Board Characteristics, Accounting Report Integrity, and the Cost of Debt

Abstract

Creditor reliance on accounting-based debt covenants suggests that debtors are potentially concerned with board of director characteristics that influence the integrity of financial accounting reports. In a sample of S&P 500 firms, we find that the cost of debt is inversely related to board independence and board size. We also find that fully independent audit committees are associated with a significantly lower cost of debt financing. Similarly, yield spreads are also negatively related to audit committee size and meeting frequency. Overall, these results provide market-based evidence that boards and audit committees are important elements affecting the reliability of financial reports.

Full text

Board Characteristics, Accounting Report Integrity, and the Cost of Debt
November 15, 2003
Ronald C. Andersona, Sattar A. Mansib, and David M. Reebc*
a Kogod School of Business, American University, Washington, DC 20016
b Pamplin College of Business, Virginia Tech, Blacksburg, VA 24061
c Fox School of Business, Temple University, Philadelphia, PA 19122
Abstract
Creditor reliance on accounting-based debt covenants suggests that debtors are potentially
concerned with board of director characteristics that influence the financial accounting process. In a
sample of S&P 500 firms, we find that the cost of debt financing is inversely related to board
independence and board size. We also examine the impact of audit committee characteristics on
corporate yields spreads as audit committees are the direct mechanism that boards use to monitor
the financial accounting process. We find that fully independent audit committees are associated
with a significantly lower cost of debt financing. Similarly, yield spreads are also negatively related to
audit committee size and the number of audit committee meetings. Overall, these results provide
market-based evidence that boards and audit committees are important elements affecting the
reliability of financial reports.
JEL Classification: M4, K0, G3
Key Words: Accounting Process, Debt Covenants, Audit Committee Composition, Board
Composition, Corporate Governance, Financial Statements, and Accounting Information
We would like to thank Anup Agrawal, Augustine Duru, Scott Lee, Bob Thompson, and seminar
participants at American University, Temple University, Texas Tech, Virginia Tech and an
anonymous referee for their helpful suggestions. All remaining errors are the sole responsibility of
the authors.
* Contact Information: Fox School of Business, 205 Speakman Hall, Temple University,
Philadelphia, PA 19122, E-mail Address: dreeb@temple.edu

1
Accounting-based numbers are a persistent and traditional standard that creditors use to
assess firm health and viability. Smith and Warner (1979) note for instance, that such criteria have
been used in lending agreements and debt covenants for hundreds of years. Firms violating these
accounting-based standards allow debt holders, as senior claimants, to liquidate projects or
renegotiate lending contracts (DeFond and Jiambalvo (1994)). Managers as such, may have
incentives to issue misleading financial statements to conceal negative news and thereby provide
private personal benefits or potential shareholder benefits (Dechow, Sloan, and Sweeney (1996)).
The importance creditors place on accounting numbers and the countervailing managerial incentives
to manipulate these reports suggests that bondholders potentially exhibit great concern over factors
influencing the reliability and validity of the financial accounting process (Smith (1993) and Leftwich
(1983)).1
From a creditor’s perspective, perhaps one of the most important factors influencing the
integrity of the financial accounting process involves the board of directors.2 Boards of directors,
among other tasks, are charged with monitoring and disciplining senior management, and lending
agreements typically require that boards supply audited financial statements to the firm’s creditors
(Daley and Vigeland (1983), DeFond and Jiambalvo (1994), and Dichev and Skinner (2002)). Klein
(2002a), Carcello and Neal (2000), Beasley (1996), and Dechow, Sloan, and Sweeney (1996) examine
the importance of directors monitoring the financial accounting process and document a relation
between board characteristics and manipulation of accounting information. Board attributes that
influence the validity of accounting statements may thus be of great importance to creditors. Smith
and Warner (1979) suggest that creditors price the firm’s debt to reflect the difficulties in ensuring
1 Anecdotal accounts in the popular press are illustrative. For example, recent announcements concerning the
reliability of the financial statements at Levi Straus led to a sharp reduction in the price of Levi’s bonds (see Wall
Street Journal, June 2, 2003, p. A3).
2 Swiss Venture Funds for instance, typically requires a non-voting seat on a firm’s board of directors (board
observation rights) for the firm to obtain Swiss Venture’s private-unsecured debt. Similarly, Caltius, UPS, Alliance
Capital, and numerous other firms in the private debt market typically seek observation rights for board of director
meetings. Thus, as Standard & Poor’s note in their credit rating documentation, board oversight of the accounting
information process is a paramount concern in assessing firm default risk.

2
the validity of the lending agreement, indicating that if board structure is an important oversight
element in the financial accounting process, debt prices may be sensitive to board of director
characteristics.
In this study, we examine the relation between board structure and the cost of debt
financing. Based on the proposition that independent directors are superior monitors of
management and likely to provide credible financial reports, we test the prediction that the firm’s
cost of debt (yield spread) is inversely related to the proportion of independent directors on the
board. We also examine the relation between board size and the cost of debt financing. Klein
(1998, 2002b) indicates that the number of directors on the board affects committee assignments
and board monitoring. Similarly, Adams and Mehran (2001) suggest that bigger boards increase
monitoring effectiveness and provide for greater board expertise. As such, we posit that debt yields
are negatively related to board size as larger boards may increase the level of managerial monitoring
(i.e., a greater number of guards) and enhance the financial accounting process.3
For most large firms, boards of directors delegate direct oversight of the financial accounting
process to a subcommittee of the full board, the audit committee. Audit committees are responsible
for recommending the selection of external auditors to the full board; ensuring the soundness and
quality of internal accounting and control practices; and monitoring external auditor independence
from senior management. Recent regulations put forth by the major stock exchanges requiring that
a minimum of three independent directors serve on the audit committee suggest that committee
independence and size may be integral factors for firms in delivering meaningful financial reports
(Klein (2002a)). Carcello and Neal (2000) provide support for this argument by documenting a
relation between greater audit committee independence and the quality of financial reporting. If
3 Lipton and Lorsch (1992), Jensen (1993), and Yermack (1996) argue that larger boards are less effective in group
decision-making and strategy formulation, which suggests that equity holders would have divergent interests from
debtors on board size. We explore this issue in greater detail in Section I.

3
audit committee composition influences the financial accounting process, we then anticipate that
corporate debt yields exhibit an inverse relation to committee independence and size.
Using a sample of 252 industrial firms from the Lehman Brothers Fixed Income database
and the S&P 500, we find that board independence is associated with a lower cost of debt financing.
After controlling for industry and firm specific attributes, our analysis indicates that debt costs
(using non-provisional publicly traded debt) are 17.5 basis points lower for firms with boards
dominated by independent directors (51 percent independents) relative to firms with insider-stacked
boards (25 percent independents). We also find a negative relation between board size and the cost
of debt financing. Specifically, we find that an additional board member is associated with about a
10 basis point lower cost of debt financing. The results are robust to various measures of board
independence, board size, endogeneity, non-linear specifications, and are both economically and
statistically significant. Overall, our empirical results indicate that bondholders view board
independence as an important element in the pricing of the firm’s debt, suggesting that creditors are
sensitive to board attributes that affect reporting validity.
The analysis also indicates that creditors view audit committees and their characteristics as
important elements in the financial accounting process. Specifically, we find that the cost of debt is
about 15 basis points lower for firms with fully-independent audit committees relative to those with
insiders or affiliates on the committee. Focusing on the size of audit committees, we find that
committee size ranges from one to 12 directors, with most committees having either 4 or 5
members. The analysis indicates that for the average-size audit committee, an additional board
member is associated with a 10.6 basis point lower cost of debt. Although, the reduction in the cost
of debt for this additional audit committee member appears large, an additional member results in
approximately a 20% change in committee size.4 One implication is that creditors view audit
4 One potential concern is that the audit committee size results are an artifact of firm size. Although we control
for firm size in our primary regressions, we also perform subsequent tests in Section 5 that use audit committee
size scaled by firm size and size-based subset tests (with similar results). In addition, it is important to note that
this lower cost of debt is for an additional audit committee member in the average size committee. As audit
committee size continues to increase, we find a diminishing cost of debt for each additional member. For

4
committees with 4 or 5 members very differently than those with only 1 or 2 members. Overall,
these results suggest that creditors view audit committees as an important device in ensuring the
reliability of accounting reports.
We also conduct supplemental analysis on independent director attributes. Monks and
Minnow (1995) and Beasley (1996) suggest that director expertise or occupational characteristics
may influence the board’s ability to effectively monitor management and the firm. Our results
indicate that independent-director employment characteristics (executive, retired, academic, etc.),
while all significantly related to lower debt costs, are not substantively different from one another.
Recent passage of the Sarbanes-Oxley Act by the US Congress also requires that at least one
“financial expert” serve on the firm’s audit committee.5 In light of this new regulation, we also
investigate whether financial experts influence the cost of debt financing. Consistent with our earlier
results on director employment characteristics, we find no relation between debt costs and financial
experts serving on the audit committee. In sum, these tests suggest that the primary concern of
creditors is the presence of independent directors on the board and audit committee, as opposed to
director expertise.
The investigation also suggests that director equity ownership is not related to the cost of
debt financing. In contrast, board tenure is positively related to corporate yield spreads, suggesting
that as director tenure increases, managers are potentially more able to influence or sway board
opinion. Audit-committee meeting frequency also exhibits a negative relation to debt costs,
indicating bondholder concern with directors actively monitoring the financial accounting process.
example, at the top decile of the distribution for audit committee size, the analysis suggests only a 5 basis point
lower cost of debt.
5 The Sarbanes-Oxley Act specifies that for a director to be classified as a financial expert that the individual
should have knowledge through education or work experience of: (i) Generally Accepted Accounting Principles
(GAAP), (ii) preparing or auditing public company financial statements, (iii) applying GAAP in connection with
accounting estimates, accruals and reserves, (iv) internal accounting controls and, (v) the audit committee function.
Due to a large number of registrant complaints, the SEC expanded the definition of a financial expert to include
any director that has supervised any finance or accounting personnel – essentially suggesting that any executive
who has managed a financial/accounting employee is a financial expert. For the tests in this paper, we use a
narrower definition of financial expert. Section II provides greater detail on our categorization.

