Journal of Corporate Finance 28 (2014) 116–134

Contents lists available at ScienceDirect

Journal of Corporate Finance

journal homepage: www.elsevier.com/locate/jcorpfin

Does the independence of independent directors matter?

Rafel Crespí-Cladera 1, Bartolomé Pascual-Fuster ⁎
Departament d'Economia de l'Empresa, Universitat de les Illes Balears, Cra. de Valldemossa km 7.5, 07122 Palma (Illes Balears), Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper analyzes the characteristics of firms that declare board directors as independents,
Received 26 September 2013 although the directors are not strictly independent, and examines the consequences in terms
Received in revised form 26 November 2013 of performance and corporate governance outcomes. Based on publicly available information,
Accepted 9 December 2013
eight criteria of “independence” used to examine a panel of Spanish listed firms classify 14.2%
Available online 14 December 2013
of the directors as strictly independent, whereas the firms classify 32.5% of the board as
independent directors. Firms with dispersed ownership structures misclassify directors more
JEL classification: frequently than do firms with large controlling owners. In terms of consequences, we find
G30
weak evidence of a negative relation between misclassification and a firm's future operating
G34
performance. However, no relation is found between independents' misclassification and
K22
several relevant outcomes of the primary delegated committees with monitoring roles: the
audit committee and the nomination and remuneration committee. There is no significance
with regard to the non-strictly independent measures explaining executive directors'
Keywords:
Board strict independence
compensation, CEO turnover, audit qualifications or earning management behavior.
Corporate governance © 2013 Elsevier B.V. All rights reserved.
Executive compensation
CEO turnover
Audit qualifications
Earnings management

1. Introduction

Independent directors are perceived as minimizing the potential opportunism of managers, or large controlling owners, in a
principal agent setting. In the context of Anglo-American dispersed ownership, independents are key members of the board in
avoiding managerial misappropriation. In the concentrated ownership setting of continental Europe, the role assigned to
independent directors is to limit the extraction of private benefits by the controlling large shareholders, who usually appoint the
remaining members of the board of directors. The underlying fundamental concept is that the monitoring activity of the
boardroom depends on the effectiveness of the independent members. This view, which is widely accepted in the academic world
(Adams and Ferreira, 2007; Adams et al., 2008; Bhagat and Black, 2002; Dyck and Zingales, 2004; Hermalin and Weisbach, 2003),
is in the spirit of regulations such as the OECD Principles of Corporate Governance of 2004, the Commission of the European
Communities Recommendation,2 the final corporate governance rules of the New York Stock Exchange of 2009, and nearly all
existing corporate governance codes or guidelines. Effectively, the European Commission Recommendation of 2005 asserts that
independent directors play a role in both dispersed ownership companies, in which the concern is how to make managers
accountable to shareholders, and in companies with large controlling shareholders that must account for the interests of minority
shareholders.

⁎ Corresponding author. Tel.: +34 971172652.

E-mail addresses: rafel.crespi@uib.es (R. Crespí-Cladera), tomeu.pascual@uib.es (B. Pascual-Fuster).
1
Tel.: +34 971171323.
2
The Commission of the European Communities Recommendation of 15 February 2005 on the role of non-executive or supervisory directors of listed
companies and on the committees of the supervisory board.

0929-1199/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jcorpfin.2013.12.009
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 117

In line with best practices codes and guidelines, the majority of companies report increasing proportions of independents
among their board members (Gordon, 2007; Linck et al., 2009). Rating agencies also account for the presence of a qualified
number of independent directors as an element in agency rating outputs, as Santella et al. (2006) indicate. One long-term
perspective regarding the changing trend of executives and independents in US boardrooms comes from Gordon (2007), who
indicates that from 1950 to 2005, the average percentage of independent directors has risen from 20% to levels above 70%.
Using publicly available information, our paper compares firms' declarations regarding independent directors with an
unbiased measure of strictly independent board members for a panel of Spanish listed firms. In addition to the contribution of the
descriptive information regarding the firms' deviation between declared and strictly independent directors, the paper also
addresses the following two main questions:

Question 1: What are the characteristics of firms that misclassify independent directors more frequently?
Question 2: What are the consequences of firms declaring directors as independents when these directors are not strictly
independent? This paper will examine firms' future operating performance and the outcomes of the main board committees in
terms of monitoring: the audit committee and the nomination and remuneration committee. Any negative impact on firms'
performance or on the corporate governance outcomes would mean that there is an agency problem that should be corrected
with law enforcement. In addition, the lack of effect of independent directors' misclassification on firms' performance would
indicate that the usual recommendations of large proportions of independents should be reexamined.
The previous empirical literature addresses the relevance and the role of independent board members, usually by considering
the number of directors declared as independent (e.g., Boone et al., 2007; Coles et al., 2008; Duchin et al., 2010; Kim et al., 2007;
Linck et al., 2008). No question regarding the quality of the directors' independence was examined in these papers. However,
several other papers address the qualitative aspect of the independence of the board of directors. Independent directors who
joined the board after the CEO are assumed to be less independent in Core et al. (1999) and in Coles et al. (2010). Santella et al.
(2006) quantify the extent to which corporate disclosure for the financial year 2003 allows for verification of the independence
of directors formally declared as independents by the 40 Italian blue chip firms. Santella et al. (2007) extend the previous
descriptive results comparing financial and non-financial firms in terms of ownership structure. Byrd et al. (2009) examines a
more behavioral definition of director independence to exclude problematic directors, who are unlikely to stand up to
management, to test the directors' impact on firm performance and management compensation. Hwang and Kim (2009) analyze
social independence between independent directors and the CEO and the implications for CEO compensation, pay performance
sensitivity and turnover. Cohen et al. (2012) analyze firms that appoint independent directors who are overly sympathetic to the
management. Finally, Ansari et al. (2014) in this special issue, analyze the links of independent directors with the controlling
family to understand the determinants of CEO successor choice in family firms.
Our contribution to the body of knowledge regarding strict board independence, including determinants and consequences, is
threefold. First, based on the best practices definitions of independent directors, we provide a testable set of eight formal
independence criteria and make the set operational for a panel of Spanish listed companies. We evaluate the strict independence
of each declared independent director and find a relevant difference between strict independence and declared board
independence. Second, we estimate a model to disentangle the characteristics of the companies that misclassify their
independent directors. We find that the proxies of managerial control increase with the presence of non-strictly independent
directors. Third, we test the impact of this misclassification on firms' future operating performance and on firms' corporate
governance practices, with particular attention on the impact of misclassification in the audit committee and in the nomination
and remuneration committee. We find only weak empirical evidence of a negative effect on future operating performance.
This paper begins with a discussion of the academic literature regarding the role of board independence and recommendations
regarding the codes of good governance. The next section links the international recommendations and the Spanish regulation
with the proposed empirical measure of independence. Section 4 includes the data description and the most relevant statistics.
The governance characteristics of firms with non-strictly independent directors are discussed in Section 5. Section 6 analyzes the
consequences of the presence of non-strictly independent directors on future operating performance, executive directors'
compensation, CEO turnover, audit qualifications, and earnings management. Section 7 presents the discussion of results and
Section 8 concludes.

2. Independent directors in the codes of good governance and the previous literature

An increasing number of corporate governance codes globally propose explicit recommendations regarding the structure of
boards of directors, and more specifically, regarding the desirable proportion that independent directors should hold. Aguilera
and Cuervo-Cazurra (2009) indicate that 46 countries issued 196 distinct codes stemming from the first code of 1978, including
the relevant 1992 UK Cadbury Report and leading up to the middle of 2008. These researchers indicate that the majority of codes
have explicit or implicit recommendations on six governance practices; the first practice examines the balance of executive and
non-executive directors, which includes the independent non-executive directors.
The corporate governance rules of the New York Stock Exchange approved by the SEC on 2003 (amended on 2009) and
codified in Section 303A of the NYSE's Listed Company Manual describe the first rule as follows: “Listed companies must have a
majority of independent directors”. The justification is that “effective boards of directors exercise independent judgment in
118 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

carrying out their responsibilities. Requiring a majority of independent directors will increase the quality of board oversight and
lessen the possibility of damaging conflicts of interest”.
Indeed, the proportion of independent directors has increased over time. Gordon (2007) indicates that the focus on
shareholder value and stock market prices have contributed to such an increase. More recently, Linck et al. (2009) noted the way
in which the Sarbanes–Oxley Act in the US substantially increased the presence of independents, indicating that a substantial
(although smaller) increase was also detected in firms for which this act was not mandatory. However, although the previous
literature analyzing the influence of independent directors on companies' behavior or performance is extensive, the literature is
not conclusive. On the one hand, some studies support positive outcomes. For example, Byrd and Hickman (1992) show a positive
relation between independent outside directors and shareholders' interests. Cotter et al. (1997) find that shareholders' wealth
increased during tender offers when a larger proportion of independent directors were on the board. Weir and Laing (2000)
report that for the UK, best corporate governance practices occur when independent directors play an important role. Benkel et al.
(2006) found that a large proportion of independent directors on the board (and in the audit committee) reduced the level of
earnings management, particularly for large firms.
On the other hand, Agrawal and Knoeber (1996) found empirical evidence of a negative relation between the proportion of
independent directors and firms' performance. Dalton et al. (1998) could not support a significant relationship between financial
performance and board composition. In the same direction, Bhagat and Black (2002) find no correlation between the measures of
long-term performance and independent directors on the board. Ferris and Yan (2007) found no relationship either between
board independence and performance in mutual funds or between board independence and scandals in these firms after the SEC
changed its corporate governance rules. Bhagat and Bolton (2009) also found that larger board independence negatively affected
operating performance before the Sarbanes–Oxley Act; however, after passage of the act, the relationship was found to be positive
and significant.
Several theoretical developments have attempted to clarify the role played by board independence, developing what could be
called an optimal board independence theory (Hermalin and Weisbach, 1998; Raheja, 2005; Adams and Ferreira, 2007; Harris and
Raviv, 2008; Kumarand and Sivaramakrishnan, 2008). This theory shows that board independence is not always in shareholders'
interest, partially explaining the apparently conflicting empirical evidence found regarding the role of board independence. For
example, Hermalin and Weisbach (2003) show that past performance is a determinant of optimal board independence. In
well-performing firms, it is in the shareholders' interest to allow boards with less independent directors. Several empirical papers
on board composition provide evidence consistent with the optimal board independence theory (e.g., Boone et al., 2007; Coles et
al., 2008; Linck et al., 2008). In addition, the previous literature that controls for this endogeneity problem finds that board
independence is in shareholders' interest when independence aligns with the optimal board independence theory. These findings
are supported by Duchin et al. (2010), who analyze the effect of regulatory modifications regarding board independence, and by
Nguyen and Nielsen (2010), who analyze the effect of the sudden deaths of independent directors. Adopting a different
methodology by using laboratory experiments, Gillette et al. (2003) indicate that board independence is in the shareholders'
interest even when non-informed independent directors are present.
However, to empirically determine the impact of board independence, we require an accurate measure of the firms' strict
independence level. Our approach relies on the formal independence criteria reported publicly by the firms. These criteria differ
from the informal independence criteria that consider social ties, i.e., Hwang and Kim (2009), or that excludes as independents
the directors involved in scandals in previous years, i.e., Byrd et al. (2009). Unlike Santella et al. (2006, 2007), who analyze
whether disclosure is sufficient to corroborate the declared independence by firms, our formal independence approach applies to
all directors in an explicit and measurable way. Furthermore, we analyze the relationship of independent directors with all
controlling shareholders, not just with the controlling family as in Ansari et al. (2014) in this special issue. This approach allows
us to detect the characteristics of firms that are more prone to misclassify directors as independents than others, and the
consequences of such misclassification.

