regulation and corporate board composition · enjoyable chats with siri valseth, kjell jørgensen,...
TRANSCRIPT
1
Regulation and Corporate Board Composition
PhD Dissertation
Siv Staubo
August 30, 2013
Department of Financial Economics
Norwegian Business School, BI
2
Acknowledgements
First of all, I would like to thank Øyvind Bøhren, my supervisor, for his guidance, advice,
patience throughout my time as a PhD student. His input has been invaluable for the
completion of this dissertation. Next, I am grateful to the former and present Head of
Department of Financial Economics, Dag Michalsen and Richard Priestley for making it
possible to combine my work in the department with PhD studies. I thank Øystein Strøm and
Charlotte Østergaard for valuable comments and suggestion on my pre-doc defense. I
appreciate valuable comments from my colleagues at the Norwegian Business School, in
particular I thank, Janis Berzins, Bogdan Stacescu, and Øyvind Norli. Encouraging and
enjoyable chats with Siri Valseth, Kjell Jørgensen, and Andreea Mitrache have been of great
importance throughout the years working with this dissertation. Finally, I want to thank my
wonderful family and my really good friends for their support.
3
Contents Page:
1 Introduction 7 1.1 Does mandatory gender balance work? Changing organizational form to avoid board upheaval. 8 1.2 Female directors and board independence. Evidence from boards with mandatory gender balance. 9 1.3 Determinants of board independence in a free contracting environment. 10
2 Does mandatory gender balance work? Changing organizational form to avoid board upheaval. 11 2.1 Introduction 12
2.2 Predictions 19
2.2.1 Compliance costs 20 2.2.2 Compliance benefits 22 2.2.3 Benefits regardless of the GBL 22
2.3 Data and descriptive statistics 23
2.4 Statistical tests 26
2.4.1 The base case 28 2.4.2 Robustness 29 2.4.3 Entry 31
2.5 Summary and conclusions 33
3 Female directors and board independence. Evidence from boards with mandatory gender balance. 49 3.1 Introduction 50
3.2 Predictions 55
3.2.1 Board independence 56 3.2.2 Regulatory determinants 58 3.2.3 Non-regulatory determinants 59
4
3.3 Data and summary statistics 61 3.4 Empirical methodology and base-case results 65 3.5 Robustness 67
3.5.1 Econometric techniques 67 3.5.2 Board independence 68 3.5.3 Non-regulatory determinants of board independence 69
3.6 Conclusions 71
4 Determinants of board independence in a free contracting environment. 89 4.1 Introduction 90 4.2 Theory and predictions 94 4.2.1 Measures of board independence 96 4.2.2 Determinants of board independence 97 4.3 Data, sample, and summary statistics 100 4.4 Research design, methodology, and estimation results 102 4.5 Robustness 105 4.5.1 Alternative econometric techniques 105 4.5.2 Alternative proxy for board independence 108 4.5.3 Non-linear determinants of board independence 108 4.6 Conclusion 110
5 Summary 132
5
Figures, tables, and appendices
1 Does mandatory gender balance work? Changing organizational form to avoid board upheaval. 37 1.1 Tables
1.1.1 Responding to the gender balance law by choosing organizational form 37 1.1.2 Sample size by listing status, exit behavior, and entry behavior 38 1.1.3 Characteristics of exit firms and non-exit firms 39 1.1.4 Exit propensity in Norway and in neighboring countries 40 1.1.5 The base-case estimates 41 1.1.6 Alternative estimation methods 42 1.1.7 Alternative definitions of family control 43 1.1.8 Definition of exit status 44 1.1.9 The entry decision 45
1.2 Appendices 1.2.1 Regulatory differences between limited liability firms
with alternative organizational forms 46 1.2.2 Regulation of gender balance in corporate boards across the world 47 1.2.3 The empirical variables 48
2. Female directors and board independence. Evidence from boards with mandatory gender balance. 76 2.1 Figures
2.1.1 The fraction of female directors in Norwegian firms exposed to the gender balance law 76
2.2 Tables 2.2.1 The empirical proxies 77 2.2.2 Distributional properties of key variables 78 2.2.3 Characteristics of listed and non-listed firms 79 2.2.4 Director types 80 2.2.5 Board size 81 2.2.6 Multiple directorships 82 2.2.7 Bivariate correlation coefficients between the determinants
of board independence 83
6
2.2.8 Estimates of the base-case model 84 2.2.9 Alternative econometric techniques 85 2.2.10 Alternative proxies for board independence 86 2.2.11 Alternative non-regulatory determinants of board independence 87
2.3 Appendices 2.3.1 Classifying directors as inside, grey or outside: Examples 88
3. Determinants of board independence in a free contracting environment. 117 3.1 Tables
3.1.1 The empirical variables 117 3.1.2 Distributional properties of the variables 118 3.1.3 Firm characteristics by ownership 119 3.1.4 Bivariate correlation coefficients between the determinants
of board independence 121 3.1.5 Estimates of the base-case model – Agency problem 1 122 3.1.6 Estimates of the base-case model – Agency problem 2 123 3.1.7 Estimates of the base-case model in subsamples 124 3.1.8 Alternative estimation methods – Agency problem 1 126 3.1.9 Alternative estimation methods – Agency problem 2 127 3.1.10 Alternative proxy for board independence – Agency problem 1 128 3.1.11 Estimates of the base-case model with non-linear determinants
– Agency problem 1 129
3.2 Appendices 3.2.1 Characteristics of non-listed firms and listed firms 130 3.2.2 Properties of the instrumental variable (IV) 131
7
1. Introduction
This thesis consists of three essays investigating the effects of regulations on board composition. Two regulations affecting the selection of board members have been put into effect during the last decade:
Regulation 1: The Gender Balance Law (GBL). In 2003, the Norwegian government passed a law requiring at least 40% of each gender in the board of directors of ASA firms.
Regulation 2: The Independence Code (IC). In 2006, boards of listed ASA firms were recommended by a corporate governance code from the Oslo Stock Exchange to appoint at least 50% independent directors to their boards.
The GBL is unique to Norway. Although ‘women on boards’ is a hot topic in countries across the world, Norway is the first country to establish a 40% gender quota by law. Firms that do not comply with the law will be liquidated.
The IC is a recommendation, following the principle of comply-or-explain. This recommendation is one of the codes in ‘The Norwegian code of practice for Corporate Governance’. Similar codes exist in most countries across the world. An independent director is neither a member of the firm’s management team nor family-related to members of the management team. A more detailed definition of independent directors is given in the second essay.
This thesis is in the field of corporate governance. The corporate governance structure involves laws, rules, and regulations on the distribution of rights and responsibilities among the different stakeholders in the firm. Due to several financial crises in the late 1990s and early 2000s, there has been an increased interest in the regulation of corporate governance. In particular, the composition of the corporate board has achieved extensive attention.
The first essay in the thesis investigates stockholders’ reactions to the GBL.
The second essay mainly addresses a supposedly unintended consequence of the GBL. Furthermore, this essay explores the link between the GBL and the IC.
Finally, using a sample of firms that are neither exposed to the GBL nor the IC, the third essay explores which firms will benefit and which firms will suffer if they had to comply with the IC.
The motivation for writing three essays on the regulation of board composition is that such regulations come with both costs and benefits. Therefore, regulation might be costly to some firms and beneficial to others. This possibility is analyzed in detail in the three essays. If we assume that stockholders are able to compose an optimal board for their firm, restrictions to board composition might result in non-optimal boards for some firms. This happens if a regulation makes it costly for stockholders to compose a board with the same qualities as the board had before the firm was exposed to the regulation.
8
To better understand why the GBL and the IC might have both costs and benefits, we explain the background of the two regulations.
The GBL, was proposed by a politician who believed that more value for stockholders can be created by increasing diversity in top management and on corporate boards. Other politicians argue that it makes a fairer society when firms include more women in their board room. For example, Prime Minister David Cameron recently stated that: “There is clear evidence that ending Britain’s male-dominated business culture would improve performance, and that Britain’s economic recovery is being held back by a lack of women in the boardroom” (The Guardian 2012).
The IC was first proposed in the United States, and later included in the Sarbane-Oxeley Act (SOX). Prior to SOX, boards in United States, as well as in most other countries, were dominated by insiders who were members of the management team. During the last decade, most countries have developed corporate governance codes that recommend stockholders to appoint a majority of independent directors. Following several financial scandals, policy makers suggested that enhancing the boards’ monitoring role would help prevent further scandals. Independent directors are assumed to be better at monitoring management than dependent directors. To illustrate, the European Commission’s Recommendation from October 6, 2006 states the following: ‘The role of independent non-executive directors features prominently in corporate governance codes. The presence of independent representatives on the board, capable of challenging the decisions of management, is widely considered a means of protecting the interests of stockholders and, where appropriate, other stakeholders.’
We believe that the corporate governance rationale for these opinions need to be further investigated. As far as we can judge, existing research provides no robust support for these opinions. That is, there is neither convincing theory nor convincing empirical evidence showing that more board independence unconditionally improves firm value. The reason is simply that more board independence has both benefits and costs and that the costs may outweigh the benefits. Moreover, this relationship between costs and benefits may vary from firm to firm.
1.1 Does mandatory gender balance work? Changing organizational form to avoid board upheaval
In the first essay, we study stockholders’ reactions to the gender balance law (GBL). We find that, after the GBL ruled that the firm will be liquidated unless at least 40% of each gender is present on the board, half the firms exited to an organizational form that is not exposed to the law. In Norway, as in many other countries, there are two organizational forms for limited liability firms. All firms operating in the most advanced organizational form (ASA) had to change their boards by 2008. The only way to avoid the GBL was to exit to a less advanced organizational form (AS), where gender diversity in the board room is not regulated. It is
9
reasonable to infer that the new regulation is costly to many firms, since the stockholders of half the firms decided to exit the exposed organizational form.
We also show that the costs are firm-specific. Exit is more common when the firm is non-listed, successful, small, young, has powerful owners, no dominating family owner, and few female directors. It is important to notice that listed firms have to delist if they change organizational form. That is, all listed firms have to operate in the most advanced organizational form. Consequently, if the benefits of being listed are greater than the costs of changing the board, the GBL is still costly even though the firm does not exit. Correspondingly, certain unexposed firms may hesitate to become exposed because the expected benefits of operating in the most advanced organizational form are lower than the cost of changing the board.
Overall, we find that mandatory gender balance may produce firms with either inefficient boards or inefficient organizational forms.
1.2 Female directors and board independence. Evidence from boards with mandatory gender balance.
The second essay explores whether gender quotas have other effects on the composition of corporate boards than the implied upward shift in gender diversity. We analyze the impact on board independence of the GBL that requires at least 40 percent of a firm’s directors to be of each gender. We find that the average fraction of independent directors grows by 20 percentage points after the passage of the law. This upwards shift occurs because 84 percent of the female directors are independent, while only 50 percent of the men are.
This large increase in board independence may be costly to some firms because the demand for monitoring by independent directors is firm-specific. That is, optimal board independence requires a trade-off between monitoring by independent directors and advice by dependent directors. This conflict between monitoring and advice suggests that board quality will suffer if forced gender balance pushes the board’s independence level above its optimal level.
We find that demand for an independent board is lowest in small, young, profitable, non-listed firms with few female directors and powerful stockholders. Such firms need monitoring by independent directors the least and advice by dependent directors the most. These firms are hit hardest by excessive board independence, which may be an unintended side effect of mandatory gender balance.
One may wonder whether increased board independence is driven not by the GBL, but rather by the IC, which was introduced in the middle of our sample period. This code is soft law based on the principle of comply-or-explain, recommending that half the firm’s directors be independent. However, the code applies to the listed (public) firms, but not to the non-listed (private). Hence, whereas the GBL imposes the same indirect restriction on board independence regardless of listing status, the IC restricts board independence only in listed
10
firms. We exploit this difference to separate the effects on independence stemming from the two regulations. Roughly, half the firms in the population are listed and implicitly exposed to both the GBL and the code. The other half consists of non-listed firms and is exposed only to the GBL. Therefore, the smaller the difference in growth of board independence between listed and non-listed firms, the higher the likelihood that the regulatory effect on independence is due to the GBL rather than the IC.
Our evidence shows that the impact does not come from the IC, but rather from the GBL. The GBL produces the same upward shift in board independence regardless of the firm’s listing status. That is, because the entire pool of female director talent has so few dependent candidates, one cannot select both many women and many dependent women simultaneously. Therefore, choosing a female director very often means having to choose an independent director, even though that was not the intention. 1.3 Determinants of board independence in a free contracting environment
The essay is the first to explore the demand for monitoring and advice on the board by the owners of firms that are not required by regulation to appoint independent directors. The sample of firms in this study is regulated neither by the GBL nor by the IC. Our focus is on the potential conflict between monitoring and advice and on the idea that the relative value of these two board functions varies across firms.
We explore the board’s monitoring role not just relative to the CEO, but also relative to potential conflicts between large and small stockholders. The first of these two monitoring functions is the only focus in the existing literature. This function of the board is to reduce the principal-agent problem that arises when managers exploit their control rights at the stockholders’ expense. This situation is called the first agency problem in the literature, and directors who are independent of management are supposed to be better at reducing this problem.
The board’s second monitoring function is to oversee the conflict between majority and minority stockholders, which has been called the second agency problem. Directors who are independent of influential stockholders are supposed to be better at protecting the rights of minority stockholders. As far as we are aware, we are the first to study the second monitoring function in a board independence setting.
Our evidence shows that well established, small, and profitable firms with concentrated ownership need advice from dependent directors more than monitoring of their management by independent directors. The analysis shows similar results when we investigate the demand for board independence driven by potential conflicts between large and small stockholders. Unlike earlier research, we find that female directors are just as likely to be advisors as monitors when the firm operates in a free contracting environment regarding gender balance and independence. Our results support the idea that optimal board independence is firm-specific.
11
2.
Does mandatory gender balance work? Changing organizational form to avoid board upheaval
by
Øyvind Bøhrena,b
Siv Stauboa
March 20, 2013
Abstract Norway is the first, and so far only, country to mandate a minimum fraction of female and
male directors on corporate boards. We find that after a new gender balance law surprisingly
stipulated that the firm must be liquidated unless at least 40% of its directors are of each
gender, half the firms exit to an organizational form not exposed to the law. This response
suggests that forced gender balance is costly. These costs are also firm-specific, because exit
is more common when the firm is non-listed, successful, small, young, has powerful owners,
no dominating family owner, and few female directors. These characteristics reflect high costs
of involuntary board restructuring and low costs of abandoning the exposed organizational
form. Correspondingly, certain unexposed firms hesitate to become exposed. Overall, we find
that mandatory gender balance may produce firms with either inefficient organizational forms
or inefficient boards.
Keywords: Corporate governance. Organizational form. Regulation. Boards. Gender quota JEL classification codes: G30. G38
a BI Norwegian Business School, Nydalsveien 37, N0442 Oslo, Norway. b Corresponding author. Telephone: +4746410503. Email address: [email protected].
12
1. Introduction
The choice of organizational form determines regulatory constraints on firms’ governance
system, such as stockholders’ ability to design the board, to separate cash flow rights from
voting rights, and to choose the principles for financial reporting. Therefore, a regulatory shift
may change the optimal way to organize the firm (Hansmann 1996, p. 151). This paper
analyzes how a large and unexpected shift in corporate law, with a liquidation penalty for
non-compliers, influences the firm’s choice of organizational form. In particular, we are the
first to study how a new law for mandatory gender balance in the boardroom induces firms to
exit from or not enter into the organizational form that suddenly becomes exposed to the
stricter regulation.
We find that half the initially exposed firms choose to exit, and that exit propensity is driven
by firm characteristics. This result suggests that the regulation is costly for firms in general,
more costly for some firms than for others, and that even non-exiting firms may end up with
suboptimal boards because the benefit of keeping their exposed organizational form exceeds
the inherent cost of forced gender balance. Correspondingly, our findings for the entry
decision indicate that firms choosing not to enter may keep their optimal board composition,
but fail to obtain their best organizational form. Thus, the observed changes in exit and entry
propensities do not reflect the full corporate costs of mandatory gender balance.
The Norwegian Parliament passed a regulation in 2003 requiring that at least 40% of the
firm’s directors be of each gender. Ahern and Dittmar (2012) argue that this gender balance
law (GBL) represents a massive, surprising shock to the stockholders’ ability to optimally
design their firm’s board. The authors also notice that the GBL represents a natural
experiment that allows the researcher to study the choice of corporate governance
mechanisms with less worry than usual about endogeneity problems (Adams, Hermalin, and
Weisbach 2010). Ahern and Dittmar document the magnitude of the shock by observing that
the average proportion of female directors in listed firms was about 10% when the GBL was
passed. During the next five years until the end of the transition period in 2008, firms
complying with the 40% quota replaced about one third of their male directors by females.
The number of female directorships increased by 260% (from 165 to 592 seats), while the
number of male directorships dropped by 38% (from 1,516 to 938 seats).
13
Our paper identifies firm characteristics that separate firms that chose to comply with the
GBL by making this large board restructuring, from the firms that avoided it by exiting their
current organizational form. We also consider the flip side of the exit decision by analyzing
how the GBL’s passage influenced the tendency of unexposed firms to enter the exposed form.
Table 1 shows how firms in our sample can respond to the GBL by changing or not changing
their current organizational form.
Table 1
Norwegian firms with limited liability can choose between two organizational forms called
ASA and AS, respectively.1 This dual system is dominant worldwide, although exceptions
exist in Canada, the United States, and a few other countries (Lutter 1992). The Norwegian
ASA and AS forms resemble, respectively, A/S and ApS in Denmark, the S.A. and S.A.R.L.
in France, the AG and GmbH in Germany, the AB (publ.) and AB in Sweden, and the Plc. and
Ltd. in the United Kingdom.
The GBL applies to every ASA, but to no any AS. Hence, an ASA may respond to the GBL
by keeping its current organizational form. If it does, the 40% gender quota must be filled.2
This choice response corresponds to the first row of table 1 (Stay). Alternatively, the ASA
may convert into an AS (Exit), which is the response shown in the second row. Unlike for
Stay, the Exit option allows the firm to continue having a board with the preferred gender mix.
AS firms in rows 3 and 4 are not exposed to the GBL. If the AS chooses to become an ASA
(Enter), it must meet the gender quota. Alternatively, the firm remains an AS (Do not enter)
and chooses whatever gender balance owners prefer.
Existing research has focused on firms in the first row of table 1, which are the ASA firms
that choose to remain ASA and hence comply with the GBL. The findings suggest that the
large, forced upwards shift in the demand for female directors by ASA firms made it difficult
to design post-GBL boards with pre-GBL qualities. For instance, 69% of the retained male
directors had CEO experience, compared to 31% of the entering females. The new female
1 ASA (AS) is short for allmennaksjeselskap (aksjeselskap). The dual system was introduced in 1996 to align Norwegian corporate law with legislation in the European Union. 2 The 40% quota applies only to boards with more than nine members. For smaller boards the quota is specified as a minimum number of directors per gender. There must be at least one director of each gender if the board has two or three members, at least two of each if there are four or five members, at least three of each if there are six to eight members, and at least four of each gender if the board has nine members. These thresholds imply that the quota may vary between 33% and 50% in the cross-section of compliers.
14
directors had less board experience and were on average eight years younger than their male
co-directors (Ahern and Dittmar 2012). Thus, most female directors in ASAs post-GBL differ
from their male colleagues in terms of less experience and younger age.
This difference means that although the GBL regulates only gender balance per se, the law
may effectively restrict stockholders’ ability to choose a board with desired qualities. The
reason is that such director qualities may correlate with gender. In particular, the two pools of
potential male and female directors differ considerably along dimensions that may matter for
the board’s ability to create firm value, such as work experience in general and leadership
experience in particular.
This impression of restricted board competence in ASAs after the GBL is supported by Ahern
and Dittmar. They estimate an average announcement return of –3.5% for listed firms with no
female directors when the Minister of Trade and Industry announced his plans to mandate
gender balance. The remaining firms experienced no abnormal announcement return, but
firms with no women on the board represented about three quarters of all listed firms at that
time. This result is consistent with evidence from a period the period 1989–2002, which is
before the GBL was announced. Listed firms would most likely lose value in that period if
they had voluntarily improved their boards’ gender balance (Bøhren and Strøm 2010). The
subsequent value drop when the regulatory intent was announced in 2002 shows that
stockholders did indeed expect a prospective GBL to be likely and costly. Moreover, this
value drop does not appear to be a temporary overreaction. Typically, firms that had to change
their boards the most experienced an abnormal 15% drop in their market-to-book ratio during
the five subsequent years.
The four alternative responses to the GBL, shown in table 1, suggest that the cost of gender
balance may differ across firms. First, the compliance costs may vary among ASA firms that
choose to keep their organizational form (Stay). For instance, the reported announcement
returns support the notion that boards with more female directors must sacrifice less board
competence to reach the 40% quota. Second, firms converting from ASA to AS (Exit) may
experience different exit costs depending on the firm’s listing status. This is because the GBL
applies to all ASAs regardless of whether they are listed (public) or non-listed (private).
However, only listed firms are required to be an ASA. Therefore, exit to avoid the quota
automatically triggers delisting for a listed ASA, but obviously not for a non-listed firm. Third,
15
because the GBL changes the benefit of having the exposed organizational form, the law may
not influence just the exit decision, but also may influence when an unexposed firm chooses
to become exposed (Enter or Do not enter).
To improve the understanding of how this one-size-fits-all regulation of gender balance has
heterogeneous effect across firms, we study how the GBL affects the choice of organizational
form of all exposed (ASA) and unexposed (AS) firms during nine years. This approach
provides new insight for four reasons. First, following the firms’ behavior during an extended
time turns out to be important. For instance, we find that among the ASAs that existed when
the GBL was passed in 2003 and that did not subsequently merge or go bankrupt, 51% had
chosen to exit into AS by the time the law became binding five years later. Second, including
non-listed firms is essential not just for a priori reasons, but also because the exit propensity
turns out to be much higher for non-listed firms than for listed firms.
Third, we find that the tendency to enter the ASA form does not constitute a mere mirror
image of the tendency to exit the ASA form. Thus, studying both exit and entry deepens
insight into the regulatory effect. Finally, the two existing studies on valuation effects of the
GBL report conflicting results. Ahern and Dittmar (2012) find negative valuation effects,
while Nygaard (2011) finds positive effects using a different event date and a different sample.
We avoid these ambiguities by analyzing how firms respond to the regulatory shift by
changing organizational form. This direct evidence on altered firm behavior may improve the
understanding of what forced gender balance does to different firms and why announcement
returns may vary across firms. The key is to identify how certain characteristics enable the
firm to influence the cost of the regulatory shock by either keeping or changing its
organizational form.
The GBL was implemented on January 1, 2006 with a two-year grace period.3 Among the 309
ASAs in 2002 that did not subsequently merge, fail, or exit for other reasons unrelated to the
GBL, we find that 151 firms existed in 2008. This exit behavior represents a drop of 51%.
These exiting firms chose the unexposed AS form. As shown by Appendix 1, the financial
reporting and the corporate governance mechanisms are less tightly regulated for AS than for 3 The law as passed in 2003 would have been withdrawn if the firms had voluntarily filled the gender quota by July 1, 2005. Because that did not happen, the GBL became mandatory in 2008. All firms had complied by April 2008, including the 72 firms that violated the January 2008 deadline (Nygaard 2011). The regulation states that the firm will be liquidated three months after non-compliance, although the government may abstain from liquidation if the firm is considered particularly important for society. No firm has been liquidated for non-compliance so far. A likely reason is that the alternative to fill the quota as an exposed ASA is exit into the unexposed AS.
16
ASA. For instance, an ASA must have ten times larger minimum share capital, produce
financial reports containing greater detail, and provide compensation data containing greater
specificity about its officers and directors. Unlike for most AS firms, CEO-chair duality is
illegal in ASAs, and not more than half their share capital can be non-voting.
In all, 42% of the ASAs were non-listed by year-end in 2009. Appendix 1 shows that there is
less discretion in the design of corporate governance mechanisms when the ASA is listed. For
instance, only listed ASAs are subject to comply-or-explain governance codes, flagging
requirements, and tender offer rules.
We find that unlike before the GBL and unlike in neighboring countries where firms were not
exposed to gender balance regulation, exit is much more common than entry. However, this
exit from ASA to AS primarily happens among the non-listed firms. For instance, the number
of listed ASAs in our sample increases by 11% from 2002 to 2008, while the number of non-
listed ASAs decreases by 49%. Thus, listed firms, which cannot remain listed unless they
keep the ASA form, exit much less often. Also, and unlike before the GBL, the propensity for
an AS to enter the ASA form and hence become exposed to the GBL is higher if the firm
immediately goes public upon ASA entry rather than staying private. These findings support
the argument that the GBL more often induces a change of organizational form when the
firm’s listing benefits are low.
This evidence shows that a study of exits by listed firms only would miss most of the
interesting cases. Moreover, because the change in the number of ASAs is the net of exits and
entries, both exit from ASA to AS and entry from AS to ASA must be addressed.
Regardless of listing status, we find that most firms converting are those that perform well
and have powerful owners. This finding supports the idea that independently of the GBL,
profitable firms with low agency costs benefit the least from the strictest regulatory standards
for transparency and governance. Exit is also more common among non-family firms. This
may indicate that family owners are better able than others to radically change the board’s
gender balance. Moreover, most firms that exit have few female directors, suggesting that
regulatory costs are higher the more the board must be restructured. Finally, most firms that
convert are small or young firms. This result may reflect that the compliance cost is fixed
relative to firm size, and that the cost of changing organizational form grows as the firm
matures.
17
Most of these relationships are supported by the evidence for firms converting from AS to
ASA. The exception is that unlike for exiting firms, the fraction of female directors is not a
significant predictor of conversion from AS to ASA. This result may be driven by the fact that
whereas an ASA must either meet the mandatory gender quota or exit to AS, an AS faces no
such pressure. An AS enters only if it expects the benefits will exceed the compliance costs.
Radically changing the gender mix is apparently not an important compliance-cost driver for
AS firms that voluntarily choose to enter. A possible reason is that they have had sufficient
time to ensure easy access to the pool of qualified female directors.
Our results are robust to alternative econometric techniques, to the definition of family control,
and to how we measure performance. The definition of exit matters, however. The fraction of
female directors is a strong determinant of exit if the firm is classified as an exit firm also in
the years before it actually exits. The relationship is considerably weaker if the firm is
classified as an exit firm only in the actual exit year. This result may reflect the empirical fact
that after the GBL was passed, gender balance gradually increased also in firms that
ultimately exited. Thus, ignoring the years before the ASA actually exits misses the general
trend towards more gender balance in all ASAs before 2008. In particular, the approach
misses the cost this increasing trend imposes on firms that gradually improve their gender
balance, but ultimately decide to exit.
These findings do not imply that the GBL is more costly for firms that exit than for those that
stay. The reason is that the non-exiting firms may find the cost of changing organizational
form to be even higher than the cost of complying with the GBL. Thus, abandoning the more
strongly regulated ASA form may be more burdensome than being forced to radically change
the board’s gender balance. This happens particularly often to the listed firms in our sample,
because exit implies losing the listing benefit. Correspondingly, AS firms that choose not to
enter the ASA form may still incur a GBL-related cost. This is because these firms will not
enter whenever the cost of complying with the GBL exceeds the ASA benefits that are
independent of the GBL. Examples of such benefits are easier access to financing and
stronger legal protection of minority stockholder rights.
Our paper is related to the empirical literature on the economics of corporate governance
regulation. Bushee and Leuz (2004) study the effect of stricter SEC disclosure requirements
for firms trading on the OTC Bulletin Board. They find that almost 75% of the firms either go
18
private or exit to the pink sheets market, which is not exposed to the new regulation. The exit
propensity is strongest for small, profitable firms with low leverage. Engel, Hayes, and Wang
(2004) analyze corresponding effects of the 2002 Sarbanes-Oxley Act, finding a slightly
increased tendency to go private. Small firms with high ownership concentration go private
more often than others. A study of 17 European countries shows that firms go private more
often when corporate governance codes are introduced and when minority protection is
increased. Exit is more common among small and profitable firms (Thomsen and Vinten
2007). These results are generally consistent with ours.
Ahern and Dittmar (2012) briefly address exits after the GBL in their valuation study, but
consider only listed firms. Because ownership characteristics are not included in their study,
Ahern and Dittmar ignore agency costs as a determinant of exit. This bias also applies to the
valuation study of Nygaard (2011), who makes a robustness test of whether the firm’s listing
status influences the relationship between exit and the fraction of female directors. Moreover,
both approaches are biased towards finding excessive exit because they include financial
firms. Such firms were allowed to convert from ASA to AS one year before the gender quota
became mandatory. Finally, the entry decision is not addressed.
Although Norway was the first country to regulate gender balance in the private sector,
mandating gender quotas for corporate boards is currently a hot political topic internationally.
The obvious reason is that corporate boards are strongly dominated by men. The highest
fraction of female directors in listed firms outside Norway is 27% (Sweden), and 70% of the
43 countries included in a recent survey have fewer than 10% of their board positions filled
by women (www.catalyst.org). France, Iceland, Netherlands, and Spain will implement quotas
in 2013–2016. Proposals along the same lines have recently been made in Australia, Belgium,
Canada, Italy, and the EU Commission. Gender balance rules for state-owned firms have
already been launched in Ireland, South Africa, and Switzerland.
Some of these countries consider whether to use mandatory law like in Norway or the softer
comply-or-explain system, which is the common standard in national corporate governance
codes among more than 100 countries worldwide (www.ecgi.org). Appendix 2 shows the
details. So far, however, only Norway has experience with gender quotas in the private sector,
and no other country has chosen a mandatory system with liquidation penalty. Hence, our
findings from the first country to adopt a radically new regulation on board diversity may
19
contribute to a more informed choice elsewhere. In particular, we can document effects from
a regulatory regime that is mandatory rather than voluntary, dictates the same gender balance
in all boards rather than allows for firm-specific discretion, ensures full compliance by a
liquidation penalty, and applies to listed firms as well as to some non-listed firms rather than
to all firms or just to listed firms.
