Download Effects of Local Director Markets on Corporate Boards

Effects of Local Director Markets on Corporate Boards*
Anzhela Knyazeva
University of Rochester
Diana Knyazeva
University of Rochester
Ronald Masulis†
University of New South Wales
First version: November, 2008
This version: August 4, 2011
Abstract
This study explores the effects of local director labor markets on corporate board structure. We
examine whether firms can more easily attract independent directors to their boards when a larger
pool of prospective directors (executives of other firms) is located nearby a firm’s headquarters.
Empirically, firms close to large pools of local director talent have more independent boards and
more local independent directors. Proximity to local pools of specialized expertise - academic, legal,
financial, technological or similar growth experience - is associated with greater board representation
of such directors. Local labor markets are less important for well established, highly visible firms.
Using the local director pools as an exogenous predictor of board composition, we reexamine the
relations between board independence and firm value, performance and CEO compensation for small
and medium size firms.
JEL: G30, G34
Keywords: board of directors, board expertise, director labor market, location, firm value, firm
performance, CEO compensation
*
We thank Bernard Black, Qianqian Du, Kose John, Jason Kotter, Rafael LaPorta, John Long, Vikram Nanda, Subrata Sarkar, George
Serafeim, Cliff Smith, Jerry Warner, Toni Whited, Wanli Zhao, and conference and seminar participants at the 2009 European Financial
Management Symposium on Corporate Governance and Control, 2009 Financial Management Association Annual Meetings, 2010 Asian
Financial Management Association Meetings, 2010 Australasian Finance and Banking Conference, 2011 SFS Cavalcade, University of
Michigan, Frontiers of Finance Conference at ISB Hyderabad, Dartmouth College, and University of Queensland for valuable comments.
The authors gratefully acknowledge the financial support of the William E. Simon Graduate School of Business Administration and the
Financial Markets Research Center at Vanderbilt University.
†
Corresponding author: Ronald Masulis, Australian School of Business, University of New South Wales, Sydney, NSW 2052, Australia. Email: [email protected].
1. Introduction
Recent research on the role and effectiveness of boards of directors has emphasized the
optimal choice of board composition, while recognizing that one size does not fit all in the
design of corporate governance mechanisms (e.g., Adams and Ferreira, 2007; Coles, Daniel and
Naveen, 2008; Boone et al., 2007; Raheja, 2005). One stream of research focuses on independent
director representation on the board and concludes that the nature of a firm’s investment
opportunities affects its demand for outside directors with particular attributes that enhance a
board’s advisory and monitoring roles. Another stream of research emphasizes private benefits
of control and CEO influence over board nominations to explain the level of board
independence. In contrast, this study explores an important overlooked question, namely the role
of supply in the labor market for directors in affecting board composition. Specifically, we
investigate whether a firm’s ability to recruit qualified independent directors with particular skill
sets to its board is affected by the supply of nearby prospective directors.
We conjecture that the local labor market for prospective directors has an important
bearing on board appointment decisions due to the need to recruit outside directors, who have
lower costs of serving on boards of local firms. Empirically, we find that the local supply of
potential directors near a firm’s headquarters has important effects on board structure and
independent director representation, even after controlling for demand related determinants of
board design found in earlier research. Firms located in the vicinity of larger pools of prospective
directors have a higher percentage of locally employed independent directors and higher
independent director representation on their boards. Firms constrained by a limited local pool of
qualified prospective directors rely more heavily on gray directors and insiders for their board
appointments, in place of independent directors. The size of the local pool of active executives
1
also has a positive effect on a firm’s proportion of independent directors with executive
experience. Furthermore, firms located near major universities, law firms, technology firms, and
financial institutions have higher percentages of independent directors with academic, legal,
technological and financial expertise, respectively. Our main results establish a significant role
for the supply of prospective directors in explaining board composition. Given these results and
the general stationarity of headquarters locations, we use the local supply of prospective directors
to predict the level of board independence and then use it to reexamine the relation between
board composition and firm value and performance and CEO compensation and turnover.
The availability of local director talent is relevant for the firm’s board selection process
for several reasons. First, the ability to hire non-local directors is reduced by their higher costs of
attending board meetings, including obvious transportation costs as well as their higher time
costs spent traveling to board meetings far from their home firms’ offices. Such opportunity
costs could be considerable for directors with full-time positions in distant locations.
Second, local directors can acquire more soft information and better quality information
about managerial decisions and operating performance at a lower cost. Such directors have the
advantage of location specific knowledge and expertise that enhances their capacity to monitor
and advise management and participate actively on the board. In contrast, remotely located
directors must expend greater effort to oversee managers effectively from a distance.
Existing evidence corroborates a nontrivial impact of distance in a variety of situations
where monitoring is valuable. For example, Lerner (1995) argues that proximity to a venture
capitalist’s (VC’s) office affects its willingness to invest in a firm or join a firm’s board since it
is generally less costly to oversee local firms than more distant ones. Bengtsson and Ravid
(2009) argue that the VC’s proximity to a portfolio firm facilitates soft information transfers and
2
lowers monitoring costs and find that proximate VCs offer entrepreneurs less harsh incentive
contracts, with fewer investor-friendly cash flow contingencies. Masulis et al. (2011) find that
foreign independent directors, who are far removed from a firm, are less likely to attend board
meetings and that boards that include them tend to offer excessive CEO compensation and
restate earnings more often due to financial misreporting and firms with these directors are
associated with significantly poorer firm performance. Coval and Moskowitz (2001) find that
mutual fund managers earn higher abnormal returns on nearby investments, while Bae et al.
(2008) illustrate the information advantage of local stock analysts.
While there are several reasons why firms may want to recruit independent directors
locally, qualified directors are a scarce human resource and locating a willing candidate can be
time consuming. For example, replacing a director following an unexpected departure takes a
considerable amount of effort and time (on average 185 days, according to Nguyen and Nielsen
(2010)). Moreover, an average senior executive holds less than one outside board seat,
suggesting a reluctance by executives to become overcommitted with outside responsibilities
(Perry and Peyer, 2005; Ferris et al. 2003). If the local market for prospective directors is thin,
firms may be unable to obtain the desired number of experienced independent directors in the
local market.
Given the weaker director reputation benefits that small, less established firms offer, as
well as the challenges posed by firms in poorly accessible areas, such firms may face greater
difficulties in attracting qualified independent experts to their boards. Faced with a combination
of limited supply of local director talent (especially independent directors with specialized
expertise) and high costs of finding and attracting non-local independent directors and assuring
their active board participation, firms may appoint fewer independent directors or fewer directors
3
with specific expertise to their boards. Thus, we hypothesize that the availability of prospective
independent directors in the local labor market affects a firm’s choice of board structure and that
this effect is much stronger for small and medium size firms. This hypothesis is tested holding
constant a firm’s demand for independent directors with specific skills, which is a function of the
costs and benefits a firm realizes from intensive oversight and advisory services of these
independent directors.
The above discussion motivates our main empirical predictions, which are specified
below. First, firms located near larger pools of prospective directors are expected to have a
higher fraction of independent directors on their boards, holding other factors constant. Based on
the Guner et al. (2008) and Linck et al. (2008b) findings that most outsiders come from nonfinancial executive or similar industry backgrounds, we focus on the pool of current officers and
directors at nearby firms as the primary source of prospective directors likely to be easiest to
recruit. Further, firms located near larger such local director pools should draw a higher
proportion of independent directors with executive expertise, especially from nearby firms, all
else equal.
Second, while the median independent director is a Main Street executive, other common
career paths for outside directors include executive positions at financial institutions and legal or
non-corporate backgrounds such as academia, politics, government service, non-for-profits etc.
Averaging Linck et al.’s (2008b) annual samples, directors with executive experience account for
over forty percent of independent directors (with retired executives accounting for another
twenty-five percent), financial directors - ten percent, lawyers - seven percent, academic
directors - four and a half percent, and the remainder - under ten percent. Holding other factors
4
constant, a firm’s proximity to these types of director talent is expected to have a positive effect
on the presence of independent directors with specialized expertise on firm the board.
Third, different classes of firms need outside directors with differing types of expertise.
Growth (mature) firms are likely to seek outside directors with a similar growth (assets-in-place)
profile and if they are near many other firms with similar attributes, they should find it easier to
recruit outside directors with similar corporate experience, improving their outside directors’
overall fit with the firm’s needs. Similarly, proximity to directors with specialized business
backgrounds (high-tech and R&D experience) is expected to facilitate inclusion of such directors
on the board.
Fourth, we consider firm and industry characteristics that can affect the dependence of a
firm’s board composition on the local director labor market. Qualified prospective directors have
opportunity costs of joining a company’s board. A prospective director is more likely to accept a
board appointment at a company with more visibility, which offers greater director reputation
benefits. We therefore expect directors to be more willing to assume distance-related costs to
join the boards of larger, more established firms. Since potential directors have limited time to
serve on outside boards and realize greater reputation benefits from serving on boards of larger
firms, they will be more reluctant to serve on smaller firms’ boards. Thus, firms with less
visibility are expected to face greater hurdles in attracting non-local directors and so are more
likely to rely on the local supply of director talent as a primary source of independent directors.
We also assess whether local labor market constraints are less binding for firms that are
more readily accessible due to their proximity to a major airport. The presumption is that
prospective directors will be expected to bear lower travel costs and thus will be willing to travel
further to join a firm’s board. As a consequence, these firms could be less constrained by local
5
labor markets. Weather and climate could also affect how a prospective director views the
expected travel conditions and the potential for serious delays. So, we also assess if the weather
or climate at a firm’s headquarters is an exogenous factor which affects a nonlocal director
candidate’s willingness to serve on a board.
The prior analysis raises an important question as to whether we effectively control for a
firm’s demand for independent directors generally and those with specific expertise in our
analysis of the supply effects of the local director pools. Our primary approach is to include in
our analysis a number of determinants of a firm’s choice of board composition documented in
existing studies of boards. Following Coles et al. (2008), Denis and Sarin (1999), Linck et al.
(2008a), and related work, we include as control variables: firm size, growth opportunities, firm
age, intangible assets, R&D intensity, monitoring by institutional investors, the extent of
takeover defenses, the strength of managerial ownership incentives, and CEO tenure. To account
for unobservable time-invariant differences in demand for outside directors or specific types of
outside director expertise, we include industry fixed effects. We also incorporate time fixed
effects to account for aggregate trends in demand for board independence and director expertise.
Furthermore, for several tests we exploit the relatively greater demand for particular types of
outside director expertise in certain industry sectors, such as knowledge intensive and high tech
industries, regulated industries, and the financial industry. We separately examine whether firms
more likely to benefit from local directors with particular expertise, such as growth firms and
firms in competitive industries, actually rely heavily on these local markets for this expertise.
Next, we examine whether there is a measurable change in the relation between board
composition and local director markets following an exogenous shock to board composition due
to recent governance reforms that place greater priority on board independence. On the one hand,
6
as the proportion of independent directors on boards increases to comply with these laws and
regulations, firms are likely to search the local labor pool for prospective directors more
intensively. On the other hand, an exogenous increase in aggregate demand for independent
directors could lead many firms to exhaust local labor markets and thus, be forced to pursue a
regional or nationwide search.
Finally, if local labor markets have a significant effect on board characteristics, while
firm location is largely predetermined and thus independent of later corporate decisions, we can
use the exogenous variation in local director market conditions to achieve a more powerful
experimental design for examining the empirical association of board composition with firm
value and other outcomes. We focus on the subsample of small and medium-sized companies,
for which local labor market conditions are most likely to serve as a constraint on the ability to
attract independent directors. Given that we show that local labor market pools significantly
affect board composition, and this variable is missing in most, if not all existing studies of board
composition and firm value, this is a very important question to explore.
Our main findings are as follows. Firms located near larger pools of prospective directors
have a higher proportion of independent directors on their boards. The effect is strongest for
small and medium-sized firms that may come to rely more heavily on local director markets for
board members due to their lower visibility and prestige. Further, firms in locations that are less
readily accessible by air are more reliant on local director pools due to complicated logistics of
attending meetings and otherwise being an active outside director. We also find support for the
conjecture that firms in locations with worse climates (defined as having a lower percentage of
sunshine, more snow days, higher temperature extremes, and higher temperature variability
throughout the year) are more constrained by local director pools. Our results are consistent with
7
the prediction that weather, an arguably exogenous factor, influences a qualified nonlocal
director’s willingness to join the firm’s board (due to weather-related transportation delays and a
less comfortable living environment).
Empirically, firms with access to deeper local labor markets for particular types of
expertise have a higher proportion of such experts among their independent directors and a
higher fraction of these directors are drawn locally. Local director pools continue to have an
effect on board structure following governance reforms associated with Sarbanes-Oxley,
particularly for growth firms and firms in competitive industries. Finally, we examine the effect
of board independence on firm performance and value in a two-stage setting using local director
pool as an instrument for board independence and confirm positive effects of board
independence on firm performance, valuation, CEO incentive pay, and CEO turnover.
Our findings contribute to the extensive literature on corporate boards (see, e.g.,
Rosenstein and Wyatt, 1990; Yermack, 1996; Yermack, 2004; Guner et al., 2008; Fich, 2005;
Linck et al., 2008a; Boone et al., 2007; Masulis and Mobbs, 2011; Brickley et al., 1994). While
these prior studies focus on the demand for independent directors (or particular types of
independent directors), we examine the supply of potential directors in the local labor market and
show that it is an important determinant of board structure.
This study is related to a recent strand of literature that explores the implications of
geography for the quality of financial monitoring. Distance and remote location have been linked
to reduced effectiveness of information collection, monitoring and advising and to lower
institutional ownership and analyst following (Loughran and Schulz, 2005, 2006; Almazan et al.,
2008; Kedia et al., 2008; John et al., 2010; Becker et al., 2011; Coval and Moskowitz, 1999,
2001; Ivkovic and Weisbenner, 2005; Malloy, 2005; Bae et al., 2008; Landier et al., 2007).
8
Becker et al. (2008) use the density of wealthy individuals near a firm’s headquarters to predict
the presence of a large individual blockholder in a publicly held firm. Ang et al. (2009) study the
effects of social circles on CEO pay. Bouwman (2010) documents executive pay spillovers
among neighboring firms.
Our study examines the density of local labor markets for prospective directors with
specific types of expertise on the quality and composition of boards, a critically important
corporate governance mechanism. Two other studies examine proximity of current board
members to a firm’s headquarters. Wan (2008) shows that current board members living close to
headquarters are better informed and realize higher abnormal returns on inside trades, however,
they are less effective as monitors and are associated with lower firm value. Alam et al. (2011)
report that firms with lower asset tangibility and capital intensity have more local directors and
they find that local directors reduce the sensitivity of both CEO pay and turnover to performance.
Our analysis differs in several important ways from this prior literature. We do not
examine how independent directors’ geographic distance from the firm affects CEO
compensation structure or turnover or value. Instead, we evaluate the role that the pool of
potential directors in a firm’s vicinity plays for the structure of the board and the ability to
attract different types of outside expertise, and demonstrate a supply effect on board composition.
Besides documenting the overall importance of local director markets for board composition, we
isolate firm characteristics that increase its dependence on the local pool of director talent.
The remainder of the paper is organized as follows. The second section describes the
sample, data, and variables. The third section presents empirical results and robustness checks.
The fourth section concludes.
9
2. Data
Sample
The sample includes Compustat / CRSP firms with available RiskMetrics data on board
characteristics and takeover provisions, 13f data on institutional holdings, and Execucomp data
on CEO characteristics and share ownership. Financial firms (6000-6999), regulated utilities
(4900-4999), small firms with total assets under $20 million, foreign firms, and firms
headquartered outside the continental US are excluded. The sample period is 1996-2006, and
unless otherwise specified, the final sample includes 1,661 firms. In several tests, sample
selection criteria are modified for robustness. Where director titles are required to identify
executive experts, the sample period starts in 1998 due to availability of primary titles in
RiskMetrics. Measures of academic, legal, and financial expertise and director education are
based on BoardEx data beginning in 2002 and ending in 2008. Geographic coordinate data is
obtained from the 2000 Census and climate data is obtained from the National Climactic Data
Center.
Explanatory Variables
Board characteristics
The main measure of board composition is board independence, defined as the
proportion of the board represented by independent (non-gray) outside directors. Gray directors
are outside board members with familial or business ties to a firm or its senior management,
which create conflicts of interests that can compromise a board’s major functions.1 Robustness
tests also consider inside directors, defined as the proportion of firm officers on the board. For a
1
Gray directors are identified by RiskMetrics based on firm proxy statements and disclosures of related transactions. Examples include
professional service providers; customers; suppliers; former employees of the firm or subsidiaries; designees under an agreement with a group,
directors designated by a significant shareholder, majority holders; family members of executives; recipients of the firm’s gifts; and interlocking
directors (if an inside director of a firm sits on the board of another firm that has an inside director sitting on the first firm’s board).
10
representative sample firm, the board is comprised of 9 directors, of whom 65% are independent
directors, 14% are gray directors, and 21% are officer-directors (including the CEO).
Local independent directors are independent directors employed at companies located
within a sixty-mile radius of the sample firm, as a fraction of independent directors holding
corporate jobs.2 For the average firm in our sample, about forty percent of independent directors
are identified as holding executive positions, of which a third are employed at local firms.
Director expertise contributes to a board’s ability to effectively advise the CEO (Fich,
2005; Adams and Ferreira, 2007; Raheja, 2005). Director expertise is measured by the
proportion of executive, financial, technology, and academic experts among outside board
members. Executive experience is defined as being a current chief executive officer, chief
financial officer, chief operating officer, or inside director on another firm’s board.3 We also
distinguish among directors with more specialized business expertise in R&D intensive, hightech, and high or low growth firms. Financial, legal, and academic expertise is defined by titles a
director holds or held in the recent past, as reported in BoardEx. Detailed definitions are shown
in Appendix A. Summary statistics for the main variables are presented in Table 1.
[Table 1]
Characteristics of local director labor markets
We construct several variables to characterize the availability of prospective directors in
the vicinity of a firm’s headquarters. According to Guner et al. (2008), the most common outside
director career is an executive role at another nonfinancial firm, followed by finance and non 2
We choose the sixty mile (one hundred kilometer) threshold as a rough approximation for the short duration of a one-way commute.
