structure statistics for a matched sample of firms where the percentage
difference in a firm’s market-to-book ratio over the period (between 1999 and 2005) is less than 40%. This controlled
sample has 391 observations. Differences in means (and medians in brackets) denoted in bold represent statistical
significance at the 95% confidence level.
Panel A:
Matched
Sample
Board
Size
Indep Board
Meetgs
Frac
Comp
Frac
Nom/Gov
Frac
Audit
Frac
Misc
Mon
Frac
Exec
Work
in
Exec
Com
Meet
Exec
Com
Unadjusted
Matched
Sample
2005
Sample
9.73 0.831 8.59 0.390 0.313 0.520 0.061 0.055 0.087 0.195
[9] [0.875] [8] [0.400] [0.333] [0.533] [0] [0]
1999
Sample
9.75 0.790 8.54 0.321 0.134 0.301 0.046 0.093 0.143 0.318
[9] [0.818] [8] [0.333] [0.111] [0.300] [0] [0]
Difference
Btwn
Periods
-0.02 0.041 0.05 0.069 0.179 0.219 0.015 -0.038 -0.056 -0.123
[0] [0.057] [0] [0.067] [0.222] [0.233] [0] [0]
Adjusted
Matched
Sample
2005
Sample
9.73 0.831 8.59 0.390 0.313 0.520 0.061 0.055 0.087 0.195
[9] [0.875] [8] [0.400] [0.333] [0.533] [0] [0]
1999
Sample
9.75 0.790 9.10 0.302 0.100 0.289 0.045 0.089 0.141 0.318
[9] [0.818] [8] [0.313] [0] [0.286] [0] [0]
Difference
Btwn
Periods
-0.02 0.041 -0.51 0.088 0.213 0.231 0.016 -0.034 -0.054 -0.123
[0] [0.057] [0] [0.087] [0.333] [0.247] [0] [0]
52
Table 1.5 (Continued)
Panel B:
Market-to-
Book
Controlled
Sample
Board
Size
Indep Board
Meetgs
Frac
Comp
Frac
Nom/Gov
Frac
Audit
Frac
Misc
Mon
Frac
Exec
Work
in
Exec
Com
Meet
Exec
Com
Unadjusted
Matched
Sample
2005
Sample
9.78 0.830 8.51 0.391 0.320 0.518 0.061 0.058 0.093 0.210
[10] [0.875] [8] [0.400] [0.333] [0.529] [0] [0]
1999
Sample
9.76 0.794 8.33 0.328 0.142 0.304 0.045 0.092 0.148 0.331
[10] [0.818] [8] [0.333] [0.125] [0.300] [0] [0]
Difference
Btwn
Periods
0.02 0.036 0.18 0.063 0.178 0.214 0.016 -0.034 -0.055 -0.121
[0] [0.057] [0] [0.067] [0.208] [0.229] [0] [0]
Adjusted
Matched
Sample
2005
Sample
9.78 0.830 8.51 0.391 0.320 0.518 0.061 0.058 0.093 0.210
[10] [0.875] [8] [0.400] [0.333] [0.529] [0] [0]
1999
Sample
9.76 0.794 8.89 0.308 0.104 0.291 0.043 0.088 0.143 0.331
[10] [0.818] [8] [0.333] [0] [0.286] [0] [0]
Difference
Btwn
Periods
0.02 0.036 -0.38 0.083 0.216 0.227 0.018 -0.030 -0.050 -0.121
[0] [0.057] [0] [0.067] [0.333] [0.243] [0] [0]
53
representation of just how control over the monitoring decision-making processes changed over
this time period. First, the mean number of meetings in which the CEO had a voting stake
decreased from 9.10 to 8.59 between 1999 and 2005, significant at the 5% level.
43
Next, the
mean number of meetings held by outside directors in the independent audit, compensation, and
nominating committees increased from 3.48, 3.98, 1.11 to 9.15, 5.45, 3.82, respectively between
1999 and 2005 (all differences significant). Adding these monitoring committee meetings
together, the CEO had a voting stake in more meetings (9.10) than meetings held by outside
directors in independent committees (8.56) in 1999. In contrast, by 2005, the CEO had a voting
stake in 8.59 meetings while outside directors held 18.42 meetings in the three primary
independent monitoring committees.
Next, across all monitoring measures previously detailed, similar, yet slightly more
significant results hold given the adjusted matched sample. The mean (median) difference in the
fraction of meetings in the compensation committee over this time period is 0.088 (0.087). This
amounts to a 30% increase in the time spent discussing compensation issues outside of meetings
in which the CEO has a voting stake. Similarly, the average fraction of time spent in the
nominating committee increased over 200% (mean difference of 0.213), and the average fraction
of meetings in the audit committee increased 80% (mean difference of 0.231). The final three
columns of the adjusted matched sample also provide supporting evidence pertaining to the
decline in the CEO’s ability to exercise control over board investment decisions through the
executive committee. The fraction of the meetings held in the executive committee fell by 0.034
(38%) between 1999 and 2005. In addition, noted in the final column of the table, in 1999 31.8%
43
Note, further adjusting this measure by weighting each board/committee meeting observation by the voting stake of
the CEO in the particular meeting serves to increase the spread between meetings controlled in 1999 and 2005 (since
by construct the CEO has greater voting power in any given committee meeting).
54
of firms held at least one executive committee meeting, while in 2005 only 19.5% of firms held at
least one executive committee meeting (a significant difference).
For robustness, in Panel B of Table 1.5 I also examine a matched sample of firms where
the percent difference in a given firm’s market-to-book ratio over the period (between 1999 and
2005) is less than 40%. This matched sample is constructed to control for possible differences in
meeting structure which may result purely from differences in firm performance states between
1999 and 2005.
44
This new market-to-book matched sample includes 391 observations. The
results presented in Panel B follow in a qualitatively identical manner to those detailed in Panel
A. For the new unadjusted matched sample, the average fraction of meetings in the
compensation, nominating, and audit committees increased by 0.063, 0.178, 0.214, respectively,
between 1999 and 2005 (all significant at the 5% level). In addition, the fraction of time spent in
the executive committee decreased by 0.034 over this period (significant at the 5% level).
Further robustness checks to partition the starting sample by firm size (at the median level of firm
total assets in 1999), also yields equivalent changes in the meeting structure of the board over the
time period - both small and large firms underwent this drastic shift in internal board operations.
Moreover, given the cross-sectional determinants of the operational form of the board
presented in Tables 1.2-1.4, it is also important to detail how differences in firm characteristics
over the time period cannot explain the documented changes in internal board control in Table
1.5. Table 1.6 presents changes in CEO and firm attributes for the matched sample over the
1999-2005 period. First, the median market-to-book ratio for the matched sample weakly
increased from 1.55 to 1.58 between 1999 and 2005, while median CEO tenure remained at 5
44
Though, as previously demonstrated in Table 1.3, market-to-book appears to have no significant relation to
monitoring work allocation.
55
years over the time period.
45
In addition, the fraction of shares held by the CEO also did not
change in a material fashion, with median ownership decreasing from 0.29% to 0.26% between
1999 and 2005. Firm size (total assets) did increase in a statistically significant manner over this
period, with the average total assets amounting to 6798 ($MM) in 1999 and 9949 ($MM) in 2005.
Yet, consider the actual economic impact that this 46% change in total assets would have on the
fraction of meetings held in committees. Implementing the cross-sectional results in Table 1.3,
such a shift in total assets would imply an approximate 1% increase in the fraction of meetings in
the compensation committee, and an approximate 1.2% increase in the fraction of time spent in
the nominating committee (relative to median levels). These two implied changes are nowhere
near the magnitude of the changes which took place between 1999 and 2005, and hence the
increase in firm total assets may at best supplement the documented trends. Similar differences in
CEO and firm characteristics persist across the market-to-book controlled sample.
In Panel B of Table 1.6 I document changes in miscellaneous investment related features
of the board. First, the mean fraction of meetings held in the finance committee weakly increased
from 0.053 to 0.063, and the mean fraction of meetings held in miscellaneous investment
committees weakly increased from 0.011 to 0.018 (insignificant). Though the fraction of board-
time spent in these committees did not change in a material manner, the composition of both sets
of committees did shift in a significant manner. Outsider representation on the finance
committee, and miscellaneous investment committees increased 0.098 and 0.081, respectively.
Hence, allocating such observations where the CEO holds positions on these committees to the
number of meetings which the CEO has a voting stake in (sum of full board meetings, executive
committee meetings, and meetings of miscellaneous investment/finance committees where the
45
Average CEO tenure decreased from 6.79 years to 6.15 years between 1999 and 2005, while average market-to-book
decreased from 2.02 to 1.84 over the period.
56
Table 1.6: Changes in Firm, CEO and Board Characteristics
This table reports differences in firm, CEO and board statistics between 1999 and 2005 for a matched sample of firms.
The sample comprises 586 firms from the original 2005-2006 data set which have available board and financial data for
fiscal years 1999 and 2005. The variables presented below in Panel A for the full matched sample are the following:
total assets, market-to-book, CEO tenure, and CEO ownership. Panel B presents changes in board structure statistics
for the unadjusted matched sample which include: the fraction of meetings held in miscellaneous investment
committees, the fraction of meetings held in the finance committee, miscellaneous investment committee
independence, finance committee independence, and standing executive committee (the fraction of firms which have a
standing executive committee). Differences in means (and medians in brackets) denoted in bold represent statistical
significance at the 95% confidence level.
Panel A:
Firm/CEO
Characteristics
Assets Market-
to-Book
CEO
Tenure
CEO
Ownership
2005 Sample 9949.58 1.84 6.15 1.61
[2650.41] [1.58] [5] [0.258]
1999 Sample 6798.27 2.02 6.79 2.12
[1628.27] [1.55] [5] [0.291]
Difference Btwn
Periods
3151.31 -0.18 -0.64 -0.51
[1022.14] [0.03] [0] [-0.033]
Panel B:
Misc
Committee
Structure
Frac Misc
Investment
Frac
Finance
Misc
Investment
Independence
Finance
Independence
Standing
Exec
2005 Sample 0.018 0.063 0.762 0.791 0.402
[0] [0] [0.75] [0.80]
1999 Sample 0.011 0.053 0.681 0.693 0.415
[0] [0] [0.67] [0.71]
Difference
Btwn Periods
0.007 0.010 0.081 0.098 -0.013
[0] [0] [0.08] [0.09]
57
CEO holds a seat) serves to increase the difference in the measure to -0.60 (as compared to the -
0.51 difference presented in Table 1.5).
In total, the matched sample results in Table 1.5 provide evidence that not only did
outside board members work more following the regulatory changes (Linck et al. (2009)), but
more importantly, the internal operations of the board shifted in such a manner where CEO
participation on the board was abated. In addition to the CEO presiding over fewer board
meetings, the fraction of time spent by outside directors discussing monitoring issues removed
from the influence of the CEO (outside of meetings in which the CEO has a voting stake) also
increased dramatically during this period.
46
Following this, the decrease in the fraction of
meetings held in the executive committee lends further support to the notion that the CEO’s
ability to exercise control over investment decisions outside of the full board was significantly
curtailed between 1999 and 2005.
To examine additional board time trends, in Table 1.7 I investigate whether firms which
were non-compliant with NYSE rulings in 1999 were still associated with board differences by
2005. If the 2003 NYSE listing rulings only mandated the independence of monitoring
committees, did non-compliant firms in 1999 simply shift operations so that outside directors
spent less time in such committees by 2005? To test this idea, I partition my matched sample by
NYSE compliance. First, 54% of firms in the sample were not in compliance with NYSE rulings
regarding the structure of the nominating committee - 34% of firms did not have such a
committee and 20% of firms had insiders serving on the committee. Taking this partitioned
sample and looking at differences in board form between compliant and non-compliant firms in
2005 demonstrates that the two sub-samples were moderately different in terms of internal
operations by 2005. Compliant firms had a mean (median) fraction of meetings held in the
46
Further, while only 2% of firms reported having outside executive sessions in 1999, the average firm held 3.5
meetings in independent executive sessions in 2005.
58
Table 1.7: Compliant v. Non-Compliant Firms
This table reports board structure statistics for a matched sample of firms. The sample comprises 586 firms from the
original data set which have available board and financial data for fiscal years 1999 and 2005. The statistics below
detailed board structure for the sample in 2005 partitioned by the firm’s compliance to the NYSE rulings in 1999.
Comp Non-Compliance denotes firms where the CEO or an insider held a compensation committee position in 1999.
Nom/Gov Non-Compliance denotes firms where the CEO or an insider held a nominating/governance committee
position or such a committee did not exist in 1999. The variables presented below are the following: full board
meetings (including executive committee meetings), the fraction of compensation committee meetings (Frac Comp),
and the fraction of nominating/governance committee meetings (Frac Nom/Gov). In addition to the full matched
sample, the table also presents board structure statistics for a matched sample of firms where the percent difference in a
firm’s market-to-book ratio over the period (between 1999 and 2005) is less than 40%. This controlled sample has 391
observations. Differences in bold represent statistical significance at the 95% confidence level.
Comp Non-
Compliance
Nom/Gov
Non-
Compliance
Board
Meetings
Frac Comp Frac
Nom/Gov
Board
Meetings
Frac Comp Frac
Nom/Gov
Matched
Sample
Compliant 8.55 0.390 0.313 8.54 0.390 0.323
[8] [0.400] [0.333] [8] [0.400] [0.333]
Not Compliant 9.70 0.381 0.301 8.65 0.389 0.300
[9] [0.400] [0.301] [8] [0.400] [0.300]
Difference -1.15 0.009 0.012 -0.11 0.001 0.023
[-1] [0] [0.022] [0] [0] [0.033]
Market-to-Book
Matched
Sample
Compliant 8.46 0.392 0.320 8.26 0.393 0.337
[8] [0.400] [0.333] [7] [0.400] [0.333]
Not Compliant 10.02 0.369 0.322 8.73 0.390 0.304
[9] [0.400] [0.333] [8] [0.400] [0.310]
Difference -1.56 0.023 -0.002 -0.47 0.003 0.033
[-1] [0] [0] [-1] [0] [0.023]
59
nominating committee of 0.323 (0.333). Non-compliant firms had a mean (median) fraction of
meetings held in the nominating committee of 0.300 (0.300). This mean (median) difference of
0.023 (0.033) is significant at the 5% level.
Next, Table 1.7 also presents the results for compensation committee compliance. Just
over 5% of firms in the matched sample were not in compliance with NYSE rulings regarding
compensation committee independence in 1999. The resulting differences between the compliant
and non-compliant firms are negligible. Table 1.7 serves to provide some evidence that non-
compliant firms did not fully adjust to the intentions of the NYSE mandates (i.e. non-compliant
firms were still not equivalent to compliant firms in terms of board and committee meeting
structure by 2005).
47
In particular, Table 1.7 demonstrates that although the NYSE rulings
required the complete independence of nominating committees, non-compliant firms were still
structured to have a greater fraction of the nominating work handled in the presence of the CEO
by 2005.
1.4 Conclusion
This paper functions to extend our understanding of board control beyond independence. In
particular, using a hand-collected dataset of 586 NYSE firms, I provide supporting evidence that
the primary change in the structure of the board between 1999 and 2005 was the reduction in the
CEO’s influence and control over the decision-making processes of the board. Over this time
period, not only did the CEO have a voting stake in fewer board meetings, but more importantly,
the fraction of meetings held by outside directors (removed from the CEO’s voting influence) in
the nominating, audit, and compensation committees increased 200, 80, and 30 percent,
47
Though, it is clearly evident that non-compliant firms did change to the greater degree over this time period. Since
non-compliant firms started with all monitoring meetings (nominating or compensation) being presided over by
insiders and shifted to a level of 0.31 (Frac Nom), and 0.40 (Frac Comp), this does highlight that these firms underwent
the greatest transformation in terms of board operational form.
60
respectively. In addition, the fraction of board meetings held by the CEO in the executive
committee decreased by 40%. Hence, while board independence increased 5% between 1999 and
2005, the empirical findings lend support to the notion that the principal governance reform
following the corporate malfeasance scandals/regulatory events of 2000-2003 was through an
alternative channel of ‘independence’ on the board - the internal control over board monitoring
and investment operations.
Following this, to provide greater clarity on how the internal operations of the board are
related to the bargaining position of the CEO and other firm-level determinants, I extend and test
three primary hypotheses in the boards literature: the scope of operations hypothesis, the
monitoring hypothesis, and the negotiation hypothesis. Consistent with past empirical work, I
find strongest support for the scope of operations hypothesis and the negotiation hypothesis.
First, complex firms (large and older firms) allocate a greater percentage of their monitoring and
investment tasks to be performed in committees, outside of full board meetings. Next, CEOs with
considerable power over the operations of the board are associated with a lower fraction of
meetings handled in independent monitoring committees (removed from the CEO). With respect
to board investment decisions, high power CEOs are also associated with a greater fraction of
meetings held in the executive committee. Together these results offer support for the idea that
CEOs who have a greater ability to affect the structure of the board will pull the monitoring
operations of the board away from independent committees (meetings which they cannot control),
and will also pull the investment operations of the board toward the executive committee, thus
avoiding the scrutiny of the full board as it pertains to policy decisions. Further, in conjunction,
these two points suggest that the documented changes in board operational form between 1999
and 2005 were contrary to the preferences of CEOs.
61
The empirical results in this investigation lend new insight into a primarily unexplored
area of board structure. In a post-SOX environment, where all boards are ‘dominated’ by outside
directors and 87% independence is the norm, the findings presented here provide evidence that
the internal operations of the board are an important structural feature to consider when
discussing issues concerning board governance. In addition, the control over the internal
decision-making processes may offer a more complete picture of ‘independence’ on the board.
Moreover, the results have strong implications for the validity of prior theoretical
investigations into the determinants of board structure, as they apply to the modern board. For
instance, the basis for our understanding of optimal board size follows from the argument that
the trade-off between the extra monitoring ability (expertise) of additional outside directors and
the free-riding (moral hazard) problems associated with additional members defines the size of
the board (Raheja (2005); Harris and Raviv (2008)). Yet, such a specification of the determinants
of board size assumes that board members monitor the CEO in a full board meeting structure (i.e.
where the CEO and all board members interact in a particular meeting). This presumption ignores
the internal (endogenous) decision to allocate work to committees, which may foil or nullify any
‘free-riding’ explanation. While such a trade-off theory aptly applies to the board of the 1960s-
1990s, its use with regard to the determinants of the modern board may not be suitable. Given
that the median NYSE board now holds over 18 meetings in individual monitoring committees
and only 7 full board meetings in a given year, the assumption that each director continuously
interacts with all other directors within the boardroom is no longer valid, especially for
large/complex firms. Future theoretical research into the determinants of board structure cannot
ignore the inherent option of work allocation on the board - to do so would only produce a theory
of a board meeting and not a comprehensive theory of the modern board.
62
Thus, if the average NYSE board in 2005 operates in a highly fragmented manner, where
the division of board tasks to committees is prevalent, this implies that board-wide empirical
measures of director characteristics may no longer be appropriate for investigations into board
policies and performance. If most board members now spend a significant fraction of their board-
time interacting only with other members on a given committee, then board-level measures of
director attributes (i.e. all inclusive measures of director differences, preferences and biases) will
not accurately capture the interactions among board members which take place in the boardroom,
and yield faulty conclusions when used in empirical studies. Hence, if most board work is now
handled at the committee level, the structure and composition of distinct committees is as crucial
a feature to understand as the structure of the overall board of directors.
63
Chapter 2
Monitoring the Monitors
2.1 Introduction
While past research has documented that outside directors have strong incentives to represent the
interests of shareholders (Gilson (1990); Yermack (2004); Srinivasan (2005)), one would be hard-
pressed to challenge the notion that directors still function as partially aligned agents of the firm
who desire to retain their board seats, and may undertake measures to secure their positions.
48
If
directors derive reputational and monetary benefits from their board positions, and are ultimately
responsible for the formation and composition of the board, then what actions can the board
collectively take to insulate themselves from turnover?
49
Further, if certain boards do implement
measures to provide greater job security to their members, does this lower risk of removal affect
the type and quality of decisions made by the board? To address these issues, I focus on how the
board structures its operations to monitor and evaluate the productivity of its own members. In
particular, I examine how the size of the nominating committee and the cross-membership of
48
Yermack (2004) documents that incentive mechanisms borne upon by compensation, replacement and the
opportunity to secure other directorships provide directors with wealth increases of 11 cents per 1000 dollar rise in firm
value. Fama and Jensen (1983) conjecture that poor monitoring will result in decreased future job positions for
directors. Moreover, see Gilson (1990), Kaplan and Reishus (1990), Harford (2003), Srinivasan (2005), and Fich and
Shivdasani (2007) for further evidence validating the costs of poor monitoring to directors.
49
Bebchuk (2003) documents that incumbent directors face very little risk of being removed via the ballot box, and
hence, the governance and composition of the board is primarily determined by the board itself.
64
directors between the nominating committee and other monitoring committees affect the effort
level of directors and the self-evaluation process of the board.
If the role of the nominating committee is to make recommendations regarding board
composition, then how does such a committee properly function given the task of self-
assessment? Anecdotal evidence supports such a concerned inquiry: Charles Elson notes that
given the multitude of functions a board must perform, ‘Another problem is the issue of board
members effectively evaluating themselves.’ William Allen echoes these sentiments with the
remark that, ‘The hardest thing to ask a director to do is to assess the productivity or contributions
of other board members.’
50
If having directors effectively evaluate their own performance
constitutes a difficult assignment, it stands to reason that the challenges associated with self-
assessment should increase in proportion to the size of the nominating committee. In other
words, if directors value their board positions, the greater the number of outside directors serving
on the nominating committee, the lower the likelihood that a director will be removed from the
board.