The rest of the paper is organized as follows: Section 2 specifies our predictions, and section 3
presents the data and summary statistics. We explain the methodology and test the predictions
in section 4, and we summarize and conclude in section 5.
2. Predictions
The firm should transform itself from the ASA (exposed) to the AS (unexposed)
organizational form when the firm benefits from doing so. This benefit, B(Exit), has three
components:4
(1) B (Exit) = Compliance costs – Compliance benefits – Benefits regardless of the GBL
Exit is optimal if B(Exit) is positive. If negative, the best choice is to continue being an ASA
in compliance with the GBL. Section 4 deals with the entry decision, where the benefit of
entry is the negative of (1). Either way, changing organizational form requires a two-thirds
majority vote at the stockholder meeting.
If there are no market imperfections such as irrational owners or conflicts of interest between
owners and managers, the compliance benefits in the second term of (1) are zero. The only
effect of the GBL in such a case is to add a new constraint to the owners’ value maximization
problem by ruling out in an ASA any board design involving fewer than 40% of the positions
going to each gender. This added restriction will at best leave the owners’ opportunity set
unchanged.
Consequently, the GBL must be rationalized economically by its ability to reduce negative
effects on stockholder wealth of market imperfections. In the absence of such benefits, the
new regulation produces only compliance costs for owners as reflected in the first term of (1).
Consistent with this view, Ahern and Dittmar (2012) argue that lack of leadership experience 4 Similar logic is used by Engel, Hayes, and Wang (2004).
20
among female directors may be a major driver of compliance costs and the resulting loss of
firm value.
The third term in (1) explains why an ASA with low compliance benefits and high
compliance costs may still decide not to exit. This happens when the net compliance cost of
the GBL (the first two terms) is smaller than the benefit of being an ASA for reasons
unrelated to the GBL (the third term). These latter reasons are independent of board
composition and were also valid before the GBL. Examples are the benefit of having a liquid
stock if the ASA is listed and of having the option to go public without first changing
organizational form if the ASA is non-listed. Regardless of listing status, any ASA may also
benefit from regulation ensuring more transparency and stronger protection of minority
stockholders.
The composite nature of the exit benefit in (1) has two immediate implications. First, cost
measures based on the exiting firms alone will underestimate the full cost of the GBL. This
happens because regulatory costs based on just exiting firms ignore the costs for firms that
stay. The latter costs are particularly relevant for the listed firms in our sample, because they
cannot exit without simultaneously delisting. Second, firms more likely to exit are not just
those with high compliance costs and low compliance benefits. Exit to AS is also optimal for
firms with low benefits from being an ASA in the first place. For such firms, the third
component in (1) is too small to overcome even a moderate cost of the GBL.
We next hypothesize how firm characteristics will influence the three components of B(Exit)
in (1) and hence the likelihood of switching from the ASA to the AS organizational form after
the GBL.
2.1 Compliance costs
The costs of complying with the GBL consist of search costs for new directors, increased
compensation costs for these directors once hired, and reduced private benefits for owners,
who lose control because the board is restructured.
If the owners have chosen the optimal board composition before the GBL, forced board
changes, and hence the compliance costs, will be higher the fewer women the board has. This
logic is supported by Ahern and Dittmar (2012), who find that on average, only firms with no
21
female directors lose market value at the GBL announcement. We predict that the lower the
fraction of female directors, the higher the propensity to exit (H1).
If earlier top management experience matters for director quality, the findings by Ahern and
Dittmar (2012) that women have less such experience than men do imply that new qualified
directors must be drawn from a smaller pool than earlier. Thus, the GBL will increase both
search costs and compensation costs. Because these increased costs seem rather independent
of firm size, however, compliance costs will more often produce a positive exit benefit when
the firm is small. We expect that the smaller the firm, the higher the propensity to exit (H2).
Family firms often have members of the controlling family in board and CEO positions
(Anderson and Reeb 2003). To illustrate, family-controlled firms in our sample have a median
ownership concentration of 50%, a family chairperson in 38% of these cases, and a family
CEO in 30%. In contrast, non-family firms have a median ownership concentration of 26%,
and the largest owner is chairperson or CEO in 17% of the cases. The GBL may therefore
threaten the family’s ability to extract private benefits in the ASA whenever the gender mix
among the family’s director candidates does not match the mandated gender quota. Such
concerns for family-internal recruiting to the board suggest that ASAs controlled by families
convert more often to the AS form than other firms after the GBL.
Nevertheless, this concern for family-internal recruiting may not tell the full story about the
family firm’s compliance costs. The high ownership concentration and the family’s
governance involvement during extended periods suggest that family firms often have
particularly powerful and committed owners. The long and deep experience with the firm and
its environment may have enabled the family to establish a rich network with resourceful
individuals outside the firm. Therefore, the controlling owner may know the outside pool of
potential female directors particularly well. This argument suggests that unlike for the
family’s ability to recruit female directors from inside the family, it may be relatively easy to
fill the gender quota by recruiting from outside the family. Hence, compared to other firms,
family-controlled firms may exit the exposed organizational form less often rather than more
often. We define a family-controlled firm as one where ultimate owners by blood or marriage
hold more than half the equity. The two conflicting arguments imply that the expected
relationship between family control and exit propensity is unspecified (H3).
22
2.2 Compliance benefits
The hypotheses discussed so far assume that owners always know their best interest,
including the ability to establish an optimally designed board before the GBL. Allowing for
imperfections in terms of gradual learning, however, firms may need time to locate the pool of
director candidates and pick the best team. Such a limited ability to choose the optimal board
may be particularly relevant for gender mix, because boards and recruiting committees were
strongly dominated by men before the GBL (Rosener 2011). Hence, older firms with a long
learning history pre-GBL may have been closer to their value-maximizing gender balance
than were younger firms with a shorter history.
This logic suggests that older firms will be hurt by a rule mandating the same gender mix for
every firm, while younger firms may benefit from being forced to establish a more gender-
balanced board. On the other hand, older firms may find it harder to change organizational
form because they are more complex and rigid (Boone et al. 2007). This argument suggests
older firms are less rather than more prone to exit. Thus, the expected relationship between
firm age and exit propensity is unspecified (H4).
2.3 Benefits regardless of the GBL
The ASA firms in our sample are subject to tighter reporting requirements than the AS firms
are. This higher transparency of ASAs reduces the asymmetric information between old and
new stockholders, between majority and minority stockholders, and between borrowers and
lenders. Thus, being organized under the most demanding organizational form may reduce the
cost of raising outside finance. This option is more valuable the more financially constrained
the firm (Myers and Majluf 1984). Using leverage to proxy for financial constraints, we
predict that the weaker the financial constraint, the higher the propensity to exit (H5).
For similar reasons, profitable firms may suffer less after exit because they can more easily
finance growth internally. Measuring profitability as operating returns to assets after taxes
(ROA), we expect that the more profitable the firm, the higher the propensity to exit (H6).
The higher transparency of ASAs than ASs because of regulatory differences may induce less
costly and more intense monitoring by financiers, analysts, and the media. The resulting lower
information asymmetry in ASA firms is more valuable the higher the potential agency costs,
that is, the weaker the owners’ incentives and power to monitor management (Morck, Shleifer,
and Vishny 1989). Hence, an ASA with low potential agency costs has fewer governance
23
benefits from being an ASA. Moreover, these low agency costs increase the likelihood that
the firm will rationally choose organizational form according to the value-maximizing exit
criterion in (1). We relate agency costs to ownership concentration, which we measure as the
fraction of outstanding equity held by the firm’s largest ultimate owner. We hypothesize that
the higher the ownership concentration, the higher the propensity to exit (H7).
Unlike a listed ASA, a non-listed ASA does not change listing status when exiting to AS.
Thus, owners of listed firms have more to lose by not having their stock traded in a liquid
market (Bahrat and Dittmar 2006). Listed firms also have a much wider stockholder base,
which makes them more vulnerable to free-rider and coordination problems when concerted
action would benefit all stockholders as a group (Shleifer and Vishny 1986). For instance,
Norwegian listed and non-listed ASAs of similar size have on average roughly 4,000 and 10
stockholders, respectively (Bøhren 2011). We expect that non-listed firms will exit more often
than listed firms do (H8).
Summarizing predictions H1–H8, we hypothesize that a firm with the ASA organizational
form, which is exposed to the GBL, will exit more often to the unexposed AS form when the
firm has low leverage, high profitability, high ownership concentration, small size, few
female directors, and when the firm is non-listed. The expected effects on exit of firm age and
family control are left unspecified.
3. Data and descriptive statistics
The official initiatives to regulate gender balance in corporate boards were made in 1999 and
once more in 2001 through public hearings about possible overhaul of the Equal
Opportunities Act from 1978. The first public announcement of the planned regulation was
made in February 2002. The regulation was passed as corporate law by Parliament in
December 2003 and once more in June 2005, with the added provision of a liquidation
penalty for non-compliers. The transition period from the old to the new regime ended at
year-end 2007, although 77 firms were allowed to postpone compliance until the end of
February 2008.
To allow for approximately two non-event years at the beginning and end of these regulatory
events, our sample period is 2000–2009. The sample for the analysis of exits is based on the
24
population of ASA firms by year-end.5 We ignore firms that exit due to merger or bankruptcy.
Financial firms are also excluded because they had to choose the exposed form until a new
law lifted this requirement in 2007.6 Because both merging firms and financial firms may also
have left the ASA form partly because of the GBL, this sample restriction biases our tests
towards accepting the null hypothesis that the GBL has no effect on the choice of
organizational form.
Table 2 shows the number of sample firms by year-end during the sample period. The total
number of sample firms in panel A (All) represents 264 observations per year on average,
which is 53% of the population.7 This large difference between their population and the
sample suggests that our filters are important for eliminating firms that have probably not
exited because of the GBL.
The number of ASAs is largest in 2001, monotonically decreasing thereafter to a minimum in
2009. Although not shown in the table, the peak in 2001 becomes more obvious if we also
include every year from when the dual system of ASA and AS was established in 1996. The
number of ASA firms starts at 177 in 1996 and grows every year until 2001.
Table 2
Panel A also documents that the decline after 2001 only happens in the sub-sample of non-
listed firms. The number of non-listed firms drops by 56% from 2001 to 2009, while the
number of listed firms grows by 6%. This large difference suggests that if the underlying exit
and entry decisions are partially driven by the GBL’s introduction, the benefit of changing
organizational form to avoid the GBL is considerably larger for non-listed firms.
The change in the number of firms from one year to the next in panel A reflects the difference
between entering firms (from AS to ASA) and exiting firms (from ASA to AS) during the
year. Panel B shows the exits, entries, and net exits. As already documented by panel A, net
exit (exit minus entry) is generally positive and increasing. There were altogether 217 exit 5 Our data set is organized by the Centre for Corporate Governance Research (www.bi.edu/ccgr). The data on family relationships are delivered by the tax authorities (www.skatteetaten.no), while Experian (www.experian.no) has delivered the accounting data and the corporate governance data. 6 Financials are also regulated differently than other firms regarding capital structure and corporate governance. For instance, the risk-adjusted leverage of banks cannot exceed 92% according to the Basel regulation, and Norwegian banking law stipulates that no investor can own more than 10% of a bank’s equity without the government’s permission. 7 The population of ASA firms averages 482 firms per year. Excluding financial ASAs reduces this number to 340, dropping further to 264 when we also exclude ASAs that go bankrupt or become AS because of a merger.
25
firms and 146 entry firms, producing a net exit of 71 firms during the sample period. Panel C
confirms that this tendency to exit is much stronger among non-listed firms. For instance,
while panel A shows that the number of firm years is rather independent of listing status, non-
listed firms account for over four times more of the exits (175 vs. 42, respectively). In all,
12.1% of the non-listed firms exit yearly on average, while only 3.3% of the listed firms do.
Panel D shows that when an AS decides to become ASA and hence becomes exposed to the
GBL, the firm more often enters as non-listed (no IPO) than listed (IPO), where the latter
means going public in the entry year (92 firms vs. 54 firms, respectively). However, this
tendency is reversed after the passage of the GBL, when it becomes much more common to
go public directly. Whereas 21% of the entry firms go public directly up to 2003, 54% do so
subsequently. This shift in IPO propensity by entry firms suggests that although compliance
with the GBL may produce similar costs and benefits regardless of listing status, listed firms
earn more benefits that are unrelated to the GBL. Therefore, firms considering to become
exposed after the GBL increasingly find that entry does not pay off unless the firm quickly
reaps the listing benefits as well. We focus on the exits in the following, leaving discussion of
the entry decision until section 4.
The empirical variables are defined in Appendix 3. Unreported summary statistics for these
variables show that non-listed firms account for 55% of our observations, the largest
stockholder owns on average 43% of the equity, the average board has 17% of its positions
filled by female directors and 5.6 members, while 20% of the firms are majority controlled by
a family.
A firm is classified as exiting if it transforms from ASA to AS during the sample period. An
exiting firm leaves the sample the year it actually exits. The firm is called non-exiting if it
never abandons the ASA form. Table 3 shows that compared to ASA firms that stay exposed
to the GBL, the ASAs that convert to AS and hence become unexposed are different
according to most of the hypothesized determinants. For instance, exiting firms have
significantly fewer women on the board (10% vs. 20%), are younger (20 years vs. 29 years),
have higher ownership concentration (53% vs. 38%), and are more often non-listed (78% vs.
42%). Compared to our hypotheses from section 2, these univariate relationships are
consistent with H1, H4, H7, and H8, respectively.
Table 3
26
Unreported tests show that when we split the sample based on listing status rather than exit
status, a similar pattern emerges as for exit vs. non-exit firms. This similarity suggests that
whether or not the firm is listed correlates both with the exit/non-exit choice and with other
determinants of exit beyond listing status. Thus, not controlling for listing status may create a
serious omitted variable bias in a regression where exit is the dependent variable. That does
not imply, however, that any other hypothesized determinant than listing status is redundant in
an exit model. This argument is supported by the bivariate correlation coefficients, which
show that listing status does not correlate alarmingly with any other determinant.8
4. Statistical tests Table 2 shows that the number of firms exposed to the GBL has been dropping every year
since 2002, when the intention to regulate gender balance in corporate boards was announced.
Although table 2 ignores all ASA firms that become AS because of merger, bankruptcy, or
regulatory change for financials, our filtering criteria might still have failed to exclude other
exogenous exit determinants that are unrelated to the GBL. To account for this possibility, we
compare the exit propensity for ASA firms in Norway with the exit propensity from the
similar organizational form in the neighboring countries Denmark and Sweden. These two
countries do not mandate gender-balanced boards, but they have the same system of dual
organizational forms as Norway has.
We use a difference-in-difference approach to test whether the tendency to change
organizational form by Norwegian firms differs from what it is in the neighboring countries.
The event is the passage of the GBL in 2003, the event group is the Norwegian ASA firms,
and the alternative non-event groups are the firms in Denmark, Sweden, or both, with similar
organizational form as the firms in the event group. The post-event period is 2003–2009,
while the pre-event period is 1996–2002. Hence, all firms in our sample have the option to
change organizational form any time during the sample period. Only Norwegian firms may
consider exiting to avoid the GBL, however. Moreover, this can be a valid reason only in the
post-event period.
8 The listed/non-listed dummy correlates the strongest with ownership concentration. The Pearson correlation coefficient is -0.42, which is nevertheless well below the rule-of-thumb critical limit of +/- 0.8 (Greene 2007).
27
The statistic of interest is the difference-in-difference statistic D ≡ ΔNorway – ΔForeign, where
ΔNorway is the difference between the number of Norwegian firms in the post-event period and
the pre-event period, respectively. Correspondingly, ΔForeign is the difference between the
number of firms in the two periods in the foreign country (Denmark, Sweden, or both). We
estimate D by the model
(2) 0 1 2 3 ,= + + + ⋅ +it i t i t ity EG PE EG PEβ β β β ε
where yit is the number of firms in group i at time t. EGi is a dummy variable which is 1 if the
firm is in the event group and 0 if the firm is in the non-event group. Similarly, PEt is 1 if t is
in the post-event period and 0 if t is in the pre-event period.
The estimator of D is the OLS estimate of β3 in (2). This coefficient reflects the effect on the
number of firms if the observation is from Norway (rather than a neighboring country) in the
event period (rather than in the non-event period).
The number of firms in the three countries is shown in panel A of table 4. Two major factors
explain why the number of firms in Denmark is much higher than in the two other countries.
First, Norway and Sweden wrote their laws for ASA type of firms around 1995 based
primarily on an EU directive. Denmark, however, wrote its ASA law twenty years earlier
primarily based on its existing AS law. Second, both Danish laws are less restrictive than their
Norwegian and Swedish counterparts are (Gomard and Schaumburg-Müller 2011). For these
reasons, we will estimate (2) using alternatively Denmark, Sweden, and both as the non-event
group.
Table 4
Panel B shows the estimate of β3, which is negative and significant in every case. This result
reflects that the drop in the number of firms with organizational form exposed to the GBL as
observed in tables 1 and 2 is a unique Norwegian phenomenon.
The findings in table 4 strengthen the impression of an inverse relationship between the
introduction of the GBL and the choice of ASA as organizational form. We analyze this link
more closely in the following by relating the exit and entry behavior to firm characteristics as
specified in section 2. We first report the findings from the base case, followed by a series of
robustness tests. Finally, we present estimates of an entry model.
28
4.1 The base case Our base-case model is the following:
(3) 1 2 3 4
5 6 7 8
it it it it it
it it it it it
Exit Female directors Firm size Family control Firm ageFinancial constraints Performance Ownership concentration Listed u
α β β β ββ β β β
= + + + +
+ + + + +
Exitit is a dummy variable that equals 1 if firm i leaves the exposed form (ASA) during the
sample period and 0 otherwise. We estimate (3) as a logit model, using the GLM and a sample
from 2000–2009.
Table 5 shows the estimates. Consistent with our prediction based on the compliance cost
component of B(Exit) in (1), the table documents that firms with few female directors exit
more often from the exposed to the unexposed organizational form (H1). This inverse
relationship suggests that the costs of complying with the GBL are higher the more the board
must be restructured in general, and the more that male directors must be replaced by females
in particular. The finding is also in line with Ahern and Dittmar (2012), who report that firms
with no female directors lost value when plans for a GBL were announced.
Table 5
Smaller firms exit more often than other firms do. This finding supports the economies of
scale argument that compliance costs are fixed relative to firm size, such as the cost of
searching for new female directors and having to pay them higher compensation because of
short supply (H2). Moreover, firms controlled by families exit less often than other firms do.
This result is inconsistent with the logic that the family’s ability to extract private benefits is
threatened by a GBL that mandates a gender-based board composition the family cannot
match. Rather, the finding supports the argument that family owners have better access than
others to female directors who can protect the owners’ interests (H3).
Turning next to potential compliance benefits, the estimates show that younger firms exit
more often than older firms do. This result is inconsistent with the ignorant-owner argument,
but supports the notion that mature firms find it more costly to change organizational form
(H4).
The third component of B(Exit) in (1) is benefits of having the exposed organizational form
that are independent of the GBL. Table 5 shows that unlike what we predicted, the exit
decision is not significantly related to financial constraints as measured by leverage (H5). A
29
more profitable firm is more willing to leave the exposed form, however, possibly because it
can more easily self-finance investments by high earnings and can afford the higher financing
costs as the firm becomes less transparent (H6). Exit is also more common when ownership
concentration is high. This is evidence that strong owners can be a substitute for the
disciplining effect of a stricter regulatory regime (H7).
Finally, non-listed firms are more prone to exit. This result supports with the notion that listed
firms have more to lose by exiting for reasons unrelated to the GBL, such as better stock
liquidity, continuous pricing of their stock, and closer following by financial analysts and the
media (H8).
4.2 Robustness Table 6 estimates the base-case model (1) with five alternative econometric techniques. The
table documents that the results are insensitive to whether we use logit (the base case), probit,
a standard panel method with random effects, a logit panel method with random effects,
pooled OLS, or pooled OLS with standard errors adjusted for clustering at the firm level.
Thus, the choice of econometric technique is not driving the base-case results.
Table 6
Family control may be operationalized in several ways. Table 7 shows what happens when we
measure family control by other proxies than the one used in table 5, which is whether the
family holds a majority ownership stake (family firm). We alternatively measure family
control by the fraction of board seats held by the family (family board), by whether the CEO
is recruited from the family (family CEO), by the number of family members owning stock in
the firm (family size), and by whether a family member heads the board (family chair), and by
the family’s equity in the firm (family ownership). The first column of results copies the base-
case result from table 5, where we measure family control by whether the largest family by
ownership has a majority stake.
The table shows that the relationship between the exit decision and all other variables than
ownership concentration is insensitive to how we measure family control. As in the base case,
higher ownership concentration always increases the expected exit propensity, but the
statistical significance is weaker when we use family control variables that do not directly
reflect the family’s ownership rights. This result suggests that formal power at the stockholder
30
meeting is the key ownership determinant of the exit decision. Overall, table 7 reflects that the
estimates of (3) are robust to how family control is measured.
Table 7
Unreported tests use alternative empirical measures for financial constraints and for
performance. Instead of measuring financial constraints by leverage, we use annual real sales
growth during either the current, the last two, or the last three years. The estimates are
equivalent to those in table 5. We find the same result if we measure annual ROA during the
current or the last two years.
We have so far used the convention that if the firm exits at time t, it is classified as an exit
firm also before t. Table 8 repeats the base-case result from table 5 in model I, while model II
classifies the firm as exiting only in the year it actually switches from ASA to AS. The
estimates show that female directors and firm performance become insignificant determinants
in model II, and that firm size becomes positive and significant at the 4% level. The roles are
unaltered for ownership concentration, family control, listing status, and firm age.
The insignificant relationship between exit and female directors may turn up because all firms
in our sample tend to increase their use of female directors over time, regardless of whether
the firm ultimately exits. For instance, firms that exit increase the average fraction of females
on their boards before exit from 8% in 2002 to 30% in 2008. This increasing use of female
directors regardless of exit behavior means that when a firm is classified as exiting only in the
year it actually exits, it is an exit firm in our test only when its female director fraction is the
highest and hence the closest to the female fraction for non-exiting firms. This is probably
why this firm characteristic is unable to separate exiting firms from the non-exiting.
Table 8
Summarizing, we have shown that the base-case results are independent of whatever
econometric technique we use, how we define family control, how we measure return on
assets, and whether we use leverage or growth to measure financial constraints. The definition
31
of an exit firm matters, because the fraction of female directors is not a significant predictor of
exit behavior when the firm is classified as exiting only in the year it actually exits.9
4.3 Entry The entry decision is expected to be driven by almost the same firm characteristics as for the
exit decision. We specify the following model:
(4) 1 2 3 4
5 6 7 8
it it it it it
it it it it it
Entry Female directors Firm size Family control Firm ageFinancial constraints Performance Ownership concentration IPO u
α β β β ββ β β β
= + + + ++ + + + +
Entry is a dummy variable that is 1 for firms that enter the exposed organizational form and 0
otherwise. Based on the theoretical arguments for the exit decision in section 2 as specified in
hypotheses H1–H8, we predict that entry is more likely if the firm has many female directors
(β1 > 0), large size (β2 > 0), binding financial constraints (β5 > 0), low performance (β6 < 0),
and low ownership concentration (β7 < 0). Like we did for exit, we leave unspecified the
expected effects on entry of family control (β3) and firm age (β4).
Because an AS firm considering entry must necessarily be non-listed, listing status is an
irrelevant determinant. However, and as suggested by table 2, we expect the GBL will
increase the tendency of entering firms to become listed (make an IPO; go public) directly
upon entry rather than to stay non-listed in their new ASA form. The reason is that listed
firms enjoy more of the benefits of being an ASA that are independent of the GBL. Hence, the
GBL makes it relatively more attractive to be listed than non-listed once the firm is already an
ASA. The dummy variable IPO in (4) is 1 if the entering firm chooses to become listed in the
entry year and 0 otherwise. We expect a positive relationship between the propensity to enter
and the tendency to become listed upon entry (β8 > 0).
The sample is AS firms during the period 2000–2009 and ASA firms in their entry year. To
qualify as an entry candidate, the AS must have at least three board members and sales not
less than the lowest sales observed among the ASA firms that year.
The estimates of the relationship in (4) are shown in table 9, which reports the results from
two models. In model A, an AS that converts to ASA during the sample period is classified as 9 Boards in firms with more than 200 employees have one third of their directors elected by and from the employees. We find no evidence that the exit decision depends on whether the board has employee directors. Hence, even if employee directors may be more positive to the GBL than stockholders are, stockholders do not seem so be influenced by this view when making the exit decision.
32
entering every year until the year it enters, which is the firm’s last sample year. Model B uses
the alternative definition, where the firm is classified as entering only in the year it actually
enters. Unlike in the exit case, where every exit candidate already is an ASA, it seems more
appropriate for the entry case to use the definition in model B. This definition is more
appropriate because an entry candidate is not exposed to the GBL until it voluntarily chooses
to enter. Moreover, many non-entering firms that qualify for entry may not even consider
entry a relevant option. Finally, firms choosing to enter know what new regulation they must
comply with. Hence, model B seems to reflect the more reasonable definition of an entering
firm.
Table 9
The estimates are consistent with our predictions for the effect of firm size, family control,
firm age, ownership concentration, and the decision to go public directly. The negative
coefficient for financial constraints is inconsistent with the hypothesis, strengthening the
impression gathered from the exit model that financial constraints do reliably predict changes
in organizational form. More surprisingly, the fraction of female directors is not a significant
determinant of entry in model B and is even significantly negative in model A.10
This finding suggests that non-entering firms are not held back more than entering firms are
by the GBL’s requirement to replace a large portion of the male directors. This result is
apparently puzzling, given our earlier finding in table 5 that the existing gender mix in the
board is a strong determinant of exit. However, and as already discussed, exiting and entering
firms are fundamentally different in a GBL compliance sense. An ASA firm is already
exposed to the GBL and has no choice in the sense that unless it exits to AS, it must comply
with the 40% quota or accept to be liquidated. An AS, however, chooses to enter and hence
comply only if it wants to. Therefore, AS firms that voluntarily decide to become ASA do so
because their owners think the cost of filling the gender quota is small relative to the benefit
of being an ASA. Apparently, the owners also consider this cost to be independent of the
board’s current gender mix.
Finally, tables 5 and 9 jointly document that family-controlled firms are more hesitant than
are other firms both to exit from and to enter into the ASA form. This finding suggests that 10 Unreported regressions show that the fraction of female directors continues to be insignificant when we re-estimate the two models on the subsample of firms that enter after December 2005, when Parliament decided to punish non-compliers with liquidation.
33
when new regulation changes the benefit of the status quo, family firms are inclined to keep
their organizational form, whatever that is. One possible reason is that family-controlled firms
have transaction costs of organizational change that are not well accounted for by the other
independent variables in our entry and exit models.
Overall, we have shown that the radical board restructuring mandated by the GBL has strong
effects on the choice of organizational form. This evidence suggests the regulation is costly,
and estimates made by others may indicate the magnitude of this cost. First, Ahern and
Dittmar (2012) find that listed firms with no female directors experienced an abnormal price
drop of 3.5% when the intention to regulate was announced. This estimate is an upper limit
for the average GBL cost to ASAs, because it reflects the listed ASAs only, including the
three quarters of them that must restructure their boards the most if they decide not to exit.
Second, Fjesme (2012) finds an average first-day IPO return of 2.7% at the Oslo Stock
Exchange during 2000–2008. This evidence of IPO underpricing may represent the minimum
average benefit for an ASA firm of being listed instead of non-listed. Hence, 3% may be a
rough estimate of the average cost of the GBL for owners of listed firms. The average cost for
non-listed ASAs is below 3%, because they lose less when value when exiting than listed
firms do. Importantly, however, all these estimates are averages and do not necessarily apply
to an individual firm in the sample. This limitation follows from our finding that the cost and
benefit of exiting or entering the organizational form exposed to the GBL depends on a series
of firm characteristics.
5. Summary and conclusions
The findings of this paper support the idea that firms may respond to more restrictive
regulation by changing their organizational form. Such change occurs when the added cost of
the new regulatory constraint makes the firm’s current organizational form less attractive than
the best alternative. Strikingly, we find that when a new law mandates at least 40% of men
and women in Norwegian boardrooms, half the firms choose to exit into an organizational
form that is not exposed to the law. This tendency to avoid costly regulation by changing
organizational form varies systematically with firm characteristics. We find that exit is
significantly more likely when the firm is profitable, small, young, and non-listed. Exiting
firms also tend to have powerful owners, no controlling family, and few female directors.
34
Most of these characteristics also influence the decision to enter the exposed organizational
form by firms that are not exposed to the gender balance law. Even though far fewer listed
firms exit than do non-listed firms, listed firms that do not exit may nevertheless have to bear
the highest cost of the new regulation.
This evidence is consistent with theoretical predictions and existing empirics on how firms
respond to regulatory shocks. The results are also in line with earlier findings that board
composition matters for firm value, and that compulsory gender balance in the boardroom
shrinks the pool of competent directors and reduces stockholder wealth. Our evidence
supports the notion that optimal board composition and the best response to regulatory shocks
varies from firm to firm. Moreover, gender balance regulation may be less disruptive if firms
have the option to exit into organizational forms where the law does not apply.