Identification of directors with executive expertise is based on concurrent positions at financial and nonfinancial firms within the RiskMetrics
S&P1500 universe of companies during our sample period, which does not cover very small or privately held firms. Compared to their peers at
large firms, executives of small firms and banks are more likely to be invited to local firms’ boards, so if this produces any effect, it would be to
bias empirical results against finding a significant effect of local director markets on board expertise. As an alternative way of addressing this
issue, we use a more comprehensive executive expertise measure based on the BoardEx database.
3
11
corporate backgrounds. Similarly, in Linck et al. (2008b) corporate directors with nonfinancial
executive backgrounds are significantly more prevalent than financial, nonprofit, consultant or
academic backgrounds.
Thus, our main measure of the local pool of prospective directors (local director pool) is
the density of nonfinancial firms located within a sixty-mile radius of the sample firm.4 Logs are
used to reduce the right skewness of the densities. Since executives of direct competitors are
unlikely to join the board due to concerns about proprietary information and collusion, we
exclude firms in the same four-digit SIC industry. In robustness tests, we expand the local pool
definition to include firms in the same industry, use a hundred-mile radius, include firms from
Alaska, Hawaii, and Canada, and, to account for the possibility that large firms are the primary
source of directors, exclude small firms from the local director pool.
In addition to looking at the size of the local pool of potential executive expert directors,
we consider prospective directors with other backgrounds. Specifically, we examine the log of
the number of (i) financial institutions in the firm’s vicinity, as a source of financial experts; (ii)
academic institutions, defined as nationally ranked business schools and universities listed in US
News and World Report, as a source of directors with academic backgrounds and advanced
degrees; (iii) main offices of law firms, as a source of legal expertise; and (iv) high-tech and
R&D firms, as a source of technology experts.
We use firm headquarters locations reported in Compustat. Headquarters locations are
generally chosen in the early life of a firm, many years prior to the board structure choices we
4
This formulation implicitly assumes that prospective directors employed in top positions at other firms are generally concentrated at a firm’s
headquarters, which seems plausible, and also that the number of executives available to serve on outside boards is comparable across firms. The
former assumption is common in other studies of geographic location cited earlier. We do not expect the latter assumption to have a significant
effect on the interpretation of our findings: it is reasonable to suggest that the number of persons holding top positions at corporate headquarters
(CEO, CFO, COO) is going to be similar across firms. However, to address the possibility that firms of different size supply varying numbers of
prospective directors, we redefine local director pools to contain large companies only in a robustness test in Table 5.
12
examine, and typically for reasons unrelated to board composition, such as close proximity to
major production facilities, important inputs (suppliers, labor), customers, transportation
networks etc. Furthermore, Pirinsky and Wang (2006) find that relocations of headquarters are
infrequent.5 Thus, we treat firm location as a predetermined variable and use the concentration of
companies and organizations in the firm’s vicinity as primarily a source of exogenous variation,
unrelated to the demand for a particular board structure. Headquarters locations are likely to be
most relevant for the hiring of independent directors and the cost of managerial oversight and
board activity. Geographic coordinates are obtained from the US Census (2000) Gazetteer.
Control variables
We use various controls to capture costs and benefits of independence and firm-specific
demand for independent directors. Complex firms are expected to have a greater need for advice
from skilled outside experts (Boone et al., 2007; Coles et al., 2008; Linck et al., 2008a). Our
main proxy for firm complexity is firm size. To capture firm complexity, some of our tests also
use the degree of a firm’s business and geographic diversification, measured by the number of
industry segments and a foreign segment indicator, respectively.
Although independent directors have fewer conflicts of interest than insiders, they also
have less firm-specific knowledge (Fama and Jensen, 1983). For firms with valuable firmspecific growth options, firm-specific knowledge can be crucial and the costs of transferring it to
outsiders can be high, resulting in fewer independent directors on the board (Coles et al., 2008;
5
Even if managers could affect the choice of firm locations, it is not clear why CEOs that face less outsider scrutiny from the board would want
to locate farther from other companies. The extent of hubris may in fact increase with proximity to other companies (networking). We note that
potential preference of poorly monitored CEOs to locate in large cities, which also tend to have a higher density of firm headquarters, was
addressed by controlling for location in a metropolitan area. Pirinsky and Wang (2006) identify 118 relocating firms in 1992-1997 (for
comparison, the full sample included approximately 4,000-5,000 firms per year).
We further checked locations reported in Compact Disclosure for nonfinancial firms with assets of at least twenty million (in line with our
selection criteria in the main sample). We analyzed the locations of firms contained both in December 1995 and January 2001 files, the latest
accessible through our data service at the moment. In line with Pirinsky and Wang, the overwhelming majority of firms remained in the same city
and state between the start and the end of the period (eighty-five percent of firms); some firms have moved to a different city but remained within
a sixty-mile radius of the old location (an additional nine percent); only a small fraction of firms have relocated to a different city more than sixty
miles from their prior location (four percent). Data on coordinates of the new location could not be matched for two percent of firms.
13
Linck et al., 2008a). Sales growth, R&D intensity, and intangible assets (and to some extent, the
standard deviation of returns) are used to capture growth options and the importance of firmspecific information.
Other controls reflect governance mechanisms that could serve as substitutes for or
complements to board monitoring. The Gompers et al. (2003) G Index of takeover defense
provisions is often used to capture managerial entrenchment. A classified board indicator is used
as an alternative measure for robustness. Institutional ownership captures potential monitoring by
blockholders, although it could also reflect institutional preferences towards firm governance.
When managerial and shareholder interests are aligned through higher CEO ownership, the need
for board monitoring is expected to decrease (Raheja, 2005). CEO characteristics may also affect
optimal board composition. More influential CEOs with longer tenure may require greater board
monitoring (Raheja, 2005). Alternatively, if a manager’s tenure is a function of ability, then
CEOs with longer tenure require fewer outside experts on the board. Firms with older CEOs
nearing retirement may add inside board members to facilitate internal succession (Linck et al.,
2008a; Hermalin and Weisbach, 1998).
It should be recognized that governance controls such as the G Index, institutional
ownership or CEO characteristics could be chosen simultaneously with board independence, so
their coefficients are more indicative of an association than of a causal relation with board
composition. Indicators for firms located in large (top ten) and medium-sized (next forty) cities,
based on the US Census (2000), are added in some robustness tests. To assess whether our
findings are affected by potentially endogenous controls, we also report estimates for a baseline
specification that only includes our variables of interest and industry and year fixed effects. All
14
specifications include three-digit SIC industry and year fixed effects to capture industry and
temporal variation.
Two-stage analysis
In the latter portion of our analysis, we explore the effects of board independence. Twostage least squares estimation is used to estimate the effects of board independence on firm
decisions and outcomes such as total CEO pay, the proportion of incentive pay in total pay, CEO
turnover and firm performance measured by Tobin’s q (market-to-book ratio) and its operating
performance (return on assets (ROA)). In the first stage estimation, board independence is
predicted by the size of the local director pool, our measure of interest, as well as industry
median level of board independence (in the spirit of the John and Kadyrzhanova (2008) study of
peer governance effects) and large and medium-sized city indicators. These variables affect
board independence but do not directly influence firm outcomes. The relevance of geographic
predictors of board structure is shown by the main board independence regressions and first stage
identification statistics. In the second stage, firm performance and other variables are regressed
on predicted board independence and a set of controls.6 Director labor markets near a firm’s
location are viewed as predetermined given that firms rarely relocate headquarters. The analysis
excludes the largest firms (the top quartile of sample firms based on total asset size), for which
local director markets are less likely to be a constraint, and focuses instead on small and
medium-sized companies, which represent about 3/4rds of the sample of S&P 1500 firms. To
6
One might be concerned that local director markets are capturing underlying variation in economic conditions, and therefore, affecting firm
value directly. Conceptually, geographic pools in the area of the firm’s location should not be directly related to individual valuations given the
predetermined nature of firm location, the few headquarters relocations and, the choice of initial location being made early in a firm’s life for
reasons such as location of suppliers, production inputs and not to secure prospective directors. To address this concern empirically, in unreported
tests we control for valuation in the firm’s state in valuation regressions and obtain similar results; we also add instruments to valuation
regressions and find they are jointly insignificant after other factors are considered; this mitigates potential concerns about direct effects or
validity of exclusion requirement.
15
account for the possibility of correlated observations for a given firm, we use robust standard
errors clustered at the firm level, following Petersen (2009).
3. Empirical Results
As a first step before turning to our multivariate analysis, we examine descriptive
evidence pertaining to our main premise that firms with access to larger local director markets
draw a larger proportion of independent directors locally. As seen from Fig. 1 in Appendix B,
companies near larger local pools of prospective directors have a higher average proportion of
locally employed independent directors. Further, the proportion of locally employed independent
directors with identifiable corporate positions is three times greater for companies in the top
quartile of local director pools compared to companies in the bottom quartile.
Main results
Our multivariate analysis begins with an examination of the association of local director
markets and board composition, controlling for other firm characteristics. OLS estimates of the
percentage of independent directors on the board are presented in Table 2. They show that board
independence has a strong association with the local supply of prospective directors. In column I,
a larger local pool of prospective directors is associated with greater board independence, after
controlling for industry and year fixed effects. This finding supports the hypothesis that board
governance is affected by the size of the prospective director pool in a firm’s vicinity.
[Table 2]
Controls for firm characteristics (size, growth, ROA, age, risk, asset tangibility, R&D
intensity, institutional ownership, and the G index) and CEO characteristics (ownership, age,
16
tenure) are included in columns II and III. The main effect remains unchanged in significance
and the parameter estimates on the local director pool fall only marginally in size.
Fig. 2 of Appendix B summarizes economic magnitudes of the determinants of board
independence from one standard deviation changes in each determinant. The local director pool
effect is larger in magnitude than the individual effects of sales growth, risk, ROA, and CEO
characteristics. It is comparable in size to the effects of asset tangibility and size and has roughly
half the impact of CEO ownership or the G index. Consistent with Coles et al. (2008), Denis and
Sarin (1999), and Linck et al. (2008a), large firms, which tend to be more complex, rely more on
outsiders. Young, high growth firms have fewer independent directors, consistent with such
firms having a greater need for firm-specific inside knowledge. Consistent with Coles et al.
(2008), R&D intensive firms have a higher fraction of independent directors on the board.
Monitoring by institutional investors appears to complement board oversight, although we are
careful to interpret the relation as indicative of association rather than causality. Firms with
managers protected by more extensive takeover defenses tend to have more independent boards,
consistent with the corporate control market acting as a substitute for board monitoring and
disciplining managers. Higher managerial ownership lowers the need for oversight by
independent directors.
Which firms are most dependent on local director markets?
The relation between local director labor markets and board composition could vary
based firm visibility. Directors, especially those holding full-time jobs, internalize the costs of
service on nonlocal boards, including transportation costs and opportunity costs of time spent
traveling to and from an average of seven board meetings a year, plus the expected additional
costs of their time and reputation if the firm faces serious legal or financial difficulties.
17
Prospective directors might be less willing to join the boards of nonlocal firms if such firms have
low visibility, since reputational and career growth benefits are likely to be smaller than
opportunity costs of director time.7
On the other hand, a prospective director interested in reputation building, career growth,
and additional board appointments is more likely to overlook distance costs and accept a
nonlocal board appointment if the firm is large, established, or otherwise highly visible.
Intuitively, a board appointment to a more prominent firm carries with it greater reputational
benefits.8 Since large firms tend to be more complex and given that performing board functions
entails substantial learning about firm practices, service on a larger firm’s board can also lead to
greater gains in director expertise. Finally, larger firms generally offer more generous director
compensation (e.g., Brick et al., 2006; Linn and Park, 2005). Thus, potential directors may be
more willing to accept board seats at larger, higher visibility firms, even when they are located
further away. Therefore, we expect larger, more established firms are more likely to overcome
local labor market constraints and succeed in a national search.
Firm size, firm age, and NYSE listing may all raise a firm’s profile and attractiveness to
potential directors.9 Univariate evidence on local hiring of independent directors by firm
categories is presented in Appendix B, Table B1. Low visibility firms have significantly more
locally employed directors among independent board members with managerial positions at
7
Small firms also pay their directors less (Linck et al., 2008b). In turn, Vafeas (1999) finds no effect of firm size on the number of board
meetings, so director time costs associated with being on these boards appears to be similar.
8
Low visibility firms could counter-balance independent directors’ smaller reputation incentives to join their boards by increasing compensation
for board service, which could attract more non-local directors. However, this possibility is not supported by the evidence since we find the level
of director pay is significantly lower at smaller firms (see, e.g., Brick et al., 2006; Linn and Park, 2005). Further, without asserting causality, we
find in an unreported test, that when we control for director pay based on Execucomp, we find that it does not eliminate the local director pool
effect. We also do not find a significant relation between local director pools and director pay. Potential reputation or human capital benefits are
likely to play a larger role than the pay itself in a director’s utility. Low visibility firms could also draw directors by using attractive off-site
locations for some board activities. However, the main meetings would still be held at headquarters and directors would still weigh opportunity
costs of time against slim human capital gains; moreover, small firms likely have less generous budgets for off-site meetings.
9
Boards of small recently formed firms may also have a greater presence of venture capitalists, and venture capitalists’ board positions were
shown to be affected by distance and location (Lerner, 1995).
18
other firms. The result extends to outside directors more generally (independent as well as gray
directors) and the subgroup of gray directors. Our findings on mean distance between the firm
and outside directors’ full-time position are also consistent with this argument. Based on
univariate tests, high visibility firms have the luxury of tapping a wider, national director pool,
while low visibility firms are constrained by the local supply of prospective directors.
Potential firm benefits of having local directors and, hence, demand for them, could vary
across firms, with some firms being more dependent on directors’ local knowledge and expertise.
Growth opportunities and a competitive industry environment could both increase the need for
directors with local knowledge. Local knowledge could help achieve a higher return on a firm’s
growth options and be essential for promptly identifying and responding to evolving market
conditions in competitive industries where firms operate on narrow margins. Univariate tests
across subsamples based on firm growth opportunities and industry competition are inconclusive
(the significance of the differences in means does not meet the 5% threshold and the tests are
sensitive to the use of the median or quartile thresholds in subsample definitions).
Multivariate evidence on the effect of the local director pool on board independence is
presented in Table 3. In the first set of tests, presented in Panel A, the local director pool has a
stronger effect on board independence for smaller, younger firms, firms with few blockholders,
and non-NYSE listed firms, i.e., firms with less overall visibility.10 Since these visibility proxies
are correlated (see Table B2), we repeat the analysis using a combined visibility measure derived
from a principal components analysis. We obtain similar results. In summary, local director
pools affect board independence at firms with low to moderate visibility (roughly three-quarters
10
In an unreported robustness test, we separately consider firms deemed more vulnerable to takeovers and hence potentially less attractive to
prospective directors due to career concerns. Specifically, we expect firms in industries with more merger and acquisition related delistings and
firms publicly listed less than ten years ago to be more vulnerable to takeovers or liquidations. For such firms, the local director pool has a
significant positive effect on board independence. For all other firms (older firms in industries with fewer merger and acquisition related
delistings), the local director pool effect is only marginally significant.
19
of the sample). High visibility firms appear to be much less constrained by the local supply of
director talent. Intuitively, a firm’s visibility is a plausible source of variation in prospective
directors’ willingness to bear the costs of distant board meetings and as a consequence, the extent
to which a firm is constrained to search locally for prospective directors. In addition, we examine
a proxy for the visibility or economic activity level of the firm’s region, on the assumption that it
is easier to attract directors if the firm’s state is experiencing relatively high levels of business
activity, which could offer prospective directors valuable synergies and networking benefits,
offsetting the disutility of serving on a nonlocal board. We use an indicator that equals one if
total corporate assets of public firms located in the state exceed the national median, and zero
otherwise. All else equal, reliance on a local director pool is greatest for firms in states with low
levels of business activity, as measured by low overall assets under corporate control.
[Table 3]
Table 3B exploits variation in the difficulty of attracting nonlocal directors due to varying
accessibility of air travel depending on the firm’s location, as well as in the overall attractiveness
of a firm’s location. Similar to John et al. (2010), we use distance to the closest airport as a proxy
for direct transportation costs. We use sixty- and thirty-mile thresholds to identify firms that are
located “far from airports”. We expect such firms to rely more heavily on local director talent
than firms with similar-size local director pools that are headquartered closer to airport hubs.
Indeed, we find that proximity to a large airport reduces the binding nature of a local director
pool. Board composition of firms close to airport hubs is less dependent on local director pools.
We further conjecture that nonlocal directors are more willing to join boards of firms in
attractive locations. We examine a more exogenous measure that is unlikely to be related to
economic factors affecting the firm’s or a prospective director’s opportunity set. Metrics based
20
on climate at the firm’s main location seek to capture the attractiveness of the location itself to a
prospective nonlocal director, who might travel there for board meetings or on other occasions to
monitor the firm. Although directorships may very well be strictly about ‘business first’, there is
some asset pricing evidence, that, for example, sunshine affects investor mood (Hirshleifer and
Shumway, 2003). Furthermore, inclement weather conditions are more likely to result in travel
delays and less pleasant traveling conditions, holding distance constant. We conjecture that the
following climate characteristics can affect nonlocal director willingness to join a board:
frequency of snow days and average snowfall (values below median are classified as favorable);
percent sunshine (values above median or alternatively, values above sixty percent, are classified
as favorable); an indicator for extreme versus moderate temperatures (the first measure
characterizes annual temperature means outside the 25th-75th percentile range as extreme; the
second measure characterizes areas with the coldest winters (January lows in the bottom 25%) or
hottest summers (July highs in the top 25%) as extreme, and other days as moderate); and an
indicator for high versus low within-year variation in temperatures (whether the standard
deviation of monthly temperature means or lows is above or below median).
Empirically, we find that firms in areas with the worst climates are more constrained by
local director markets. Local director labor constraints are relaxed for firms in locations with
better climates, all else equal. The evidence is consistent with the notion that considerations
related to the attractiveness of the firm’s location, in this case, typical weather in the area, impact
local director labor market constraints.
Table 3C examines differences in the effect of local director pools on board
independence by segmenting firms in terms of their demand for local director talent. We find that
21
high growth firms and, to a lesser extent, firms in competitive industries are marginally more
reliant on local director pools in attracting independent directors.