3. What should an independent director be?

The observed number of independent board directors comes from the firm's self-classification when filing the required forms
according to their country or stock market regulations or disclosing their corporate governance report. The Spanish Unified Good
Governance Code of Listed Companies specifies that independent directors are “directors appointed for their personal or
professional qualities who are in a position to perform their duties without being influenced by any connection with the
company, its shareholders or its management”. The spirit of this definition does not differ from the current definitions in the
NYSE, the European Union or the UK regulations. Nevertheless, the details in the definition that international regulatory bodies
propose as independent directors' characteristics are heterogeneous. Aguilera and Cuervo-Cazurra's (2009) comparison of
corporate governance codes remarks that the definitions of independent director change across countries and even across firms.
Table 1 shows four proposals regarding what an independent director should be, and more specifically, what an independent
should not be. These proposals are found in the NYSE Listed Company Manual (amended November 25, 2009), the European
Union Commission recommendation of 15 February 2005 on the role of non-executive directors, the UK Corporate Governance
Code form 2012, and the Spanish Unified Good Governance Code form 2006. In the Spanish case, the definition has been
mandatory since 2007.
The criteria to check each director's independence in a set of Spanish listed companies fits into the Spanish regulation that is
similar to the majority of international locations, as Table 1 indicates. Furthermore, the Spanish regulatory environment allows us
Table 1
Definitions of Independent directors' requirements according to NYSE Listed Company Manual, European Union commission recommendation, UK unified corporate governance code and Spanish unified good governance
code.
NYSE Listed Company Manual. Section 303A.02
(Amended November 25, 2009)
An independent director is
No director qualifies as “independent” unless the
board of directors affirmatively determines that the
director has no material relationship with the listed
company (directly or as a partner, shareholder or
officer of an organization that has a relationship
with the company).
An independent director is not
(a) The director is, or has been within the last three
years, an employee of the listed company, or an
immediate family member is, or has been within
the last three years, an executive officer of the listed
company.
(b) The director has received, or has an immediate
family member who has received, during any
twelve-month period within the last three years,
more than $120,000 in direct compensation from
the listed company, other than director and
committee fees and pension or other forms of
deferred compensation for prior service (provided
such compensation is not contingent in any way on
continued service).
(c) (A) The director is a current partner or
employee of a firm that is the listed company's
internal or external auditor; (B) the director has an
immediate family member who is a current partner
of such a firm; (C) the director has an immediate
family member who is a current employee of such a
firm and personally works on the listed company's
European Union Commission recommendation of 15 The UK Corporate Governance Code September 2012.
February 2005 on the role of non-executive or
supervisory directors of listed companies and on the
committees of the (supervisory) board
A director should be considered to be independent
only if he is free of any business, family or other
relationship, with the company, its controlling
shareholder or the management of either, that
creates a conflict of interest such as to impair his
judgment.
(a) not to be an executive or managing director of
the company or an associated company, and not
having been in such a position for the previous five
years;
(b) not to be an employee of the company or an
associated company, and not having been in such a
position for the previous three years, except when
the non-executive or supervisory director does not
belong to senior management and has been elected
to the (supervisory) board in the context of a
system of workers' representation recognized by
law and providing for adequate protection against
abusive dismissal and other forms of unfair
treatment;
(c) not to receive, or have received, significant
additional remuneration from the company or an
associated company apart from a fee received as
non-executive or supervisory director. Such
additional remuneration covers in particular any
participation in a share option or any other
performance-related pay scheme; it does not cover
The board should identify in the annual report each
non-executive director it considers to be independent.
The board should determine whether the director is
independent in character and judgment and whether
there are relationships or circumstances which are
likely to affect, or could appear to affect, the director's
judgment. The board should state its reasons if it
determines that a director is independent notwith-
standing the existence of relationships or circumstances
which may appear relevant to its determination.
(a) has been an employee of the company or group
within the last five years;
(b) has, or has had within the last three years, a
material business relationship with the company
either directly, or as a partner, shareholder, director
or senior employee of a body that has such a
relationship with the company;
(c) has received or receives additional
remuneration from the company apart from a
director's fee, participates in the company's share
option or a performance-related pay scheme, or is a
member of the company's pension scheme;
(d) has close family ties with any of the company's
advisers, directors or senior employees;
(e) holds cross-directorships or has significant links
with other directors through involvement in other
companies or bodies;
(f) represents a significant shareholder; or has
served on the board for more than nine years from
the date of their first election.
R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134
Unified Good Governance Code of listed companies.
May 19, 2006. Spain
Directors appointed for their personal or
professional qualities who are in a position to
perform their duties without being influenced by
any connection with the company, its shareholders
or its management.
(a) Past employees or executive directors of group
companies, unless 3 or 5 years have elapsed,
respectively, from the end of the relation.
(b) Those who have received some payment or
other form of compensation from the company or
its group on top of their directors' fees, unless the
amount involved is not significant. Dividends or
pension supplements received by a director for
prior employment or professional services shall not
count for the purposes of this section, provided such
supplements are non contingent, i.e. the paying
company has no discretionary power to suspend,
modify or revoke their payment, and by doing so
would be in breach of its obligations.
(c) Partners, now or on the past 3 years, in the
external auditor or the firm responsible for the audit
report, over the said period, of the listed company
or any other within its group.
(d) Executive directors or senior officers of another
company where an executive director or senior
officer of the company is an external director.
(continued on next page)
119

Table 1 (continued)
NYSE Listed Company Manual. Section 303A.02
(Amended November 25, 2009)
audit; or (D) the director or an immediate family
member was within the last three years a partner
or employee of such a firm and personally worked
on the listed company's audit within that time.
(d) The director or an immediate family member is,
or has been with the last three years, employed as
an executive officer of another company where any
of the listed company's present executive officers at
the same time serves or served on that company's
compensation committee.
(e) The director is a current employee, or an
immediate family member is a current executive
officer, of a company that has made payments to, or
received payments from, the listed company for
property or services in an amount which, in any of
the last three fiscal years, exceeds the greater of $1
million, or 2% of such other company's consolidated
gross revenues.
General Commentary to Section 303A.02(b): An
“immediate family member” includes a person's
spouse, parents, children, siblings, mothers and
fathers-in-law, sons and daughters-in-law, brothers
and sisters-in-law, and anyone (other than domestic
employees) who shares such person's home. When
applying the look-back provisions in Section
303A.02(b), listed companies need not consider
individuals who are no longer immediate family
members as a result of legal separation or
divorce, or those who have died or become
incapacitated.
In addition, references to the “listed company” or
“company” include any parent or subsidiary in a
consolidated group with the listed company or
such other company as is relevant to any
determination under the independent standards
set forth in this Section 303A.02(b).
European Union Commission recommendation of 15 The UK Corporate Governance Code September 2012.
February 2005 on the role of non-executive or
supervisory directors of listed companies and on the
committees of the (supervisory) board
the receipt of fixed amounts of compensation under
a retirement plan (including deferred compensa-
tion) for prior service with the company (provided
that such compensation is not contingent in any
way on continued service);
(d) not to be or to represent in any way the
controlling shareholder(s) (control being
determined by reference to the cases mentioned in
Article 1(1) of Council Directive 83/349/EEC (1));
(e) not to have, or have had within the last year, a
significant business relationship with the company
or an associated company, either directly or as a
partner, shareholder, director or senior employee of
a body having such a relationship. Business
relationships include the situation of a significant
supplier of goods or services (including financial,
legal, advisory or consulting services), of a signifi-
cant customer, and of organizations that receive
significant contributions from the company or its
group;
(f) not to be, or have been within the last three
years, partner or employee of the present or former
external auditor of the company or an associated
company;
(g) not to be executive or managing director in
another company in which an executive or manag-
ing director of the company is non-executive or
supervisory director, and not to have other signif-
icant links with executive directors of the company
through involvement in other companies or bodies;
(h) not to have served on the (supervisory) board as
a non-executive or supervisory director for more
than three terms (or, alternatively, more than
12 years where national law provides for normal
terms of a very small length);
(i) not to be a close family member of an executive
or managing director, or of persons in the situations
referred to in points (a) to (h);
120
Unified Good Governance Code of listed companies.
May 19, 2006. Spain
(e) Those having material business dealings with
the company or some other in its group or who have
had such dealings in the preceding year, either on
their own account or as the significant shareholder,
director or senior officer of a company that has or
has had such dealings. Business dealings will
include the provision of goods or services, including
R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134
financial services, as well as advisory or consultancy
relationships.
(f) Significant shareholders, executive directors or
senior officers of an entity that receives significant
donations from the company or its group, or has
done so in the past 3 years. This provision will not
apply to those who are merely trustees of a
Foundation receiving donations.
(g) Spouses, partners maintaining an analogous
affective relationship or close relatives of one of the
company's executive directors or senior officers.
(h) Any person not proposed for appointment or
renewal by the Nomination Committee.
(i) Those standing in some of the situations listed in
a), e), f) or g) above in relation to a significant
shareholder or a shareholder with board
representation. In the case of the family relations set
out in letter g), the limitation shall apply not only in
connection with the shareholder but also with his or
her proprietary directors in the investee company.
Proprietary directors disqualified as such and
obliged to resign due to the disposal of shares by the
shareholder they represent may only be re-elected
as independents once the said shareholder has sold
all remaining shares in the company. A director
with shares in the company may qualify as
independent, provided he or she meets all the
conditions stated in this Recommendation and
the holding in question is not significant.
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 121

to measure strict independence director by director using public data provided by the listed firms, compared with their declared
independence.
Under the Spanish regulation, since 2004 listed companies have been required to release a standardized annual report on
corporate governance practices (ARCG) to the Spanish securities and exchange commission (CNMV). The report is filed
electronically and is publicly available at the CNMV web page. The report includes information about board directors' names and
their typology according to the company's judgment and specifically defines executive directors, proprietary directors,
independent directors and others. The ARCG also includes information on the board members' relationship with significant
shareholders, the firm, its associates and subsidiaries.
The ARCG allows us to empirically test eight independence criteria for all directors declared as independent. These criteria,
shown in Table 2, can be interpreted as international standard definitions specifically focused on the Spanish regulation as stated
in Table 1. The selection of these criteria, based on the Spanish mandatory definition of independent directors and the
international standards, is adapted to the availability of public information to check the majority of the aspects of this definition.3
Criterion 1 addresses the 2007 Spanish corporate governance code that states that the appointment by the nomination committee
is compulsory for the director to be classified as independent. Nevertheless, international best practices codes include the
recommendation of giving this committee tasks such as the appointment of the independent directors. Shivdasani and Yermack
(1999) found empirical evidence that CEOs' involvement in the appointments of new directors when no nomination committee
existed yielded more ‘gray’, non-strictly independent directors with conflicts of interest. Limited tenure of up to twelve years is
the second criterion, which is also included in the UK combined code and the EU recommendations. The six remaining criteria are
standard in the international regulations, as Table 2 confirms. The third criterion restricts independence to those directors who do
not have a significant business relationship with the company. The relationship with the controlling shareholders is an element
that the Spanish code, the NYSE rules and other codes define as essential for the independence of directors. The condition of being
a director, a manager or employee of a significant shareholder (fourth criterion), having other types of relevant relationship (i.e.,
family) with a significant shareholder (fifth criterion), or receiving compensation by the company, its subsidiaries or its
associates, for other functions apart from the directorship (sixth criterion) must exclude the director from being qualified as a real
independent member. Companies can also be formally board members through a representative, although the seventh criterion
obviously does not accept this type of director as an independent. Our last criterion, the eighth, avoids classifying as independent
any recent company executive. Table 2 summarizes these criteria, showing the fit with the corresponding rule in the codes
revised in Table 1.