Recent political signals indicate that the exit option we analyze in this paper may soon
disappear. In particular, gender balance in corporate boards may be made mandatory for more
than just one organizational form.11 If that happens, Norway will not just be special for being
the first and only country to mandate a massive, rapid shift in the composition of corporate
boards and to punish non-compliers with liquidation. The regulators may also decide to
eliminate the possibility firms currently have to mitigate the costs of regulatory shocks by
transforming into organizational forms that are not exposed to the law. Every other country
considering gender balance regulation seems to favor the comply-or-explain system or
considerably milder sanctions than liquidation. Such regulatory regimes would leave the
gender balance choice to the firm’s discretion and hence allow for firm heterogeneity in board
design. Our findings suggest that compared to this more flexible alternative, the mandatory
approach, and particularly one without exit options, is a costly way to the regulate gender
balance of corporate boards.
11 http://e24.no/jobb/naa-vil-regjeringen-ha-kvinner-i-alle-styrer/20060520.
35
Acknowledgements We are grateful for valuable feedback from Janis Berzins, Tore Bråthen, Ilan Cooper, Daniel Ferreira, Miguel Garcia-Cestona, Tom Kirchmaier, Espen Moen, Øyvind Norli, Charlotte Ostergaard, Luc Renneboog (the editor), Amir Sasson, R. Øystein Strøm, Danielle Zang, and from participants at BI’s Brownbag Seminar in Economics, the CBS Conference on Board Diversity and Economic Performance, London School of Economics, and the 13th Workshop on Corporate Governance and Investment at Cardiff Business School.
36
References Adams, Renée B., Benjamin E. Hermalin, and Michael S. Weisbach. 2010. "The Role of
Boards of Directors in Corporate Governance: A Conceptual Framework and Survey." Journal of Economic Literature no. 48 (1):58–107.
Ahern, Kenneth R., and Amy Dittmar. 2012. "The Changing of the Boards: The Impact on Firm Valuation of Mandated Female Board Representation." Quaterly Journal of Economics no. 127 (1):137–197.
Anderson, Ronald C., and David M. Reeb. 2003. "Founding-Family Ownership and Firm Performance: Evidence from the S&P 500." Journal of Finance no. 58 (3):1301–1328.
Bahrat, Sreedhar T, and Amy K. Dittmar. 2006. "To Be or Not to Be (Public)." SSRN eLibrary.
Boone, Audra L., Laura C. Field, Jonathan M. Karpoff, and Charu G. Raheja. 2007. "The Determinants of Corporate Board Size and Composition: An Empirical Analysis." Journal of Financial Economics no. 85 (1):66–101.
Bushee, Brian J., and Christian Leuz. 2004. "Economic Consequences of SEC Disclosure Regulation: Evidence from the OTC Bulletin Board." SSRN eLibrary.
Bøhren, Øyvind. 2011. "Eierne, Styret og Ledelsen: Corporate Governance i Norge." Bergen: Fagbokforlaget.
Bøhren, Øyvind, and R. Øystein Strøm. 2010. "Governance and Politics: Regulating Independence and Diversity in the Board Room." Journal of Business Finance & Accounting no. 37 (9–10):1281–1308.
Engel, Ellen, Rachel M. Hayes, and Xue Wang. 2004. "The Sarbanes-Oxley Act and Firms' Going-Private Decisions." SSRN eLibrary.
Fjesme, Sturla F. 2012. "Laddering in Initial Public Offering Allocations." AFA 2012 Chicago Meetings Paper. SSRN eLibrary.
Gomard, Bernhard, and Peer Schaumburg-Müller. 2011. "Kapitalselskaber." Copenhagen: Jurist- og Økonomforbundets Forlag.
Greene, William. H. 2007. "Econometric Analysis." Upper Saddle River, NJ: Prentice Hall. Hansmann, Henry. 1996, page 151. "The Ownership of Enterprise." Cambridge, MA: Harvard
University Press. Lutter, Marcus. 1992. "Die Entwicklund der GmbH in Europa und in der Welt.” In Marcus
Lutter, Peter Ulmer and Wolfgang Zollner (eds), “Festschrift 100 Jahre GmbH-Gesetz." Cologne: Verlag Dr. Otto Schmidt KG.
Morck, Randall, Andrei Shleifer, and Robert W. Vishny. 1989. "Alternative Mechanisms for Corporate Control." American Economic Review no. 79 (4):842–852.
Myers, Stewart C., and Nicholas S. Majluf. 1984. "Corporate Financing and Investment Decisions when Firms have Information that Investors Do Not Have." Journal of Financial Economics no. 13 (2):187–221.
Nygaard, Knut. 2011. "Forced Board Changes: Evidence from Norway." SSRN eLibrary. Rosener, Judy B. 2011. "The 'Terrible Truth' About Women On Corporate Boards."
ForbesWoman no. 06072011. Shleifer, Andrei, and Robert W. Vishny. 1986. "Large Shareholders and Corporate Control."
The Journal of Political Economy no. 94 (No. 3, Part 1):461–488. Thomsen, Steen, and Frederik Vinten. 2007. "Delisting in Europe and the Cost of
Governance." SSRN eLibrary.
37
Table 1: Responding to the gender balance law by choosing organizational form
Response Before gender balance law After gender balance law
Stay ASA ASAExit ASA ASEnter AS ASADo not enter AS AS
Organizational form
This table shows the relationship between alternative responses to the gender balancelaw (Stay, Exit, Enter, and Do not enter) and the inherent choice of alternativeorganizational forms (ASA and AS). The gender balance law applies to every ASA, butto no AS.
38
Table 2: Sample size by listing status, exit behavor, and entry behavior
Year All Non-listed Listed Exit Entry Net exit All Non-listed Listed All IPO No IPO
2000 313 187 126 8 33 -25 8 6 2 33 29 42001 317 190 127 14 18 -4 14 13 1 18 15 32002 309 187 122 17 9 8 17 13 4 9 6 32003 292 175 117 23 6 17 23 18 5 6 2 42004 283 166 117 19 10 9 19 13 6 10 5 52005 270 146 124 25 12 13 25 23 2 12 3 92006 263 130 133 28 21 7 28 24 4 21 11 102007 248 110 138 46 31 15 46 36 10 31 17 142008 231 95 136 22 5 17 22 16 6 5 3 22009 217 83 134 15 1 14 15 13 2 1 1 0Average 274 147 127 22 15 7 22 18 4 15 9 5Total 2,743 1,469 1,274 217 146 71 217 175 42 146 92 54
This table shows the number of sample firms by alternative classification criteria. Panel A shows the total number of ASA firms (All)and the number of such firms by listing status (Non-listed and Listed), while Panel B shows the number of ASA firms that exit to AS(Exit), the number of AS firms that enter into ASA (Enter), and the difference between the two (Net exit). Panel C splits the number ofexits from panel B into non-listed and listed firms, while panel D splits the sample of entry firms from panel B according to whether thefirm makes an initial public offering (IPO) in the entry year. ASA firms are exposed to the gender balance law, while AS firms are not.Listed firms are quoted on the Oslo Stock Exchange. The sample is all Norwegian AS firms entering the ASA form and all NorwegianASA firms that are not financials or have not exited the ASA form because of takeover or bankruptcy during the sample period.
A: Listing status B: Entry and Exit C: Exit by listing status D: Entry by IPO propensity
39
Table 3: Characteristics of exit firms and non-exit firms
Exit less Exit Non-exit Non-exit t-value (p-value)
General firm characteristics
Listed 0.221 0.582 -0.361 -
20.982 (0.000) Financial constraints 0.542 0.519 0.023 2.108 (0.035) Growth 1.659 1.847 -0.188 -1.112 (0.268) Performance 6.864 6.813 0.051 0.129 (0.893) Firm age 19.800 29.428 -9.628 -8.288 (0.000) Firm size 658.930 3,225.282 -2,566.352 -4.299 (0.000)
Ownership characteristics Ownership concentration 52.768 37.699 15.069 11.783 (0.000) Family ownership 35.132 30.754 4.378 4.078 (0.000) Inside ownership 13.934 12.042 1.892 2.144 (0.032)
Board characteristics
Female directors 0.100 0.201 -0.101 -
15.911 (0.000)
Board size 5.129 5.926 -0.797 -
10.545 (0.000)
Family characteristics Family firm 0.222 0.189 0.033 1.874 (0.062) Family size 1.847 2.109 -0.262 -5.072 (0.000) Family chair 0.200 0.221 -0.021 -1.244 (0.214) Family CEO 0.184 0.214 -0.030 -1.923 (0.055) Family board 0.134 0.124 0.010 1.639 (0.101)
N 1,100 1,900
This table compares exit firms to non-exit firms in terms of their mean value for general firm, ownership, board, and family characteristics. The difference between the two mean values, the t-value, and the p-value (in parentheses) of this difference are reported in the three right-most columns. Performance is censored at the 0.5% tail and then winzorized at the 1% and 99% tails. Financial constraints and growth are winzorized at the 1% and 99% tails. Appendix 3 defines the variables, and the sample is all Norwegian ASA firms in 2000–2009 that are not financials or have not exited to the AS form because of takeover or bankruptcy during the sample period. Unlike AS firms, ASA firms must comply with the gender balance law.
40
Table 4: Exit propensity in Norway and in neighboring countries
A: Number of firms
Year Norway Denmark Sweden Denmark and Sweden
1996 177 22,572 286 22,858 1997 226 22,898 448 23,346 1998 253 23,589 420 24,009 1999 288 24,855 460 25,315 2000 313 25,978 470 26,448 2001 317 29,164 452 29,616 2002 309 30,540 430 30,970 2003 292 31,894 420 32,314 2004 283 32,519 447 32,966 2005 270 38,021 487 38,508 2006 263 39,525 533 40,058 2007 248 41,087 595 41,682 2008 231 41,778 592 42,370 2009 217 41,280 575 41,855 Average 263 31,836 473 32,308
B: Difference-in-difference regressions
Non-event group Estimate (p-value)
Adjusted R2 N
Denmark -0.426 (0.000) 0.593 27 Sweden -0.271 (0-027) 0.494 27 Denmark and Sweden -0.424 (0.000) 0.592 27
This table compares the number of ASA firms in Norway with the number of firms in Denmark and Sweden that have a similar organizational form, but that are not subject to gender balance regulation. Panel A reports the number of firms per year, while Panel B shows estimates of the difference-in-difference equation defined in model (2) of the main text. The estimate in the first column of results reflects the difference between the number of Norwegian firms (the event group) in the event period and the non-event period relative to the corresponding difference in Denmark, Sweden, or both (the non-event group). The sample period is 1996–2009, where 1996–2002 is the pre-event period and 2003–2009 is the post-event period. We exclude Norwegian ASAs that are financials or firms that have exited the ASA form because of takeover or bankruptcy during the sample period. The Danish, Swedish, and the combined Danish and Swedish samples consist of all firms with the organizational form that is similar to the Norwegian ASA form. Sources for the Danish and Swedish data are www.dst.dk and Finbas, respectively.
41
Independent variable Prediction Estimate
Female directors (-) -3.064(0.000)
Firm size (-) -0.104(0.000)
Family control (-/+) -0.701(0.000)
Firm age (-/+) -0.011(0.000)
Financial constraints (-) 0.349(0.222)
Performance (+) 0.025(0.000)
Ownership concentration (+) 0.010(0.000)
Listed (-) -1.182(0.000)
Constant 1.868(0.000)
N 1,560LR chi2(8) 377.470Prob > chi2 (0.000)Pseudo R2 0.182
Table 5: The base-case estimates
This table shows the estimated coefficients for a logit regression of ownership, board, family, andgeneral firm characteristics on the decision to exit or not exit the organizational form exposed to thegender balance law. The relationship is specified in model (3) of the main text. The predicted signs ofthe coefficients are shown in the second column, and the p-values of the estimated coefficientsreported in the third column are stated in parentheses underneath. The dependent variable is 1 if thefirm is an exit firm and 0 otherwise. Female directors is the proportion of shareholder-elected boardmembers who are women. Firm size is the log of sales in constant 2009 millions of NOK. Family controlis a dummy variable which equals 1 if the largest family owns more than 50% of the equity and 0otherwise. Firm age is the number of years since the firm was founded. Financial constraints is totaldebt divided by total assets. Performance is the average real return on assets per year from year t-3 tot. Ownership concentration is the fraction of equity held by the largest stockholder. Listed is a dummyvariable which is 1 if the firm is quoted on the Oslo Stock Exchange and 0 otherwise. Performance iscensored at the 0.5% tail and then winzorized at the 1% and 99% tails. Financial constraints iswinzorized at the 1% and 99% tails. The sample is all Norwegian ASA firms in 2000–2009 that are notfinancials or have not exited to the AS form because of takeover or bankruptcy during the sampleperiod. Unlike AS firms, ASA firms must comply with the gender balance law.
42
Independent variable Logit Probit Standard panel Logit panel Pooled OLS Clustered OLS
Female directors -3.064 -1.809 -0.897 -9.547 -0.561 -0.561(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Firm size -0.104 -0.061 -0.026 -0.65 -0.019 -0.019(0.001) (0.001) (0.032) (0.003) (0.001) (0.039)
Family control -0.701 -0.421 -0.158 -4.387 -0.139 -0.139(0.000) (0.000) (0.028) (0.000) (0.000) (0.005)
Firm age -0.011 -0.007 -0.001 -0.493 -0.002 -0.002(0.000) (0.000) (0.223) (0.001) (0.000) (0.007)
Financial constraints 0.349 0.215 -0.022 1.054 0.058 0.058(0.222) (0.209) (0.859) (0.603) (0.273) (0.499)
Performance 0.025 0.015 0.008 0.186 0.005 0.005(0.000) (0.000) (0.009) (0.000) (0.000) (0.018)
Ownership concentration 0.01 0.006 0.003 0.351 0.002 0.002(0.000) (0.000) (0.002) (0.062) (0.000) (0.033)
Listed -1.182 -0.706 -0.162 -3.945 -0.248 -0.248(0.000) (0.000) (0.009) (0.000) (0.000) (0.000)
Constant 1.868 1.105 0.581 9.093 0.856 0.856(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
N 1,560 1,560 1,560 1,246 1,560 1,560LR chi2(8)/Wald chi2(8) 377.47 377.99 407.48Prob > chi2 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)F(t,n) 53.97 24.28Prob > F(t,n) (0.000) (0.000)Pseudo R2 0.182 0.183 0.213 0.214 0.212Random firm effects no no yes yes no no
Table 6: Alternative estimation methods
This table shows the effect of estimating the base case model (3) in the main text with alternative econometric techniques. The dependent variable is 1 if the firm is anexit firm and 0 otherwise. Logit is the base case from table 5, Probit is a probit model, Standard panel is a random effects model with exit considered a continuousvariable, Pooled OLS uses no panel method and treats exit as a continuous variable, while clustered OLS treats exit as a continuous variable and uses standard errorsadjusted for dependence between observations at the firm level. The p-values are stated in parentheses. Female directors is the proportion of shareholder-elected boardmembers who are women. Firm size is the log of sales in constant 2009 millions of NOK. Family control is a dummy variable which equals 1 if the largest family ownsmore than 50% of the equity and 0 otherwise. Firm age is the number of years since the firm was founded. Financial constraints is total debt divided by total assets.Performance is the average real return on assets per year from year t-3 to t. Ownership concentration is the fraction of equity held by the largest stockholder. Listed is adummy variable which is 1 if the firm is quoted on the Oslo Stock Exchange and 0 otherwise. Performance is censored at the 0.5% tail and then winzorized at the 1% and99% tails. Financial constraints is winzorized at the 1% and 99% tails. The sample is all Norwegian ASA firms in 2000–2009 that are not financials or have not exited tothe AS form because of takeover or bankruptcy during the sample period. Unlike AS firms, ASA firms must comply with the gender balance law.
Method
43
Family Family Family Family Family FamilyIndependent variable firm board CEO size chair ownership
Female directors -3.064 -2.956 -2.876 -2.830 -2.790 -2.841(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Firm size -0.104 -0.129 -0.125 -0.113 -0.107 -0.125(0.001) (0.000) (0.001) (0.002) (0.003) (0.001)
Family control -0.701 -1.575 -0.807 -0.195 -0.174 -0.0090.000 0.001 0.000 0.264 0.003 0.002
Firm age -0.011 -0.013 -0.012 -0.013 -0.011 -0.014(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Financial constraints 0.349 0.219 0.167 0.227 0.106 0.272(0.222) (0.514) (0.621) (0.497) (0.751) (0.417)
Performance 0.025 0.027 0.029 0.026 0.025 0.029(0.000) (0.002) (0.001) (0.002) (0.004) (0.001)
Ownership concentration 0.010 0.004 0.003 0.004 0.004 0.008(0.000) (0.112) (0.294) (0.200) (0.150) (0.015)
Listed -1.182 -1.156 -1.208 -1.097 -1.114 -1.171(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Constant 1.868 2.120 2.051 2.231 1.955 2.235(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
N 1,560 1,145 1,145 1,145 1,145 1,145LR chi2(8) 377.470 269.910 279.890 259.940 267.920 268.820Prob > chi2 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)Pseudo R2 0.182 0.179 0.186 0.173 0.178 0.179
Table 7: Alternative definitions of family control
Family control definition
This table shows estimates of model (3) in the main text under six alternative proxies for family control.The first column of results reproduces the base-case findings from table 5. The dependent variable is 1if the firm is an exit firm and 0 otherwise.The p-values are stated in parentheses. Female directors isthe proportion of shareholder-elected board members who are women. Firm size is the log of sales inconstant 2009 millions of NOK. Performance is the average real return on assets per year from year t-3 to t. Ownership concentration is the fraction of equity held by the largest stockholder. Listed is adummy variable which is 1 if the firm is quoted on the Oslo Stock Exchange and 0 otherwise.Performance is censored at the 0.5% tail and then winzorized at the 1% and 99% tails. Financialconstraints is winzorized at the 1% and 99% tails. The sample is all Norwegian ASA firms in2000–2009 that are not financials or have not exited to the AS form because of takeover or bankruptcyduring the sample period. Unlike AS firms, ASA firms must comply with the gender balance law.
44
Independent variable I II
Female directors -3.064 -0.541(0.000) (0.503)
Firm size -0.104 0.124(0.001) (0.040)
Family control -0.701 -0.651(0.000) (0.022)
Firm age -0.011 -0.014(0.000) (0.010)
Financial constraints 0.349 -0.377(0.222) (0.480)
Performance 0.025 -0.002(0.000) (0.878)
Ownership concentration 0.010 0.017(0.000) (0.000)
Listed -1.182 -1.881(0.000) (0.000)
Constant 1.868 -4.194(0.000) (0.000)
N 1,560 1,060LR chi2(8) 377.47 109.77Prob > chi2 (0.000) (0.000)Pseudo R2 0.182 0.162
Table 8: Defining exit status
This table shows the effect of defining exit in two alternative ways. Model I is thebase case from table 5, where the firm is classified as an exit firm every year until itleaves the ASA oganizational form and enters the AS form at some point during thesample period. Model II assigns exit status to the firm only in the year it becomes anAS. The dependent variable is 1 if the firm is an exit firm and 0 otherwise.The p-values are stated in parentheses. Female directors is the proportion of shareholder-elected board members who are women. Firm size is the log of sales in constant 2009millions of NOK. Family control is a dummy variable which equals 1 if the largestfamily owns more than 50% of the equity and 0 otherwise. Firm age is the number ofyears since the firm was founded. Financial constraints is total debt divided by totalassets. Performance is the average real return on assets per year from year t-3 to t.Ownership concentration is the fraction of equity held by the largest stockholder.Listed is a dummy variable which is 1 if the firm is quoted on the Oslo Stock Exchangeand 0 otherwise. Performance is censored at the 0.5% tail and then winzorized at the1% and 99% tails. Financial constraints is winzorized at the 1% and 99% tails. Thesample is all Norwegian ASA firms in 2000–2009 that are not financials or have notexited to the AS form because of takeover or bankruptcy during the sample period.Unlike AS firms, ASA firms must comply with the gender balance law.
Model
45
Independent variable Prediction A B
Female directors (+) -1.381 -0.336(0.001) (0.546)
Firm size (+) 0.226 0.178(0.000) (0.000)
Family control (-/+) -1.455 -1.705(0.000) (0.000)
Firm age (-/+) -0.130 -0.149(0.000) (0.000)
Financial constraints (+) -1.640 -2.115(0.000) (0.000)
Performance (-) -0.009 -0.003(0.206) (0.775)
Ownership concentration (-) -0.013 -0.012(0.000) (0.002)
IPO (+) 1.062 2.019(0.000) (0.000)
Constant -5.277 -5.160(0.000) (0.000)
N 126,152 126,152LR chi2(8) 813.858 484.060Prob > chi2 (0.000) (0.000)Pseudo R2 0.232 0.280
Table 9: The entry decision
Model
This table shows the estimated coefficients from a logit regression of ownership, family, and general firm characteristics on the decision of an AS firm to become an ASA and thereby become exposedto the gender balance law. The predicted signs of the coefficients are shown in the second column.Model A classifies the AS firm as an entry firm every year before entry if it becomes ASA duringthe sample period. Model B assigns entry status to the firm only the year it actually enters. Thedependent variable is 1 if the firm is an entry firm and 0 otherwise. Female directors is theproportion of shareholder-elected board members who are women. Firm size is the log of sales inconstant 2009 millions of NOK. Family control is a dummy variable which equals 1 if the largestfamily owns more than 50% of the equity and 0 otherwise. Firm age is the number of years sincethe firm was founded. Financial constraints is total debt divided by total assets. Performance is theaverage real return on assets per year from year t-3 to t. Ownership concentration is the fraction ofequity held by the largest stockholder. IPO is a dummy variable which is 1 if the firm becomeslisted in the entry year and 0 otherwise. We censor financial constraints and performance at the +/-2% tails. The p-value of the estimated coefficient's t-statistic is shown in parentheses. The sampleis AS firms that are entry candidates and ASA firms in their entry year during the period2000–2009. To qualify as an entry candidate, the AS firm must have at least three board membersand sales not smaller than the lowest sales among the ASA firms that year.
46
Appendix 1: Regulatory differences between limited liability firms with alternative organizational forms
Exposed listed (Public ASA) Exposed non-listed (Private ASA) Unexposed (AS)
1 Minimum share capital 1 million Norwegian kroner 1 million Norwegian kroner 0.1 million Norwegian kroner2 Corporate governance code
(comply-or-explain)The annual report must specify, item by item, whetherthe firm complies with the corporate governance codeof the OSE.
No corporate governance code No corporate governance code
3 CEO-chair duality Illegal Illegal Legal if share capital is below 3 millionNOK.
4 Gender balance on board At least 40% of each gender* At least 40% of each gender* No gender balance requirement
5 Non-voting shares Up to 50% of the shares can be non-voting. Up to 50% of the shares can be non-voting. No restriction on non-voting shares
6 Mandatory flagging An investor passing up and down through thethresholds of 5%, 10%, 20%, 1/3, 50%, 2/3, and 90% ofthe outstanding cash flow rights or voting rights mustnotify both the firm and the OSE no later than the nextmorning.
No flagging rule No flagging rule
7 Mandatory tender offer An investor passing the 1/3 ownership thresholdmust offer to buy all the remaining stock in the firm.
No tender offer requirement No tender offer requirement
8 Reporting of trades by corporate insiders
Insiders (the firm's officers and directors) must reporttheir trades to the OSE no later than the next morning.
Insiders must report to the board. Theinformation is not public.
No insider reporting required
9 Ownership recording The firm must report every transaction in itsoutstanding equity securities to the VPS. Thenotification must specify the identity of the buyer andseller, the exact time of the transaction, the number ofsecurities traded, and the security price.
The firm must report every transaction in itsoutstanding equity securities to the VPS. Thenotification must specify the identity of thebuyer and seller, the exact time of thetransaction, the number of securities traded,and the security price.
The firm must have a register that keeps track of every trade in the firm's stock.The information is not public.
10 Accounting rules The firm must report in detail on the executives' anddirectors' compensation, and on the stockholdings ofthe officers, directors, and their close family. The firmmust report the interest adjustment dates of theirbonds.
Less detailed reporting requirements than forlisted firms on compensation, ownership, anddebt
Less detailed reporting requirementsthan for listed firms on compensation,ownership, and debt
Regulation
This table shows regulatory differences between Norwegian limited liability firms with alternative organizational forms. An exposed firm is subject to the gender balance law, while anunexposed firm is not. OSE is the Oslo Stock Exchange. VPS is Verdipapirsentralen (The Securities Registry).
Organizational form
* The minimum fraction of each gender varies between 33% and 50%, depending on board size. The minimum is 40% when the board has at least ten members.
47
Percentage of Gender Comply or explain(ce)/Country female directors quota Regulatory activity Mandatory(m)Australia 8.4 no Firms should adopt and publicly explain a diversity policy. ce
Belgium* 7.7 yes A draft for a law was filed in December 2009 requiring boards of listed firms and certain non-listed firms to chooseat least one third of their directors from each gender. No sanction
pending
Canada 10.3 no The Canadian Board of Diversity was launched in November 2009 with a goal of improving diversity on boards,including gender diversity.
pending
France* 12.7 yes Parliament proposed a law in January 2011. The law requires a quota whereby 25% of the directors are femalewithin 3 years and 40% within 6 years for firms employing at least 500 and with revenues over 50 million EUR. Nosanction
ce
Germany* 11.2 no The German Corporate Governance Code, which applies to listed firms, has recommendations aimed at promotingmore women on boards.
ce
Iceland 3.8 yes Law passed in March 2010; quota of 40%. No sanction pending
Italy* 3.7 yes A law was passed in December 2010 requiring one third of each gender on boards of listed firms and state-ownedfirms. The law needs Senate approval and will will apply to new board nominations six months after suchapproval. No sanction
pending
Netherlands* 14.0 yes A law on gender quotas for the executive and the supervisory boards received govermental approval inDecember 2009. The law proposes at least 30% of each gender for the board of listed firms and for the boards ofnon-listed firms that meet certain financial and employment criteria. No sanction
ce
New Zealand 7.5 no The New Zealand Shareholders' Association will make board diversity one of three priorities. pendingNorway 40.1 yes A law passed in 2003 and implemented in 2008 mandates at least 40% of each gender on the board of listed firms
and certain non-listed firms.** Non-complyers are liquidated.m
Spain* 9.3 yes Parliament passed the Law of Equality in 2007, which requires listed firms to appoint females to 40%–60% of theboard positions. Firms are allowed until 2015 to comply. No penalty ***
ce
United Kingdom* 15.0 no The Corporate Governance Code recommends gender diversity on boards. The Conservative Party hasannounced that it will require that females constitute at least 50% of the candidates on the long list ofdirectorship candidates.
pending
United States 16.1 no The SEC approved a rule in December 2009 requiring disclosure of whether and how board nominationcommittees consider diversity when identifying director candidates. If the committee or the board has a diversitypolicy, the SEC rule requires disclosure of how this policy is implemented and how the nomination committee orthe board assesses the policy's effectiveness. The rule was implemented in February 2010.
ce
Appendix 2: Regulation of gender balance in corporate boards across the world
This table shows the regulatory status on gender quotas and the actual fraction of females in boards across the world as of 2012. In addition to the countries specified above, Ireland, SouthAfrica, and Switzerland have gender quotas for state-owned firms. Sources: Ahern and Dittmar (2012), Catalyst (2012), www.corpgov.deloitte.com, and www.nho.no.
* A EU draft from November 2012 proposes a 40% target for each gender by 2020 on boards of listed firms with at least 250 employees. Each members state is supposed to decide whether the quota should be mandatory and what sanctions should be used for non-compliers.** Boards with less than ten members have the quota stated as a minimum number of members per gender.*** No formal penalty will apply to non-compliers, but the government will take compliance into account when awarding state contracts to private firms.
48
Variable DefinitionExit Dummy variable which equals 1 if the firm leaves the organizational form
exposed to the gender balance law during the sample period and 0 otherwiseEntry Dummy variable which equals 1 if the firm enters the organizational form
exposed to the gender balance law during the sample period and 0 otherwise
General firm characteristicsListed Dummy variable which is 1 if the firm is listed on the Oslo Stock Exchange
and 0 otherwiseFinancial constraints Total debt divided by total assets (leverage)Growth The average percentage increase in real sales per year from year t-3 to tPerformance The average real return on assets (ROA) per year from year t-3 to tFirm age The number of years since the firm was foundedFirm size Sales in constant 2009 millions of NOK. Log transformed in regressionsIPO Dummy variable which is 1 if the firm becomes listed the same year it enters
the exposed organizational form and 0 otherwise
Ownership characteristicsOwnership concentration The fraction of equity held by the largest stockholderFamily ownership The fraction of equity held by the family with the largest equity stakeInside ownership The fraction of equity held by the officers and directors
Board characteristicsFemale directors The proportion of shareholder-elected board members who are womenBoard size The number of shareholder-elected board members
Family characteristicsFamily firm Dummy variable which equals 1 if the largest family owns more than 50% of
the equity and 0 otherwiseFamily size The number of owners in the largest family by ownershipFamily chair Dummy variable which equals 1 if the chair belongs to the largest family by
ownership and 0 otherwiseFamily CEO Dummy variable which equals 1 if the CEO belongs to the largest family by
ownership and 0 otherwiseFamily board The fraction of directors coming from the largest family by ownership
Appendix 3: The empirical variables
This table defines the variables used in the empirical analysis. The ownership characteristics are basedon ultimate (direct plus indirect) equity holdings.
49
3.