In the last set of tests, reported in Table 3D, we reevaluate the relation between local
director pools and board independence in the aftermath of governance reforms (Sarbanes-Oxley
and governance rules in exchange listing requirements). More stringent board governance
standards, including a majority of independent directors and independent director representation
on key committees, could have prompted firms to expand their director search beyond their local
director pools in the post-reform period. However, we find that local director pools remain a
significant, albeit smaller, factor in the appointment of independent directors, even as
noncompliant firms increase board independence in the post-reform period.
Alternative measures of local director talent
Next we examine the sensitivity of our main findings to variable definitions, sample
selection criteria, and additional controls. Alternative definitions of the local director pool
variable are examined in Table 4. Our main measure of the local pool of prospective directors
includes executives at nearby US firms. We expand the sixty-mile radius (roughly 100
kilometers) around a firm’s headquarters to a hundred-mile radius in column I. Canadian firms
are added to the pool in column II. The local director pool coefficient remains positive and
significant and retains its economic magnitude.
[Table 4]
Our measure of the local pool of prospective directors includes executives from local
firms of varying sizes. However, firms may prefer to hire independent directors with executive
experience at larger firms, or at least similar size firms. As a sensitivity test, we measure the
director pool with only firms of similar or larger size in column III. It is possible that all firms
22
prefer directors with experience at medium to large-size firms. In column IV, the local director
pool measure excludes small firms.11 Our main coefficients remain positive and significant.
Our main measure of prospective director supply excludes firms from the same
industry.12 In column V, the local pool measure includes nonfinancial firms from all industries,
including the firm’s own industry. Coefficient estimates are similar to the earlier results. While
our main local directors pool measure includes only Main Street executives, in columns VI-VII
prospective directors from financial firms are included in the measure, either without or with
executives from the firm’s own industry also included. Finally, in column VIII, the local director
pool is allowed to have a different association depending on whether it is below or above the
sample median. The effect remains positive and significant. The difference in coefficients is not
statistically significant (not shown), consistent with board independence having a linear
relationship with the local directors pool measure.
Proportion of inside and gray directors on the board and other board characteristics
In Table 5A, we supplement our previous findings on the proportion of independent
directors by examining the proportion of gray directors and inside directors on the board. In
column I, we test the possibility that firms with weak local director markets rely more heavily on
gray directors. These types of directors may not be as effective in their role as monitors, but they
may offer valuable advisory services and be easier to recruit. We find that the proportion of gray
directors is decreasing in the local director pool. In column II, the fraction of inside directors is
decreasing in the size of the local director pool.
11
We exclude firms with assets below hundred million, which approximately corresponds to the first and second quintiles of assets in the full
sample of nonfinancial firms (including firms with missing governance data) for our sample period.
A small minority of independent directors hold executive positions at another firm in the same industry (among independent directors with
executive positions, only 2% work for a firm in the same four-digit SIC industry and just 3.5% - in the same three-digit SIC industry). This is
likely due to concerns about potential conflicts of interest that may arise over proprietary information or strategic motives involving competitors.
12
23
We further investigate whether the local director pool is related to board size.
Specifically, we check whether proximity of a larger local pool of prospective directors leads
firms to add more independent directors (resulting in larger board size overall) or to replace
inside or gray directors with independent directors (resulting in zero net effect on board size). In
column III, board size is not significantly related to the local director pool, suggesting that firms
near large local director pools substitute independent directors for gray and inside directors.
Finally, in column IV, the proportion of independent directors that hold other
directorships is found to be positively related to the size of the local director pool, reflecting low
opportunity costs of accepting additional board appointments when these firms are nearby.13
Further robustness analysis
Alternative sample selection criteria, variable definitions, and controls are considered in
Table 5B. Since the fraction of independent directors is a proportion, it is bounded between zero
and one percent. For robustness, we regress our control variables on a logit transformation of
board independence (logarithm of the ratio y/(1-y), where y is board independence expressed as
a proportion; see, e.g., Maddala (1983)). Estimates for this specification are reported in column I.
By construction, coefficient magnitudes vary from the earlier regressions of the untransformed
dependent variable. However, the local director pool coefficient retains its sign and significance.
To account for the possibility that location in a big city is correlated with board structure, we
exclude firms in top ten metropolitan areas and then firms located in California, Illinois,
Massachusetts, and New York (columns II-III). In column IV, we lag local director pool one
period to mitigate simultaneity concerns. In columns V-VIII, we control for classified board
presence, business segments, foreign segments, exchange listing, and quadratic firm size term, as
13
This is consistent with sociological research by Kono et al. (1998) that reveals an increased incidence of interlocks among proximate firms.
24
well as substitute market value for book value as a measure of firm size.14 In unreported tests, we
redefine board independence to exclude directors with 5% or larger stakes, cluster errors both by
firm and by year, add Alaska and Hawaii firms (a handful of observations) to the sample,
exclude firms that moved headquarters (alternatively, firms that moved headquarters more than
sixty miles away from their former location). The local director market effect remains
statistically and economically significant.
Several other alternative explanations for the local director pool effect are considered in
Table 5C. Geographic clusters of same-industry firms (columns I-II), variation in population
density (columns III-IV), and regional differences in college education per capita (column V) do
not explain variation in independent director representation. At the same time, local director pool
remains significant. Accounting for state variation in economic conditions (profitability,
investment opportunities, and unemployment) and governance practices (average board
independence) similarly does not explain the local director pool effect, which retains significance
and economic importance.
Executive expertise
The prior analysis of board characteristics focuses on the proportion of outside directors
on a firm’s board. While the proportion of independent directors is an indicator of the degree of
board oversight of a CEO, it is not a sufficient metric of the quality of such oversight or of the
board’s ability to provide expert advice to top management. Outside directors with executive
experience are crucial to shareholder wealth creation (see, e.g., Fich, 2005). Furthermore,
executives of other local firms comprise a significant proportion of independent directors and are
14
Allowing for a quadratic firm size specification, the linear term becomes negative and the quadratic term enters with a positive coefficient: as
small companies grow, they are able to meet their expertise needs internally, but as firms continue to expand and their complexity increases, they
are less able to rely on internal expertise and appoint more independent directors. In another robustness check (not reported), inclusion of a linear
and quadratic market value terms, or replacing market value terms with equity market capitalization does not affect the results.
25
a key component of the local pool of prospective directors. Thus, we next explore the importance
of the local director pool of potential executive directors.
Consistent with the notion that firms look to local executives as a primary source of
outside directors, the proportion of executives among independent directors on a firm’s board is
positively associated with the size of the local pool of executives (column I of Table 6A). The
relation of independent director executive expertise with the pool of potential executives from
similar-size or larger local firms is even stronger (column II). We obtain similar findings for the
proportion of executives among all outside directors (columns III-IV). The result remains
qualitatively similar when we draw executive expertise data from BoardEx15, in order to also
include executives of small publicly listed firms and privately held companies (column V).
[Table 6]
We have seen that board representation by independent directors and those with
managerial expertise increases with the density of the local director pool, consistent with our
hypothesis that local labor markets are important in determining board structure. In our next test
we look at the representation of local executive experts on the board to test the channel through
which the local director pool affects a firm’s board composition. Univariate evidence in Fig. 1 of
Appendix B suggests that firms located near larger local director pools have a higher proportion
of locally employed directors. Consistent with the univariate evidence, the multivariate test in
column I of Table 6B shows that the proportion of independent directors with executive positions
at other local firms is increasing in local director pool. Similarly, the proportion of local outside
directors (independent and gray directors) is increasing in the local director pool (column II).
15
A greater proportion of directors with executive experience are identified using BoardEx data for the sample of boards since BoardEx lists
executive affiliations at smaller firms, outside of S&P1500, as well as at privately held (unlisted) companies. The proportion of executive experts
among outside directors using BoardEx data rises to over sixty percent.
26
The proportion of local gray directors alone is marginally related to the size of the local director
pool (column III). Further, average distance to an independent director’s primary employer is
decreasing in the depth of the local pool of prospective directors (column IV). Overall, firms
located near larger director pools can more readily recruit executives of nearby firms to serve as
independent board members. This finding supports the hypothesis that local director markets
affect director appointments and board independence.
Next, we consider the effect of firm size on the decision to appoint local independent
directors. Univariate evidence in Table B1 indicates that the proportion of locally employed
directors among independent directors with identified executive jobs is at least a third higher for
small firms, relative to large firms. The difference is statistically significant. Consistent with this
univariate evidence, we find in multivariate analysis that the proportion of independent directors
who are local executives is highest for small firms (column I). Also, R&D intensive firms appear
to be more likely to hire local directors, consistent with distance interfering with timely firmspecific information acquisition required by such firms.
We have seen that small firms draw more heavily on local directors. A related question is
which types of firms deliver local director talent. We conjecture that large firms supply more
prospective directors with executive expertise.16 Univariate evidence in Table 1B shows that
officers of large firms are on average twice as likely to hold outside local directorships. This
finding is confirmed in the multivariate analysis in Table 6C. The proportion of insiders with
outside board seats at other local firms is higher for larger firms, even after controlling for large
firms’ possible tendency to be situated near larger local director pools (column I).17 Consistent
16
In unreported univariate statistics, the median firm’s assets are approximately 24% smaller than equal-weighted assets of the companies where
the firm’s independent directors work.
17
At an average firm in our sample, 31% (14%) insiders hold a concurrent appointment at another (local) firm. The average number of
appointments at other (local) firms in our sample is 0.45 (0.17) per insider. Local firms are defined based on headquarters locations within a
27
with this, executives of large firms are more likely to sit on the boards of other local firms and on
average hold more local directorships (columns II-III). Overall, large firms appear to serve as net
suppliers of outside directors with executive expertise to other smaller local firms (even beyond
the direct effect of large firms having more officers).
Retired executives, classified as directors over sixty who have previously held an
executive position, comprised about twelve percent of independent directors, based on Table 1.
These retired executives generally secure directorships while they are active executives at nearby
firms and then they are reappointed after retirement.18 This suggests that our local director pool
measure may also predict board representation of retired executives. On the other hand, some
executives undoubtedly will move elsewhere after retirement and may resign from the board.
Furthermore, some boards can have mandatory or informal retirement rules for directors. Thus,
the extent of retired executives on the board might not be explained by our measure of local pool
size. Empirically, we do not find that firms near larger local pools have more retired executives
as independent directors. Thus, it appears that firms first and foremost turn to the local pool of
current executives when looking for outside directors with general managerial expertise.
Specialized board expertise
Corporate boards require more than general managerial experience from their
independent directors; they also require individuals with specialized knowledge and skills to
advise the board and CEO (see, e.g., Linck et al., 2008b). Similarly, given the heavy time
demands on major classes of potential directors, firms can be expected to first approach local
sixty-mile radius of the firm. When all directors (inside, gray and independent) are considered, the average number of appointments at other
(local) firms in our sample is 0.65 (0.08) per director. Compared to all directors, insiders tend to hold fewer concurrent board appointments,
presumably due to the demands of their main job, however, their board appointments are disproportionately concentrated among local firms.
Figures are based on an inside director’s concurrent independent or gray director status on the board of another firm as reported by RiskMetrics.
18
Using our sample from RiskMetrics, more than four fifths of outside directors classified as retired were active executives earlier in their tenure
as a director.
28
candidates with such expertise to be directors. Thus, recruiting outside directors, firms can be
forced to forego certain valuable types of director expertise, such as academic, technical, legal,
or financial, when there is limited local supply. However, if such specialized expertise is viewed
as essential for corporate boards, representation of such experts on the board could be less
sensitive to local labor market conditions, as firms would search more widely for potential
director experts. Further, demand for such experts is likely to be related to firm characteristics
reflecting the nature of its investment opportunities, expenditures on technology, asset
characteristics, uncertainty, and of course, a firm’s industry, and thus these firms’ choices of
directors are likely to be less sensitive to geographical conditions.
To empirically investigate the demand for independent director expertise, we focus on
subgroups of firms that have similar needs for particular types of outside board expertise.19 We
then examine the relations of particular types of expertise available in local director markets to
the expertise found in the firm’s current group of outside directors.20 After controlling for major
firm characteristics that act as a proxy for a firm’s demand for outside directors with specific
knowledge and skills, we expect the supply of local director expertise to be positively related to
the overall level of this expertise found among a firm’s independent directors.
[Table 7]
In Table 7A, we focus on the specialized professional expertise of outside directors
gained from their current and prior work experience. We first examine the presence of outside
directors with academic work experience, termed academic experts. Besides likely having fewer
potential conflicts of interest than seasoned industry veterans, academic directors can add
19
An alternative approach to control for firm demand for director expertise is to use the full population of firms and then include
firm characteristics as controls in a multivariate regression framework. This alternative approach requires that the required
controls are observable and affect board composition linearly. 20
Due to data availability, information on several types of specialized director skill sets - educational qualifications, legal,
academic, and financial expertise – is obtained from BoardEx (see the Appendix for more details).
29
specialized skills and knowledge and state of the arts perspectives gained from their fields of
expertise to help address a firm’s strategic and operational challenges. Intuitively, we expect the
availability of prospective academic directors to be greater in the vicinity of major university
campuses. Indeed, we find that having a firm’s headquarters near a number of universities is
positively associated with the proportion of outside directors having academic positions. In
columns III-IV, we allow for the possibility that R&D intensive firms are more reliant on
academic directors than other, less R&D oriented firms. The expectation is that R&D intensive
firms have a greater need for outside directors with specialized knowledge, who can offer expert
advice to management on technologically intensive investment projects and critically evaluate
relevant research and firm product innovation. Consistent with this line of reasoning, R&D
intensive firms located near major academic centers include a higher proportion of academic
directors on their boards.
In Table 7B, we examine the local availability of legal and financial experts. Consistent
with a local labor supply effect, we find that the board’s overall legal expertise is positively
related to the local density of the main offices of major law firms, a proxy for the local supply of
prospective directors with legal experience. The effect is concentrated among regulated firms
(financials and utilities), which arguably require a greater degree of legal and regulatory
expertise from board members advising top management. For regulated firms, the presence of
legal experts is also positively associated with proximity to universities, a measure of the pool of
candidates with legal training, some of whom will practice law locally after graduating. In the
last two columns we investigate the financial expertise of outside directors in relation to the local
availability of these types of experts. Outside financial experts are particularly valuable for audit
committees and for providing independent advice to top management on a firm’s financial
30
health, capital structure, and payout decisions. Thus, independent directors with financial
expertise should be important to most, if not all, corporate boards. At the margin, the density of
financial institution headquarters in a firm’s vicinity, which serves as a measure of the
availability of potential local directors with this type of expertise, has a positive effect on the
presence of financial experts on a corporate board. Interestingly, while financial firms are more
likely to appoint outside financial experts to their own boards, they are not more likely to rely on
the local pool of potential financial directors than other firms to fill this need, possibly because
of anti-trust and competitive concerns of having a competitor’s executive on the board.
We next consider directors with technological and R&D expertise in relation to the local
availability of prospective directors with those attributes (measured by the local pool of
executives and outside directors in R&D intensive and high tech firms). In Table 7C, we see that
consistent with our hypothesis about local supply effects, the proportions of outside directors
with technology and R&D expertise are both increasing in the size of the local pool of such
prospective directors. We further conjecture that an important hiring criterion for boards might
be the “fit” between a director’s expertise and the firm’s growth opportunities. For example,
growth firms might seek outside directors able to overcome information asymmetries associated
with growth options and to advise management on innovative projects. In contrast, mature firms
are in need of experts who can help a firm to efficiently manage large free cash flows and
implement strong cost controls, given the scarcity of positive-NPV investment opportunities.
Consistent with this intuition, in columns V and VI, the proportion of independent directors with
experience at firms in a similar growth phase is increasing in the local pool of potential directors
with such expertise (defined as the same quartile of growth opportunities). Overall, a limited
31
supply of local candidates with experience in firms with similar growth patterns clearly affects
board composition with respect to the representation of these independent director traits.
Finally, Table 7D examines the relation of firm proximity to major academic institutions
and outside director educational qualifications. Indeed, the proportion of outside directors with
advanced graduate and professional degrees is increasing in the density of nearby major
universities. Breaking down advanced degrees into different types, we observe that boards have
more outside directors with MBAs, other masters degrees, and doctoral degrees when they are
located near more major universities or business schools (the effect on law degrees is not
statistically significant). To the extent that educational credentials proxy for director expertise,
this evidence further supports the conclusion that a firm’s ability to recruit outside experts to its
board is positively affected by the density of the local pool of academically qualified candidates.
4. New evidence on how board independence affects CEO incentives and firm performance
A number of prior studies report evidence that greater representation of outside directors
on boards lead to gains in shareholder wealth (e.g., Gompers et al., 2003; Cremers and Nair,
2005; Bebchuk and Cohen, 2005; Masulis and Mobbs, 2011; Yermack, 1996; Rosenstein and
Wyatt, 1990; Brickley et al., 1994). However, other studies have raised concerns about
endogeneity of board composition and challenges in identifying the board independence-firm
performance relation, which are summarized below (e.g., Adams et al., 2009; Coles et al., 2008;
Boone et al., 2007). First, an omitted variable bias can arise when certain omitted characteristics
correlated with firm value are also correlated with board independence. For example, growth
firms, which tend to have higher valuations, may appoint more insiders due to the high costs of
conveying proprietary information to outside directors, whereas mature firms, which tend to have
32
lower valuations, may appoint more outsiders to overcome agency conflicts. Unless the relation
is properly identified, the resulting confounding effect can dominate the direct effect of board
independence. Second, a reverse causality concern can arise when strong firm performance is
due to a lack of internal agency problems, which reduces the need for outside monitoring by the
board. Alternatively, top performing firms can be more appealing to qualified prospective
directors, resulting in them making more appointments of outside directors with strong
qualifications. Endogeneity concerns cast doubt on the ability of OLS estimation to establish
causation. A recent study by Nguyen and Nielsen (2010) seeks to address this concern by
looking at market reactions to sudden director deaths, a type of unexpected exogenous shock to
board composition. Although the approach yields interesting insights consistent with our findings
that independent directors add to firm value, it is limited to a small sample of 229 sudden deaths
of directors, including 108 independent directors; moreover, a large majority of the deceased
independent directors are replaced with new independent directors within a year.