5
To assess the robustness of our results, we conduct regressions using first differences in
bond yields against first differences in board characteristics. The analysis suggests that changes in
yield spreads are significantly (and negatively) related to increases in board size (13.0 basis points per
board member added) and independence (11.4 basis points for an increase of 10 percent in board
independence). These findings imply that debtors view larger, more independent boards (audit
committees) as more effective monitoring mechanisms and thereby enhance the financial accounting
process.
Although our results are consistent with the hypothesis that board structure influences the
accounting reports that creditors use in managing lending agreements, an alternative explanation for
the observed relation focuses on firm performance. Specifically, Monks and Minnow (1995) argue
that board monitoring can improve the quality of managerial decision-making and lead to better firm
performance; suggesting that better firm performance results in lower yield spreads. Prior literature
however, provides little evidence of independent boards improving firm performance (see Hermalin
and Weisbach (2003)). Still, we include several proxies for firm performance (e.g. cash flows, credit
ratings, etc.) in our analysis to alleviate this concern. In addition, the results indicate that audit
committee structure – a direct link between boards and financial reporting – affects the cost of debt.
Finally, because audit committee characteristics may simply capture full board attributes, we examine
whether audit committee characteristics exhibit incremental explanatory power over full board traits.
Again we observe a negative relation between audit committee attributes and the cost of debt
financing; suggesting that the link between debt costs and audit/board characteristics is the financial
accounting process. However, to the extent that these measures do not fully capture firm
performance, both the accounting process hypothesis and firm performance potentially explain the
documented relation.
This research contributes to the literature in several important ways. First, our analysis
suggests that debt holders exhibit interest in board and audit committee monitoring of the financial
accounting process. Second, our analysis provides support for the notion that board independence

6
and board size influence the cost of debt financing. We interpret this to suggest that bondholders
are concerned with governance mechanisms that limit managerial opportunism and improve the
financial accounting process. Third, our analysis suggests that larger, more independent audit
committees provide a measurable and significant benefit to the firm, namely through a lower cost of
debt financing. These results provide some support for recent regulatory and listing requirements
(see NYSE, NASDAQ, recent SEC proposals, and the Sarbanes-Oxley Act) concerning audit
committee independence; as well as calls for more actively involved audit committees. Fourth, we
find that director independence, rather than director expertise, is the more relevant issue in the cost
of debt capital. In aggregate, our analysis provides market-based evidence to suggest that boards
and audit committees are important mechanisms in overseeing the financial accounting process.
The remainder of this paper is organized as follows. Section I develops our testable
hypothesis and section II describes our sample and gives summary statistics. Section III provides
the multivariate analysis and section IV examines alternative specifications and test procedures.
Section V concludes the paper.
I. Board Characteristics and Monitoring the Financial Accounting Process
The Securities and Exchange Commission, the Financial Accounting Standards Board, and
the major stock exchanges regularly emphasize the role of board of directors in overseeing the
financial accounting process. Boards comprising mostly employee or employee-related directors
may be more willing to conceal negative information to gain private benefits or to limit stakeholder
intervention in the firm. Recent reports in the financial press suggest that some boards “shut their
eyes when the numbers are squishy or even fraudulent,” leading to several well-publicized scandals.6
Yet, Beasley (1996), Dechow, Sloan, and Sweeney (1996), and Fama and Jensen (1983) suggest that
independent directors are more willing to provide effective oversight and disclosure due to their
desire to maintain their reputations. In the debate over director efficacy, prior literature primarily
6 New York Times (1/26/03) pages B1 and B12 (The Revolution That Wasn’t).

7
focuses on four board characteristics; (i) board independence, (ii) board size, (iii) committee
structure, and (iv) specific occupational characteristics or expertise of independent directors. In the
following sub-sections, we develop testable hypotheses on the relation between debt yields and
board structure.
A. Board Independence, the Financial Accounting Process, and the Cost of Debt Financing
Smith and Warner (1979) and Kalay (1982) observe that bondholders’ concerns lie with
protecting their investment.7 One of the more important elements in bondholders’ ability to protect
their investments is the firm’s financial accounting numbers. Creditors use accounting numbers to
judge compliance with debt covenants and to administer lending agreements (DeFond and
Jiambalvo (1994) and Daley and Vigeland (1983)).
Boards of directors have a primary responsibility of overseeing the firm’s financial reporting
process. Boards meet routinely with the firm’s accounting staff and external auditors to review
financial statements, audit procedures, and internal control mechanisms (Klein (2002a)). As such,
bondholder’s potentially view boards of directors and, in particular, board structure as critical
elements in delivering credible and relevant financial statements.
Prior literature generally posits that board of director independence from senior
management provides, among other things, the most effective monitoring and control of firm
activities. Byrd and Hickman (1992) for instance, suggest that independent directors contribute
expertise and objectivity that minimizes managerial entrenchment and expropriation of firm
resources. Beasley (1996) and Dechow, Sloan, and Sweeney (1996) find that the proportion of
independent directors on the board (board independence) is inversely related to the likelihood of
financial statement fraud. More recently, Klein (2002a) documents a negative relation between
abnormal accruals and director independence from senior management. If independent boards
7 Branch (2000) and Perumpral, Davidson, and Sen (1999) discusses creditor rights and fiduciary responsibilities in
bankruptcy. Mansi, Maxwell, and Miller (2003) discuss the impact of auditor choice and tenure on creditors, while
Begley (1990) discusses debt covenants and accounting choices. Betker (1995) examines creditor and board issues

8
provide superior oversight of the financial accounting process, then we expect bondholders to
directly benefit through greater transparency and validity in accounting reports. This leads to our
first testable hypothesis:
Hypothesis 1: Greater board independence is associated with lower corporate-debt yield spreads.
B. Board Size, the Financial Accounting Process, and the Cost of Debt Financing
Recent research also indicates that board size may play an important role in directors’ ability
to monitor and control managers. Lipton and Lorsh (1992) and Jensen (1993) for instance, argue
that because of difficulties in organizing and coordinating large groups of directors, board size is
negatively related to the board’s ability to advise and engage in long-term strategic planning. In
contrast, Adams and Mehran (2002) and Yermack (1996) suggest that some firms require larger
boards for effective monitoring. Chaganti, Mahajan, and Sharma (1985) posit that large boards are
valuable for the breadth of their services. Klein (2002b) for instance, finds that board committee
assignments are influenced by board size since large boards have more directors to spread around.
As such, she suggests that board monitoring is increasing in board size due to the ability to distribute
the work load over a greater number of observers. Monks and Minow (1995) and Lipton and
Lorsch (1992) extend this argument by suggesting that larger (smaller) boards are able to commit
more (less) time and effort to overseeing management.8 If large boards are more effective monitors
of the financial accounting process, then bondholders should benefit through improved financial
transparency and reliability. This leads to our second testable hypothesis:
Hypothesis 2: Larger boards of directors are associated with lower corporate-debt yield spreads.
in default and Duke and Hunt (1990) discuss creditor demands for monitoring.
8 Monks and Minow (1995) note that most companies typically have several committees and that larger boards
allow for fewer committee assignments per director. In this context, a larger board provides for greater task
sharing and potentially better monitoring. However, the negative aspects to larger boards may be quite relevant to
shareholders, such as increased formalism, slower decision-making, and greater inflexibility.

9
C. Audit Committee Structure and Cost of Debt Financing
Although boards of directors are responsible for oversight of the financial accounting
process, this task is often delegated to a subcommittee of the full board, the audit committee. The
audit committee plays an important role because it is concerned with establishing and monitoring
the accounting processes to provide relevant and credible information to the firm’s stakeholders
(Pincus et al (1989), Beasley (1996)). The 1999 Blue Ribbon Committee Report (co-sponsored by the
New York Stock Exchange (NYSE) and the National Association of Security Dealers (NASD))
indicates that independent audit committee members are better able to protect the reliability of the
accounting process. Following the report, the NYSE and the NASD, along with the SEC, proposed
that listed firms maintain standing audit committees with at least three independent directors; with
the express purpose of monitoring the accounting information process (Klein (2002a)). If
independent audit committees provide more reliable accounting information (relative to insider-
stacked committees), then we expect the cost of debt to be related to audit committee composition.
This leads to our third hypothesis:
Hypothesis 3: Greater audit committee independence is associated with lower corporate-debt yield spreads.
D. Audit Committee Size and Cost of Debt Financing
The recent regulations put forth by the major stock exchanges stipulating that audit
committees comprise at least three members implies that governing bodies deem audit committee
size as an integral attribute in controlling the accounting process. Pincus et al (1989) suggest that
audit committees are an expensive monitoring mechanism and that firms with greater agency costs
are potentially more willing to bear these expenses. In this context, firms with larger audit
committees are willing to devote greater resources to overseeing the financial accounting process. A
firm with an audit committee composed of only a couple of members would, on average, have less
time to devote to overseeing the hiring of auditors, questioning management, and meeting with

10
internal control system personnel. If large audit committees better protect and control financial
standards than small committees, we then expect greater accounting transparency and a lower cost
of debt financing. This leads to our fourth testable hypothesis:
Hypothesis 4: Larger audit committees are associated with lower corporate-debt yield spreads.
E. Director Characteristics, the Financial Accounting Process, and the Cost of Debt Financing
Effective monitoring also requires both expertise and proper incentives (Beasly (1996)).
Fama and Jensen (1983) suggest that independent directors are effective monitors because of
reputation concerns and their desire to obtain additional director positions. Jensen and Meckling
(1976) argue that director equity-ownership creates powerful incentives for directors to monitor
management. Generally, the literature suggests that professional directors and directors with equity
stakes are associated with greater monitoring.
Brickly, Coles, and Terry (1994) report that retired executives from other companies are also
effective monitors. Similarly, Monks and Minow (1995) suggest that academics are less effective
directors relative to those with business experience. As monitoring expertise increases, managerial
opportunism becomes less prevalent, causing the value of investor claims to increase. Furthermore,
effective monitoring is potentially an acquired skill, suggesting boards with greater tenure provide
greater monitoring. However, as board tenure increases, managers may be better able to influence
or sway director opinion, indicating director tenure exhibits an inverse relation to oversight of the
financial accounting process. If director experience, tenure, or equity ownership creates incentives
for independent directors to more closely monitor firm management, then we expect bondholders
to benefit through credible and transparent financial statements. This leads to our fifth testable
hypothesis:
Hypothesis 5: Greater board expertise (ownership) is associated with lower corporate-debt yield spreads.