Female directors and board independence:
Evidence from boards with mandatory gender balance
by*
Øyvind Bøhren Siv Staubo
April 8, 2013
Abstract
This paper explores if gender quotas have other effects on the composition of corporate
boards than the implied upwards shift in gender diversity. We analyze the impact on board
independence of an unexpected and radical law in Norway requiring that at least 40 percent of
a firm’s directors be of each gender. Our evidence shows that this regulatory shock has strong
and firm-specific effects. The average fraction of independent directors grows by 20
percentage points. This upwards shift occurs because 84 percent of the female directors are
independent, while only 50 percent of the men are. We find that demand for an independent
board is lowest in small, young, profitable, non-listed firms with few female directors and
powerful stockholders. Such firms need monitoring by independent directors the least and
advice by dependent directors the most. These firms are hit hardest by excessive board
independence, which may be an unintended side effect of mandatory gender balance.
Keywords: corporate governance; regulation; board independence; gender balance
JEL classifications: G30; G38.
* Department of Financial Economics, BI Norwegian Business School, N-0442 Oslo, Norway. Our email addresses are [email protected] and [email protected]. We are grateful for valuable comments to an earlier draft from Renée Adams, Rosemarie Koch, Øyvind Norli, Charlotte Østergaard, and R. Øystein Strøm.
50
1. Introduction The low proportion of females on corporate boards has attracted widespread attention in many
countries from practitioners, regulators, researchers, and the media (Farrell and Hersch 2005;
McKinsey & Company 2007; Terjesen, Sealy, and Singh 2009; Langli 2011; Adams and
Kirchmaier 2013). Norway was the first country to act politically on this issue by mandating
gender balance in the boardroom, and other countries have followed suit. France, Iceland,
Netherlands, and Spain will implement gender quotas in 2013–2016, and proposals along
similar lines have recently been made in Australia, Belgium, Canada, the EU Commission,
and Italy. 12 Prime Minister David Cameron recently stated, “There is clear evidence that
ending Britain’s male-dominated business culture would improve performance, and that
Britain’s economic recovery is being held back by a lack of women in the boardroom” (The
Guardian 2012).
The Norwegian gender balance law was announced to the surprise of many in 2002, was
passed by Parliament in 2003, and became mandatory from 2008. Figure 1 shows that the
average fraction of directorships filled by females increased monotonically from 11 percent
when the law was passed to 42 percent five years later, when the 40 percent quota became
mandatory, including a liquidation penalty after three months for non-compliers.
Figure 1
This paper explores whether such a massive, involuntary shift in gender balance changes
other board characteristics than just the gender balance. In particular, we analyze the impact
of the gender balance law on board independence, which regulators consider the key
characteristic of a board with high monitoring skills (Bhagat and Black 1998; Adams,
Hermalin, and Weisbach 2010). However, the idea that board independence is always
beneficial has no support in the research literature. As has been shown both theoretically
(Adams and Ferreira 2007) and empirically (Linck, Netter, and Yang 2008; Duchin,
Matsusaka, and Ozbas 2010), optimal board independence requires a tradeoff between the
value of monitoring provided by independent (outside) directors and the value of advice
provided by dependent (inside) directors. This inherent conflict between monitoring and
advice suggests that board quality will suffer if forced gender balance pushes the board’s
independence above its optimal level.
12 Gender balance rules for the boards of state-owned firms have been implemented in Ireland, South Africa, and Switzerland.
51
We show that the Norwegian gender balance law (GBL) increases board independence quite
dramatically. Whereas the average fraction of independent directors was 46 percent when the
GBL was passed, the fraction rose to 67 percent when the gender quota became mandatory
with a non-compliance penalty five years later. Even though all firms exposed to the GBL had
to eventually fill the 40 percent quota, we document that the effect of the law on board
independence varies in the cross-section.13 The firms affected the most are those that need
independence the least, because they have low demand for the monitoring provided by
independent directors and high demand for the advice provided by dependent directors. We
find that those firms that ultimately have the most excessive board independence tend to be
small, young, private, profitable, owned by powerful stockholders, and to have had few
female directors before the quota became mandatory.
This finding means that although the GBL regulates just one board characteristic per se
(gender balance), the law affects another characteristic (independence) as well. This shift
from dependence towards independence also shifts the balance of skills away from advice
towards monitoring. This effect occurs because independence is a much more widespread
characteristic among female director candidates than among males. That is, the pool of male
director candidates and the pool of female director candidates are not equal regarding the two
fundamental director skills, which are monitoring and advice.
One may wonder whether increased board independence is driven not by the GBL, but rather
by the corporate governance code, which was introduced in the middle of our sample period.
This code is soft law based on the principle of comply-or-explain, recommending that half the
firm’s directors be independent.14 However, the code applies to the listed (public) firms, but
not to the non-listed (private). Hence, whereas the GBL imposes the same indirect restriction
on board independence regardless of listing status, the governance code restricts board
independence only in listed firms.
13 The exact 40 percent quota applies only to boards with more than nine members. The quota for smaller boards is specified as a minimum number of directors per gender. There must be at least one director of each gender if the board has two or three members, at least two if there are four or five members, at least three if there are six to eight members, and at least four of each gender if the board has nine members. These thresholds imply that the minimum fraction of each gender may vary between 33 percent and 50 percent in a cross-section of compliers. 14 The code was implemented in 2006 and states, “The majority of the stockholder-elected members of the board should be independent of the company’s executive personnel and material business contacts.” Moreover, at least two stockholder-elected directors should be independent of the main stockholder (www.nues.no). According to corporate law, one third of the board must be elected by and from the employees in firms with more than 200 employees. Employee-elected directors are probably dependent by nature.
52
We exploit this difference to separate the effects on independence stemming from the two
regulatory sources. Roughly half the firms in the population are listed and hence exposed to
both the GBL and the code. The other half consists of non-listed firms and hence is exposed
only to the GBL. Therefore, the less the growth in board independence differs between listed
and non-listed firms, the higher the likelihood that the regulatory effect on independence is
due to the GBL rather than to the code. Our evidence shows that the impact does not come
from the code, but rather from the GBL. The GBL produces the same upward shift in board
independence regardless of the firm’s listing status. Specifically, the average independence
level rises from 52 percent to 72 percent in listed firms and from 41 percent to 59 percent in
non-listed firms.
This finding also suggests that unlike what has been argued, it is not the independence
requirement in governance codes that causes the high fraction of independent female directors
in countries that have governance codes but no gender quotas. The idea behind this argument
is that if independence is easier to find among female director candidates than among males,
stockholders will rationally choose females rather than males in order to meet the code’s
independence requirement (Beecher-Monas 2007). However, we show that female directors
are as often independent in firms that are not exposed to the independence code, but only
exposed to the GBL. Hence, the high independence among female directors in countries other
than Norway may not be the outcome of stockholders who look for female directors to fill the
independence quota. Rather, the entire pool of female director talent has so few dependent
(inside) candidates that one cannot select both many women and many dependent women
simultaneously. Therefore, choosing a female director very often means having to choose an
independent director, even though that was not the intention. This problem becomes clear for
firms that try to fill the gender quota.
Stockholders could have complied with the GBL in at least two ways that would have
increased board independence less dramatically. We find that these options were not widely
used. First, just adding female directors to the existing board increases independence less than
when independent females replace dependent males. While extensive use of this option would
have increased board size by about 50 percent in our sample, this response is not what we
observe. The average board, excluding employee directors, has 5.6 members of whom about
10 percent are women at the start of the sample period. There are 5.9 members of whom about
40 percent are women at the period’s end. Nevertheless, increasing board size to mitigate the
53
independence effect may be rational. The reason is that studies in several countries have
found an inverse relationship between board size and firm performance regardless of firm size
(Yermack 1996; Eisenberg, Sundgren, and Wells 1998; Bøhren and Strøm 2010).15
A second way to mitigate excessive board independence is by recruiting females with
multiple directorships, who have been shown to be less independent (Fields and Keys 2003).
However, the only trend we find is that men hold multiple seats considerably less often after
the GBL than before. There is no clear evidence in our sample that women’s holding of
multiple directorships becomes more common.
Ahern and Dittmar (2012) find that Norwegian firms without female directors, which
represent three quarters of their sample, lost on average 3.5 percent in market value when the
plan for the GBL was announced. The authors argue that this value drop is not a temporary
overreaction. The Tobin’s Q ratio of the firms with no female directors at the GBL
announcement typically fell by 15 percent when the law was passed. These firms continued
underperforming relative to the other firms until the law was implemented five years later.
The number of female directorships increased by 260 percent (from 165 to 592 seats) during
this period, while the number of directorships held by males dropped by 38 percent (from
1,516 to 938 seats). Ahern and Dittmar also find that the incoming female directors had
considerably less leadership experience than the exiting males. While 69 percent of the
retained male directors had CEO experience, this was the case for only 31 percent of the new
female directors, who were on average also eight years younger.
Our evidence suggests that this loss of firm value partly happens because the GBL produces
excessive board independence. Most incoming female directors are more independent than
most exiting males are. This greater independence is because the incoming females generally
have less prior experience with the firm and its environment. Hence, the value drop may not
be driven by the shift in gender balance per se. Rather, the drop happens because skills that
are relevant for board work correlate with gender. According to Adams and Kirchmaier
(2013), a key reason for this correlation is that men and women with equal formal education
choose different subsequent careers. This choice matters for the ability to build director skills.
15 Coles, Daniel, and Naveen (2008) have recently challenged this result by finding a positive empirical relationship between Tobin’s Q and board size for firms with high leverage or many business segments, both of which are used to measure firm complexity.
54
Empirical tests of the Adams and Ferreira (2007) model find that, as predicted, the demand
for independent directors with monitoring skills varies from firm to firm (Linck, Netter, and
Yang 2008; Duchin, Matsusaka, and Ozbas 2010). We extend this literature by showing that a
regulatory floor on gender diversity may produce a side effect in terms of excessive board
independence relative to dependence and hence excessive monitoring relative to advice. This
distortion of optimal board design is particularly strong in firms that need monitoring the least
and advice the most. Our evidence shows that firms with low demand for monitoring have
low agency costs related to the manager-stockholder relationship. Those firms have few and
strong stockholders (private firm and high ownership concentration) who can easily stay well
informed (small and young firm). Hence, they have little need for monitoring by non-owner
directors. However, these firms may have high need for advice, because their low age and
small size suggest they lack the internal resources to successfully make strategic decisions
and to establish key external networks without the help from experts on their boards.
Earlier research on the role of women on corporate boards has studied only settings where the
proportion of female directors is both low and unregulated (Farrell and Hersch 2005; Adams
and Ferreira 2009b). Using the same independence measure as we do, Adams and Ferreira
classify 84 percent of the female directors in US firms as independent, as opposed to 40
percent of the males. 16 These figures make the authors warn against mandating gender
diversity, fearing that a larger proportion of female directors may unduly increase board
independence.
This argument assumes that the pool of potential female directors is biased towards high
independence. However, unlike what we do us, existing research studies only firms where the
governance code recommends that half the firm’s directors be independent, and where about
10 percent of the directors are females. This limited empirical context means that the
unobserved, complete pool of female director candidates may very well be balanced regarding
dependence vs. independence. Nevertheless, firms may still want to recruit more heavily from
the subpool of independent females in order to comply with the independence code.
We show, however, that even firms facing no such independence requirement increase their
board independence by almost 45 percent when one third of their male directors must be
replaced by women. Thus, stockholders find it difficult to avoid excessive board 16 Their sample is an unbalanced panel of S&P 500, S&P MidCap, and S&P SmallCap firms during the period 1996–2003. The data are collected by the Investor Responsibility Research Center.
55
independence when gender quotas are mandated. Therefore, it seems that the complete pool of
female director candidates is indeed heavily biased towards independence. This bias means
that mandatory quotas will be a more serious problem for optimal board design the more
gender balance the regulator enforces.
We present our base-case model and the predictions in Section 2. The data and summary
statistics are discussed in Section 3, while Section 4 explains the methodology and presents
the base-case results. Section 5 analyzes alternative board independence proxies and
alternative independence determinants. We summarize and conclude in Section 6.
2. Predictions
Board independence is affected by regulatory and non-regulatory determinants. Our paper
focuses on the former, which are exogenous restrictions on board composition. Specifically,
we analyze how board independence is influenced by mandatory law in terms of the GBL and
by soft law in terms of the governance code. Nevertheless, our basic framework comes from
the non-regulatory determinants, which are the firm-specific characteristics we will use to
explain why the response to the same regulatory constraint differs across firms. The literature
has so far addressed only non-regulatory determinants, such the firm’s ownership structure,
profitability, complexity, and risk.
Our base-case model is the following
1 2 3
4 5 6
7 8 9
*
A Reg B Non-regit
it it it it
it it it
it it
(1) Board independence X XFemale directors Listed Female directors Listed
Inside ownership Outside ownership PerformanceLeverage Risk
α β βα β β ββ β ββ β β
= + +
= + + +
+ + +
+ + + 10it it itFirm size Firm age u β+ +
The regulatory determinants in the vector XReg are specified in the second row of the model,
while the non-regulatory determinants in XNon-reg are spelled out in the two bottom rows. The
starting point is the tradeoff theory, in which optimal board independence reflects the value-
maximizing combination of monitoring and advice. We outline the tradeoff theory and
operationalize the board independence concept in section 2.1. We predict how board
independence relates to regulatory determinants in section 2.2, and to non-regulatory
determinants in section 2.3. Table 1 summarizes the empirical proxies.
56
Table 1
2.1 Board independence
Independent directors may create firm value because there is a potential conflict of interest
between the manager who runs the firm and the owners who delegate control rights to the
manager (Jensen and Meckling 1976). The more serious the resulting separation between
ownership and control, the more value the directors may create by monitoring management
more efficiently inside the boardroom than the non-director owners can do from the outside.
Different director types have different incentives to fill this monitoring function. Compared to
directors with professional ties to the firm or personal ties to the manager, directors without
such ties have less to lose by challenging and criticizing the manager. The former director
type is called dependent or inside, while the latter is called independent or outside (Bhagat
and Black 1998; Hermalin and Weisbach 1998; Adams, Hermalin, and Weisbach 2010). The
independent director’s monitoring incentives on a specific board are strengthened further if
the value of his or her human capital primarily depends on a reputation for offering
monitoring services on any board.
The board’s second function is to advise the firm’s management. High advisory skills require
deep insight into the firm, its customers, suppliers, competitors, and industry. Dependent
directors have such skills because of their closeness to the firm, while independent directors
lack them because of their arms-length distance from the firm (Bhagat and Black 1998, 2002).
Because dependent directors may lose reputation by monitoring (control) and build reputation
by advising (support), they have stronger incentives to advise than independent directors do,
and lower incentives to monitor.
This setting implies that the value of the board’s monitoring function stems from reduced
agency costs in the relationship between owners and managers (“monitoring prevents bad
projects”). In contrast, the value of advice stems from the board’s ability to generate new
ideas for strategy and operations that management can develop further and implement
(“advice creates good projects”).
Finally, one would expect that managers operating in this environment would dislike being
intensively monitored. The reason is that monitoring reduces managements’ discretion
regarding the firm’s resources as represented by the free cash flow (Jensen 1986). In contrast,
57
managers like advice, because more advice may increase the free cash flow. Therefore,
management may respond to more monitoring by providing the board with less information.
Adams and Ferreira (2007) formalize this setting and show that optimal board independence
involves a tradeoff between the value of monitoring provided by independent directors and
the value of advice provided by dependent directors. Two of their results are particularly
important for our context. First, over-optimal (too much) independence reflects a board where
the advisory skills are too weak relative to the monitoring skills. There is too much control
and too little support. Firm value is lost because the board generates too few new ideas and
too strongly restricts management.
The second important result is that too much independence pushes not only the board’s
advisory skill below its optimal level. Excessive independence may also reduce the board’s
ability to monitor. Such a loss of both advisory value and monitoring value may occur when
managers respond to being monitored by reducing information flow to the board. This
response creates a problem because an uninformed board cannot properly fulfill its two roles.
Therefore, greater independence caused by new regulation may reduce board quality because
the board gets too low advisory skills and also too little information for advice as well as for
monitoring. This imbalance means the radical change in board composition mandated by the
GBL in our setting may produce uninformed boards that have underoptimal focus on advice
and overoptimal focus on monitoring.
We measure board independence empirically in line with earlier research from the United
States, which classifies a director as either inside, grey, or outside (Baysinger and Butler 1985;
Weisbach 1988; MacAvoy and Millstein 1999; Adams and Ferreira 2009b). Inside directors
are defined as the firm’s full-time employees, former employees, or employees of closely
related firms. Grey (affiliated) directors have professional relationships with management, or
are likely to have business relationships with the firm, such as investment bankers and
lawyers. Outside directors are neither inside nor grey.
We measure board independence in the base case as the fraction of outside directors.
Although this is the most common measure in the literature, our robustness tests will use
several alternatives. Specifically, we alternatively define independent directors as the fraction
of outside directors minus the fraction of inside directors (Bhagat and Black 2002) and as the
fraction of outside plus grey directors (Linck, Netter, and Yang 2008; Ahern and Dittmar
58
2012). Finally, we use an independence measure based on whether a firm’s CEO has a seat on
the board (Carter and Lorsch 2004).
These independence measures may all be criticized for overestimating true independence. The
reason is that the measures ignore social (i.e., non-business) relationships between the
directors and the CEO (Hwang and Kim 2009; Cohen, Frazzini, and Malloy 2012). We think
this is a minor problem in our sample, where men fill 96 percent of the CEO positions. This
fact suggests that if anything, measures that ignore social relationships will overestimate
independence more for men than for women. Consequently, the true difference in
independence between men and women is even larger than what we measure. This bias
strengthens the power of our test, where the null hypothesis is that independence and gender
are unrelated.
2.2 Regulatory determinants
We expect that board independence relates positively to the GBL, which mandates the same
minimum fraction (40 percent) of female directors in every firm. This prediction is supported
by Farrell and Hersch (2005), who examine the characteristics of female directors in 300
firms on the Fortune 1000 list from 1990 to 1999. The use of female directors increased by 7
percentage units during these ten years, and more than 90 percent of the incoming females
were classified as outside directors. The authors argue that because women have limited
experience as managers and stockholders, one would expect board independence to increase
when female directors replace males. Adams and Ferreira (2009b) support this logic by
arguing that because women seldom belong to the so-called old boys’ network, women are
also closer to the theoretical notion of independent directors being arm’s-length monitors. The
authors estimate that while large firms in the United States had about 10 percent females on
the boards in 2004, 84 percent of them were classified as outside directors.17
17 Along similar lines, it has been argued that independence is easier to achieve by ensuring diversity in ethnicity and gender (Fields and Keys 2003; Beecher-Monas 2007). It may be a general belief that diverse boards are more independent (Adams and Ferreira 2009b). Several behavioral differences between men and women reported in the literature support this idea. Female directors may reduce earnings management (Gul, Scrinidhi, and Tsui 2007), women seem to adopt a more democratic, transformational, and trust-building leadership style (Cohen, Pant, and Sharp 2001; Klenke 2003; Trinidad and Normore 2005), females may exhibit higher ethical standards in their decision making (Betz, O'Connell, and Shepard 1989; Mason and Mudrack 1996; Clikeman, Geiger, and O'Connell 2001), and women may be more risk averse (Riley and Chow 1992; Powell and Ansic
59
We expect a positive relationship between having independent directors and being listed
because every sample firm is exposed to the GBL, but only the listed are subject to the
independence code. Nevertheless, listed firms may comply with both the GBL and the code in
one move by appointing female directors who are also independent. To account for this
possibility, we relate the GBL and the code by means of an interaction term for the fraction of
female directors and the firm’s listing status. This interaction term allows us to determine
whether the GBL drives independence differently in firms that must also comply with the
code. If listed firms choose female directors to comply with both the code and the GBL, we
expect a positive coefficient for the interaction term. Conversely, the coefficient will be
negative if female directors are more often independent in non-listed firms than in the listed.
This latter case would occur if dependent female director candidates prefer to sit on listed
firms’ boards, possibly because of higher visibility and pay. Such preferences would tend to
produce a higher fraction of independent female candidates in the talent pool for non-listed
firms than for the listed.
2.3 Non-regulatory determinants We group the non-regulatory determinants of board independence into board, ownership, and
general firm characteristics. Because the inclusion of partly overlapping characteristics in the
same model may cause multicollinearity problems, our base-case model uses only a subset of
the non-regulatory determinants in Table 1 that are common in the literature.
Empirical research has shown that board independence declines as inside ownership increases
(Bhagat and Black 2002; Linck, Netter, and Yang 2008). This finding can be rationalized by
the theoretical argument that there is less need for monitoring when the board and the
stockholders are aligned (Raheja 2005). We measure inside ownership by the aggregate
equity fraction held by the firm’s officers and directors. The equity fraction is measured as
ultimate ownership, which is the investor’s direct equity holding in the firm plus any indirect
holdings through intermediaries such as holding companies. We predict that board
independence and inside ownership are negatively correlated.18
1997; Sundén and Surette 1998). Adams and Funk (2009) have recently challenged the latter point by finding that female directors in Sweden are slightly less risk averse than males are. 18 The term “inside owner” is very different from the term “inside director.” An inside owner is a stockholder in the firm who is also on the board or on the management team. An inside (dependent) director is a board member with close relationships to the firm that are unrelated to ownership. The same relationship goes for outside
60
The principal’s power and incentives to monitor the agent increase with outside ownership
concentration (Shleifer and Vishny 1997). This means powerful owners outside the
boardroom have strong incentives to ensure that their directors monitor management properly
on the owners’ behalf. Hence, board independence may relate positively to outside ownership
concentration. On the other hand, large outside owners may monitor management directly
rather than indirectly through the board. Such monitoring can happen both in the stockholder
meeting and through informal contact with management. The use of such channels reduces
the demand for monitoring directors. Hence, we do not specify the expected relationship
between board independence and outside ownership concentration, which we measure by the
Herfindahl index based on every ultimate ownership stake in the firm.
Research shows both theoretically and empirically that firms may respond to poor
performance by appointing a higher fraction of independent directors (Hermalin and
Weisbach 1991; Bhagat and Black 2002). The rationale is that a CEO who dislikes
monitoring is supposed to have less influence over board composition the weaker the firm’s
performance. We measure performance as the past three years’ average return on assets,
predicting an inverse relationship between board independence and firm performance.
Complex firms are thought to have a high need for monitoring and have also been found to
have more independent boards (Coles, Daniel, and Naveen 2008; Linck, Netter, and Yang
2008). We measure firm complexity by firm size and by firm age, expecting higher values of
both characteristics to correlate positively with board independence. Managers of firms with
high debt have less free cash flow to waste on value-destroying projects. Therefore, high debt,
and the resulting strong monitoring by creditors, is a substitute for the monitoring carried out
by an independent board (Jensen 1986). We predict an inverse relationship between board
independence and financial leverage.
Finally, the optimal fraction of independent directors decreases as the cost of monitoring
increases (Adams and Ferreira 2007). Such monitoring costs are particularly high when firms
with strong information asymmetry are monitored by independent directors (Maug 1997).
Moreover, empirical research has found that the information asymmetry is higher the more
volatile the firm’s stock returns (Fama and Jensen 1983). Because we have stock return data
for only the listed firms in our sample, we use the standard deviation of the book return on owners vs. outside (independent) directors, where an outside owner is not on the board or on the management team.
61
assets to proxy for information asymmetry and hence for monitoring costs. We expect board
independence to correlate negatively with risk in a volatility (total risk) sense.
Summarizing section 2, we predict that the board will have a higher fraction of independent
directors when the firm is listed, has many female directors, low inside ownership
concentration, low performance, low leverage, low risk, and when the firm is large and old.
1 2 3
4 5 6
7 8 9
*
A Reg B Non-regit
it it it it
it it it
it it
(1) Board independence X XFemale directors Listed Female directors Listed
Inside ownership Outside ownership PerformanceLeverage Risk
α β βα β β ββ β ββ β β
= + += + + ++ + ++ + + 10it it itFirm size Firm age u β+ +
3. Data and summary statistics
Our sample is all firms exposed to the GBL, and the data set is an unbalanced panel from
2003 to 2008. Except for the data on director independence, the source is the CCGR database
(www.bi.edu/ccgr). 19 Norwegian firms with limited liability are legally obliged to publish
accounting statements every year. The firm must also report the identity and the equity
holdings of its owners, directors, and CEO. Failure to submit this information within 17
months after fiscal year-end triggers automatic liquidation by the court. The law also
mandates a standardized set of accounting statements certified by a public auditor, regardless
of the firm’s listing status, size, and industry.
The data used to manually classify directors as inside, grey, and outside are from
Brønnøysundregistrene (www.brreg.no) and Proff (www.proff.no). We obtain supplementary
information on director characteristics by manually searching the annual reports. Appendix 1
illustrates how we use this data set to classify the directors.
Table 2 shows distributional properties of the variables. Because the governance code applies
only to listed firms, Table 3 splits the sample by listing status to uncover whether board
composition depends on organizational form.
Table 2
Table 3
According to Table 2, board independence as measured by the fraction of outside directors is
59 percent on average during the sample period. Table 3 documents that this fraction is 64 19 The database includes every limited liability firm registered in Norway from 1994 to the present.
62
percent in listed firms and 50 percent in the non-listed, respectively. Because non-listed firms
are not subject to the governance code on independence, this lower degree of independence in
non-listed firms is as expected from a regulatory perspective. The difference may also reflect
that optimal independence is lower in non-listed firms because the value of monitoring is
lesser. This rationale is supported by the table, which suggests that potential agency problems
between owners and managers are generally smaller in non-listed firms. Ownership is
concentrated in fewer hands, which implies fewer free-riding problems and more incentives
and power to discipline management. For instance, the largest stockholder in non-listed firms
holds on average 63 percent of the equity, compared to 29 percent in the listed. Boards of
non-listed firms also have more owner presence. Officers and directors hold on average 16
percent of the equity in non-listed firms and 8 percent in the listed.
These characteristics mean that demand for independent directors who monitor management
on the owners’ behalf is lower in non-listed firms. Risk as measured by asset return volatility
is also considerably higher. These characteristics increase demand for advice relative to
monitoring, which also follows from the fact that non-listed firms in our sample are generally
smaller and younger.
Table 4 shows the prevalence of outside, grey, and inside directors for the sample as a whole
in panel A. The fraction of outside directors increases every year, growing from 46 percent in
2003 to 67 percent in 2008. An opposite monotone decline occurs for the fraction of inside
directors, which drops from 44 to 27 percent. Finally, the fraction of grey directors drops from
10 to 6 percent. Hence, the large growth in outside directors primarily happens at the expense
of inside directors rather than grey directors.
This pattern means that the fraction of outside directors, which is our base-case measure of
board independence, may capture a fundamental shift in director skills towards more
monitoring at the expense of less advice. Such a shift towards monitoring would have been
more questionable if most of the decline had happened among the grey directors, who are
more independent than the inside directors.
Table 4
Panel B shows that the fraction of outside directors grows from 52 percent to 72 percent in the
listed firms and from 41 to 59 percent in the non-listed. Hence, the difference remains
63
constant, documenting that the increased independence is not due to the governance code,
which was introduced in the middle of the sample period. Linck, Netter, and Yang (2008)
follow a large sample of listed firms in the United States over time. They estimate that the
average fraction of inside directors is around 30 percent in 2003, which is their next-to-last
sample year, and also after the 2002 Sarbanes-Oxley Act was implemented. The
corresponding fraction in our sample that year is 45 percent. Remarkably, however, it takes
only another five years in our sample until the fraction of inside directors has dropped to 28
percent.
Panel C documents the relationship between director type and gender. Females are on average
outside directors in 84 percent of the cases, while the corresponding figure for men is 50
percent. Hence, the average female director is more than 60 percent more likely than a male
director is to be in the outside category. Similarly, whereas on average 13 percent of the
women are inside directors, men are inside directors in 43 percent of the cases. These
relationships remain practically constant over time. Although not reported in the table, this
pattern is also the same regardless of listing status.
This difference in independence between men and women is very close to the difference
estimated by Adams and Ferreira (2009b) and Farrell and Hersch (2005) in the United States,
using the same definition of outside directors as we do. They report that 84 percent of the
females are in this category, as opposed to 40 percent of the males. Our figures are 84 percent
and 50 percent, respectively. This evidence strengthens the impression that owners of
Norwegian firms could not have mitigated excessive independence after the GBL by
recruiting more dependent (inside) female directors than earlier. It seems there is no such
untapped pool of talent better qualified for advice than for monitoring.
Nevertheless, stockholders could have dampened the GBL’s impact on board independence
by merely adding female directors to the existing, male-dominated board. According to Table
5, however, this is not a widespread strategy, because board size does not increase noticeably
during the sample period. The average board has 5.60 members in 2003 and 5.86 members
five years later. In fact, panel B shows that board size even decreases somewhat in non-listed
firms, where the board also tends to be smaller than in the listed (4.97 vs. 6.13 members on
average, respectively).
Table 5
64
The second way to reduce excessive board independence is by appointing female directors
who already hold board seats elsewhere. Directors with multiple seats are likely to be less
independent than those with just one seat because of the former’s links to board members and
top management in other firms. Table 6 documents the holding of directorships by men and
women in 2003, 2006, and 2008, respectively.
Table 6
Panel A documents, as expected, a trend towards a higher total number of multiple seats held
by women and a lower number of such seats held by men. In particular, multiple seats are
held by as many men as women at the end of the sample period (66 vs. 68, respectively),
while men dominate very strongly five years earlier (320 vs. 18, respectively). This pattern
follows almost by implication when 30 percent of the men are to be replaced by women in
short supply during a brief period. A similar impression is given by panel B, where the
number of seats per director increases over time for females and decreases for males. For
instance, the mean number of seats per director increases from 1.16 to 1.22 for women, while
decreasing from 1.25 to 1.10 for men.
Nevertheless, panel C shows that when we consider only directors with multiple seats, there is
no clear tendency that the fraction of women who hold multiple seats increases over time.
Rather, the striking feature is that men have multiple seats considerably less often than earlier.