To investigate this conjecture, I analyze the relation between outside director turnover
and board sub-structure for a set of post-Sarbanes-Oxley NYSE firms over the 2004 to 2008 time
period. Following the Sarbanes-Oxley Act of 2002 (SOX), the NYSE altered its listing
requirements by mandating the complete independence of the audit, compensation, and
nominating/governance committees. While the intention of this ruling was to improve the
governance standards of the board, it also allows for the construction of a ‘clean’ dataset, where
the influence and effect of outside directors can be isolated from that of inside directors.
Controlling for firm and board level determinants demonstrated to have an effect on director
turnover (Denis and Sarin (1999); Yermack (2004); Fich and Shivdasani (2007); Ertimur et al.
50
See ‘Emerging Trends in Corporate Governance’, a supplement to Corporate Board Member, 2001.
65
(2010)), I document that boards with small nominating committees (fewer than 60% of directors
on the board holding the position) have 25% higher director turnover rates as compared to boards
with large nominating committees. In addition, examining outside director turnover at the audit
(compensation) committee level, I observe that high cross-membership between the nominating
and audit (compensation) committees is associated with significantly lower levels of turnover as
well.
Next, considering a variety of performance measures used in the prior literature,
including financial restatements (Srinivasan (2005)), compensation granted to the CEO (Hartzell
et al. (2004)), attendance standards (Cai et al. (2009)), and firm performance (Huson et al.
(2001)), I find that large nominating committees (high cross-membership between the nominating
and other monitoring committees) exhibit a lower sensitivity of outside director turnover to
performance, yet not significant at a conclusive level. The predominant result remains that large
nominating committees serve as insulating devices for outside directors (lower risk of removal)
over all performance-states.
Moreover, if large nominating committees are associated with lower rates of director
turnover, does this translate in any manner to the type of decisions that are implemented by such
boards? If the role of the board is to act on behalf of shareholders, constraining the self-interested
actions of management, then does the greater level of board-position security provided by large
nominating committees lead to lower effort (quality of decisions) throughout the board? To
examine this issue, I analyze how nominating committee size and the cross-membership between
the nominating committee and other monitoring committees relate to board-level governance,
auditing, and compensation decisions. First, extending the past literature on the relation between
governance and CEO compensation (Core et al. (1999)), boards with high cross-membership
between the nominating committee and the compensation committee appear to weakly allocate
66
lower levels of equity-based compensation to the CEO. Yet, I find no significant evidence that
boards with high cross-membership between the nominating committee and the compensation
committee pay higher levels of cash-based compensation to the CEO. Next, addressing the
quality of decisions made by the board with regard to financial reporting, I observe no strong
association between the probability of a restatement being issued and the cross-membership
between the nominating committee and audit committee.
51
Finally, turning to the issue of charter
provisions and governance, I test whether boards with large nominating committees are less likely
to bargain with the CEO to remove antitakeover provisions (e.g. poison pills, classified boards).
52
Again, I find no significant results which would indicate that large nominating committees are
associated with lower levels of effort (lower incidence of provision removal) when representing
shareholders’ interests.
The notion that board policy and monitoring decisions are a function of not only the
composition of the board as a whole, but also of the structure of the board’s committees has been
asserted by a variety of researchers. Kesner (1988) argues that the most important board
decisions originate at the committee level, while Vance (1983) maintains that corporate decisions
are primarily influenced by four board committees: audit, executive, compensation, and
nominating.
53
Recently, this idea has been empirically advanced by numerous authors. Newman
and Mozes (1999) provide evidence that boards structured with a greater fraction of insiders on
the compensation committee allocate higher levels of compensation to the CEO. Klein (2002)
51
Additional tests investigating how the size of the nominating committee relates to accruals quality (Dechow and
Dichev (2002)) and fraud allegations also yield no significant results.
52
See Cai et al. (2009) for a similar examination into how lower shareholder votes affect directors’ incentives to
remove value destroying provisions.
53
Anecdotal support for these ideas comes from a director interviewed by Lorsch and MacIver (1989, p. 59): ‘In my
experience, I have observed that the work of the board is done in committees.’ In addition, the American Bar
Association's Corporate Director’s Guidebook (1994) emphasizes that the nominating, audit, and compensation
committees are the primary committees in which directors carry out their monitoring duties (see the Committee on
Corporate Laws (1979) for more information on the issue).
67
finds that audit committee independence is negatively related to abnormal accruals (earnings
management).
54
While these past works primarily focus on the issue of inside director incentive
misalignment, this paper serves to document the decisions of outside directors in isolation from
insider influence.
The results presented here contribute to an established empirical literature on board
structure (Boone et al. (2007); Coles et al. (2008); Linck et al. (2008); Duchin et al. (2010)) by
detailing how the size of the nominating committee and the cross-membership between
committees affect board-level policies and decisions in a post-SOX environment. The empirical
findings provide evidence that large nominating committees function to protect outside directors
from turnover. This highlights an alternative, new role for the nominating committee in the
modern board - one tending toward director job security, as opposed to its more traditional
function of seeking new director talent. Yet, given the inconclusive relation between nominating
committee size (cross-membership) and the quality of board-level decisions, the self-assessment
issue associated with large nominating committees is neither to be interpreted as universally
efficient nor inefficient. If board member heterogeneity, differences in opinions, and board-level
disagreement are valued structural elements in a particular board, then this may yield a situation
where shareholders would demand the formation of a large nominating committee to protect all
members from each other. Alternatively, shareholders may also permit the construction of a large
nominating committee to retain and attract top director talent, using the security offered by the
large nominating committee as a bargaining device. Of course, in opposition to this efficient
view is the notion that shareholder inattentiveness may allow outside directors to create large
nominating committees and implicitly collude within the committee to protect each other from
54
See Klein (1998) for further board sub-structure work highlighting the positive association between the percentage of
inside directors on the finance/investment committee and firm performance. Further, see Shivdasani and Yermack
(1999) for evidence pertaining to the CEO’s control over the nominating committee and its impact on the selection of
directors.
68
removal. Nevertheless, the primary result remains the same whether the interpretation is efficient
or inefficient - large nominating committees ultimately serve the interests of outside directors by
providing greater board-position security.
This paper proceeds as follows. Section 2.2 presents the data construction and summary
statistics. Section 2.3 presents the empirical methodology and results. Section 2.4 concludes the
paper.
2.2 Data, Variable Specification and Summary Statistics
In the proceeding section, I summarize how recent changes in the U.S. regulatory environment
lend itself to the formation of a ‘clean’ boards dataset. Following this, I explicitly detail the
construction of the sample used in the forthcoming empirical analysis. Finally, I provide a
summary of the salient features of the sample.
2.2.1 Regulatory Environment and Dataset Construction
Following the Enron and WorldCom malfeasance scandals of 2000-2001, the Sarbanes-Oxley Act
of 2002 (SOX) was adopted with the intention of being a thorough solution to the governance
inefficiencies which engendered the corporate scandals. Included in the regulation were rules to
improve the quality of financial statements, formally legislate audit committee independence, and
strengthen the enforcement of securities laws.
55
In 2003, the NYSE and NASDAQ took the SOX
requirements one step further by mandating that publicly listed firms have a majority of
independent directors on their boards. In addition, both exchanges set rules for the composition
of board committees, refined the SOX definition of ‘independent’ board member, and required
the financial literacy of audit committee members.
55
For a full listing of the objectives of SOX see The Practitioner’s Guide to Sarbanes-Oxley Act, Volume 1, The
American Bar Association (2004).
69
While the two exchanges set very similar standards regarding audit committee
composition, their listing requirements and rules regarding other committees
(nominating/governance, compensation) differed. NYSE required that all firms establish audit,
nominating/governance, and compensation committees composed entirely of independent
directors. NASDAQ took similar measures, yet allowed more flexibility in the composition of
these committees. First, NASDAQ did not explicitly require that firms have a nominating or
compensation committee. Yet, they did mandate that compensation payable to the CEO had to be
approved either by a majority of the independent directors on the board or a compensation
committee of independent directors. Similarly, for these companies, nominations must be
approved either by a majority of the independent directors on the board, or a nominating
committee of independent directors. Further, if a company elects to establish compensation and
nominating committees of at least three members, then one director who is not independent under
NASDAQ’s rules may hold a position. Under both exchanges’ requirements, certain entities (e.g.
controlled companies, limited partnerships, foreign private issuers, companies in bankruptcy, and
other passive organizations) are exempt from a number of the rules regarding committee
independence and board independence. In addition, both exchanges detailed timetables in which
firms were to comply with the rulings by October 2004.
56
Given NYSE’s more definitive rulings and listing requirements pertaining to the
complete independence of all major committees (audit, compensation, nominating), the sample
used in this investigation is constructed from a set of post-SOX NYSE firms. In addition, foreign
private issuers, controlled companies, and other passive organizations that are not in compliance
with the 2003 NYSE rulings (nor required to be) are eliminated from consideration in the sample.
56
Exceptions were granted to firms with staggered boards, allowing such firms until late 2005 to implement the
changes.
70
Below, I explicitly detail the databases used to form the sample and the necessary firm
information required to be included in the final dataset.
To construct the sample of post-SOX NYSE firms, I utilize six primary databases: IRRC,
Compustat, the Thomson Financial Institutional Ownership database, ExecuComp, the Corporate
Library, and CRSP. To access data needed to create the necessary proxies for the governance
standards within the firm I use the IRRC and Thomson Financial Institutional Ownership
databases. The IRRC database provides annual data for approximately 1,500 firms, primarily
from the S&P 500 and other large corporations.
57
Included in the database is information on firm-
level provisions regarding poison pills, staggered boards, and other charter-bylaws governance
characteristics. Following Bebchuk et al. (2009), the E-Index is constructed from these
provisions and used to proxy for the external corporate governance standards of the firm.
58
Next,
Form 13-F statements from the Thomson Financial Institutional Ownership database are used for
information on firm-level institutional holdings. Data is collect on a yearly basis (4th quarter) and
used to construct two variables: the largest single blockholder and the sum of institutional
holdings (aggregate firm shares held by all institutions).
From Compustat I access the following firm-specific information necessary for the
analysis: total assets, firm age, number of business segments, book leverage, R&D intensity
(R&D/sales), ROA, market-to-book (ratio of the market value to book value of assets).
59
In
addition, all prices and returns are taken from the CRSP monthly files. Since a full year of
57
Since the data is collected sporadically, the methodology implemented in Gompers et al. (2003) is used to fill in
missing observations.
58
Bebchuk et al. (2009) construct a takeover defense index, the entrenchment index (E-Index), based on charter
amendments, supermajority requirements, golden parachutes, poison pills, limits to shareholder bylaw amendments,
and staggered boards.
59
Specifically, return on assets (ROA) is operating income before depreciation over assets. Market-to-book is the book
value of assets minus book value of equity plus the market value of equity normalized by the book value of assets.
Book leverage is the ratio of debt (long term total debt plus debt in current liabilities) to shareholders equity.
71
monthly return data is necessary for the analysis, firm-year observations where the firm enters the
CRSP files half-way through are removed. To ensure that outliers do not have a material impact
on the results, variables are winsorized at the 1% level.
Most importantly, data from the Corporate Library is used for information on board
membership and outside director characteristics. The Corporate Library provides data on board
size, director affiliation, director tenure, director age, attendance, director shares held, director job
title, and other board positions held by each director. In addition, the Corporate Library provides
a full breakdown of each board’s committee structure, detailing which director holds which
committee position.
60
To supplement the Corporate Library with information on CEO
characteristics and compensation, I utilize the ExecuComp database. ExecuComp provides CEO
and officer data, including salary, bonus, equity compensation (fair value of restricted stock
grants plus Black-Scholes value of options awarded), total compensation, CEO age, CEO
ownership, and CEO tenure.
For a given firm-year observation to be included in the final dataset of post-SOX NYSE
firms (2004 to 2008), data on a firm’s board structure must be available from the Corporate
Library (including full information on committee structure). In addition, necessary financial data
(total assets, market-to-book) must be available from Compustat, the firm must be present in the
IRRC and ExecuComp databases, and there must be available institutional ownership information
on Thomson. Since the empirical analysis requires lagged observations, each firm must have
available information for two consecutive years to be included in the final dataset. In addition, all
regulated firms (utilities and financials) are removed from consideration in the forthcoming
60
Schedule 14A (Item 7(e)(1)) requires firms to disclose the names of committee members, the functions performed by
their committees (audit, nominating, compensation), and the number of committee meetings during the year. In the few
cases where firms have separate nominating and governance committees, the nominating committee composition is
used in the forthcoming analysis.
72
empirical analysis. These necessary conditions result in 3,341 firm-year observations and 26,688
director-firm-year observations over the span of the sample period (2004 to 2008).
2.2.2 Summary Statistics
Before examining the relationship between board sub-structure and director decisions, it is
important to summarize how the composition of this sample of post-SOX NYSE firms relates to
prior samples. Table 2.1 provides summary statistics for firm-year and director-firm-year
observations in the dataset. Panel A includes the mean, median, standard deviation, 25th
percentile, and 75th percentile for various firm financial and governance measures. The mean
value of firm total assets is 10,092 ($MM), while the median value is 2,665 ($MM).
61
The
average firm in the sample has a book leverage of 35%, is 29 years old and has an R&D intensity
of 0.02. The mean (median) ROA and market-to-book of a firm in the sample is 0.148 (0.138)
and 1.86 (1.59), respectively. The mean (median) G-Index and block ownership for the sample
firm is 9.58 (9) and 10% (9.4%), respectively.
Panel B presents summary statistics for board structure, director characteristics, and CEO
compensation. The median board size in the sample is 9, while the median level of independence
(fraction of non-employee directors on the board) is 87.5%. This is consistent with levels
documented by other post-SOX boards studies and studies surrounding the SOX event (Linck et
al. (2008), Linck et al. (2009)). The average CEO owns 1.26% of the firm, yet the median CEO
owns only 0.26% of the firm. The mean (median) tenure of a CEO in the sample is 6.18 (5)
years. Mean director ownership (the average percent of shares held by outside directors, by firm)
is about one-fifth the level of mean CEO ownership (0.26% v. 1.26%). In addition, the tenure of
61
The average firm in this study is larger in comparison to firms in previous boards studies (Boone et al. (2007); Linck
et al. (2008)), yet this is not surprising given the NYSE restrictions and the strict sample requirements previously
detailed.
73
Table 2.1: Summary Statistics
This table reports summary statistics for the sample of 3341 firm-year observations from 2004 to 2008. The firm
financial descriptive statistics in Panel A include: assets ($MM), book leverage, R&D intensity (R&D/Sales), standard
deviation of monthly returns, business segments, firm age, ROA, and the ratio of the market value to book value of
assets (market-to-book). The governance descriptive statistics in Panel A include: the Gompers, Ishii, and Metrick
(2003) antitakeover index (G-Index), the Bebchuk, Cohen, and Ferrell (2009) entrenchment index (E-Index),
institutional holdings (total), and block (top blockholder). Panel B presents summary statistics for the sample board
structure. The board descriptive statistics include: board size, the ratio of outsiders to board size (independence), the
fraction of board outsiders holding three or more board seats (fraction busy), the size of various committees (audit,
compensation, nominating/governance), the number of committee positions held by each director, and the cross-
membership between committees (fraction of members on the audit committee who also hold a nominating committee
position, and the fraction of members on the compensation committee who also hold a nominating committee position).
In addition, the summary statistics for director characteristics and outside director turnover are also presented. These
statistics include: mean director ownership (mean holdings of outside directors by firm), median tenure (by firm),
median age (by firm), and the fraction of outside director turnover in a given firm-year. CEO summary statistics
include: salary, total compensation, percent ownership, CEO tenure, and CEO age. Panel C presents board structure
and director turnover statistics partitioned by board size. The board structure and turnover statistics include: the
fraction of board members sitting on various committees, the cross-membership between committees, director turnover,
mean director ownership, median director tenure and median director age.
Panel A: Firm Statistics Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Financial & Governance
Information
Assets 10092.18 36418.17 1102.72 2665.90 7354.00
Book Leverage 0.351 0.225 0.200 0.341 0.476
R&D Intensity 0.021 0.047 0 0.000 0.022
Std of Returns 0.084 0.041 0.057 0.075 0.101
Segments 3.39 1.99 1 3 5
Firm Age 28.77 14.29 15 33 43
ROA 0.148 0.082 0.099 0.138 0.187
Market-to-Book 1.86 0.982 1.27 1.59 2.12
G-Index 9.58 2.47 8 9 11
E-Index 2.51 1.16 2 3 3
Institutional Holdings 0.786 0.165 0.695 0.813 0.914
Block 0.100 0.045 0.069 0.094 0.124
74
Table 2.1 (Continued)
Panel B: Board Statistics Mean Std Dev 25
th
Percent Median 75
th
Percent
Board & Committee Structure
Board Size 9.61 2.11 8 9 11
Independence 0.841 0.078 0.80 0.875 0.90
Fraction Busy 0.378 0.240 0.181 0.375 0.555
Audit Committee Size 3.91 1.02 3 4 5
Nom/Gov Committee Size 4.04 1.51 3 4 5
Comp Committee Size 3.79 1.12 3 4 4
Fraction on Audit 0.502 0.147 0.40 0.500 0.571
Fraction on Nom/Gov 0.517 0.194 0.375 0.500 0.600
Fraction on Comp 0.487 0.161 0.375 0.454 0.571
Fraction of Audit on Nom 0.446 0.320 0.250 0.400 0.666
Fraction of Comp on Nom 0.550 0.309 0.333 0.500 0.75
Committee Positions Held 1.44 0.815 1 1 2
Director Information
Fraction of Director Turnover 0.076 0.108 0.00 0.00 0.125
Mean Director Ownership 0.258 0.919 0.008 0.027 0.096
Director Tenure 7.15 3.94 4 6 9
Director Age 61.44 4.26 59 62 64
CEO Information
CEO Salary 869.11 357.12 630 837.50 1018.84
CEO Total Comp 6684.16 7951.54 2334.43 4465.57 8219.01
CEO Ownership (%) 1.26 3.39 0.090 0.266 0.754
CEO Tenure 6.18 6.10 2 5 8
CEO Age 55.69 6.68 51 56 60
75
Table 2.1 (Continued)
Panel C: Board Structure
Partitioned by Board Size
Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Large Boards (Size>=10)
Fraction on Audit 0.445 0.110 0.363 0.444 0.500
Fraction on Nom/Gov 0.461 0.164 0.333 0.444 0.545
Fraction on Comp 0.427 0.119 0.333 0.428 0.500
Fraction of Audit on Nom 0.369 0.293 0.167 0.333 0.500
Fraction of Comp on Nom 0.485 0.298 0.250 0.500 0.667
Committee Positions Held 1.31 0.760 1 1 2
Fraction of Director Turnover 0.083 0.105 0.00 0.083 0.125
Mean Director Ownership 0.239 0.955 0.006 0.022 0.088
Director Tenure 7.10 3.32 5 6 9
Director Age 61.60 3.58 59 62 64
Small Boards (Size < 10)
Fraction on Audit 0.554 0.157 0.428 0.500 0.625
Fraction on Nom/Gov 0.568 0.206 0.428 0.500 0.667
Fraction on Comp 0.542 0.174 0.428 0.500 0.625
Fraction of Audit on Nom 0.518 0.327 0.333 0.500 0.750
Fraction of Comp on Nom 0.610 0.307 0.333 0.667 1
Committee Positions Held 1.61 0.857 1 2 2
Fraction of Director Turnover 0.069 0.110 0.00 0.00 0.143
Mean Director Ownership 0.276 0.885 0.011 0.033 0.104
Director Tenure 7.19 4.43 4 6 9
Director Age 61.29 4.79 58 61 64
76
the average director is slightly higher than the tenure of the average CEO in the sample (7.15 v.
6.18 years).
In addition, Panel B also details committee size and various measures of board structure.
The average audit, nominating, and compensation committees have 3.91, 4.04, and 3.79
members, respectively. This is in accordance with Hayes et al. (2004) who find that most
committees have a median size of four directors. Most important for the purposes of this
empirical investigation is the fraction of outside directors who hold the nominating committee
position. The median firm in the sample has 50% of their outside directors holding a position on
the nominating committee, while the 25th and 75th percentiles are 37% and 60%, respectively.
Although not directly listed in Table 2.1, 6.81% of firms in the sample have a board structure
where every single outside member sits on the nominating committee. Together, these two
summary statistics indicate that there is a significant degree of heterogeneity across boards in
terms of the formation and size of the nominating committee. In addition, the mean (median)
fraction of audit committee members who also hold a nominating committee position is 44%
(40%), and the mean (median) fraction of compensation committee members who also hold a
nominating committee position is 55% (50%).
Another important feature of the dataset is the level of outside director turnover in a
given year (Panel B). The fraction of outside directors who leave the board in a given year is
7.6%.
62
Past pre- and post-SOX studies investigating the rate of director turnover document
similar levels of outside director turnover. Through a study into the consequences of option back-
dating, Ertimur et al. (2010) highlight that outside directors have a 15% chance of departing the
board over a two year period (2006 to 2008). Prior to this study, Yermack (2004) shows in a
62
Consistent with past studies, all director turnover events due to de-listing are removed from consideration. Such
firm-year observations are removed from the dataset as well.
77
1994 boards sample that 28% of directors leave their seat over a five year period. This is
equivalent to a 5.6% rate of removal on a yearly basis. The rate of outside director turnover in
this study is at a similar level of magnitude to these two studies and other past works.