11
Finally, we focus on audit-committee director attributes and committee meeting frequency.
The Sarbanes-Oxley Act requires that audit committees include at least one “financial expert.”
Similar to our arguments on director experience, we posit that financial experts on the audit
committee lead to greater rigor in financial reporting. The 1999 Blue Ribbon Committee Report likewise
advocates that the audit committee, as the watchdog of the financial accounting process, can best
assure the quality of the financial statements by having at least 4 meetings a year (Morrissey (2000)).
If financial expertise on the audit committee or committee meeting frequency improves the financial
accounting process, we anticipate a negative relation between these attributes and debt costs. This
leads to our final two hypotheses:
Hypothesis 6: Financial expertise on the audit committee is associated with lower corporate-debt
yield spreads.
Hypothesis 7: Audit committee meeting frequency is associated with lower corporate-debt yield spreads.
F. Research Focus
Our study tests the proposition that bondholders, as important end users of accounting
information, are sensitive to board of director characteristics that influence the financial accounting
process. We address seven specific questions relating to board effectiveness in monitoring the
reporting process. First, is there a relation between board independence and the cost of debt
financing? Second, is board size related to the cost of debt financing? Third, do independent audit
committees influence the cost of debt financing? Fourth, is audit committee size related to the cost
of debt? Fifth, does director experience, expertise, ownership, or tenure affect the cost of debt
financing? Finally, are financial experts on the audit committee or committee meeting frequency
related to corporate-debt yield spreads? Our empirical analysis of these issues on the relation
between board monitoring effectiveness and the cost of debt financing uses firm level data from
1993 through 1998 on a sample of S&P 500 firms.

12
II. Data Description
A. The Sample
For our sample, we collect information on firms that are in both the Lehman Brothers Fixed
Income database (LBFI) and the S&P 500 Industrial Index (as of December 31, 1992). The LBFI
provides month-end security-specific information on bonds that are in the Lehman Brothers
Indices. The goal of the database is to provide a representative sample of outstanding publicly
traded debt. Information is provided on coupons, yields, maturities, credit ratings from Moody’s
and S&P, bid prices, durations, convexities, holding period returns, call and put provisions, and
sinking fund provisions. Lehman Brothers selects bonds for inclusion in the database based on firm
size, liquidity, credit ratings, subordination, and maturity. The database contains non-provisional
bonds of differing maturities, differing credit ratings, and differing debt claims (senior and
subordinated debt). Although the database does not contain the universe of traded debt, we have
no reason to suspect any systematic bias within the sample. We exclude financial and utility firms
from the sample because of the potential effect of regulations on debt yields.
We manually collect data from corporate proxy statements on board structure, audit
committee composition, and other governance characteristics for the S&P 500 Industrial firms. To
gather firm-specific financial data not already included in the Lehman Brothers Database, we use the
Compustat Industrial Files. Combining the three datasets yields a sample of 1,052 firm-year
observations on 252 firms for the period 1993 through 1998.9
B. Measuring Board Structure and Yield Spreads
We categorize directors similar to Brickley, Coles, and Terry (1994). Directors employed by
the firm, retired from the firm, or immediate family members are insiders. Affiliate directors are
directors with existing or potential business ties to the firm, but are not full time employees.
Examples of affiliated directors are consultants, lawyers, financiers, and investment bankers.
9 A potential concern is survivalship bias. To minimize this concern we allow firms to exit and reenter the sample.

13
Independent directors are individuals whose only business relationship with the firm is their
directorship.
Our primary measure of board independence is the number of independent directors divided
by board size (fraction of independent directors). We also use two alternative proxies for the
influence of independent directors on the board. First, we develop a binary variable that equals one
when independent directors hold over 50 percent of the board seats and zero otherwise, denoted as
Independent Dominated board. Second, we use the number of independent directors on the board.
The actual number of independent directors may be important because of how committee
assignments are allocated, relative differences in expertise, or because it increases the probability that
someone will ask the tough/right questions. In subsequent testing, we also examine the fraction of
seats held by both inside and affiliated directors.
For consistency with prior research, our primary measure of board size is the natural log of
the total number of directors serving on the board. In addition, because firm size exhibits a positive
correlation with board size, we also develop two other measures to assess the robustness of our
results. Our first alternative measure is the ratio of the number of board members to the natural log
of total assets. Our second alternative measure of board size uses binary variables that denote
boards as either large (top quartile of board size) or small (bottom quartile of board size).
We use similar procedures for measuring audit committee independence and size. Our
primary measure of audit committee independence is the number of independents on the audit
committee divided by audit committee size (fraction of independent directors on the audit
committee). We also develop two binary variables to denote audit committee independence. The
first binary variable, fully-independent audit committees, equals one when the committee consists of
only independent directors and zero otherwise. The second binary variable, audit committees
dominated by independent directors, equals one when the committee comprises at least 50.1%
independent directors and zero otherwise. Our primary measure of audit committee size is the
natural log of the total number of directors on the committee. We also develop two alternative

14
measures of audit committee size. The first is the number of directors serving on the committee
divided by the natural log firm size. The second alternative measure of committee size uses binary
variables that denote committees as either large (top quartile of committee size) or small (bottom
quartile of committee size).
We also gather information on board qualifications such as average board age, average board
tenure, and the occupation of independent directors. Board age is the sum of all director ages
divided by the number of directors and proxies for director business experience. Board tenure is the
sum of the number of years that the directors serve on the board divided by the number of
directors. This measure captures the ability of managers to influence directors; longer tenure
potentially allows managers greater influence over directors’ decisions. As in Brickley, Coles, and
Terry (1994), we categorize independent directors into one of four occupations: executives from
other firms, retired executives from other firms, academics, and other.
The Sarbanes-Oxley Act also requires that at least one “financial expert” serve on the audit
committee. The definition of a financial expert, as stipulated by the act, includes directors with
educational or occupational experience in; generally accepted accounting principles (GAAP),
auditing public companies, applying GAAP in connection with accruals or reserves, internal
accounting controls, and the audit committee function. The SEC later expanded the definition of
the Act to include supervisory experience of individuals involved with these functions, suggesting
that any director who manages (or managed) an accounting/financial employee can be classified as a
financial expert. Although the Sarbanes-Oxley Act provides for a wide breadth of experience to be
classified as a financial expert, we use a narrower and more empirically manageable definition for our
tests. Specifically, we classify the following types of directors serving on the audit committee as
financial experts; chief financial officers (CFOs), investment bankers, investment managers (mutual
funds, etc.), bankers, financial consultants, auditors, and chief executive officers (CEOs) of financial
firms. Because of the varying background of many academics that serve on audit committees, some

15
ambiguity arises as to their specific financial expertise. Consequently, we conduct our tests both
including and excluding academics as financial experts.
Our dependent variable, yield spread (Spread), is measured as the difference between the
weighted-average yield to maturity on the firm’s outstanding (non-provisional) publicly traded debt
and the yield to maturity on a Treasury security with a corresponding duration, where the weight of
each debt issue is the fraction of amount outstanding for that issue divided by the total market value
of all outstanding traded debt for the firm. The yield on a corporate debt security is defined as the
discount rate that equates the present value of the future cash flows to the security price. The yields
on Treasury securities are constant maturity series published by the Federal Reserve Bank of New
York in its H15 release. 10
C. Control Variable Measures
We incorporate control variables into the analysis on firm and security specific attributes.
Firm specific measures include firm size (Size), leverage (Leverage), risk (Volatility), Firm
Performance, and blockholdings (Block Holdings). Security specific measures relate to our
dependent variable (Spread) and include duration (Duration), credit ratings (Rating), age of the debt
(Age), and non-linear credit ratings (NLCredit).
Firm size is the natural log of sum of the firm’s debt and equity. That is
)( EquityDebtLnSize
+
= (1)
where Debt is the sum of the firm's publicly traded (market value) and non-traded (book value)
debt, and Equity is the market capitalization of the firm. The market value of debt is computed by
multiplying the face value of the outstanding debt by its trading debt price (as a fraction of par). The
market value of equity is computed by multiplying the number of shares outstanding by the traded
closing stock price. Our results are robust to alternative measures of firm size (log of total sales or
10 In cases with no equivalent Treasury maturity, the yield is computed using the Nelson and Siegel (1987)
interpolation function.

16
total assets). Furthermore, adding additional terms for firm size (such as the square and cube of
size) also lead to similar results with regards to the hypotheses, as does size based subset regressions.
We measure leverage as the ratio of long-term debt (LT Debt) to total capital (Debt and
Equity). That is
)(
EquityDebt
DebtLT
Leverage +
= (2)
where Debt is measured as the sum of the firm's publicly traded (market value) and non-traded
(book value) debt, and Equity is the market value of equity multiplied by the number of shares
outstanding.
Although leverage controls for variations in the firm's capital structure and may proxy for
default risk (as do credit ratings, firm size, and duration), to reinforce our results, we also control for
bankruptcy risk using a measure of stock return volatility. Volatility is the standard deviation of
stock returns for the prior 60 months11. We measure firm performance (Perform) as the ratio of
cash flows (net income plus depreciation and amortization) to total assets. Finally, equity
blockholdings (Block) represent the holdings of large shareholders who own five percent or more of
the firm’s outstanding equity.12 Because shareholders can benefit from a lower long-term cost of
debt, blockholders potentially have incentives to protect bondholder interests.
Security specific variables include duration, credit ratings, and liquidity. We use duration to
control for differences in maturity and coupon of the firm's outstanding debt. Duration refers to
Macaulay duration, computed as the discounted time weighted cash flow of the security divided by
its price. That is
∑
=+
×
=
K
t
t
t
YP
CFt
Durat
1)1( (3)
11 Alternative measures of firm risk include cash flow volatility and Ohlson’s (1980) bankruptcy prediction model.
The results are robust to using either of these methods (see section IV.C.).
12 Shleifer and Vishny (1997), Anderson, Mansi, and Reeb (2003), and Bhojraj and Sengupta (2002) suggest
decomposing blockholders into founding family and institutional ownership. Our results are robust to using either
approach.

