Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6 671 AN ANALYSIS OF THE DETERMINANTS OF CORPORATE GOVERNANCE DISCLOSURE POLICIES IN MULTINATIONAL ENTERPRISES: A MULTI- MEDIUM STUDY Daniel Zeghal*, Manel Moussa** Abstract Purpose – This research aims to identify the factors underlying the corporate governance disclosure policies of the world’s largest multinational companies (MNCs) based on the following: (1) national factors related to the MNCs’ home countries (2) governance factors related to their governance systems and (3) operational factors arising from the operational characteristics of the MNCs. Methodology – Our sample includes 159 MNCs from 24 countries representing three geographic regions. The corporate governance disclosure policy is examined in terms of level and quality of disclosed information in two different mediums (traditional i.e .paper vs. websites). Results – Multiple linear regressions indicate that national factors, especially cultural ones, are important determinants of MNCs corporate governance disclosure policy in the traditional print mediums. National factors, however, seem to play no part in governance disclosures on the internet but can rather be explained by the international MNCs listing status. Practical implications – This study could guide the harmonization efforts of international standard setters in identifying factors leading to different governance disclosure behaviors and the disclosure medium most influenced by these factors. Keywords: MNCs, Disclosure, Policy, Corporate Governance, Cultural Characteristics, MNCs’ Characteristics, Traditional Mediums, Websites * Telfer School of Management, University of Ottawa ** Higher Institute of Business Administration, University of Gafsa, Tunisia Acknowledgement We acknoledge the support of the CPA-Canada Accounting and Governance Research Center at the University of Ottawa 1 Introduction A small number of large MNCs dominate the global economy. Consequently, the global economy relies upon their stable functioning. When governance mechanisms break down, the impact is felt not only in the home country but around the world (UNCTAD, 2011). A good example of this is the bankruptcies, frauds, social and environmental implications of MNCs businesses and the subprime crisis that was at the origin of the world economic crisis. The harmful practices of a small number of MNCs called into question the reputation of the majority of them and showed the importance of good corporate governance systems for the stability of economies and wellbeing of society. Good governance practices strengthen the trust between a company and its partners and contributes to creating more value for the company and its shareholders as well as contributing to the economic development of these countries. Governance disclosure represents the most effective tool that regulators can use to encourage companies to comply to best corporate governance practices (Winter Report, 2002) 1 . Furthermore, 1 Report of the High Level Group of Company Law Experts on a Modern Regulatory Framework for Company Law in markets are more effective when investors have sufficient information to properly assess corporate governance and make better investment decisions (OECD, 2004) 2 . Recently, governance disclosure has become an important aspect of corporate transparency, particularly, for those involved in global business and aiming to improve their reputation in international markets (Radebaugh et al. 2006; Markarian et al. 2007; Kolk, 2008). Previous studies on governance disclosure focused on domestic companies whose disclosure practices are largely influenced by the legal and financial contexts of their home countries (Bushman et al. 2004; Vander Bauwhede and Willekens, 2008; Ben Othman and Zeghal, 2008; Khan, 2009). A few studies have examined whether governance disclosure by multinationals, facing different legal, social and regulatory parameters, could be affected by factors other than those identified for national companies. Furthermore, it should be noted that these studies Europe, 4 November 2002. Available at: http://ec.europa.eu/internal_market/company/ docs/modern/report_en.pdf 2 Corporate government principles of the OECD, 2004
24
Embed
AN ANALYSIS OF THE DETERMINANTS OF · PDF fileAN ANALYSIS OF THE DETERMINANTS OF CORPORATE GOVERNANCE DISCLOSURE POLICIES IN MULTINATIONAL ... seem to play no part in governance ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
671
AN ANALYSIS OF THE DETERMINANTS OF CORPORATE GOVERNANCE DISCLOSURE POLICIES IN MULTINATIONAL
ENTERPRISES: A MULTI- MEDIUM STUDY
Daniel Zeghal*, Manel Moussa**
Abstract Purpose – This research aims to identify the factors underlying the corporate governance disclosure policies of the world’s largest multinational companies (MNCs) based on the following: (1) national factors related to the MNCs’ home countries (2) governance factors related to their governance systems and (3) operational factors arising from the operational characteristics of the MNCs. Methodology – Our sample includes 159 MNCs from 24 countries representing three geographic regions. The corporate governance disclosure policy is examined in terms of level and quality of disclosed information in two different mediums (traditional i.e .paper vs. websites). Results – Multiple linear regressions indicate that national factors, especially cultural ones, are important determinants of MNCs corporate governance disclosure policy in the traditional print mediums. National factors, however, seem to play no part in governance disclosures on the internet but can rather be explained by the international MNCs listing status. Practical implications – This study could guide the harmonization efforts of international standard setters in identifying factors leading to different governance disclosure behaviors and the disclosure medium most influenced by these factors.