Given the lack of consensus concerning the effect of board independence on shareholder
wealth, we offer new evidence less susceptible to endogeneity concerns. Specifically, we exploit
the exogenous variation in the density of local labor markets for director talent, documented in
the previous tables, to instrument for board independence. We then employ this instrument to
estimate the effects of board independence on CEO compensation and turnover, and firm value
and operating performance. All our regression models include industry and year effects to filter
out industry variation and any general time trend in CEO compensation and turnover, and firm
performance and investment opportunities. As discussed above, a limited pool of local director
talent imposes the most binding external constraints on board selection in small and mediumsized firms. Thus, we concentrate our attention on this large subsample (75% of the full sample)
33
in our two-stage instrumental variables model. In Table 8 we use variation in local director pools
and other controls to predict board independence and then in the second stage regressions reexamine the relation of board independence to important board decisions and firm performance
outcomes, such as executive pay and turnover and firm value and performance.
[Table 8]
Existing work reports that firms with more independent boards use more incentive pay to
better motivate managers (e.g., Mehran, 1995). We reexamine the relation between board
composition and CEO pay using local availability of prospective directors as a source of external
variation in board structure. We find clear evidence that more independent boards use more
equity based compensation as a proportion of total CEO pay (column I). All else the same, a one
standard deviation increase in board independence results in a 5.5% increase in the percent of
stock option pay in total CEO pay. The result holds with a broader definition of performance
based pay that includes both stock options and restricted stock grants (column II). In contrast,
there are no significant differences in total pay levels (column III). Thus, rather than curtail
overall pay, independent boards strengthen CEO incentives through greater weight on equity
compensation.
Finally, column IV shows that the proportion of independent directors is positively
associated with CEO turnover, which includes all CEO changes, whether voluntary or forced.
Since firms commonly disguise forced turnover when reporting CEO changes (Weisbach, 1988),
the distinction between forced and voluntary departures is difficult to draw in practice. However,
some instances of CEO turnover are clearly unrelated to performance, namely, CEO deaths.
34
Since they do not offer evidence of board control over agency conflicts, we exclude these
relatively infrequent cases of CEO turnover.21
Next, we evaluate whether board independence, as measured by the local director pool,
has significant effects on firm value (measured by market-to-book ratio) and operating
performance (measured by ROA). We find that the proportion of independent directors on the
board has significant positive effects on firm value and operating performance (columns V-IX).
The documented effects are economically important. All else the same, a one standard deviation
(17%) increase in the proportion of independent directors on a firm’s board results in a 0.17 rise
in a firm’s market-to-book ratio (median is 1.66) and a 1.3% rise in ROA (median ROA is 14%).
The above findings are based on controlling for firm size, firm age, growth opportunities,
volatility, the G index, institutional ownership, dividend yield, CEO characteristics, as well as
industry fixed effects. The signs of these controls are generally consistent with the findings of
prior studies of firm valuation and operating performance. For instance, the G Index enters with
a negative sign in firm value regressions, consistent with Gompers et al. (2003). Institutional
ownership is associated with better firm value and operating performance. CEO tenure, a
possible measure of managerial quality, is associated with better valuation and operating
performance. Board size is negatively related to firm value, consistent with Yermack (1996). In
Table 8, column IX, we add controls for several other corporate governance characteristics,
namely dual class firms, independent director blockholders, and institutional blockholders. These
added controls do not appear to be associated with firm valuation improvements, nor do they
alter the positive relation of director independence to performance.
21
One might wonder if dense local director pools indirectly proxy for a more competitive executive labor market, thus, affecting CEO turnover.
When local director pool is added to the regressions of turnover, it has no significant relation with turnover, which alleviates this concern.
35
The relevance of local director pools for board composition shown earlier (see Table 2) is
confirmed by the first-stage statistics in Table 8. To obtain stronger evidence of causation in our
two-stage analysis, we focus on small and low visibility firms, which are most reliant on local
labor markets for board appointments. The instrument’s excludability is economically clear;
local director labor markets are unlikely to affect firm value, performance or CEO pay and
turnover other than through board composition. Since corporate headquarters relocations are very
infrequent, location can be viewed as a predetermined characteristic, exogenous to decisions
concerning independent director selection. For comparison, OLS estimates of board effects are
shown at the bottom of Table 8. Except for incentive pay, OLS estimates are smaller in
magnitude than IV estimates.
5. Conclusions
In this study, we examine how local director labor markets affect board composition
choices and the quality of corporate governance. We document that the supply of potential
directors in the local labor market strongly affects board composition, especially the proportion
and expertise of independent directors. For instance, in an average sample firm, a third of
independent directors in S&P 1500 firms holding executive positions are employed locally. Our
empirical tests show that access to a larger local pool of prospective directors has a positive
effect on the proportion of independent directors and a negative effect on the proportion of gray
and inside directors found on a typical board of directors. At firms located near larger local pools
of prospective directors, a significantly higher fraction of independent directors are employed
locally. Overall, boards of firms located near larger pools of managerial talent include a larger
percentage of independent directors who are executives from other nearby firms.
36
Our findings provide strong evidence of the importance of local director labor markets
and the notion that distance matters in spite of technological developments that facilitate
teleconferencing and other forms of information sharing over long distances. Although board
meetings typically average only seven a year, the presence of directors at board meetings appears
to be important for the effective execution of their oversight and advisory roles. At the margin, it
appears that distance negatively affects the willingness of prospective directors to serve on
boards, especially those who are full-time corporate and financial executives or holders of other
professional positions, who face heavy demands on their time. Of course, firms may also benefit
from selecting local outside directors to the extent that soft information about director quality
cannot be fully communicated over long distances.
The analysis of our small- and medium-sized firm subsamples indicates that low visibility
firms are most likely to rely on local director labor markets in their search for independent
directors. Prospective directors appear more willing to accept the costs of distance in exchange
for the benefits of being affiliated with the boards of large, established companies. In a similar
vein, firms in more accessible locations (closer to major airports) are less constrained by local
director pools due to the lower expected travel times for nonlocal directors. Further, the climate
near a firm’s headquarters can also affect the strength of the local labor market constraint. Firms
in areas with better climates, measured by less snowfall, more sunshine, temperature extremes,
and less within-year variation in weather, are likely to face fewer expected travel disruptions and
shorter travel times. Thus, firms in these locations are less constrained by local labor markets as
prospective directors expected lower travel costs of serving on these boards.
To corroborate the channel through which the local director pool exerts its influence on
board composition, we examine the proportion of locally employed independent directors on
37
boards and find that it similarly increases in the size of the local pool of prospective directors. In
addition, we examine several dimensions of director expertise, including executive, technology,
financial, academic, and legal expertise, and director education, to proxy specialized skills and
abilities to perform effective monitoring and advisory roles. We find that firms located closer to
a large pool of potential director talent (general managerial or specialized) have a higher
proportion of independent directors with these respective types of expertise on their boards. We
include a number of controls to account for the possibility that a third variable explains variation
in both local labor markets and board structure.
Finally, since the local pool of prospective directors is a powerful predictor of board
independence, we use it to reexamine the effect of board independence on the firm’s bottom line.
We focus on small and medium-sized firms (75% of our sample), which are more likely to be
limited to the local pool of prospective directors due to their low visibility or prestige. Using this
supply constraint on director selection, we find board independence has a significant positive
effect on firm value and operating performance, and CEO turnover and proportion of equity
based pay, but it has no effect on a CEO’s total compensation per se.
38
References
Adams, R., Ferreira, D., 2007. A theory of friendly boards. Journal of Finance 62(1), 217–250.
Adams, R., Hermalin, B., Weisbach, M., 2009. The role of boards of directors in corporate
governance: A conceptual framework and survey. Working paper.
Alam, Z., Chen, M., Ciccotello, C., Ryan, C., 2011. Does the location of directors matter? CEO
influence, information acquisition and board decisions. Working paper.
Almazan, A., de Motta, A., Titman, T., Uysal, V., 2008. Financial structure, liquidity and firm
locations. Working paper.
Ang, J., Nagel, G., Yang, J., 2009. Is there a social circle premium in CEO pay? Working paper.
Bae, K.-H., Stulz, R., Tan, H., 2008. Do local analysts know more? A cross-country study of the
performance of local analysts and foreign analysts. Journal of Financial Economics 88(3),
581–606.
Baginski, S., Hassell, J., Kimbrough, M., 2004. Why do managers explain their earnings
forecasts? Journal of Accounting Research 42(1), 1–29.
Bebchuk, L., Cohen, A., 2005. The costs of entrenched boards. Journal of Financial Economics
78(2), 409–433.
Becker, B., Cronqvist, H., Fahlenbrach, R., 2008. Estimating the effects of large shareholders
using a geographic instrument. Working paper.
Becker, B., Ivkovic, Z., Weisbenner, S., 2011. Local dividend clienteles. Journal of Finance
66(2), 655–683.
Bengtsson, O., and A. Ravid, 2009. The geography of venture capital contracts. Working paper.
Boone, A., Field, L. C., Karpoff, J., Raheja, C., 2007. The determinants of corporate board size
and composition: An empirical analysis. Journal of Financial Economics 85(1), 66–101.
Bouwman, C., 2010. The geography of executive compensation. Working paper.
Brick, I., Pamon, O., Wald, J. K., 2006. CEO compensation, director compensation, and firm
performance: Evidence of cronyism? Journal of Corporate Finance 12(3), 403–423.
Brickley, J., Coles, J., Terry, R., 1994. Outside directors and the adoption of poison pills. Journal
of Financial Economics 35(3), 371–390.
Coles, J., Daniel, N., Naveen, L., 2008. Boards: Does one size fit all? Journal of Financial
Economics 87(2), 329–356.
Coval, J., Moskowitz, T., 1999. Home bias at home: Local equity preference in domestic
portfolios. Journal of Finance 54(6), 2045–2073.
Coval, J., Moskowitz, T., 2001. The geography of investment: Informed trading and asset prices.
Journal of Political Economy 109(4), 811–841.
Cremers, M., Nair, V., 2005. Governance mechanisms and equity prices. Journal of Finance
60(6), 2859–2894.
Denis D., Sarin, A., 1999. Ownership and board structures in publicly traded corporations.
Journal of Financial Economics 52(2), 187–223.
39
Fama, E., Jensen, M., 1983. Separation of ownership and control. Journal of Law and Economics
26(2), 301–325.
Ferris, S., Jagannathan, M., Pritchard, A., 2003. Too busy to mind the business? Monitoring by
directors with multiple board appointments. Journal of Finance 58(3), 1087–1111.
Fich, E., 2005. Are some outside directors better than others? Evidence from director
appointments by Fortune 1000 firms. Journal of Business 78(5), 1943–1971.
Gompers, P., Ishii, J., Metrick, A., 2003. Corporate governance and equity prices. Quarterly
Journal of Economics 118(1), 107–155.
Guner, B., Malmendier, U., Tate, G., 2008. Financial expertise of directors. Journal of Financial
Economics 88, 323–354.
Hermalin, B., Weisbach, M., 1998. Endogenously chosen boards of directors and their
monitoring of the CEO. American Economic Review 88(1), 96–118.
Hirshleifer, D., Shumway, T., 2003. Good day sunshine: stock returns and the weather. Journal
of Finance 58(3), 1009–1032.
Ivkovic, Z., Weisbenner, S., 2005. Local does as local is: Information content of the geography
of individual investors’ common stock investments. Journal of Finance 55(1), 267–306.
John, K., Kadyrzhanova, D., 2008. Peer effects in corporate governance. Working paper.
John, K., Knyazeva, A., Knyazeva, D., 2010. Does geography matter? Firm location and
corporate payout policy. Journal of Financial Economics, forthcoming.
Kedia, S., Panchapagesan, V., Uysal, V., 2008. Geography and acquirer returns. Journal of
Financial Intermediation 17, 256–275.
Kono, C., Palmer, D., Friedland, R., Zafonte, M., 1998. Lost in space: The geography of
corporate interlocking directorates. American Journal of Sociology 103(4), 863–911.
Landier, A., Nair, V., Wulf, J., 2007. Trade-offs in staying close: Corporate decision making and
geographic dispersion. Review of Financial Studies (forthcoming).
Lerner, J., 1995. Venture capitalists and the oversight of private firms. Journal of Finance 50(1),
301–318.
Linck, J., Netter, J., Yang, T., 2008a. The determinants of board structure. Journal of Financial
Economics 87(2), 308–328.
Linck, J., Netter, J., Yang, T., 2008b. The effects and unintended consequences of the SarbanesOxley Act on the supply and demand for directors. Review of Financial Studies 22(8),
3287–3328.
Linn, S., Park, D., 2005. Outside director compensation policy and the investment opportunity
set. Journal of Corporate Finance 11(4), 680–715.
Loughran, T., Schulz, P., 2005. Liquidity: Urban versus rural firms. Journal of Financial
Economics 78, 341–374.
Loughran, T., Schulz, P., 2006. Asymmetric information, firm location, and equity issuance,
Working paper.
40
Maddala, G., 1983. Limited Dependent and Qualitative Variables in Econometrics. Cambridge
University Press, Cambridge, England.
Malloy, C., 2005. The geography of equity analysis. Journal of Finance 60(2), 719–755.
Masulis, R., Mobbs, S., 2011. Are all inside directors the same? Journal of Finance, 66(3), 823872.
Masulis, R., Wang, C., Xie, F., 2011. Are foreign directors valuable advisors or ineffective
monitors? Working paper.
Mehran, H., 1995. Executive compensation structure, ownership, and firm performance. Journal
of Financial Economics 38(2), 163–184.
Nguyen, B.D., Nielsen, K., 2010. The value of independent directors: evidence from sudden
deaths. Journal of Financial Economics 98, 550–567.
Perry, T., Peyer, U., 2005. Board seat accumulation by executives: A shareholder’s perspective.
Journal of Finance 60(4), 2083-2123.
Petersen, M., 2009. Estimating standard errors in finance panel data sets: Comparing approaches.
Review of Financial Studies 22(1), 435-480.
Pirinsky, C., Wang, Q., 2006. Does corporate headquarters location matter for stock returns?
Journal of Finance 61, 1991-2015.
Raheja, C., 2005. Determinants of board size and composition: A theory of corporate boards.
Journal of Financial and Quantitative Analysis 40(2), 283–306.
Rosenstein, S., Wyatt, J., 1990. Outside directors, board independence, and shareholder wealth.
Journal of Financial Economics 26(2), 175–191.
Vafeas, N., 1999. Board meeting frequency and firm performance. Journal of Financial
Economics 53(1), 113–142.
Wan, H., 2008. Are local directors better monitors? Working paper.
Weisbach, M., 1988. Outside directors and CEO turnover. Journal of Financial Economics 20,
431–460.
Yermack, D., 1996. High market valuation of companies with a small board of directors. Journal
of Financial Economics 40(2), 185–211.
Yermack, D., 2004. Remuneration, retention, and reputation incentives for outside directors.
Journal of Finance 59(5), 2281–2308.
41
Appendix A. Variable definitions.
The main sample covers 1996-2006 and includes Compustat/CRSP firms with available RiskMetrics data on directors and takeover provisions, 13f institutional holdings data, and Execucomp executive
compensation data. Firms with total assets below twenty million, firms in financial industries (SIC codes 6000-6999), firms in regulated utilities industries (SIC codes 4900-4999), and firms
headquartered outside continental US are excluded. Tests using director education and certain types of professional expertise based on BoardEx information about outside directors cover a period
beginning in 2002 and ending in 2008. Firms with missing CRSP/Compustat or Execucomp data, firms with total assets below twenty million, and firms headquartered outside continental US, as well as
boards with fewer than three directors reported, are excluded; financials and regulated utilities are included, unless indicated otherwise.
Variable
Definition
Board composition and other board characteristics
Independent Directors (%)
Percentage of independent directors on the board. Source: RiskMetrics.
Percentage of gray (affiliated) directors on the board. Source: RiskMetrics. Gray directors are identified in proxies and disclosures of related transactions and include
professional service providers; customers and suppliers; former employees of the firm and subsidiaries; designees under an agreement with a group (such as a union),
Gray Directors (%)
directors designated by a significant shareholder, and majority holders; family members of executives; recipients of the firm’s gifts; certain interlocking directors (a
director and an executive of our firm sit on a board of another firm that has an executive and a director on our firm’s board).
Inside Directors (%)
Percentage of inside (employee) directors on the company’s board. Source: RiskMetrics.
Board Size
Log of the number of directors on the company’s board. Source: RiskMetrics.
Independent Directors with
Percent of independent directors with other board seats, as reported by RiskMetrics.
Other Seats (%)
Percentage of local independent directors among independent directors with identifiable corporate positions. Corporate positions are identified as employee status on
Local Independent
another company’s board as reported by RiskMetrics. Independent directors with corporate positions are classified as local if their position is at a company headquartered
Directors (%)
within sixty miles of the firm’s headquarters. Only observations where name and headquarters location of the company of outside employment is known were used.
Local Gray Directors (%)
Defined similarly to Local Independent Directors (%), except gray directors are considered instead.
Local Outside
Defined similarly to Local Independent Directors (%), except both independent and gray directors are considered.
Directors (%)
Independent Dir. Distance
Log of one plus average distance in miles to the main executive job held by an independent director on the firm’s board, based on independent directors holding
to Exec. Job (Mean)
executive jobs.
Outside Dir. Distance
Defined similarly to Independent Dir. Distance to Exec. Job (Mean), except both independent and gray directors are considered.
to Exec. Job (Mean)
Insiders with Local
Percentage of insiders with a (outside) seat on the board of another local firm (firms located within sixty miles of our firm), as reported in RiskMetrics; defined for firms
Directorships (%)
with at least one insider. Source: RiskMetrics.
D(Insiders with Local
Indicator variable equal to one if Insiders with Local Directorships (%) is positive, zero otherwise.
Directorships>0)
Local Directorships per
Average number of seats on the boards of other local firms (firms located within sixty miles of our firm) held by the firm’s insiders, as reported in RiskMetrics; defined
Insider
for firms with at least one insider. Source: RiskMetrics.
Insider Distance to
Log of one plus average distance to directorship for the firm’s insiders with directorships elsewhere, as reported in RiskMetrics; defined for firms with at least one insider
Directorship
holding a seat on the board of another firm. Source: RiskMetrics.
Percentage of outside directors with executive expertise on the board. Executive expertise is defined as having the CEO, CFO, CIO, COO, President, VP, Executive VP,
Executive Expertise (%)
Senior VP, Partner, Managing Director, or Treasurer title, or being an inside board member at another firm. Where specified, independent directors are used. Source:
RiskMetrics or BoardEx (where indicated)
Percentage of directors with academic expertise on the board. Academic expertise is defined as having the title of “Professor”, “Faculty Member”, “Lecturer”,
Academic Expertise (%)
“Instructor”, “Researcher”, “Fellow”, “Dean” or “Provost” in the present or prior years. Data on outside directors reported in BoardEx is used.