4. Data and descriptive statistics

The data set consists of 752 firms/year observations ranging from 2004 to 2009. This six-year dataset includes all listed firms
in the Spanish stock market, ranging from 118 to 135 firms depending on the year. Information regarding board composition,
corporate governance practices and individual information on board members, such as tenure or their relationship with
significant shareholders, is obtained from the standardized ARCG that firms must fulfill. Stock market data are from the Thomson
Financial database and data on firms' annual financial reports are from the Bureau Van Dijk database on financial reports for
Spanish firms.
Panel A of Table 3 describes the most relevant statistics of board independence of the Spanish listed companies. Although the
firms declared an average of 32.5% of independent directors on the board, this figure decreases to 14.2% when considering strictly
independent directors over board size; that is, 56.3% of the declared independent directors do not comply with at least one of the
eight independence criteria described in Table 2. The proportion of non-strictly independents is nearly uniform across firm size
and industries and decreases along time (panel C), particularly in 2007 when the Spanish mandatory definition of independent
directors became compulsory.
Panel B of Table 3 analyzes the pattern of new independent directors appointed each year. The proportion of non-strictly
independent new directors supports the notion that the findings of panel A are not merely driven by the stock of appointed
directors, when independent directors' appointments were less rigorous. The shift in the level in 2007 is consistent with the
mentioned new mandatory definition of an independent director under the Spanish regulation. The Spanish best practices code
also recommends that the audit committee and the nomination and remuneration committee should be chaired by an
independent director and formed by a majority of independents, with no executives in place. Panel D of Table 3 shows that the
average percentage of declared independent directors in both commissions is substantially higher than the percentage of
independents on the board (one-half versus one-third, respectively). However, after filtering the non-strictly independent
directors, these percentages decrease significantly. The declared average of 49.4% of independent members in the audit
committee and the average of 50.4% in the nomination and remuneration committee decrease to 21.2% and 24.1% of strictly
independents, respectively. The majority of firms appointed a declared independent director for the chair of both committees,
although on average in less than one-third of firms were chaired by strictly independent directors. Panel E indicates that the rate
of misclassified independents in these committees is decreasing over time. Nevertheless, for the year 2009 the proportion of
strictly independents is always below 50% when one-third of the firms still have non-strictly independent directors chairing these

3
The potential bias of our empirical test of independence is toward the overestimation of true independence, never toward an excess of strictly non
independents.
122 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

Table 2
Empirical definition of independent directors of Spanish listed firms. Includes the corresponding rule under different international legal settings from Table 1.

NYSE EU UK Spain

[1] Proposed for appointment or renewal by the Nomination Committeea (h)

[2] Tenure as independent director for up to twelve years (h) (f)
[3] Not having a significant business relationship with the company (b) (e) (b) (e)
[4] Not holding a directorship, to be a manager or an employee of significant (d) (g) (f) (i)
shareholder or a shareholder with board representation
[5] Not having other relevant relationship (different than those in point 4) with (d) (g) (d) (i)
significant shareholder or a shareholder with board representation
[6] Not being a director or executive in subsidiaries or associated companies (c) (a) (e) (d)
[7] Not to be a company as board director (a) (b) (b) (c)
[8] Not being executive director of the firm in the previous year b (a) (a) (a) (a)
a
In 2007 the CNMV (the Spanish Securities and Exchange Commission) modified the information requirements regarding director proposals. Firms must
communicate who proposed every director, except for independent directors. Therefore, after 2006 we assume that all independent directors have been proposed
by the nomination committee, except when this committee does not exist, or if the director has not been formally renewed and was not promoted by this
committee before 2007.
b
As we have information starting from 2004, this criterion does not operate for this year.

committees. One-sixth of the firms have non-strictly independents chairing both committees. These findings remain relevant in
firms of all sizes and industry activities.
Table 3 captures the values of the dependent variables in the empirical model of misclassification. In a step further, these
measures of non-strictly independent directors become a part of the set of explanatory variables of the operating performance
model and also of the models of executive directors' pay, CEO turnover, audit qualification and earnings management.
Table 4 describes the basic statistics of the variables included in the posterior analysis, as performance measures, corporate
governance practices or control variables. This table contains a short definition of these variables. The justification of their
inclusion in the analysis and the full definition will be found in the sections analyzing the empirical models.
The descriptive statistics show that the average ownership concentration is high. Non-reported Pearson correlation
coefficients show high values among variables such as sales and market capitalization, both proxies of firm size, and are also
positively correlated with board size.

5. The governance of firms with non-strictly independents

Firms controlled by managers may appoint non-strictly independent directors to achieve the regulator's recommended level
of board independence and to reach the desired level of board real independence (a friendly board). In firms controlled by large
block holders, independents are supposed to protect the interest of minority shareholders, and large shareholders may also use
non-strictly independents to achieve a desired low level of real board independence at the same time as the recommended level
of declared board independence. Therefore, we analyze whether ownership structure and managerial control proxies are related
to the presence of non-strictly independent directors. We estimate the following logit model:

PNSIDi;t ¼ α þ α 1  Log Salesi;t þ α 2  C1i;t þ α 3 CEO is board chairi;t þ α 4 Board sizei;t þ

þα 5  Voting Capi;t þ α 6  %Busy non‐executive directorsi;t þ ð1Þ
þα 7  %Interlocked executive directorsi;t þ X i;t β þ εi;t

where PNSIDi,t takes a value of 1 if firm i has at least one non-strictly independent director in year t. We use the ownership of the
largest shareholder (C1) as a proxy of ownership concentration (it is highly correlated with the accumulated ownership of the
three and five largest shareholders, 0.92 and 0.84, respectively). As proxies of managerial power, we use the following: a dummy
variable to identify firms in which the CEO also chairs the board of directors; board size (to detect board coordination problems in
its monitoring functions, Yermack, 1996); a dummy variable to detect firms with voting caps (a maximum number of votes that a
shareholder may facilitate); the percentage of busy non-executive directors (those with three or more directorships, with no time
to properly monitor executives, Fich and Shivdasani, 2006); and the percentage of interlocked executive directors (serving in a
second firm as non-executives with a non-executive director of the first firm serving as executive, Hallock, 1997). Finally, we add
size (the log of sales), and year and industrial sector fixed effects (in Xi,t) to detect any systematic differences in the presence of
non-strictly independents related to such factors.
Table 5 shows the results of the following three logit specifications: pooled logit models with robust standard errors clustered
by firm (Huber–White), panel data random effects logit models, and generalized estimating equations (GEE) logit models (Liang
and Zeger, 1986; Zeger and Liang, 1986) to account for any persistence in the decision to have misclassified directors within the
same firm.4 This method is equivalent to feasible generalized least squares for lineal models. The panel data models show 50
replication bootstrap standard errors to obtain robust statistical inference (Mooney and Duval, 1993).5 Models 1, 4, and 7 include

4
See Ballinger (2004) for a description of this method in organizational research.
5
Huber–White standard errors are not feasible for logit panel data models with random effects.
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 123

Table 3
The strict independence of directors. Panel A: “% declared” is the average percentage of independent directors over board size declared by firms in their Annual
Report of Corporate Governance (ARCG). The remaining percentages refer to the average of strictly independents over the board size once pass the eight joint
independence criteria [1..8] and individually [i]. In italics, there is the percentage of firms that do not comply with the independence criteria. Panel B: Number of
new appointed independent directors and the proportion that meet the criteria. Panel C: Longitudinal perspective of Panel A and Panel B data contents. Panel D:
The average percentage of independent directors over committees' size as declared and the average percentage of strictly independents in the audit, and in the
nomination and remuneration committee over committees' size. % firms chair is the percentage of firms with a declared independent director chairing the
committee and the percentages of firms where there is a strictly independents after the eight criteria. Panel E: Longitudinal perspective of Panel D data contents.
Data for a pool of all listed firms in Spain from 2004 to 2009: 118 firms in 2004, 119 firms in 2005, 126 in 2006, 135 for 2007, 130 during 2008, and 124 in 2009.

Independence criteria

% declared [1…8] [1] [2] [3] [4] [5] [6] [7] [8]

Panel A: % of existing declared and strictly independent director's

32.5% 14.2% 20.3% 28.0% 30.6% 32.1% 32.1% 28.0% 31.6% 32.5%
69.8% 43.0% 28.7% 10.5% 3.7% 2.8% 27.3% 7.0% 0.5%

Panel B: # New appointed declared and % strictly independent director's

419 64.70% 83.80% 95.50% 95.90% 98.10% 99.30% 88.10% 95.50% 99.00%

Panel C: Declared and strictly independent director's along time

Existing directors New appointed

% declared [1…8] # New [1…8]

2004 33.2% 8.3%

2005 33.5% 10.5% 56 58.9%
2006 32.2% 10.7% 85 29.4%
2007 30.8% 16.1% 117 74.4%
2008 32.4% 18.4% 82 73.2%
2009 33.3% 20.5% 79 83.5%

Panel D: Declared and strictly independent director’s in committees

Audit committee Nomination and remuneration committee

% Independents % firms Chair Independent % Independents % firms Chair Independent

Declared After [1…8] Declared After [1…8] Declared After [1…8] Declared After [1…8]

49.4% 21.2% 71.8% 30.9% 50.4% 24.1% 64.0% 32.9%

Panel E: declared and strictly independents in committees along time

2004 47.1% 9.8% 64.4% 10.2% 48.7% 15.1% 52.8% 18.0%

2005 48.3% 13.5% 63.9% 15.1% 49.7% 18.4% 56.0% 20.9%
2006 48.6% 17.0% 68.3% 20.6% 50.3% 18.7% 56.9% 24.5%
2007 49.7% 26.3% 75.6% 40.7% 48.3% 26.0% 61.9% 36.5%
2008 51.7% 28.7% 76.2% 45.4% 52.6% 30.4% 72.7% 46.3%
2009 50.6% 30.5% 81.5% 50.0% 52.6% 31.4% 78.0% 43.2%

one proxy of shareholder power (C1) and managerial power (CEO chairs the board). Models 2, 5, and 8 include the remaining
corporate governance variables. Because board size might proxy firm size, we also estimate models 3, 6, and 9 without this
variable as a robustness test.
The models in Table 5 clearly show that the stronger the control by the largest shareholder is, the lower the proportion of firms
misclassifying independents is.6 We also obtain weak empirical evidence (for the pooled logit models) of a positive interaction
among firms with non-strictly independents and voting caps limiting the power of potential new entrant shareholders. Board size
is also positively related to non-strictly independents, except in the GEE panel data logit specification.
In sum, the empirical evidence of these logit models suggests that large shareholders do not use non-strictly independent
directors to decrease independents' influence. Firms controlled by managers (with dispersed ownership structures) tend to
reduce real board independence with non-strictly independents.