Panel C presents a breakdown of certain statistics partitioned by board size. Given that
large boards (10 or more member boards) have a greater number of directors to pick from when
forming committees, it naturally follows that large boards have a lower fraction of outside
directors sitting on the nominating committee (46% v. 57%). This also entails that large boards
have a lower level of cross-membership between the nominating committee and other committees
(fraction of audit committee members on the nominating committee, and fraction of
compensation committee members on the nominating committee). Hence, since board size plays
such a significant role in defining committee structure, much of the analysis in subsequent
sections is partitioned by board size. Finally, since board size is strongly correlated with firm size
(Boone et al. (2007); Coles et al. (2008)), the results presented in Table 2.1 Panel C are
qualitatively the same if the sample is partitioned by firm size instead of board size.
2.3 Empirical Design
In the following section, I examine the relation between board sub-structure and the decisions of
the board. First, I investigate how the size of the nominating committee and the cross-
membership between the nominating committee and other committees is associated with outside
director turnover. Next, I turn to the policy decisions of the board and detail how the cross-
membership between the nominating and other monitoring committees relates to the type and
quality of board decisions.
78
2.3.1 Outside Director Turnover and Nominating Committee Structure
To examine the effect of board sub-structure on director turnover, I utilize the previously
constructed dataset of post-SOX NYSE firms. This set of 3,341 firm-year observations (26,688
outside director-firm-year observations) constitutes a ‘clean’ dataset where the effect of director
committee cross-membership on outside director turnover is easy to discern due to the 2003
NYSE listing requirements. Table 2.2 reports the results for a series of logit regressions for the
sample of 26,688 outside director-firm-year observations. The dependent variable in all columns
takes a value of one if an outside director leaves the board in a given year. Following the existing
literature, I control for a variety of director and firm characteristics found to affect director
turnover (Denis and Sarin (1999); Yermack (2004); Fich and Shivdasani (2007); Ertimur et al.
(2010)). In particular, all controls denote measures taken from the beginning of each director-
firm-year observation and include: firm size (log of total assets), firm age, institutional
ownership, board size, independence, gray director (affiliate), industry-adjusted returns (returns
over the prior two years, adjusted by the median return in the Fama-French 48 grouping), change
in industry-adjusted ROA over the prior year, busy director (three or more concurrent board
seats), old director (over 68 years in age), director tenure, high director ownership (greater than
0.1% ownership in the firm).
63
I also include indicator variables for the fraction of outside
directors holding certain committee positions (audit, compensation, nominating). High fraction
of outside directors on the nominating committee (High Frac on Nom) denotes firm-year
observations where the firm is at the 75th percentile or higher in terms of the fraction of directors
sitting on the nominating committee - which equates to 60% or more of the outside directors
holding the nominating committee position. In a similar fashion, high fraction of outside
directors on the audit (compensation) committee denotes firm-year observations where the firm is
63
Consistent with the prior literature, affiliated outside directors are denoted as outsiders (Coles et al. (2008), Huson et
al. (2001)) and independence is constructed as the number of outsiders divided by the total number of directors.
79
at the 75th percentile or higher in terms of the number of directors sitting on the audit
(compensation) committee - which equates to 57% (57%) or more of the outside directors holding
the audit (compensation) committee position. Industry and year fixed effects are included in all
logit models and heteroskedasticity-consistent p-values clustered at the director-firm level are
reported throughout the table. In addition, it should be noted that the inclusion of other director
characteristics (job title, financial expertise, gender), CEO departures, and other measures of
governance (E-Index, block ownership) do not alter the results pertaining to the primary variables
of interest and are excluded for brevity.
First, although not directly reported in Table 2.2, it is important to summarize the
univariate differences in director turnover between firms with small nominating committees and
firms with large nominating committees (greater than or equal to 60% of outsiders on the
nominating committee). Small nominating committees have an average yearly turnover rate of
8.14%, while large nominating committees have an average yearly turnover rate of 6.53%
(difference significant at the 1% level). This equates to a 25% increase in the probability of
departure if a director sits on a small nominating committee as opposed to a large nominating
committee. All differences between large and small nominating committees in the predicted
probability of director turnover presented in Table 2.2 closely equate to this univariate result (i.e.
setting all other control variables at their mean levels in the logit model to assess the effect of
nominating committee size).
Column (1) of Panel A (Table 2.2) presents the relation between director turnover and
various firm-director characteristics found to be associated with the rate of director turnover.
Consistent with the prior literature, independent boards and director age (old director) are
positively related to outside director turnover, and firm performance (industry-adjusted returns)
and busy directors are negatively associated with turnover (omitted in Panel A). Column (1) also
80
Table 2.2: Director Turnover and Firm-Level Committee Structure
This table reports results for a series of logit regressions for 26688 director-firm-year observations from 2004 to 2008.
The dependent variable takes a value of one if the director leaves the board in a given year, and zero otherwise. The
following director and firm-level variables are implemented as controls: firm size (log of total assets), firm age,
institutional ownership, board size, independence, gray director, busy director (indicator of one if the director holds
three or more positions), old director (greater than 68 years in age), director tenure (log of tenure), industry-adjusted
returns over the prior two years, change in industry-adjusted ROA over the prior year, director ownership (indicator of
one if the director holds greater than .1% of the common shares outstanding), and the fraction of directors holding
committee positions within the firm. High fraction on Nom (Audit, Comp) denotes firm-year observations where the
number of outside directors holding the nominating (audit, compensation) committee position is at or above the 75
th
percentile for the sample. Panel A presents the results for the full sample, while Panel B presents the results for large
board sample (10 or more members) and the small board sample (fewer than 10 members). In columns (3) and (4) of
Panel A the performance measure used in interaction terms is the industry-adjusted returns to the firm over the prior
two years and an indicator if the director failed attendance standards in the previous year, respectively. In columns (3)
and (6) of Panel B the performance measure used in interaction terms is the industry-adjusted returns to the firm over
the prior two years. Industry (Fama-French 48 classification) and year fixed effects are included in all regressions.
Standard errors are computed using robust methods and p-values are reported below coefficients in parentheses.
Panel A: Full Sample (1) (2) (3) (4)
Firm Size 0.0108 0.0193 0.0185 0.0192
(0.68) (0.48) (0.49) (0.48)
Firm Age -0.0021 -0.0039 -0.0039 -0.0038
(0.34) (0.10) (0.09) (0.11)
Busy Director -0.0808 -0.0767 -0.0772 -0.0819
(0.12) (0.14) (0.14) (0.11)
Old Director 1.337 1.198 1.198 1.198
(0.00) (0.00) (0.00) (0.00)
High Frac on Nom -0.1842 -0.1911 -0.2059 -0.1822
(0.01) (0.01) (0.00) (0.02)
Director Tenure 0.1704 0.1533 0.1710
(0.00) (0.00) (0.00)
High Own Dir 0.0341 0.0354 0.0393
(0.68) (0.67) (0.63)
High Frac on Audit 0.0391 0.0361 0.0391
(0.64) (0.67) (0.64)
High Frac on Comp -0.0426 -0.0443 -0.0505
(0.62) (0.60) (0.55)
Ret*High Nom Frac 0.1430
(0.20)
Ret*Board Size -0.0166
(0.65)
Ret*Independence 0.7671
(0.43)
Ret*Dir Tenure 0.2187
(0.00)
Failed Standards 1.389
(0.40)
Failed Standards*High Nom -0.4863
(0.19)
Failed Standards *Board Size 0.0081
(0.91)
Failed Standards *Indep -0.3107
(0.88)
Failed Standards *Dir Tenure 0.0009
(0.98)
N 26688 26688 26688 26688
χ
2
948.41 986.62 1004.01 1034.21
81
Table 2.2 (Continued)
Panel B: Small and Large
Boards
Large
Boards
(1)
Large
Boards
(2)
Large
Boards
(3)
Small
Boards
(4)
Small
Boards
(5)
Small
Boards
(6)
Firm Size 0.0298 0.0385 0.0368 -0.0170 -0.0125 -0.0118
(0.36) (0.26) (0.28) (0.69) (0.79) (0.79)
Firm Age -0.0017 -0.0032 -0.0034 -0.0031 -0.0052 -0.0053
(0.59) (0.35) (0.31) (0.29) (0.08) (0.08)
Inst Own -0.0915 -0.0644 -0.0666 0.3342 0.3827 0.4010
(0.74) (0.82) (0.82) (0.27) (0.23) (0.20)
Board Size 0.0247 0.0221 0.0169 0.0951 0.1060 0.1110
(0.44) (0.49) (0.58) (0.07) (0.04) (0.04)
Independence 1.041 1.095 0.9903 1.723 1.933 1.909
(0.08) (0.08) (0.09) (0.01) (0.01) (0.01)
Gray Director 0.4158 0.3296 0.3313 0.4186 0.2908 0.2923
(0.00) (0.00) (0.00) (0.00) (0.01) (0.01)
Ind Adj Ret -0.3251 -0.3127 -3.163 -0.2594 -0.2584 -1.127
(0.00) (0.00) (0.01) (0.00) (0.00) (0.39)
Change in ROA -0.2354 -0.2261 -0.1968 1.895 1.901 1.798
(0.83) (0.83) (0.85) (0.08) (0.10) (0.11)
Busy Director -0.1405 -0.1416 -0.1431 0.0122 0.0295 0.0324
(0.03) (0.03) (0.02) (0.88) (0.73) (0.71)
Old Director 1.458 1.331 1.334 1.164 1.016 1.016
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
High Frac on Nom -0.2347 -0.2297 -0.2338 -0.1685 -0.1829 -0.1992
(0.02) (0.03) (0.03) (0.08) (0.07) (0.06)
Director Tenure 0.1561 0.1431 0.1901 0.1707
(0.00) (0.00) (0.00) (0.00)
High Own Dir 0.0348 0.0354 0.0805 0.0791
(0.77) (0.76) (0.48) (0.49)
High Frac on Audit -0.0081 0.0005 0.1272 0.1275
(0.95) (0.98) (0.19) (0.19)
High Frac on Comp -0.0530 -0.0448 -0.0643 -0.0641
(0.75) (0.78) (0.51) (0.50)
Ret*High Nom Frac 0.0015 0.1901
(0.94) (0.29)
Ret*Board Size 0.0660 -0.0119
(0.49) (0.91)
Ret*Independence 2.122 0.3457
(0.17) (0.78)
Ret*Dir Tenure 0.1323 0.3172
(0.13) (0.00)
N 15415 15415 15415 11273 11273 11273
χ
2
704.78 723.47 731.14 328.37 350.52 367.19
82
shows that boards with large nominating committees (High Frac on Nom) are associated with
lower levels of director turnover (coefficient negative and significant at the 1% level). In Column
(2), I also include controls for the fraction of directors holding other committee positions (audit,
compensation) at the firm level. The coefficients on High Frac on Audit and High Frac on Comp
are insignificant, while the coefficient on High Frac on Nom remains negative and significant at
the 1% level.
While the first two columns of Panel A appear to suggest that large nominating
committees serve as insulating devices for outside board members (lower probability of director
removal), the next question to address is whether or not such a board feature is also associated
with a lower sensitivity of director turnover to performance. In other words, while large
nominating committees may remove fewer directors, are they also less likely to dismiss directors
following poor performance? To examine this question, I consider two primary measures of
performance: firm performance and director attendance. First, following Huson et al. (2001) I
consider a return based measure of performance: industry-adjusted returns over a two year period
where returns to each firm are adjusted by subtracting the median return for all firms in the same
FF48 classification during the same two year period. While Huson et al. (2001) implement this
performance measure to study CEO turnover, it may also be the case that directors are held
partially responsible for the performance of the firm. Columns (1) and (2) of Panel A lend
support to this notion - industry-adjusted returns are negatively related to director turnover. In
such, in Column (3) I interact this measure of firm performance with four board characteristics:
board size, independence, director tenure, and High Frac on Nom. Column (3) highlights that
high tenure directors are less likely to be removed following poor firm performance (significant at
the 1% level), and boards with a high fraction of members on the nominating committee are also
83
less likely to remove members following poor performance, though not significant at the 10%
level.
64
Next, in Column (4) I consider an alternative measure of performance at the director
level: board meeting attendance. While individual measures of director performance are difficult
to construct, one easily observable measure is whether a director attends at least 75% of board
meetings.
65
To capture director performance on an individual basis, I create an indicator variable
which takes a value of one if a director fails attendance standards in the prior year. It is important
to note that although this indicator variable constitutes an appropriate measure of director
performance, only 1.5% of directors in this sample fail attendance standards in a given year.
Hence, given this highly skewed distribution, it may be difficult to observe any economically or
statistically relevant results. Nevertheless, Column (4) includes this performance measure
interacted with various board characteristics. Again, although boards with large nominating
committees are less likely to dismiss members following poor attendance, the coefficient on the
interaction term is not significant at the 10% level.
Since committee structure is correlated with board size, in Panel B of Table 2.2 I repeat a
similar series of logit regressions where the initial sample is partitioned by board size. Columns
(1) - (3) present the model results for the large boards sample (greater than or equal to 10 member
boards), and Columns (4) - (6) present the results for the small boards sample. Similar results to
those presented in Panel A persist across control variables. In addition, comparing the relation
between nominating committee size to turnover across the two samples, it appears that the
64
In untabulated results, I also consider industry-adjusted ROA as a performance measure and observe insignificant
results when considering the interaction of this performance measure with High Frac on Nom.
65
Cai et al. (2009) show that directors who fail attendance standards receive 14% fewer votes in director elections.
Since the authors demonstrate that director votes withheld are primarily determined by director performance indicators
(meeting attendance), instead of using shareholder votes as a measure of director performance it suffices to use a direct
measure of performance - the director’s board attendance.
84
coefficient on High Frac on Nom is slightly more significant for the large boards sample as
compared to the small boards sample. Across both of the sub-samples, there again is no strong
evidence that large nominating committees are less likely to remove directors following poor
performance, using industry-adjusted returns as the performance measure.
Moreover, the results detailed in Panel A and B of Table 2.2 are robust to other sample
constructions, including the removal of affiliated directors, and using only young directors (under
68 years old) to define the sample. Further, although the intent of this empirical investigation is
to document how nominating committee size at the firm-level affects director turnover,
substituting the High Frac on Nom indicator variable with a director-specific indicator variable
(value of one if the director holds a nominating committee position) yields an equally negative
and significant coefficient when implemented in the logit models of Table 2.2. In essence, large
nominating committees are associated with lower director turnover rates for the exact reason that
more outside directors are holding nominating committee seats.
While Table 2.2 presents the sensitivity of director turnover to performance using
attendance standards and firm returns as performance measures, directors sitting on certain
committees may be evaluated using different metrics. To address this issue, I focus on audit
committee directors and compensation committee directors separately, applying distinct
performance measures for each group. First, I examine outside director turnover among
compensation committee members with respect to committee level performance. Cai et al.
(2009) and Hartzell et al. (2004) consider abnormal CEO compensation as a measure of corporate
governance, where firm-level excess CEO compensation is defined as the residual from a
compensation regression with firm size and prior year firm returns as control variables. Though
abnormal total compensation granted to the CEO may serve as an indicator of compensation
committee performance, it may also be an indeterminate measure depending on the form in which
85
the compensation is granted to the CEO. If the role of the compensation committee is to
represent the interests of shareholders by tying the incentives of the CEO to that of the firm
through equity-based compensation, it is unclear if excess total compensation is necessarily a
measure of poor committee performance if the primary component of the compensation given to
the CEO is restricted stock and options. Therefore, I opt to concentrate on the form of the
compensation granted by the compensation committee as a measure of performance.
66
I follow
the same general procedure as Cai et al. (2009) and Hartzell et al. (2004) and define abnormal
equity-based compensation as the residual from a regression of all ExecuComp firms over the
2004-2008 period. The dependent variable (sum of the value of restricted stock and options
awarded normalized by salary) is regressed on firm size (log of assets), prior year returns
(industry-adjusted), market-to-book, CEO age, CEO tenure, industry fixed effects, and year fixed
effects. The lagged residual extracted from this regression (Res EBC) indicates that the
compensation committee is performing its duties if the measure is positive, expending high effort
by forcing the CEO to take on more equity-based as opposed to cash-based compensation.
Table 2.3 reports the association between director turnover and firm characteristics for
12,040 compensation committee director-firm-year observations. Previously, the central
independent variable of interest was the fraction of outside directors holding the nominating
committee position. Now, given that the set of director observations only includes compensation
committee members, I redefine this variable of interest as the fraction of compensation committee
members within the firm that also hold the nominating committee position. Following the
previous methodology, High Frac of Comp on Nom denotes firm-year observations where the
fraction of compensation members holding the nominating committee position is at the 75th
66
Mehran (1995) shows that equity ownership and equity-based compensation to the CEO are associated with higher
levels of firm performance. Datta et al. (2001) demonstrate that equity-based compensation is associated with greater
returns around acquisition announcements and better long-run performance.
86
Table 2.3: Compensation Director Turnover and Firm-Level Committee Structure
This table reports results for a series of logit regressions for 12040 compensation director-firm-year observations from
2004 to 2008. The dependent variable takes a value of one if a director on the compensation committee leaves the
board in a given year, and zero otherwise. The following director and firm-level variables are implemented as controls:
firm size (log of total assets), firm age, institutional ownership, board size, independence, gray director (omitted), busy
director (indicator of one if the director holds three or more positions), old director (greater than 68 years in age),
director tenure (log of tenure), industry-adjusted returns over the prior two years, change in industry-adjusted ROA
over the prior year, residual of EBC (residual from equity-based compensation on a set of controls), director ownership
(indicator of one if the director holds greater than .1% of the common shares outstanding), and the fraction of
compensation committee directors who hold a nominating committee position. High Frac of Comp on Nom denotes
firm-year observations where the number of compensation committee members holding the nominating committee
position is at or above the 75
th
percentile. Columns (1) – (2) include the full sample, columns (3) – (4) include only
large boards (boards greater than or equal to 10 members), and columns (5) – (6) include only small boards (boards less
than 10 members in size). Industry (Fama-French 48 classification) and year fixed effects are included in all
regressions. Standard errors are computed using robust methods and p-values are reported below in parentheses.
Full
Sample (1)
Full
Sample (2)
Large
Boards (3)
Large
Boards (4)
Small
Boards (5)
Small
Boards (6)
Firm Size 0.0519 0.0493 0.0587 0.0610 0.0398 0.0426
(0.17) (0.19) (0.17) (0.15) (0.59) (0.56)
Firm Age -0.0069 -0.0070 -0.0056 -0.0057 -0.0086 -0.0086
(0.02) (0.02) (0.17) (0.16) (0.05) (0.05)
Inst Own 0.0138 0.0240 -0.1447 -0.1193 0.3861 0.4119
(0.96) (0.93) (0.71) (0.76) (0.39) (0.36)
Board Size 0.0632 0.0646 0.0526 0.0376 0.1933 0.1926
(0.01) (0.01) (0.21) (0.37) (0.01) (0.01)
Independence 1.581 1.602 1.373 1.130 1.448 1.541
(0.00) (0.00) (0.08) (0.16) (0.08) (0.06)
Residual EBC -0.0177 0.0967 -0.0277 0.5151 -0.0005 -0.1603
(0.08) (0.47) (0.08) (0.02) (0.97) (0.53)
Change in ROA -1.027 -1.097 -2.598 -2.772 0.8413 0.8726
(0.34) (0.31) (0.09) (0.08) (0.61) (0.59)
Ind Adj Ret -0.2091 -0.2082 -0.2245 -0.2148 -0.1909 -0.1919
(0.02) (0.02) (0.08) (0.09) (0.14) (0.13)
Busy Director -0.0050 -0.0051 -0.0314 -0.0264 0.0135 0.0099
(0.94) (0.94) (0.75) (0.79) (0.91) (0.93)
Old Director 1.283 1.284 1.405 1.417 1.141 1.146
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
High Frac of Comp on Nom -0.1927 -0.1864 -0.2134 -0.1885 -0.1652 -0.1750
(0.02) (0.04) (0.06) (0.12) (0.19) (0.18)
Director Tenure 0.2067 0.2057 0.2060 0.2088 0.1993 0.1967
(0.00) (0.00) (0.00) (0.00) (0.01) (0.01)
High Own Dir 0.0009 -0.0044 -0.0140 -0.0328 0.1080 0.1116
(0.98) (0.97) (0.95) (0.88) (0.52) (0.51)
Res EBC*High Nom Frac 0.0419 0.0594 0.0103
(0.09) (0.13) (0.79)
Res EBC*Board Size -0.0040 -0.0144 0.0031
(0.44) (0.13) (0.86)
Res EBC *Independence -0.0771 -0.4375 0.1904
(0.62) (0.05) (0.49)
Res EBC*Dir Tenure -0.0102 -0.0021 -0.0135
(0.40) (0.91) (0.39)
N 12040 12040 6374 6374 5666 5666
χ
2
516.17 522.99 359.41 370.04 210.93 212.39
87
percentile or higher for the sample (0.75). In Column (1), for the full sample of compensation
committee director-firm-year observations, the coefficient on Res EBC is negative and weakly
significant, while the coefficient on High Frac of Comp on Nom is negative and significant at the
2% level. These results indicate that lower levels of equity-based compensation (poor
compensation committee performance) are associated with higher director turnover, and that
higher levels of cross-membership between the nominating committee and the compensation
committee correspond to lower compensation committee turnover. In Column (2), the interaction
term between the performance measure (Res EBC) and High Frac of Comp on Nom is included in
the logit model. The coefficient attached to the interaction term is positive and significant (at the
10% level), indicating that boards with a high fraction of their compensation members holding
the nominating committee position will weakly remove fewer members following poor
performance (low equity-based compensation). Columns (3) - (6) detail the results for the boards
sub-samples. Similar results persist across the two sub-samples, with a slightly more significant
coefficient on High Frac of Comp on Nom for the large boards sample as compared to the small
boards sample.