Percentage of directors with legal expertise on the board. Legal expertise is defined as holding an attorney, counsel, or a similar law-related professional title at present or
Legal Expertise (%)
in prior years or holding a law degree. Data on outside directors reported in BoardEx is used.
Percentage of directors with financial expertise on the board. Financial expertise is defined as having the title of CFO or Treasurer or a banking, finance, investment or
Financial Expertise (%)
accounting title (such as a comptroller) at present or in prior years. Data on outside directors reported in BoardEx is used.
Percentage of outside directors with corporate experience at R&D firms among outside directors with identifiable corporate positions. Corporate positions are identified
R&D Experience (%)
as employee status on another company’s board as reported by RiskMetrics. R&D firms are firms that had positive research & development expenditures in a given year.
Only observations where name and headquarters location of the company of outside employment is known were used.
Tech. Experience (%)
Percentage of outside directors with corporate experience at high-tech firms among outside directors with identifiable corporate positions. Corporate positions are
42
Similar Growth Experience
(%)
Retired Executives (%)
identified as employee status on another company’s board as reported by RiskMetrics. High-tech firms are identified by SIC codes 2833-2836, 3570-3577, 3600-3674,
7371-7379 or 8731-8734, following Baginski et al. (2004). Only observations where name and headquarters location of the company of outside employment is known
were used.
Percentage of outside directors with corporate experience at firm in the same quartile of growth opportunities as the sample firm among outside directors with
identifiable corporate positions. Corporate positions are identified as employee status on another company’s board as reported by RiskMetrics. Growth opportunities are
measured by the market-to-book ratio. Only observations where name and headquarters location of the company of outside employment is known were used.
Percentage of outside directors who are retired executives (identified as directors over age sixty who have held an executive position in the past but do not hold one any
longer in the current year and in up to two subsequent years), computed based on RiskMetrics (1998-2006) data. Where specified, independent directors are used.
Percent of outside directors with at least one advanced graduate or professional degree (MBA, non-business master’s degree, doctorate, law degree, or medical degree),
as reported in BoardEx. For the categories below, a director with multiple advanced degrees will be counted within each category.
Director Education
(Advanced Degree) (%)
Director Education (MBA)
Percent of outside directors with a master’s degree in business or commerce, as reported in BoardEx.
(%)
Director Education (Misc.
Percent of outside directors with a non-business master’s degree, as reported in BoardEx.
Master’s) (%)
Director Education (Law)
Percent of outside directors with a law degree, as reported in BoardEx.
(%)
Director Education
Percent of outside directors with a doctoral degree (non-medical), as reported in BoardEx.
(Doctorate) (%)
Characteristics of local director labor markets
Local director pool
Log of one plus the number of US nonfinancial firms headquartered within sixty miles of the firm’s headquarters, excluding firms in the same four-digit SIC industry.
Local director pool
Equal to Local director pool if Local director pool is below sample median and 0 otherwise.
(below median)
Local director pool
Equal to Local director pool if Local director pool is above sample median and 0 otherwise.
(above median)
Local director pool
Log of one plus the number of US and Canadian nonfinancial firms with headquarters located within sixty miles of the sample firm’s headquarters, excluding firms in the
(US & Canada)
same four-digit SIC industry as the sample firm.
Local director pool
Log of one plus the number of US and Canadian nonfinancial firms with headquarters located within hundred miles of the sample firm’s headquarters, excluding firms in
(100 mi)
the same four-digit SIC industry as the sample firm.
Local director pool
Log of one plus the number of US nonfinancial firms in the same or higher quartile of total assets with headquarters located within sixty miles of the sample firm’s
(similar or larger size)
headquarters, excluding firms in the same four-digit SIC industry as the sample firm.
Local director pool
Log of one plus the number of US nonfinancial firms with headquarters located within sixty miles of the sample firm’s headquarters, excluding firms in the same 4-digit
(excl. small)
SIC industry as the sample firm and firms with total assets less than hundred million.
Local director pool
Log of one plus the number of US nonfinancial firms with headquarters located within sixty miles of the sample firm’s headquarters.
(all industries)
Local director pool
Log of one plus the number of US nonfinancial and financial firms with headquarters located within sixty miles of the sample firm’s headquarters, excluding firms in the
(incl. financials)
same 4-digit SIC industry as the sample firm
Local director pool
Log of one plus the number of US nonfinancial and financial firms with headquarters located within sixty miles of the sample firm’s headquarters.
(all ind. incl. financials)
Universities
Log of one plus the number of universities (top 130 national universities in the 2008 US News and World Report) within a sixty-mile radius of the firm’s headquarters.
B-schools
Log of one plus the number of business schools (425 US b-schools in the 2008 US News and World Report) within a sixty-mile radius of the firm’s headquarters.
B-schools (top)
Log of one plus the number of business schools (top 65 US b-schools in the 2008 US News and World Report) within a sixty-mile radius of the firm’s headquarters.
Law firms
Log of one plus the number of law firms (250 largest law firms in the 2008 Internet Legal Research Group ranking) within a sixty-mile radius of the firm’s headquarters.
Financial institutions
Log of one plus the number of financial institutions (SIC codes 6000-6999) located within a sixty-mile radius of the firm’s headquarters.
Log of one plus the number of US nonfinancial firms with positive R&D expenditures with headquarters located within sixty miles of the sample firm’s headquarters,
Local director pool (R&D)
excluding firms in the same four-digit SIC industry as the sample firm.
Log of one plus the number of US high-tech firms (SIC codes 2833-2836, 3570-3577, 3600-3674, 7371-7379 or 8731-8734, following Baginski et al. (2004))
Local director pool (tech)
headquartered within sixty miles of the sample firm’s headquarters, excluding firms in the same four-digit SIC industry as the sample firm.
Local director pool (similar Log of one plus the number of US nonfinancial firms in the same quartile of growth opportunities with headquarters located within sixty miles of the sample firm’s
43
growth)
headquarters, excluding firms in the same four-digit SIC industry as the sample firm. Growth opportunities are measured by the market-to-book ratio.
Other geographic characteristics and climate at the firm’s location
Distance in miles between the firm’s headquarters and the closest airport. We focus on large primary commercial service airport hub that accounted for at least one
Distance to the closest
percent of annual revenue passenger boardings (ATL, BOS, BWI, CLT, CVG, DCA, DEN, DFW, DTW, EWR, FLL, HNL, IAD, JFK, LAS, LAX, LGA, MCO, MDW,
airport
MIA, MSP, OAK, ORD, PHL, PHX, PIT, SAN, SEA, SFO, SLC, STL, TPA). Source: FAA.
Number of days of
Average annual number of days with snowfall exceeding one tenth of an inch or one inch, as indicated. Source: National Climactic Data Center (NCDC), 1971-2000
snowfall
climate normals.
Mean snowfall
Average annual snowfall in inches. Source: NCDC 1971-2000 climate normals.
Sunshine
Average percent possible sunshine. Source: NCDC percent sunshine information through 2002.
Annual mean temperatures expressed in Farenheit, constructed by averaging twelve monthly normals. The variable is a thirty-year average (over 1971-2000 period)
Annual temperature means
provided in the NCDC climate normals data for the closest weather station.
Minimum temperatures for the month of January expressed in Farenheit. The variable is a thirty-year average (over 1971-2000 period) provided in the NCDC climate
January temperature lows
normals data for the closest weather station.
Maximum temperatures for the month of July expressed in Farenheit. The variable is a thirty-year average (over 1971-2000 period) provided in the NCDC climate
July temperature highs
normals data for the closest weather station.
Monthly temperature
Standard deviation of monthly temperature means (or monthly temperature lows, where indicated) expressed in Farenheit within a year. Each month’s data point is a
variability
thirty-year average for that month (over 1971-2000) provided in the NCDC climate normals data.
Control variables
Firm size
Sales growth
ROA
Firm age
Firm risk
Number of blockholders
Firm visibility
Firm uncertainty
Institutional ownership
Log of total assets. Source: Compustat.
Annual change in net sales divided by the previous year’s net sales. Source: Compustat.
Ratio of operating income before depreciation to total assets. ROA(%) is ROA expressed as a percent of total assets. Source: Compustat.
Log of one plus the number of years since the first listing of the firm’s shares in CRSP. Source: CRSP monthly.
Standard deviation of daily excess returns expressed in percent in a given year. Source: CRSP daily.
Number of 5% institutional blockholders with stakes in the firm. Source: CDA Spectrum/Thomson Financial.
Linear combination of Firm size (0.60), Firm age (0.65), and Number of blockholders (0.46) based on the principal components analysis of these three variables.
Standard deviation of the firm’s excess return over the value-weighted market index. Source: CRSP monthly.
Percentage stake of all institutional investors in the firm. Source: CDA Spectrum/Thomson Financial.
The index of 24 takeover defense provisions from Gompers et al. (2003). Data is reported for years 1995, 1998, 2000, 2002, 2004, and 2006. Similarly to existing work,
G Index
gap years are filled in with data for adjoining years. Source: RiskMetrics.
CEO age
Equals 1 for CEO aged sixty-five and over; 0 otherwise. Source: Execucomp.
CEO ownership
Percentage ownership stake of the CEO in the firm. Source: Execucomp.
CEO tenure
Log of CEO tenure. Source: Execucomp.
R&D intensity
Ratio of research and development expenditures to assets; zero if missing.
R&D intensity indicator
Indicator variable equal to one if the firm reports positive research and development expenditures and zero otherwise.
High-tech indicator
High-tech firms are identified by SIC codes 2833-2836, 3570-3577, 3600-3674, 7371-7379 or 8731-8734, following Baginski et al. (2004).
Tangible asset intensity
Ratio of property, plants, and equipment to total assets. Source: Compustat.
Robustness control variables and other location related characteristics
Classified board
Indicator variable equal to 1 if the firm has a classified board provision and 0 otherwise. Source: RiskMetrics.
Business segments
Log of the number of business segments. Source: Compustat Segments.
Foreign segment
Indicator variable equal to 1 if the firm has a foreign geographic segment and 0 otherwise. Source: Compustat Segments.
Major exchange
Indicator variable equal to 1 if the firm’s shares are listed on NYSE, Amex, or NASDAQ.
Indicator variable equal to 1 if the firm’s headquarters are located in one of the top 10 metropolitan statistical areas by population size (New York City, Los Angeles,
Big city
Chicago, Washington-Baltimore, San Francisco, Philadelphia, Boston, Detroit, Dallas, and Houston, and their suburbs) and 0 otherwise. Source: Compustat and US
Census (2000).
Medium-sized city
Indicator variable equal to 1 if the firm’s headquarters are located in one of top 11-50 metropolitan statistical areas by population size and 0 otherwise. Source:
44
Compustat and US Census (2000).
Log of number of US nonfinancial firms in the same four-digit SIC industry as the sample firm with headquarters located within sixty miles of the sample firm’s
headquarters.
Population density (60mi)
Log of population density in the counties located within sixty miles of the sample firm’s headquarters. Source: US Census (2000).
College graduates
Percent of college graduates and holders of advanced degrees in the population ages 25 and over, in the counties located within sixty miles of the sample firm’s
(%)(60mi)
headquarters. Source: US Census (2000).
Unemployment (%)
Percent of unemployed in total civilian labor force in the county of the firm’s headquarters. Source: US Census (2000).
ROA (state)
Median of ROA across all firms in the state of the firm’s headquarters location in a given year. Source: Compustat.
Market-to-book (state)
Median of Market-to-book across all firms in the state of the firm’s headquarters location in a given year. Source: Compustat.
Additional variables used in firm value and compensation regressions
Ratio of the firm market value to the book value of total assets. Market value is defined as the book value of total assets minus the book value of equity plus the product
Market-to-book ratio
of year-end price and the number of common shares outstanding. Source: Compustat.
Incentive/Total CEO pay
Percent of value of CEO option grants in total CEO compensation (including value of option grants). Source: Execucomp.
Incentive/
Percent of value of CEO option grants and restricted stock grants in total CEO compensation (including value of option grants). Source: Execucomp.
Total CEO pay (II)
Total CEO pay
Total CEO compensation (including value of option grants), in million, divided by total assets. Source: Execucomp, Compustat
CEO turnover (excl.
Indicator variable equal to 1 if a change in the CEO has occurred compared to the previous year, according to Execucomp. CEO deaths are excluded.
deaths)
Dividend yield
Cash dividends per share divided by price at year-end, times hundred. Source: Compustat.
Past return (lag 1)
Annual average of monthly excess stock return, lagged one year. Source: CRSP monthly.
Past return (lag 2)
Annual average of monthly excess stock return, lagged two years. Source: CRSP monthly.
Dual class firm
Indicator variable equal to 1 if the firm has dual classes of shares and 0 otherwise. Source: RiskMetrics.
Institutional block.
Indicator variable equal to 1 if the firm has a 5% institutional blockholder and 0 otherwise. Source: CDA Spectrum/Thomson Financial.
Independent dir. block.
Indicator variable equal to 1 if the firm has an independent director with a 5% or larger stake and 0 otherwise. Source: RiskMetrics.
Industry cluster
45
Appendix B. Figures and descriptive evidence.
Fig. 1. Local director markets and presence of local independent directors on boards
Mean local independent directors (%) (on the Y-axis) by Local director pool quartile. Local Independent Directors (%) is the percent of
independent directors holding corporate executive positions who are locally employed (within a sixty-mile radius of the firm’s headquarters). For
consistency, only observations where name and headquarters location of the company of outside employment is known were used. Variable
definitions and sample selection criteria are presented in Appendix A.
50
45
40
35
30
25
20
15
10
5
0
Local Independent
Directors (%)
Local
Local
Local
Local
director pool director pool director pool director pool
Q1
Q2
Q3
Q4
46
Table B1. Presence of local directors on boards (by subsample)
Representation of local directors on corporate boards, by subsample. Local Independent/Gray/Outside directors (%) is the percent of independent/gray/outside
(independent plus gray) directors employed within a sixty-mile radius of the firm’s headquarters among independent/gray/outside directors with identified corporate
positions. For consistency, only observations where name and headquarters location of the company of outside employment is known were used. Variable definitions
and sample selection criteria are presented in Appendix A. Subsamples are identified as follows. Small firms (large) are firms with total assets below (above) the
sample median. Young (mature) firms are firms with firm age below (above) the sample median. High (low) growth firms are firms with sales growth above (below)
the sample median. Competitive (concentrated) industries are three-digit SIC industries with sales-based Herfindahl index below (above) the sample median. Twosided t-tests of differences in means are performed; statistical significance at 1%, 5%, and 10% levels is denoted with ***, **, and *, respectively.
Firm visibility
Firm size:
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Firm age:
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Number of blockholders:
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Firm visibility:
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
NYSE listing:
Local Independent Directors
Local Gray Directors
Local Outside Directors (Indep. & Gray)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Firm benefits from local knowledge
Growth opportunities:
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Industry concentration:
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Governance reforms
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Indep. Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Small
(bottom 75%)
39.44
57.74
42.77
4.70
4.52
Young
(bottom 75%)
37.77
51.54
40.67
4.80
4.65
Few
(bottom 75%)
34.04
51.19
37.02
5.02
4.88
Low
(bottom 75%)
39.06
55.19
42.25
4.71
4.54
No
42.10
56.73
44.88
4.70
4.52
Large
(top 25%)
26.32
38.64
27.78
5.48
5.45
Old
(top 25%)
28.75
45.93
30.68
5.33
5.27
Many
(bottom 75%)
33.97
41.67
35.57
5.04
4.95
High
(bottom 75%)
26.86
40.52
28.39
5.47
5.43
Yes
32.01
47.88
34.65
5.10
4.99
Strong
(top 75%)
35.64
51.01
38.27
4.96
4.81
Competitive
(HI - bottom
75%)
34.66
49.05
37.46
5.00
4.88
Before
34.91
49.469
37.629
4.968
4.8482
Weak
(bottom 25%)
32.70
48.55
35.50
5.07
4.96
Concentrated
(HI - top
25%)
32.53
50.97
35.09
5.07
4.93
After
30.80
57.89
35.94
5.14
4.90
Diff.
Sig.
13.12
19.10
14.99
-0.78
-0.94
***
***
***
***
***
Diff.
Sig.
9.02
5.61
9.99
-0.54
-0.62
***
Diff.
Sig.
0.07
9.53
1.45
-0.02
-0.07
***
***
***
**
Diff.
Sig.
12.20
14.67
13.86
-0.76
-0.89
Diff.
10.09
8.85
10.23
-0.40
-0.46
***
Diff.
Sig.
2.94
2.45
2.78
-0.12
-0.15
*
Diff.
Sig.
2.13
-1.92
2.36
-0.07
-0.05
Diff.
4.11
-8.43
1.69
-0.17
-0.05
***
***
***
***
Small
(bottom 50%)
42.09
65.13
46.59
4.51
4.28
Young
(bottom 50%)
37.55
45.61
39.40
4.90
4.78
Few
(bottom 50%)
33.61
54.06
36.96
5.06
4.91
Low
(bottom 50%)
40.24
56.36
43.85
4.64
4.46
Large
(top 50%)
30.2
41.9
31.9
5.26
5.2
Old
(top 50%)
32.15
52.03
35.27
5.09
4.96
Many
(top 50%)
34.6
43
36.5
4.96
4.87
High
(top 50%)
30.9
45.7
32.9
5.21
5.12
Strong
(top 50%)
35.64
51.01
38.27
4.96
4.81
Weak
(bottom 50%)
32.7
48.6
35.5
5.07
4.96
Competitive
(HI – bottom 50%)
Concentrated
(HI – top 50%)
Diff.
Sig.
39.11
49.56
40.95
4.88
4.80
29.9
49.7
33.4
5.14
4.97
9.20
-0.12
7.51
-0.26
-0.17
***
Diff
11.85
23.21
14.71
-0.75
-0.91
Sig
***
***
***
***
***
Diff.
Sig.
5.40
-6.42
4.13
-0.19
-0.17
***
Diff.
Sig.
-1.04
11.09
0.50
0.10
0.04
***
**
**
***
Diff.
Sig.
9.39
10.68
10.91
-0.57
-0.66
***
Diff.
Sig.
***
***
***
***
Sig.
***
*
***
*
***
*
**
2.94
2.45
2.78
-0.12
-0.15
*
***
**
Sig.