17
where CFt is the security cash flows at time t, t is the number of periods until the cash flow, P is the
market bid price of the security, Y is the yield to maturity, and K is the number of cash flows. For
our analysis, we compute Duration, or the weighted average duration of the outstanding debt, as a
linear combination of the weighted duration of each bond for each firm.
Credit ratings are used to control for differences in default risk. This measure is computed
as the average of both Moody’s and S&P bond ratings and represents the average of the firm’s credit
rating at the date of the yield observation. Reeb, Mansi, and Allee (2001) suggest using the average
of the Moody’s and S&P ratings to proxy for the default risk premium. Bond ratings are computed
using a conversion process in which AAA rated bonds are assigned a value of 22 and D rated bonds
receive a value of one. For example, a firm with an A1 rating from Moody’s and an A+ from S&P
would receive an average score of 18. The bond conversion numbers for both Moody’s and S&P
ratings are provided in Appendix A.
A potential problem is that credit ratings may already incorporate the impact of board
structure. To mitigate this concern we use an estimate of the debt credit rating without the board
structure component. We achieve this by regressing credit ratings on board size and independence.
The error term from this regression incorporates the credit rating information without the influence
or impact of board structure. We label the error term from this regression as Rating and use it as
our primary measure of credit ratings in the multivariate analysis. In addition, credit ratings may
exhibit non-linearities as many institutions are barred from holding securities below a certain grade.
Therefore, we also include a binary variable (NLCredit) to denote those firms with non-investment
grade debt. We consider alternative approaches of incorporating credit ratings and firm risk in
section IV.
For bond liquidity, the fixed income literature provides three proxies: the age of the bond,
the amount available for trade, and the bid-ask spread (Sarig and Warga (1989)). We use age of the
bond as a measure of liquidity. Bond age (Age) is the number of years that a bond has been
outstanding. This is a weighted-average difference between the settlement date and the original

18
bond issue date. For example, a bond with a settlement date of April 30, 1997, and an issue date of
January 31, 1994, has an age of 3.25 years.
D. Descriptive and Univariate Statistics
Table 1 provides descriptive information for our sample. Panel A shows means, medians,
standard deviations, and minimum and maximum vales for our key variables, Panel B provides an
industry breakdown of the firms in our sample, and Panel C presents a correlation matrix between
bond yields and board characteristics.
The average yield spread for the publicly traded debt in our sample is about 136 basis points
over its duration equivalent Treasury Security. The standard deviation of the yield spread is about
110 basis points, with a maximum and minimum value of 1,067 and two basis points, respectively.
The average bond in our sample has duration of about 6.3 years, has been outstanding for about 3.9
years, and has Moody’s credit rating of A3.
The average board size in our sample is 12.1 directors, of which 3.3 are inside directors (27
percent), 6.9 are independent directors (57 percent), and the remaining two are affiliate directors (15
percent). Board independence varies widely across our sample from zero to 92.9 percent. Average
director age on the boards is 60.3 years and average tenure is 9.2 years. There is also substantial
variability in board size, ranging from a minimum of six directors to a maximum of 24 directors.
With respect to audit committees, the average size is 4.5 directors, ranging from a minimum of one
director to a maximum of 12 directors. Independent directors hold 70.1% of all audit committee
seats and the median audit committee meets 3 times per year.
Firm size, the natural log of total capital (in millions), has a mean of about $8.88, a standard
deviation of $1.28, and a maximum and a minimum size of $12.78 and $4.40, respectively. Long-
term debt, on average, comprises 21 percent of our sample-firms total capital.

19
Panel B of Table 1 provides a breakdown of the number of firm-year observations based on
Standard Industry Classification (SIC) codes. Industries in the sample include: agriculture, forestry
and fishing, construction, manufacturing, transportation, wholesale trade, retail trade, and services.
Panel C of Table 1 provides correlation coefficients between yield spreads and board
characteristics. In general, the board structure variables exhibit a negative relation to debt yields
with the exceptions of audit committee independence and board tenure. This analysis indicates that
firms with large, independent boards are more likely to have lower debt costs, which is consistent
with the hypothesis that independent directors provide superior monitoring of the financial
accounting process, leading to lower debt costs. However, because firm size has an effect on board
independence, board size, and debt yields, we use a multivariate framework to further explore our
hypotheses.
[Insert Table 1 about here]
III. Multivariate Testing Results
A. Bond Yield Spread and Board Structure
In the primary specification, we test the cross-sectional relation between board structure
variables and the cost of debt financing, and various control measures. That is
Spreadi,t = A0 + A1 (Fraction Independenti,t ) + A2 (Log Board Sizei,t ) + A3 (Firm Sizei,t ) + A4 (Leveragei,t )
+ A5 (Durationi,t) + A6 (Bond Agei,t ) + A7 (Ratingi,t ) + A8 (Blocki,t ) + A8 (NLCrediti,t )
+ A9 (Volatilityi,t ) + A10 (Performi,t ) + A11 (Time_Dumi,t ) + A12 (Ind_Dumi,t ) +
ε
i,t (4)
where Spread is the weighted average debt yield to maturity in excess of the duration equivalent
Treasury yield. Our primary interest lies in the coefficient estimates on board independence (A1)
and board size (A2). Negative estimates are consistent with the hypotheses that independent
directors and board size improve the financial accounting process and thus, are associated with a
lower cost of debt financing.
Our firm specific control variables include: firm size, leverage, blockholdings, volatility, and
performance. We expect firm size and performance to be negatively related to yield spread as larger,

20
more profitable firms enjoy greater stability and cash flows and therefore may have a lower cost of
debt financing. Firm leverage should be positively related to yield spreads, as firms with high debt
usage are associated with higher bankruptcy costs, causing an increase in the required rate of return
to the bondholders. We expect stock return volatility to exhibit a positive relation to yield spreads,
as price fluctuations are associated with higher risk and higher yield spreads.
The remaining control variables, duration, age, rating, and non-linear credit are debt specific.
The fixed income literature suggest that duration should be increasing in maturity, but decreasing in
coupon and yield to maturity. Since the interaction of all these variables may produce a negative or
positive sign, we have no expectation on the sign of the relation between duration and yield spreads.
Also, as the yield spread is computed using the duration-equivalent Treasury security, the
construction of the dependent variable may mitigate this concern. Age of the firm’s outstanding
debt should be positively related to yield spread, as seasoned securities demand higher prices and
lower yields. The variable Ratings should be negatively related to the yield spread as firms with lower
ratings have a higher cost of debt financing. An indicator variable (NLCredit) is added to account
for those firms with debt below investment grade status. We expect these firms to be positively
associated with yield spreads as lower quality firms demand higher yields and higher spreads.
Finally, we include year and industry dummy variables to control for possible time and industry
effects. Column 1 of Table 2 provides the predicted sign for each of the coefficient estimates.
[Insert Table 2 about here]
Column 2 of Table 2 provides the primary regression results using Equation (4). The t-
values are corrected for heteroskedasticity using White standard errors.13 Our results indicate that
greater board independence is associated with a lower cost of debt financing. The coefficient
estimate on independent directors is –67.2 with a t-statistic of –4.0, consistent with the concept that
board independence provides greater managerial oversight. Economically, the coefficient indicates
13 We control for serial correlation by subtracting the Treasury security yield from the firm yield and by including
yearly dummy variables. Repeating the analysis using a random effects model and using Fama-MacBeth

21
that debt costs are about 17.47 basis points higher for firms with inside boards (25 percent
independent) relative to those with independent boards (51 percent independent directors).
The coefficient estimate on the natural log of board size is –122.24 with a t-statistic of -7.1,
indicating that firms with larger boards enjoy a lower cost of debt financing. Thus, comparing a
firm with the average board size (12 members) to a firm with one additional board member (13
members), suggests about a 9.8 basis point lower cost of debt financing. These results are also
economically significant and are consistent with the hypothesis that larger boards improve the
financial accounting process, resulting in lower debt yields.
In terms of the control variables, the coefficient estimates for rating and firm size are
negative and significant at the one percent level, while the leverage coefficient estimate is positive
but not significantly different from zero. However, credits rating potentially incorporate differences
in capital structure suggesting that leverage will be unrelated to debt spreads. The coefficients on
duration, age, and volatility are positive and significant at the one percent level. The variable
blockholdings is insignificant suggesting that large blockholders (e.g., mutual funds, insurance
companies, and investment bankers, etc.) do not have an effect on the cost of debt financing.
Finally, the non-linear credit indicator is positive and significant suggesting that investment and non-
investment grade debt are positively associated with yield spreads, but that non-investment grade
debt requires a higher rate of return and a therefore higher spread.
B. Alternative Specifications for Board Structure Variables
To further test the relation between board structure and the cost of debt financing, we
consider alternative measures of board independence and board size. Column 3 of table 2 presents
the results on board independence using a binary variable that equals one for boards dominated by
independent directors (fraction of independent directors greater than 50 percent) and zero
otherwise. Again, we find strong evidence that firms with independent boards experience a lower
regressions both lead to similar results (See section IV C. and Appendix B).

22
cost of debt financing (about 19 basis points). Column 4 of Table 2 presents the results using the
number of independent directors on the board to measure board independence. These results
suggest that an independent director, relative to an inside director is associated with about a 5.5 basis
point lower cost of debt financing.
Because board size and firm size are positively correlated, we also use alternative measures of
board size in columns 5 and 6 of Table 2 to correct for any firm size effects. Using the number of
directors divided by firm size, we again find that larger boards are associated with lower debt yields.
Column 6 of Table 2 shows regression results using two indicator variables for board size; big
boards and small boards. Big board equals one when board size falls in the largest quartile of the
board of director sample and zero otherwise. Small board equals one when board size falls in the
smallest quartile of our board of director sample. Consistent with our prior results, we find that
firms with larger boards have a lower the cost of debt financing. Overall, our analysis indicates that
board structure influences the cost of debt financing, consistent with our prediction that board
monitoring affects the financial accounting process.
C. Yield Spreads and Audit Committee Structure
To examine the relation between audit committee structure and the cost of debt financing,
we use the specification in Equation (4) and replace the board variables with those for audit
committees. Column 1 of Table 3 provides the results of regressing audit committee independence
(fraction of independents) and audit committee size (log of committee size) on corporate yield
spreads. The results indicate that greater audit committee independence is associated with a lower
cost of debt financing. The coefficient estimate on independent directors is –33.2 with a t-statistic
of –2.8, consistent with the concept that audit committee independence provides greater oversight
of the financial accounting process. The coefficient estimate on the natural log of audit committee
size is –57.9 with a t-statistic of -4.2, indicating that firms with larger audit committees have a
significantly lower cost of debt financing. Thus, comparing a firm with 6 committee members

23
relative to a firm with 5 members, suggests a debt savings of about 10.6 basis points. These results
are consistent with the hypothesis that larger, more independent audit committees are important to
creditors, resulting in lower debt yields.
[Insert Table 3 about here]
To further test the relation between audit committee structure and the cost of debt
financing, we consider alternative metrics of audit committee independence and size. Column 2
presents the results of using an alternative measure of audit committee independence (fully
independent), suggesting that firms with greater audit committee independence experience a lower
cost of debt financing (about 14.7 basis points for fully independent committees). Column 3 gives
the results of using an alternative proxy for audit committee size (binary variables to denote large
and small committees). Again the results indicate a significant and negative relation between
corporate yield spreads and the board size.14 Overall, this analysis indicates that audit committee
structure influences the cost of debt financing, consistent with the prediction that audit committee
monitoring of the financial accounting process is important to creditors.
D. Board Member Characteristics
Column 2 of Table 4 provides the results of regressing board age, board tenure, and officer
and director holdings on yield spread. Board age is the average age of all directors and proxies for
business experience. Board tenure is the average number of years that directors serve on the board
and proxies for the ability of managers to influence director opinion - with longer director tenure,
managers potentially capture director decision-making. The officers and director equity ownership
variables measure directors’ immediate economic interests in the firm. The coefficient estimates for
board age, and officer and director holdings are not statistically significant, suggesting that directors’
14 We also repeat the tests using additional measures of audit committee characteristics (binary variable for
independent dominated audit committees and the ratio of audit committee size to firm size) and find similar results
to those reported in Table 4.