In untabulated analysis, I consider two alternative measures of compensation committee
performance: excess total compensation and excess cash-based compensation (following the
methodology of Hartzell et al. (2004)). The director performance-turnover relationship for boards
with high levels of cross-membership between the compensation committee and nominating
committee is again weak and not significant at the 5% level.
Turning to the issue of audit committee performance and director turnover, Table 2.4
presents the results for 10,390 audit committee director-firm-year observations spanning 2004 to
2007. Srinivasan (2005) demonstrates that for firms that overstate earnings, the likelihood of
director departure increases with restatement size, particularly for audit committee directors. In
88
such, to define my audit committee performance measure, I collect data taken from the U.S.
Government Accountability Office (GAO), which reports financial restatements by firms through
2007. The restatement indicator variable used in all columns takes a value of one if there is a
restatement declared in the contemporaneous or prior year (relative to the director-firm-year
observation). In addition, I define High Frac of Audit on Nom as an indicator taking a value of
one if the fraction of audit committee members holding the nominating committee position is at
or above the 75th percentile for the sample (0.66). Corroborating past findings, in Column (1) of
Table 2.4, the relation between audit committee turnover and the incident of a financial
restatement is positive and significant at the 2% level. Following this, the coefficient on Frac of
Audit on Nom is negative and significant, which indicates that high cross-membership between
the audit committee and the nominating committee is associated with lower audit committee
turnover. In subsequent columns throughout the table, the interaction term between High Frac of
Audit on Nom and the restatement indicator exhibits no significant relation to audit committee
turnover.
The results documented in Tables 2.2-2.4 provide evidence of the problems of self-
evaluation faced by large nominating committees. Large nominating committees and boards with
high cross-membership between the nominating committee and other monitoring committees are
associated with lower levels of director turnover. In addition, while a negative association does
generally persist, results pertaining to the sensitivity of director turnover to performance for large
nominating committees are ultimately inconclusive (i.e. results are not statistically significant at
the 5% level across all measures implemented). The primary consistent and significant result is
that high cross-membership between the nominating committee and other monitoring committees,
and large nominating committees are associated with lower director turnover rates over all
performance-states.
89
Table 2.4: Audit Director Turnover and Firm-Level Committee Structure
This table reports results for a series of logit regressions for 10390 audit director-firm-year observations from 2004 to
2007. The dependent variable takes a value of one if a director on the audit committee leaves the board in a given year,
and zero otherwise. The following director and firm-level variables are implemented as controls: firm size (log of total
assets), firm age, institutional ownership, board size, independence, gray director, busy director (indicator of one if the
director holds three or more positions), old director (greater than 68 years in age), director tenure (log of tenure),
industry-adjusted returns over the prior two years, change in industry-adjusted ROA over the prior year, restatement
(indicator taking a value of one if there was a restatement in the contemporaneous or prior year), director ownership
(indicator of one if the director holds greater than .1% of the common shares outstanding), and the fraction of audit
committee directors who hold a nominating committee position. High Frac of Audit on Nom denotes firm-year
observations where the number of audit committee members holding the nominating committee position is at or above
the 75
th
percentile. Columns (1) – (2) include the full sample, columns (3) – (4) include only large boards (boards
greater than or equal to 10 members), and columns (5) – (6) include only small boards (boards less than 10 members in
size). Industry (Fama-French 48 classification) and year fixed effects are included in all regressions. Standard errors
are computed using robust methods and p-values are reported below coefficients in parentheses.
Full Sample
(1)
Full Sample
(2)
Large
Boards (3)
Large
Boards (4)
Small
Boards (5)
Small
Boards (6)
Firm Size -0.0059 -0.0065 -0.0148 -0.0164 -0.0298 -0.0330
(0.87) (0.86) (0.76) (0.73) (0.64) (0.61)
Firm Age -0.0044 -0.0047 0.0016 0.0012 -0.0087 -0.0089
(0.17) (0.15) (0.74) (0.80) (0.07) (0.06)
Inst Own 0.0252 0.0409 -0.2702 -0.2588 0.6253 0.6270
(0.92) (0.88) (0.50) (0.52) (0.11) (0.11)
Board Size 0.0303 0.0472 0.0199 0.0268 0.0969 0.1057
(0.27) (0.11) (0.68) (0.60) (0.21) (0.19)
Independence 2.027 1.869 1.576 1.475 2.146 1.929
(0.00) (0.01) (0.09) (0.14) (0.03) (0.06)
Gray Director -0.1080 -0.1334 0.0302 0.0164 -0.2653 -0.2894
(0.52) (0.43) (0.89) (0.94) (0.29) (0.26)
Restatement 0.2677 0.6641 -0.0092 -0.3256 0.5174 -0.5651
(0.02) (0.70) (0.95) (0.92) (0.00) (0.85)
Change in ROA 3.153 3.287 2.268 2.236 4.102 4.046
(0.02) (0.01) (0.19) (0.19) (0.02) (0.03)
Ind Adj Ret -0.2450 -0.2391 -0.1309 -0.1161 -0.3074 -0.2960
(0.01) (0.01) (0.44) (0.50) (0.01) (0.02)
Busy Director -0.0023 -0.0003 -0.0217 -0.0172 0.0155 0.0147
(0.97) (0.99) (0.86) (0.88) (0.90) (0.91)
Old Director 1.197 1.195 1.339 1.338 1.061 1.065
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
High Frac Audit on Nom -0.1798 -0.1874 -0.1515 -0.1879 -0.2020 -0.2097
(0.06) (0.05) (0.19) (0.13) (0.13) (0.10)
Director Tenure 0.2942 0.3449 0.2356 0.2809 0.3702 0.4135
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
High Own Dir 0.1731 0.1716 -0.0239 -0.0269 0.3297 0.3153
(0.21) (0.21) (0.92) (0.91) (0.10) (0.09)
Restate*High Nom Frac 0.1620 0.3744 0.0653
(0.68) (0.53) (0.88)
Restate*Board Size -0.1291 -0.0401 -0.0589
(0.03) (0.74) (0.76)
Restate *Independence 1.628 1.5366 2.321
(0.40) (0.59) (0.43)
Restate *Dir Tenure -0.3062 -0.3281 -0.2370
(0.02) (0.07) (0.18)
N 10390 10390 5369 5369 5021 5021
χ
2
388.44 400.73 252.97 257.19 218.14 220.62
90
2.3.2 Board Decisions and Committee Structure
While the previous findings suggest that large nominating committees and high cross-
membership between the nominating committee and other monitoring committees serve to protect
outside directors from turnover, how does the sub-structure of the board affect the type and
quality of board policy decisions? If large nominating committees appear to function as
insulating devices, do outside directors who sit on such boards perform their oversight duties in a
different manner given their increased job security? To investigate this issue, I examine the
relationship between nominating committee size (cross-membership) and board decisions through
the three primary functions of the board: compensation, accounting quality, and changes in
governance.
If the proper role of outside directors on the compensation committee is to bargain in the
interests of shareholders for lower CEO cash-based compensation and higher equity-based
compensation, then how does the sub-structure of the board relate to such decisions?
67
Does
greater board-position security via the nominating committee translate to lower levels of effort on
the part of compensation committee members - allocating higher cash-based compensation, and
granting the CEO lower levels of equity-based compensation? In Table 2.5, I detail the
relationship between the form of the compensation granted to the CEO and the fraction of
compensation committee members who hold the nominating committee position. Columns (1) -
(3) present a set of OLS regressions for the full sample of 3,341 firm-year observations where the
dependent variables over the span of the columns are equity-based compensation (sum of the fair
value of restricted stock grants and Black-Scholes value of option grants normalized by salary),
cash-based compensation (salary plus bonus), and total compensation, respectively. Following
67
See Mehran (1995) and Datta et al. (2001) for evidence documenting how higher levels of CEO equity-based
compensation serve shareholders’ interests.
91
Table 2.5: CEO Compensation and Board Committee Structure
This table reports results for a series of regressions for 3341 firm-year observations from 2004 to 2008. The three
primary dependent variables investigated in this table are equity based compensation (total CEO equity compensation
normalized by salary), cash compensation (salary plus bonus), and total compensation in a given year. The following
director and firm-level variables are implemented as controls: firm size (log of total assets), firm age, segments (log of
firm business segments), book leverage, R&D (indicator of one if R&D intensity is above the sample median), standard
deviation of monthly returns, institutional block ownership (largest single block owner), CEO tenure (log of tenure),
CEO age, board size, independence, market-to-book, industry-adjusted returns over the prior two years, industry-
adjusted ROA, change in industry-adjusted ROA over the prior year, and high fraction of compensation committee
members on the nominating committee (indicator of one if the ratio is at the 75
th
percentile or greater). Columns (1) –
(3) present the results for the full sample, columns (4) – (6) present the results for the large board sample (10 or more
members), and columns (7) – (9) present the results for the small board sample. Industry (Fama-French 48
classification) and year fixed effects are included in all regressions. Standard errors are computed using robust
methods and p-values are reported below coefficients in parentheses.
EBC
(1)
Cash
Comp
(2)
Total
Comp
(3)
EBC
(4)
Cash
Comp
(5)
Total
Comp
(6)
EBC
(7)
Cash
Comp
(8)
Total
Comp
(9)
Firm Size 1.24 470.03 2956.94 1.15 485.77 3108.23 1.44 435.38 2822.06
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Firm Age -0.0129 0.2732 -3.51 -0.0112 3.52 2.78 -0.0155 -1.10 -15.02
(0.01) (0.85) (0.65) (0.16) (0.20) (0.82) (0.02) (0.61) (0.12)
Firm Segments -0.0092 48.15 54.02 0.3694 89.35 504.72 -0.2881 -17.17 -192.20
(0.90) (0.16) (0.74) (0.02) (0.09) (0.05) (0.06) (0.71) (0.37)
Leverage -0.6272 18.64 -1601.60 -0.3480 -294.15 -1838.53 -0.7861 246.94 -1066.23
(0.07) (0.85) (0.00) (0.47) (0.07) (0.02) (0.07) (0.08) (0.10)
R&D 0.0901 -6.67 74.74 -0.5022 -136.89 -451.61 0.5434 65.13 243.13
(0.66) (0.92) (0.80) (0.10) (0.17) (0.35) (0.03) (0.45) (0.54)
Std of Ret 2.95 -122.45 2926.64 2.22 729.85 6096.52 4.26 -424.36 1007.04
(0.15) (0.84) (0.32) (0.47) (0.49) (0.23) (0.12) (0.57) (0.77)
Block Own 0.1138 -1234.30 -4330.27 1.67 -619.84 -1904.17 -0.9987 -1660.92 -5482.90
(0.95) (0.01) (0.06) (0.46) (0.42) (0.61) (0.62) (0.01) (0.05)
CEO Tenure 0.0846 112.03 468.02 0.1374 120.50 536.60 0.0230 90.73 362.44
(0.28) (0.00) (0.00) (0.21) (0.00) (0.00) (0.83) (0.01) (0.01)
CEO Age -0.0130 6.24 20.50 0.0020 7.97 31.54 -0.0168 5.34 12.62
(0.22) (0.06) (0.19) (0.89) (0.13) (0.23) (0.25) (0.20) (0.52)
Board Size -0.1135 -8.41 -85.29 -0.0508 5.86 -26.87 -0.1654 15.08 -158.49
(0.01) (0.49) (0.14) (0.45) (0.79) (0.81) (0.10) (0.61) (0.24)
Independence 1.65 -397.04 1606.98 3.41 412.77 6039.25 -0.0124 -1053.06 -2827.20
(0.05) (0.15) (0.22) (0.01) (0.35) (0.01) (0.98) (0.01) (0.10)
Market-to-Book 0.5812 38.91 875.75 0.5804 23.71 476.72 0.5483 26.01 945.28
(0.00) (0.17) (0.00) (0.00) (0.67) (0.07) (0.00) (0.43) (0.00)
Ind Adj ROA 0.5713 597.97 1014.30 3.40 1100.08 9558.21 -0.6647 423.06 -2748.77
(0.61) (0.08) (0.54) (0.11) (0.10) (0.00) (0.63) (0.30) (0.15)
Ind Adj Ret 0.4337 198.86 1370.29 0.3563 239.37 1675.41 0.4517 211.69 1259.80
(0.01) (0.00) (0.00) (0.15) (0.00) (0.00) (0.03) (0.00) (0.00)
Change in ROA -2.22 780.41 1016.92 -1.75 847.35 1098.25 -2.43 652.72 731.15
(0.23) (0.18) (0.71) (0.56) (0.39) (0.82) (0.31) (0.37) (0.82)
High Frac of
Comp on Nom
-0.2340 22.12 -394.51 -0.1495 -9.95 -745.51 -0.3075 25.54 -226.66
(0.09) (0.60) (0.09) (0.45) (0.90) (0.03) (0.10) (0.65) (0.41)
N 3147 3190 3180 1527 1543 1539 1620 1647 1641
R
2
0.2157 0.3843 0.4095 0.2436 0.4285 0.4441 0.2218 0.3480 0.3451
92
the existing literature on governance and compensation levels, I include various firm, board, and
CEO characteristics to control for factors that have been previously identified to affect CEO
compensation (Core et al. (1999)). These control variables include: firm size, firm age, segments,
leverage, R&D, standard deviation of prior year returns, block ownership, CEO tenure, CEO age,
board size, independence, market-to-book, industry-adjusted ROA, change in ROA (past year),
industry-adjusted returns, and the cross-membership between the nominating and compensation
committee.
68
Column (1) demonstrates that boards with a high fraction of compensation
members on the nominating committee are weakly associated with lower levels of equity-based
compensation (coefficient significant at the 10% level).
69
In Column (2) there does not appear to
be any significant relation between cash-based compensation levels and the fraction of
compensation committee members on the nominating committee. Given these two results, it is
not surprising that Column (3) documents a weak negative association between High Frac of
Comp on Nom and total CEO compensation. This result follows as a natural extension to the
finding that boards with high cross-membership between the compensation and nominating
committees allocate less compensation in the form of equity and equal levels of cash-based
compensation. Columns (4) - (9) document similar findings where the sample has been
partitioned by board size. In sum, Table 2.5 documents that high cross-membership between the
compensation and nominating committees has only a marginal effect on the form of
compensation granted to the CEO.
Turning to the role of the audit committee, the simplest and most tangible measure of
audit committee effort is the observance of a financial restatement. Implementing the GAO
68
Throughout Columns (1) - (3) coefficients attached to the control variables are in congruence with past empirical
findings.
69
An alternative definition of equity-based compensation, equity granted normalized by total compensation, yields
quantitatively identical results, where High Frac of Comp on Nom is negatively related to the measure at the 10% level.
93
Table 2.6: Probability of Restatement and Board Committee Structure
This table reports results for a single cross-sectional logit regression for 644 firm observations in 2004. The dependent
variable in all columns takes a value of one if a given firm issued a restatement in either 2005 or 2006 and zero
otherwise. The following director and firm-level variables are implemented as controls: firm size (log of total assets),
firm age, segments (log of firm business segments), book leverage, R&D (indicator of one if R&D intensity is above
the sample median), standard deviation of monthly returns, board size, independence, market-to-book, industry-
adjusted returns over the prior two years, industry-adjusted ROA, change in industry-adjusted ROA over the prior year,
and high fraction of audit committee members on the nominating committee (indicator of one if the ratio is at the 75
th
percentile or greater). Industry (Fama-French 48 classification) fixed effects are included in all regressions. Standard
errors are computed using robust methods and p-values are reported below coefficients in parentheses.
Full Sample
(1)
Large Boards
(2)
Small Boards
(3)
Firm Size -0.0122 -0.0987 0.0664
(0.92) (0.57) (0.74)
Firm Age 0.0149 -0.0053 0.0233
(0.18) (0.78) (0.13)
Firm Segments 0.1897 0.1640 0.2488
(0.35) (0.63) (0.44)
R&D 0.0462 0.0890 -0.1293
(0.89) (0.88) (0.85)
Std of Ret 5.115 2.679 9.882
(0.10) (0.59) (0.04)
Board Size 0.0589 0.1091 0.3639
(0.42) (0.49) (0.11)
Independence -1.561 -0.9662 -2.309
(0.33) (0.70) (0.39)
Market-to-Book -0.5269 -0.4904 -0.4326
(0.04) (0.19) (0.15)
Industry-adjusted ROA 0.8613 -2.106 0.4745
(0.74) (0.70) (0.86)
Industry-adjusted Returns -0.3156 -1.394 -0.1833
(0.39) (0.06) (0.72)
Change in ROA -2.555 4.020 -5.905
(0.55) (0.65) (0.39)
High Frac of Audit on Nom 0.1741 0.3424 0.2575
(0.59) (0.43) (0.56)
N 644 287 357
χ
2
115.79 58.95 102.67
94
dataset of audit restatements over the 2005-2006 period, I run a single cross-sectional logit
regression where the dependent variable takes a value of one if the audit committee files a
restatement during this period.
70
All independent variables in Table 2.6 denote firm and board
characteristics in 2004. Across the full sample (Column (1)) and the two sub-samples (Columns
(2) and (3)), the relation between the probability of a restatement and the cross-membership
between the audit and nominating committees is positive but not significant. In untabulated
analysis, I also consider two alternative indicators of audit committee effort. Following the
methodology of Dechow and Dichev (2002) (and an alternative specification following Klein
(2002)), I examine whether accrual quality/earnings management is associated with the cross-
membership between the audit and nominating committees, controlling for a variety of firm-level
factors (i.e. firm size, intangible assets, operating cycle). No consistent and significant results
manifest which would indicate an association between the two measures. In addition, an
investigation into the relation between board sub-structure and the revelation of fraud (AAERs)
also yields no significant findings.
71
In conjunction, the results do not establish any material
association between audit committee performance, or accounting quality, and the cross-
membership between the audit and nominating committees.
Finally, an additional measure of director effort stems from the governance decisions of
the board. A variety of empirical works have documented the valuation effects of antitakeover
provisions. Gompers et al. (2003) and Bebchuk et al. (2009) develop aggregate measures of
antitakeover protection and demonstrate how firms with such provisions underperform compared
70
91 out of the 644 firm observations declare a restatement between 2005 and 2006. 40 of these observations come
from large boards (ten or more member boards), and 51 come from small boards.
71
Data for the revelation of fraud (charges brought against the firm regarding financial reporting violations) comes
from a publicly available repository of Accounting and Auditing Enforcement Releases (AAERs) issued by the SEC
(http://www.sec.gov/divisions/enforce/friactions.shtml). Typically charges pertain to managers overstating income or
manipulating costs and liabilities in statements.
95
Table 2.7: Changes in Corporate Governance and Board Committee Structure
This table reports results for a single cross-sectional logit regression for 615 firm observations in 2004. The dependent
variable in column (1) takes a value of one if a firm removes a poison pill provision between 2004 and 2006, and takes
a value of zero otherwise. The dependent variable in column (2) takes a value of one if a firm removes a classified
board provision between 2004 and 2006, and takes a value of zero otherwise. The dependent variable in column (3)
takes a value of one if a firm removes a classified board or poison pill provision between 2004 and 2006, and takes a
value of zero otherwise. The dependent variable in column (3) takes a value of one the E-Index of a firm decreases
between 2004 and 2006, and takes a value of zero otherwise. The following director and firm-level variables are
implemented as controls: firm size (log of total assets), E-Index, CEO tenure (log of tenure), board size, independence,
market-to-book, industry-adjusted returns over the prior two years, director mean ownership, and high fraction of board
members on the nominating committee (indicator of one if the fraction of board members on the nominating committee
is at the 75
th
percentile or greater). Industry (Fama-French 48 classification) fixed effects are included in all
regressions. Standard errors are computed using robust methods and p-values are reported below coefficients in
parentheses.
Removal of
Poison Pill
(1)
Removal of
Class Board
(2)
Removal of
Class Board or
Poison Pill
(3)
Decrease in
E-Index
(4)
Firm Size 0.8077 0.6455 0.6801 0.4721
(0.01) (0.01) (0.00) (0.00)
E-Index -0.0980 -0.2152 0.1537 0.6187
(0.73) (0.47) (0.45) (0.00)
CEO Tenure 0.0960 -0.2447 -0.1692 -0.0559
(0.75) (0.33) (0.37) (0.73)
Board Size -0.0676 0.1533 0.0912 0.0091
(0.64) (0.24) (0.33) (0.91)
Independence 5.61 1.51 3.30 2.10
(0.19) (0.63) (0.19) (0.30)
Market-to-Book -0.0861 0.0334 0.0854 0.1737
(0.75) (0.88) (0.61) (0.20)
Ind Adj Ret -0.9762 -1.22 -1.34 -0.8056
(0.19) (0.10) (0.03) (0.05)
Director Own -18.27 0.3445 0.1892 0.0302
(0.08) (0.10) (0.26) (0.82)
High Frac on Nom 0.4419 0.1368 0.3872 0.1255
(0.51) (0.85) (0.49) (0.81)
N 358 345 489 615
χ
2
80.12 67.09 89.35 77.91
96
to their benchmarks. Bebchuk and Cohen (2005) explicitly examine one central determinant of
managerial entrenchment, classified boards, and demonstrate the costs and negative valuation
effects of this board feature. If antitakeover provisions (poison pills, classified boards, etc.) serve
to protect management from the discipline of the market for corporate control, then it stands to
reason that if outside directors represent the interests of shareholders, it is their duty to remove
such provisions.
72
Following the methodology of Cai et al. (2009), I examine how the size of the
nominating committee relates to the probability of removing particular entrenchment provisions.