6. The impact of non-strictly independents in the boardroom

Based on the agency theory, the approach regarding the relevance of independent directors in the boardroom is that the
monitoring of managerial power is a key element in preventing shareholders' wealth expropriation. A potential manager's

6
Even basic statistics are consistent with this result: 497 observations (firm/year) belong to firms in which the largest shareholder owns less than half of
shares; in 77.4% of these observations there are non-strictly independent directors. The largest shareholder owns more than half of the shares in 255 observations
and in 54.9% of them there are misclassified independent directors.
124 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

Table 4
Descriptive statistics of firms' characteristics: mean, median and standard deviations. NCR refers to the nomination and remuneration committee, AC refers to the
audit committee. Percentages of non-strictly independents are computed over board size and over the committee size, respectively. Average Executive
compensation is the total compensation, of any kind, of executive directors divided by the number of executive directors on the board. CEO turnover refers to the
percentage of cases where there is a change of a CEO between two consecutive years. Audit qualification captures the “not clean” opinion of the external auditor
(no data for banks). Loss is a dummy variable for the firms that report losses in the earnings statement. Leverage is the level of long term debt over total assets (no
data for banks). Liquidity is current assets over current liabilities (no data for banks). Operating Cycle is the sum of days receivable and days inventory (no data for
banks, neither Days payable). Capital Intensity is fixed assets over total assets (no data for banks). Executives' ownership and Non-executives' ownership is the sum
of the ownership stakes of executive and non executive directors respectively. C1 is the percentage of shares owned by the largest shareholder. CEO is board Chair
is a dummy variable to identify firms where the CEO chairs the board of directors. Board size is the number of directors. Tenure is the average number of years that
board executives remain in board since their appointment. Operations directors-firm is a dummy that captures if the firm reports having transactions or
commercial operations with its directors. % busy non-executive directors (a director is busy if she holds a position in three or more boards of directors) is computed
over board size, as it is the percentage of interlocked executive directors (those who are also non-executive directors in a firm where a non-executive director of
the first firm is an executive director). Golden parachute is a dummy variable that captures firms protecting executives against dismissal. Voting cap is a dummy
variable identifying firms with a maximum number of votes that a shareholder may exercise independently of the number of shares she has.

% Observ. Mean Median Std Dev

% non-strictly independents 18.31 14.29 17.89

% non-strictly independents NRC 26.34 25.00 27.20
% non-strictly independents AC 28.13 25.00 29.53
NRC non-strictly independent chair 31.07
AC non-strictly independent chair 40.96
Average Executive Compensation (1000 €) 969.12 554.00 1252.47
CEO turnover events 14.04
Audit Qualification 11.72
Sales (Mill €, annual) 3,509.33 618.80 9,262.89
Market Capitalization (Mill €, yearend) 4,655.18 833.71 12,074.34
ROA (Return on Assets, annual) 4.62 4.68 9.13
Stock Return (annual) 14.11 11.18 59.13
Market to Book (equity, yearend) 3.26 2.13 4.48
Loss 13.16
Leverage 0.27 0.25 0.18
Liquidity 3.63 1.23 35.22
Operating Cycle (days) 329.37 118.08 573.86
Days Payable (days) 213.99 149.60 211.26
Capital Intensity 0.55 0.58 0.23
Executives ownership 11.21 0.10 22.23
Non-executives ownership 12.55 3.88 18.98
C1 (Ownership Largest Shareholder) 36.39 29.54 26.05
CEO is board chair 57.98
Board size 11.04 10.00 3.86
% Executives on the board 20.16 19.38 12.38
Tenure of board executives (years) 8.56 7.02 6.88
Operations directors-firm 37.77
% Busy non-executive directors 9.64 0.00 14.01
% Interlocked executive directors 12.02 0.00 26.20
Golden parachutes 57.58
Voting Cap 12.77

expropriation behavior is more influential under dispersed ownership structures, where shareholders' coordination is difficult
and costly, than under concentrated ownership structures. For firms without significant large owners, our findings show that
incentives to appoint non-strictly independents are stronger. However, this scenario does not imply that the managerial power is
the ultimate reason for the widespread presence of non-strictly independent directors in our sample. The optimal board
independence theory may be this ultimate reason. For example, our results may be explained by Hermalin and Weisbach (2003),
who found that it is optimal in shareholders' interest to reduce board independence for CEOs with positive records of past
performance. Whenever the optimal board independence level of a firm is below the recommended level of board independence,
firms have incentives to use non-strictly independent directors to achieve the optimal level of real board independence at the
same time as the recommended level of declared independence. Indeed, Cicero et al. (2013) found empirical evidence of firms
declaring board independence above its optimum level.
Therefore, we next analyze the consequences of independents' misclassification in terms of future firms' performance and in
terms of several relevant outputs of the main board committees with monitoring roles, which include the audit committee and
the nomination and remuneration committee. Negative consequences will be interpreted as support for the managerial power
approach, whereas positive consequences or a lack of consequences will support the optimal board independence theory. First, for
the nomination and remuneration committee, we test whether executive directors' compensation practices and CEO turnover are
dependent on the level of non-strictly independent directors. Second, for the audit committee, we test whether the number of
non-strictly independent directors affects firms' audit qualification and earnings management behavior.
Nguyen and Nielsen (2010) show that the contribution to the firm value by independent directors is greater when the
directors play relevant roles on the board committees, particularly on the audit committee. Under the Spanish Securities and
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 125

Table 5
Corporate governance and non-strictly independents. The dependent variable takes value 1 when a firm has a director that does not meet the eight criteria for
independence described in Table 2. Panel A shows Pooled logit model estimations with Huber (1967) and White (1980, 1982) robust t statistics (in parenthesis),
clustering by firm (Petersen, 2009). Panel B shows Panel data logit estimations with random effects where t statistics are computed by bootstrap techniques with
50 replications (Mooney and Duval, 1993). Panel C presents the GEE panel data logit models allowing correlation among the error terms of the same firm, with 50
replications bootstrap standard errors. The explanatory variables are the logarithm of sales, the percentage of shares owned by the largest shareholder, a dummy
variable detecting if the CEO also chairs the board of directors, the board size, a dummy variable identifying firms with voting caps (a maximum number of votes
that a shareholder may exercise independently of the number of shares she has), the percentage of busy non-executive directors (a director is busy if she holds a
position in three or more boards of directors), the percentage of interlocked executive directors (those who are also non-executive director in a firm where a
non-executive director of the first firm is an executive director), dummy variables for industries (6) and years. Sample of all firms listed in the main trading
mechanism of the Spanish Stock Exchange (SIBE). Chi2 is a Wald test of the statistical significance of all the explanatory variables. Chi2-II is a Wald test of the joint
statistical significance of Board size, Voting Cap, % Busy non-executive directors, and % Interlocked executive directors. Chi2-III is a Wald test of the joint statistical
significance of Voting Cap, % Busy non-executive directors, and % Interlocked executive directors. Likelihood ratio test rho = 0 is a test on the statistical
significance of the variance of the unit-specific residual. *** denotes significance at the 1% level; ** denotes significance at the 5% level; * denotes significance at the
10% level.

Panel A: Pooled logit models Panel B: Panel data logit with random Panel C: GEE Panel data logit models
effects

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Log Sales 0.106 −0.089 0.079 0.161 −0.111 0.136 −0.081 −0.092 −0.083
(1.229) (−0.824) (0.924) (0.778) (−0.443) (0.809) (−1.448) (−1.620) (−1.358)
C1 −0.022*** −0.017*** −0.020*** −0.037*** −0.030** −0.036*** −0.009** −0.008** −0.008**
(−3.830) (−2.809) (−3.368) (−2.768) (−2.236) (−3.007) (−2.282) (−2.079) (−2.109)
CEO is board chair 0.341 0.430 0.312 0.565 0.563 0.571 0.078 0.079 0.081
(1.127) (1.364) (1.044) (0.784) (0.997) (1.098) (0.663) (0.645) (0.691)
Board size 0.148*** 0.263*** 0.010
(2.877) (2.883) (0.480)
Voting Cap 1.134* 1.186** 1.078 1.251 0.262 0.264
(1.830) (1.993) (1.092) (0.344) (1.117) (1.107)
% Busy non-executive directors 1.316 1.769* 0.638 1.043 0.120 0.138
(1.489) (1.940) (0.399) (0.536) (0.327) (0.372)
% Interlocked executive directors −0.001 0.064 −0.648 −0.376 −0.116 −0.108
(−0.002) (0.103) (−0.554) (−0.330) (−0.427) (−0.302)
Year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes
Industry fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 752 752 752 752 752 752 752 752 752
chi2 41.51*** 52.10*** 50.66*** 36.53*** 40.21*** 57.99*** 111.42*** 95.72*** 112.91***
chi2-II 15.52*** 14.37*** 1.68
chi2-III 7.27* 0.50 1.3
Likelihood ratio test rho = 0 197.23*** 172.4***

Market Law, all listed firms must have an audit committee, whereas the Spanish corporate governance code merely recommends
the presence of a nomination and remuneration committee (86% of observations, firm/year).

6.1. Future operating performance

If non-strictly independent directors are appointed to mask an agency problem among managers and owners (or large and
minority owners), we should expect a negative impact on firm future performance. As operating performance, we evaluate return
on assets, calculated as the net income plus interest payments, net of tax effects, over the amount of the previous year's total
assets. The following five measures capture the power of non-strictly independents: the proportion of non-strictly independent
directors over board size; the proportion over the committee size of non-strictly independent directors in the nomination and
remuneration committee and in the audit committee; and two dummy variables to identify firms in which the chair of these
committees is a non-strictly independent director. Given the relevance of the ownership structure found in the previous section,
these measures are also interacted with the ownership of the largest shareholder to investigate any different impact of
non-strictly independents. This interaction is also conducted in our posterior analyses. The structure of the model and the control
variables follow Hwang and Kim (2009):


Performancei;tþ1 ¼ α þ α 1  NSIDi;t þ α 2  C1i;t × NSIDi;t þ X i;t β þ εi;t ð2Þ

where Performancei,t + 1 is the ROA of firm i in year t + 1, NSIDi,t is the non-strictly independent directors measure of firm i in
year t, and Xi,t is a set of control variables including proxies of firm size, growth opportunities, and corporate governance variables.
The measure of firm size is the log of sales, and growth opportunities are taken as one period lagged market to book ratio
(of equity). To capture the potential impact of an uncontrolled agency problem we use the ownership structure and a group of
board of directors and corporate governance characteristics. Ownership structure proxies are the ownership stake of the largest
shareholder, the ownership of executive directors and the shares owned by non-executive directors. The board of directors'
characteristics are; a dummy variable to identify firms in which the CEO chairs the board, the percentage of executive directors,
126 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

the presence of “golden parachutes” to protect executive directors against dismissals (Brick et al., 2006), the board size to detect
board coordination problems (Yermack, 1996), the percentage of busy directors with dedication problems (Fich and Shivdasani,
2006), and the percentage of interlocked executive directors to detect the potentially lower monitoring effort of non-executive
directors (Hallock, 1997). With a dummy variable, we also control whether directors conduct relevant economic transactions
with the firm (as declared by firms in their ARCG) as a potential rent extraction mechanism; this condition is particularly relevant
in our sample given the predominance of owner-controlled firms (Faccio et al., 2001). We estimate these models with time and
firm fixed effects to control for any systematic time effect and control for non-observable constant firm characteristics affecting
future firm performance. We include one period lagged return on assets to address any endogeneity concern; however, the results
do not change without this variable (available on request). Inference is based on Huber–White robust t statistics.
The initial 752 firm/year observations decrease to 667 because of the lags in the market to book ratio; missing values are
generated by mergers and acquisitions in the period and the new listing firms that do not report lagged values.7 When
information on the nomination and remuneration committee is used, the sample decreases to 580 firm/year observations because
14% have no such committee.
Table 6 shows that the presence of misclassified independent directors is not related to future operating performance except
when a non-strictly independent director chairs the board committees, although with low statistical significance. In this case,
future operating performance tends to be lower (models 4 and 5) than for firms with strictly independents chairing these
committees. When interacting the non-strictly independents variables with the ownership stake of the largest shareholder we do
not detect relevant differences, except the higher statistical significance of the negative effect of a non-strictly independent
chairing the audit committee.8
The results remain robust using different specifications as the model of future operating performance used in Core et al.
(1999), which includes the variance of past operating performance and one period lagged sales as control variables instead of the
lagged market to book ratio and of contemporaneous sales, results are omitted for brevity but are available on request.
Criticism of the performance models in corporate finance occurs primarily regarding the potential endogeneity of some of the
explanatory variables or regarding the omission of relevant variables. Although our panel data estimations tend to uncover these
issues, additional measures of the impact of non-strictly independents on the outcomes of board committees are analyzed.9 The
expected outcomes of a properly performing nomination and remuneration committee include, among others, a scenario in
which the presence of non-strictly independents does not distort executive compensation or CEO turnover. Similarly, for the audit
committee, given the lack of agency problems because of non-strictly independent directors, we do not expect that the presence
of non-strictly independents affect the external auditor qualifications or the earnings management behavior of the firm.