24
business experience and immediate economic interests have little discernible impact on the cost of
debt financing. However, board tenure is positive and significant; indicating that as director tenure
increases, managers are potentially more able to influence or sway board opinion. The magnitude of
the coefficient estimate is relatively small (when compared to the coefficients on board
independence and size), suggesting that an increase in average board tenure from seven to eight
years increases debt costs by 2.5 basis points.
We also examine whether specific occupational characteristics of independent directors are
associated with different costs of debt. Executives of other companies may have different incentives
to monitor (or not to monitor) compared to other board members or have incentives to vote with
management in the hopes of justifying actions at their own firms. We classify independent directors
into four groups: independent directors who are executives from other firms, retired executives
from other firms, academics, and other. Other includes clergy, government officials, and charity and
community workers. Column 2 of Table 4 provides the regression results. Interestingly, we find
that each director type is associated with a lower cost of debt financing. From a bondholder’s
perspective, there appears to be little difference in monitoring among academics, executives, retired
executives, or other job categorizations. In general, our analysis suggests that bondholders’ primary
concern is independent director monitoring rather than the specific expertise of the directors.
[Insert Table 4 about here]
Finally, we directly examine the impact of having financial experts on the audit committee
and the frequency of audit committee meetings on corporate yield spreads. We denote CFOs,
financial consultants, investment bankers, investment managers, bankers, auditors, and CEOs of
financial companies as financial experts. Because academics may or may not be considered financial
experts we measure them separately (similar results if combined). Column 3 of Table 4 shows the
results of including financial experts in the analysis. We find little evidence to indicate that financial

25
experts are an important consideration to creditors in our sample.15 However, in regards to audit
committee meeting frequency, the evidence suggests that audit committee activity is quite important
to creditors. The coefficient estimate (Table 4, Column 4) on the natural log of audit committee
meetings is –25.2 with a t-statistic of -3.9, indicating that firms with more audit committee meetings
are associated with a significantly lower cost of debt financing. Thus, comparing a firm with 4
committee meetings relative to a firm with only 3 meetings, suggests a reduction in debt costs of 7.3
basis points. These results are consistent with the hypothesis of the 1999 Blue Ribbon Report that
more actively involved audit committees are associated with greater financial statement reliability.16
IV. Endogeneity, First Differences, and Other Alternative Specifications
In this section, we examine several alternative specifications and models to assess the validity
of our results. Specifically, we examine the nature of causality in our tests, consider the incremental
impact of audit committees on the cost of debt, and investigate the relation between changes in
board structure and changes in corporate debt yield spreads.
A. Endogeneity
Our analysis potentially suffers from an endogeneity problem. Specifically, the issue is
whether board structure improves managerial monitoring thereby leading to lower yields or instead,
whether bond yields influence board structure in some way. We use two approaches to address this
15 Passage of the Sarbanes-Oxley Act bans firms from hiring former-audit firm personnel in the positions of CEO,
CFO, or controller within one year of the leaving the audit firm. Although the Act does not extend to the board
of directors, we investigate whether former audit-firm employees sitting on the firm’s audit committee influence
debt costs to ascertain the degree of the committee’s independence. Of the 824 directors serving as audit
committee members in our sample during 1993, we find only one instance (American Stores) of a former audit-
firm employee sitting on a firm’s audit committee. Interestingly, for the same set of 824 audit committee
members, we find only seven directors that worked for any audit firm. The implication is that former auditors
rarely appear on the audit committees of S&P 500 firms.
16 Another potential issue related to disclosure quality is auditor choice (Big 4/5) and whether the firm receives a
clean audit opinion (Mansi et al 2003). Within our sample of S&P 500 firms, we find that all firms employed a Big
5 auditor and received an unqualified audit opinion or an unqualified audit opinion with explanation. The
distinction between unqualified and unqualified with explanation is unrelated to the cost of debt.

26
concern. First, to the extent that board structure could be a function of bond yields, we use two-
stage least squares, instrumental variable (2SLS-IV) regressions to estimate the relation between
board structure and bond yields. Specifically, we use instrumental variables for board independence
(the natural log of total assets, fractional blockholdings, board tenure, and diversification) and board
size (the natural log of total assets, firm age, board tenure, and diversification). The results of this
analysis are reported in column 1 of Table 5 and are consistent with the results in Table 2. That is,
the cost of debt is negatively (significant at the five percent level) related to both board
independence and board size.
[Insert Table 5 about here]
Second, following Klein (1998) we also control for simultaneity by incorporating the yield
spread from the prior period (first lag) into the regression. These results are reported in column 2 of
Table 5. Consistent with the previous results, we find that board independence and board size are
both related to a lower cost of debt financing. We also examined additional control variables, such as
intangible assets, CEO pay-mix, CEO on the nominating committee, and corporate diversification
in Equation (4). The results of adding these control variables confirm our initial findings that board
structure is associated with a lower cost of debt.
B. Changes in Board Independence and Size on Changes in Yield Spread
Our primary research focus concerns the cross-sectional relation between board monitoring
of the financial accounting process and the cost of debt financing. However, for robustness, we also
consider the relation between changes in board structure and changes in yield spreads. Specifically,
we regress changes in yield spread on changes in board size and changes in board independence,
using the following specification. That is
∆Spreadi,t = B0 + B1 ( ∆Board Independencei,t ) + B2 (∆Board Sizei,t ) + ∆ Control Variables + e (5)

27
Where ∆ Spread is the change in the yield spread from t-1 to t, ∆ Board Independence is the change in
the fraction of independent to total directors from t-1 to t, ∆ Board Size is the change in the number
of board members from t-1 to t, and ∆ Control Variables represents changes in our control variables
for both firm specific and security specific attributes (change in firm size, credit ratings, leverage,
bond age, volatility, and duration) again from t-1 to t. For our full sample, we have 339 changes in
board size and board independence. With respect to board size, the median (mean) absolute change
is one (1.71) member(s). Similarly, the median (mean) absolute change in independence is 6.3 (7.9)
percent.
[Insert Table 6 about here]
The results of the first-difference regressions are reported in Table 6. Consistent with our
levels tests, we find that changes in yield spreads are negatively associated with changes in board size
and independence. Specifically, we find that increasing board size by one additional member, on
average, is associated with about a 13 basis point reduction in the cost of debt financing (median
change in board size is one member). Similarly, increasing the fraction of independent board
members by 10 percent (e.g., from 30 percent to 40 percent board independence) is associated with
about an 11.5 basis point reduction in the cost of debt. In terms of the control variables, the results
are consistent with our prior tests.
C. Incremental Impact of Audit Committees
Although our results are consistent with the hypothesis that board structure influences the
financial accounting process, another potential explanation focuses on the notion that board
structure improves firm performance and thereby reduces the cost of debt. While prior research
(see Hermalin and Weisbach (2003)) provides little evidence of independent boards improving firm
performance, we attempt to disentangle these competing explanations by examining whether audit-
committee characteristics exhibit additional explanatory power over full-board attributes. If the
direct link between boards and the financial accounting process – audit committees – continues to

28
exhibit a negative relation to debt costs even after controlling for full board attributes, then this
provides additional evidence that the relation between debt costs and boards/audit committees is
the financial accounting process rather than firm performance. We use two approaches to examine
the incremental explanatory power of audit committees relative to the full board.17 First, we repeat
the analysis in Table 3 but include variables for both board and audit committee characteristics in
the same regression. Since this test is motivated by the strong correlation between audit and board
structure variables (nearly 0.60 in Table 1, Panel C) it suggests that multicollinearity is a concern.
Consequently, we orthogonalize the board and audit committee variables in this test. Column 4 of
Table 3 shows the results and indicates that audit committee characteristics continue to exhibit a
negative relation to debt costs even after including full-board characteristics.
In a second approach to assess whether audit committees exhibit additional explanatory
power over full boards, we investigate those firms where insiders control a substantial portion of all
board seats and then examine whether audit committee characteristics continue to exhibit a negative
association with debt costs. For this test, we examine firms in the top quintile of insider control of
the board of directors. Full-board independence for this subsample is 0.30 and audit committee
independence is 0.47, with a correlation between the two variables of 0.37. The regression results
for this subsample are shown in Column 5 of Table 3 and indicate a negative relation between debt
costs and audit committee independence. Thus, even with a non-independent full board, we
continue to find an inverse relation between audit committees and the cost of debt financing,
suggesting that the financial accounting process (rather than improved firm performance) is the
linkage between board structure and debt costs.
17 Another test is to directly examine whether board structure affects future firm performance. After including a
one-year lead for firm performance, we continue to observe a negative relation between board (audit committee)
structure and the cost of debt.