73
Table 2.7 presents the results for a series of logit models where the dependent variable takes a
value of one if a particular provision is removed between 2004 and 2006. A single cross-
sectional regression is run in each column, where all independent variables are constructed in
2004. For the 615 firm observations, 31 firms removed classified board provisions over this time
period, and 32 firms removed poison pill provisions over this period. Column (1) presents the
results pertaining to the probability of removing a poison pill (conditioned on the firm having
such a provision in 2004). The size of the nominating committee (High Frac on Nom) has no
significant relation to the probability of removing such a provision. Columns (2) - (4) document
similar results for the probability of removing a classified board provision and the probability of a
decrease in the E-Index (indicator variables). In total, given these findings and the results
documented in prior tables, it is difficult to conclude that the size of the nominating committee or
the cross-membership between the nominating committee and other monitoring committees has a
significant material impact on the type or quality of board policy decisions.
72
See Scharfstein (1988) Jensen and Ruback (1983) for further evidence on how antitakeover provisions protect and
entrench management from outside control.
73
In the context of shareholder votes, Cai et al. (2009) investigate how increased pressure from shareholders (lower
votes) affects the board’s ability to improve governance standards (remove entrenchment provisions).
97
2.4 Conclusion
In this empirical investigation, I examine how the sub-structure of the board relates to outside
director turnover and board-level policies. In particular, I analyze how the size of the nominating
committee and the cross-membership between the nominating committee and other monitoring
committees influence the self-evaluation process of the board and the quality of board-level
decisions.
First, I observe that large nominating committees are associated with lower levels of
outside director turnover. Directors sitting on boards with small nominating committees (under
60% outside director membership) have a 25% higher probability of being removed as compared
to directors sitting on boards with large nominating committees. Further, when examining
director turnover at the committee level, boards structured with high cross-membership between
the nominating committee and other monitoring committees (audit, compensation) are also
associated with lower turnover rates. In essence, the self-assessment issue faced by large
nominating committees serves the interests of outside directors by providing greater levels of
board-position security (insulation).
Next, I examine whether this greater level of security offered by large nominating
committees translates to lower effort on the part of outside directors. Investigating a variety of
audit, compensation, and governance decisions, I find no conclusive evidence that high cross-
membership between the nominating and other monitoring committees (large nominating
committees) is associated with lower board effort or weaker policy decisions that are not in the
interests of shareholders.
While past empirical work has documented the effect that inside directors have on board
policies when they hold committee positions (Newman and Mozes (1999); Shivdasani and
Yermack (1999); Klein (2002)), the results presented in this investigation extend our
98
understanding of the operations of the board by focusing on the board-level decisions of outside
directors in isolation from the impact of inside directors. The findings provide evidence that in
post-SOX environment where all monitoring committees are entirely comprised of outside
directors, the manner in which the board structures its committees does have an effect on the
decisions of the board. In particular, the empirical analysis extends the literature on board
structure by detailing the self-evaluation problems of large nominating committees, and how this
ultimately functions to protect outside directors’ board positions. In total, the documented results
serve as a basis for future investigations into issues regarding shareholder proxy access, the
composition/formation of the nominating committee and the internal governance of the board.
99
Chapter 3
The Internal Governance of the Board
3.1 Introduction
The Sarbanes-Oxley Act of 2002 (SOX) and the subsequent NYSE/NASDAQ listing requirement
changes constituted a significant shift in the governance standards applied to U.S. public firms.
Recent empirical studies demonstrate that these regulatory events not only had a mechanical
influence on the responsibilities of directors, but also a real impact on the operations of the board,
the risks associated with holding board positions, and most importantly for the purposes of this
study, the role of the nominating committee within the board.
74
In this paper, I extend our
understanding of the structure of the nominating committee and explore how this committee is
constructed to handle the board’s internal governance in the post-SOX period. In particular, I
develop and test three primary hypotheses regarding the association between the bargaining
position of the board (directors), firm determinants, and the composition of the nominating
committee. First, I document that high demand and strong performing directors have a greater
likelihood of serving on the nominating committee. Next, I show that firms that attract these high
demand directors and firms associated with higher risk levels construct larger nominating
74
Linck et al. (2009) provide evidence that the work required of directors, the compensation granted to directors, and
the liability risks associated with holding board positions all increased substantially following SOX.
100
committees, offering greater board-position security to retain their directors. In total, the results
are consistent with the contracting relationship between outside board members and shareholders.
Past research and anecdotal evidence suggest the self-evaluation issue faced by the
nominating committee within the board. William Allen notes that ‘The hardest thing to ask a
director to do is to assess the productivity or contributions of other board members.’
75
If having
directors effectively monitor their own performance constitutes a difficult task, then the
challenges associated with self-assessment should increase in proportion to the size of the
nominating committee. Examining director departures over the post-SOX period, Horstmeyer
(2011a) lends support to this conjecture by demonstrating that nominating committee size is
negatively related to outside director turnover. Hence, this result suggests an alternative,
secondary role for the nominating committee in the modern board - aside from its responsibilities
associated with seeking new director talent, the nominating committee may also aid outside
directors through a lower risk of removal and greater position security on the board.
Moreover, recent empirical studies have also explored the involvement of shareholders in
the director election process. Bebchuk (2003) documents that most director elections are
uncontested, and the few elections which are contested usually result with an unsuccessful rival.
76
Further, Cai et al. (2009) show that votes exceeding 90% are the norm, even for poorly
performing directors, and that lower levels of votes have little effect on the election of directors.
77
Together, these two works demonstrate that the monitoring and replacement of board members is,
by-and-large, left to the discretion of the board itself.
75
See ‘Emerging Trends in Corporate Governance’, a supplement to Corporate Board Member, 2001.
76
See Bebchuk (2007) for a similar set of findings. In particular, the author highlights that over a ten-year period
(1996 to 2005) only 118 director elections were contested and two-thirds of such elections resulted in a rivals defeat.
77
Though, Cai et al. (2009) document that directors attending less than 75% of board meetings and those receiving a
negative ISS recommendation receive 14% and 19% fewer votes, respectively.
101
In addition, the formation and existence of large nominating committees is quite
common. In fact, the average NYSE board has 50% of its outside directors serving on the
nominating committee and 7% of NYSE boards are structured such that all outside directors hold
the nominating committee position. Thus, if shareholders rarely intervene in director elections
and large nominating committees suit the interests of outside directors by insulating them from
turnover, how can we reconcile such findings with the prevalence of large nominating
committees? This paper functions to explain these ostensibly conflicting observations.
To address these issues, I detail how the bargaining position of outside directors and the
board is associated with the formation of the nominating committee using a sample of NYSE
firms over the 2005 to 2009 period. To begin, I examine how director performance, director
bargaining power, and a director’s control within the board relate to the probability of holding a
nominating committee position. To proxy for the bargaining position of the outside director I
consider three measures: the age of the director, the number of other directorships held (busy
director), and the director’s job title (CEO director).
78
To capture the director’s performance
within the board and the director’s control within the board I use the tenure of the director and the
shares held by the director, respectively. Implementing these five proxies, I find strong support
for the idea that the director’s bargaining position and performance are positively related to the
probability of holding a nominating committee position, and ultimately inconclusive evidence
regarding the association between the director’s control over board proceedings (director
ownership) and the probability of holding a nominating committee position.
Following this, I explicitly test how the size of the nominating committee (at the firm-
level) relates to the bargaining position of the board and firm determinants. I find strong support
78
Fahlenbrach et al. (2010) demonstrate that CEOs serving on boards are the most coveted of all board members and
CEOs are more likely to join boards of large established firms. Fich (2005) finds high market reactions to director
appointments when appointees are CEOs of other firms.
102
for the notion that boards structured with a greater number of these high demand directors (high
fraction of CEO directors and old directors) and strong performing directors (high tenure boards)
form larger nominating committees to attract and secure their directors’ services.
79
In addition,
firms that are performing well and firms associated with higher risk levels also create larger
nominating committees. The results demonstrate that the bargaining position of the board, the
performance of the board/firm, and the risk levels inherent to the firm are all positively associated
with nominating committee size - outside directors who hold seats on such boards are granted
greater levels of job security via the nominating committee.
Finally, I examine the relationship between outside director compensation and the size of
the nominating committee.
80
If allocating more outside directors to the nominating committee
functions as a form of security for board members, then, conditional on the bargaining position of
the board, outside directors may accept lower levels of compensation if offered larger nominating
committees. Investigating this conjecture, I observe no definitive relation between nominating
committee size and director salary, equity-based compensation, and total compensation (once
controlling for measures of board bargaining power). I cannot conclude that the security of the
nominating committee and director compensation levels are substitutes.
This paper contributes to a long and established literature on board structure (Hermalin
and Weisbach (1988, 1998); Raheja (2005); Boone et al. (2007); Harris and Raviv (2008)) by
addressing the issue of nominating committee construction within the board. The results extend
our understanding of board sub-structure in two ways. First, the findings help explain the
apparent conflict between the prevalence of large nominating committees and the insulating
79
These results persist to the greatest extent when examining small boards (small firms). Considering that small boards
are less prestigious positions, in all likelihood such firms will have to bargain to a greater extent to attract high demand
directors.
80
See Brick et al. (2006), and Ryan and Wiggins (2004) for similar investigations into the determinants of director
compensation.
103
nature of this board feature. Second, the empirical results demonstrate how the formation of the
nominating committee is consistent with the bargaining/contracting procedure between
shareholders and outside directors.
This paper proceeds as follows. Section 3.2 details the development of the hypotheses
and the form/role of the nominating committee. Section 3.3 presents the data construction and
summary statistics. Section 3.4 presents the empirical methodology and results. Section 3.5
concludes the paper.
3.2 Role of the Nominating Committee & Development of Hypotheses
3.2.1 Regulatory Environment and Background Literature
Before addressing the issue of nominating committee formation, it is important to describe the
U.S. regulatory environment and how it relates to this investigation. In 1999, NASDAQ and
NYSE implemented exchange listing rules requiring the complete independence of audit
committees. Three years later, following the corporate scandals at WorldCom and Enron, the
Sarbanes-Oxley Act of 2002 (SOX) was adopted with the intention of being a comprehensive
solution to corporate malfeasance. Included in the law were rules to formally declare audit
committee independence, improve the quality of financial statements, and strengthen the
enforcement of securities law. NYSE and NASDAQ took the regulatory requirements one step
further in 2003 by mandating that publicly listed firms have a majority of independent directors
on their boards. Moreover, NYSE and NASDAQ both set rules for the composition of other
committees, refined the SOX definition of ‘independence’, and required the financial literacy of
audit committee members.
Though the two exchanges implemented very similar standards regarding audit
committee composition, their rules pertaining to membership on the compensation and
104
nominating/governance committees differed slightly. NYSE required that all firms establish
audit, nominating/governance, and compensation committees composed entirely of independent
directors. NASDAQ issued similar requirements, though allowed more flexibility in the
formation and composition of these committees. In particular, NASDAQ did not explicitly
require that firms have nominating or compensation committees, yet compensation payable to the
CEO and other officers had to be approved either by a majority of the independent directors on
the board or a compensation committee of independent directors. Similarly, for these companies,
nominations had to be approved either by a majority of the independent directors on the board or
a nominating committee of independent directors.
81
Aside from these rulings having a mechanical effect on the responsibilities of outside
directors, recent work suggests that these regulatory changes also had a significant impact on
board operations. First, Linck et al. (2009) document that between 1998 and 2004 workloads for
outside directors on the board increased significantly. In addition, the authors also demonstrate
that director compensation and D&O insurance premiums went up dramatically following the
regulatory events. While an increase in pay may purely reflect the fact that directors were asked
to work more post-SOX, the later finding suggests that the risks associated with holding board
positions increased over this time period.
More importantly, various empirical studies provide evidence that these new regulations
also altered the role of the nominating committee within the board. Prior to 2003, inside directors
were not barred from holding the nominating committee position; thus, many boards were
structured such that the CEO had a voting stake in nominating committee decisions. In fact, 20%
of NYSE firms in 1999 had an inside director or the CEO holding a nominating committee seat
and 34% had no nominating committee in place at all. In other words, over half of NYSE firms
81
Also, if a company elects to establish compensation and nominating committees of at least three members, then one
director who is not independent under NASDAQ’s rules may hold a position.
105
had the CEO or insiders presiding over and voting on board composition issues in the pre-SOX
period.
82
Further, Shivdasani and Yermack (1999) demonstrate that the structure of the
nominating committee also had an impact on the selection of directors during this time period. In
particular, the authors show that when the CEO holds a nominating committee position, or no
such committee exists, more affiliated directors are selected to fill board positions.
Yet, this role and form which the nominating committee exhibited prior to the regulatory
events changed drastically between 1999 and 2005. Horstmeyer (2011b) documents that
surrounding the 2003 NYSE mandate requiring the complete independence of the nominating
committee, not only was the CEO removed as a voting member of the committee, but the fraction
of meetings held outside of the CEO’s voting control in the nominating committee increased
substantially.
83
This finding suggests that the operations of the committee shifted away from the
CEO’s influence and toward outside directors working apart from the CEO’s oversight.
Moreover, Horstmeyer (2011a) investigates the relationship between nominating committee size
and outside director turnover in the post-SOX period. The author finds that small nominating
committees (fewer than 60% of outsiders holding the position within the board) are associated
with 25% higher outside director turnover rates as compared to large nominating committees.
84
This indicates that large nominating committees serve the interests of outside directors by
operating as insulating devices (lower probability of director removal).
Taken together, the evidence suggests an alternative role for the nominating committee in
the post-SOX environment. While many contend that the renomination of directors was more-or-
82
Klein (1998) documents similar summary statistics for the nominating committee over the early 1990s.
83
The author also documents that the meetings held by directors on the nominating committee are the fewest out of all
of the primary monitoring committees (compensation and audit committees), highlighting that the workload associated
with holding a nominating committee seat is quite low.
84
Though, Horstmeyer (2011a) finds no conclusive evidence which would indicate that boards with large nominating
committees are associated with lower observable director effort (i.e. poor policy decisions).
106
less a formality in the pre-SOX board, these findings signify that the nominating committee
operates in a far different manner after 2005. In particular, the results highlight that the role of
the nominating committee has, at least partially, transitioned toward one of job security, aiding
outside directors with a lower risk of removal from the board. The implications of this shift in the
role of the nominating committee constitute the central focus of this investigation.
3.2.2 Nominating Committee Formation
In this section, I develop three hypotheses regarding the formation of the nominating committee
in the post-SOX board. In particular, I detail how the size and composition of the nominating
committee relate to firm determinants and the bargaining power of the board (director).
To define the relationship between the bargaining position of the board (director) and the
construction of the nominating committee, it suffices to establish how the preferences of the
board (director) and shareholders relate to the size of the nominating committee. If board
members within any particular board derive benefits from their board positions, either in a
compensation or reputational sense, and such benefits are positively associated (in expectation)
with the size of the nominating committee, it follows that boards that have a greater ability to
influence their own structure (high bargaining position boards) will demand larger nominating
committees. Further, if directors desire to control their own fate with respect to their board
position, then directors that have a greater ability to influence board proceedings (high bargaining
position directors) will demand to hold the nominating committee position.
To demonstrate the first point, consider the interaction between the nominating
committee and shareholders. The nominating committee proposes a slate of directors to be
approved by shareholders. If the slate of directors is not to the liking of shareholders, they then
must pay an intervention cost (or verification cost) to remove directors and change the board to
their desired form. It naturally follows that the greater the size of the nominating committee, the
107
higher the intervention cost to shareholders. Why is this? If we assume that all board members
are better off, in terms of expected benefits, if they propose themselves for board membership as
opposed to stepping down from the board, then this creates a board environment where, if given
the opportunity to do so (i.e. all members on the nominating committee), the board will propose
all incumbent members. On an individual basis, this assumption is not unfounded. Consider the
case of a low ability board member who holds the nominating committee position. As long as the
compensation and reputational benefits of holding the board position outweigh any valuation
gains from removing himself from the board (via personal equity held in the firm), the low ability
board member will still propose to shareholders that he should retain his board position.
85
Given
that board members will propose retaining themselves if given the opportunity to do so,
intervention costs to shareholders will scale with nominating committee size since nominating
committee members will be better able to defend their seats. Shareholders will have to expend
lower effort to remove board members, if they desire to do so, from a ten member board with two
nominating committee members than a ten member board with ten nominating committee
members.
Hence, a board of any particular characteristic (i.e. ability), will always prefer the
security of a large nominating committee. If a board is granted a large nominating committee,
more board members will be able to propose themselves for re-election, and shareholders will
face a higher intervention cost should they try to change the proposed board.
86
This higher
intervention cost implies that shareholders are more likely to approve any slate of directors given
85
See Yermack (2004) for estimates of director incentives via compensation, the opportunity to gain other board seats,
and the risk of replacement.
86
Further, while shareholders might benefit from having more directors on the nominating committee due to an
increase in the board’s ability to identify better potential directors from outside the board, such a positive effect is
minimal in the post-SOX board. Shareholders can always mitigate intervention costs and still access strong search
committees for new director talent through the use of external consultants. High expenditures on board composition
consultants is common in the modern firm/board and aids shareholders by identifying top director talent without the
downside of the self-evaluation issue that comes with nominating committee members.
108
to them, and thus board turnover is lower (expected benefits to the board is higher). As
previously noted, Horstmeyer (2011a) provides supporting evidence for this role of the
nominating committee in the post-SOX environment. The author demonstrates that large
nominating committees are associated with lower rates of outside director turnover. Therefore,
large nominating committees function to provide greater job security to directors.
If, all else being equal, any board collectively stands to benefit from the insulating nature
of a large nominating committee, then one should observe a positive association between the
bargaining position of the board and nominating committee size. If the formation of the board
and the nominating committee follows from a negotiation process between shareholders and
board members, then boards with a higher bargaining position should have an increased
likelihood of successfully demanding large nominating committees. Further, if firms exposed to
higher risk levels need to bargain to a greater extent to attract outside directors, then firm risk and
nominating committee size should also be positively associated. To underscore this shareholder-
board negotiation procedure, consider the work of Hermalin and Weisbach (1998). The authors
present a CEO-board bargaining model where CEOs use their influence, via surplus production,
to negotiate for monitoring dilution on the board. As the bargaining position of the CEO
increases, more insiders are placed on the board. In much the same manner, if high influence
boards can affect production through their policy decisions or monitoring of the CEO, then
shareholders will be more inclined to negotiate with such boards, and offer larger nominating
committees to retain their talent.
Formally, this leads to the following two hypotheses:
H1: Given that directors value the job security provided by large nominating committees
and holding the committee position (Horstmeyer (2011a)), the performance of the director and the
109
bargaining power of the director will be positively associated with the likelihood of holding a
nominating committee position.
H2: Given that directors (and boards as a whole) stand to benefit from the board-position
security provided by large nominating committees (Horstmeyer (2011a)), the bargaining position
of the board, the performance of the firm/board, and firm risk levels will be positively related to
nominating committee size.
Following these two hypotheses, the structure of the nominating committee should also
have an effect on the compensation granted to outside directors. If the board-position security
provided by large nominating committees acts as an alternative to compensation for directors,
then boards with large nominating committees should accept lower levels of compensation, once
controlling for the bargaining position of the board. This leads to the third and final hypothesis in
this exercise:
H3: Conditional on the bargaining position of the board, boards with large nominating
committees will accept lower levels of director total and cash-based compensation.
With these three hypotheses well defined, it is important to note before proceeding that it
is not within the scope of this investigation to address whether firms benefit from certain
nominating committee structures, or which firms are better off with large or small nominating
committees. The central point of this empirical investigation is to examine how the shareholder-
board contracting relationship defines the formation and composition of the nominating
committee.
3.3 Empirics - Data, Variable Specification and Summary Statistics
Following the preceding discussion on the formation of the nominating committee, I now
empirically address the relationship between board member characteristics and the composition
110
of the nominating committee. In the following section I detail the construction of the dataset, the
specification of variables, and summary statistics for the sample.
3.3.1 Data Set Construction and Variable Specification
As previously discussed, the basis for the sample in this study follows from post-SOX NYSE
firms over the 2005 to 2009 time period.
87
NYSE’s definitive 2003 listing requirements
regarding nominating committee independence enables the construction of a ‘clean’ boards
dataset, where the effect and influence of outside directors on the nominating committee is
distinguishable from that of inside directors.
88
Explicitly, to create the sample of post-SOX
NYSE firms I utilize six databases: Compustat, the Thomson Financial Institutional Ownership
database, ExecuComp, the Corporate Library, CRSP, and the IRRC database. From Compustat I
access firm-specific information including: firm age, number of business segments, total assets,
book leverage, R&D intensity (R&D/Sales), ROA (operating income before depreciation over
assets), and market-to-book (ratio of the market value to book value of assets). In addition, all
firm prices and returns are taken from CRSP. To ensure that outliers do not have a material
impact on the results, variables are winsorized at the 1% level.
For data needed to create the necessary proxies for the governance standards within the
firm, I use the IRRC database and the Thomson Financial Institutional Ownership database.
From the Thomson Financial Institutional Ownership database I access Form 13-F statements
which are used for information on the institutional holdings at the firm level. Two primary
variables are constructed from this database - the largest single blockholder and the aggregate
level of institutional ownership by firm (sum of institutional holdings). The IRRC database
87
NYSE’s 2003 listing requirement generally allowed firms until 2004 to comply with the mandates, though some
firms (e.g. those with staggered boards) were granted compliance extensions until 2005.
88
Since the CEO and insiders are prohibited from holding the nominating committee position, this reduces one
dimension of the problem and makes the implementation of the three hypotheses relatively straightforward.