For instance, the average number of seats held by men decreases from 1.25 to 1.10 during the
sample period (panel B), while the fraction of men who hold multiple seats decreases from 22
percent to 8 percent (panel C).
Overall, Table 6 shows that there is a certain tendency for some women to hold multiple
directorships more often. Nevertheless, the only clear trend is that men as a group hold
multiple seats less often than earlier. Although unreported tests we have made show that
women with multiple directorships are generally more dependent than are women holding just
one seat, it seems that recruiting female directors with multiple seats is not a widespread
strategy to reduce board independence. One possible reason is if stockholders share the view
that busy directors may easily become overstretched and therefore have less value.
65
Alternatively, the pool of females who can potentially hold multiple directorships is so small
that this recruiting source does not matter much.20
Overall, it seems the two alternative strategies for recruiting female directors are not widely
used to dampen the growth in board independence caused by the GBL. Instead, stockholders
comply with the new law by recruiting from the talent pool of female directors with relatively
little leadership experience. These female directors replace males, thereby leaving board size
unchanged.
Table 7 shows bivariate correlation coefficients between the independent variables in the
sample we will use to estimate model (1). Multicollinearity should not be a serious concern,
because no correlation coefficient exceeds the critical limit of 0.8 (Studenmund 2000). Also,
the variance inflation factor (VIF) equals 4.88, which is below the limit of 5 considered a
sufficient reason for not suspecting multicollinearity problems.
Table 7
4. Empirical methodology and base-case results Our data set contains multiple observations of the same firm over time. Therefore, we will use
panel data techniques to reduce potential endogeneity problems caused by unobservable
determinants of board independence (Hsiao 2003).21 We use the fixed-effects approach to
account for firm-specific unobservables.22 Moreover, we control for the possibility that the
20 The effect of women on board independence may also be dampened by appointing experienced women from other countries. Nygaard (2011) shows, however, that the proportion of foreign female directors was 11.7 percent in 2003, while the corresponding number for males was 13.3 percent. Five years later, 12.8 percent of the female directors and 15.7 percent of the males were foreigners. Thus, there is no clear shift from national to international recruitment of female directors. One possible reason is that even though such board members are experienced, they may nevertheless be outside directors in a monitoring sense because their network is from a different country. 21 Endogeneity caused by reverse causation seems to be a minor concern in our study. First, the large increase in female directors is caused by an exogenous shock in terms of an unexpected new law. Second, the board’s independence in listed firms is partially driven by an exogenous corporate governance code that recommends the same minimum level of independence in every firm. Third, firms may certainly reduce inside ownership to ensure that a majority of the directors are independent. Although the correlation between board independence and insider ownership is negative and significant in our sample, the correlation between inside ownership and time is insignificant. Hence, increased inside ownership is not likely to be caused by increased independence over time. Finally, firms exposed to the GBL can avoid the law by exiting to an unexposed organizational form. However, such exit is not an option for a listed firm unless it also chooses to delist. Bøhren and Staubo (2013) show that although half the firms exposed to the GBL in 2003 had left this organizational form by 2008, roughly three quarters of the exiting firms were non-listed. 22 We use fixed firm effects even though the base-case model contains a dummy variable for a firm’s listing status. Because some firms change listing status during the sample period, however, this dummy variable is not a constant over time for all firms. We prefer the fixed-effects approach because it allows the unobservable firm
66
GBL’s effect on board independence in all firms may change as 2008 approaches, when the
law became mandatory. Table 8 shows the estimates of (1), using OLS estimation with fixed
firm effects and fixed time effects.
Table 8
The presence of female directors is associated positively with board independence and is
highly statistically significant. Thus, as predicted, mandating a large change in gender balance
towards more females does not alter gender balance alone. The regulation increases the
board’s independence, too. This result is consistent with that shown in Table 4, which shows
that the sample’s average for monitoring skills in any year is much higher for female directors
than for males. Conversely, women have a comparative disadvantage as advisors. The
stockholders’ problem is, however, that the tradeoff between monitoring and advice after the
GBL must be made within a severely restricted opportunity set regarding the female director
candidates.
Listed firms have boards with significantly greater independence than non-listed firms do.
This result is as predicted, given the regulatory fact that the independence rule in the
governance code applies only to listed firms.23 Moreover, the negative coefficient for the joint
effect of female directors and listing status on independence suggests that listed firms do not
fill the gender quota and the independence quota in one go. That is, a listed firm does not
systematically recruit a female who is independent more often than a non-listed firm does.
Rather, it is even more common for a female director to be independent in non-listed firms,
despite the fact that such firms are subject only to the GBL rather than to both the GBL and
the independence code. The reason is that independence and gender are so closely related in
the talent pool and possibly also because dependent female director candidates prefer listed
firms.
Turning next to the non-regulatory determinants, the table shows that while outside ownership
concentration is not a significant determinant, inside ownership concentration relates
negatively to board independence. This finding is in line with our prediction that the demand
effect to be correlated with the error term. The robustness tests will also use the random-effects approach, which assumes the unobservable firm effect is a random variable that is uncorrelated with the error term. 23 We get the same result both for female directors and for listing status if we exclude all observations after 2005, that is, after Parliament decided not to withdraw the law, but to instead add a liquidation penalty for non-compliers.
67
for directors who primarily monitor is weaker when the CEO or these directors are owners as
well.
As expected, firm performance and board independence are negatively correlated, suggesting
that monitoring becomes stricter when the firm does poorly. Independence decreases as
leverage grows, which is consistent with substitutability between a monitoring board and a
monitoring creditor. However, board independence is not systematically related to asset risk.
This finding suggests that if monitoring costs are reflected in the volatility of asset returns,
such costs are not an important concern in the tradeoff between monitoring and advice.
Finally, and as we predicted, firm complexity as measured by the firm’s size and age
correlates positively with board independence. Hence, the smaller and younger the firm, the
lesser the demand for an independent board.
Summarizing, we have tested the multivariate model in (1) using panel data techniques to
account for unobservable firm and time effects on board independence. Most relationships are
statistically significant, and every significant relationship is consistent with our predictions of
how the demand for an independent vs. a dependent board depends on regulatory and non-
regulatory characteristics. We find that the demand for board independence tends to be lower
when the firm has few female directors, strong owners, high profits, and when the firm is
small, young, and non-listed. Such firms trade off monitoring and advice in a way that reflects
a low value of being monitored and a high value of being advised.
5. Robustness
We first re-estimate the base-case model in (1) using alternative econometric techniques.
Subsequently, we analyze what happens when we use alternative measures of board
independence. Finally, we test the sensitivity to using different empirical proxies than those
used in (1) for some of the non-regulatory determinants of board independence.
5.1 Econometric techniques
The base-case regression in Table 8 uses fixed-effects estimation to control for the influence
of unobservable firm characteristics that remain constant over time. Table 9 repeats these
estimates as the point of reference under technique I. Technique II is random-effects
68
regressions with ML (maximum likelihood) methodology, and III controls for fixed firm
effects by using sample period averages for each variable in every firm.24
Table 9
The estimates reported in II and III are mostly consistent with those of the base case in I. First,
the significant relationships between board independence and female directors, listing status,
inside ownership concentration, and firm size are robust to the econometric technique used.
Second, none of the remaining variables is significant with opposite signs under alternative
techniques. These results show that our major findings in the base-case model are not driven
by the chosen estimation method. Therefore, we continue using the fixed firm effect and the
fixed time effect approaches in the following.
5.2 Board independence
The independence measure in the base case comes from partially hand-collected data that we
use to classify each director as either outside, grey, or inside according to the definitions from
section 2. The data collection and the classification system are both subject to potential bias
because they rely on our judgment. Moreover, independence is an elusive concept because it
is hard to define precisely in terms of a specific empirical proxy. For these reasons, we test
three alternative measures of independence.
The first alternative is whether the CEO is a member of the firm’s board.25 The rationale is
that the party being monitored should not have a say in the monitoring body (Carter and
Lorsch 2004). The CEO is not a board member in 88 percent of our sample firms. This
happens in 92 percent of the listed firms and in 80 percent of the non-listed, respectively. We
measure board independence by a dummy variable that is 1 if the CEO is not a director and 0
otherwise.
24 Technique II reduces standard errors by accounting for cross-sectional variance in the between estimator. Technique III reduces the standard errors by accounting for cross-sectional variance in the between and within estimators. Technique IV is appropriate when the dependent variable can take on values only in a restricted interval, such as in our case, where board independence varies between zero and one. 25 Of course, the CEO is always present in board meetings except when a case cannot be openly discussed unless he or she is absent, as when the board evaluates the CEO’s performance and decides compensation. A law introduced two years after the end of our sample period makes it illegal for the CEO to be a board member.
69
The sample as well as the he determinants of independence correspond to what we used in the
base case, and we estimate the relationship as a logit model with random effects.26 Table 10
reports the estimates under model A. The table shows that model A and the base-case model
from Table 8 have the same set of significant determinants at the 5 percent level or lower
(female directors, listing status, performance, leverage, firm age, firm size), whereas two
determinants are insignificant in both models (outside ownership concentration and risk). Two
determinants are significant only at the 10 percent level in the base case (the interaction term
and inside ownership concentration).
Table 10
Model B follows Bhagat and Black (2002) by defining independent directors as the fraction of
outside directors minus the fraction of inside directors. Model C follows Ahern and Dittmar
(2012) and Linck, Netter, and Yang (2008) by classyfying both outside and grey directors as
independent.27 Again, the deviation from the base case is minor in both models. Reassuringly,
therefore, our findings are not sensitive to alternative ways of measuring board independence.
5.3 Non-regulatory determinants of board independence
The set of potential determinants of board independence is large when moving from
theoretical constructs to empirical proxies. Given the purpose of this paper, however, our
major focus is on determinants that directly or indirectly reflect regulatory effects on board
independence. These determinants are listing status as a direct determinant through the
governance code and female directors as an indirect determinant through the GBL. The
remaining variables in (1) are potential non-regulatory determinants. To analyze whether the
choice of such non-regulatory determinants matters for our findings for the regulatory effects,
we specify a model of board independence where the non-regulatory determinants are as close
as possible to those used by two recent studies of board composition in the United States
(Adams and Ferreira 2009a; Linck, Netter, and Yang 2008):
26 Fixed-effects estimation is not feasible in a logit model (Hsiao 2003). 27 The definition of independent directors used in model C comes closest to the definition used in the Norwegian governance code. This code recommends that half the stockholder-elected directors be unrelated to the executive team or to key business partners.
70
1 2 3
4 5 6
7 8 9
(3) *
it it it it it
it it it
it it
Board independence Female directors Listed Female directors ListedBoard size CEO ownership CEO ownership squaredCEO tenure CEO tenure squared Firm age
α β β ββ β ββ β β
= + + ++ + ++ + +
10 it
it itFirm age squared uβ+ +
The first line of independent variables in (3) specifies the regulatory determinants, which are
identical to those used in (1). Lines 2–4 hold the non-regulatory determinants, which differ
from their counterparts in (1).
Research shows that larger boards are more independent (Boone et al. 2007). Moreover,
larger firms are usually more complex and have larger and more diverse boards (Eisenberg,
Sundgren, and Wells 1998; Linck, Netter, and Yang 2008; Yermack 1996). Accordingly, we
use board size instead of firm size to proxy for firm complexity in (3), expecting board
independence to grow with growing board size.
Board independence has been found to decline with increasing CEO ownership and CEO
tenure (Boone et al. 2007; Coles, Daniel, and Naveen 2008; Linck, Netter, and Yang 2008;
Shivdasani and Yermack 1999; Weisbach 1988). However, Adams and Ferreira (2009a) show
theoretically that one should reconsider the evidence that board independence relates
inversely to CEO ownership and CEO tenure. Adams and Ferreira also find empirically that
board independence does indeed correlate non-monotonically with CEO ownership and with
tenure. Specifically, board independence first decreases and then increases as CEO ownership
and tenure grow. This evidence of a v-shaped relationship is interesting because it suggests
that a change in board independence comes with benefits and costs that vary with the pre-
change level of independence (Adams and Ferreira 2007).
We include CEO ownership, CEO tenure, and their squared values in (3). We expect the
coefficient for both proxies to be negative for the linear term and positive for the quadratic
term. Finally, firm complexity may not increase linearly with firm age, and complex firms
may need boards with greater independence (Boone et al. 2007). We account for this possible
non-linearity by including firm age squared, which we expect to have a negative coefficient.
The estimates are shown in Table 11. The two new measures of CEO characteristics and the
new measure of board diversity are all significant at the 5 percent level with the predicted
signs. Also, the role of female directors corresponds to the base case, while listing status is no
longer significant. Although the results as a whole are generally weaker than in the base case,
71
the relationship between gender balance regulation and board independence is insensitive to
the non-regulatory determinants of independence we account for.
Table 11
6. Conclusions
This paper is the first to analyze the empirical relationship between mandatory gender quotas
and director independence in corporate boards. We show that forcing every firm to radically
and quickly change the board’s gender mix produces a strong upwards shift in board
independence. This happens because director independence is a much more common
characteristic among females than among males. We also find that board independence varies
considerably from firm to firm before the gender quota became mandatory. This result
supports the idea of a firm-specific demand for monitoring skills vs. advisory skills in the
boardroom. As predicted, the evidence shows that firms needing monitoring the least and
advice the most are small, young, private, profitable firms with strong owners and low gender
diversity on the board. Such firms are likely to be hurt the most by mandatory gender balance.
A major political argument for mandating that female directors hold at least 40 percent of the
positions in Norwegian boards was that it would improve corporate governance and economic
performance. The idea was to force supposedly irrational owners to choose human talent from
a wider pool, that is, from both female and male candidates. It was also argued that gender
quotas ensure a more inclusive and fair business society that better reflects basic values in
modern society (Langli 2011). Up from 11 percent when the gender balance law was passed
in 2003, females currently hold about 40 percent of the board seats in firms that must fill the
mandatory gender quota.
Our findings suggest that because the fraction of dependent directors has dropped so sharply,
the gender balance law may have weakened boards’ ability to fill their advisory role. Also,
and despite the fact that independent directors have stronger monitoring incentives than
dependent directors do, the law may even have weakened a board’s ability to monitor. Such
regulatory costs of excessive independence are consistent with theoretical predictions and
earlier findings. Overall, the case of mandatory gender balance on corporate boards illustrates
72
how a large regulatory shift can have strong and unintended side effects that were perhaps not
even considered when the law was written.
73
References
Adams, Renée B., and Daniel Ferreira. 2007. "A Theory of Friendly Boards." Journal of Finance no. 62 (1):217–250.
Adams, Renée B., and Daniel Ferreira. 2009a. "Strong Managers, Weak Boards?" CESifo Economic Studies no. 55 (3–4):482–514.
Adams, Renée B., and Daniel Ferreira. 2009b. "Women in the Boardroom and Their Impact on Governance and Performance." Journal of Financial Economics no. 94 (2):291–309.
Adams, Renée B., and Patricia Funk. 2009. "Beyond the Glass Ceiling; Does Gender Matter?" ECGI Finance Working Paper no. No. 273/2010.
Adams, Renée B., Benjamin E. Hermalin, and Michael S. Weisbach. 2010. "The Role of Boards of Directors in Corporate Governance: A Conceptual Framework and Survey." Journal of Economic Literature no. 48 (1):58–107.
Adams, Renée B., and Tom Kirchmaier. 2013. "From Female Labor Force Participation to Board Gender Diversity." ECGI-Finance Working Paper No. 347/2013.
Ahern, Kenneth R., and Amy Dittmar. 2012. "The Changing of the Boards: The Impact on Firm Valuation of Mandated Female Board Representation." Quaterly Journal of Economics no. 127 (1):137–197.
Baysinger, Barry D., and Henry N. Butler. 1985. "Corporate Governance and the Board of Directors: Performance Effects of Changes in Board Composition." Journal of Law, Economics, and Organization no. 1 (1):101–124.
Beecher-Monas, Erica. 2007. "Marrying Diversity and Independence in the Boardroom: Just How Far Have You Come, Baby?" SSRN eLibrary.
Betz, Michael, Lenahan O'Connell, and Jon M. Shepard. 1989. "Gender Differences in Proclivity for Unethical Behavior." Journal of Business Ethics no. 8 (5):321–324.
Bhagat, Sanjai, and Bernard S. Black. 1998. "Independent Directors." The New Palgrave Dictionary of Economics and the Law. no. 2:283–287.
Bhagat, Sanjai, and Bernard S. Black. 2002. "The Non-Correlation Between Board Indpendence and Long-Term Firm Performance." Journal of Corporation Law no. 27 (2):231–273.
Boone, Audra L., Laura C. Field, Jonathan M. Karpoff, and Charu G. Raheja. 2007. "The Determinants of Corporate Board Size and Composition: An Empirical Analysis." Journal of Financial Economics no. 85 (1):66–101.
Bøhren, Øyvind, and R. Øystein Strøm. 2010. "Governance and Politics: Regulating Independence and Diversity in the Board Room." Journal of Business Finance and Accounting no. 37 (9–10):1281–1308.
Bøhren, Øyvind, and Siv Staubo. 2013. "Does Mandatory Gender Balance Work? Changing Organizational Form to Avoid Board Upheaval." Working Paper, Norwegian Business School.
Carter, Colin B., and Jay W. Lorsch. 2004. "Back to the Drawing Board: Designing Corporate Boards for a Complex World." Boston, MA: Harvard Business School Press.
Clikeman, Paul M., Marshall A. Geiger, and Brendan T. O'Connell. 2001. "Student Perceptions of Earnings Management: The Effects of National Origin and Gender." Teaching Business Ethics no. 5 (4):389–410.
Cohen, Jeffrey R., Laurie W. Pant, and David J. Sharp. 2001. "An Examination of Differences in Ethical Decision-Making Between Canadian Business Students and Accounting Professionals." Journal of Business Ethics no. 30 (4):319–336.
74
Cohen, Lauren, Andrea Frazzini, and Christopher Malloy. 2012. "Hiring Cheerleaders: Board Appointments of "Independent" Directors." Management Science no. 58 (6):1039–1058.
Coles, Jeffrey L., Naveen D. Daniel, and Lalitha Naveen. 2008. "Boards: Does One Size Fit All?" Journal of Financial Economics no. 87:329–356.
Duchin, Ran, John G. Matsusaka, and Oguzhan Ozbas. 2010. "When are Outside Directors Effective?" Journal of Financial Economics no. 96 (96):215–237.
Eisenberg, Theodore, Stefan Sundgren, and Martin T. Wells. 1998. "Larger Board Size and Decreasing Firm Value in Small Firms." Journal of Financial Economics no. 48 (1):35–54.
Fama, Eugene F., and Michael C. Jensen. 1983. "Separation of Ownership and Control." Journal of Law & Economics no. 26 (2):301–326.
Farrell, Kathleen A., and Philip L. Hersch. 2005. "Additions to Corporate Boards: The Effect of Gender." Journal of Corporate Finance no. 11 (1/2):85–106.
Fields, M. Andrew, and Phyllis Y. Keys. 2003. "The Emergence of Corporate Governance from Wall St. to Main St.: Outside Directors, Board Diversity, Earnings Management, and Managerial Incentives to Bear Risk." Financial Review no. 38 (1):1–24.
Gul, Ferdinand A., Bin Scrinidhi, and Judy Tsui. 2007. "Do Female Directors Enhance Corporate Board Monitoring? Some Evidence from Earnings Quality." The Hong Kong Polytechnic University.
Hermalin, Benjamin E., and Michael S. Weisbach. 1991. "The Effects of Board Composition and Direct Incentives on Firm Performance." Financial Management no. 20 (4):101–112.
Hermalin, Benjamin E., and Michael S. Weisbach. 1998. "Endogenously Chosen Boards of Directors and their Monitoring of the CEO." American Economic Review no. 88 (1): 96–118.
Hsiao, Cheng. 2003. "Analysis of Panel Data." Cambridge, UK: Cambridge University Press. Hwang, Byong-Hyoun, and Seoyoung Kim. 2009. "It Pays to Have Friends." Journal of
Financial Economics no. 93 (1):138–158. Jensen, Michael C. 1986. "Agency Costs of Free Cash Flow, Corporate Finance, and
Takeovers." American Economic Review no. 76 (2):323–329. Jensen, Michael C., and William H. Meckling. 1976. "Theory of the Firm: managerial
Behavior, Agency Costs and Ownership Structure." Journal of Financial Economics no. 3 (4): 305–360.
Klenke, Karin. 2003. "Gender Influences in Decision-Making Processes in Top Mangement Teams." Management Decision no. 41 (10):1024–1034.
Langli, John C. 2011. "Kjønnskvoterte Styrer." Praktisk Økonomi & Finans (4):67–92. Linck, James S., Jeffry M. Netter, and Tina Yang. 2008. "The Determinants of Board
Structure." Journal of Financial Economics no. 87 (2):308–328. MacAvoy, Paul W., and Ira M. Millstein. 1999. "The Active Board of Directors and its Effect
on Performance of the Large Publicly Traded Corporation." Journal of Applied Corporate Finance no. 11 (4):8–20.
Mason, E. Sharon, and Peter E. Mudrack. 1996. "Gender and Ethical Orientation: A Test of Gender and Occupational Socialization Theories." Journal of Business Ethics no. 15 (6):599–604.
Maug, Ernst. 1997. "Boards of Directors and Capital Structure: Alternative Forms of Capital Restructuring." Journal of Corporate Finance no. 3 (2):113–139.
McKinsey & Company. 2007. "Women Matter: Gender Diversity, A Corporate Performance Driver." McKinsey & Company, Inc.
75
Nygaard, Knut. 2011. "Forced Board Changes: Evidence from Norway." SSRN eLibrary. Powell, Melanie, and David Ansic. 1997. "Gender Differences in Risk Behaviour in Financial
Decision-Making: An Experimental Analysis." Journal of Economic Psychology no. 18 (6):605–628.
Raheja, Charu G. 2005. "Determinants of Board Size and Composition: A Theory of Corporate Boards." Journal of Financial & Quantitative Analysis no. 40 (2):283–306.
Riley, William B., and K. Viktor Chow. 1992. "Asset Allocation and Individual Risk Aversion." Financial Analysts Journal no. 48 (6):32–37.
Shivdasani, Anil, and David Yermack. 1999. "CEO Involvement in the Selection of New Board Members: An Empirical Analysis." Journal of Finance no. 54 (5):1829–1853.
Shleifer, Andrei, and Robert W. Vishny. 1997. "A Survey of Corporate Governance." Journal of Finance no. 52 (2):737–783.
Studenmund, A. H. 2000. Using Econometrics: A Practical Guide. Vol. 4. Boston, MA: Addison Wesley Longman, Inc.
Sundén, Annika E., and Brian J. Surette. 1998. "Gender Differences in the Allocation of Assets in Retirement Savings Plans." American Economic Review no. 88 (2):207–211.
Terjesen, Siri, Ruth Sealy, and Val Singh. 2009. "Women Directors on Corporate Boards: A Review and Research Agenda." Corporate Governace: An International Review no. 17 (3):320–337.
The Guardian. 2012. "Britain's Boardrooms Need More Women, Cameron says." Guardian News and Media Limited (guardian.co.uk).
Trinidad, Cristina, and Anthony H. Normore. 2005. "Leadership and Gender: A Dangerous Liaison?" Leadership and Organization Development Journal no. 26 (7):574–590.
Weisbach, Michael S. 1988. "Outside Directors and CEO Turnover." Journal of Financial Economics no. 20:431–460.
Yermack, David. 1996. "Higher Market Valuation of Companies with a Small Board of Directors." Journal of Financial Economics no. 40 (2):185–211.
76
Figure 1: The fraction of female directors in Norwegian firms exposed to thegender balance law
This figure shows the average ratio of stockholder-elected female directors to all stockholder-elected directors at year end in firms subject to the Norwegian gender balance law. The lawwas passed in December 2003 and was mandatory from January 2008.
0.0
0.1
0.2
0.3
0.4
0.5
2003 2004 2005 2006 2007 2008
Rat
io o
f fem
ale
dire
ctor
s to
all d
irec
tors
Year
77
Table 1: The empirical proxies
Theoretical variable Empirical proxy
Board characteristicsBoard independence The number of outside board members divided by the number of
stockholder-elected board membersOutside director 0/1 dummy variable which is 1 if and only if the board member is either
not a full-time employee in the firm, a former employee, employed by aclosely related firm, related to a member of management, or has nobusiness relationship with the firm
Board size The number of stockholder-elected board membersFemale director number The number of stockholder-elected directors who are womenFemale directors The proportion of stockholder-elected board members who are womenFemale age The average number of years since the female directors were bornMale age The average number of years since the male directors were born
Ownership characteristicsInside concentration The ultimate fraction of equity owned by the firm's officers and directorsOutside concentration The sum of squared ultimate equity fractions in the firm (Herfindahl Largest owner The ultimate equity fraction held by the firm's largest stockholder
General firm characteristicsListed 0/1 dummy variable which is 1 if the firm is public and 0 otherwiseCEO age The number of years since the CEO was bornCEO tenure The number of years since the CEO took officePerformance The average real book return on assets per year over the last three yearsRisk The standard deviation of performance during the last three yearsLeverage Total debt divided by total assetsFirm age The number of years since the firm was foundedFirm size Sales in constant 2009 millions of NOK. Log-transformed in regressions
This table defines the variables used in the empirical analysis.
78
Table 2: Distributional properties of key variables
Variable Mean Std. 0 5 25 50 75 95 100 N
Board characteristicsBoard independence 0.59 0.30 0.00 0.00 0.40 0.60 0.80 1.00 1.00 750Board size 5.69 1.67 1.00 3.00 5.00 5.00 6.00 9.00 12.00 750Female director number 1.51 1.05 0.00 1.00 1.00 2.00 2.00 3.00 5.00 748Female directors 0.24 1.16 0.00 0.00 0.13 0.25 0.38 0.45 0.67 748Female age 45.72 6.20 30.50 35.00 42.00 45.58 49.67 56.00 72.00 630Male age 51.37 5.86 30.50 41.58 47.50 51.50 55.33 61.60 67.50 747Ownership characteristicsInside concentration 0.12 0.21 0.00 0.00 0.00 0.00 0.16 0.59 1.00 750Outside concentration 0.19 0.24 0.00 0.01 0.04 0.09 0.25 0.91 1.00 719Largest owner 0.42 0.31 0.05 0.08 0.15 0.32 0.56 1.00 1.00 620
General firm characteristicsListed 0.64 0.48 0.00 0.00 0.00 1.00 1.00 1.00 1.00 750CEO age 47.09 7.03 30.00 36.00 42.00 47.00 52.00 59.00 72.00 745CEO tenure 5.13 3.87 0.00 1.00 2.00 3.00 6.00 12.00 12.00 511Performance 0.06 0.16 -4.48 -0.22 0.01 0.05 0.12 0.32 0.69 612Risk 0.18 0.47 0.00 0.01 0.03 0.06 0.15 0.39 0.95 630Leverage 0.43 0.30 0.00 0.03 0.22 0.40 0.63 0.91 3.93 630Firm age 25.13 33.44 0.00 1.00 4.00 11.00 29.00 107.80 161.00 743Firm size 2,513 22,256 0 2 33 155 515 4,298 452,370 743Average 696
This table shows distributional properties of the variables used to measure board, ownership, and generalfirm characteristics. Table 1 defines the variables. Performance, Risk, and Leverage are censored at the 1%and 99% tails. The sample consists of all firms exposed to the Norwegian gender balance law from 2003 to2008.
Percentile, %
79
Table 3: Characteristics of listed and non-listed firms
Variable All Listed Non-listed Difference t-value (p-value)
Board characteristicsBoard independence 0.59 0.64 0.50 0.14 5.63 (0.00)Outside director 3.63 3.78 3.36 0.42 9.21 (0.00)Board size 5.69 5.91 4.95 0.96 7.49 (0.00)Female director number 1.51 1.68 1.21 0.47 5.85 (0.00)Female directors 0.24 0.28 0.23 0.05 3.80 (0.00)Female age 45.72 46.22 44.59 1.63 2.87 (0.00)Male age 51.37 52.00 50.25 1.75 3.87 (0.00)
Ownership characteristicsInside concentration 0.12 0.08 0.16 -0.08 -7.99 (0.00)Outside concentration 0.19 0.13 0.32 -0.19 -8.70 (0.00)Largest owner 0.42 0.29 0.63 -0.34 -11.40 (0.00)
General firm characteristicsCEO age 47.09 47.35 46.62 0.73 1.35 (0.18)CEO tenure 5.13 5.26 4.06 1.20 4.18 (0.00)Performance 0.06 0.06 0.07 -0.01 -0.07 (0.95)Risk 0.18 0.11 0.29 -0.18 -3.74 (0.00)Leverage 0.43 0.41 0.46 -0.05 -1.93 (0.06)Firm age 25.13 31.97 12.88 19.09 9.02 (0.00)Firm size 2,513 2,915 2,262 658 3.07 (0.00)Average N 696 418 278 140
This table compares the mean values of board, ownership, and general firm characteristics forlisted and non-listed firms. The differences between mean values across the two groups, their t-values and p-values (in parentheses) are reported in the three right-most columns. Table 1 definesthe variables. Performance, Risk, and Leverage are censored at the 1% and 99% tails. Thesample consists of all firms exposed to the Norwegian gender balance law during the period2003–2008.