29
D. Additional Robustness Testing
An assumption of our analysis is that the specifications and proxies adequately measure the
appropriate attributes. As such, we further examine non-linearities in board structure and credit
ratings, outliers and influential observations, and macro-economic interest rate shocks. We find that
our results are also robust to various alternative specifications.
First, we included the square of board size to examine the issue of non-linearities in our
board size analysis. We find the square term (of board size) is insignificantly different from zero,
while board size continues to be significant. While our results unambiguously indicate that larger
boards are better, a potential caveat is that the largest board in our sample consists of 24 directors.
To further examine the issue of non-linearities in board independence and board size we also use
piecewise regressions. Specifically, we repeat the analysis using binary variables to indicate the
second through fifth quintiles for both board independence and board size. The results of this
analysis again indicate that large, independent boards are associated with a lower cost of debt
financing.
We also allow for a non-linear relation between bond yield spreads and credit ratings.
Specifically, we use a piecewise linear regression with 22 breakpoints and 21 dummy variables to
proxy for credit ratings. The results of these regressions are consistent with the primary
specification. Although our analysis includes leverage, stock volatility, firm size, and credit ratings
we consider additional alternative approaches to capturing default risk such as including cash flow
volatility, the square of credit ratings, the square of leverage, and the coverage ratio, as well as
creating composite bankruptcy prediction scores (e.g. Ohlson’s 1980 measure) or modeling credit
ratings. Using these procedures we find similar results to those reported.
In further tests, we consider an alternative method of measuring bond liquidity. Our primary
specification utilizes age as a proxy for liquidity. However, a non-linear specification may be more
appropriate, as bond liquidity decays exponentially with age (Beim (1992)). Therefore, we replace
age with the natural log of age. We also measure liquidity using a binary variable to denote bonds

30
less than three years old, as it has been found that bonds lose a third of their liquidity in the first
three years of their issuance (Beim (1992)). Regressions using these two alternate measures are
consistent with our primary regression, suggesting that board structure is associated with lower cost
of debt.
To test the sensitivity of our analysis to serial correlation and to the impacts of outliers and
influential observations, we consider several alternative techniques. First, we repeat the tests using a
random effects model, year-by-year regressions, and using the year-by-year results in the Fama-
MacBeth procedure. All three approaches lead to similar inferences (see Appendix B for the
random effects and Fama-MacBeth results). To identify observations that are outliers and/or
influential observations, we use the use the R-Student statistic and the DFFITS statistic. These tests
examine a sample to determine if any observations have a dramatic effect on the fitted least-squares
function. The results were similar to those reported in Tables 2 and 3 and do not change
substantively when truncated for outliers at the largest one percent, three percent, and five percent
levels at each tail of the distribution for each variable in the model. In addition, both least absolute
deviation and least median deviation regressions give similar results (see Appendix B).
Finally, because the relation between board structure and yield spread may not be stable over
time, especially during the 1994 period when the Federal Reserve raised interest rates seven times in
succession, we use a dummy variable to denote the pre-1995 period along with period subset
regressions. The results corroborate those reported. Although the credit ratings are debt specific
measures, we also control for debt structure by including the proportion of senior debt to total debt
in our analysis and with subset regressions based on duration. Once again we find a negative
relation between board structure (audit committee structure) and the cost of debt financing.
V. Conclusion
Boards of directors are responsible for monitoring, evaluating, and disciplining firm
management. Perhaps one of the most important responsibilities of the board from a creditor’s

31
perspective is oversight of financial reporting. Because debt holders rely on accounting based
covenants in lending agreements, creditors may have concerns with board and audit committee
monitoring of the financial accounting process. Consistent with this idea, we find that board and
audit committee independence are associated with significantly lower debt financing costs.
Our analysis also indicates that board and audit committee size are inversely related to the
cost of debt, suggesting that larger boards and audit committees provide greater monitoring of the
financial accounting process. The results are statistically and economically significant, and robust to
a variety of board size and board independence measures. Additional tests (non-linear specifications
and changes regression) provide additional support for the hypothesis the board structure is relevant
to debtholders.
Finally, we investigate the relation between independent director attributes and debt yields.
Our results indicate that employment characteristics (executive, retired, academic, etc.), while all
significantly related to lower debt costs, are not statistically different from one another. In addition,
the presence of financial experts on the audit committee is not related to the costs of debt in our
sample. The investigation also suggests that director equity-ownership does not exhibit a significant
relation to the cost of debt financing. We also document that the frequency of audit committee
meetings is associated with lower debt costs. We interpret this to suggest that active monitoring by
independent directors of the financial accounting process is quite important to creditors.
In conclusion, we find compelling evidence that board and audit committee monitoring
substantially impact the cost of debt financing. The results indicate that firms with large
independent boards and audit committees are associated with a lower cost of debt financing,
suggesting that boards of directors are an important element of the financial accounting process.

32
References
Adams, R., and H. Mehran, Board Structure and Banking Firm Performance, working paper, Federal Reserve
Bank of New York
Anderson, R.., S. Mansi., and D. Reeb, 2003, Founding Family Ownership and the Agency Costs of Debt,
Journal of Financial Economics 68, 263-287.
Beasley, M., 1996, An empirical analysis of the relation between the board of director composition and
financial statement fraud, Accounting Review 71, 443-465.
Begley, J., 1990, Debt Covenants and accounting choice, Journal of Accounting and Economics 12, 125-139.
Beim, D., 1992, Estimating bond liquidity, Working Paper, Columbia University.
Betker, B., 1995, Management’s incentives, equity’s bargaining power, and deviations from absolute priority in
chapter 11 bankruptcy, Journal of Business 68, 161-183.
Bhojraj, S. and P. Sengupta, 2002, Effects of corporate governance on bond ratings and yields: The role of
institutional investors and outside directors, Journal of Business, Forthcoming.
Branch, B., 2000, Fiduciary duty: Shareholders versus creditors, Financial Practice and Education, 10,8-13.
Brickley, J., J. Coles, and R. Terry, 1994, Outside directors and the adoption of poison pills, Journal of Financial
Economics 35, 371-390.
Byrd, J. and K. Hickman, 1992, Do outside directors monitor managers? Evidence from tender offer bids,
Journal of Financial Economics 32, 195-222
Carcello, J. and T. Neal, 2000, Audit committee composition and auditor reporting, Accounting Review 75,453-
467.
Chaganti, R. S., V. Mahajan, and S. Sharma, 1985, Corporate board size, composition, and corporate failures
in retailing industry, Journal of Management Studies 22, 400-417
Daley, L. and R. Vigeland, 1983, The effects of debt covenants and political costs on the choice of accounting
methods: The case of accounting for R&D costs, Journal of Accounting and Economics 5, 195-211.
Dechow, P., R. Sloan, A. Sweeney, 1996, Causes and consequences of earnings manipulation: An analysis of
firms subject to enforcement actions by the SEC, Contemporary Accounting Research 13, 1-36.
DeFond, M. and J. Jiambalvo, 1994, Debt covenant violations and manipulation of accruals, Journal of
Accounting and Economics 17, 145-176.
Dichev, I. and D. Skinner, 2002, Large-sample evidence on the debt covenant hypothesis, Journal of Accounting
Research, Forthcoming.
Duke, J. and H. Hunt, 1990, An empirical examination of debt covenant restrictions and accounting related
debt proxies, Journal of Accounting and Economics 12, 45-63.
Fama, E. and M. Jensen, 1983, Separation of ownership and control, Journal of Law and Economics 26, 301-326.
Hermalin, B. and M. Weisbach, 2003, Board of directors as an endogenously determined institution: A Survey
of the economic literature, Economic Policy Review, Forthcoming.
Jensen, M., 1993, Presidential address: The modern industrial revolution, exit and the failure of internal
control systems, Journal of Finance 48, 831-880.

33
Jensen M. and W. Meckling, 1976, Theory of the firm: Managerial behavior, agency costs and ownership
structure, Journal of Financial Economics 3, 305-360.
Kalay, A., 1982, Stockholder-bondholder conflict and dividend constraints, Journal of Financial Economics 10,
211-233.
Klein, A. 1998, Firm performance and board committee structures, Journal of Law and Economics 41, 275-303.
Klein, A., 2002a, Audit committee, board of director characteristics, and earnings management, Journal of
Accounting and Economics 33, 375-400.
Klein, A., 2002b, Economic determinants of Audit Committee Independence, Accounting Review 77, 435-452.
Leftwich, R., 1983, Accounting Information in Private Markets: Evidence from private lending agreements,
Accounting Review 58, 23-43.
Lipton, L. and J. Lorsch, 1992, A modest proposal for improved corporate governance, The Business Lawyer
48, 59-77.
Morrissey, J., 2000, Corporate Responsibility and the Audit Committee, Deputy Chief Accountant SEC, Remarks at
the General Audit Management Conference New Orleans, LA (available at http://www.sec.gov).
Monks, R. and N. Minow, 1995, Corporate Governance, Blackwell Publishing, New York, NY.
Nelson C. and A. Siegel, 1987. Parsimonious modeling of yield curves, Journal of Business 6, 473-489.
Ohlson, J., 1980, Financial ratios and the probabilistic prediction of bankruptcy, Journal of Accounting Research
18, 109-131.
Pincus, K., M. Rusbarsky, and J. Wong, Voluntary formation of corporate audit committees among
NASDAQ firms, Journal of Accounting and Public Policy 8, 239-265.
Perumpral, S., D. Davidson, and N. Sen, 1999, Event risk covenants and shareholder wealth: Ethical
implications of the “Poison Put” provisions in bonds, Journal of Business Ethics 22, 119-132
Reeb, D., S. Mansi, and J. Allee, 2001, Firm internationalization and the cost of debt financing: Evidence
from non-provisional publicly traded debt, Journal of Financial and Quantitative Analysis, 36, 395-414.
Sarig, O. and Warga, A., 1989, Bond price data and bond market liquidity, Journal of Financial and Quantitative
Analysis 24, 367-378.
Shleifer, A. and R. Vishny, 1997, A survey of corporate governance, Journal of Finance 52, 737-783.
Smith, C., 1993, A perspective on accounting-based debt covenant violations, The Accounting Review 68, 289-
303.
Smith, C. and J. Warner, 1979, On financial contracting: An analysis of bond covenants, Journal of Financial
Economics 7, 117-161.
Yermack, D., 1996, Higher market valuation companies with a small board of directors, Journal of Financial
Economics 40, 185-212.