111
provides annual data for approximately 1,500 firms, primarily from the S&P 500 and other large
corporations. Since the data is collected sporadically, the methodology implemented in Gompers
et al. (2003) is used to fill in missing observations. The E-Index used in subsequent sections, a
proxy for firm-specific external corporate governance (shareholder rights), follows from Bebchuk
et al. (2009).
To collect data on board membership and outside director characteristics, I use the
Corporate Library. This database provides information on director affiliation, director tenure,
board size, director age, director shares held, director job title, and the other board positions held
by each director. In addition, the Corporate Library provides a full breakdown of the board
committee structure - detailing which director holds which committee position. To supplement
the Corporate Library with director compensation data, I employ the ExecuComp database.
ExecuComp provides information for each director’s salary (cash-based fees, and other
compensation), total compensation, and equity-based compensation (stock and options awards).
To be included in the final dataset of post-SOX (2005 to 2009) NYSE firms, data on a
firm’s board structure must be available from the Corporate Library. In addition, necessary
financial data (total assets, market-to-book) must be available from Compustat, and the firm must
be present in the IRRC (E-Index), Thomson (institutional ownership), and ExecuComp databases
(director compensation). Since the empirical analysis requires lagged observations, each firm
must have available information for two consecutive years to be included. In addition, all
regulated firms (utilities and financials) are removed from consideration in the forthcoming
empirical analysis.
89
These necessary conditions result in 3,378 firm-year observations and
27,685 director-firm-year observations over the span of the sample period (2005 to 2009).
89
Due to exemptions granted to certain NYSE firms regarding nominating committee independence (i.e. controlled
companies, foreign private issuers, limited partnerships, and other passive organizations), such non-compliant firms are
also removed from the final sample.
112
3.3.2 Summary Statistics
In this section, I provide summary statistics for firm-year and director-firm-year observations in
the sample. Panel A of Table 3.1 includes the mean, median, standard deviation, 25th percentile,
and 75th percentile for all variables detailed. First, firm total assets in the sample exhibits strong
skewness (mean of 10,582 ($MM), median of 2,822 ($MM)), and hence the logarithmic
transform of assets is used throughout the subsequent analysis. Next, the mean (median) ROA
and market-to-book in the sample is 0.147 (0.140) and 1.79 (1.54), respectively. Further, the
average firm is 30 years old (based on Compustat files), has 36% book leverage, and has an R&D
intensity of 0.02. The governance and ownership summary statistics are also consistent with
past studies. The mean (median) E-Index in the sample is 2.48 (2) and the average CEO holds
1.23% of the common shares outstanding. Next, the sum of institutional holdings for a firm-year
observation averages 81% and the average single largest blockholder owns 10% of the firm.
Panel B presents summary statistics for board structure and director characteristics. The
median board size in the sample is 9, while the median level of independence (fraction of non-
employee directors on the board) is 0.875. This finding regarding board independence highlights
that the most common board structure for NYSE firms post-SOX is one in which the CEO serves
as the single insider on the board. Following this, the average board has 20% of its outside
members currently serving as CEOs of other firms, and 34% of its outside members are holding
three or more board positions (busy members). The average board has a set of outside directors
who have held the board position for 6 years, and are 62 years old. With respect to
compensation, the average director receives 182 thousand dollars in compensation per year, and
half of that received is in the form of equity-based compensation (49%).
In addition, Panel B also details statistics for committee size and structure. The average
audit committee has 3.91 members. The average nominating committee has 4.04 members. The
113
average compensation committee has 3.78 members. These committee size statistics are in
accordance with prior studies over the pre-SOX period (Hayes et al. (2004)). Most important for
the purposes of this empirical investigation is the fraction of outside directors who hold the
nominating committee position. The median firm in the sample has 50% of their outside directors
holding a nominating committee seat, while the 25th and 75th percentiles are 37% and 60%,
respectively.
Although not directly noted in the table, 6.81% of firms in the sample have a board
structure where every single outside member sits on the nominating committee. Together, these
results highlight that a significant number of firms have large nominating committees (high
fraction of outside directors holding the position), and that across firms there is a significant
degree of heterogeneity in nominating committee size.
Moreover, since the size of the board may very well affect committee structure and
director compensation, Panel C presents a breakdown of certain statistics partitioned by board
size. Large boards (greater than or equal to 10 members) have a greater fraction of busy and
CEO directors as compared to small boards (fewer than 10 members). In addition, large boards
pay higher levels of director salary and total compensation as compared to small boards. By
construction, large boards have a lower fraction of outside directors sitting on the nominating
committee than small boards (46% v. 57%). Yet, it is important to note that across both sub-
samples (small and large boards) significant heterogeneity in the fraction of directors serving on
the nominating committee still persists across firms - the 25th and 75th percentiles are 33% and
50% for the large boards sample, while the 25th and 75th percentiles are 42% and 66% for the
small boards sample. Further, since board size is a significant determinant of committee
structure, much of the analysis in subsequent sections is partitioned by board size.
114
Table 3.1: Summary Statistics
This table reports summary statistics for the sample of 3378 firm-year observations from 2005 to 2009. The firm
policy descriptive statistics in Panel A include: assets ($MM), book leverage, R&D intensity (R&D/Sales), standard
deviation of monthly returns, business segments, firm age, ROA, and the ratio of the market value to book value of
assets (market-to-book). The governance and compensation descriptive statistics in Panel A include: the Bebchuk,
Cohen, and Ferrell (2009) entrenchment index (E-Index), institutional ownership (aggregate), block (top blockholder),
CEO percent ownership, and CEO tenure. Panel B presents summary statistics for the sample board structure. The
board descriptive statistics include: board size, the ratio of outsiders to board size (independence), the fraction of board
outsiders holding three or more board seats (fraction busy), the fraction of CEO outsiders (by firm), number of board
meetings, the size of various committees (audit, compensation, nominating/governance), the number of committee
positions held by directors, and the fraction of outside directors holding particular committee positions. In addition, the
summary statistics for director characteristics and compensation are also presented. These statistics include: median
director ownership (median holdings of outside directors by firm), median tenure (by firm), median age (by firm), mean
outside director salary, mean equity-based compensation, and mean director total compensation (by firm). Panel C
presents board structure and director compensation statistics partitioned by board size. The board structure and
compensation statistics include: fraction of busy directors, fraction of CEO outsiders, fraction of board members sitting
on various committees, director compensation, median director ownership, median director tenure and median director
age.
Panel A: Firm Statistics Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Financial & Investment Policies
Assets 10581.76 37305.56 1201.06 2822.36 7588.66
Book Leverage 0.357 0.228 0.201 0.344 0.482
R&D Intensity 0.021 0.041 0 0.000 0.022
Std of Returns 0.090 0.045 0.059 0.078 0.110
Segments 3.34 1.95 1 3 4
Firm Age 29.55 14.39 15 34 44
ROA 0.147 0.086 0.101 0.140 0.188
Market-to-Book 1.79 0.936 1.22 1.54 2.06
Governance & Ownership
E-Index 2.48 1.15 2 2 3
Institutional Holdings 0.814 0.153 0.736 0.839 0.935
Block 0.102 0.044 0.071 0.095 0.125
CEO Ownership (%) 1.23 3.32 0.092 0.267 0.737
CEO Tenure 6.28 6.02 2 5 8
115
Table 3.1 (Continued)
Panel B: Board Statistics Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Board & Committee Structure
Board Size 9.66 2.08 8 9 11
Independence 0.845 0.076 0.80 0.875 0.90
Fraction Busy 0.341 0.233 0.143 0.333 0.50
Fraction CEO Directors 0.202 0.156 0.100 0.182 0.30
Fraction Academic Directors 0.017 0.045 0.00 0.00 0.00
Board Meetings 7.98 3.56 6 7 9
Audit Committee Size 3.91 1.03 3 4 5
Nom/Gov Committee Size 4.04 1.51 3 4 5
Comp Committee Size 3.78 1.12 3 4 4
Fraction on Audit 0.502 0.146 0.40 0.500 0.571
Fraction on Nom/Gov 0.517 0.194 0.375 0.500 0.600
Fraction on Comp 0.486 0.161 0.375 0.444 0.555
Committee Positions Held 1.44 0.815 1 1 2
Director Compensation
Director Salary 81.32 43.86 55.00 73.06 96.62
Director Equity-Based Comp 0.497 0.193 0.398 0.510 0.620
Director Total Comp 182.37 104.89 119.72 166.22 216.97
Median Director Ownership 0.034 0.268 0.003 0.008 0.024
Director Tenure 7.12 3.88 4 6 9
Director Age 61.46 4.16 59 62 64
116
Table 3.1 (Continued)
Panel C: Board Structure
Partitioned by Board Size
Mean Std Dev 25
th
Percentile Median 75
th
Percentile
Large Boards (Size>=10)
Fraction Busy 0.381 0.223 0.222 0.363 0.555
Fraction CEO Directors 0.224 0.146 0.111 0.222 0.333
Fraction on Audit 0.445 0.107 0.363 0.444 0.500
Fraction on Nom/Gov 0.460 0.164 0.333 0.444 0.500
Fraction on Comp 0.426 0.118 0.333 0.428 0.500
Committee Positions Held 1.31 0.759 1 1 2
Director Salary 88.11 43.51 62.12 80.57 104.29
Director Equity-Based Comp 0.498 0.183 0.408 0.509 0.606
Director Total Comp 191.77 95.17 136.11 177.39 224.38
Median Director Ownership 0.022 0.246 0.002 0.006 0.017
Director Tenure 7.00 3.14 5 6 9
Director Age 61.57 3.54 59 62 64
Small Boards (Size < 10)
Fraction Busy 0.302 0.237 0.143 0.285 0.50
Fraction CEO Directors 0.181 0.163 0 0.143 0.285
Fraction on Audit 0.557 0.157 0.428 0.500 0.625
Fraction on Nom/Gov 0.572 0.205 0.428 0.500 0.667
Fraction on Comp 0.544 0.175 0.428 0.500 0.625
Committee Positions Held 1.61 0.856 1 2 2
Director Salary 74.22 43.12 51.20 65.28 87.70
Director Equity-Based Comp 0.497 0.203 0.382 0.512 0.636
Director Total Comp 172.53 113.39 102.21 146.97 209.74
Median Director Ownership 0.044 0.287 0.004 0.011 0.031
Director Tenure 7.23 4.49 4 6 9
Director Age 61.36 4.16 59 62 64
117
3.4 Empirical Design
In the proceeding section, I empirically test how the contracting procedure between directors and
shareholders affects the formation of the nominating committee. First, I investigate the
characteristics of outside directors who tend to hold the nominating committee position. Next, I
explicitly test what firm-level determinants lead to large nominating committees. Finally, I
investigate the relationship between the size of the nominating committee and director
compensation.
3.4.1 Director Characteristics and the Formation of the Nominating Committee
First, I address the association between director characteristics and the nominating committee at
the director level. In accordance with H1 if directors value their board seats and negotiate
amongst themselves when forming the nominating committee, then those directors with the
greatest bargaining power (performance/seniority) should be more likely to demand and receive
the security of the nominating committee position. To examine this issue, I focus on five primary
proxies: director tenure, director age, busy directors, CEO directors, and director ownership.
To capture the bargaining position of the director, I opt to consider busy directors, old
directors (greater than 68 years in age), and CEO directors as indicators of high bargaining
position within the board.
90
While past research has documented the downside of busy directors
(Core et al. (1999), Fich and Shivdasani (2006)), recently Lowry et al. (2011) provide supporting
evidence that a busy director’s network of contacts and expertise are valuable for particular
firms.
91
Hence, though it may be ultimately unclear how busy members affect firm performance,
it is appropriate to consider the busy director to be an individual whose services and experience
90
20% of directors in the sample are classified as old directors.
91
Core et al. (1999) demonstrate that busy directors pay CEOs higher salaries and that this excess pay is connected to
lower performance. Fich and Shivdasani (2006) find that busy directors are associated with lower valuation levels.
118
are in high demand. Further, if each of the board positions that a busy director holds is contingent
on firm/board performance at all other board assignments, then we can expect that a busy director
will be more likely to demand the security offered by a nominating committee position to protect
all board seats.
92
Moreover, past empirical work has demonstrated that directors who hold the CEO
position at other firms are highly coveted board members. Fich (2005) finds that investor
reactions to director appointments are significantly higher when appointees are CEOs of other
firms than when they are not. Fahlenbrach et al. (2010) show that CEOs are more likely to join
boards of large established firms. If CEO directors are in high demand, then their increased
bargaining position may allow them to request greater levels of control over board composition
decisions. In a similar fashion, directors who are senior to other board members (old directors)
may also use their elevated position within the board to demand the security of the nominating
committee position.
To proxy for director performance and control over board proceedings, I focus on
director tenure and director ownership, respectively. A high tenure director (10 or more years of
service) may be considered a director that is valuable to the firm and has a proven record of
excellent performance within the board. Thus, such a director is in a greater position to demand
control over the nominating process. To capture a director’s ability to control and influence the
operations of the board, I construct an indicator variable for high ownership directors.
Specifically, high ownership directors constitute directors who hold greater than 0.1% of the
shares outstanding.
93
One may contend that since high ownership directors have greater control
over the policies of the board/firm, their risk of removal from the board is lower. Given a lower
92
Often, firm proxy documents state that a director's board seat is contingent on performance at other jobs (board
positions) and if a material change should occur, then the director may be asked to resign.
93
High ownership directors comprise 11% of director-firm-year observations.
119
risk of removal, such a director may not need to demand the board-position security that comes
with a nominating committee seat. Yet, contrary to this view, since a director can always be
better off in terms of job security when holding a position on the nominating committee, a high
ownership director may still demand and have a high probability of receiving the committee
assignment, despite an ex-ante low chance of removal from the board.
Table 3.2 provides a univariate look at the probability of holding a nominating committee
position partitioned by various director characteristics. In Panel A, each outside director-firm-
year observation constitutes one observation in the table (affiliated directors included), leading to
27,685 total observations over the 2005-2009 sample period. First, busy directors are
significantly more likely to hold the nominating committee position as compared to non-busy
directors (51.6% probability of serving on the nominating committee v. 48.2% probability of
serving on the nominating committee position). In a similar manner, old directors and high
tenure directors are also more likely to hold the position (a difference in probabilities of 7.6% and
7.0%, respectively), while high ownership directors are less likely to hold a nominating
committee position (a difference in probabilities of -6.8%). CEO directors are weakly more
likely to hold a position, but not significant at the 5% level. Partitioning the sample of director-
firm-year observations by board size, similar results hold across large (greater than or equal to 10
members) and small boards.
94
In particular, the differences between busy and non-busy, old and
young, high tenure and low tenure, high ownership and low ownership directors all remain
significant. Yet, the difference between CEO and non-CEO directors is again weak across the
split sample.
94
Using firm-specific measures of ‘director age’, ‘high tenure’, and ‘director ownership’ (i.e. a ranking classification
within each firm) yields qualitatively similar results as those presented.
120
Table 3.2: Univariate Analysis of Nominating Committee Positions
This table reports nominating committee statistics for the full sample of 27685 outside director observations from the
2005-2009 period, partitioned by various director characteristics. For each director characteristic, the probability of
holding a nominating committee position is denoted in the table below. The following director characteristics are
investigated: busy member, age, tenure, ownership, and job title. Busy director takes a value of one if the director
holds three or more board positions. Old director denotes individuals greater than 68 years old. High tenure director
takes a value of one if the director has been on the board for ten or more years. A CEO director is an individual who
concurrently holds the CEO position at another firm. Director ownership takes a value of one if the director holds more
than .1% of the common shares outstanding. The table presents differences in the probability of holding a nominating
committee position across each director characteristic. Panel A presents the results for the full sample of 27685 outside
director observations (including affiliates), and Panel B presents the results for the sample of 24269 outside director
observations (excluding affiliates). Differences in bold represent statistical significance at the 95% confidence level.
Panel A: All Outside
Directors
Busy
Director
Old
Director
Director
Tenure
CEO
Director
Director
Ownership
Full Sample
High (indicator=1) 0.516 0.556 0.543 0.495 0.435
Low (indicator=0) 0.482 0.480 0.473 0.494 0.503
Difference 0.034 0.076 0.070 0.001 -0.068
Large Boards
(Size>=10)
High (indicator=1) 0.486 0.502 0.506 0.467 0.367
Low (indicator=0) 0.430 0.442 0.429 0.451 0.463
Difference 0.056 0.060 0.077 0.016 -0.096
Small Boards
(Size<10)
High (indicator=1) 0.565 0.628 0.598 0.540 0.499
Low (indicator=0) 0.543 0.531 0.531 0.552 0.558
Difference 0.022 0.097 0.067 -0.012 -0.059
121
Table 3.2 (Continued)
Panel B: Outside
Directors (Excluding
Affiliate Directors)
Busy
Director
Old
Director
Director
Tenure
CEO
Director
Director
Ownership
Full Sample
High (indicator=1) 0.527 0.599 0.597 0.514 0.586
Low (indicator=0) 0.507 0.495 0.481 0.514 0.510
Difference 0.020 0.104 0.116 0.000 0.076
Large Boards
(Size>=10)
High (indicator=1) 0.499 0.539 0.557 0.483 0.525
Low (indicator=0) 0.454 0.458 0.437 0.471 0.471
Difference 0.045 0.081 0.120 0.011 0.054
Small Boards
(Size<10)
High (indicator=1) 0.575 0.679 0.657 0.559 0.629
Low (indicator=0) 0.568 0.544 0.538 0.572 0.564
Difference 0.007 0.135 0.119 -0.013 0.065
122
Since affiliated directors (as categorized by the Corporate Library) may in some cases not
meet the NYSE definition of ‘independent director’ and hence be prohibited from serving on the
nominating committee, I eliminate such observations and repeat the univariate analysis in Panel
B.
95
This treatment reduces the number of director-firm-year observations to 24,269. Similar
differences in probabilities persist for busy directors, old directors, CEO directors, and high
tenure directors. Yet, whereas director ownership was negatively associated with the probability
of holding a nominating committee position in Panel A, it is now positively related to the
probability of holding the position (58.6% probability of serving on the nominating committee if
high ownership v. 51.0% probability of serving on the nominating committee position if not high
ownership). Since high ownership directors (outside directors with greater than 0.1% ownership
in the firm) are often former executives and representative blockholders, such affiliated directors
often do not meet the NYSE listing requirements to be eligible to hold committee positions. In
fact, removing the 3,416 affiliated director-firm-year observations from the sample reduces the
number of directors in the sample which are categorized as high ownership by 40%.
Nevertheless, Panels A and B present conflicting results regarding the association between
director ownership and the probability of holding a nominating committee position. It may be the
case that the results documented in Panel A follow from the high correlation between director
ownership and the inability to serve on committees (not meeting the NYSE’s definition of
‘independent director’), yet given the existing board data, I cannot conclusively state whether
director ownership is positively or negatively related to the probability of holding a nominating
committee seat. Moreover, using alternative thresholds of ownership to define high ownership
95
The Corporate Library may categorize directors as affiliates (gray or outside related directors) if the director has a
material relationship with the firm which does not preclude him from holding committee positions (i.e. meets the
NYSE definition of ‘independent director’), or if the director is a non-employee but does have a relationship with the
firm such that he does not meet the NYSE definition of ‘independent director’. The database combines these two types
of directors into one classification denoted as ‘Outside Related’ and does not provide an alternative indicator to
distinguish between the two.
123
outside directors (0.01%, and 0.05%) yields similar conflicting results to those documented in
Panels A and B.
Following the univariate results, I now investigate the relationship between director
characteristics and the probability of holding a nominating committee position in a multivariate
framework through a series of logit regressions. Table 3.3 documents the results where the
dependent variable takes a value of one for a given director-firm-year observation if the outside
director holds a nominating committee position and takes a value of zero otherwise. The primary
independent variables of interest are busy director, CEO director, high ownership director,
director age, and director tenure.
96
Additional controls include board size, independence, an
indicator variable for affiliated directors (gray directors), an indicator for foreign directors, firm
size (log of assets), market-to-book, and firm age.
97
All control variables are taken at the end of
the prior year. Industry (Fama-French 48 classification) and year fixed effects are also included in
all logit models and heteroskedasticity-consistent p-values clustered at the director-firm level are
reported below coefficients.
Columns (1) and (2) present the results for the full sample. First, board size, affiliated
director, and board independence are all negatively associated with holding the nominating
committee seat. Next, in congruence with the results from Table 3.2, the coefficients on the
primary variables of interest - busy director, director age, and director tenure - are all positive and
significantly related to nominating committee membership (at the 1% level). Concurrently
holding a CEO position at another firm is weakly positively related to the probability of holding a
nominating committee position (significant at 7% and 10% levels in Columns (1) and (2),
96
Director age and director tenure are treated as continuous variables, as opposed to indicator variables as in Table 3.2.
Results throughout the table are significant at the 1% level whether this treatment is used or whether ‘high director
tenure’ and ‘old director’ indicators are implemented.
97
Consistent with the prior literature, affiliated outside directors are denoted as outsiders (Coles et al. (2008), Huson et
al. (2001)) and independence is constructed as the number of outsiders divided by the total number of directors.