80
Table 4: Director types
AverageDirector type 2003 2004 2005 2006 2007 2008 Mean Diff. z p(z) NOutside 0.46 0.52 0.54 0.58 0.64 0.67 0.57 0.21 11.98 (0.00) 427Grey 0.10 0.09 0.09 0.08 0.07 0.06 0.08 -0.04 -3.13 (0.00) 60Inside 0.44 0.40 0.37 0.34 0.29 0.27 0.35 -0.17 -5.16 (0.00) 263N 706 702 757 825 863 645 750 750
Outside Grey InsideYear Listed Non-listed Listed Listed Non-listed2003 0.52 0.41 0.03 0.10 0.45 0.492004 0.56 0.47 0.03 0.09 0.41 0.442005 0.64 0.50 0.02 0.09 0.34 0.412006 0.64 0.52 0.01 0.09 0.33 0.392007 0.71 0.57 0.01 0.08 0.28 0.352008 0.72 0.59 0.00 0.07 0.28 0.34Average 0.63 0.51 0.02 0.09 0.35 0.40Difference 0.12 -0.07 -0.05z-value 15.36 -27.11 -8.12p(z) (0.00) (0.00) (0.00)2008 less 2003 0.20 0.18 -0.03 -0.03 -0.17 -0.15z-value 5.97 4.13 -3.63 -1.23 -5.09 -3.49p(z) (0.00) (0.00) (0.00) (0.22) (0.00) (0.00)
Outside Grey InsideYear Female Male Female Male Female Male2003 0.84 0.49 0.02 0.07 0.14 0.442004 0.83 0.50 0.03 0.08 0.14 0.422005 0.85 0.50 0.02 0.06 0.13 0.442006 0.83 0.48 0.03 0.09 0.14 0.432007 0.83 0.50 0.04 0.08 0.13 0.422008 0.83 0.51 0.04 0.04 0.13 0.43Average 0.84 0.50 0.03 0.07 0.13 0.43Difference 0.34 -0.04 -0.30z-value 30.70 -25.50 -38.40p(z) (0.00) (0.00) (0.00)2008 less 2003 -0.01 0.02 0.02 -0.03 -0.01 -0.01z-value -0.16 0.51 0.63 -0.84 0.00 -0.18p(z) (0.87) (0.61) (0.55) (.040) (0.99) (0.85)
This table reports the average fraction of outside, grey, and inside directors by listing status and gender.Inside directors are the firm’s full-time employees, former employees, or employees of closely related firms.Grey (affiliated) directors are related to a member of management, or are likely to have businessrelationships with the firm. Outside directors are neither inside nor grey. Listed firms are quoted on the OsloStock Exchange. The sample is all firms exposed to the Norwegian gender balance law.
A. All2008 less 2003
Non-listed
B. By listing status
C. By gender
81
Year Mean Median Mean Median Mean Median2003 5.60 5 0.56 0 5.04 52004 5.65 5 0.89 1 4.76 42005 5.77 6 0.88 1 4.89 52006 5.67 5 1.12 1 4.55 42007 5.56 5 1.60 2 3.96 32008 5.86 5 2.25 2 3.61 3Average 5.69 5 1.22 1 4.47 42008 less 2003 0.26 0.00 1.69 2.00 -1.43 -2.00t-value 23.63 0.00 19.59 18.86 -18.06 -15.89(p-value) (0.00) (1.00) (0.00) (0.00) (0.00) (0.00)
Year All Females Males All Females Males2003 5.91 0.61 5.30 5.12 0.51 4.612004 6.11 0.90 5.21 5.11 0.63 4.482005 6.19 1.21 4.98 5.05 0.66 4.392006 6.22 1.53 4.69 4.93 1.04 3.892007 6.12 1.95 4.17 4.70 1.50 3.202008 6.23 2.52 3.71 4.88 1.95 2.93Average 6.13 1.45 4.68 4.97 1.05 3.922008 less 2003 0.32 1.91 -1.59 -0.24 1.44 -1.68t-value 5.83 31.83 -27.33 -4.90 29.39 -34.29(p-value) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)Listed less non-listedt-value(p-value) (0.00)
Table 5: Board size
1.17
This table shows board size as measured by the number of shareholder-elected directors. Thesample is all firms exposed to the Norwegian gender balance law. Listed firms are quoted on theOslo Stock Exchange.
A. All firms
Non-listed firms
B. Mean by listing status
MalesFemalesAll
Listed firms
11.99
82
Table 6: Multiple directorships
Directorships 2003 2006 2008 Average 2003 2006 2008 Average1 124 341 417 294 1,134 1,164 788 1,0292 15 33 47 32 290 93 53 1453 1 16 13 10 24 20 10 184 2 5 3 3.3 4 6 1 3.75 2 2 1.3 2 6 2 3.36 2 0.77 1 0.38 1 0.3 1 0.3All 142 397 485 341 1,454 1,291 854 1,199
Directorships 2003 2006 2008 Average 2003 2006 2008 AverageMean 1.16 1.22 1.22 1.20 1.25 1.15 1.10 1.17Median 1 1 1 1 1 1 1 1Maximum 4 5 8 5.7 5 8 5 62008 less 2003 0.06 -0.15t-value 15.50 -20.50(p-value) (0.00) (0.00)N 165 485 592 414 1,812 1,479 938 1,410
Directorships 2003 2006 2008 Average 2003 2006 2008 AverageNumber 18 23 68 36 320 127 66 171Fraction 0.13 0.06 0.14 0.11 0.22 0.10 0.08 0.142008 less 2003 0.01 -0.14z-value 0.41 -8.75(p-value) (0.68) (0.00)N 142 397 485 341 1,454 1,291 854 1,199
MalesFemalesA. Totals
B. Seats per directorFemales Males
This table shows the total number of directorships and the number per director heldby females and males in 2003, 2006, and 2008. Panel A shows the total number ofdirectors with 1 to 8 seats. Panel B shows the number of seats per director, whilepanel C shows the number and fraction of directors who held more than one seat.The sample is all firms exposed to the Norwegian gender balance law during theperiod 2003–2008.
C. Directors with multiple seats
Females Males
83
Table 7: Bivariate correlation coefficients between the determinants of board independence
Female Female directors Inside Outsidedirectors Listed *Listed concentration concentration Performance Leverage Risk Firm age
Listed 0.141Female directors*Listed 0.612 0.733Inside concentration -0.073 -0.156 -0.147Outside concentration -0.099 -0.052 -0.277 0.203Performance 0.143 0.109 0.035 0.032 0.014Leverage -0.006 -0.085 -0.027 0.035 0.054 0.200Risk -0.233 -0.170 -0.168 -0.014 -0.030 -0.077 0.024Firm age 0.086 0.274 0.237 -0.090 -0.122 0.064 0.143 -0.133Firm size 0.122 0.133 0.149 -0.160 0.077 0.260 0.301 -0.171 0.276
This table shows pairwise Pearson correlation coefficients between the hypothesized determinants of board independence as specified in model (1) ofthe main text. Listed is a dummy variable which is 1 if the firm is public and 0 otherwise. Female directors is the proportion of stockholder-electedboard members who are women. Inside concentration is the ultimate fraction of equity owned by the firm's officers and directors. Outsideconcentration is the sum of squared ultimate equity fractions in the firm (Herfindahl index). Performance at time t is the average real book return onassets per year over the last three years. Leverage is total debt divided by total assets. Risk at time t is the standard deviation of performance overthe last three years. Firm age is the number of years since the firm was founded. Firm size is sales in constant 2009 millions of NOK. Performance,Risk, and Leverage are censored at the 1% and 99% tails. The sample is all Norwegian firms that are exposed to the gender balance law during theperiod 2003–2008.
84
Table 8: Estimates of the base-case model
Determinant Prediction Coefficient (p-value)
Female directors (+) 0.308 *** (0.003)
Listed (+) 0.137 *** (0.003)
Female directors*Listed (+/-) -0.185 * (0.077)
Inside concentration (-) -0.002 * (0.056)
Outside concentration (+/-) -0.076 (0.357)
Performance (-) -0.002 *** (0.007)
Leverage (-) -0.137 ** (0.015)
Risk (-) 0.001 (0.196)
Firm age (+) 0.028 *** (0.000)
Firm size (+) 0.026 ** (0.020)Firm fixed effects YesYear fixed effects Yesp-value (F) (0.000)R2 0.300N 429
This table shows the base-case estimates of model (1) in the main text, which specifies boardindependence as a function of the determinants in column 1. The predicted signs of thecoefficients are shown in column 2, column 3 reports the coefficient estimates, and thecorresponding p-values (in parentheses) are in column 4. Female directors is the proportion ofstockholder-elected board members who are women. Listed is a dummy variable which is 1 if thefirm is public and 0 otherwise. Inside concentration is the ultimate fraction of equity owned by thefirm's officers and directors. Outside concentration is the sum of squared ultimate equity fractionsin the firm (Herfindahl index). Performance at time t is the average real book return on assets per year during the last three years. Leverage is total debt divided by total assets. Risk at time t is thestandard deviation of performance during the last three years. Firm age is the number of yearssince the firm was founded. Firm size is the log of sales in constant 2009 millions of NOK.Performance, Risk, and Leverage are censored at the 1% and 99% tails. Statistical significance at the 1%, 5%, and 10% levels is labeled ***, **, and *, respectively. The sample is all firmsexposed to the Norwegian gender law during the period 2003–2008.
85
Table 9: Alternative econometric techniques
Determinant I II III
Female directors 0.308 *** 0.410 *** 0.410 ***
Listed 0.137 *** 0.094 ** 0.098 **
Female directors*Listed -0.185 * -0.115 -0.119
Inside concentration -0.002 * -0.002 * -0.002 *
Outside concentration -0.076 -0.125 * -0.120 *
Performance -0.002 *** -0.001 ** -0.001 **
Leverage -0.137 ** 0.066 -0.068
Risk -0.001 0.001 0.001
Firm age 0.028 *** 0.000 0.000
Firm size 0.026 ** 0.018 ** 0.017 **
Constant 0.403 *** 0.433 ***
Random firm effects No Yes NoFixed firm effects Yes No NoFixed time effects Yes Yes Nop-value (Wald chi2) (0.000) (0.000)p-value (F) (0.000)R2 0.300 0.220 0.170N 429 429 429
This table shows the base-case estimate from table 8 under technique I. Random-effects ML andOLS with sample averages per firm are used in techniques II and III, respectively. The dependentvariable is board independence, which measures the fraction of outside directors. Female directors isthe proportion of stockholder-elected board members who are women. Listed is a dummy variablewhich is 1 if the firm is public and 0 otherwise. Inside concentration is the ultimate fraction of equityowned by the firm's officers and directors. Outside concentration is the sum of squared ultimateequity fractions in the firm (Herfindahl index). Performance at time t is the average real book returnon assets per year during the last three years. Leverage is total debt divided by total assets. Risk attime t is the standard deviation of performance during the last three years. Firm age is the number ofyears since the firm was founded. Firm size is the log of sales in constant 2009 millions of NOK.Performance, Risk, and Leverage are censored at the 1% and 99% tails. Statistical significance at the1%, 5%, and 10% levels is labeled ***, **, and *, respectively. The sample is all firms exposed tothe Norwegian gender law during the period 2003–2008.
Technique
86
Table 10: Alternative proxies for board independence
Definition of board independence
Determinant Prediction
Female directors (+) 0.576 *** 0.571 *** 0.351 ***
Listed (+) 0.192 *** 0.154 ** 0.125 **
Female directors*Listed (-) -0.294 -0.176 -0.092
Inside concentration (-) -0.001 -0.001 0.000
Outside concentration (+/-) 0.002 0.042 0.043
Performance (-) -0.002 *** -0.004 ** -0.004 **
Leverage (-) -0.106 ** -0.096 ** .0.102 **
Risk (-) -0.097 -0.088 -0.092
Firm age (+) 0.017 ** 0.011 * 0.016 **
Firm size (+) 0.026 ** 0.032 ** 0.027 **
Constant 0.505 ***Random firm effects Yes No NoFixed firm effects No Yes YesFixed time effects Yes Yes Yesp-value (Wald chi2)/(F) (0.000) (0.000) (0.000)R2 0.13 0.28 0.24N 332 429 429
This table shows the estimates using alternative measures of board independence. Model A uses a dummyvariable which equals 0 if the CEO is a board member and 1 otherwise. Model B uses the fraction of outsidedirectors minus the fraction of inside directors, while model C uses the fraction of outside and grey directors.The predicted signs of the coefficients are shown in column 2. Female directors is the proportion of stockholder-elected board members who are women. Listed is a dummy variable which is 1 if the firm is public and 0otherwise. Inside concentration is the ultimate fraction of equity owned by the firm's officers and directors.Outside concentration is the sum of squared ultimate equity fractions in the firm (Herfindahl index). Performanceat time t is the average real book return on assets per year during the last three years. Leverage is total debtdivided by total assets. Risk at time t is the standard deviation of performance during the last three years. Firmage is the number of years since the firm was founded. Firm size is the log of sales in constant 2009 millions ofNOK. Performance, Risk, and Leverage are censored at the 1% and 99% tails. Statistical significance at the1%, 5%, and 10% levels is labeled ***, **, and *, respectively. The sample is all firms exposed to theNorwegian gender law during the period 2003–2008.
A: The CEO is not a director
B: Fraction outside less fraction inside directors
C: Fraction outside and grey directors
87
Table 11: Alternative non-regulatory determinants of board independence
Determinant Prediction (p-value)
Female directors (+) 0.038 ** (0.033)
Listed (+) 0.055 (0.483)
Female directors*Listed (+/-) 0.157 (0.487)
CEO ownership (-) -0.009 *** (0.009)
CEO ownership squared (+) 0.000 ** (0.018)
CEO tenure (-) -0.056 ** (0.011)
CEO tenure squared (+) 2.030 ** (0.042)
Firm age (+) 0.000 (0.999)
Firm age squared (-) 0.000 (0.758)
Board size (+) 0.154 ** (0.016)
Fixed firm effects YesFixed time effects Yesp-value (Wald chi2) (0.000)R2 0.190N 407
This table shows the estimates using alternative proxies for non-regulatorydeterminants of board independence compared to those used for the base case inmodel (1) of the main text. Female directors is the proportion of stockholder-elected board members who are women. Listed is a dummy variable which is 1 ifthe firm is public and 0 otherwise. CEO ownership is the ultimate fraction ofequity held by the firm's CEO. CEO tenure is the number of years since the CEOwas hired. Firm age is the number of years since the firm was founded. Firm sizeis the log of sales in constant 2009 millions of NOK. The predicted signs of thecoefficients are shown in column 2, the coefficient estimates are in column 3, andcolumn 4 reports the corresponding p-values (in parentheses). Statisticalsignificance at the 10%, 5%, and 1% levels is labeled ***, **, and *,respectively. The sample is all firms exposed to the Norwegian gender balancelaw during the period 2003–2008.
Coefficient
88
Appendix 1: Classifying directors as inside, grey or outside: Examples
Employee of Related to a Business Board Full-time Former closely member of relation
Name Gender position employee employee related firms management with the firm Inside Grey Outside
Synnøve Finden ASA, 2007Svein Sundsbø Male Chair xRichard Olav Aa Male Director xMimi K. Berdal Female Director xGro Mykling Female Director xPer Arne Eggen Male EmployeeAnne-Mette Hoel Female Director xGeir Dalsegg Male EmployeeLine Rugsveen Female Employee
Eitzen Chemical ASA, 2006Axel C. Eitzen Male Chair x xJohn G. Bernander Male Director xMai-Lill Ibsen Female Director xJames Stove Lorentzen Male Director xAnnette Malm Justad Female Director x x
Norse Energy Corp. ASA, 2007Petter M. Andresen Male Chair xLise H. Langaard Female Director xJoey S. Horn Female Director xJon-Axel Torgersen Male Director x
Fred. Olsen Production ASA, 2008Per-Oscar Lund Male Chair x xSiv Staubo Female Director xAnette Olsen Female Director x xAngar Gravdal Male Director x x
This table illustrates how we classify a director as inside, grey, or outside. Inside directors are defined as the firm's full-time employees, formeremployees, or employees of closely related firms. Grey directors are related to the firm's management or are likely to have business relationshipswith the firm. Outside directors are neither inside nor grey. The data used to manually classify directors are from Brønnøysundregistrene(www.brrg.no) and Proff (www.proff.no). We obtain supplementary information on director characteristics by manually searching the annualreports. We do not classify employee directors.
89
4.
Determinants of board independence in a free contracting environment
Siv Staubo*
August 28, 2013
Abstract
This paper is the first to explore the demand for monitoring and advice on the board by the owners of firms that are not required by regulation to appoint independent directors. Our focus is on the potential conflict between monitoring and advice and on the idea that the relative value of these two board functions varies across firms. We show that well established, small, and profitable firms with concentrated ownership demand advice from dependent directors more than monitoring of their management from independent directors. Similar results are obtained when we investigate the demand for board independence triggered by the potential conflict between large and small stockholders. Unlike earlier research, we find that female directors are just as likely to be advisors as monitors, particularly in family firms. Our results, which are robust to alternative definitions of board independence and alternative econometric techniques, strongly support the idea that optimal board independence is firm-specific.
Keywords: Corporate governance, Regulation, Board independence, Agency problems,
Ownership concentration, Family firm
JEL classification codes: G30, G38
*Department of Financial Economics, BI Norwegian Business School, N-0442 Oslo, Norway. The email address is [email protected]. I am grateful for valuable comments on earlier drafts from Mike Burkart, Øyvind Bøhren, and seminar participants at BI Norwegian Business School.
90
1. Introduction
The directors’ role as monitors of the CEO is widely described in the research literature
(Shleifer and Vishny 1997, Hermalin and Weisbach 2003, Adams, Hermalin, and Weisbach
2010). In contrast, there is hardly any research on the board’s advisory role. The only
theoretical model that addresses both board roles is Adams and Ferreira (2007), who
recognize the potential conflict between monitoring and advice.28 They show that for CEOs
who dislike monitoring and like advice, the information they provide decreases with the
directors’ monitoring intensity and increases with their advice intensity. Since more
independent directors have stronger incentives to monitor, information production decreases
with increasing board independence. Hence, information provision responds endogenously to
board independence. This relationship means that owners who design board face a tradeoff
between the need for monitoring and the need for advice. Moreover, since the relative value
of monitoring and advice may differ from firm to firm, a board’s optimal independence may
be firm-specific.
This paper explores the determinants of board independence, focusing on the potential
conflict between monitoring and advice. The first premise is the theoretical finding by Adams
and Ferreira (2007) that optimal board independence is firm-specific. The second premise is
the empirical fact that listed firms in most countries cannot freely choose the independence
level of their board because of regulation (ECGI 2012, Sarbane-Oxely 2002). These two
premises suggest that a regulatory floor on board independence is costly for firms as a whole,
and particularly costly for firms when the optimal degree of board independence is
considerably below the mandated one. To ensure that board independence is determined in a
free contracting environment, we analyze non-listed (private) firms. Unlike listed (public)
firms, non-listed firms are not subject to regulation of their board independence. We show
empirically that board independence varies systematically with a series of observable firm
characteristics.
It is a common belief among policy makers and regulators that independent directors reduce
the incidence of corporate scandals. In fact, policy makers and regulators often assume that
good governance more generally requires independent boards (Bhagat and Black 2002, Coles, 28 Other theoretical models also emphasize that directors’ need information in order to be active decision-makers (Raheja 2005, Song and Thakor 2006, Harris and Raviv 2008). However, Adams and Ferreira (2007) provide the first model to address the conflict between information sharing by the CEO and monitoring intensity by the directors.
91
Naveen, and Naveen 2008, Duchin, Matsusaka, and Ozbas 2010). For example, corporate
governance reforms in most countries put particular emphasis on the value of board
independence:
‘The role of independent non-executive directors features prominently in corporate governance codes. The presence of independent representatives on the board, capable of challenging the decisions of the management, is widely considered a means of protecting the interests of stockholders and, where appropriate, other stakeholders.’ European Commission's Recommendation (provisional text), October 6, 2004.
‘The common hallmark of corporate governance reforms proposed by the NYSE,3 NASDAQ4 and AMEX5 is the emphasis placed on a board of directors having the capacity to exercise independent judgment while performing its responsibilities. For example, the NYSE Corporate Accountability and Listing Standards Committee, convened in 2002 to recommend ways to enhance the accountability, integrity and transparency of NYSE-listed companies, stated its belief that having a majority of independent directors would increase the quality of board oversight and lessen the possibility of damaging conflicts of interest.' PriceWaterHouseCoopers: The Sarbanes Oxley Act 2002.
Existing research provides no robust support for these opinions. That is, there is neither
convincing theory nor convincing empirical evidence showing that more board independence
unconditionally improves firm value (Adams, Hermalin, and Weisbach 2010). The reason is
that more board independence brings both benefits and costs, that the costs may dominate the
benefits, and that this relationship between costs and benefits is not constant across firms.
Two recent empirical studies build on the Adams and Ferreira (2007) model, analyzing if
board characteristics like size and independence vary with the firm’s demand for director
expertise on strategic issues (Coles, Naveen, and Naveen 2008, Linck, Netter, and Yang
2008). Both studies find that complex firms have larger boards with more independent
directors, whereas growth firms have smaller boards and a larger demand for advice by
dependent directors. These findings support the idea that board independence is firm-specific.
However, these studies analyze listed firms in the United States, which are required by law to
appoint a majority of independent board members.29 Although the Sarbanes-Oxley Act was
not introduced until 2002, board composition in the United States was influenced by
compliance requirements on board independence introduced many years earlier.30 Similarly,
29 Coles, Daniel, and Naveen (2008) analyze public firms in the United States from 1992 to 2001. Linck, Netter, and Yang (2008) analyze US public firms from 1990 to 2004. 30 The Sarbanes-Oxley Act (SOX) is a federal law that set new or enhanced standards for all boards in public firms. Prior to SOX, board independence was not mandated, but the SEC encouraged firms to increase board
92
European stock exchanges recommend by soft law that listed firms compose boards with at
least 50 percent independent directors. 31 Thus, empirical studies of such regulatory
environments investigate determinants of board independence in firms that are required by
law or recommended by code to appoint a majority of independent directors. That is, the firm
cannot set its optimal board independence level unless this level is at or above the regulatory
minimum, i.e., unless the regulation is not binding relative to the first-best optimum. Since the
firms analyzed in the existing literature are regulated to appoint a majority of independent
directors, this literature may not reveal the relationship between firm characteristics and board
independence that would be observed in a free contracting environment.
Our study avoids this problem. We use a dataset consisting of 16,100 non-listed Norwegian
firms from 2000 to 2011. Board independence is not regulated in these firms in any part of the
sample period. There is a large variation in firm characteristics across the sample. This
heterogeneity produces large cross-sectional variations in the determinants of optimal board
independence. Hence, unlike earlier papers, we explore the determinants of board
independence in firms where stockholders are free to construct the optimal board from a
monitoring and advice perspective. The average fraction of independent directors is close to
30% in the sample period, compared to more than 50% in listed firms, where independence is
regulated.
We make four contributions to the literature. First, unlike existing research, we investigate the
trade-off between monitoring and advice in a setting where board independence is not
regulated. Second, we show that board independence is more important to some firms than to
others. This result suggests that one-size-fits-all regulation is costly.
Our third contribution is to explore the board’s monitoring role not just relative to the CEO,
but also relative to potential conflicts between large and small stockholders. The first of these
two monitoring functions is to reduce the principal-agent problem that arises when managers
exploit their control rights at the stockholders’ expense (Jensen and Meckling 1976). This
situation is called the first agency problem (Agency problem 1) in the literature, and directors
who are independent of management are supposed to be better at reducing this problem
(Villalonga and Amit 2006). The existing literature on board independence focuses independence. For example, Linck, Netter, and Yang (2008) notice that Harold Williams, the SEC chairman from 1971 to 1981, placed significant pressure on NYSE firms to have a majority of outside directors on their boards. 31 A soft law is based on the principle of comply or explain.
93
exclusively on the first agency problem. The board’s second monitoring function is to oversee
the conflict between majority and minority stockholders, which has been called the second
agency problem (Agency problem 2) in the literature (Villalonga and Amit 2006). Directors
who are independent of the main stockholder are supposed to be better at protecting the rights
of minority stockholders.32
We measure board independence relative to the first monitoring function (Agency problem 1)
in terms of the arm’s-length distance between directors and management. We use standard
empirical proxies from existing papers (Carter and Lorsch 2004, Bøhren and Strøm 2010).
Moreover, we establish a new empirical measure related to the monitoring function caused by
the second agency problem. As far as we are aware, we are the first to study the second
monitoring function in a board independence setting.
There is large variation in ownership concentration in our dataset. This fact allows us to
separate Agency problems 1 and 2. Agency problem 1 is expected to be most prominent when
ownership is dispersed, whereas Agency problem 2 is more likely to occur when ownership is
concentrated. Therefore, we divide our data set into two subsamples according to ownership
concentration in order to improve insight into firm-specific demand for board independence.
In the first subsample, the firms have no controlling owner. Hence, the firms in this
subsample are primarily exposed to the first agency problem. The firms in the second
subsample are controlled by one owner. Because these firms are majority owned, there is a
potential conflict between majority and minority stockholders, which creates the second
agency problem. The conflict between owners and management is expected to decrease as the
equity stake of the largest owner increases. That is because the largest owner has strong
power and incentives to control the management. Hence, the first agency problem may not be
as prominent in the second subsample as in the first. We exclude single-owner firms since
these firms have no majority-minority problem and seldom face a separation between
ownership and control.
Our forth contribution is to investigate board independence in family firms. The relationship
between board independence and information provision as addressed by Adams and Ferreira
(2007) needs specific attention in these firms. The owners of family firms are often managers 32 The Norwegian Corporate Governance Codes explicitly address these two dimensions of independence: ”The majority of the stockholder-elected members of the board should be independent of the company’s executive management and material business contracts” and ”At least two of the members of the board elected by stockholders should be independent of the company’s main stockholder(s)” (NUES 2012).
94
and directors, and they are often particularly well informed about the business. The existing
literature suggests that family firms either mitigate or exacerbate agency problems. Anderson
and Reeb (2003b) and Andre and Ben-Amar (2006) argue that families are better monitors of
management than other stockholders because owners and managers are better aligned in
family firms than in non-family firms. However, expropriation of minority stockholders’
wealth by the controlling family is probably the most critical agency cost in family firms
(Claessens et al. 2002, Masulis, Pham, and Zein 2011). Our dataset contains a comprehensive
set of family characteristics, and 82% of the firms in our sample are controlled by a family.33
Our results show that the optimal level of board independence is lower than 50% for more
than half the firms in this study. That is, when we investigate a sample of non-listed firms, we
find that recommending a majority of independent directors to all firms will hurt firms in
which stockholder value advice by the board of directors more than monitoring. In particular,
we show this is the case for well established, small, and profitable firms with concentrated
ownership. Unlike earlier research the relationship between firm age and board independence
is negative and the relationship between female directors and board independent is not
significant in the study. We find similar results when we explore the determinants of board
independence related to the second agency problem. Since we do not know any other studies
that analyses the determinants of board independence related to independence between board
members and the largest owner, we believe that these results are new in literature. Finally, we
find that family firms have less independent boards than non-family firms.
Section 2 in the following reviews the literature and specifies the predictions. The data and
summary statistics are discussed in Section 3, while Section 4 explains the methodology and
presents the results. We estimate the base-case model by alternative econometric techniques
in section 5, where we also use an alternative proxy for board independence as well as
alternative determinants of board independence. Section 6 summarizes and concludes.
2. Theory and predictions
Even though our paper is based on a theoretical model of board composition (Adams and
Ferreira 2007), there is not much theory on this topic compared to the empirics (Hermalin and 33 We define a firm as a family firm if the largest family by ultimate ownership holds more than 50% of the firm’s equity. A family is a group of individuals related by blood or marriage.
95
Weisbach 2003). Hermalin and Weisbach (2003) divide the empirical research on the board of
directors into three groups. The first group studies the impact of board characteristics on
board behavior, such as the relationship between board independence and CEO replacement.
The second group studies whether boards’ actions influence firm performance. Our study
belongs to the third group, which analyzes the determinants of board composition. For
instance, Weisbach (1988), Denis and Sarin (1999) and Hermalin and Weisbach (2003) find
that small, young, and closely held firms tend to have less independent boards. In contrast,
larger, older, and dispersedly owned firms are more likely to have more independent boards.
We want to investigate the impact of these and other determinants on board independence
used in earlier research (Linck, Netter, and Yang 2008, Adams 2009). Unlike earlier studies,
however, our sample firms can choose the level of board independence without concerns for
regulatory restrictions.
Our base-case model is the following:
1 2
3 4 5
6 7 8
+ +
it it it
it it it
it it it
Independence CEO ownership PerformanceLeverage Female directors GrowthInformation costs Firm age Firm size
α β ββ β ββ β β
= + ++ +
+ + + itu (1)
The empirical variables are defined in Table 1.
Table 1
Most of the explanatory determinants used in empirical governance studies are endogenous.
Section 5 addresses endogeneity problems in model (1).
We define measures of board independence in section 2.1. Section 2.1.1 defines the measure
for board independence related to Agency problem 1, whereas section 2.1.2 defines measure
for board independence related to Agency problem 2. We predict how board independence
relates to its determinants in section 2.2.
96
2.1 Measures of board independence
2.1.1 Independence – Agency problem 1 (Independence-A1)
There are several definitions of board independence in the corporate governance literature.
The following definition is from Wikipedia:
An Independent Director (also sometimes known as an outside director or non-executive director) is a director (member) of a board of directors who does not have a material or pecuniary relationship with the company or related persons, except sitting fees. Independent Directors do not own shares in the company.
Earlier research measures board independence in different ways. One of the most common
measures is the fraction of outside directors on the firm’s board (Baysinger and Butler 1985,
Weisbach 1988, MacAvoy and Millstein 1999, Adams and Ferreira 2009b).34 In this context
board independence refers to independence between directors and management. A board is
more independent the higher the fraction of outside directors. An independent board is
assumed to be better at reducing the first agency problem.
We do not have the data to measure the fraction of outside directors. Instead, we follow
earlier research by using a different measure. In the base-case model we use a proxy that
identifies whether or not the CEO is a board member. That is, although the CEO is always
present in the board meetings, this proxy measures whether or not the CEO has voting rights.