6.2. Outcomes of the nomination and remuneration committee

6.2.1. Executive director compensation

In this sub-section we analyze the relation between non-strictly independent directors and the average individual
compensation of executive directors, giving special attention to non-strictly independents in the nomination and remuneration
committee.10 The conjecture in this case would be that amid agency problems related to non-strictly independents' control in the
nomination and remuneration committee, executive compensation would be larger compared to firms without non-strictly
independents in the boardroom (see Goergen and Renneboog, 2011, for the relation between executives' compensation and weak
corporate governance). The proposed model of compensation includes fixed pay, bonuses, cash from exerted stock options,
retirement benefits, and any additional remuneration from the firm. We estimate the following regression:


Compensationi;t ¼ α þ α 1  NSIDi;t þ α 2  C1i;t × NSIDi;t þ X i;t β þ εi;t ð3Þ

where Compensationi,t is the log of the average executives compensation of firm i in year t, NSIDi,t is the non-strictly independent
directors measure, and Xi,t is a set of control variables. The structure of control variables follows Core et al. (1999) with size (the
log of one period lagged sales), growth opportunities (the average of the past three years' market to book ratio), past performance
(one period lagged ROA and stock return), and risk (the variance of the past three years' ROA and stock return) as the economic
determinants of compensation based on firm characteristics (see also Aggarwal and Samwick, 1999; Core and Guay, 1999; Cyert
et al., 1997; Jackson et al., 2011; Lambert and Larcker, 1987). The model then considers the ownership structure and other
corporate governance characteristics as proxies of a potentially uncontrolled agency problem. The measure of ownership
structure is, as previously, the amount of the ownership stake of the largest shareholder, the ownership of executive directors and
of non-executive directors (see Bertrand and Mullainathan, 2001; Elston and Goldberg, 2003; Hartzell and Starks, 2003; Sun et al., 2009).

7
In Section 6, firms resulting from mergers and acquisitions are analyzed as new entities.
8
Table 6 does not show control variables due to space constrains. Among these variables, the one period lagged market to book ratio is the only one that has a
statistically significant coefficient in all models (positive, with 5% or 1% statistical significance depending on the model).
9
Alternatively, Andres et al. (2014) and Liu et al. (2014), both in this special issue, also use differences in differences and instrumental variables respectively to
address the endogeneity concern when operating performance is the dependent variable.
10
As in other countries, such as Germany till 2006, the ARCG in Spain just reports aggregate compensation of executive directors. Indeed, with Germany data,
Andres et al. (2014) in this special issue also analyze the average individual compensation of executive directors.
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 127

Table 6
Future firm's performance and independents’ misclassification. Firm fixed effects panel data regressions of future Return on Assets one year ahead explained by
the percentage of misclassified independents on the board of directors, in the nomination and remuneration committee (NRC), in the audit committee (AC), and a
dummy variable to identify firms where the chair of the previous committees is a misclassified independent director. Models 6 to 10 also contain the previous
variables interacted with the ownership of the largest shareholder (C1). Control variables (omitted to save space) are the log of sales, the log of lagged market to
book ratio, the lagged return on assets, and the following Corporate Governance variables; the percentage of shares owned by executives, by non-executives, and
by the largest shareholder of the firm, a dummy variable identifying whether the CEO is also the chairman of the board of directors, the size of this board, the
percentage of executives in this board, a dummy variable identifying whether directors have done commercial transactions with the firm, the percentage of busy
non-executive directors (a director is busy if she holds a position in three or more boards of directors), the percentage of interlocked executive directors (those
who are also a non-executive director in a firm where a non-executive director of the first firm is an executive director), a dummy variable identifying where
there are golden parachutes protecting top executives against dismissal, and a dummy variable identifying firms where there are voting caps (a maximum
number of votes that a shareholder may exercise independently of the number of shares she has). Year dummy variables are introduced. Huber (1967) and White
(1980, 1982) robust t statistics are in parenthesis. Models 2,4,7 and 9 have a lower number of observations since just 86% of observations (firm/year) have
nomination and remuneration committee. *** denotes significance at the 1% level; ** denotes significance at the 5% level; * denotes significance at the 10% level.

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

% non-strictly indep. 1.571 1.302

(0.469) (0.264)
% non-strictly indep. NRC −0.533 −1.367
(−0.380) (−0.653)
% non-strictly indep. AC −1.167 −1.6717
(−0.60) (−0.564)
NRC chair non-strictly indep. −1.486* −2.134*
(−1.812) (−1.663)
AC chair non-strictly indep. −2.269* −4.201**
(−1.947) (−2.291)
% non-strictly indep. x C1 0.009
(0.095)
% non-strictly indep. NRC × C1 0.026
(0.498)
% non-strictly indep. AC × C1 0.016
(0.266)
NRC chair non-strictly indep. × C1 0.021
(0.838)
AC chair non-strictly indep. × C1 0.060
(1.532)
Year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Firm fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 677 580 677 580 677 677 580 677 580 677
R2 0.121 0.135 0.121 0.139 0.132 0.121 0.135 0.121 0.140 0.137
Adjusted R2 0.094 0.104 0.095 0.108 0.106 0.093 0.103 0.093 0.107 0.109

The remaining board and corporate governance control variables are those of the performance model (Eq. (2)), except that we also
include the tenure of executive directors as a determinant of their compensation.11
The results shown in Table 7 refer to 535 firm/year observations instead of the initial 752 because of the existence of firms
with no executives on their board (46 observations), missing lagged stock market data because of new listings, mergers and
acquisitions (132 observations), and several firms that do not report executives' compensation in the Annual Report of Corporate
Governance (39 observations). When referring to the nomination and remuneration committee, the final sample is reduced to
482 firm/year observations because of firms that do not possess this particular committee. The empirical models are estimated
with year and firm fixed effects and report Huber–White robust t statistics.
The results do not detect a relation between non-strictly independent directors and executive compensation, even in the
nomination and remuneration committee. The interaction between the misclassification measures and the ownership structure
also generates no statistically significant results. Models with the lagged dependent variable as an additional control variable to
address potential endogeneity or estimations dropping corporate governance controls except non-strictly independents
measures do not bring any significance to the non-strictly independent measures.12
Therefore, we must conclude that no malfunction is detected in the full board of directors and on the nomination and
remuneration committee regarding executive compensation when non-strictly independent directors are present.13

11
There may be significant differences between CEO compensation and other executive compensation, and the number of executive directors may distort
mechanically the average compensation of executive directors, our dependent variable. The number of executive directors may control for this mechanical effect,
and it is in our model through the board size and the percentage of executive directors. Furthermore, the introduction of the number of executive directors as an
additional explanatory variable does not affect the results. Results are available on request.
12
These estimations are available upon request. Unreported results also show that non-strictly independent directors in the audit committee are also not
statistically related to executive directors' compensation. In Table 7 we present the results without the lagged dependent variable, as an additional control,
because there is no data on compensation before 2004 and we lose one year of observations.
13
Control variables' coefficients are omitted in Table 7. Only firm size (positive coefficient), the ownership of the largest shareholder (positive coefficient), and
the percentage of executive directors (negative coefficient) are statistically significant (5%, 1% and 1% of statistical significance respectively). As expected,
compensation is higher in larger firms, the control of large shareholders, with their involvement in management, may explain the positive effect of ownership in
compensation, and the negative relation between the percentage of executive directors and their compensation is explained as follows; the larger the number of
executive directors is, the lower their average compensation is.
128 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

Table 7
Executives' compensation and non-strictly independents. Firm fixed effects panel data estimation. Dependent variable: log of average individual executive board
members' remuneration. The key explanatory variables are the percentage of non-strictly independents on the board, in the nomination and remuneration
committee (NRC), and the chairing of the NRC by a non-strictly independent. Control variables (omitted for space considerations) are one period lagged log of
sales, stock return, and of return on assets (ROA), the average of the market to book ratio in the previous three years, the variance of ROA and the variance of stock
return for the three previous years. Corporate governance control variables (omitted) are the percentage of shares owned by executives, by non-executives, and
by the largest shareholder of the firm, a dummy variable identifying whether the CEO is the chairman of the board of directors, the board size, the percentage of
executives on the board, the average tenure of executive directors, a dummy variable identifying whether directors have done commercial transactions with the
firm, the percentage of busy non-executive directors (a director is busy if she holds a position in three or more boards of directors), the percentage of interlocked
executive directors (those who are also a non-executive director in a firm where a non-executive director of the first firm is an executive director), a dummy
variable identifying where there are golden parachutes protecting top executives against dismissal, and a dummy variable identifying firms where there are
voting caps (a maximum number of votes that a shareholder may exercise independently of the number of shares she has). Year dummy variables are introduced.
Huber (1967) and White (1980, 1982) robust t statistics are in parenthesis. *** denotes significance at the 1% level; ** denotes significance at the 5% level;
* denotes significance at the 10% level.

(1) (2) (3) (4) (5) (6)

% non-strictly indep. 0.406 0.444

(1.62) (1.351)
% non-strictly indep. NRC 0.296 0.259
(1.375) (1.09)
NRC chair non-strictly indep. 0.117 0.141
(1.21) (1.267)
% non-strictly indep. × C1 −0.002
(−0.177)
% non-strictly indep.NRC × C1 0.001
(0.233)
NRC chair non-strictly indep. × C1 −0.0008
(−0.316)
Year fixed effects Yes Yes Yes Yes Yes Yes
Firm fixed effects Yes Yes Yes Yes Yes Yes
Observations 535 482 482 535 482 482
R2 0.2478 0.2646 0.2609 0.2478 0.2647 0.2611
Adjusted R2 0.2124 0.226 0.2221 0.2109 0.2244 0.2205

6.2.2. CEO turnover

The evidence regarding the use of independent and outside directors as a method of increasing turnover-sensitivity to past
performance is mixed, as indicated by Franks et al. (2001). CEO turnover and board independence are treated as substitute
mechanisms in Easterwood et al. (2012) subsequent to firm underperformance. The Hermalin and Weisbach (1998) model shows
that the CEO's role in selecting board members decreases discipline when firms underperform. This evidence allows us to indicate
that non-strict independence in the boardroom might decrease the sensitivity of CEO turnover for underperforming firms,
reflecting an uncontrolled agency problem.
Because of the lack of standardized Annual Reports of Corporate Governance before 2004, we can only detect CEO turnover
events from 2005 to 2009, resulting in a database with 634 firm/year observations,14 14% of them corresponding to CEO turnover
events (89 observations). This number is observed to increase over time, from 14 in 2005 to 20 in 2009, and is regularly
distributed across industries. The CEO turnover of 4 events overlaps with a merger or acquisition.
We analyze the relationship between the presence of non-strictly independent directors and CEO turnover events with an
empirical logit model in which the dependent variable identifies the turnover events. As in Hwang and Kim (2009) the
explanatory variables are one period lagged stock return, a group of corporate governance variables lagged one period to proxy
any uncontrolled agency problem, and the interaction with the lagged stock return:



Turnoveri;t ¼ α þ α 1  Ret i;t−1 þ α 2  NSIDi;t−1 þ α 3  Ret i;t−1 × NSIDi;t−1 þ X i;t−1  β1 þ Ret i;t−1 × X i;t−1  β2 þ εi;t ð4Þ

where Ret is stock return, NSID is the non-strictly independent directors measure, and X is the set corporate governance control
variables, the same as in the performance model (Eq. (2)). 15
Given that explanatory variables are lagged one period, 77 observations are lost because of missing past stock returns in new
listings and as a result of mergers and acquisitions, leaving a final sample of 557 observations. The firms resulting from mergers
and acquisitions are analyzed as new entities. Finally, firms with no nomination and remuneration committee generate 88
additional missing observations in the estimation in which this committee is relevant.