34
Table 1
Sample Description
Panel A: Descriptive Statistics for Variable Measures
This table provides summary statistics for the data employed in our analysis. The data set is comprised of
1,052 firm-year observations for the period 1993 through 1998. The yield spread (Spread) is the difference
between the yield to maturity on the firm’s debt less its duration equivalent Treasury rate. Independent
directors represents the number of independents on the board of directors; Board Independence is the ratio
of independent directors to total directors; Inside directors represent the number of insiders on the board of
directors; Board (Audit) size is the number of directors on the board (audit committee); Board tenure is the
average length of time a director has been on the board; Board age is the average age of the board of
directors. The variables academic, retired, executive, and other represent the last occupation of each
independent director on the board.
Audit Committee Independence is the fraction of independents on audit committee, Audit
Committee Size is the number of directors on the audit committee, Audit Committee meetings is the number
meetings held each year, Financial Experts on Audit Committee is someone who is a CFO, investment
banker, financial consultant or CEO of a financial firm.
Firm size is the natural log of total assets; Leverage is the ratio of long-term debt to total capital;
Volatility is the stock return variability for each firm for the past five years; Duration is the Macualy duration
and proxies for effective maturity; Credit ratings is the average of Moodys and S&P ratings, and Age is the
difference between the bond issue date and quote date in years. Firm performance is the ratio of the firms
cash flows divided by total assets.
Variables Mean Median Std. Dev Maximum Minimum
Spread (bps) 135.961 102.701 109.454 1067.866 2.239
Board Structure Variables
Independent Directors (Number of) 6.903 7.000 2.417 16.000 0.000
Board Independence 0.573 0.583 0.173 0.929 0.000
Number Inside Directors 3.295 3.000 1.877 12.000 1.000
Board Size 12.056 12.000 2.524 24.000 6.000
Tenure on Board 9.177 8.850 3.080 24.860 0.330
Age of Directors 60.269 60.270 2.651 67.780 49.570
Independent Board Member Occupation
Number Academic 0.918 1.000 0.988 6.000 0.000
Number Retired 1.742 2.000 1.440 7.000 0.000
Number Executive 3.851 4.000 1.953 10.000 0.000
Number Other 0.365 0.000 0.645 6.000 0.000
Audit Committee Variables
Audit Committee Independence 0.697 0.750 0.263 1.000 0.000
Audit Committee Size 4.475 4.000 1.364 12.000 1.000
Audit Committee Meetings 3.490 3.000 1.419 14.000 1.000
Financial Expert on Audit Committee 0.875 1.000 0.837 4.000 0.000
Firm & Security Specific Variables
Firm Size 8.877 8.705 1.28 12.782 4.398
Firm Leverage 0.219 0.203 0.133 0.943 0.000
Duration (years) 6.287 6.276 2.454 13.621 0.083
Credit Ratings 15.979 16.000 3.209 22.143 1.000
Bond Age (years) 3.917 3.511 2.689 25.655 0.033
Volatility 0.259 0.236 0.078 0.173 0.128
Perform 0.138 0.127 0.075 0.786 -0.119

35
Table 1 - Continued
Sample Statistics
Panel B: Industry Data
This panel gives the number of firm-year observations for each industry group in the sample.
SIC
Code
Titles of Industries
Number of
Firm-Year
Observations
1 Mining 41
2 Construction 359
3 Manufacturing 322
4 Transportation 62
5 Wholesale Trade 138
6 Retail Trade 82
7 Agricultural Services 42
8 Forestry 6
Panel C: Correlation Matrix
This table provides correlation data for the yield spread (spread), board structure (independent directors,
board size, board tenure, and board age) variables, and firm size. Spread is the difference between the
weighted average yield to maturity on the firm’s debt less its duration equivalent Treasury rate. Fraction
Independent is the ratio of independent directors to total directors; Board Size is the number of directors on
the board; Audit Independence is the fraction of independent directors on the committee; Audit Comm. Size
is the number of directors on the committee, Board Tenure is the average length of time a director has been
on the board; Board Age is the average age of board of directors; Firm Size is the natural log of total assets.
Yield
Spread
Board
Ind.
Board
Size
Audit
Ind.
Audit
Size
Firm
Size
Board
Tenure
Yield Spread 1.000
Board Independence -0.119 1.000
Board Size -0.261 -0.001 1.000
Audit Independence 0.014 0.595 0.044 1.000
Audit Comm. Size -0.130 0.119 0.389 0.063 1.000
Firm Size -0.355 0.149 0.372 0.061 0.244 1.000
Board Tenure 0.039 -0.339 0.184 0.063 -0.067 -0.155 1.000
Board Age -0.003 -0.037 0.062 -0.238 0.081 0.051 0.363

36
Table 2
Yield Spread and Board structure (n= 1,052)
This table provides the estimated coefficients from regressing yield spreads on board structure variables (the fraction of
independent directors and the natural log of board size) and various firm and security specific controls.
Spreadi,t = A0 + A1 (Board Independencei,t ) + A2 (Log Board Sizei,t ) + A3 (Firm Sizei,t ) + A4 (Leveragei,t ) + A5 (Durationi,t) + A6 (Bond Agei,t )
+ A7 (Ratingi,t ) + A8 (Blocki,t ) + A8 (NLCrediti,t ) + A9 (Volatilityi,t ) + A10 (Performi,t ) + A11 (Time_Dumi,t ) + A12 (Ind_Dumi,t ) +
ε
i,t
Column 2 reports the results for the primary specification. Board Independence is the ratio of independent directors to
total directors. Board size is the natural log of the number of directors on the board. Firm size is the natural log of total
capital. Leverage is the ratio of long-term debt to total capital. Duration is the Macualy duration. Age is the difference
between the bond issue date and quote date in years. Rating is the security specific adjusted credit rating. Volatility is the
stock return variability for each firm for the past 60 months. Firm performance is the ratio of the firms cash flows
divided by total assets. The time period and industry dummy variable results are not reported. In columns 3 through 6,
we use different measures of independent director influence and board size variables for robustness.
Variables Sign Dependent Variable = Yield Spread
(1) (2) (3) (4) (5) (6)
Intercept
374.679*
(6.161)
239.921*
(5.565)
222.363*
(5.264)
288.032*
(5.709)
86.029**
(2.324)
Board Independence ? -67.180*
(-4.019)
-64.511*
(-3.979)
-50.432*
(-3.163)
Independent Dominated
? -19.185*
(-3.575)
Number of Independents
? -5.457*
(-4.091)
Board/Firm Size Ratio
? -69.289*
(-6.565)
-53.053*
(-5.221)
-77.968*
(-6.531)
Log (Board Size)
? -122.241*
(-7.054)
Big Boards
? -32.665*
(-5.372)
Small Boards
? 23.123*
(2.686)
Firm Size
- -9.899*
(-4.172)
-21.777*
(-8.531)
-18.695*
(-7.945)
-22.574*
(-8.547)
-13.138*
(-5.291)
Leverage
+ 0.405
(1.566)
0.574**
(2.381)
0.461***
(1.818)
0.444***
(1.723)
0.569**
(2.265)
Duration
+ 3.923*
(4.811)
4.121*
(5.048)
4.041*
(4.931)
4.029*
(4.945)
4.301*
(5.303)
Bond Age + 7.849*
(6.534)
8.048*
(6.838)
7.945*
(6.731)
7.914*
(6.678)
8.012*
(6.764)
Rating
- -8.185*
(-6.590)
-6.459*
(-6.429)
-7.451*
(-6.445)
-7.499*
(-6.400)
-5.095*
(-5.426)
Non-Linear Credit + 121.052*
(9.123)
128.970*
(9.871)
125.218*
(9.431)
124.724*
(9.321)
136.643*
(10.051)
Block Holdings + 1.961
(0.133)
9.063
(0.607)
4.961
(0.336)
3.736
(0.253)
3.081
(0.209)
Volatility + 203.461*
(4.014)
227.880*
(4.606)
220.212*
(4.378)
220.856*
(4.381)
238.361*
(4.902)
Perform
- -71.878**
(-2.442)
-89.507*
(-3.015)
80.802*
(-2.728)
-79.403*
(-2.662)
-99.852*
(-3.355)
Adjusted R Squared 0.643 0.637 0.639 0.639 0.635
*, **, *** denote significance at the one percent, five percent, and ten percent levels, respectively. The t-values given in
parenthesis below each estimate are corrected for heteroskedasticity.

37
Table 3
Yield Spread and Audit Committee structure (n= 1,052)
This table provides the estimated coefficients from regressing yield spreads on audit committee structure variables (the
fraction of independent directors and the natural log of audit committee size) and various controls.
Spreadi,t = A0 + A1 (Audit Independencei,t ) + A2 ( Audit Sizei,t ) + Control Variables +
ε
i,t
Column 1 reports the results for the primary specification. Audit Independence is the ratio of independent directors to
total directors on the audit committee. Audit size is the natural log of the number of directors on the audit committee.
In columns 2 and 3 we use alternative measures of audit committee independence and size. Columns 4 adds board
independence and board size to examine the incremental explanatory power of audit committee characteristics over
board characteristics. Because board and audit committee characteristics are correlated (motivating this test) we
orthogonalize the board and audit committee variables in this particular regression. In column 5, we take a different
approach and focus on those boards that are in the top quintile of corporate insider control (repeat analysis using only
top quintile). Both approaches provide additional evidence on whether the channel through which board structure
affects the cost of debt is improved financial reporting and disclosure. The control variables are described in Table 2.
Variables Dependent Variable = Yield Spread
Audit Com. Ind . & Size Incremental Impacts
(1) (2) (3) (4) (5)
Intercept
181.099*
(3.956)
154.161*
(3.753)
90.589**
(2.511)
190.511*
(4.398)
385.390**
(4.025)
Audit Independence -33.168*
(-2.759)
-21.534**
(-1.967)
-37.583*
(-3.149)
-55.258**
(-2.459)
Fully Independent
-14.675**
(-2.381)
Log (Audit Size)
-57.899*
(-4.155)
-53.492*
(-4.134)
-13.183***
(-1.723)
-128.327*
(-2.293)
Board Independence
-9.048*
(-4.563)
Log (Board Size)
-10.496*
(-6.486)
Big Committee
-34.478*
(-3.584)
Small Committee
18.504**
(2.521)
Firm Size
-13.715*
(-5.609)
-14.278*
(-5.976)
-15.301*
(-6.576)
-10.386*
(-4.429)
-10.282
(-1.632)
Leverage
0.503***
(1.915)
0.589**
(2.337)
0.660*
(2.694)
0.4574***
(1.796)
0.191
(0.364)
Duration
4.343*
(5.284)
4.439*
(5.369)
4.349*
(5.234)
4.162*
(5.217)
4.013
(1.581)
Bond Age 7.904*
(6.567)
8.171*
(6.831)
8.020*
(6.702)
7.748*
(6.460)
8.982**
(2.331)
Rating
-6.244*
(-5.111)
-5.170*
(-5.334)
-4.539*
(-4.504)
-8.198*
(-6.456)
-9.919**
(-2.397)
Non-Linear Credit 129.224*
(9.657)
132.764*
(10.095)
137.195*
(10.557)
121.258*
(9.273)
101.241*
(2.808)
Block Holdings 7.796
(0.525)
9.261
(0.606)
8.554
(0.569)
-0.851
(-0.057)
11.933
(0.357)
Volatility 237.363*
(4.576)
246.298*
(4.928)
254.768*
(5.033)
201.746*
(4.043)
33.579
(0.264)
Perform
-86.749*
(-2.851)
-97.241*
(-3.253)
-103.092*
(-3.448)
-74.330*
(-2.508)
-114.747
(-1.330)
Adjusted R Squared 0.636 0.635 0.633 0.646 0.667
*, **, *** denote significance at the one percent, five percent, and ten percent levels, respectively. The t-values given in
parenthesis below each estimate are corrected for heteroskedasticity.