124
Table 3.3: Determinants of Nominating Committee Holdings
This table reports results for a series of logit regressions on the pooled sample of 27685 outside director-firm-year
observations from 2005 to 2009. The dependent variable takes a value of one if a director holds a nominating
committee position in a given year. The following director and firm-level variables are implemented as controls: firm
size (log of total assets), firm age, market-to-book, board size, independence, gray director (indicator), busy director
(holds three or more board positions), CEO director, foreign director, director age, director tenure, and director
ownership (indicator for greater than .1% ownership in the firm). Columns (1) and (2) denote logit regressions over
the full sample. Columns (3) and (4) denote logit regressions over large board observations (board size greater or equal
to ten). Columns (5) and (6) denote logit regressions over small board observations (board size less than ten). Industry
(Fama-French 48 classification) and year fixed effects are included in all regressions. Standard errors are computed
using robust methods (clustered at the director-firm level) and p-values are reported below coefficients in parentheses.
Full Sample
(1)
Full Sample
(2)
Large Boards
(3)
Large Boards
(4)
Small Boards
(5)
Small Boards
(6)
Board Size -0.1230 -0.1234 -0.0909 -0.0940 -0.2360 -0.2201
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Independence -1.1906 -1.1414 -0.2788 -0.2941 -1.9633 -1.8273
(0.00) (0.00) (0.39) (0.31) (0.00) (0.00)
Gray Director -0.7286 -0.7368 -0.7125 -0.7143 -0.7581 -0.7684
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Busy Director 0.1597 0.1541 0.2054 0.2019 0.0884 0.0861
(0.00) (0.00) (0.00) (0.00) (0.05) (0.06)
CEO Director 0.0625 0.0609 0.1087 0.0997 -0.0097 -0.0091
(0.07) (0.10) (0.01) (0.02) (0.89) (0.90)
Foreign 0.1295 0.1123 0.0205 0.0120 0.3442 0.3332
(0.21) (0.32) (0.85) (0.93) (0.07) (0.10)
High Own -0.2841 -0.2762 -0.2970 -0.2817 -0.3084 -0.3135
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Director Age 0.0148 0.0151 0.0094 0.0095 0.0208 0.0216
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Dir Tenure 0.0259 0.0261 0.0296 0.0296 0.0221 0.0224
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Firm Size 0.0063 0.0089 -0.0296
(0.58) (0.54) (0.23)
Market-to-
Book
0.0873 0.0311 0.1223
(0.00) (0.25) (0.00)
Firm Age -0.0012 -0.0001 -0.0014
(0.24) (0.94) (0.48)
N 27685 27685 15897 15897 11788 11788
χ
2
1306.47 1344.48 598.14 600.40 642.85 684.92
125
respectively). Again, once partitioned by board size, similar results persist. In Columns (3) - (6),
the coefficients on busy director, director age, high ownership director, and director tenure all
have similar levels of significance as compared to Columns (1) and (2). Further, repeating the
analysis with the removal of affiliated directors from the sample yields qualitatively identical
results to those presented in Table 3.3, with the single exception being the results associated with
director ownership. As previously demonstrated in Panel B of Table 3.2, the exclusion of
affiliated directors from the sample functions to change the relationship between director
ownership and the probability of holding a nominating committee position from a negative to a
positive association. A similar transition in the sign of the coefficient occurs in this multivariate
framework, yet all other results remain the same in terms of the significance of coefficients
attached to the bargaining position variables (busy director, director age, and director tenure).
Together, the results of Tables 3.2 and 3.3 provide support for the first hypothesis H1 that
directors with a high bargaining position (old directors, high tenure directors, and busy directors)
are more likely to demand and receive the security of a nominating committee seat.
3.4.2 Determinants of Nominating Committee Structure at the Firm Level
Next, I explicitly examine how firm characteristics and the bargaining position of the board relate
to the structure of the nominating committee (H2). As established in Table 3.3, board members
with greater bargaining power and strong performing directors are more likely to hold nominating
committee positions, but this finding does not necessarily imply that firms create larger
nominating committees to accommodate their presence. To test how board structure and firm-
level determinants relate to the size of the nominating committee, I run a series of OLS
regressions for 3,378 firm-year observations where the dependent variable is the fraction of
outside directors who hold the nominating committee position at the firm in a given year (number
of outside directors on the nominating committee divided by the number of outside directors).
126
Table 3.4: Determinants of Nominating Committee Structure at the Firm-Level
This table reports results for a series of regressions for 3378 firm-year observations from 2005 to 2009. The dependent
variable is the fraction of outside directors who hold a nominating committee position at the firm in a given year. The
following director and firm-level variables are implemented as controls: firm size (log of total assets), firm age, market-
to-book, E-Index, institutional ownership, R&D (indicator of one if the firm has greater than the sample mean level of
R&D intensity), standard deviation of monthly returns, high CEO tenure (indicator of one if the CEO has held the
position for 10 or more years), ROA, board size, independence, fraction of gray directors, fraction of busy directors,
fraction of CEO directors, median director age, median director tenure, and director ownership (median ownership of
the outside directors). Industry (Fama-French 48 classification) and year fixed effects are included in all regressions.
Standard errors are computed using robust methods (clustered by firm) and p-values are reported below coefficients in
parentheses.
Panel A: Full
Sample
(1) (2) (3) (4)
Firm Size 0.0024 0.0051 0.0025 0.0048
(0.65) (0.34) (0.66) (0.40)
Firm Age 0.0001 0.0002 0.0001 0.0001
(0.91) (0.70) (0.92) (0.88)
E-Index 0.0001 -0.0005 0.0001
(0.94) (0.91) (0.98)
Inst Own -0.0055 -0.0301 -0.0057
(0.86) (0.36) (0.86)
Market-to-Book 0.0275 0.0260
(0.00) (0.00)
R&D 0.0269 0.0175
(0.08) (0.21)
Std of Ret 0.0390 0.0628
(0.75) (0.62)
High Tenure CEO 0.0105 0.0137 0.0101
(0.44) (0.32) (0.45)
ROA -0.0904 0.1409 -0.0594
(0.29) (0.04) (0.49)
Board Size -0.0316 -0.0317 -0.0324 -0.0322
(0.00) (0.00) (0.00) (0.00)
Independence -0.3876 -0.3695 -0.3701 -0.3575
(0.00) (0.00) (0.00) (0.00)
Fraction Gray -0.0608 -0.0831 -0.0752 -0.0941
(0.19) (0.08) (0.10) (0.05)
Fraction Busy 0.0021 0.0059 0.0019 0.0016
(0.93) (0.81) (0.94) (0.96)
Director Own 0.0358 0.0379 0.0331 0.0357
(0.02) (0.01) (0.02) (0.02)
Fraction CEO
Directors
0.0738 0.0414 0.0727 0.0637
(0.03) (0.14) (0.03) (0.05)
Director Age 0.0043 0.0044 0.0036
(0.01) (0.01) (0.02)
Director Tenure 0.0035 0.0022
(0.01) (0.09)
N 3358 3358 3358 3358
R
2
0.2063 0.2164 0.2142 0.2238
127
Table 3.4 (Continued)
Panel B: Board
Size Partitioned
Large
Boards (1)
Large
Boards (2)
Large
Boards (3)
Small
Boards (4)
Small
Boards (5)
Small
Boards (6)
Firm Size 0.0052 0.0054 0.0055 -0.0048 -0.0053 -0.0021
(0.41) (0.40) (0.45) (0.53) (0.49) (0.78)
Firm Age 0.0004 0.0004 0.0004 -0.0002 -0.0005 -0.0002
(0.45) (0.42) (0.43) (0.74) (0.43) (0.69)
E-Index -0.0031 0.0070
(0.59) (0.25)
Inst Own 0.0632 -0.0283
(0.12) (0.56)
Market-to-Book 0.0081 0.0309
(0.40) (0.00)
R&D 0.0323 0.0031
(0.14) (0.89)
Std of Ret 0.0278 0.0307
(0.89) (0.84)
High Tenure
CEO
0.0118 0.0035
(0.50) (0.85)
ROA -0.0314 -0.0384
(0.85) (0.71)
Board Size -0.0215 -0.0215 -0.0210 -0.0501 -0.0496 -0.0513
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Independence -0.0831 -0.0905 -0.1009 -0.5352 -0.5276 -0.4938
(0.38) (0.33) (0.28) (0.00) (0.00) (0.00)
Fraction Gray -0.0984 -0.0989 -0.0959 -0.0457 -0.0398 -0.0621
(0.10) (0.10) (0.11) (0.50) (0.55) (0.32)
Fraction Busy 0.0220 0.0250 0.0280 0.0146 -0.0081 -0.0112
(0.55) (0.49) (0.46) (0.65) (0.84) (0.64)
Director Own 0.0047 0.0044 -0.0018 0.0439 0.0456 0.0473
(0.51) (0.54) (0.87) (0.00) (0.00) (0.00)
Fraction CEO
Directors
0.0002 0.0042 0.0963 0.0945
(0.99) (0.91) (0.02) (0.02)
Director Age 0.0011 0.0010 0.0061 0.0062
(0.71) (0.78) (0.00) (0.00)
Director Tenure 0.0035 0.0042 0.0047 0.0037 0.0011 0.0013
(0.08) (0.06) (0.05) (0.04) (0.49) (0.41)
N 1604 1604 1604 1754 1754 1754
R
2
0.1161 0.1175 0.1287 0.2301 0.2448 0.2702
128
Control variables implemented in Table 3.4 include firm size (log of total assets), firm
age, high tenure CEO (indicator of one if the CEO has held the position for 10 or more years),
board size, independence, and the fraction of gray (affiliated) directors on the board. In addition,
to control for the governance standards at the firm, I use the E-Index (entrenchment index) and
institutional ownership (aggregate percent of shares held by institutions). To capture the
valuation and performance of the firm, I use the firm’s market-to-book ratio and the firm’s
operating performance (ROA). While, the market-to-book ratio of the firm may also proxy for
the long-run risk that directors are exposed to, both interpretations of the results I present
(forthcoming) are in congruence with shareholder-director bargaining procedure. To proxy for
the risk levels inherent to the firm I use an indicator variable for high R&D firms which takes a
value of one if the firm has greater than the sample mean level of R&D intensity (0.021).
98
In
addition, as an alternative measure of firm risk levels (on a short horizon), I also consider the
standard deviation of monthly returns over the prior year. Finally, the central board-level
variables of interest are the fraction of outside directors who are busy directors (holding three or
more board positions), the fraction of outside directors who are CEO directors (outside directors
who currently hold a CEO position), the median outside director age within the board, the median
tenure of the outside directors, and the median ownership of the outside directors. Industry
(Fama-French 48 classification) and year fixed effects are also included as controls and
heteroskedasticity-consistent p-values clustered at the firm level are reported in all models.
Panel A of Table 3.4 presents the results for the full sample of 3,358 firm-year
observations (20 observations are removed due to incomplete pricing information). In Column
98
Since R&D levels have been used in previous boards research as a measure of information asymmetry between the
firm and investors, outside directors are at a disadvantage in terms of job security at high R&D firms since firm-level
information costs are high. Hence, since information asymmetry between the firm and outside directors is high, such a
measure does appropriately capture risk levels (uncertainty) for directors. See Coles et al. (2008) and Boone et al.
(2007) for more information on the relation between R&D and board structure. See Linck et al. (2008) for information
related to the use of the standard deviation of monthly returns as a measure of information asymmetry.
129
(1) minimal firm-level controls are implemented. First, the coefficients on board size and
independence are negative and significant at the 1% level. By construction, boards that are larger
in size and boards with a greater percentage of outsiders on the board will have a lower total
fraction of outsiders sitting on the nominating committee. Next, Column (1) highlights that firms
with a greater fraction of outside directors currently serving as CEOs and firms with older
directors (higher median outside director age) are associated with a greater fraction of outside
directors holding the nominating committee position (significant at the 3% and 1% level,
respectively). In addition, median director ownership on the board is positively associated with
the fraction of outside directors holding the nominating committee position (significant at the 2%
level). This provides support for the notion that boards where outside directors have greater
control over the proceedings will be better able to demand higher job security in the form of a
larger nominating committee. Yet, contrary to the board-shareholder bargaining hypothesis, the
fraction of busy outside directors serving on the board does not appear to be significantly related
to the fraction of directors on the nominating committee.
Following this, I run a similar set of OLS regressions in Columns (2) - (4), with the
addition of certain controls. In Column (2), median outside director ownership and median
outside director tenure are both positively related to the fraction of directors holding the
nominating committee position (both significant at the 1% level). In addition, Column (2)
demonstrates that high market-to-book firms are associated with larger nominating committees.
In Column (3), I observe that firms with higher risk levels (R&D) and firms with strong operating
performance (ROA) are more likely to form larger nominating committees (significant at the 8%
and 4% level, respectively). However, the coefficient on the standard deviation of firm returns is
insignificant in the model. Finally, Column (4) presents the results with all control variables
implemented. Simultaneously controlling for the market-to-book ratio of the firm and the R&D
130
levels at the firm serves to reduce the significance attached to the coefficient on the R&D
indicator. Yet, Column (4) again provides evidence that the fraction of CEO directors on the
board, the median age of the directors, and the median tenure of the outside directors are all
positively associated with nominating committee size.
Following the results presented in Panel A, I now partition the sample by board size and
run separate OLS regressions across each subset of data (Panel B). Similar findings hold across
each subset of data. One important difference between the two subsets is the observation that the
results pertaining to the relationship between the bargaining power of the board (director age,
director tenure, fraction CEO directors) and the fraction of directors who sit on the nominating
committee appear to hold in a more consistent and significant manner for the small boards sample
(boards with fewer than 10 members) as compared to the large boards sample. In particular, the
coefficient attached to median outside director tenure is significant at the 8% level for the large
boards sample (Column (1)), while significant at the 4% level for the small boards sample
(Column (4)). The coefficients attached to the fraction of CEO directors and director age are
significant at the 91% and 78% level for the large boards sample (Column (3)), while significant
at the 2% and 1% level for the small boards sample (Column (6)). Yet, this finding is not
surprising if we consider which types of firms are more likely to bargain with outside directors
when forming the nominating committee. If small boards proxy for small firms (Boone et al.
(2007)), and small firms are less prestigious board appointments (Fahlenbrach et al. (2010)), then
in all likelihood, small boards will be forced to negotiate to a greater extent to attract and retain
high demand directors. The results of Panel B serve to affirm such a statement.
Numerous robustness checks are also implemented to test the validity of the results
presented in Table 3.4. As a first treatment I remove all affiliated directors from the sample and
repeat the construction of all variables. In addition, as opposed to using continuous measures for
131
outside director age and outside director tenure, I construct indicator variables for the fraction of
old directors on the board (fraction of outside directors over the age of 68) and the fraction of
high tenure directors (fraction of outside directors who have served on the board for more than 10
years). Finally, consistent with previous boards research (Boone et al. (2007)), since board
structure is relatively persistent over time, I eliminate every other year from the sample and repeat
the tests.
99
These three treatments yield qualitatively similar results pertaining to the primary
variables of interest (i.e. all previously detailed outside director bargaining proxies remain
significant at the 5% level).
Moreover, an additional important issue to address is the fact that there was an
insignificant association between CEO directors and the probability of holding the nominating
committee seat (Table 3.2 and Table 3.3), yet Table 3.4 documents a positive relationship
between the fraction of CEO directors on the board and nominating committee size (fraction of
outside directors holding the nominating committee position). How can one reconcile these two
observations? To do so, I present the following example: consider a board with 8 outside
directors, none of whom are CEOs, and three individuals sit on the nominating committee (Board
A). Now, consider a board the same as Board A, but it realizes that 4 of its current 8 members are
of no value (none of them sit on the nominating committee) and it wants to replace them with 4
CEO directors (Board B). To attract these new members shareholders need to offer at least 2 of
them the nominating committee position. Now I have created two boards where we observe the
results documented in Table 3.2/3.3 and Table 3.4. Board A: 0 CEOs, 8 non-CEOs, 3 non-CEOs
sit on the nominating committee. Board B: 4 CEOs, 4 non-CEOs, 2 CEOs sit on the nominating
committee, and 3 non-CEOs sit on the nominating committee. This entails that in a pooled sense
99
Similarly, Boone et al. (2007) also document that board size is highly correlated with firm size. Partitioning the
sample by firm size in Panel B of Table 3.4, instead of board size, yields similar results on the bargaining variables of
interest.
132
(at the director-level), CEOs have an equal probability of sitting on the nominating committee as
non-CEOs (2/4 v. 6/12), yet firms that need to attract CEOs to their boards must create larger
nominating committees to do so - Board A (low bargaining position board with 0 CEOs) has
only 3 members on the nominating committee, while Board B (high bargaining position board
with 4 CEOs) must create a larger nominating committee of size 5 to attract the CEOs as
directors.
The results of Table 3.4 provide support for the second hypothesis H2 - strong
performing firms (market-to-book, ROA) and firms associated with higher risk levels for their
outside directors (R&D expenditures) form larger nominating committees. In addition, firms with
high bargaining position boards and high demand directors (high fraction of CEO directors on the
board, older outside directors and high tenure boards) are more likely to construct larger
nominating committees to attract and retain their directors. These results persist to the greatest
extent for firms that are, in all likelihood, forced to bargain the most to secure their directors
(small boards).
3.4.3 Director Compensation and the Nominating Committee
In this section, I investigate the relationship between director compensation and the size of the
nominating committee (H3). If the creation of a large nominating committee serves as a form of
security for board members, then it may follow that directors will accept lower levels of
compensation if granted large nominating committees, conditional on the bargaining power of the
board. To address this conjecture, I consider three measures of director compensation: salary
(cash-based compensation), equity-based compensation (the fraction of total compensation that is
in the form of equity), and total compensation. All compensation statistics come from
133
ExecuComp over the 2006 to 2009 period and represent the average of all outside directors at the
firm-level (providing 2,391 firm-year observations).
100
In Table 3.5, I construct a matched sample of firms and take differences across matched
pairs to analyze how nominating committee size relates to compensation measures. First, I
partition my sample by nominating committee size at the median fraction of outside directors
holding the nominating committee position (50%). Next, I cross-sectionally match each firm
denoted as having a large nominating committee (greater than 50%) to its closest counterpart in
terms of the bargaining position of the board, conditional on the counterpart having a small
nominating committee (lower than 50%). I take the difference in various compensation measures
between this matched pair to assess the effect of nominating committee size on director
compensation, having controlled for the bargaining position of the board. Table 3.5 matches
firms by three measures of board bargaining power: fraction of CEOs on the board, median
outside director age, and median outside director tenure.
First, in Panel A of Table 3.5, firms are matched by the fraction of CEOs on the board,
and differences in mean director compensation between large and small nominating committees
are presented. Although, not directly noted in Panel A, for the full sample, large nominating
committees (greater than 50% membership) are associated with a mean director salary of 78.93
thousand, mean equity-based compensation of 0.495, and mean total salary of 181.77 thousand.
The corresponding matched sample of small nominating committee firms (with closest fraction of
CEO directors) has a mean director salary of 82.56 thousand, mean equity-based compensation of
0.499, and mean total salary of 182.74 thousand. This leads to a difference in means of -3.63, -
0.004, -0.97, for director salary, equity-based compensation, and total compensation, respectively
100
Given the change in ExecuComp reporting standards between 2005 and 2006, data is taken from 2006 and forward.
The inclusion of 2005 data does not alter results throughout the forthcoming analysis.
134
Table 3.5: Matched Sample Differences in Director Compensation
This table reports differences in outside director compensation statistics between boards with large and small
nominating committees, controlling for the bargaining position of the board. The full sample of 2391 firm-year
observations is partitioned by median nominating committee size for the sample, creating a set of large nominating
committee observations and small nominating committee observations. Each large nominating committee observation
is matched to a corresponding small nominating committee observation with the smallest difference in bargaining
position. The mean and median (in brackets) difference in various compensation measures between matched pairs is
denoted in the panels below. The three director compensation variables of interest are salary, equity-based
compensation and total compensation. Each compensation measure represents the average over the set of outside
directors in a given firm-year. Panel A reports differences for matched pairs where boards are matched by the fraction
of CEO directors on the board. Panel B reports differences for matched pairs where boards are matched by median
outside director age. Panel C reports differences for matched pairs where boards are matched by median outside
director tenure. In addition to the full sample, sub-samples based on large and small boards are also constructed and
corresponding differences in matched pairs are presented. Differences in bold represent statistical significance at the
95% confidence level.
Panel A: Matching by Fraction of
CEO Directors on the Board
Salary Equity-Based
Comp
Total Comp
Full Sample
Difference Btwn Large and Small
Nominating Committees
-3.63 -0.004 -0.97
[-4.35] [-0.006] [-2.84]
Large Boards (Size>=10)
Difference Btwn Large and Small
Nominating Committees
-0.16 0.006 3.48
[2.47] [0.005] [7.70]
Small Boards (Size<10)
Difference Btwn Large and Small
Nominating Committees
3.15 -0.021 6.37
[-0.63] [-0.024] [0.65]
135
Table 3.5 (Continued)
Panel B: Matching by Median Outside
Director Age
Salary Equity-Based
Comp
Total Comp
Full Sample
Difference Btwn Large and Small
Nominating Committees
-1.25 0.004 0.08
[-2.58] [-0.007] [-1.52]
Large Boards (Size>=10)
Difference Btwn Large and Small
Nominating Committees
-1.87 0.010 -0.84
[-0.62] [0.009] [0.25]
Small Boards (Size<10)
Difference Btwn Large and Small
Nominating Committees
2.68 -0.014 2.54
[1.51] [-0.020] [0.95]
Panel C: Matching by Outside
Director Tenure
Salary Equity-Based
Comp
Total Comp
Full Sample
Difference Btwn Large and Small
Nominating Committees
1.58 0.009 1.22
[-1.69] [0.011] [0.58]
Large Boards (Size>=10)
Difference Btwn Large and Small
Nominating Committees
-0.48 -0.002 -1.01
[1.98] [-0.015] [-2.21]
Small Boards (Size<10)
Difference Btwn Large and Small
Nominating Committees
2.51 0.008 5.11
[0.51] [0.014] [3.22]
136
(all insignificant at the 5% level). Further, once partitioning the original sample by board size,
similar weak differences hold. In Panels B and C, I repeat the matched sample construction by
matching firms based on the median outside director age on the board and mean board tenure,
respectively. Again, differences in compensation between large and small nominating
committees (conditioned on bargaining position) do not manifest at a significant level.