14
The Spanish standardized Annual Report of Corporate Governance does not explicitly identify the CEO. Our identification procedure includes five steps. First,
it is when a firm declares that the top executive of the firm chairs the board (441 identifications). Second, for firms without chair-CEO duality, the CEO is
identified as the “Consejero Delegado” among board directors (214 CEOs identified). Third, if no CEO is still identified, the figure of “director general” among the
list of non-director top executives is taken as CEO (41 CEOs). Fourth, if the CEO is still not identified the top executive director is selected (49 CEOs). In seven
observations (year/firm) no CEO is still identified; however, the CEO of the posterior year is the same as the CEO of the previous year, and she/he is identified as
the CEO.
15
The tenure of executive directors is not included because several firms (46 observations) do not have executive directors.
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 129

Table 8 shows the estimation of the CEO turnover model with GEE panel data logit estimations to account for any unobserved
persistence in the residuals, with robust (Huber–White) standard errors, and time and industrial sector fixed effects. Models
account for the percentage of non–strictly independent directors over board size, the percentage of non-strictly independent
directors in the nomination and remuneration committee, and a dummy variable for the non-strictly independents chairing this
committee. Corporate governance control variables are omitted in Table 8 to make it shorter; the complete results are available on
request.
The results show the expected negative relation between past performance and the probability of a CEO turnover event
(see also Fiordelisi and Ricci, 2014; Liu, 2014 in this special issue), and no statistically significant relation between non-strictly
independent directors and CEO turnover events. This non-significant relation remains when panel data random effects logit
models with bootstrap standard errors are estimated, not reported for brevity. The unreported results in Table 8 show that CEO
turnover events are less probable when the CEO chairs the board of directors (1% statistical significance in models 1 and 3, and 5%
in model 2) and the higher the proportion of executive directors is (10% statistical significance in models 2 and 3). Finally, the
higher probability of CEO turnover after firm underperformance is reduced the larger the size of the board of directors is
(5% statistical significance in model 1, 10% in models 2 and 3), consistent with monitoring coordination problems.16

6.3. Outcomes of the audit committee

6.3.1. Audit qualifications

There is evidence that board and audit committee independence affects the control of managers through the quality of
accounting information. Anderson et al. (2004) suggest that more independent audit committees provide a measurable and
significant benefit to firms because these committees provide more reliable accounting information and affect the cost of debt.
Knechel and Willekens (2006) argue that independent board members apply pressure to enhance the external audit function.
Independent board members may be more concerned about their personal exposure if managers misbehave. Therefore, these
members are more interested in an extensive audit activity to minimize the risk of managerial misbehavior, which could affect
the members' personal reliability. For firms with non-strictly independent directors, the pressure on reliable accounting
information is expected to be lower compared to boards without misclassified independents.
The external auditor opinions in the annual financial report, available for the audited firms, allow us to analyze whether the
presence of non-strictly independents, particularly in the audit committee, is related to the probability of a qualified opinion (not
clean). We follow the empirical model of audit qualification by Sánchez-Ballesta and García-Meca (2005) tested on a Spanish
sample from 1999 to 2002:


Audit Qualificationi;t ¼ α þ α 1  NSIDi;t þ α 2  C1i;t ×NSIDi;t þ X i;t  β þ ε i;t ð5Þ

where Audit Qualificationi,t is a dummy variable equal 1 if there is a qualified opinion, NSIDi,t is the non-strictly independent
directors measure, and Xi,t is the set of control variables. The accounting literature shows that several characteristics influence the
auditor opinion. Firm size in our paper is measured as the log of sales. Larger firms are more likely to receive a qualified opinion
because the opportunity cost of auditors increases if a misstatement is detected. Although larger firms tend to have more
developed internal financial control systems, and auditors may hesitate to issue a qualified opinion because of a potential loss of
significant fees. High leverage (long-term debt over total assets) as a proxy of company risk makes the auditors stricter, expecting
a positive relation with audit qualifications. Operating performance (ROA) and liquidity (current assets over current liabilities)
proxy the probability of firm failure and are expected to be negatively related with audit qualifications. Lower performance and
liquidity make the auditors stricter to avoid litigation costs in case of firm failure. Finally, in line with Sánchez-Ballesta and
García-Meca (2005), we include ownership structure (C1 and the ownership of executive and nonexecutive directors) and the
board of directors' coordination problems (board size) as additional control variables. When owners care about the auditor
qualified opinion, the expected consequences are an effective monitoring over management.
We eliminate banks from the sample because of the special regulations in terms of financial reporting (78 observations). In the
remaining 674 observations, we find 79 audit qualifications. This number increases from 10 in 2004 to 18 in 2009 and similarly
affects all industrial sectors. The low frequency of audit qualifications and its persistence over time reduces the power of standard
econometric explanatory techniques. Therefore, following accounting literature standards, we estimate the logit empirical model
in a reduced sample of audit qualifications with a matching sample (see also Dopuch et al., 1987). For each firm, we use the first
audit qualification (31 observations). The matching sample selection is among firms with no audit qualification in the same
industry and year with the closest amount of sales. A firm enters into the matching sample just one time.
Table 9 shows the estimation of the empirical logit model of audit qualifications on this reduced sample with robust
(Huber–White) standard errors and year and industrial sector fixed effects.17
Models 1 to 3 in Table 9 measure non-strictly independence as the percentage of non-strictly independent directors over
board size, the percentage of non-strictly independent directors in the audit committee over its size, and a dummy variable to
identify firms in which the chair of the audit committee is a non-strictly independent director. The results show no statistically

16
Non-tabulated results also show no statistically significant relation between non strictly-independents in the audit committee and CEO turnover events.
Available on request.
17
No clustering is introduced to compute the robust standard errors and no panel data techniques are used because there is just one observation per firm in this
reduced sample.
130 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

Table 8
CEO turnover and independents misclassification. GEE panel data logit models allowing persistence in the residuals with Huber (1967) and White (1980, 1982)
robust t statistics (in parenthesis), where the dependent variable is a dummy variable to identify CEO turnover events. The key explanatory variables are one
period lagged stock return, one period lagged percentage of non-strictly independents on the board of directors, in the nomination and remuneration committee
(NRC), a dummy variable to identify firms where the chair of the previous committee is a misclassified independent director, and its interaction with the lagged
stock return. Control variables (omitted to save space) are lagged one period; percentage of shares owned by executives, by non-executives, and by the highest
shareholder of the firm, a dummy variable identifying whether the CEO is also the chairman of the board of directors, the size of this board, the percentage of
executives in this board, a dummy variable identifying whether directors have done commercial transactions with the firm, the percentage of busy non-executive
directors (a director is busy if she holds a position in three or more boards of directors), the percentage of interlocked executive directors (those who are also a
non-executive director in a firm where a non-executive director of the first firm is an executive director), a dummy variable identifying where there are golden
parachutes protecting top executives against dismissal, and a dummy variable identifying firms where there are voting caps (a maximum number of votes that a
shareholder may exercise independently of the number of shares she has), and the interaction between these variables and one period lagged stock return.
Finally, industrial sector and year dummy variables are introduced. Chi2 is a Wald test of the statistical significance of all the explanatory variables. *** denotes
significance at the 1% level; ** denotes significance at the 5% level; * denotes significance at the 10% level.

(1) (2) (3)

Stock returnt − 1 −4.569*** −3.644*** −3.645***

(−2.789) (−2.262) (−2.245)
% non-strictly indep. t−1 0.3715
(0.446)
% non-strictly indep. t−1 × Stock returnt − 1 1.8175
(0.944)
% non-strictly indep.NRC t−1 0.448
(0.804)
% non-strictly indep.NRC t−1 × Stock returnt − 1 0.312
(0.280)
NRC chair non-strictly independent t−1 0.1448
(0.445)
NRC chair non-strictly independent t−1 × Stock returnt − 1 0.0002
(0.000)
Year fixed effects Yes Yes Yes
Industry fixed effects Yes Yes Yes
Observations 557 469 469
Chi2 70.026*** 61.190*** 61.312***

significant relationship among non-strict independence and audit qualifications. In models 4 to 6 of Table 9, the interaction term
between the independents' misclassification measures and the ownership of the largest shareholder also show no statistically
significant relations. Control variables are omitted in Table 9, where the measures of operating performance and liquidity are
statistically significant (both at 5% of confidence level) and behave as expected: the lower the operating performance and the
liquidity (current assets over current liabilities) are, the higher the firm failure risk is, and the higher the probability of financial
statements audit qualification is.
Additional estimations to check the robustness of the previous findings include a logit model in the full sample (674
observations) with panel data random effects. In addition, a set of GEE panel data logit models have been tested to account for any
unobserved persistence in the residuals, with no statistically significant relation found between non-strictly independents and
audit qualifications.18

6.3.2. Earnings management

Klein (2002) suggests that boards structured to be more independent are more effective in monitoring the corporate financial
accounting process. This assertion is based on evidence that reductions in board or audit committee independence are
accompanied by large increases in abnormal accruals. Xie et al. (2003) find a particularly relevant role of the audit committee in
the propensity of managers to engage in earnings management. Therefore, if the presence of non-strictly independent directors is
the result of an uncontrolled agency problem, we may expect more earnings management, particularly with non-strictly
independents on the audit committee.
Earnings management may be indicative of information provision or information manipulation. Some previous accounting studies
used discretionary accruals models to decompose the proxies of earnings management into an information component and a
manipulation component (Jones, 1991). Given the criticism of these models (e.g., Dechow et al., 1995; Guay et al., 1996), however,
recent accounting literature introduced a regression approach to decompose the informative component of earnings management
from the manipulative component (Gopalan and Jayaraman, 2012). We follow this approach to detect the manipulation component
of earnings management through the fundamental determinants of the informative portion, as differences in firms' underlying
business process or in the length of the operating cycle (Dechow, 1994). We estimate the following empirical model:


Earnings Management Measurei;t ¼ α þ α 1  NSIDi;t þ α 2  C1i;t × NSIDi;t þ X i;t  β þ εi;t ð6Þ

18
Statistical inference obtained with bootstrap standard errors. Additionally, in the full and in the reduced sample, we also estimate the logit model only with
the misclassified independents measures as explanatory variables and no statistically significant results are found.
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 131

Table 9
Audit qualifications and independents' misclassification. Logit models with Huber (1967) and White (1980, 1982) robust t statistics (in parenthesis), where
dependent variable is a dummy variable to identify a qualified opinion of the external auditor in the annual financial report. The key explanatory variables are the
percentage of non-strictly independents on the board of directors, on the audit committee (AC), a dummy variable to identify firms where the chair of the
previous committee is a misclassified independent director, and its interaction with the ownership of the largest shareholder. Control variables (omitted to save
space) are the ownership of the largest shareholder, return on assets, leverage (long term debt over total assets), the log of sales, the size of the board of directors,
the ownership of executive directors and of non executive directors, and Liquidity (current assets over current liabilities). Finally, industrial sector and year
dummy variables are introduced. All models are estimated with a sample of 31 first audit qualifications with a matching sample by year, industrial sector, and
sales. Chi2 is a Wald test of the statistical significance of all the explanatory variables. *** denotes significance at the 1% level; ** denotes significance at the 5%
level; * denotes significance at the 10% level.