**
47
Table B2. Correlation of measures of firm visibility and benefits from local knowledge
Firm size
Firm age
Number of
blockholders:
1.00
0.38
1.00
Number of blockholders
-0.24
-0.14
Visibility
0.72
0.71
NYSE listing
0.40
0.33
Growth opportunities
0.01
-0.12
Industry concentration
0.12
0.20
Correlations significant at 5% are italicized.
Visibility
NYSE listing
1.00
0.40
-0.07
0.19
1.00
-0.08
0.20
Growth
opportunities
Industry
concentration
1.00
-0.07
1.00
Firm size
Firm age
1.00
0.29
-0.07
-0.03
0.01
Fig. 2. Economic significance of local director markets relative to other determinants of Independent
Directors (%)
Economic effect of a one standard deviation increase in the X variable on Independent Directors (%)
2.5
2.0
1.5
1.0
0.5
-2.5
CEO tenure
CEO age
CEO ownership
G Index
Firm risk
Firm age
ROA
Tangible asset
intensity
R&D intensity
indicator
-2.0
Institutional ownership
-1.5
Sales growth
-1.0
Firm size
-0.5
Local director pool
0.0
Economic effects are computed based on coefficient estimates from Table 2, column III and standard deviations of right-hand-side variables
reported in Table 1.
For each determinant of Board independence, the Y-axis displays the expected change in independent director representation on the board,
Independent Directors (%), in response to a one standard deviation increase in the determinant, holding other determinants constant.
48
Table 1. Summary statistics and variable definitions
Variable definitions and sample selection criteria are presented in Appendix A.
Panel A: Full sample
Variable
Obs
Mean
Med
SD
Board Characteristics:
Independent Directors (%)
Inside Directors (%)
Gray Directors (%)
Board Size
Board Size [num]
Independent Directors with Other Seats (%)
Local Independent Directors (%)
Local Gray Directors (%)
Local Outside Directors (%)
Independent Dir. Distance to Exec. Job (Mean)
Independent Dir. Distance to Exec. Job [miles] (Mean)
Outside Dir. Distance to Exec. Job (Mean)
Outside Dir. Distance to Exec. Job [miles] (Mean)
Insiders with Local Directorships (%)
D(Insiders with Local Directorships>0)
Local Directorships per Insider
Executive Expertise (%)
Academic Expertise (%)
Legal Expertise (%)
Financial Expertise (%)
R&D Experience (%)
Tech. Experience (%)
Similar Growth Experience (%)
Retired Executives (%)
Director Education (Advanced Degree) (%)
Director Education (MBA) (%)
Director Education (Misc. Master’s) (%)
Director Education (Law) (%)
Director Education (Doctorate) (%)
Local Director Pool Characteristics:
9693
9693
9693
9693
9693
8181
3365
785
3842
3365
3365
3842
3842
9628
9628
9628
8217
9694
9694
9694
3813
3813
3811
8217
9694
9694
9694
9694
9694
65.14
20.88
13.98
2.17
9.11
50.98
34.03
49.64
36.76
5.02
534
4.89
511
13.54
0.20
0.17
33.35
13.16
12.75
16.56
50.48
16.83
28.90
12.79
59.92
31.41
18.42
12.24
11.05
66.67
16.67
11.11
2.20
9.00
50.00
0.00
50.00
0.00
5.74
311
5.71
302
0.00
0.00
0.00
33.33
11.11
11.11
14.29
50.00
0.00
0.00
10.00
62.50
30.00
16.67
11.11
10.00
17.49
11.39
13.83
0.27
2.48
28.10
43.34
48.85
43.92
2.26
616
2.34
604
30.20
0.40
0.40
23.42
13.59
12.87
13.84
44.96
34.32
40.87
15.63
22.09
19.37
16.28
12.89
12.74
Local director pool
Local director pool [num]
Local director pool (100 mi)
Local director pool (100 mi) [num]
Local director pool (US & Canada)
Local director pool (US & Canada) [num]
Local director pool (similar or larger size)
Local director pool (similar or larger size) [num]
Local director pool (excl. small)
Local director pool (excl. small) [num]
Local director pool (all industries)
Local director pool (all industries) [num]
Local director pool (incl. financials)
Local director pool (incl. financials) [num]
Local director pool (all ind. incl. financials)
Local director pool (all ind. incl. financials) [num]
Local director pool (R&D)
Local director pool (tech)
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
3.75
105
3.88
116
3.76
105
2.65
32
3.32
61
3.87
111
4.25
138
4.32
145
3.92
3.23
4.04
56
4.19
65
4.04
56
2.83
16
3.58
35
4.08
58
4.49
88
4.51
90
4.09
3.47
1.65
116
1.63
128
1.64
116
1.46
44
1.53
65
1.52
125
1.37
143
1.31
151
1.48
1.60
49
Local director pool (similar growth)
Universities
Universities [num]
B-schools
B-schools [num]
B-schools (top)
B-schools (top) [num]
Law firms
Law firms [num]
Financial institutions
Financial institutions [num]
Other Location Related Characteristics:
9693
9694
9694
9694
9694
9694
9694
9694
9694
3.78
115.96
3.42
1.23
3.50
1.99
8.29
0.82
1.66
1.93
11.14
3.58
35.00
3.66
1.10
2.00
1.95
6.00
0.69
1.00
1.95
6.00
1.54
153.66
1.21
0.76
3.25
0.72
6.13
0.58
1.56
1.12
13.30
3.78
115.96
Large city
Medium-sized city
Distance to closest airport [mi]
Industry cluster
Industry cluster [num]
Population density (60mi)
Population density (60mi) [mln]
College graduates (%) (60mi)
Unemployment (%)
Firm Characteristics:
Firm size
Firm size [mln]
Sales growth
ROA
Market-to-book ratio
Dividend yield
G Index
Institutional ownership
Major exchange listing
NYSE listing
Firm age
Firm age [years]
Number of blockholders
Firm visibility
Firm risk
Tangible asset intensity
R&D intensity
R&D intensity indicator
High-tech indicator
Business segments
Business segments [num]
Foreign segment
Dual class firm
Institutional block.
Independent dir. block.
9693
9693
9693
9693
9693
9693
9693
9693
9693
0.52
0.34
52.84
1.36
7
15.28
6.44
28.33
5.37
1.00
0.00
19.79
1.10
2
15.35
4.64
28.78
5.02
0.50
0.47
68.56
0.92
14
0.99
5.51
4.91
1.81
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
9693
8650
8650
9693
9693
9693
8201
7.32
5571
0.12
0.14
2.12
0.96
9.22
67.83
0.96
0.67
2.93
24.12
2.11
1.25
2.57
0.30
3.34
0.53
0.21
0.71
2.56
0.74
0.10
0.83
0.04
7.18
1312
0.08
0.14
1.66
0.33
9.00
69.72
1.00
1.00
2.94
18.00
2.00
1.20
2.26
0.24
0.43
1.00
0.00
0.69
2.00
1.00
0.00
1.00
0.00
1.41
18219
0.28
0.11
1.44
1.35
2.60
17.93
0.20
0.47
0.79
19.33
1.32
0.84
1.28
0.21
5.78
0.50
0.41
0.68
1.72
0.44
0.29
0.38
0.19
CEO Characteristics:
CEO ownership
CEO age
CEO tenure
CEO tenure [years]
Incentive/total CEO pay
Incentive/total CEO pay (II)
Total CEO pay
CEO turnover (excl. deaths)
9693
9693
9693
9693
9640
9640
9499
9481
2.34
0.09
1.74
7.33
36.18
43.99
3.27
0.12
0.31
0.00
1.79
5.00
35.25
47.20
1.80
0.00
5.62
0.29
0.90
7.61
28.96
29.09
4.66
0.32
50
Panel B: Statistics of the main variables by subsamples (firm size and age)
Firms with total assets below the sample median (small) and above the sample median (large); firms with firm age below the sample median (young) and above the sample median (mature).
Firm size
Independent Directors (%)
Inside Directors (%)
Gray Directors (%)
Executive Expertise (%)
Local Independent Directors (%)
Local Outside Directors (%)
Independent Dir. Distance to Exec. Job (Mean)
Insiders with Local Directorships (%)
Board Size
Local director pool
Local director pool (all ind. incl. financials)
Obs
4857
4857
4857
4090
1076
1275
1076
4835
4857
4857
4857
Small
Mean
Med
62.83 66.67
22.94 20.00
14.24 12.50
38.04 33.33
42.09
0.00
46.59 50.00
4.51
5.12
8.99
0.00
2.05
2.08
3.73
4.04
4.27
4.47
SD
17.42
12.00
14.03
30.23
46.86
47.06
2.44
25.60
0.24
1.66
1.33
Obs
4836
4836
4836
4087
2289
2567
2289
4793
4836
4836
4836
Firm age
Large
Mean
Med
67.46 70.00
18.80 16.67
13.72 11.11
39.64 37.50
30.24
0.00
31.87
0.00
5.26
6.00
18.13
0.00
2.29
2.30
3.78
4.03
4.36
4.55
SD
17.25
10.33
13.62
27.78
41.05
41.42
2.13
33.61
0.24
1.63
1.30
Obs
5008
5008
5008
4214
1171
1383
1171
4983
5008
5008
5008
Young
Mean
Med
62.53 66.67
22.97 20.00
14.50 12.50
37.10 33.33
37.55
0.00
39.40
0.00
4.90
5.74
9.44
0.00
2.07
2.08
3.88
4.29
4.39
4.67
SD
18.02
11.97
14.73
30.28
45.69
46.03
2.51
25.64
0.26
1.66
1.35
Obs
4685
4685
4685
3963
2194
2459
2194
4645
4685
4685
4685
Mature
Mean
Med
67.92 70.00
18.63 15.79
13.43 11.11
40.69 40.00
32.15
0.00
35.27
0.00
5.09
5.74
17.94
0.00
2.28
2.30
3.62
3.78
4.23
4.41
SD
16.45
10.26
12.78
27.56
41.92
42.62
2.11
33.89
0.24
1.62
1.27
Panel C: Subsamples (firm growth opportunities and industry competition)
Firms with sales growth above sample median (strong) and below sample median (weak); firms in three-digit SIC industries with sales Herfindahl index below sample median (competitive) and above sample median (concentrated).
Independent Directors (%)
Inside Directors (%)
Gray Directors (%)
Executive Expertise (%)
Local Independent Directors (%)
Local Outside Directors (%)
Independent Dir. Distance to Exec. Job (Mean)
Insiders with Local Directorships (%)
Board Size
Local director pool
Local director pool (all ind. incl. financials)
Obs
4845
4845
4845
4082
1522
1744
1522
4820
4845
4845
4845
Growth opportunities
Strong
Weak
Mean
Med
SD
Obs
Mean
Med
63.54 66.67 17.94 4848 66.73 69.23
22.15 20.00 11.95 4848 19.61 16.67
14.32 12.50 14.12 4848 13.65 11.11
38.65 33.33 29.74 4095 39.03 33.33
35.64
0.00 44.30 1843 32.70
0.00
38.27
0.00 44.91 2098 35.50
0.00
4.96
5.74
2.37 1843
5.07
5.75
12.14
0.00 28.55 4808 14.94
0.00
2.15
2.20
0.27 4848
2.19
2.20
3.82
4.22
1.66 4848
3.69
3.91
4.36
4.60
1.33 4848
4.28
4.44
Industry competition
SD
16.88
10.64
13.53
28.33
42.50
43.04
2.16
31.72
0.27
1.63
1.30
Obs
4882
4882
4882
4124
1508
1696
1508
4847
4882
4882
4882
Competitive
Mean
Med
64.64 66.67
21.37 18.18
14.00 12.50
38.44 33.33
39.11
0.00
40.95
0.00
4.88
5.68
12.92
0.00
2.12
2.08
4.17
4.64
4.65
5.00
SD
17.32
11.24
14.04
29.70
44.85
45.14
2.40
29.15
0.28
1.55
1.27
Obs
4811
4811
4811
4053
1857
2146
1857
4781
4811
4811
4811
Concentrated
Mean
Med
65.65 66.67
20.37 16.67
13.97 11.11
39.25 33.33
29.91
0.00
33.44
0.00
5.14
5.83
14.17
0.00
2.22
2.20
3.33
3.47
3.98
4.09
SD
17.65
11.51
13.62
28.36
41.63
42.64
2.13
31.23
0.26
1.63
1.27
51
Table 2. Local director labor markets and board composition.
The dependent variable is independent director representation on the board, Independent Directors (%). Variable definitions and sample selection
criteria are presented in Appendix A. Ordinary least squares regressions. Three-digit SIC industry effects and year effects are included. Robust tstatistics adjusted for clustering by firm are italicized.
I
Local director pool
0.740
2.84
Firm size
II
***
0.600
III
**
0.646
***
0.840
***
-2.707
2.47
1.189
2.78
3.99
-3.189
-3.49
-3.12
ROA
-4.426
-3.938
-1.25
2.085
Firm risk
1.919
3.88
0.409
0.160
1.31
0.190
G Index
0.833
***
0.54
***
0.152
***
0.710
8.49
5.77
***
-1.15
***
4.10
Institutional ownership
***
2.90
Sales growth
Firm age
***
***
6.96
***
5.06
Tangible asset intensity
4.378
R&D intensity indicator
4.412
1.60
***
3.99
CEO ownership
-0.354
CEO age
-4.576
CEO tenure
-0.264
***
-5.59
***
-5.27
-0.85
Obs.
9693
9693
R2
0.24
0.30
9693
0.33
Adj. R2
0.22
0.29
0.31
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
52
Table 3. Local director labor markets and board independence: Subsample analysis.
The dependent variable is independent director representation on the board, Independent Directors (%). Variable definitions and sample selection
criteria are presented in Appendix A. Ordinary least squares regressions by subsample. Control variables from Table 2, column III are included but
not shown for brevity. Subsamples are identified based on the following characteristics: Firm size (total assets); Firm age (number of years since
initial listing in CRSP); Number of blockholders (number of 5% institutional blockholders with stakes in the firm identified in CDA Spectrum); Firm
visibility (a linear combination of Firm size (0.63), Firm age (0.66), and Number of blockholders (0.40) based on the principal components analysis of
these three variables); NYSE listing (indicator for stock listing on NYSE); Total corporate assets in the state (total assets of public firms in the firm
headquarters state); Distance to the closest airport; Number of days of snowfall; Mean snowfall in inches; Percent possible sunshine; Annual
temperature means; January temperature lows and July temperature highs; Monthly temperature variability is based on monthly means or alternatively
lows within a year; Growth opportunities (sales growth); Industry concentration (Herfindahl index based on sales in the three-digit SIC industry). For
continuous variables subsamples based on sample quartiles and/or sample medians are used, except as indicated in the table for Distance to the closest
airport, Average annual temperatures, and Percent possible sunshine. Observations before governance reforms include firm-years 1999-2002.
Observations after governance reforms include firm-years 2003-2006, for firms that failed to meet one of the following governance requirements by
2001 – majority of independent directors on the board, independent directors on nominating/governance and compensation committees and three
independent directors on the audit committee. Three-digit SIC industry effects and year effects are included. Robust t-statistics adjusted for clustering
by firm are italicized.
Panel A: Firm visibility
(1) Firm visibility
Firm size:
Local director pool
Firm age:
Local director pool
Number of blockholders:
Local director pool
Firm visibility:
Local director pool
Small
(bottom 75%)
0.590 **
2.36
Large
(top 25%)
0.374
0.70
Small
(bottom 50%)
0.686 **
2.30
Large
(top 50%)
0.431
1.24
Young
(bottom 75%)
0.586 **
2.22
Old
(top 25%)
0.280
0.62
Young
(bottom 50%)
0.605 **
2.08
Old
(top 50%)
0.673
1.85
Few
(bottom 75%)
0.751 ***
3.03
Many
(top 75%)
0.141
0.34
Few
(bottom 50%)
0.856 ***
2.92
Many
(top 50%)
0.343
1.21
Low
(bottom 75%)
0.640 **
2.56
High
(top 75%)
0.252
0.54
Low
(bottom 50%)
0.612 **
2.09
High
(top 50%)
0.311
0.94
*
NYSE listing:
No
0.750
1.98
Local director pool
Profile of the firm’s region
(total corporate assets in the state)
Local director pool
**
Below national
median
1.311 ***
3.58
Yes
0.461
1.53
Above national
median
0.171
0.52
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
53
Panel B: Accessibility of the firm’s location and attractiveness of climate in the firm’s area
(1) Ease of access by air
Distance to the closest airport:
Local director pool
(2) Typical weather at the firm’s location
Number of days of snowfall:
Local director pool
Mean snowfall:
Local director pool
Sunshine:
Local director pool
Temperature extremes:
Local director pool
Temperature variability
within a year:
Local director pool
Over 30 miles
away
1.262 ***
3.65
Within a 30-mile
radius
0.372
0.71
Over 60 miles
away
1.966 ***
3.12
Within a 60-mile
radius
0.486
1.21
Unfavorable
Good
Unfavorable
Good
More snow days
(days with snowfall
>0.1” in top 50%)
0.868 ***
2.90
Few snow days
(days with snowfall
>0.1” in bottom 50%)
0.425
0.98
More snow days
(days with snowfall
>1” in top 50%)
0.939 ***
3.12
Few snow days
(days with snowfall
>1” in bottom 50%)
0.359
0.85
More snowfall
(mean snowfall (”)
in top 50%)
1.072 ***
3.42
Limited snowfall
(mean snowfall (”)
in bottom 50%)
0.571
1.41
Less sunshine
(pct. possible sunshine in bottom 50%)
0.896 ***
2.80
More sunshine
(pct. possible sunshine in top 50%)
0.595
1.53
Less sunshine
(pct. possible sunshine is below 60)
0.793 ***
2.78
More sunshine
(pct. possible sunshine is above 60)
0.396
0.80
Temp. extremes
(avg. temp. outside
25%-75% range)
1.154 ***
3.42
Moderate temp.
(avg. temp. in
25%-75% range)
-0.017
-0.05
Temp. extremes
(Jan. lows <25%tile
or July highs >75%tile)
0.984 ***
2.94
Moderate temp.