38
Table 4
Yield Spread and Independent Director Attributes (n= 1,052)
This table provides the estimated coefficients from regressing corporate yield spreads on board and audit
committee director attributes and various control variables. Columns 1 & 2 focus on director attributes of
the whole board. Specifically, in column 1 we examine the impact of board age, board tenure, and officer and
director holdings on the cost of debt financing. In column 2 we test for differences based on independent
director occupation on corporate yield spreads. Columns 3 & 4 focus on audit committee director attributes.
In column 3 we examine the impact of Financial Experts and if a committee member was a former employee
of on the firm’s auditor. In column 4, we focus on the natural log of the number of meetings held by the
audit committee (Audit Committee Meetings as a measure of audit committee activity.
Firm size is the natural log of total capital. Leverage is the ratio of long-term debt to total capital.
Duration is the Maculay duration and proxies for effective maturity. Age is the difference between the bond
issue date and quote date in years. Rating is the board structure adjusted credit rating based on Moodys and
S&P ratings. Volatility is the stock return variability for each firm for the past 60 months. Firm performance
is the ratio of the firms cash flows divided by total assets. We also include time period and industry dummy
variables. The results for the control variables are similar to prior tables and are not reported.
Dependent Variable = Yield Spread
Variable
Board of Director Attributes Audit Committee Attributes
(1) (2) (3) (4)
Intercept
306.463*
(3.963)
226.595*
(5.325)
201.474*
(3.981)
197.042*
(4.575)
Board Independence
-64.482*
(-4.025)
Board Size
-83.903*
(-6.750)
-58.672*
(-5.781)
Audit Com Independence
-33.712*
(-2.624)
-22.192*
(-2.601)
Audit Com Size
-55.681*
(-3.922)
-66.053*
(-5.733)
Board Age
-0.888
(-0.915)
Board Tenure
2.465**
(2.336)
Officer & Dir Holdings
-31.442
(-1.529)
Academic
-7.324*
(-3.223)
Retired
-4.224**
(-2.489)
Executive
-5.389*
(-3.632)
Other
-6.401***
(-1.673)
Financial Expert on Audit
-3.436
(-0.751)
Academic On Audit
-1.045
(-0.222)
Audit Committee Meetings
-25.231*
(-3.857)
Adjusted R Square 0.639 0.629 0.637 .651
*, **, *** denote significance at the one percent, five percent, and ten percent levels, respectively. The t-
values given in parenthesis below each estimate are corrected for heteroskedasticity.

39
Table 5
Yield Spread and Board Structure: Endogeneity
This table provides the estimated coefficients from regressing yield spreads (Spread) on board structure
variables (controlling for endogeneity) and various firm and security specific controls.
Spreadi,t = B0 + B1 ( Board Independencei,t ) + B2 ( Log Board Sizei,t ) + Control Variables + e
In column 1 we use instrumental variables for both board independence (natural log of total assets,
blockholdings, board tenure, and diversification) and board size (natural log of total assets, firm age, board
tenure, and diversification). In column 2, following Klein (1998), we incorporate the lagged dependent
variable to control for endogeneity. In both regressions we use several control variables. Firm size is the
natural log of total capital. Leverage is the ratio of long-term debt to total capital. Duration is the Macualy
duration and proxies for effective maturity. Age is the difference between the bond issue date and quote date
in years. Rating is the board structure adjusted credit rating based on Moodys and S&P ratings. Volatility is
the stock return variability for each firm for the past 60 months. Firm performance is the ratio of the firms
cash flows divided by total assets. The time period and industry dummy variable results are not reported.
Variables Dependent Variable = Yield Spread
(1) (2)
Intercept
772.193*
(3,559)
216.492*
(3.604)
IV for Board Independence -154.368**
(-2.207)
IV for Log of Board Size
-300.368**
(-2.955)
Fraction Independent
-47.481*
(-3.168)
Log (Board Size)
-69.482*
(-3.828)
Firm Size
2.844
(0.448)
-5.267**
(-2.573)
Leverage
0.836*
(3.464)
0.300
(1.424)
Duration
4.282*
(5.115)
3.165*
(3.731)
Bond Age 7.902*
(6.237)
5.310*
(4.211)
Rating
-1.342**
(-2.077)
-4.323*
(-3.528)
Non-Linear Credit 141.665*
(10.264)
57.875*
(3.763)
Block Holdings -68.728**
(-2.383)
-0.239
(-0.019)
Volatility 246.666*
(4.529)
114.535*
(2.529)
Perform
-159.128*
(-5.491)
-38.663
(-1.562)
Lag of Spread 0.414*
(5.627)
Adjusted R Squared 0.616 0.710
*, **, *** denote significance at the one percent, five percent, and ten percent levels, respectively.
The t-values given in parenthesis below each estimate are corrected for heteroskedasticity.

40
Table 6
Yield Spread and Board Structure: Changes Regression
This table provides the estimated coefficients from regressing the change in yield spreads (∆Spread) on the
change in board independence (∆Board Independence) and board size (∆Board Size) and control variables
(changes in).
∆Spreadi,t = B0 + B1 ( ∆Board Independencei,t ) + B2 (∆Board Sizei,t ) + ∆ Control Variables + e
Where ∆Spread is the change in the yield spread, ∆Board Independence is the change in the fraction of
independent to total directors, ∆Board Size is the change in the number of board members, and ∆Control
Variables represents changes our control variables (change in firm size, credit ratings, leverage, bond age,
volatility, and duration). Firm size is the natural log of total capital. Leverage is the ratio of long-term debt to
total capital. Duration is the Macualy duration and proxies for effective maturity. Age is the difference
between the bond issue date and quote date in years. Rating is the board structure adjusted credit rating
based on Moodys and S&P ratings. Volatility is the stock return variability for each firm for the past 60
months. Firm performance is the ratio of the firms cash flows divided by total assets.
Variables Change in Yield Spread
Intercept
-37.040*
(-5.294)
Change in Board Independencee -115.283*
(-2.672)
Change in Board Size
-13.051*
(-3.398)
Change in Firm Size
-47.028***
(-1.895)
Change in Leverage
0.845
(1.373)
Change in Duration
4.606
(1.323)
Change in Bond Age 9.062*
(3.422)
Change in Rating
-15.797*
(-4.270)
Change in Volatility 116.064
(1.091)
Change in Performance
-38.943
(-0.652)
Adjusted R Squared 0.225
*, ** denote significance at the one percent and five percent levels.

i
Appendix A
Bond Rating Conversions
This table provides the bond rating categories for both Moody's and S&P ratings. Bond ratings are computed
using a conversion process in which AAA rated bonds are assigned a value of 22 and D rated bonds receive a
value of 1.
Conversion
Number
Moody's
Ratings
S&P
Ratings
22 Aaa AAA
21 Aa1 AA+
20 Aa2 AA
19 Aa3 AA-
18 A1 A+
17 A2 A
16 A3 A-
15 Baa1 BBB+
14 Baa2 BBB
13 Baa3 BBB-
12 Ba1 BB+
11 Ba2 BB
10 Ba3 BB-
9 B1 B+
8 B2 B
7 B3 B-
6 Caa1 CCC+
5 Caa2 CCC
4 Caa3 CCC-
3 Ca CC
2 C C
1 D D

ii
APPENDIX B
Serial Correlation, Outliers, and Non-Normal Errors
This table provides the estimated coefficients from regressing yield spreads on board structure variables and various
firm and security specific controls.
Spreadi,t = A0 + A1 (Fraction Independenti,t ) + A2 (Board Size / Log of Total Assetsi,t ) + A3 (Firm Sizei,t ) + A4 (Leveragei,t ) + A5 (Durationi,t) +
A6 (Bond Agei,t ) + A7 (Ratingi,t ) + A8 (Blocki,t ) + A8 (NLCrediti,t ) + A9 (Volatilityi,t ) + A10 (Performi,t ) + A11 (Ind_Dumi,t ) +
ε
i,t
Alternative adjustments for non-spherical serial correlation are provided in columns 2 and 3 (random effects &
Fama-MacBeth Regressions). Column 4 gives the results after truncating outliers, while columns 5 and 6 give non-
parametric regression results (Least Absolute Deviation [LAD] regression & Least Median Regression [LMR]).
Variables Sign Dependent Variable = Yield Spread
ALTERTNATIVE
ADJUSTEMENTS -
SERIAL CORRELATION
OUTLIERS AND NON-NORMAL
ERRORS
(1)
(2)
RANDOM
EFFECTS
MODEL
(3)
FAMA-
MACBETH
MODEL
(4)
OUTLIERS
REMOVED
MODEL
(5)
LEAST
ABSOLUTE
DEVIATION
(6)
LEAST
MEDIAN
REGRESSION
Intercept
356.663*
(5.375)
467.672*
(4.450)
413.999*
(7.026)
352.565*
(10.957)
269.402*
(5.265)
Fraction Independent ? -66.708*
(-3.331)
-78.738*
(-2.937)
-70.923*
(-4.837)
-61.155*
(-6.852)
-77.144*
(-8.425)
Board Size
? -74.194
(-3.613)
-139.051*
(-8.010)
-127.42*
(-7.956)
-118.481*
(-11.979)
-132.516*
(-8.425)
Firm Size
- -21.053*
(-6.209)
-8.793**
(-2.230)
-12.235*
(-4.626)
-5.302*
(-4.119)
-9.035*
(-4.413)
Leverage
+ 0.649*
(2.605)
0.362
(1.222)
0.350
(1.346)
0.281**
(2.405)
0.987*
(5.319)
Duration
+ 2.116***
(1.793)
1.766*
(3.595)
4.409*
(5.359)
6.702*
(12.483)
12.945*
(15.162)
Bond Age + 7.412*
(8.194)
3.418*
(3.164)
8.715*
(7.985)
6.389*
(13.746)
1.165
(1.575)
Rating
- -6.429*
(-4.113)
-9.489*
(-6.558)
-8.612*
(-8.002)
-8.185*
(-11.581)
-8.081*
(-7.191)
Non-Linear Credit + 96.782*
(8.971)
112.671*
(4.128)
114.588*
(9.111)
133.184*
(24.151)
144.600*
(16.489)
Block Holdings - -40.974**
(-2.260)
12.715
(0.463))
-1.751
(-0.115)
9.339
(1.118))
-8.097
(-0.608)
Volatility + 257.400*
(6.102)
220.340*
(2.587)
202.847*
(3.942)
38.752***
(1.813)
106.247*
(3.125))
Perform
- -191.451*
(-5.392)
-59.515*
(-0.817)
-63.334**
(-2.152)
-16.529
(-0.884)
98.598*
(3.317)
Adjusted R Squared 0.571 - 0.665 0.628 0.263
*, **, *** denote significance at the one percent, five percent, and ten percent levels, respectively.