(1) (2) (3) (4) (5) (6)

% non-strictly indep. −1.716 0.647

(−0.688) (0.194)
% non-strictly indep.AC −0.621 1.071
(−0.495) (0.554)
AC chair non-strictly indep. −0.426 0.398
(−0.498) (0.302)
% non-strictly indep. × C1 −0.083
(−1.158)
% non-strictly indep.AC × C1 −0.053
(−1.188)
AC chair non-strictly indep. × C1 −0.024
(−0.883)
Year fixed effects Yes Yes Yes Yes Yes Yes
Industry fixed effects Yes Yes Yes Yes Yes Yes
Observations 62 62 62 62 62 62
Chi2 18.512 19.090 17.220 17.666 17.236 16.540

were X is the set of proxies of the determinants of the informative portion of earnings management. Following Gopalan and
Jayaraman (2012) and Leuz et al. (2003), earnings management is defined as the absolute value of accruals (ACC) divided by the
absolute value of cash flows from operations. The use of accruals is interpreted as managerial discretion to conceal true economic
performance from outsiders (Healey and Wahlen, 1999). Cash flow from operations is defined as operating income minus
accruals. Accruals are calculated as follows:
h i h i
ACCi;t ¼ ΔCAi;t −ΔCashi;t − ΔCLi;t −ΔSTDi;t −Depi;t ð7Þ

where ΔCAi,t is the change in total current assets for firm i in year t, ΔCashi,t is the change in cash, ΔCLi,t is the change in total
current liabilities, ΔSTDi,t is the change in debt included in current liabilities, and Depi,t is the depreciation and amortization
expense.
These accruals are naturally affected (with no manipulation of information) by differences in firms’ operating cycle, the credit
of suppliers, the volatility of the operating environment, capital intensity, and profitability (Dechow, 1994; Dechow and Dichev,
2002; Hribar and Nichols, 2007). Consequently, we include the following as explanatory variables of earnings management (X in
Eq. (6)): the sum of days receivable and days inventory (operating cycle), days payable (the credit of suppliers), the variance of
ROA in the previous three years (the volatility of the operating environment), fixed assets over total assets (capital intensity), and
a dummy variable to identify firms with negative net income (profitability).19 Aligning with Gopalan and Jayaraman (2012), we
also include differences in growth opportunities (the average of the market to book ratio over the previous three years), in size
(the log of sales), and in leverage (long-term debt over total assets) as additional determinants.
From the initial sample of 752 observations, we eliminate banks (78 observations) given this sector's special regulations on
financial reporting and 89 additional observations because of missing data on stock market information for new firms and new
listings in the stock market (firms resulting from mergers and acquisitions are analyzed as a new entity), which results in a
sample of 585 observations.
Table 10 shows the estimated empirical model in which the dependent variable is the log of the earnings management
measure, and omits the coefficients of control variables. The key explanatory variables are the percentage of non-strictly
independent directors over board size, the percentage of non-strictly independents in the audit committee over its size, a dummy
variable to identify firms with a non-strictly independent chairing this committee, and the interaction of these variables with the
ownership of the largest shareholder. Controlling for the above mentioned variables, no significant relation is found between
earnings management and the presence of misclassified independent directors. Consistent with Gopalan and Jayaraman (2012)
we find that reporting losses and financial leverage are positively related with earnings management and that growth
opportunities are negatively related with earnings management.
The estimation method, pooled OLS with year and industrial sector fixed effects and robust t statistics (Huber–White)
clustered by firm, is selected to be comparable with Gopalan and Jayaraman (2012). Alternative estimations with firm fixed

19
Days receivable is computed by dividing 360 by the ratio of sales divided by the average accounts receivable, days inventory by dividing 360 by the ratio of
cost of goods sold divided by the average inventory, and days payable by dividing 360 by the ratio of the cost of goods sold divided by the average accounts
payable. Averages are computed with the previous annual financial statements.
132 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

Table 10
Earnings management and independents' misclassification. Pooled OLS regression with year and industry fixed effect models with Huber (1967) and White
(1980, 1982) robust t statistics (in parenthesis), clustering by firm (Petersen, 2009). Dependent variable: the log of absolute value of accruals over absolute value
of operating cash flow. The key explanatory variables are the percentage of non-strictly independents on the board of directors, on the audit committee (AC), a
dummy variable to identify firms where the chair of the previous committee is a misclassified independent director, and its interaction with the ownership of the
largest shareholder (C1). Control variables (omitted to save space) are a dummy variable to identify firms reporting negative net income, Operating Cycle as the
sum of days receivable and days inventory times 10−4, Days Payable multiplied by 10−4, the average of the market to book ratio in the previous three years, the
log of sales, the variance of ROA using the annual data of the three previous years times 10−2, leverage (long term debt over total assets), and Capital Intensity
(fixed assets over total assets). Year and industrial sector dummy variables are also introduced. *** denotes significance at the 1% level; ** denotes significance at
the 5% level; * denotes significance at the 10% level.

(1) (2) (3) (4) (5) (6)

% non-strictly indep. −0.064 0.129

(−0.190) (0.295)
% non-strictly indep.AC −0.017 0.069
(−0.096) (0.288)
AC chair non-strictly indep. 0.050 0.035
(0.474) (0.235)
% non-strictly indep. × C1 −0.007
(−0.72)
% non-strictly indep.AC × C1 −0.003
(−0.46)
AC chair non-strictly indep. × C1 0.000
(0.129)
Year fixed effects Yes Yes Yes Yes Yes Yes
Industry fixed effects Yes Yes Yes Yes Yes Yes
Observations 585 585 585 585 585 585
R2 0.062 0.062 0.062 0.063 0.062 0.062
Adjusted R2 0.031 0.030 0.031 0.030 0.029 0.029

effects also leave the measures of independents misclassification as non-significant to explain earnings management. Finally,
when the measure of lagged earnings management is used as an additional explanatory variable to address endogeneity, no
differences are detected. The results are available upon request.

7. Discussion

According to the optimal board independence theory, non-strictly independents in the boardroom could be explained as the
adjustment to the recommendations of regulators and shareholder advocates regarding the desirable level of independents on the
board through corporate governance codes (the declared level) when the level is higher than the optimal degree of independence
(strictly independents).
Under this theory, an endogenously determined higher power of executives or of the CEO in front of shareholders would induce a
stronger presence of non-strictly independents. Effectively, according to Hermalin and Weisbach (1998) and Boone et al. (2007), the
higher this power is, the lower the optimum level of board independence for shareholders interest is.20 The findings in Table 5
indicate that the stronger the managerial power is, the higher the probability of non-strictly independents in the boardroom is.
However, the presence of non-strictly independent directors when managers have stronger power has two explanations. First,
this scenario can indicate an uncontrolled agency problem that is harmful to the firm, thereby increasing agency costs and
negatively affecting firm performance. Alternatively, this presence can be a way to adjust the oversizing of declared independents
to the optimal independence level, hence without negative impact on the firm performance.
Our findings in Table 6 show weak evidence of the impact of non-strictly independents on the board on future operating
performance. Effectively, the proportion of non-strictly independents on the board of directors, on the nomination and
remunerations committee or on the audit committee does not affect future operating performance. Nevertheless, when
non-strictly independent directors preside over either committee, we obtain weak empirical evidence of a negative impact on
future operating performance.
The supervisory role of both committees is reinforced by regulators in the best practices codes of good governance, which
recommend that committees should be chaired by an outside independent director and have a significant proportion of
independents among their members. The supervisory role of the nomination and remuneration committee is particularly valuable
with regard to two decisions: executive compensation and CEO removal when the firm underperforms. Our results in Table 7
show that the presence of non-strictly independents in the boardroom or in the nomination and remuneration committee does
not affect board executives' compensation and also does not affect whether this committee is chaired by a non-strictly
independent director. Similarly, the usual impact of poor performance on the CEO rate of turnover is neither enhanced nor
reduced by the presence of non-strictly independents on the board, on the committee or chairing the committee, as Table 8
indicates.

20
For example, in the Hermalin and Weisbach (1998) model, positive past performance generates the CEO's power with a less independent board as an optimal
output for shareholders' interest.
 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134 133

We address the supervisory behavior of the audit committee in the following two specific actions: the level of audit
qualifications and the firms' earnings management. In Table 9, we show that none of the measures of a lack of strict independence
explain the observed level of audit qualifications. Regarding the impact on the measure of earnings management and the level of
accruals, the models from Table 10 behave as usual in the accounting literature and the measures of non-strict independence do
not affect this relationship.
Based on these findings, the appointment of non-strictly independents in the boardroom appears to comply with the required
proportion of independents suggested by the codes of good governance. Because these non-strictly independents on the main
board committees do not negatively affect decisions, this approach provides support for the optimal board independence theory.
The managerial power approach receives weak empirical support based on the fact that the chairing of these committees by a
non-strict independent director decreases operating performance.

8. Conclusion

In sum, this paper provides a set of eight criteria to determine the strictness of the independence of the declared independent
directors according to the Spanish regulation, which we demonstrate is in line with international standards on the definition of
independent directors including NYSE standards, the European Union Commission, and the UK corporate governance code. The
empirical measure of these eight criteria on a panel of listed firms in the Spanish stock market shows that the level of
misclassifications of independent directors is more than 50%. This level of non-strict independence is decreasing over time and
affects firms of all sizes and industries. Firms with more dispersed ownership appoint non-strictly independents more frequently
than firms with large significant shareholders. However, the presence of non-strictly independent directors does not affect several
relevant outputs of the main board committees in terms of monitoring and only weak empirical support is found for a negative
relation with future operating performance. Therefore, low support is found for the managerial power origin of non-strictly
independents. Stronger support is found for the optimal board independence theory, in which firms appoint non-strictly
independents to simultaneously achieve the optimal real independence level and the recommended independence level. The
large amount of independents recommended in corporate governance codes does not necessarily improve corporate governance
practices, and conversely, may provide firms with incentives to appoint non-strictly independents.

Acknowledgments

We thank the scientific committee of this special issue and the anonymous reviewer for their valuable feedback. We also
thank the participants of the 2013 EFMA conference held in Reading (UK) for helpful comments, and the Fundación de Estudios
Financieros (www.fef.es) who facilitated us several meetings with corporate board directors of listed companies. We acknowledge
the financial support of the Spanish ministry projects ECO2010-18567 and ECO2010-21393-C04-02. We are responsible for all
errors.