(Jan. lows >25%tile &
July highs < 75%tile)
0.243
0.63
Within-year
variation in temp.
means is high
(in top 50%)
1.051 ***
3.31
Within-year
variation in temp.
means is low
(in bottom 50%)
0.159
0.42
Within-year
variation in temp.
lows is high
(in top 50%)
1.107 ***
3.49
Within-year
variation in temp.
lows is low
(in bottom 50%)
0.288
0.77
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
Panel C: Firm benefits from local knowledge
(1) Growth opportunities:
Local director pool
(2) Industry concentration:
Local director pool
Strong
(top 75%)
0.861 ***
3.10
Competitive
(HI - bottom 75%)
0.611 **
2.35
Weak
(bottom 25%)
0.417 *
1.67
Concentrated
(HI - top 25%)
0.580
1.36
Strong
(top 50%)
0.861 ***
3.10
Competitive
(HI – bottom 50%)
0.636 **
1.96
Weak
(bottom 50%)
0.417 *
1.67
Concentrated
(HI – top 50%)
0.551 *
1.79
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
Panel D: Governance reforms
(1) Governance reforms
Local director pool
Before reforms
All firms
0.808 ***
2.73
After reforms
All firms
0.673 *
1.87
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
54
Table 4. Local director labor markets and board composition: alternative labor market definitions.
The dependent variable is independent director representation on the board, Independent Directors (%). Variable definitions and sample selection
criteria are presented in Appendix A. Ordinary least squares regressions. Three-digit SIC industry effects and year effects are included. Robust tstatistics adjusted for clustering by firm are italicized.
Panel A: Alternative definitions of Local director pool. Column I includes firms in the hundred-mile radius of the firm’s headquarters; column II
includes US and Canadian firms in the hundred-mile radius; column III includes firms of similar or larger size; column IV excludes small firms;
column V includes nonfinancial firms from all industries, including the firm’s industry; column VI includes financial as well as nonfinancial firms;
column VII includes financial firms as well as nonfinancial firms from all industries, including the firm’s industry; column VIII allows the main effect
to vary depending on whether Local director pool is below or above sample median.
Local director pool
(100mi)
Local director pool
(US & Canada)
Local director pool
(similar or larger)
Local director pool
(excl. small)
Local director pool
(all industries)
Local director pool
(incl. financials)
Local director pool
(all ind. incl. financials)
Local director pool
(below median)
Local director pool
(above median)
Firm size
I
0.580
2.44
II
III
IV
V
VI
VII
VIII
**
0.648
2.79
***
0.740
2.85
***
0.654
2.67
***
0.839 ***
1.024 ***
0.841 ***
0.859 ***
2.97
2.90
3.53
2.91
Sales growth
-2.702 *** -2.705 *** -2.686 *** -2.701 ***
-3.11
-3.12
-3.10
-3.11
ROA
-3.963
-3.939
-3.898
-3.935
-1.15
-1.15
-1.14
-1.15
Firm age
1.909 ***
1.919 ***
1.903 ***
1.921 ***
3.85
3.88
3.85
3.88
Firm risk
0.169
0.160
0.153
0.170
0.56
0.53
0.51
0.57
0.152 ***
0.153 ***
0.152 ***
Institutional ownership
0.152 ***
6.97
6.96
7.02
6.97
G Index
0.700 ***
0.710 ***
0.710 ***
0.706 ***
5.01
5.06
5.05
5.03
Tangible asset intensity
4.237
4.381
4.389
4.344
1.55
1.60
1.60
1.58
4.414 ***
4.371 ***
4.394 ***
R&D intensity indicator
4.392 ***
3.97
3.99
3.96
3.98
CEO ownership
-0.354 *** -0.354 *** -0.354 *** -0.355 ***
-5.58
-5.59
-5.58
-5.59
CEO age
-4.571 *** -4.574 *** -4.569 *** -4.557 ***
-5.25
-5.26
-5.26
-5.24
CEO tenure
-0.264
-0.264
-0.263
-0.267
-0.85
-0.85
-0.85
-0.86
Obs.
9693
9693
9693
9693
R2
0.33
0.33
0.33
0.33
0.31
0.31
0.31
0.31
Adj. R2
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
0.692
2.73
***
0.561
2.03
**
0.581
2.01
0.838
2.90
-2.706
-3.12
-3.942
-1.15
1.923
3.89
0.159
0.53
0.152
6.95
0.711
5.06
4.385
1.60
4.390
3.97
-0.354
-5.59
-4.575
-5.26
-0.265
-0.86
9693
0.33
0.31
***
***
***
***
***
***
***
***
0.862
2.99
-2.707
-3.11
-3.848
-1.11
1.905
3.84
0.190
0.63
0.152
6.95
0.702
5.01
4.097
1.49
4.393
3.96
-0.357
-5.63
-4.553
-5.23
-0.264
-0.85
9693
0.33
0.31
***
***
***
***
***
***
***
***
0.861
2.98
-2.707
-3.11
-3.852
-1.12
1.908
3.85
0.190
0.63
0.152
6.95
0.703
5.01
4.109
1.50
4.380
3.95
-0.357
-5.63
-4.553
-5.23
-0.265
-0.85
9693
0.33
0.31
**
***
***
***
***
***
***
***
***
0.910
2.50
0.703
2.96
0.841
2.91
-2.691
-3.10
-3.986
-1.16
1.917
3.87
0.170
0.57
0.152
6.98
0.707
5.03
4.386
1.60
4.432
4.00
-0.355
-5.61
-4.578
-5.28
-0.258
-0.84
9693
0.33
0.31
**
***
***
***
***
***
***
***
***
***
55
Table 5. Local director labor markets and board composition: Robustness checks.
The dependent variable is independent director representation on the board, Independent Directors (%). Variable definitions and sample selection
criteria are presented in Appendix A. Ordinary least squares regressions. Three-digit SIC industry effects and year effects are included. Robust tstatistics adjusted for clustering by firm are italicized. Panel A: Additional dimensions of board composition are considered.
Panel B: Variable definitions and sample selection criteria are modified for robustness as follows: logistic transformation of board independence
(expressed as a proportion), ln[y/(1-y)], is used in column I; firms headquartered in ten largest metropolitan areas are excluded in column II; CA, IL,
MA, and NY are excluded in column III. Local director pool is lagged one period in column IV. Additional controls are included in columns V-VIII.
Firm size is defined as log of market value of the firm in column VIII.
Panel C: Additional controls and alternative explanations. Industry cluster is included separately in column I and jointly with the main effect in
column II; Population density is included separately in column III and jointly with the main effect in column IV; College graduates (%) (60mi) is
included in column V. Census regions dummies are included in columns III-V; ROA (state), Market-to-book (state), and Unemployment (%) are
included in columns VI-VII. Mean of the dependent variable in the state of location, Mean (state), is added in column VIII.
Panel A. Local director labor markets and other board characteristics
Gray
Directors (%)
I
Inside
Directors (%)
II
**
Local director pool
-0.677
-2.46
-1.82
Firm size
-0.030
-0.889
-0.12
-4.73
Sales growth
-0.444
Board Size
III
*
-0.002
0.079
1.906
1.36
4.24
ROA
-0.172
4.675
Firm age
-1.118
-2.64
-2.77
7.14
Firm risk
-0.060
-0.096
-0.026
Institutional ownership
-0.107
-5.63
-3.39
G Index
-0.142
-0.539
-1.18
-6.04
Tangible asset intensity
-1.124
-3.364
R&D intensity indicator
-3.535
-0.23
-0.814
-0.51
***
-0.049
-0.76
-0.012
-11.901
***
0.048
-0.27
***
-0.052
***
0.776
-0.63
-6.36
-0.002
***
0.010
5.12
3.18
*
0.059
2.445
-7.21
-1.54
0.50
6.54
0.773
3.701
1.01
4.99
0.617
0.91
***
-0.004
***
0.029
-0.100
**
-4.918
-2.83
-0.007
-2.259
-1.42
2.81
-1.54
1.439
-0.395
0.040
Large city
2.215
1.83
0.10
Medium-sized city
1.277
-1.483
Obs.
R2
Adj. R2
1.25
9693
0.18
0.16
-1.66
9693
0.32
0.31
***
***
-4.09
*
-1.044
-0.34
1.71
-0.39
0.111
-0.010
-1.522
*
***
-0.85
2.05
Major exchange
***
3.63
***
-4.32
***
**
-2.03
***
0.50
CEO age
*
-0.965
6.676
-1.12
1.00
***
1.45
0.334
-0.365
-0.493
0.015
-3.78
**
-2.06
***
-1.75
0.027
CEO tenure
-1.044
-0.889
CEO ownership
***
15.75
***
**
-0.49
***
7.189
-2.69
2.33
***
-0.024
***
2.77
***
18.91
***
0.891
-0.06
1.584
-0.38
***
Independent
Directors with
Other Seats (%)
IV
-0.51
-0.56
0.012
2.732
0.73
9693
0.49
0.48
1.21
8181
0.30
0.28
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
56
Panel B: Alternative sample selection criteria and additional control variables
Local director pool
Firm size
I
0.029
2.59
0.047
3.31
***
***
II
1.499
3.87
0.478
1.14
***
III
1.138
3.80
0.833
2.39
***
**
IV
0.655
2.82
0.839
2.90
***
***
V
0.620
2.60
0.625
2.02
***
**
VI
0.568
2.46
0.953
3.30
**
***
VII
1.125
3.44
0.915
3.18
***
***
Firm size2
Sales growth
ROA
Firm age
Firm risk
Institutional ownership
G Index
Tangible asset intensity
R&D intensity indicator
CEO ownership
CEO age
CEO tenure
-0.130
-3.12
-0.193
-1.18
0.095
3.82
0.010
0.66
0.007
6.83
0.035
5.07
0.203
1.52
0.214
4.00
-0.017
-5.70
-0.222
-5.36
-0.010
-0.70
***
***
***
***
***
***
***
-2.975
-2.01
-3.869
-0.68
1.797
2.43
-0.339
-0.75
0.180
6.25
0.968
4.63
2.318
0.61
4.463
2.80
-0.320
-3.19
-3.875
-3.21
-0.275
-0.64
**
**
***
***
***
***
***
-4.167
-3.37
-5.237
-1.17
1.749
3.00
0.142
0.38
0.176
6.81
0.997
5.93
2.895
0.92
4.622
3.57
-0.278
-3.60
-3.245
-3.09
-0.308
-0.82
***
***
***
***
***
***
***
-2.703
-3.11
-3.918
-1.14
1.920
3.88
0.162
0.54
0.152
6.95
0.712
5.07
4.380
1.60
4.408
3.99
-0.355
-5.60
-4.570
-5.26
-0.263
-0.85
***
***
***
***
***
***
***
-2.190
-2.60
-3.417
-0.98
1.891
3.63
0.251
0.82
0.155
6.87
0.761
5.16
5.422
1.91
4.096
3.56
-0.329
-5.31
-4.763
-5.40
-0.184
-0.57
Classified board
Business segments
Foreign segment
***
***
***
-2.929
-3.31
-4.124
-1.19
2.522
5.12
0.119
0.40
0.159
7.29
***
***
***
*
***
***
***
4.193
1.54
4.491
4.06
-0.383
-6.01
-4.525
-5.21
-0.329
-1.06
1.515
2.16
***
***
***
-2.787
-3.22
-4.408
-1.26
1.937
3.91
0.148
0.49
0.157
7.23
0.678
4.80
4.350
1.58
4.401
4.02
-0.347
-5.49
-4.513
-5.22
-0.246
-0.79
***
***
***
***
***
***
***
***
***
***
***
***
***
***
*
***
***
***
**
0.577
1.04
0.993
1.28
Major exchange
Big city
Medium-sized city
Obs.
4672
6437
9693
R2
0.43
0.39
0.33
Adj. R2
0.41
0.37
0.31
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
***
VIII
1.022
3.15
-7.454
-3.77
0.513
4.21
-2.761
-3.35
-5.438
-1.53
1.699
3.40
0.071
0.24
0.186
8.49
0.745
5.35
4.855
1.79
4.161
3.83
-0.332
-5.31
-4.337
-5.07
-0.232
-0.76
8650
0.34
0.32
9693
0.32
0.31
-1.064
-0.58
-2.404
-1.62
0.171
0.14
9693
0.33
0.31
-0.762
-0.42
-2.370
-1.61
0.255
0.21
9693
0.34
0.32
57
Panel C: Alternative explanations and additional control variables
I
Local director pool
Firm size
Sales growth
ROA
Firm age
Firm risk
Institutional ownership
G Index
Tangible asset intensity
R&D intensity indicator
CEO ownership
CEO age
CEO tenure
G Index
Industry cluster
Population density
(60mi)
College graduates (%)
(60mi)
ROA (state)
0.891
3.06
-2.742
-3.13
-4.117
-1.19
1.976
4.01
0.196
0.66
0.153
7.00
0.696
4.98
3.846
1.40
4.251
3.85
-0.354
-5.59
-4.529
-5.19
-0.274
-0.88
0.696
4.98
0.756
1.56
***
***
***
***
***
***
***
***
***
II
0.626
2.30
0.837
2.88
-2.706
-3.12
-3.941
-1.15
1.923
3.88
0.157
0.52
0.152
6.95
0.712
5.06
4.379
1.60
4.394
3.98
-0.354
-5.59
-4.576
-5.26
-0.265
-0.85
0.712
5.06
0.079
0.14
III
**
***
***
***
***
***
***
***
***
***
0.950
3.32
-2.756
-3.15
-4.397
-1.26
1.933
3.88
0.211
0.72
0.155
7.07
0.696
4.95
3.622
1.32
4.233
3.87
-0.347
-5.56
-4.498
-5.19
-0.273
-0.88
0.696
4.95
0.399
0.96
***
***
***
***
***
***
***
***
***
IV
0.856
2.53
0.927
3.23
-2.777
-3.19
-4.502
-1.29
1.954
3.92
0.162
0.55
0.155
7.11
0.699
4.97
4.112
1.51
4.199
3.88
-0.340
-5.39
-4.505
-5.21
-0.278
-0.90
0.699
4.97
**
***
***
***
***
***
***
***
***
***
V
0.623
2.21
0.899
3.12
-2.733
-3.14
-4.391
-1.26
1.937
3.90
0.162
0.55
0.154
7.04
0.705
4.98
4.117
1.51
4.238
3.89
-0.344
-5.50
-4.522
-5.21
-0.271
-0.88
0.705
4.98
**
***
***
***
***
***
***
***
***
***
VI
0.565
2.32
0.892
3.09
-2.737
-3.14
-4.262
-1.22
1.936
3.89
0.153
0.52
0.154
7.04
0.712
5.03
4.214
1.54
4.228
3.88
-0.344
-5.50
-4.513
-5.23
-0.272
-0.88
0.712
5.03
**
***
***
***
***
***
***
***
***
***
VII
0.631
2.67
0.870
3.03
-2.789
-3.20
-4.301
-1.25
2.029
4.11
0.103
0.35
0.151
6.92
0.708
5.05
4.442
1.64
4.225
3.91
-0.342
-5.45
-4.313
-5.03
-0.292
-0.94
0.708
5.05
***
***
***
***
***
***
***
***
***
***
VIII
0.449
2.02
0.953
3.35
-2.635
-3.04
-3.602
-1.05
1.865
3.78
0.105
0.36
0.149
6.96
0.608
4.33
3.151
1.18
4.356
4.12
-0.345
-5.43
-4.309
-5.02
-0.308
-1.01
0.608
4.33
**
***
***
***
***
***
***
***
***
***
-0.651
-1.10
-0.022
-0.22
-6.635
-0.44
0.178
0.13
Market-to-book (state)
Unemployment (%)
-0.624
-3.02
Mean (state)
9693
9693
9693
9693
Obs.
0.33
0.33
0.33
0.33
R2
0.31
0.31
0.31
0.32
Adj. R2
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
9693
0.33
0.31
9693
0.33
0.31
9693
0.33
0.32
***
0.710
8.76
9693
0.35
0.33
***
58
Table 6. Local director labor markets, executive expertise, and appointments of local directors.
Variable definitions and sample selection criteria are presented in Appendix A. Ordinary least squares regressions. Three-digit SIC industry effects
and year effects are included. Robust t-statistics adjusted for clustering by firm are italicized. In Panel A, columns I-II use data on independent
directors to construct executive expertise; columns III-IV use data on outside directors; column V uses BoardEx data on executive expertise of outside
directors. Panels B and C use the main sample. Panel B examines representation of local executives among independent, outside, and gray directors.
Panel C examines firm size in relation to firms’ propensity to supply local directors and the service of insiders on other local firms’ boards.
Panel A: Local director labor markets and executive expertise
Executive Expertise (%)
I
Local director pool
0.651
II
III
**
0.512
2.23
IV
*
0.692
ROA
0.602
Firm age
0.399
0.598
0.64
1.14
1.84
Institutional ownership
G Index
Tangible asset intensity
R&D intensity indicator
CEO ownership
**
2.29
*
1.638
1.736
-1.346
1.614
1.35
1.75
1.78
-1.13
6.347
6.392
3.296
3.347
-0.354
1.42
1.42
0.81
0.82
-0.08
0.136
0.109
-1.416
1.182
1.154
*
1.88
0.25
0.20
*
0.678
0.676
0.454
0.677
1.11
1.14
1.90
1.90
0.030
0.031
0.019
0.020
1.09
1.15
0.78
0.83
0.120
0.119
0.233
0.235
0.65
0.64
1.41
1.43
0.373
0.227
2.613
2.550
-6.173
0.11
0.07
0.83
0.81
-1.54
1.962
3.342
1.449
1.988
1.08
1.06
1.67
1.64
2.17
-0.135
-0.135
-0.099
-0.099
-0.344
-3.471
-1.51
0.678
-3.441
-1.26
**
-2.52
*
0.679
-2.784
-1.27
**
-2.38
*
0.813
-2.761
0.814
-3.939
-0.990
1.73
2.30
2.30
-2.15
Obs.