A board is expected to be more independent if the CEO has no voting right because he is then
unable to interfere formally with the directors who monitor him (Carter and Lorsch 2004).
We measure board independence by a dummy variable which is 1 if the CEO is not a director
and 0 otherwise.
34 To find the fraction of outside directors, earlier research classifies directors as inside, grey or outside. Inside directors are defined as the firm’s full-time employees, former employees, or employees of closely related firms. Grey (affiliated) directors are related to a member of management, or are likely to have business relationships with the firm, such as investment bankers and lawyers. Outside directors are neither inside nor grey. Moreover, a board is more independent the higher the fraction of outside directors.
(a)1 if the CEO is not a board member
0 if the CEO is a board memberIndependence A1 =
97
A board is more independent if (a) is 1. We use an alternative proxy for Independence - A1
under the robustness checks in Section 5.
2.1.2 Independence – Agency problem 2 (Independence - A2)
The board’s second monitoring function is to protect the minority stockholders. To address
the second agency problem, we are concerned with independence between the board and the
large stockholders. A board is more independent the larger the fraction of directors who are
independent of the largest owner. Such a board is supposedly better at protecting minority
stockholders. We have developed a new measure to capture independence between the
directors and the largest stockholder, using the fraction of directors that do not belong to the
largest family by ultimate ownership.
the number of board members - the number of board members from the largest familyIndependence - A2 = the number of board members
(b)
Independence – A2 measures the fraction of board members not connected to the largest
owner. This measure is justified by the fact that 82% of the firms in our data set have a family
as the largest owner.
The values of (b) will be in the interval [ ]0,1 . A board is assumed to be more independent the
closer (b) is to 1.
2.2 Determinants of board independence
We discuss the determinants of board independence according to the two main roles of the
board, which are monitoring (2.2.1) and advice (2.2.2).
The determinants of board independence described in this section are determinants of board
independence related to the first agency problem. These determinants are used in earlier
research. The existing research on board independence is concerned about independence
between board members and management. When we test board independence relative to the
second agency problem, we use a new empirical measure the captures independence between
board members and the largest owner as the left-hand side variable. However, we use the
same determinants of board independence as we do when we analyze board independence
98
related to Agency problem 1. That is, without support from either theoretical or empirical
research, we test if the same determinants influence board independence between directors
and managers and board independence between directors and the largest owner.
2.2.1 Monitoring
Earlier studies show empirically that board independence declines when the equity holdings
of the CEO increases (Bhagat and Black 1998, Linck, Netter, and Yang 2008). The theoretical
argument is that the need for monitoring is smaller when the CEO is aligned with
stockholders (Morck, Shleifer, and Vishny 1989). We expect CEO ownership to be negatively
related to board independence (H1).
Firms tend to appoint more independent directors after years of poor performance (Hermalin
and Weisbach 1991, Bhagat and Black 2002). We measure performance as the average return
on assets over the last three years, and we expect performance to be negatively correlated to
board independence (H2).
Firms with high debt have less free cash flow. Hence, high debt is a substitute for monitoring
effort by owners and directors (Jensen 1986). We measure debt by leverage, which is total
debt divided by total assets. This argument suggests an inverse relationship between leverage
and board independence. However, using leverage to proxy for firm complexity, Linck, Netter,
and Yang (2008) find that complex firms have more independent boards. Hence, they find
that leverage is positively related to board independence. Due to this contradictory evidence,
the relationship between leverage and board independence is unspecified (H3).
Adams and Ferreira (2009b) argue it is a general belief that diverse boards are more
independent. This view is supported by Beecher-Monas (2007) and Fields and Keys (2003),
who suggest that independence is easier to achieve by increasing ethnic and gender diversity
in the board room. We use the fraction of female directors to proxy for board diversity,
expecting that the fraction of female directors is positively related to board independence
(H4).
2.1.2 Advice
Jensen (1993) argues that it is more costly for large boards to monitor growth firms. Along
the same lines, Linck, Netter, and Yang (2008) find that both board size and board
99
independence decrease with the costs of monitoring. We use sales growth to proxy for growth,
and expect growth to be negatively related to board independence (H5).
Theoretical research by Adams and Ferreira (2007) show that the optimal fraction of
independent directors decreases as the cost of monitoring increases. Such monitoring costs are
particularly high when firms with strong information asymmetry are monitored by
independent directors (Maug 1997). Moreover, empirical research has found that the
information asymmetry is higher the more volatile the firm’s stock returns (Fama and Jensen
1983). Linck, Netter, and Yang (2008) use the standard deviation of stock returns to proxy for
information costs. Similarly, we use the standard deviation of the book return on assets to
proxy for information costs, and expect information costs to be negatively related to board
independence (H6). 35
Earlier research finds that complex firms have more independent boards (Lehn, Patro, and
Zhao 2003, Boone et al. 2007, Linck, Netter, and Yang 2008). Lehn, Patro, and Zhao (2003)
and Boone et al. (2007) use firm age and firm size to proxy for firm complexity. We follow
this earlier work and expect firm age and firm size to be positively correlated with board
independence (H7 & H8).
We extend the base-case model by including family control as a determinant for board
independence. A firm is defined as being a family firm (measured by ‘Family control’) if one
family owns more than 50% of the equity. Family firms differ from other firms in ways that
influence their governance structure (Anderson and Reeb 2003). Family members often hold
the CEO and chair positions. Since the families are in the business for a long time, they are
also often large and committed owners. These characteristics suggest that there are more often
strong ties between management and owners in family firms than in other firms. Therefore,
monitoring of Agency problem 1 is less important. On the other hand, earlier research shows
that the common succession of management positions in family firms may result in
managerial entrenchment, and hence higher agency costs (Gomez-Mejia, Nufiez-Nickel, and
Gutierrez 2001, Anderson and Reeb 2004). We do not pre-specify the relationship between
family control and board independence (H9).
35 We do not have data on stock returns since the firms in our sample are non-listed firms.
100
Summarizing section 2, we predict that the board will be more independent when the firm is
large, old, has many female directors, has low performance, growth, and information costs,
and when CEO ownership is low.
3. Data, sample, and summary statistics
Our sample is non-listed Norwegian firms with limited liability from 2000 to 2011. The data
source is the CCGR database (www.bi.edu/ccgr).36 The data on family relationships is from
the tax authorities. The law mandates a standardized set of full accounting statements certified
by a public auditor regardless of the firm’s listing status, size, and industry. They must also
report the identity of the CEO, directors and owners, as well as these individuals’ equity
holdings in the firm.
There are on average about 200,000 non-listed firms in the population. The vast majority of
the firms are of no interest to this study because they have no activity. We remove these firms
by the filter revenues ≥ 2,000,000 NOK (approximately 260,000 EUR). Next, because listed
firms in Norway are required to have at least three board members, we require a Board size ≥
3. Finally, we exclude firms in which the largest owner holds more than 90% of the equity.37
We are left with 16,100 firms on average each year from 2000 to 2011. Table 2 summarizes
the key properties of the frequency distributions for each of the determinants of board
independence used in this study.
Table 2
According to Table 2, board independence, measured with a dummy variable with a value of
1 if the CEO is not a board member and 0 otherwise is 48%. This is board independence
related to Agency problem 1. One must not be misleading to interpret from this value that
there are 48% independent directors on average in our sample firms. The value of this proxy
for board independence tells that the CEO is not a board member in 48% of the sample firms.
Listed firms are not included in this study. Nonetheless, we compare the values of the
variables in our sample with values of these variables in a sample of listed firms during the
same time period. It is important to notify the differences and similarities between listed and
36 The database includes every limited liability firm registered in Norway from 1994 to present. 37 We use 90% rather than 100% as the upper threshold because 90% is where the corporate law allows the majority to buy out the minority. Furthermore, if the majority owns more than 90%, the minority stockholders may demand the majority to buy them out. See Bøhren and Krosvik (2013) for details.
101
non-listed firm in order to compare our results to existing research and to justify that our
findings are important to all firms, see appendix 1. The CEO has no voting rights in 92% of
listed firms. However, this value is misleading because it exaggerates board independence
because the CEO was not allowed to be a board member in listed firms after 2007. However,
we can avoid this problem by comparing the values for listed and non-listed firms by the
measure we use for board independence for robustness test in Section 5. We find that board
independence is about 25 percentage points higher in listed than in non-listed firms on
average.
Similarly, we find that board independence related to Agency problem 2 is roughly 30
percentage points higher in listed than in non-listed firms.
For a better understanding of the difference in governance problems across firms, the base-
case sample is divided into subsamples. The first subsample is firms with and without a
controlling owner, whereas the second subsample consists of family firms and non-family
firms. The mean values of the variables in the two groups within each subsample are
compared by t-statistics in Table 3, panel A and panel B, respectively.
Table 3
Board independence related to Agency problem 1 is 50% in firms with majority owner and 46%
in firms with no-majority owner. On the other hand, board independence related to Agency
problem 2 is 54% in firms with majority owner and 60% in firms with no-majority owner.
CEO ownership is much greater in majority-held firms than in firms without a majority owner.
The difference in ownership structure in the two subsamples indicates, in accordance with
previous research and our prediction, that boards are more independent in firms with no-
majority owner. In contrast, our predictions are not supported by the findings that firm size
and the fraction of female directors are smaller in firms with no-majority owner. It is
important to point out that these results have to do with the average firm in each subsample,
and that the analysis is univariate. The t-statistics support the notion that board independence
is firm-specific.
The second subsample is family firms and non-family firms. We that 82% of the firms are
family firms. Board independence measured related to Agency problem 1 is 39% in firm with
102
family firms and 88% in non-family firms. The average value of board independence related
to Agency problem 2 is 54% in family firms and 77% in non-family firms. Hence, the
univariate analysis shows that boards in family firms are by far much more independent than
firms in non-family firms. These results can be driven by the large difference in sample size
of family firms and non-family firms and the average values may be strongly influenced by
some firms in the sample of non-family firms. However, family firms perform better, have
higher leverage; are smaller and younger than non-family firms on average.
Table 4 shows bivariate correlation coefficients of the determinants for board independence
included in our base-case model (1). Mulitcollinearity should not be a serious problem,
because the correlation coefficients are far away from the critical limit of 0.8 (Studenmund
2000).
Table 4
4. Research design, methodology, and estimation results
We analyze the estimates of the base-case model according to the two main roles of the board
of directors, which are monitoring and advice. The monitoring role is divided into monitoring
of Agency problem 1 and monitoring of Agency problem 2.
The first estimates of the base-case model (1) are estimates of board independence related to
Agency problem 1. The dependent variable equals 1 if the CEO is not a board member and 0
otherwise. The model is estimated with a logit model and uses GLM. Table 5 shows the
estimates.
Table 5
The estimates for CEO ownership and performance are consistent with our predictions from
H1 and H2. The demand for monitoring carried out by independent directors is not as
prominent when CEO ownership is high because the interests of stockholders and the CEO
are well aligned (H1). When firms perform well, stockholders tend to value advice from the
board of director’s more than monitoring (H2).
103
We did not specify a predicted sign for H3 due to conflicting theoretical arguments for the
impact of leverage on board independence in earlier research. We find that leverage is
negatively related to board independence in the base-case model, indicating that leverage is a
substitute for monitoring rather than a proxy for complex firms. That is, high leverage reduces
the free cash flow and thereby the principal-agent conflict. Hence monitoring by independent
directors is not prominent when leverage is high.
Earlier research suggests that female directors are more likely to be independent directors. A
recent study confirms this presumption, showing a high correlation between board
independence and female directors when boards are subject to gender quotas (Bøhren and
Staubo 2013). The negative and significant coefficient for the fraction of female directors on
board independence in table 5 is new in this literature, and contradicts to H4. One possible
interpretation of this result is that female directors are more often appointed to fill the
advisory role than the monitoring role in this sample of firms where board independence is
not regulated. In addition, unreported statistics show that the distribution of female directors
in our sample is different from the one in listed firms. About half the non-listed firms have no
female directors. The negative relationship between female directors and board independence
is confirmed when the base-case model is estimated in a sample consisting of firms with at
least one female director. Hence, female directors tend to be appointed more often as advisers
rather than monitors.
Growth and information costs are positively related to board independence, but the estimates
are not significant. These results are inconsistent with H5 and H6.
Firm age is negatively related to board independence, indicating that well established firms
demand more advice than monitoring. This is inconsistent with earlier research and our
hypothesis H7. One reason may be that the firms in our sample on average younger than firms
used in earlier research on listed firms (12 years compared to 34 years, 95% of the firms in
our sample are younger than 34 years), see appendix 1. We do not have any theoretical
argument to support this result.
Firm size relates positively to board independence. This result is consistent with H8.
104
Summarizing so far, estimates of the base-case model suggest that board independence related
to Agency problem 1 decreases in CEO ownership, performance, leverage, female directors,
and firm age. Board independence increases in firm size.
Next, we investigate the determinants of board independence relative to the second agency
problem. The dependent variable is a proxy for independence between board members and the
majority owner. This proxy measures the fraction of non-family directors on the firm’s board.
The base-case model is estimated by a random effects panel data regression because the left
hand side variable is a continuous variable. Table 6 shows the estimates.
Table 6
Board independence related to Agency problem 2 decreases in CEO ownership, performance,
female directors, and firm age. Board independence increases in firm size. The results are
similar to the results for the estimates of the determinants of board independence related to
the first agency problem shown in Table 5. The only difference is that leverage is only a
significant determinant for board independence related to the second agency problem at the 10%
level. That is, leverage may not be a substitute for monitoring related to the conflict between
majority and minority stockholders.
Next, we expand the base-case model and include family control as a determinant of board
independence. We define family firms by the variable Family control, which is 1 if the
ultimate family ownership > 50%, and zero otherwise. This criterion classifies 82% of the
sample firms as family firms. We did not pre-specify the relationship between board
independence and family control in the H9.
The estimate of this variable in tables 5 and 6 (rightmost column) shows a negative and
significant impact of family control on board independence. This result suggests that owners
and management are strongly aligned in family firms such that monitoring by independent
directors is less important. Additionally, the family control variable has similar impact on
board independence related to the second agency problem.
Finally, the base-case model is re-estimated in two sets of subsamples. Table 7 shows the
estimates.
Table 7
105
The first set of subsamples is firms with majority owner and firms with no-majority owner.
The results, which are reported in Table 7, Panel A, are consistent with the results from the
full sample. Hence, the same determinants influence board independence related to both
agency problems in the two subsamples.
The second set of subsamples is family firms and non-family firms. The results, which are
reported in Table 7, Panel B, are close to those in the full sample. The only exception is the
fraction of female directors in non-family firms which is positively related to board
independence related to Agency problem 2.
Hence, the inverse relationship between female directors and board independence, which is
new in literature, primarily occurs in family firms. Except for this, the results regarding the
determinants of board independence in subsamples are consistent with those in the full sample.
5. Robustness
We first re-estimate the base-case model in (1) using alternative econometric techniques.
Subsequently, we analyze what happens when we use an alternative proxy of board
independence. Finally, we test the sensitivity to using non-linear empirical proxies in addition
to those used in (1) as determinants of board independence.
5.1 Alternative econometric techniques
First, the base-case model is estimated for board independence related to Agency problem 1.
In addition to the logit regression (the base-case), we apply probit, a standard panel data
method with random effects and with board independence considered as a continuous variable,
a logit panel data method with random effects, and with board independence considered as a
logistically distributed variable. Moreover, we apply pooled OLS with standard errors
adjusted for clustering at the firm level which treats board independence as a continuous
variable. Finally, we estimate model (1) using instrumental variable technique. The results are
reported in Table 8.
Table 8
The results are insensitive to the alternative econometric techniques except for the fraction of
female directors, which is significant only in the logit and probit regressions. Panel data
106
regression and instrumental variable technique control for the potential endogeneity problems
in the base-case model. Endogeneity problems are discussed in more details below. We
believe that estimates of the fraction of female directors in panel data regression and the IV
regression are more reliable than the estimates from the logit and probit models since these
latter models do not control for endogeneity problems. This assumption is confirmed when we
analyze the estimates of the model for board independence related to Agency problem 2. In
addition to standard panel data method with random effects (the base-case), we also apply
standard panel data method with fixed effects, pooled OLS with standard errors adjusted for
clustering at the firm level. 38 Finally, we estimate the model using instrumental variable
technique. The results are reported in Table 9.
Table 9
The results are insensitive to the alternative econometric techniques, except for the fraction of
female directors that is insignificant in the regressions when we apply fixed effects panel data
method and instrumental variable technique. This result suggests that the base-case model is
exposed to endogeneity problems and the fraction of female directors is not related to board
independence. Moreover, females are just as likely to be appointed as advising directors as
monitoring directors.
38 The data set has a panel structure of N x T, where N is the number of firms per year and T is the number of years. There is on average about 16,100 firms per year, which are observed at least once over 11 years. General firm, ownership, ownership, and family characteristics are observed up to 12 times for the same firm. Some firms are observed less than 12 times because they enter the sample after 2000, or left the sample before 2011. Hence, the panel is unbalanced. Compared to studying the cross-section at a given point in time, panel data may improve estimation quality by increasing the number of data points and reducing multicollinearity. Moreover, fixed effects and random effects techniques reduce misspecification problems caused by unobservable determinants of board independence (Hsiao 2003). This property can be seen by separating the error term into one component which is time-invariant and firm-specific, one which is time-variant and firm-independent, and a third which is idiosyncratic and varies within and between firms. In such a context, the model can be specified as:
it it i t itY X uα β δ ψ= + + + + , Yit is the observed board independence measure for firm i at time t, itX is
the vector of observable, time-varying determinants of board independence for i at t, iδ is the unobserved, time-
independent effect of firm i on Yit, ψt is the unobserved, time-varying effect on all firms at time t, and itu is the
idiosyncratic error term. In our case, iδ represents any firm-specific need for monitoring and advice that is not reflected in the firm characteristics we explicitly account for through Xit.
The two alternative ways to account for
iδ are the fixed effect and the random effects approach, respectively. The fixed effects model allows for the unobserved variables to be correlated with the error term, while the random effects model assumes the two are independent and treats the unobservable firm effect as a random variable.
107
As mentioned above, we are aware of the possible endogeneity problems in the regression
models. In fact, we believe that the inconsistent results of the estimates for the fraction of
female directors may be caused by such problems. The problems occur because there may be
causality running from the left hand side variable to some of the explanatory variables. One
way to reduce this problem is to lag the variables we are concerned about. First, consider the
causality between board independence and performance. Firms may tend to appoint more
independent directors after years of poor performance. That is, board independence may
increase if the firm performed bad last year. Therefore, we refer to last year’s performance
because there will always be a time lag between recognition of bad performance and the
opportunity to change the board. Usually, there is one ordinary general meeting a year and a
change in the board has to be accepted by the general meeting. Moreover, to avoid that the
temporary fluctuations in the performance of the firm have any impact on board independence,
we measure performance as average performance over the three last years. By doing so, we
prevent that board independence and performance is contemporaneous.
Second, we are concerned about possible reverse causality between board independence and
CEO ownership. In particular, a firm with a powerful CEO needs advice from the directors
more than monitoring from independent directors. Hence, when CEO’s ownership is large, he
is assumed to be powerful and there is less need for monitoring by the board of directors. To
reduce this causality problem, we study the impact of CEO ownership on board independence
using an instrumental variable (IV). That is, we predict the variation in CEO ownership using
CEO age, which is randomly distributed relative to board independence but not relative to
CEO ownership. The CEO age is limited to an age interval from 17 to 93 years, with an
average close to 47 years and a standard deviation close to 10 years. We define a CEO as
young if his age is lower than the average age of CEOs in our sample, 47 years. A CEO is old
if he is 47 years or older. CEO ownership increases in CEO age, see appendix 2, panel A. We
show that that board independence is consistent for young and old CEOs in appendix 2, panel
B. That is, CEO age is correlated with CEO ownership, but it is uncorrelated with board
independence, which makes CEO age a valid instrument for CEO ownership. Furthermore,
appendix 2, panel B shows that other differences in firm and board characteristics for young
and old CEOs are small compared to the difference in CEO ownership between young and old
CEOs. That is, we find no evidence that firm and board characteristics, other than CEO
ownership, vary as a function of our instrument.
108
5.2 Alternative proxy for board independence
The base-case model is estimated using an alternative proxy for board independence related to
Agency problem 1. This proxy is developed by Bøhren and Strøm (2010) based on the
Hermalin and Weisbach (1998) model where board independence depends on the time of
entry of the directors relative to the entry of the CEO. The theory predicts that a CEO who
runs a firm that performs well is able to recruit dependent directors. Thus, the level of board
independence may depend on whether the directors were appointed before or after the CEO
took office. Board independence in firm i is defined as the difference between the average
tenure of the board’s non-CEO directors and the tenure of the CEO:
1
1Independence - A1 non-CEO director tenure CEO tenuren
n
i ij ij=
≡ −∑ (c)
Non-CEO director tenureij is the number of years since non-CEO director j entered office in
firm i and n is the number of stockholder-elected directors. The board is more independent the
higher the value of (c).
We estimate this model, where the left-hand side variable is continuous, using random effects
regression. Table 10 shows the estimates.
Table 10
The estimates are consistent with the base-case model except for female directors which has
positive but not significant impact on board independence.
5.3 Non-linear determinants of board independence
As mentioned in section 2, earlier research finds that board independence declines in CEO
ownership (Boone et al. 2007, Coles, Naveen, and Naveen 2008, Linck, Netter, and Yang
2008, Shivdasani and Yermack 1999, Weisbach 1988). However, Adams and Ferreira (2009a)
show theoretically that the evidence on an inverse relationship between board independence
and CEO ownership should be reconsidered. The authors also find empirically that board
independence does indeed correlate non-monotonically with CEO ownership since board
independence first decreases and then increases as CEO ownership grows.
109
Futhermore, Boone et al. (2007) argue that complex firms need more independent boards and
they use firm age as a proxy for complexity. But they suggest that firm complexity may not
increase linearly with firm age. Similarly, Linck, Netter, and Yang (2008) suggest that it is
not clear that a firm is more complex once it has matured. They inculde firm age squared as a
determinant of board independence to test if the relationship is nonlinear, and find that the
sign of the linear term changes even though firm age squared is not significant.
To analyse if these findings have any impact on demand for monitoring and advice, we
specify model (2) which includes non-linear determinants.
1 2
3 4 5
6 7 8
+ + costs +
it it it
it it it
it it
Independence CEO ownership CEO ownership squaredPerformance Leverage Female directorsGrowth Information Firm
α β ββ β ββ β β
= + ++
+ +
9 10 it
it it it
ageFirm age squared Firm size uβ β+ + +
(2)
In the first line of the explanatory variables in (2) we follow by Adams and Ferreira (2009a)
by including the squared value of CEO ownership. Lines 2–4 follow the base-case model,
while we also include the quadratic term of firm age as suggested by Linck, Netter, and Yang
(2008).
Model (2) uses the same proxy for the left hand side variable as the base-case model. The
model is estimated as a logit model, using GLM. That is, model (2) estimates determinants for
board independence related to the first agency problem. Table 11 shows the results.
Table 11
Our estimates for CEO ownership and CEO ownership squared are consistent with the
predictions based on earlier research showing that the relationship between CEO ownership
and board independence is not monotonic. That is, board independence first decrease and then
increases in CEO ownership.
The estimates for firm age and firm age squared are inconsistent with our predictions guided
by earlier research on listed firms. The average firm age is lower in our sample of non-listed
firms, (12 years) than in listed firms (34 years). This difference may explain why firm age is
negatively related to board independence, while firm age squared is positively related to board
independence. The estimates of the determinants used in the base-case model, (performance,
110
leverage, female directors, growth, information costs, and firm size) are consistent with the
base-case results.
According to the robustness checks, the base-case results are insensitive to alternative
econometric techniques and to how we define board independence. The only result that needs
to be reconsidered is the estimate for female directors which is negative in the base-case
model. When using other econometric techniques, including instrumental variable technique
(IV), however female directors is not significantly related to board independence. That is, the
gender of the directors has no impact on board independence. This result is new in literature.
The result is confirmed when we test an alternative proxy for board independence related to
the Agency problem 1. We believe that the negative relationship between female directors and
board independence is caused by endogeneity problems in the base-case model. The fixed
effect regression and the IV technique both reduce these problems. Taking into account that
none of the other determinants of board independence change when we use fixed effect
regression and IV technique, we conclude that the fraction of female directors has no impact
on board independence when board independence and gender balance are not regulated.
Finally, the results from the robustness section support earlier research showing that the
relationship between board independence and CEO ownership is non-monotonic, the
relationship is V-shaped.
6. Conclusion
This paper investigates the owners’ demand for monitoring and advice by the board of
directors in a large sample of firms that are not expected by regulation to appoint a minimum
fraction of independent directors. Our motivation is that earlier theoretical research shows that
board independence is firm-specific. That is, the demand for independent directors should
optimally vary from firm to firm.
The existing empirical research on determinants of board independence studies samples of
firms that are required to appoint at least 50% independent directors. Therefore, these studies
may be biased because some firms’ optimal board independence in a free contracting
environment is below the regulatory floor. We avoid this problem, as the firms in our sample
111
are not subject to any laws or corporate governance codes that restrict the board’s
independence.
Our findings show that firms that are well performing, small, well-established, and with high
leverage, are better off if the stockholders appoint directors who focus more on the advisory
role than the monitoring role. These results support the notion that the demand for monitoring
and advice is firm-specific.
Furthermore, we find that the relationship between CEO ownership and board independence
is non-monotonic as shown. In particular, board independence is first declining in CEO
ownership, but when CEO ownership increases the demand for monitoring by the board of
directors increases. Once more, this result supports the idea that the optimal level of board
independence varies across firms and even in a non-linear fashion.
We extend previous research by investigating board independence relative to the second
agency problem that concerns the independence between the directors and the firm’s main
stockholder. The estimates of the determinants for board independence relative to the first and
second agency problem are close to identical. We find that the same firm characteristics seem
to be driving the demand for monitoring the potential agency problem between owners and
managers and the demand for monitoring the potential conflict between majority and minority
owners.
Unlike earlier research, we find that the fraction of female directors is not related to board
independence. This is a surprising result compared to existing research on listed firms that
shows a strong positive correlation between female directors and board independence. We
argue that stockholders are just as likely to appoint female directors to fulfill the advisory role
as the monitoring role when board independence is not regulated and there are no regulations
on gender-quotas.
The results do not support our predictions that the monitoring of potential conflicts between
owners and managers is more prominent when firms have no controlling owner. Similarly, we
do not find evidence that monitoring of the conflict between majority and minority owners is
more prominent when the firm has a controlling owner. These results indicate that concerns
for the potential conflict between owners and managers and majority and minority owners are
just as likely to occur in all firms with multiple owners.
112
Moreover, we find that family firms are less likely to appoint independent directors than non-
family firms. This result suggests that family firms differ from other firms in ways that
influence their governance structure. The evidence is consistent with the notion that there are
strong ties between management and owners in family firms and also between large and small
stockholders. Hence, agency conflicts are not as common in family firms as in other firms.
Overall, our results suggest that board independence is firm specific. When owners face the
trade-off between monitoring and advice in a free contracting environment, they find an
optimal independence level that is lower than the one that is mandated for listed firms. Simply
put, firms value advice more than monitoring by the board of directors.
113
References
Adams, Renée B., and Daniel Ferreira. 2007. "A Theory of Friendly Boards." Journal of Finance no. 62 (1):217–250.
Adams, Renée B., and Daniel Ferreira. 2009a. "Strong Managers, Weak Boards?" CESifo
Economic Studies no. 55 (3–4):482–514. doi: 10.1093/cesifo/ifp023. Adams, Renée B., and Daniel Ferreira. 2009b. "Women in the Boardroom and their Impact
on Governance and Performance." Journal of Financial Economics no. 94 (2):291–309.
Adams, Renée B., Benjamin E. Hermalin, and Michael S. Weisbach. 2010. "The Role of
Boards of Directors in Corporate Governance: A Conceptual Framework and Survey." Journal of Economic Literature no. 48 (1):58–107.
Anderson, Ronald C., and David M. Reeb. 2003. "Founding-Family Ownership and Firm
Performance: Evidence from the S&P 500." Journal of Finance no. 58 (3):1301–1328. Anderson, Ronald C., and David M. Reeb. 2003b. "Founding Family Ownership, Corporate
Diversification and Firm Leverage." Journal of Law and Economics no. 56:653–684. Anderson, Ronald C., and David M. Reeb. 2004. "Board Composition: Balancing Family
Influence in S&P 500 Firms." Administrative Science Quaterly no. 49 (June):209–237. Andre, Paul, and Walid Ben-Amar. 2006. "Separation of Ownership from Control and
Acquiering Firm Performance: The Case of Family Ownership in Canada." Journal of Business Finance and Accounting no. 33 (3–4):517–543.
Baysinger, Barry D., and Henry N. Butler. 1985. "Corporate Governance and the Board of
Directors: Performance Effects of Changes in Board Composition." Journal of Law, Economics, and Organization no. 1 (1):101–124.
Beecher-Monas, Erica. 2007. "Marrying Diversity and Independence in the Boardroom: Just
How Far Have You Come, Baby?" SSRN eLibrary. Bhagat, Sanjai, and Bernard Black. 2002. "The Non-Correlation Between Board Indpendence
and Long-Term Firm Performance." Journal of Corporation Law no. 27 (2):231. Bhagat, Sanjai, and Bernard S. Black. 1998. "Independent Directors." The New Palgrave
Dictionary of Economics and the Law. no. 2:283–287. Boone, Audra L., Laura C. Field, Jonathan M. Karpoff, and Charu G. Raheja. 2007. "The
Determinants of Corporate Board Size and Composition: An Empirical Analysis." Journal of Financial Economics no. 85 (1):66–101.