References

Adams, R., Ferreira, D., 2007. A theory of friendly boards. J. Financ. 62, 217–250.
Adams, R., Hermalin, B., Weisbach, M., 2008. The role of boards of directors in corporate governance: a conceptual framework and survey. NBER Working Paper
14486 (http://www.nber.org/papers/w14486).
Aggarwal, R., Samwick, A., 1999. The other side of the trade-off: the impact of risk on executive compensation. J. Polit. Econ. 107, 65–105.
Agrawal, A., Knoeber, C., 1996. Firm performance and mechanisms to control agency problems between managers and shareholders. J. Financ. Quant. Anal. 31,
377–396.
Aguilera, R., Cuervo-Cazurra, A., 2009. Codes of good governance. Corp. Gov. Int. Rev. 17, 376–387.
Anderson, R., Mansi, S., Reeb, D., 2004. Board characteristics, accounting report integrity, and the cost of debt. J. Account. Econ. 37, 315–342.
Andres, C., Fernau, E., Theissen, E., 2014. Should I stay or should I go? Former CEOs as monitors. J. Corp. Finance 28, 26–47.
Ansari, I., Goergen, M., Mira, S., 2014. The determinants of the CEO successor choice in family firms. J. Corp. Finance 28, 6–25.
Ballinger, G., 2004. Using generalized estimating equations for longitudinal data analysis. Organ. Res. Methods 7, 127–150.
Benkel, M., Mather, P., Ramsay, A., 2006. The association between corporate governance and earnings management: the role of independent directors. Corp.
Ownersh. Control. 3, 65–75.
Bertrand, M., Mullainathan, S., 2001. Are CEOs rewarded for luck? The ones without principals are. Q. J. Econ. 116, 901–932.
Bhagat, S., Black, B., 2002. The non-correlation between board independence and long-term firm performance. J. Corp. Law 27, 1231–1273.
Bhagat, S., Bolton, B., Sarbanes-Oxley, J., 2009. Governance and performance. SSRN Working Paper (http://ssrn.com/abstract=1361815).
Boone, A., Field, L., Karpoff, J., Raheja, C., 2007. The determinants of corporate board size and composition: an empirical analysis. J. Financ. Econ. 85, 66–101.
Brick, I., Palmon, O., Wald, J., 2006. CEO compensation, director compensation, and firm performance: evidence of cronyism? J. Corp. Finance 12, 403–423.
Byrd, J., Hickman, K., 1992. Do outside directors monitor managers? Evidence from tender offer bids. J. Financ. Econ. 32, 195–207.
Byrd, J., Coopermand, E., Wolfe, G., 2009. Another look at director independence. IRABF 1, 1–16.
Cicero, D., Wintoki, M., Yang, T., 2013. How do public companies adjust their board structures? J. Corp. Finance 23, 108–127.
Cohen, L., Frazzini, A., Malloy, C., 2012. Hiring cheerleaders: board appointments of “independent” directors. Manag. Sci. 58, 1039–1058.
Coles, J., Daniel, N., Naveen, L., 2008. Boards: does one size fit all? J. Financ. Econ. 87, 329–356.
Coles, J., Daniel, N., Naveen, L., 2010. Co-opted boards. SSRN Working Paper. (http://ssrn.com/abstract=1699272).
Core, J., Guay, W., 1999. The use of equity grants to manage optimal equity incentive levels. J. Account. Econ. 28, 151–184.
Core, J., Holthausen, R., Larcker, D., 1999. Corporate governance, chief executive officer compensation, and firm performance. J. Financ. Econ. 51, 371–406.
Cotter, J., Shivdasani, A., Zenner, M., 1997. Do independent directors enhance target shareholder wealth during tender offers? J. Financ. Econ. 43, 195–218.
Cyert, R., Kang, S., Kumar, P., Shah, A., 1997. Corporate governance and the level of CEO compensation. Working Paper.Carnegie Mellon University.
Dalton, D., Daily, C., Ellstrand, A., Johnson, J., 1998. Meta-analytic reviews of board composition, leadership structure, and financial performance. Strateg. Manag. J.
19, 269–290.
Dechow, P., 1994. Accounting earnings and cash flow as measures of firm performance: the role of accounting accruals. J. Account. Econ. 18, 3–42.
134 R. Crespí-Cladera, B. Pascual-Fuster / Journal of Corporate Finance 28 (2014) 116–134

Dechow, P., Dichev, I., 2002. The quality of accruals and earnings: the role of accruals estimation errors. Account. Rev. 77, 35–59.
Dechow, P., Sloan, R., Sweeney, A., 1995. Detecting earnings management. Account. Rev. 70, 193–225.
Dopuch, N., Holthausen, R., Leftwich, R., 1987. Predicting audit qualifications with financial and market variables. Account. Rev. 62, 431–454.
Duchin, R., Matsusaka, J., Ozbas, O., 2010. When are outside directors effective? J. Financ. Econ. 96, 195–214.
Dyck, A., Zingales, L., 2004. Private benefits of control: an international comparison. J. Financ. 59, 537–600.
Easterwood, J., İnce, Ö., Raheja, C., 2012. The evolution of boards and CEOs following performance declines. J. Corp. Finance 18, 727–744.
Elston, J., Goldberg, L., 2003. Executive compensation and agency costs in Germany. J. Bank. Financ. 27, 1391–1410.
Faccio, M., Lang, l., Young, L., 2001. Dividends and expropriation. Am. Econ. Rev. 91, 54–78.
Ferris, S., Yan, X., 2007. Do independent directors and chairmen matter? The role of boards of directors in mutual fund governance. J. Corp. Finance 13, 392–420.
Fich, E., Shivdasani, A., 2006. Are busy boards effective monitors? J. Financ. 61, 689–724.
Fiordelisi, F., Ricci, O., 2014. Corporate culture and CEO turnover. J. Corp. Finance 28, 66–82.
Franks, J., Mayer, C., Renneboog, L., 2001. Who disciplines the management of poorly performing companies? J. Financ. Intermed. 10, 209–248.
Gillette, A., Noe, T., Rebello, M., 2003. Corporate board composition, protocols, and voting behavior: experimental evidence. J. Financ. 58, 1997–2031.
Goergen, M., Renneboog, L., 2011. Managerial compensation. J. Corp. Finance 17, 1068–1077.
Gopalan, R., Jayaraman, S., 2012. Private control benefits and earnings management: evidence from insider controlled firms. J. Account. Res. 50, 117–157.
Gordon, J., 2007. The rise of independent directors in the United States, 1950–2005: of shareholder value and stock market prices. Columbia Law and Economics
WP Nº. 301. (http://ssrn.com/abstract=928100).
Guay, W., Kothari, S., Watts, R., 1996. A market-based evaluation of discretionary-accrual models. J. Account. Res. 34, 83–105.
Hallock, K., 1997. Reciprocally interlocking boards of directors and executive compensation. J. Financ. Quant. Anal. 32, 331–344.
Harris, M., Raviv, A., 2008. A theory of board control and size. Rev. Financ. Stud. 21, 1797–1832.
Hartzell, J., Starks, L., 2003. Institutional investors and executive compensation. J. Financ. 58, 2351–2374.
Healey, P., Wahlen, J., 1999. A review of the earnings management literature and its implications for standard setting. Account. Horiz. 13, 365–383.
Hermalin, E., Weisbach, M., 1998. Endogenously chosen boards of directors and their monitoring of the CEO. Am. Econ. Rev. 88, 96–118.
Hermalin, E., Weisbach, M., 2003. Boards of directors as an endogenously determined institution: a survey of the economic literature. Economic Policy
Review.Federal Reserve Bank of New York 7–26.
Hribar, P., Nichols, D., 2007. The use of unsigned earnings quality measures in tests of earnings management. J. Account. Res. 45, 1017–1053.
Huber, P., 1967. The behaviour of maximum likelihood estimates under non-standard conditions. Proceedings of the Fifth Berkeley Symposium on Mathematical
Statistics and Probability, 1. University of California Press, Berkeley, CA, pp. 221–233.
Hwang, B., Kim, S., 2009. It pays to have friends. J. Financ. Econ. 93, 138–158.
Jackson, S., Lopez, T., Reitenga, A., 2011. Accounting fundamentals and CEO bonus compensation. J. Account. Public Policy 27, 374–393.
Jones, J., 1991. Earnings management during import relief investigations. J. Account. Res. 29, 193–228.
Kim, K., Kitsabunnarat-Chatjuthamard, P., Nofsinger, J., 2007. Large shareholders, board independence, and minority shareholder rights: evidence from Europe.
J. Corp. Finance 13, 859–880.
Klein, A., 2002. Audit committee, board of director characteristics, and earnings management. J. Account. Econ. 33, 375–400.
Knechel, W., Willekens, M., 2006. The role of risk management and governance in determining audit demand. J. Bus. Finan. Acc. 33, 1344–1367.
Kumarand, P., Sivaramakrishnan, K., 2008. Who monitors the monitor? The effect of board independence on executive compensation and firm value. Rev. Financ.
Stud. 21, 1371–1401.
Lambert, R., Larcker, D., 1987. An analysis of the use of accounting and market measures of performance in executives compensation contracts. J. Account. Res. 25,
85–125.
Leuz, C., Nanda, D., Wysocki, P., 2003. Earnings management and investor protection: an international comparison. J. Financ. Econ. 69, 505–527.
Liang, K.Y., Zeger, S.L., 1986. Longitudinal data analysis using generalized linear models. Biometrika 73, 13–22.
Linck, J., Netter, J., Yang, T., 2008. The determinants of board structure. J. Financ. Econ. 87, 308–328.
Linck, J., Netter, J., Yang, T., 2009. The effects and unintended consequences of the Sarbanes–Oxley act on the supply and demand for directors. Rev. Financ. Stud.
22, 3287–3328.
Liu, Y., 2014. Outside options and CEO turnover: the network effect. J. Corp. Finance 28, 201–217.
Liu, Y., Wei, Z., Xie, F., 2014. Do women directors improve firm performance in China? J. Corp. Finance 28, 169–184.
Mooney, C.Z., Duval, R.D., 1993. Bootstrapping: A Nonparametric Approach to Statistical Inference. Sage, Newbury Park, CA.
Nguyen, B., Nielsen, K., 2010. The value of independent directors: evidence from sudden deaths. J. Financ. Econ. 98, 550–567.
Petersen, M., 2009. Estimating standard errors in finance panel data sets: comparing approaches. Rev. Financ. Stud. 22, 435–480.
Raheja, C., 2005. Determinants of board size and composition: a theory of corporate boards. J. Financ. Quant. Anal. 40, 283–306.
Sánchez-Ballesta, J., García-Meca, E., 2005. Audit qualifications and corporate governance in Spanish listed firms. Manag. Audit. J. 20, 725–738.
Santella, P., Drago, C., Paone, G., 2006. How independent are independent directors? The case of Italy. Mimeo. March 2006.
Santella, P., Drago, C., Paone, G., 2007. Who Cares About Director Independence? Mimeo. March 2007.
Shivdasani, A., Yermack, D., 1999. CEO involvement in the selection of new board members: an empirical analysis. J. Financ. 54, 1829–1853.
Sun, J., Cahan, S., Emanuel, D., 2009. Compensation committee governance quality, chief executive officer stock option grants, and future firm performance.
J. Bank. Financ. 33, 1507–1519.
Weir, C., Laing, D., 2000. The performance–governance relationship: the effects of Cadbury compliance on UK quoted companies. J. Manag. Gov. 4, 265–281.
White, H., 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48, 817–830.
White, H., 1982. Maximum likelihood estimation of misspecified models. Econometrica 50, 1–25.
Xie, B., Davidson III, W., DaDalt, P., 2003. Earnings management and corporate governance: the role of the board and the audit committee. J. Corp. Finance 9,
295–316.
Yermack, D., 1996. Higher market valuation of companies with a small board of directors. J. Financ. Econ. 40, 185–211.
Zeger, S., Liang, K., 1986. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 42, 121–130.