In addition to the methodology used in Table 3.5, I also construct matched pairs where
large nominating committees denote boards at the 75th percentile and higher (60% of outsiders
holding the position within the board) and small nominating committees denote boards at the 25th
percentile and lower (37% of outsiders holding the position). Similar insignificant differences in
compensation statistics persist when considering this matched pairs construction. Further, aside
from an investigation into director compensation at the board level, I also consider a matching
procedure at the director level. This entails matching each individual director by their bargaining
position (tenure, age, job title) and then observing the differences in compensation between those
who hold nominating committee positions and those who do not hold nominating committee
positions (having controlled for their bargaining position). Once again, this procedure provides
similarly weak findings as to those presented in Table 3.5.
For robustness, in Table 3.6 I analyze the relationship between nominating committee
size and director compensation in a multivariate framework. Again, the three primary dependent
variables in the table are outside director salary, equity-based compensation, and total
compensation (all averaged at the board level). Following Ryan and Wiggins (2004) I include the
same primary control variables: firm size, E-Index, institutional ownership, R&D, standard
deviation of monthly returns, high CEO tenure, market-to-book, ROA, prior industry-adjusted
returns, board size, independence, fraction of affiliated (gray) directors, and other director
137
Table 3.6: Director Compensation and Firm-Level Nominating Committee Holdings
This table reports results for a series of OLS regressions for 2391 firm-year observations. The dependent variable in
column (1) is the firm-level mean salary for outside directors in a given year. The dependent variable in column (2) is
the firm-level mean equity-based compensation for outside directors. The dependent variable in column (3) is the firm-
level mean total compensation for outside directors in a given year. The following director and firm-level variables are
implemented as controls: firm size (log of total assets), market-to-book, E-Index, institutional ownership, R&D
(omitted), standard deviation of monthly returns (omitted), high CEO tenure (omitted), ROA (omitted), industry-
adjusted returns over the prior year, board size, independence, fraction of gray directors, fraction of outsiders on the
nominating committee, fraction of busy directors, fraction of CEO directors, median outside director age, median
outside director tenure, and director ownership (median ownership of the outsiders). Industry (Fama-French 48
classification) and year fixed effects are included in all regressions. Standard errors are computed using robust
methods (clustered by firm) and p-values are reported below coefficients in parentheses.
Director Salary
(1)
Equity-Based
Comp
(2)
Dir Total Comp
(3)
Firm Size 10.54 0.026 32.48
(0.00) (0.00) (0.00)
E-Index 0.466 0.006 3.49
(0.55) (0.08) (0.06)
Inst Own -14.62 0.257 50.52
(0.04) (0.00) (0.00)
Market-to-Book -0.793 0.025 9.41
(0.56) (0.00) (0.00)
Ind Adj Ret 0.556 0.037 18.67
(0.88) (0.00) (0.01)
Board Size 0.256 -0.004 -2.91
(0.63) (0.10) (0.04)
Independence -13.59 0.115 18.43
(0.28) (0.04) (0.51)
Fraction Gray 43.84 -0.035 64.96
(0.00) (0.37) (0.01)
Fraction on Nom -1.64 0.017 -4.03
(0.74) (0.42) (0.72)
Fraction Busy -2.02 -0.020 -6.61
(0.81) (0.34) (0.55)
Director Own -3.21 0.006 -6.23
(0.55) (0.76) (0.61)
Fraction CEO Directors 5.25 0.001 10.27
(0.35) (0.96) (0.41)
Director Age 0.481 -0.003 0.012
(0.06) (0.01) (0.98)
Director Tenure 0.425 -0.003 1.36
(0.11) (0.02) (0.02)
N 2391 2391 2391
R
2
0.1947 0.2202 0.2813
138
characteristics.
101
Since directors are often paid cash-based fees in proportion to the number of
board/committee meetings in a given year, and the number of meetings held have been
demonstrated to be negatively associated with performance (Vafeas (1999)), the firm’s market-to-
book ratio and industry-adjusted returns serve to proxy for the number of meetings.
Industry and year fixed effects are included in all regressions and heteroskedasticity-
consistent p-values clustered at the firm level are reported. Column (1) presents the results for
outside director salary, Column (2) presents the results for outside director equity-based
compensation, and Column (3) presents the results for total compensation. First, consistent with
the prior literature, Column (1) demonstrates that firm size and the standard deviation of returns
are positively associated with director salary, and that institutional ownership is negatively
associated with salary. Next, the median age of outside directors on the board is positively
related to director salary (significant at the 6% level). Though, with these controls in place for
the bargaining position of the board, Column (1) also highlights that the fraction of directors on
the nominating committee is insignificantly related to outside director salary. In Column (2), the
coefficients on director tenure and director age are both negative and significant. This
demonstrates that older boards and boards with high tenure directors demand lower equity-based
compensation. Yet, the relation between nominating committee size and equity-based
compensation is not significant at any meaningful level. Further, in Column (3), the size of the
nominating committee does not appear to be significantly related to total director compensation.
For brevity, I do not include tables for compensation regressions partitioned by board size -
comparable results are found in such tables.
In sum, Tables 3.5 and 3.6 provide no conclusive evidence that outside directors in large
nominating committees accept lower salaries or lower total compensation, once controlling for
101
See Brick et al. (2006) and Core et al. (1999) for related compensation works which utilize similar controls to those
implemented in this paper.
139
the bargaining position of the board (i.e. weak support for H3). Since the board itself is
responsible for setting director compensation, the lack of a significant finding regarding the
substitution effect between nominating committee size and compensation is not surprising. If
large nominating committees exercise greater control over the composition of the board, then they
may also exercise greater control over their own compensation, contrary to the demands of
shareholders.
3.5 Conclusion
If outside directors face little threat of removal via the ballot box (Bebchuk (2003)), and are
primarily left to monitor their own performance, then how does a board form and design its
nominating/governance committee given the inherent self-assessment issue within the board?
Further, can we explain how the average NYSE firm has 50% of its outside members serving on
the nominating committee, and 7% of firms have boards where all outside directors hold the
nominating committee position, if previous research suggests that large nominating committees
serve to benefit outside directors through a lower risk of removal from the board? This study
functions to reconcile this seeming conflict between the frequency of large nominating
committees and the insulating nature this board feature.
Implementing a set of post-SOX NYSE firms, I develop and test three primary
hypotheses regarding the association between the bargaining position of board members and the
formation/construction of the nominating committee. First, I find strong supporting evidence that
high tenure, old, and busy directors are more likely to hold nominating committee positions.
Next, I document that boards with a greater number of these high demand/strong performing
directors, well performing firms, and firms associated with higher risk levels form larger
nominating committees. In other words, firms with high bargaining position boards grant their
140
outside directors greater control over the nominating process and greater board-position security
via the nominating committee to attract and retain their directors’ services. Finally, examining
the relationship between nominating committee size and director compensation I find ultimately
inconclusive evidence that the two elements are substitutes, once controlling for the bargaining
position of the board.
In total, the empirical results extend our understanding of the structure of the nominating
committee in the modern post-SOX board. In particular, the findings document how the size and
composition of the nominating committee relate to board and firm determinants in a manner that
is consistent with the contracting relationship between board members and shareholders.
141
Bibliography
Adams, R. B. (2003). What Do Boards Do? Evidence from Board Committee and Director
Compensation Data. Working Paper.
Adams, R. B., H. Almeida, D. Ferreira (2005). Powerful CEOs and Their Impact on Corporate
Performance. Review of Financial Studies (18), 1403-1432.
Adams, R. B., D. Ferreira (2007). A Theory of Friendly Boards. Journal of Finance (62), 217-
250.
Adams, R. B., B. E. Hermalin, M. S. Weisbach (2010). The Role of Boards of Directors in
Corporate Governance: A Conceptual Framework and Survey. Journal of Economic Literature
(48), 58-107.
Agrawal, A., M. A. Chen (2010). Boardroom Brawls: Determinants and Consequences of
Disputes Involving Directors, Working Paper.
American Law Institute (1982). Principles of Corporate Governance and Structure: Restatement
and Recommendations, Tentative Draft no. 1. American Law Institute, Philadelphia, PA.
Anderson, R. C., D. M. Reeb, A. Upadhyay, W. Zhao (2009). The Economics of Director
Heterogeneity. Working Paper.
Baker, M., P. Gompers (2003). The Determinants of Board Structure at the Initial Public
Offering. Journal of Law and Economics (46), 569-598.
Bebchuk, L. A. (2003). The Case for Shareholder Access to the Ballot. Business Lawyer (59), 43-
66.
Bebchuk, L. A., A. Cohen (2005). The Costs of Entrenched Boards. Journal of Financial
Economics (78), 409-433.
Bebchuk, L. A. (2007). The Myth of Shareholder Franchise. Virginia Law Review (93), 675-732.
Bebchuk, L. A., A. Cohen, A. Ferrell (2009). What Matters in Corporate Governance? The
Review of Financial Studies 22 (2), 783-827.
142
Boone, A. L., L. C. Field, J. M. Karpoff, C. G. Raheja (2007). The Determinants of Corporate
Board Size and Composition: An Empirical Analysis. Journal of Financial Economics (85), 66-
101.
Brick, I.E., O. Palmon, J. K. Wald (2006). CEO Compensation, Director Compensation, and Firm
Performance: Evidence of Cronyism? Journal of Corporate Finance (12), 403-423.
Business Roundtable (1990). Corporate Governance and American Competitiveness: a Statement
of the Business Roundtable. Business Lawyer 46, 241-252.
Business Roundtable (1997). Statement on Corporate Governance.
Cai, J., J. L. Garner, R. A. Walking (2009). Electing Directors. Journal of Finance (64), 2389-
2421.
Coles, J. L., N. D. Daniel, L. Naveen (2008). Boards: Does One Size Fit All? Journal of
Financial Economics (87), 329-356.
Committee on Corporate Laws. The Overview Committees of the Board of Directors (1979).
Business Lawyer (34) 1837-1866.
Core, J. E., R. W. Holthausen, D. F. Larcker (1999). Corporate Governance, Chief Executive
Officer Compensation, and Firm Performance. Journal of Financial Economics (51) (1999), 371-
406.
Datta, S., M. Iskandar-Datta, K. Raman (2001). Executive Compensation and Corporate
Acquisition Decisions. Journal of Finance 56 (6) 2299-2336.
Davidson III, W. N., T. Pilger, A. Szakmary (1998). Golden Parachutes, Board and Committee
Composition, and Shareholder Wealth. Financial Review (33) 17-32.
Dechow, P. M., I. D. Dichev (2002). The Quality of Accruals and Earnings: The Role of Accrual
Estimation Errors. The Accounting Review (77), 35-59.
Del Guercio, D. G., L. J. Seery, T.Woidtke (2008). Do Boards Pay Attention When Institutional
Investors ‘Just Vote No’? Journal of Financial Economics (90), 84-103.
Denis, D., A. Sarin (1999). Ownership and Board Structures in Publicly Traded Corporations.
Journal of Financial Economics (52), 187-223.
Duchin, R., J. G. Matsusaka, O. Ozbas (2010). When Are Outside Directors Effective? Journal of
Financial Economics (96), 195-214.
Ertimur, Y., F. Ferri, D. Maber (2010). Reputation Penalties for Poor Monitoring of Executive
Pay: Evidence from Option Backdating. Working Paper.
Fahlenbrach, R., A. Low, R. M. Stulz (2010). Why Do Firms Appoint CEOs as Outside
Directors? Journal of Financial Economics (97), 12-32.
143
Fama, E., M. Jensen (1983). Separation of Ownership and Control. Journal of Law and
Economics (26), 301-325.
Fich, E. (2005). Are Some Outside Directors Better Than Others? Evidence from Director
Appointments from Fortune 1000 Firms. Journal of Business (78), 1943-1971.
Fich, E., A. Shivdasani (2006). Are Busy Boards Effective Monitors? Journal Finance (61), 689-
724.
Fich, E., A. Shivdasani (2007). Financial Fraud, Director Reputation, and Shareholder Wealth.
Journal of Financial Economics (86), 306-336.
Gilson, S. C. (1990). Bankruptcy, Boards, Banks, and Blockholders. Journal of Financial
Economics (27), 355-387.
Gillan, S. L., L. T. Starks (2007). The Evolution of Shareholder Activism in the United States.
Journal of Applied Corporate Finance 19 (1), 55-74.
Gompers, P., J. Ishii, A. Metrick (2003). Corporate Governance and Equity Prices. The Quarterly
Journal of Economics 118 (1), 107-155.
Harford, J. (2003). Takeover Bids and Target Directors’ Incentives: The Impact of a Bid on
Directors’ Wealth and Board Seats. Journal of Financial Economics (69), 51-83.
Harris, M., A. Raviv (2008). A Theory of Board Control and Size. Review of Financial Studies
(21), 1797-1832.
Hartzell, J. C., E. Ofek, D. L. Yermack (2004). What’s in it for me? CEOs Whose Firms are
Acquired. Review of Financial Studies (17), 37-61.
Hayes, R. M., H. Mehran, S. Schaefer (2004). Board Committee Structures, Ownership, and Firm
Performance. Working Paper.
Hermalin, B. E., M. S. Weisbach (1988). The Determinants of Board Composition. RAND
Journal of Economics (19), 589-606.
Hermalin, B. E., M. S.Weisbach (1998). Endogenously Chosen Boards of Directors and Their
Monitoring of the CEO. American Economic Review (88), 96-118.
Hermalin, B. E. (2005). Trends in Corporate Governance. Journal of Finance (60), 2351-2384.
Heron, R. A., E. Lie (2007). Does Backdating Explain the Stock Price Pattern Around Executive
Stock Option Grants? Journal of Financial Economics (83), 271-295.
Huson, M. R., R. Parrino, L. T. Starks (2001). Internal Monitoring Mechanisms and CEO
Turnover: A Long-Term Perspective. Journal of Finance 6 (56), 2265-2297.
144
Jensen, M. C. and W. H. Meckling (1976). Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure. Journal of Financial Economics 3 (4), 305-360.
Jensen, M. C., R. S. Ruback (1983). The Market for Corporate Control: The Scientific Evidence.
Journal of Financial Economics (11), 5-50.
Jensen, M. C. (1986). Agency Costs of Free Cash Flow, Corporate Finance and Takeovers.
American Economic Review (76), 323-329.
Jensen, M. C., and K. J. Murphy (1990). Performance Pay and Top-Management Incentives.
Journal of Political Economy (98), 225-264.
Jensen, M. C. (1993). The Modern Industrial Revolution, Exit, and the Failure of Internal Control
Systems. Journal of Finance (48), 830-879.
Kaplan, S. N., D. Reishus (1990). Outside Directorships and Corporate Performance. Journal of
Financial Economics (27), 389-410.
Kesner, I. F. (1988). Directors’ Characteristics and Committee Membership: An Investigation of
Type, Occupation, Tenure, and Gender. Academy of Management Journal (31), 66-84.
Klein, A. (1998). Firm Performance and Board Committee Structure. Journal of Law and
Economics (41) 275-303.
Klein, A. (2002). Audit Committee, Board of Director Characteristics, and Earnings
Management. Journal of Accounting Economics (33) 375-400.
Knyazeva, A., D. Knyazeva, C. Raheja (2009). The Benefits of Focus v. Heterogeneity: An
Analysis of Corporate Boards. Working Paper.
Lehn, K., S. Patro, M. Zhao (2005). Determinants of the Size and Structure of Corporate Boards:
1935-2000. Working Paper.
Linck, J. S., J. M. Netter, T. Yang (2008). The Determinants of Board Structure. Journal of
Financial Economics (87), 308-328.
Linck, J. S., J. M. Netter, T. Yang (2009). The Effects and Unintended Consequences of the
Sarbanes-Oxley Act on the Supply and Demand for Directors. Review of Financial Studies (32),
3287-3328.
Lorsch, J. W., E. MacIver (1989). Pawns or Potentates: The Reality of America’s Corporate
Boards. Boston: Harvard Business School Press.
Lowry, M. B., L. C. Field, A. Mkrtchyan (2011). Are Busy Boards Detrimental? Working Paper.
Mace, M. L. (1986). Directors: Myth and Reality. Boston: Harvard Business School Press.
Marshall, C. D. (2010). Are Dissenting Directors Rewarded? Working Paper.
145
Mehran, H. (1995). Executive Compensation Structure, Ownership, and Firm Performance.
Journal of Financial Economics (38), 163-184.
Murphy, K. J. (1999). Executive Compensation. In: Ashenfelter, O., Card, D. (Eds.), Handbook
of Labor Economics, Vol. 3., 2485-2563.
National Association of Corporate Directors (2001). Report of the NACD Blue Ribbon
Commission on Director Professionalism.
Newman, H. A., H. A. Mozes (1999). Does the Composition of the Compensation Committee
Influence CEO Compensation Practices? Journal of Financial Management (28), 41-53.
Raheja, C. (2005). Determinants of Board Size and Composition: A Theory of Corporate Boards.
Journal of Financial and Quantitative Analysis (40), 283-306.
Ryan Jr, H. E., R. A. Wiggins III (2004). Who is in Whose Pocket? Director Compensation,
Board Independence, and Barriers to Effective Monitoring. Journal of Financial Economics (73),
497-524.
Scharfstein, D. (1988). The Disciplinary Role of Takeovers. Review of Economic Studies (55),
185-199.
Shleifer, A., R. W. Vishny (1986). Large Shareholders and Corporate Control. Journal of
Political Economy, (94), 461-488.
Sonnenfeld, J. A. (2002). What Makes Great Boards Great. Harvard Business Review (80) 106-
113.
Srinivasan, S. (2005). Consequences of Financial Reporting Failure for Outside Directors:
Evidence from Accounting Restatements and Audit Committee Members. Journal of Accounting
Research (43), 291-334.
Vance, S. C. (1983). Corporate Leadership: Boards, Directors, and Strategy. McGraw-Hill, New
York, New York.
Vafeas, N. (1999). Board Meeting Frequency and Firm Performance. Journal of Financial
Economics (53), 113-142.
Weisbach, M. S. (1988). Outside Directors and CEO Turnover. Journal of Financial Economics
(20), 431-460.
White, H. (1980). A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct
Test for Heteroskedasticity. Econometrica (48), 817-838.
Xie, B., W. N. Davidson III, P. J. DaDalt (2003). Earnings Management and Corporate
Governance: The Role of the Board and Audit Committee. Journal of Corporate Finance (9)
295-316.
146
Yermack, D. (1996). Higher Market Valuation of Companies with a Small Board of Directors.
Journal of Financial Economics (40), 185-211.
Yermack, D. (2004). Remuneration, Retention, and Reputation Incentives for Outside Directors.
Journal of Finance (59), 2281-2308.
Abstract (if available)
Abstract
This dissertation is comprised of three essays. ❧ The first essay, titled “Beyond Independence: CEO Influence and the Internal Operations of the Board”, serves to extend our understanding of board control beyond traditional measures of independence. Using a detailed, hand-collected dataset on board sub-structure, I document that the board of directors for the average firm underwent a significant transformation between 1999 and 2005, not in terms of its size or composition (independence), but in terms of the CEO’s participation in the internal decision-making processes of the board. Over this time period, (i) the CEO presided over (had a voting stake in) fewer board meetings
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Costs and benefits of "friendly" boards during mergers and acquisitions
PDF
Executive compensation: the trend toward one size fits all
PDF
Shareholder litigation as a disciplining device: evidence from firms' financial policies
PDF
Essays in corporate finance
PDF
Internal capital markets and competitive threats
PDF
Essays in behavioral and financial economics
PDF
Slashing liquidity through asset purchases: evidence from collective bargaining
PDF
Share repurchases: how important is market timing?
PDF
Essays on the firm and stakeholders relationships: evidence from mergers & acquisitions and labor negotiations
PDF
CEO selection performance: does board experience matter?
PDF
Essays on information and financial economics
PDF
Essays on commercial media and advertising
PDF
The structure of strategic communication: theory, measurement, and effects
PDF
Incentives and relative-wealth concerns: theory and evidence
PDF
Three essays on young entrepreneurial firms
PDF
Who manages the firm matters: the incremental effect of individual managers on accounting quality
PDF
Dividend policy, earnings announcements, and analysts' earnings forecasts
PDF
Essays in empirical asset pricing
PDF
Do shared auditors improve audit quality? Evidence from banking relationships
PDF
Empirical essays on alliances and innovation in the biopharmaceutical industry
Asset Metadata
Creator
Horstmeyer, Derek
(author)
Core Title
Three essays on board structure
School
Marshall School of Business
Degree
Doctor of Philosophy
Degree Program
Business Administration
Publication Date
06/05/2012
Defense Date
06/04/2012
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Board of Directors,corporate governance,Finance,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Murphy, Kevin J. (
committee chair
), Ozbas, Oguzhan (
committee chair
), Matos, Pedro (
committee member
), Matsusaka, John G. (
committee member
), Tan, Guofu (
committee member
)
Creator Email
horstmey@usc.edu,horstmeyer@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-45888
Unique identifier
UC11289474
Identifier
usctheses-c3-45888 (legacy record id)
Legacy Identifier
etd-Horstmeyer-879.pdf
Dmrecord
45888
Document Type
Dissertation
Rights
Horstmeyer, Derek
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
corporate governance