Brennedsen, Morten, Kasper M Nielsen, Perez-Gonzalez, and Daniel Wolfenzon. 2007.
"Inside the Family Firm: The Role of Families in Succession Decisions and Performance." The Quaterly Journal of Economics no. 122 (2):647–691.
114
Bøhren, Øyvind, and Nils Erik Krosvik. 2013. "The Economics of Minority Freezeouts: Evidence from the Courtroom." International Review of Law and Economics, Forthcoming.
Bøhren, Øyvind, and Siv Staubo. 2013. "Female Directors and Board Independence: Evidence
from Board with Mandatory Gender Balance." Working Paper. Bøhren, Øyvind, and R. Øystein Strøm. 2010. "Governance and Politics: Regulating
Independence and Diversity in the Board Room." Journal of Business Finance & Accounting no. 37 (9–10):1281–1308.
Carter, Colin B., and Jay W. Lorsch. 2004. Back to the Drawing Board: Designing Corporate
Boards for a Complex World, Bosten, MA: Harvard Business School Press. Claessens, Stijn, Simon D. Djankov, Joseph P. H. Fan, and Larry H. P. Lang. 2002.
"Distangling the Incentives and Entrenchment Effects of Large Shareholdings." Journal of Finance no. 57:2741–71.
Coles, Jeffrey L., Daniel D. Naveen, and Lalitha Naveen. 2008. "Boards: Does One Size Fits
All?" Journal of Financial Economics no. 87:329–356. Denis, David J., and Atulya Sarin. 1999. "Ownership and Board Structures in Publicly
Traded Corporations." Journal of Financial Economics no. 52:187–223. Duchin, Ran, John G. Matsusaka, and Oguzhan Ozbas. 2010. "When are Outside Directors
Effective?" Journal of Financial Economics no. 96 (96):215–237. ECGI. 2012. "Index of Codes." European Corporate Governance Institute. Fields, M. Andrew, and Phyllis Y. Keys. 2003. "The Emergence of Corporate Governance
from Wall St. to Main St.: Outside Directors, Board Diversity, Earnings Management, and Managerial Incentives to bear Risk." Financial Review no. 38 (1):1–24.
Gomez-Mejia, Luis R., Manuel Nufiez-Nickel, and Isabel Gutierrez. 2001. "The Role of
Family Ties in Agency Contracts." Academy of Management Journal no. 44:81–96. Harris, Milton, and Artur Raviv. 2008. "A Theory of Board Control and Size." Review of
Financial Studies no. 21 (4):1797–1831. Hermalin, Benjamin E., and Michael S. Weisbach. 1991. "The Effects of Board Composition
and Direct Incentives on Firm Performance." The Journal of the Financial Management Association no. 20 (4):101–112.
Hermalin, Benjamin E., and Michael S. Weisbach. 1998. "Endogenously Chosen Boards of
Directors and their Monitoring of the CEO." American Economic Review no. 88 (1):96–118.
115
Hermalin, Benjamin E., and Michael S. Weisbach. 2003. "Boards of Directors as an Endogenously Determined Institution: A Survey of the Economic Literature." Economic Policy Review no. 9 (1):7.
Hsiao, Cheng. 2003. Analysis of Panel Data, Cambridge University Press. Jensen, Michael C. 1986. "Agency Costs of Free Cash Flow, Corporate Finance, and
Takeovers." The American Economic Review no. 76 (2):323–329. Jensen, Michael C. 1993. "The Modern Industrial Revolution, Exit, and Failure of Internal
Control Systems." Journal of Finance no. 48:831–880. Jensen, Michael C., and William H. Meckling. 1976. "Theory of the Firm: Managerial
Behavior, Agency Costs and Ownership Structure." Journal of Financial Economics no. 3 (4):305–360.
Linck, James S., Jeffry M. Netter, and Tina Yang. 2008. "The Determinants of Board
Structure." Journal of Financial Economics no. 7 (2):308–328. MacAvoy, Paul. W., and Ira. M. Millstein. 1999. "The Active Board of Directors and its
Effect on Performance of the Large Publicly Traded Corporations." Journal of Applied Corporate Finance no. 11 (4):8–20.
Masulis, Ronald W. M., Peter K. Pham, and Jason Zein. 2011. "Family Business groups
around the World: Financing advantages, control motivations, and organizational choises." Review of Financial Studies no. 24:3556–3600.
Morck, Randall, Andrei Shleifer, and Robert W. Vishny. 1989. "Alternative Mechanisms for
Corporate Control." American Economic Review no. 79 (4):842–852. Raheja, Charu G. 2005. "Determinants of Board Size and Composition: A Theory of
Corporate Boards." Journal of Financial and Quantitative Analysis no. 40 (2):283–306.
Sarbane-Oxely. 2002. "Sarbane-Oxely Act; 'Public Company Accounting Reform and
Investor Protection Act' and 'Corporate and Auditing Accountability and Responsibility Act'." United State Federal Law no. 116 (745):107–204.
Shivdasani, Anil, and David Yermack. 1999. "CEO Involvement in the Selection of New
Board Members: An Empirical Analysis." Journal of Finance no. 54 (5):1829–1853. Shleifer, Andrei, and Robert W. Vishny. 1997. "A Survey of Corporate Governance." Journal
of Finance no. 52 (2):737–783. Song, Fenghua, and Anjan V. Thakor. 2006. "Information Control, Career Concerns, and
Corporate Governance." Journal of Finance no. 61 (4):1845–1896. Studenmund, A. H. 2000. Using Econometrics: A Practical Guide. Vol. 4th ed.: Addison
Wesley Longman, Inc.
116
Villalonga, Belen, and Raphael Amit. 2006. "How do Family Ownership, Control and Management Affect Firm Value?." Journal of Financial Economics no. 80:385–417.
Weisbach, Michael S. 1988. "Outside Directors and CEO turnover." Journal of Financial
Economics no. 20:431–460.
117
Variable DefinitionBoard characteristicsBoard independence - A1 0/1 dummy variable that is 1 if the CEO is not a director
and 0 otherwiseBoard independence - A2 The fraction of directors that do not belong to the largest family
by ownershipBoard size The number of board membersFemale directors The proportion of board members who are womenBoD tenure The average number of years since the non-CEO directors were appointed
Ownership characteristicsOutside concentration The sum of squared equity fractions in the firm (Herfinadahl index)Inside ownership Fraction of equity held by the firm's officers and directorsCEO ownership Fraction of equity held by the CEOLargest owner Fraction of equity held by the firm's largest stockholder, counting
each family member as one ownerFamily ownership Fraction of equity held by the family with the largest equity stake
Family characteristicsFamily control 0/1 dummy variable that equals 1 if the largest family owns more
than 50% and 0 otherwiseFamily chair 0/1 dummy variable that equals 1 if the chair belongs to
the largest family by ownership and 0 otherwiseFamily CEO 0/1 dummy variable that equals 1 if the CEO belongs to the
largest family by ownership and 0 otherwiseFamily board The fraction of directors coming from the largest family by ownership
General firm characteristicsPerformance The average real return on assets from t-3 to tLeverage Total debt divided by total assetsGrowth The average percentage increase in real sales from t-3 to tInformation costs The standard deviation of performance from t-3 to tFirm age The number of years since the firm was foundedFirm size Sales in constant 2011 MNOK. Log transformed in regressionsCEO tenure The number of years since the CEO took officeCEO age The number of years since the CEO was born
Table 1: The empirical variables
This table defines the variables used in the empirical analysis. The ownership characteristics are based on ultimate (direct plus indirect) equity holdings.
118
Table 2: Distributional properties of the variables
Variable Mean Std. 0 5 25 50 75 95 100 NBoard characteristicsBoard independence - A1 0.48 0.50 0.00 0.00 0.00 0.00 1.00 1.00 1.00 193,297Board independence - A2 0.57 0.29 0.00 0.00 0.33 0.67 0.75 1.00 1.00 181,651Board size 3.80 1.08 3.00 3.00 3.00 3.00 4.00 6.00 14.00 193,297Female directors 0.15 0.21 0.00 0.00 0.00 0.00 0.33 0.67 1.00 193,297BoD tenure 3.47 2.64 0.00 0.00 1.33 3.00 5.00 8.75 11.00 193,297Ownership characteristicsOutside concentration 0.21 0.10 0.00 0.10 0.16 0.20 0.24 0.37 1.00 193,053Inside ownership 78.30 26.37 0.00 25.00 60.00 93.00 100.00 100.00 100.00 108,240CEO ownership 37.30 17.97 0.00 10.00 25.00 33.33 50.00 70.00 89.90 111,263Largest owner 45.79 17.18 0.00 20.00 33.33 49.00 55.00 78.00 89.99 193,297Family ownership 82.98 25.17 0.00 25.00 70.60 99.95 100.00 100.00 100.00 181,651Family characteristicsFamily control 0.82 0.38 0.00 1.00 1.00 1.00 1.00 1.00 1.00 193,297Family chair 0.47 0.50 0.00 0.00 0.00 0.00 1.00 1.00 1.00 181,651Family CEO 0.53 0.50 0.00 0.00 0.00 1.00 1.00 1.00 1.00 181,651Family board 0.43 0.29 0.00 0.00 0.25 0.33 0.67 1.00 1.00 181,651General firm characteristicsPerformance 8.48 12.62 -31.99 -6.99 1.22 3.42 14.62 35.41 51.99 193,297Leverage 0.74 0.24 0.00 0.32 0.60 0.76 0.89 1.09 2.10 193,297Growth 1.26 1.25 0.41 0.75 0.97 1.03 1.19 2.05 25.78 167,732Information costs 8.58 48.18 0.00 0.08 1.17 5.09 11.00 25.58 14260.00 140,595Firm age 12.28 12.48 0.00 1.00 4.00 9.00 16.00 33.00 162.00 187,487Firm size 28.06 328.98 1.62 2.18 4.04 8.20 19.09 77.61 89907.52 193,297CEO tenure 6.12 4.50 0.00 1.00 2.00 5.00 9.00 15.00 18.00 193,297CEO age 46.64 9.84 17.00 31.00 39.00 46.00 54.00 63.00 93.00 182,274
Percentile, %
This table shows distributional properties of the variables used to measure board, ownership, family, and general firm characteristics. Table 1 defines the variables. The sample is Norwegian non-listed firms, from 2000–2011, where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
119
Table 3: Firm characteristics by ownership
Panel A: Controlling ownerVariable All Majority owner Non-majority owner Difference t-value (p-value)Board independence - A1 0.48 0.50 0.46 0.04 17.58 (0.000)Board independence - A2 0.57 0.54 0.60 -0.06 -44.07 (0.000)CEO ownership 37.30 47.69 28.48 19.21 268.64 (0.000)Performance 8.48 8.42 8.54 -0.12 -2.03 (0.042)Leverage 0.74 0.75 0.73 0.02 17.52 (0.000)Female directors 0.15 0.17 0.14 0.03 30.54 (0.000)Growth 1.26 1.37 1.19 0.18 18.52 (0.000)Information costs 9.29 9.46 9.12 0.34 1.03 (0.299)Firm age 12.28 12.23 12.33 -0.10 -1.77 (0.077)Firm size 28.06 30.78 25.44 5.34 34.51 (0.000)Family control 0.82 0.81 0.83 -0.02 -11.37 (0.000)N 193,297 94,693 98,604
120
Panel B: Controlling family Variable All Family firms Non-family firms Difference t-value (p-value) Board independence - A1 0.48 0.39 0.88 -0.49 -231.83 (0.000) Board independence - A2 0.57 0.54 0.77 -0.23 -136.77 (0.000) CEO ownership 37.30 38.28 19.71 18.57 -93.52 (0.000) Performance 8.48 8.71 7.41 1.30 17.21 (0.000) Leverage 0.74 0.75 0.68 0.07 44.42 (0.000) Female directors 0.15 0.16 0.14 0.02 17.18 (0.000) Growth 1.26 1.17 1.38 -0.21 -11.40 (0.000) Information costs 9.29 9.18 9.74 -0.56 -1.00 (0.032) Firm age 12.28 11.96 13.75 -1.79 -20.82 (0.000) Firm size 28.06 19.15 68.79 -49.64 -110.39 (0.000) N 193,297 158,630 34,667 This table compares the mean values of the variables used in the base case model across subsamples. Panel A compares the mean values of Majority owner and Non-majority owner subsample. Panel B compares the mean values of Family firms and Non-family firms subsample. The differences between mean values, their t-values, and p-values (in parentheses) are reported in the three right-most columns. Table 1 defines the variables. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
121
Table 4: Bivariate correlation coefficients between the determinants of board independence CEO Information Female Firm Firm ownership Performance Growth Leverage costs directors age size Performance 0.016 Growth -0.010 -0.005 Leverage 0.050 -0.111 0.004 Information costs -0.007 0.013 0.001 0.028 Female directors 0.141 -0.018 -0.009 -0.013 -0.007 Firm age -0.004 -0.026 -0.016 -0.200 -0.026 0.070 Firm size -0.126 -0.005 0.006 -0.010 -0.011 -0.011 0.036 Family control 0.402 0.039 -0.013 0.112 -0.030 0.033 -0.055 -0.058 This table shows pairwise Pearson correlation coefficients between the hypothesized determinants of board independence as specified in model (1) of the main text. Table 1 defines the variables. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
122
Table 5: Estimates of the base-case model - Agency problem 1 (owner vs. manager) Base case model Determinant Prediction Base case model with family control CEO ownership (-) -0.072 -0.068 (0.000) (0.000) Performance (-) -0.008 -0.008 (0.000) (0.000) Leverage (+/-) -0.415 -0.380 (0.000) (0.000) Female directors (+) -0.272 -0.239 (0.000) (0.000) Growth (-) 0.000 0.000 (0.881) (0.919) Information costs (-) 0.000 0.000 (0.551) (0.667) Firm age (+) -0.014 -0.014 (0.000) (0.000) Firm size (+) 0.250 0.241 (0.000) (0.000) Family control (+/-) -0.655 (0.000) Constant -3.557 -2.950 (0.000) (0.000) LR chi2 (8) 3,389.66 3,478.44 Prob > chi2 0.000 0.000 R2 0.162 0.167 N 121,403 121,403 This table shows the estimated coefficients of the base-case model and the base-case model with Family control. The relationship is specified in model (1) of the main text. The predicted signs of the coefficients are shown in the second column. Column 3 and column 4 show the estimated coefficients for logit regressions of board, ownership, and general firm on board independence. The p-values of the estimated coefficients are stated in parentheses underneath. The dependent variable is 1 if the CEO is not a board member and 0 otherwise. Table 1 defines the determinants. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
123
Table 6: Estimates of the base-case model - Agency problem 2 (majority vs. minority)
Determinant Prediction Base-case
model Base-case model with Family control
CEO ownership (-) -0.001 -0.001 (0.000) (0.000) Performance (-) -0.002 -0.002 (0.000) (0.000) Leverage (+/-) -0.004 -0.040 (0.070) (0.086) Female directors (+) -0.212 -0.212 (0.000) (0.000) Growth (-) 0.000 0.000 (0.120) (0.120) Information costs (-) 0.000 0.000 (0.192) (0.197) Firm age (+) -0.003 -0.003 (0.000) (0.000) Firm size (+) 0.018 0.180 (0.000) (0.000) Family control (+/-) -0.031 (0.000) Constant 0.371 0.399 (0.000) (0.000) Prob > chi2 0.000 0.000 Random effects Yes Yes R2 0.147 0.153 N 76,856 76,856 This table shows the estimated coefficients of the base-case model and the base-case model with Family control. The relationship is specified in model (1) of the main text. The predicted signs of the coefficients are shown in the second column. Column 3 and column 4 show the estimated coefficients for panel data regressions of board, ownership, and general firm characteristics on board independence. The p-values of the estimated coefficients are stated in parentheses underneath. The dependent variable is the fraction of non-family directors. Table 1 defines the determinants. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
124
Table 7: Estimates of the base case model in subsamples Panel A: Controlling owner Agency problem 1 Agency problem 2 (owner vs. manager) (majority owner vs. minority owner) Determinant Majority owner Non-majority owner Majority owner Non-majority owner CEO ownership -0.050 *** -0.104 *** -0.001 *** -0.003 *** Performance -0.006 *** -0.009 *** -0.001 *** -0.001 *** Leverage -0.129 *** -0.453 *** 0.005 -0.007 ** Female directors -0.621 *** -0.173 ** -0.294 *** -0.138 *** Growth 0.000 0.000 0.000 0.000 * Information costs 0.001 0.000 0.000 0.000 Firm age -0.011 *** -0.017 *** -0.003 *** -0.003 *** Firm size 0.257 *** 0.243 *** 0.016 *** 0.016 *** Constant -4.514 *** -2.676 *** 0.377 *** 0.442 *** LR chi2 (8) 2,614.650 5,546.150 Prob > chi2 0.000 0.000 0.000 0.000 Random effects No No Yes Yes R2 0.157 0.173 0.171 0.124 N 34,769 34,765 34,765 34,765
125
Panel B: Controlling family Agency problem 1 Agency problem 2 (owner vs. manager) (majority vs. minority) Determinant Family firm Non-family firm Family firm Non-family firm CEO ownership -0.067 *** -0.089 *** -0.001 *** -0.004 *** Performance -0.007 *** -0.012 *** -0.001 *** -0.001 *** Leverage -0.297 *** -0.782 *** -0.003 -0.026 ** Female directors -0.290 *** 0.319 -0.217 *** -0.129 *** Growth -0.000 0.000 0.000 0.000 ** Information costs 0.000 -0.001 0.000 0.000 Firm age -0.014 *** -0.016 *** -0.003 *** -0.003 *** Firm size 0.276 *** 0.066 * 0.017 *** 0.014 *** Constant -4.293 *** 0.533 0.373 *** 0.636 *** LR chi2 (8) 6,265.150 710.430 Prob > chi2 0.000 0.000 0.000 0.000 Random effects No No Yes Yes R2 0.146 0.141 0.148 0.128 N 72,863 3,998 72,863 17,040 This table shows the estimated coefficients for regressions of board, ownership, and general firm characteristics on board independence in subsamples. Panel A shows the estimates for Majority owner and Non-majority owner subsamples. Columns 2 and 3 show the estimated coefficients for logit regressions related to Agency problem 1. The dependent variable is 1 if the CEO is not a board member and 0 otherwise. Columns 4 and 5 show the estimated coefficients for panel data regressions related to Agency problem 2. The dependent variable is the fraction of non-family directors. Panel B shows the estimates in Family firm and Non-family firm subsamples. Column 2 and column 3 show the estimated coefficients for logit regression related to Agency problem 1. The dependent variable is 1 if the CEO is not a board member and 0 otherwise. Column 4 and column 5 show the estimated coefficients for panel data regression related to Agency problem 2. The dependent variable is the fraction of non-family directors. The relationship is specified in model (1) of the main text. Statistical significance at 1%, 5%, and 10% levels is labeled ***, **, and *, respectively. Table 1 defines the determinants. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
126
Table 8: Alternative estimation methods - Agency problem 1 (owner vs. manager) Method Determinant Logit Probit Standard panel Logit panel Clustered OLS IV CEO ownership -0.072 -0.033 -0.003 -0.131 -0.004 -0.003 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Performance -0.008 -0.005 -0.001 -0.009 -0.001 -0.001 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Leverage -0.415 -0.206 -0.011 -0.470 -0.037 -0.011 (0.000) (0.000) (0.005) (0.006) (0.006) (0.003) Female directors -0.272 -0.130 -0.004 -0.261 -0.012 -0.004 (0.000) (0.000) (0.563) (0.213) (0.159) (0.493) Growth 0.000 0.000 0.000 0.000 0.000 0.000 (0.881) (0.800) (0.719) (0.960) (0.506) (0.493) Information costs 0.000 0.000 0.000 0.000 0.000 0.000 (0.551) (0.470) (0.625) (0.500) (0.296) (0.611) Firm age -0.014 -0.007 -0.001 -0.031 -0.001 -0.001 (0.000) (0.000) (0.001) (0.000) (0.000) (0.000) Firm size 0.250 0.145 0.025 0.721 0.028 0.025 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -3.557 -2.306 -0.157 -15.021 -0.139 -0.158 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Random effects No No Yes Yes No Yes Prob > chi2 / Prob > F 0.000 0.000 0.000 0.000 0.000 0.000 R2 0.162 0.158 0.090 0.092 0.091 N 121,403 76,861 76,861 76,861 76,861 76,861 This table shows the estimates of the base-case model under six different econometric techniques. The estimated coefficients are shown in columns 2–7, and the corresponding p-values are stated in parentheses underneath. The dependent variable is 1 if the CEO is not a board member and 0 otherwise. Table 1 defines the determinants. The sample is Norwegian non-listed firms, from 2000–2011, where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
127
Table 9: Alternative estimation methods - Agency problem 2 (majority vs. minority) Method Determinant Random effects Fixed effects Clustered OLS IV CEO ownership -0.001 -0.001 -0.003 -0.001 (0.000) (0.000) (0.000) (0.000) Performance -0.002 0.000 -0.001 -0.004 (0.000) (0.292) (0.059) (0.000) Leverage -0.004 0.000 0.005 -0.003 (0.070) (0.039) (0.537) (0.004) Female directors -0.212 -0.154 -0.336 0.002 (0.000) (0.753) (0.000) (0.656) Growth 0.000 0.000 0.000 -0.002 (0.120) (0.059) (0.613) (0.691) Information costs 0.000 0.000 0.001 -0.003 (0.192) (0.000) (0.063) (0.000) Firm age -0.003 -0.002 -0.003 -0.001 (0.000) (0.000) (0.000) (0.001) Firm size 0.018 0.012 0.015 0.013 (0.000) (0.000) (0.000) (0.000) Constant 0.371 0.399 0.511 0.464 (0.000) (0.000) (0.000) (0.000) Random effects/Fixed effects Random Fixed No Random Prob > chi2 / Prob > F 0.000 0.000 0.000 0.000 R2 0.147 0.138 0.154 0.077 N 76,856 72,863 76,856 76,856 This table shows the estimates of the base-case model under different econometric techniques. The estimated coefficients are shown in columns 2–5, where the corresponding p-values are stated in parentheses underneath. The dependent variable is the fraction of non-family directors. Table 1 defines the determinants. The sample is Norwegian non-listed firms, from 2000–2011, where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
128
Table 10: Alternative proxy for board independence - Agency problem 1 (owner vs. Manager)
Definition of board
independence Model A: Model B: Determinant Prediction CEO not director Relative tenure CEO ownership (-) -0.072 -0.037 (0.000) (0.000) Performance (-) -0.008 -0.014 (0.000) (0.000) Leverage (+/-) -0.415 -0.308 (0.000) (0.000) Female directors (+) -0.272 0.032 (0.000) (0.686) Growth (-) 0.000 0.000 (0.881) (0.928) Information costs (-) 0.000 -0.001 (0.551) (0.588) Firm age (+) -0.014 -0.078 (0.000) (0.000) Firm size (+) 0.250 -0.222 (0.000) (0.000) Constant -3.557 3.366 (0.000) (0.000) Random effects No Yes Year fixed effects No Yes Prob > chi2 0.000 0.000 R2 0.162 0.136 N 121,403 76,861 This table shows the estimates of the base-case model (1) using alternative proxies for board independence. Model A is the base-case model, which uses a dummy variable which equals 1 if the CEO is not a board member and 0 otherwise. Model B uses the average tenure of non-CEO directors minus the tenure of the CEO. The relationship is specified in model (1) of the main text. The predicted signs of the coefficients are shown in column 2 and the estimates of model A and model B are reported in columns 3 and 4, respectively, where the corresponding p-values are stated in parentheses underneath. Table 1 defines the determinants. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
129
Table 11: Estimates the base-case model with non-linear determinants - Agency problem 1 (owner vs. manager) Determinant Prediction Estimates CEO ownership (-) -0.144 (0.000) CEO ownership squared (+) 0.001 (0.000) Performance (-) -0.007 (0.000) Leverage (+/-) -0.370 (0.000) Female directors (+) -0.326 (0.000) Growth (-) 0.000 (0.928) Information cost (-) 0.000 (0.602) Firm age (+) -0.013 (0.000) Firm age square (-) -0.000 (0.248) Firm size (+) 0.222 (0.000) Constant -2.302 (0.000) LR chi2 (8) 9055.180 Prob > chi2 0.000 R2 0.183 N 76,861 This table shows the estimates of model (2) of the main text which uses additional determinants of board independence. The dependent variable is 1 if the CEO is not a board member and 0 otherwise. The predicted signs of the coefficients are shown in the second column, and the p-values of the estimated coefficients are reported in the third column are stated in parentheses underneath. Table 1 defines the determinants. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
130
Appendix 1: Characteristics of non-listed firms and listed firms
Variable Mean Median Mean MedianBoard characteristicsBoard independence - A1 0.48 0.00 0.92 1.00Board independence - A2 0.57 0.67 0.93 1.00Board size 3.80 3.00 6.20 6.00Female directors 0.15 0.00 0.21 0.20BoD tenure 3.47 3.00 1.62 1.37Ownership characteristicsOutside concentration 0.21 0.20 0.16 0.10Inside ownership 78.30 93.00 0.12 0.00CEO ownership 37.30 33.33 0.10 8.33Largest owner 45.79 49.00 26.61 22.64Family ownership 82.98 99.95 20.53 14.20Family characteristicsFamily control 0.82 1.00 0.08 0.00Family chair 0.43 0.00 0.12 0.00Family CEO 0.45 0.00 0.08 0.00Family board 0.45 0.00 0.07 0.00General firm characteristicsPerformance 8.48 3.42 5.59 5.02Leverage 0.74 0.76 0.49 0.52Growth 1.26 1.03 1.15 1.08Information costs 8.58 5.09 7.35 4.96Firm age 12.28 9.00 34.61 18.00Firm size 28.06 8.20 885.84 85.58CEO tenure 6.12 5.00 5.09 4.00CEO age 46.64 46.00 47.12 47.00N 178,721 178,721 3,427 3,427
Non-listed firms Listed firms
This table shows the mean and median values of the variables used to measure board, ownership, family, and general firm characteristics. Table 1 defines the variables. The sample is Norwegian non-listed and listed firms, from 2000–2011, where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
131
Appendix 2: Properties of the instrumental variable (IV) Panel A: CEO ownership classified by CEO age Variable All Young CEO Old CEO Difference p-value CEO ownership 37.30 36.16 38.49 -2.33 (0.000) Panel B: Board and firm characteristics classified by CEO age Variable All Young CEO Old CEO Difference p-value Board independence-A1 0.48 0.48 0.48 0.00 (1.000) Board independence-A2 0.57 0.58 0.56 0.02 (0.000) Performance 8.48 8.62 8.27 0.35 (0.000) Leverage 0.74 0.76 0.71 0.05 (0.000) Female directors 0.15 0.15 0.16 -0.01 (0.000) Firm size 16.18 16.13 16.28 -0.15 (0.000) This table shows board and firm characteristics for young and old CEOs. A young CEO is a CEO that is younger than 47 years (the average age of CEOs in our sample), an old CEO is a CEO that is 47 years or older. Panel A shows the ownership of young and old CEOs, the difference is shown in the fifth column, and the p-value is stated in the sixth column. Panel B shows board and firm characteristics for young and old CEOs, the differences are shown in the fifth column and the p-values are stated in parentheses in the sixth column. Table 1 defines the variables. The sample is Norwegian non-listed firms from 2000–2011 where revenue ≥ 2 million NOK, board size ≥ 3, and the largest owner holds < 90% of the equity. Performance is censored at 2% and 98% and leverage is censored at 99%.
132
5. Summary
The findings in this thesis show that regulation, such as the gender balance law and the board independence code, results in unintended effects which may matter for the firm’s behavior. In the first essay we find that stockholders of half the firms that suddenly become exposed to the gender balance law (GBL) choose the only alternative to changing the board, that is to exit into an organizational form where the GBL does not apply.
The second essay finds that the GBL causes a large increase in board independence. Involuntary increase in board independence is a potential problem because there is a trade-off between the board’s monitoring role and advice role. More independent directors are assumed to strengthen the board’s monitoring role, while more dependent directors are better advisors. Finally, we find that recommending a majority of independent directors in every firm by the independence code (IC) may hurt firms that are better off with a lower level of board independence.
This thesis addresses economic consequences of new regulation of board composition. Our results show that profitable, young, and small firms are hurt the most by these regulations. That is, the cost of changing the board, either by increasing the fraction of female directors or by increasing the fraction of independent directors, is particularly costly for such firms.
Recent political signals indicate that the exit option we analyze may soon disappear. In particular, gender balance in corporate boards may be made mandatory not just for ASA firms, but also for some AS firms . If that happens, Norway will not just be special for being the first and only country to mandate a massive, rapid shift in the composition of corporate boards and to punish non-compliers with liquidation. The regulators may also decide to eliminate the option firms currently have to mitigate the costs of regulatory shocks by transforming into organizational forms that are not exposed to the law. Every other country considering gender balance regulation seems to favor the comply-or-explain system or considerably milder sanctions than liquidation. Such regulatory regimes would leave the gender balance choice to the firm’s discretion and hence allow for firm heterogeneity in board design. Our findings suggest that, compared to this more flexible alternative, the mandatory approach, and particularly one without exit options, is a costly way to regulate gender balance of corporate boards.
Although the gender balance law is mandatory, the corporate governance code is not mandatory, but follows the principle of comply-or-explain. Nevertheless, there is a widespread view among policy makers that boards with at least half the directors being independent ensure good governance. Therefore, it is often argued that stockholders should ensure that their firm follows the code. Our evidence suggests that some firms are better off with a lower level of board independence. Hence, one-size-fits-all regulation is costly and should not be complied with by all firms.