8181
8181
8201
8201
7945
R2
0.35
0.35
0.41
0.41
0.21
0.34
0.34
0.39
0.39
0.19
Adj. R
***
**
-2.57
**
1.72
2
**
-3.43
**
-2.36
**
**
2.21
1.477
**
**
-2.13
*
0.443
*
***
4.70
*
1.350
-2.54
CEO tenure
0.740
1.33
-1.51
CEO age
*
1.326
1.93
Firm risk
**
2.08
0.225
*
*
1.89
**
2.16
Sales growth
0.743
1.93
Local director pool (similar or larger)
Firm size
Executive
Expertise (%)
V
Executive Expertise (%)
**
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
59
Panel B: Local director labor markets and appointments of local directors
Local
Independent
Directors
(%)
I
Local director pool
7.987
7.277
Sales growth
-4.352
***
-5.022
Independent
Director
Distance
to Exec. Job
(Mean)
Local
Grey
Directors
(%)
II
***
7.95
Firm size
Local
Outside
Directors
(%)
III
***
7.007
***
-9.643
7.74
IV
***
-0.259
***
0.255
2.67
-4.91
-3.46
-4.21
-3.02
4.07
0.006
-1.869
-6.211
-0.032
0.00
-0.56
-0.93
-0.14
ROA
13.575
16.938
28.906
-1.097
0.80
1.09
0.69
-1.28
Firm age
-1.913
-2.273
5.066
0.064
-1.02
-1.21
1.12
0.68
Firm risk
1.741
1.311
-3.038
-0.124
Institutional ownership
1.34
1.02
-0.135
-0.228
-1.48
-0.81
1.173
-0.446
*
-1.86
1.40
**
-0.946
-0.054
0.006
G Index
1.335
2.36
2.11
-0.73
-1.85
Tangible asset intensity
0.352
1.517
-4.799
0.180
R&D intensity indicator
7.802
1.69
2.59
2.19
-1.19
CEO ownership
0.296
0.351
0.940
-0.020
0.03
CEO age
0.13
*
10.936
-0.18
***
22.976
***
-1.64
**
-2.40
**
***
*
0.29
**
-0.273
0.83
1.04
1.08
-0.93
-1.833
-0.492
0.300
0.165
-0.39
-0.11
0.04
0.66
CEO tenure
0.380
0.383
2.132
0.038
0.27
0.28
0.69
0.49
Obs.
3365
3842
785
3365
R2
0.31
0.31
0.52
0.27
Adj. R2
0.27
0.27
0.41
0.23
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
60
Panel C: Firm characteristics and their propensity to supply local directors
Local director pool
Insiders
with Local
Directorships
(%)
D(Insiders
with Local
Directorships
>0)
Local
Directorships
per
Insider
I
II
III
2.107
***
5.27
Firm size
3.676
4.71
***
0.054
***
-0.039
6.17
Sales growth
-4.531
0.025
***
0.030
5.23
***
0.043
***
-0.058
7.30
-4.47
-2.79
-4.32
7.719
0.092
0.118
Firm age
2.441
2.67
2.42
Firm risk
-0.398
-0.013
-0.81
-1.81
-1.52
Institutional ownership
-0.035
-0.001
-0.001
1.33
***
-0.92
G Index
0.588
0.029
0.010
*
-0.120
2.32
0.033
*
-0.010
2.78
0.008
**
2.29
*
-0.101
-1.66
-1.81
-1.63
R&D intensity indicator
0.574
-0.009
0.013
0.30
-0.34
0.52
CEO ownership
-0.088
-0.001
-0.002
-0.97
-0.88
-1.66
CEO age
-1.619
-0.024
-0.022
-1.06
0.022
***
-0.98
***
-7.928
**
*
2.82
Tangible asset intensity
-0.92
***
1.71
**
-1.29
**
***
5.09
ROA
1.49
***
*
-1.03
***
0.024
CEO tenure
1.522
2.55
2.67
2.74
Obs.
9628
9628
9628
R2
0.17
0.18
0.17
Adj. R2
0.15
0.16
0.15
***
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
61
Table 7. Expertise in local director labor markets and board expertise.
Variable definitions and sample selection criteria are presented in Appendix A. Ordinary least squares regressions. Three-digit SIC industry effects
and year effects are included. Robust t-statistics adjusted for clustering by firm are italicized. Panels A, B, and D use Boardex data on outside director
education and professional experience and affiliations. Panel C uses the main sample. In Panels A and C, R&D firms are firms for which R&D
intensity indicator equals one. In Panel B, regulated firms are financials (SIC codes 6000-6999) and regulated utilities (SIC codes 4900-4999).
Panel A: Directors with academic expertise
Dep. var.: Academic Expertise (%)
Sample:
I
All firms
Universities
1.523
3.59
***
Universities * R&D intensity indicator
Firm size
Sales growth
ROA
Firm age
Firm risk
Tangible asset intensity
R&D intensity indicator
CEO age
CEO tenure
CEO ownership
1.755
7.54
-1.574
-2.45
-0.673
-0.22
1.067
2.57
0.168
1.01
-5.020
-2.11
2.288
2.37
-0.282
-0.34
0.807
2.87
0.068
1.10
***
**
**
**
**
***
II
All firms
0.396
0.76
2.552
3.24
1.781
7.69
-1.570
-2.45
-0.209
-0.07
1.009
2.43
0.159
0.96
-5.291
-2.25
-0.882
-0.63
-0.233
-0.28
0.764
2.72
0.068
1.08
III
R&D firms
IV
Non-R&D firms
3.039
4.56
***
0.652
1.24
2.310
6.23
-0.813
-0.81
-3.619
-0.85
0.978
1.46
0.128
0.48
1.675
0.38
***
1.319
4.41
-1.727
-2.31
-2.643
-0.69
0.654
1.16
0.049
0.24
-8.070
-2.72
***
***
**
**
**
***
Obs.
9694
9694
R2
0.24
0.24
0.22
0.22
Adj. R2
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
0.282
0.19
1.132
2.41
0.040
0.30
4287
0.30
0.27
**
-0.927
-0.96
0.590
1.76
0.061
0.93
***
**
***
*
5408
0.26
0.23
62
Panel B: Directors with legal and financial expertise
Dep. var.:
Legal Expertise (%)
Sample:
All firms
Universities
I
0.467
1.09
Legal Expertise (%)
All firms
II
Universities * Regulated
Law firms
0.601
2.15
All firms
All firms
III
0.007
0.01
2.210
2.03
IV
0.080
0.18
Legal Expertise (%)
Regulated
Regulated
firms
firms
V
VI
2.325 **
2.32
Legal Expertise (%)
Unregulated
Unregulated
firms
firms
VII
VIII
-0.023
-0.05
Financial Expertise (%)
All firms
All firms
IX
X
**
**
1.680
2.58
Law firms * Regulated
1.491
2.27
**
0.323
1.04
**
Financial institutions
0.370
1.84
*
-0.480
-2.02
0.293
0.40
0.479
0.16
-1.343
-2.98
0.390
2.09
-1.919
-0.73
0.785
0.69
**
Financial institutions * Financial
Firm size
Sales growth
ROA
Firm age
Firm risk
Tangible asset intensity
R&D intensity indicator
Regulated firm indicator
0.587
2.64
0.528
0.88
0.807
0.31
0.134
0.32
0.082
0.50
-2.630
-1.13
-0.909
-0.90
***
0.541
2.42
0.541
0.90
0.980
0.38
0.134
0.32
0.077
0.47
-2.329
-1.00
-0.853
-0.85
**
0.569
2.55
0.495
0.82
0.656
0.25
0.158
0.38
0.073
0.44
-2.295
-0.99
3.949
0.64
**
0.544
2.42
0.492
0.82
0.729
0.28
0.149
0.36
0.071
0.43
-2.300
-0.99
**
0.293
0.48
2.387
1.70
5.105
0.49
-0.553
-0.49
0.051
0.12
0.760
0.08
*
0.136
0.22
2.355
1.68
4.655
0.46
-0.610
-0.54
0.077
0.19
0.371
0.04
*
0.661
2.76
0.091
0.14
0.278
0.10
0.282
0.63
0.107
0.57
-2.99
-1.27
***
0.627
2.61
0.096
0.14
0.515
0.19
0.28
0.63
0.102
0.55
-2.76
-1.17
***
***
**
0.351
1.62
0.140
0.24
-0.482
-2.04
0.291
0.40
0.470
0.16
-1.340
-2.98
0.389
2.09
-1.939
-0.74
0.787
0.69
**
***
**
4.144
0.68
Financial firm indicator
CEO age
1.206
1.182
1.175
1.166
1.40
1.38
1.36
1.36
0.496 *
0.491 *
0.495
CEO tenure
0.497 *
1.81
1.81
1.79
1.81
CEO ownership
0.055
0.055
0.056
0.054
0.95
0.95
0.99
0.95
Obs.
9694
9694
9694
9694
R2
0.18
0.18
0.18
0.18
0.16
0.16
0.16
0.16
Adj. R2
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
*
3.360
1.42
-0.782
-1.31
0.051
0.31
1694
0.10
0.08
3.295
1.42
-0.746
-1.25
0.034
0.21
1694
0.10
0.08
0.798
0.87
0.777
2.54
0.049
0.80
8000
0.185
0.162
**
0.787
0.86
0.774
2.54
0.049
0.80
8000
0.186
0.162
**
26.898
9.74
-1.930
-2.16
-0.768
-2.81
2.E-04
0.E+00
9635
0.17
0.15
***
**
***
26.555
8.41
-1.936
-2.17
-0.768
-2.81
4.E-05
0.E+00
9635
0.17
0.15
***
**
***
63
Panel C: Directors with technology and R&D experience
Dep. var.:
R&D
Experience
Subsample
All firms
I
Local director pool (R&D)
4.048
***
3.44
R&D
Experience
R&D firms
only
II
4.287
Tech.
Experience
All firms
III
Tech.
Experience
Similar
Growth
Experience
High-tech
only
IV
V
**
2.34
Local director pool (tech)
4.295
***
5.25
10.314
***
4.99
Local director pool (similar growth)
2.644
***
2.82
-0.405
1.336
1.07
-0.24
-0.15
-2.42
2.30
Sales growth
3.257
2.707
4.619
6.616
3.469
0.85
0.54
1.46
0.78
1.04
ROA
5.353
8.230
-10.334
-11.131
-0.398
-0.72
-0.38
-0.03
-1.686
-0.654
-2.694
0.34
-0.161
0.42
**
***
-5.925
**
Firm size
2.068
Firm age
-3.975
-2.14
-2.78
-1.16
-0.16
-2.08
Firm risk
0.913
0.277
0.675
2.645
0.849
0.76
0.18
0.60
0.99
0.76
Institutional ownership
0.041
0.086
-0.080
-0.133
-0.062
0.45
0.69
-1.06
-0.76
-0.93
G Index
0.666
0.408
-0.247
-1.861
-0.671
1.18
Tangible asset intensity
25.441
R&D intensity indicator
1.843
**
2.34
-6.710
0.57
-0.57
-1.62
-1.77
25.656
-9.432
-25.166
0.975
1.42
0.39
CEO ownership
CEO age
CEO tenure
-1.13
-0.83
0.12
4.021
-6.950
-5.187
1.26
-0.46
-1.73
-0.129
-0.086
-0.266
0.022
-0.428
-0.40
-0.16
-1.07
0.04
**
-1.680
2.194
-6.686
-0.36
0.34
-2.28
-1.81
-2.292
0.090
-5.853
0.09
-2.30
-2.091
-1.72
*
-1.35
Technology firm
-13.978
**
**
*
*
-1.63
*
-2.546
**
1.496
-0.75
1.40
14.729
1.61
Obs.
3813
2206
3813
645
R2
0.31
0.25
0.29
0.38
3811
0.11
Adj. R2
0.27
0.21
0.24
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
0.35
0.06
64
Panel D: Director educational qualifications
Dep. var.:
Universities
Director Education Director Education Director Education Director Education Director Education Director Education
(Advanced Degree) (%)
(MBA) (%)
(MBA) (%)
(Misc. Master's) (%)
(Law) (%)
(Doctoral) (%)
I
II
III
IV
V
VI
3.360 ***
5.18
1.460 ***
0.434
0.994 **
2.88
1.01
2.44
2.540 ***
B-schools
4.12
2.719***
B-schools (top)
3.46
Firm size
3.475 ***
9.98
1.036 ***
1.094***
3.33
3.52
-2.835 ***
-2.865***
2.033 ***
0.643 ***
1.152 ***
7.52
2.91
4.95
-0.204
0.528
-0.459
Sales growth
-0.650
-0.64
-2.84
-2.86
-0.26
0.87
-0.70
ROA
-4.260
-3.192
-3.274
-1.359
1.316
0.960
Firm age
-0.786
-0.99
Firm risk
R&D intensity indicator
-0.76
-0.39
0.51
-1.668***
-0.502
0.069
-1.21
-2.83
-2.74
-0.98
0.485 *
0.151
0.155
0.614 ***
1.74
Tangible asset intensity
-0.74
-1.717 ***
-7.941 **
0.34
1.163 ***
0.17
2.77
0.108
0.286
0.59
0.60
2.76
0.65
-3.686
-4.158
0.188
-2.754
1.61
-4.443 **
-2.07
-1.02
-1.14
0.07
-1.17
-1.96
5.476 ***
2.431 *
2.273
3.712 ***
-0.470
1.654 *
3.68
1.74
1.63
3.09
-0.47
1.77
CEO ownership
-0.143
-0.098
-0.093
-0.061
0.058
0.002
CEO age
-4.852 ***
-1.60
-3.44
-1.33
-6.447 ***
-5.64
-1.000 ***
-1.26
-1.03
1.01
0.03
-6.389***
-1.759 *
0.788
-0.509
-5.61
-1.95
0.91
-1.019***
0.129
0.596 **
-0.61
0.608 **
CEO tenure
0.197
0.47
-2.64
-2.70
0.41
2.18
2.24
Obs.
9694
9694
9694
9694
9694
9694
R2
0.28
0.21
0.21
0.23
0.18
0.19
0.21
0.15
0.17
0.26
0.19
0.19
Adj. R2
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
65
Table 8. The effect of board independence on CEO pay and turnover, and firm value and operating performance
Variable definitions and sample selection criteria are presented in Appendix A. Observations in the top quartile of total assets are excluded from the sample. The dependent variables are: Incentive/Total
CEO pay (column I); Incentive/Total CEO pay (II) (column II); Total CEO pay (column III); indicator for CEO turnover (excluding CEO deaths) in a given year (column IV); ROA (in % terms)
(columns V-VI); Market-to-book ratio (columns VII-IX). Instrumental variables regressions are used. Independent Directors (%) (and Board Size, where included) is predicted with Local director pool,
big and medium-sized city dummies, industry median of Independent Directors (%) (and industry median of Board Size in the industry, where included), and second-stage controls. Three-digit SIC
industry effects and year effects are included. Robust t-statistics adjusted for clustering by firm are italicized. First-stage Kleibergen-Paap statistic is reported. In all regressions, the null hypothesis of
weak instruments is rejected. For purposes of comparison, coefficients of key variables estimated using ordinary least squares using the same model specifications are reported at the bottom of the table.
Incentive/
Incentive/
CEO turnover
Market-toMarket-toMarket-toTotal CEO pay
ROA
ROA
Total CEO pay
Total CEO pay (II)
(excl. deaths)
book ratio
book ratio
book ratio
I
II
III
IV
V
VI
VII
VIII
IX
Independent directors (%)
0.324
Firm size
3.929
***
0.286
***
5.261
3.02
Firm risk
1.781
Institutional ownership
0.195
-1.829
-1.999
***
1.704
***
0.215
-3.17
1.414
**
0.010
2.58
2.55
2.42
-0.015
0.002
-0.060
0.25
***
4.463
3.16
-1.056
***
0.460
***
0.027
1.47
-3.35
***
0.001
-3.239
***
-0.001
0.32
0.20
***
4.462
***
-1.060
***
-3.236
0.086
1.045
***
-0.106
***
-0.123
***
0.007
3.17
0.087
5.93
4.01
-2.59
5.04
4.95
3.87
0.006
0.002
-0.091
-0.091
-0.031
Tangible asset intensity
-7.549
R&D intensity indicator
56.622
-0.56
0.17
0.99
-6.529
-0.936
-0.067
***
54.761
5.23
CEO ownership
-0.356
5.69
***
-0.440
-3.54
-4.20
CEO age
-2.677
-4.784
CEO tenure
-1.996
-1.62
-3.86
Dividend yield
Dual class firm
-1.47
-2.932
-5.68
***
5.47
***
-0.054
6.971
-0.96
***
6.966
***
-0.437
-0.095
***
-0.129
***
0.007
**
-0.029
**
-0.509
***
0.055
8.45
-0.526
***
0.055
***
-0.136
-5.20
-5.24
3.46
**
-0.035
**
-0.499
***
0.057
-2.07
-2.30
-0.17
-5.01
-5.01
7.33
7.39
7.04
-2.20
-0.001
0.020
0.008
0.007
0.56
0.54
1.52
1.32
-0.11
***
-0.361
0.235
0.225
-0.055
-0.049
-0.018
***
-0.196
-2.18
2.19
0.43
**
0.437
2.17
-0.75
**
0.051
-0.65
*
1.78
0.004
-0.038
0.92
-2.13
0.048
-0.035
-1.87
***
-0.22
*
1.67
**
**
-2.03
0.020
0.46
**
-2.38
3.64
0.436
***
***
-0.437
**
***
-2.76
-1.82
-1.58
***
8.24
-0.137
-0.015
***
3.62
1.080
*
-1.92
-2.26
***
**
0.06
***
-3.23
-2.97
***
14.644
-0.99
*
**
-5.17
-0.126
*
1.039
-2.33
-15.01
***
-0.31
***
8.48
-3.10
-15.59
***
-0.83
***
5.33
***
0.011
-0.033
-4.37
-1.63
**
2.46
-0.335
-1.88
0.010
0.076
G Index
-1.45
***
2.52
0.010
4.76
**
0.084
4.73
3.78
0.074
1.97
-0.196
-1.36
**
0.003
0.62
***
0.076
**
-2.49
3.88
***
-15.39
1.54
***
0.002
0.03
2.051
0.47
-2.492
***
8.57
0.668
Firm age
4.E-04
2.76
6.36
Sales growth
***
0.056
*
1.83
*
-0.047
***
-2.84
-0.036
66
-0.31
Independent dir. block.
-0.101
Institutional block.
-0.044
-1.53
-0.32
Board size
0.105
-0.229
0.05
-0.69
Obs.
6566
6566
6566
6902
7282
7282
7278
7278
6162
First-stage Kleibergen-Paap stat.
63.71
63.71
63.71
79.29
76.44
59.53
76.89
58.82
61.30
0.003
-2.E-04
-0.016
-0.016
0.001
0.001
0.002
0.66
-0.69
-1.26
-1.26
0.69
0.67
1.31
OLS: Independent directors (%)
0.055
1.94
*
0.079
***
2.84
Board size
0.152
-0.159
0.17
-1.25
The symbols ***, ** and * denote significance at the 1%, 5% and 10% levels respectively.
67