Keywords: MNCs, Disclosure, Policy, Corporate Governance, Cultural Characteristics, MNCs’ Characteristics, Traditional Mediums, Websites * Telfer School of Management, University of Ottawa ** Higher Institute of Business Administration, University of Gafsa, Tunisia
Acknowledgement We acknoledge the support of the CPA-Canada Accounting and Governance Research Center at the
University of Ottawa
1 Introduction A small number of large MNCs dominate the global economy. Consequently, the global economy relies upon their stable functioning. When governance mechanisms break down, the impact is felt not only in the home country but around the world (UNCTAD, 2011). A good example of this is the bankruptcies, frauds, social and environmental implications of MNCs businesses and the subprime crisis that was at the origin of the world economic crisis.
The harmful practices of a small number of
MNCs called into question the reputation of the
majority of them and showed the importance of good
corporate governance systems for the stability of
economies and wellbeing of society. Good governance
practices strengthen the trust between a company and
its partners and contributes to creating more value for
the company and its shareholders as well as
contributing to the economic development of these
countries. Governance disclosure represents the most
effective tool that regulators can use to encourage
companies to comply to best corporate governance
practices (Winter Report, 2002)1. Furthermore,
1 Report of the High Level Group of Company Law Experts on
a Modern Regulatory Framework for Company Law in
markets are more effective when investors have
sufficient information to properly assess corporate
governance and make better investment decisions
(OECD, 2004)2. Recently, governance disclosure has
become an important aspect of corporate transparency,
particularly, for those involved in global business and
aiming to improve their reputation in international
markets (Radebaugh et al. 2006; Markarian et al.
2007; Kolk, 2008).
Previous studies on governance disclosure
focused on domestic companies whose disclosure
practices are largely influenced by the legal and
financial contexts of their home countries (Bushman et
al. 2004; Vander Bauwhede and Willekens, 2008; Ben
Othman and Zeghal, 2008; Khan, 2009). A few studies
have examined whether governance disclosure by
multinationals, facing different legal, social and
regulatory parameters, could be affected by factors
other than those identified for national companies.
Furthermore, it should be noted that these studies
Europe, 4 November 2002. Available at: http://ec.europa.eu/internal_market/company/ docs/modern/report_en.pdf 2 Corporate government principles of the OECD, 2004
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
672
focused on a single disclosure medium, namely annual
reports.
In this study, we try to overcome these
limitations by focusing on a sample of MNCs and
analyzing two main disclosure mediums namely the
annual printed reports (.i.e. annual reports and proxy
statements) and websites.
The main objective of our research is to identify
the determinants of MNCs governance disclosure
policy. Governance disclosure policy will be
considered in terms of level and quality of disclosed
information via the two types of disclosure mediums
chosen. Determinants of governance disclosure policy
will be divided into three groups of factors: (1)
national factors related to the MNCs’ home countries
(2) governance factors related to MNCs’ governance
systems and (3) operational factors arising from the
operational characteristics of MNCs. The purpose of
our analysis is to determine which factors are likely to
best explain MNCs governance disclosure policy in
each of the studied mediums. Based on a sample of
159 MNCs from 24 countries, the results of this
research show that while national factors remain
preeminent determinants of MNCs in annual printed
reports, their influence is lessened on the websites
where governance disclosure policy is better explained
by international MNCs listing status.
The rest of the paper is organized as follows: the
second section presents the theoretical framework of
governance disclosure. In the third section, our
hypotheses are developed. Section 4 presents the
sample and the methodology used to collect data and
are presented in section 5 and the conclusion is the
subject of the last section.
2 Importance of Governance Disclosure for Companies and other Decision Makers
Recent developments in governance codes and
regulations around the world requiring companies to
disclose the main aspects of their governance
structures and practices, underline the importance of
this information to the various company stakeholders,
especially following the series of accounting scandals
of some companies in recent years (Webb et al.
2008,b).
Clear and comprehensive information on
governance is useful for the investing decisions of
both actual and potential investors (Parum, 2005)
allowing them to assess the credibility and
effectiveness of the governance system of a given
society (OECD, 2004) which, in turn, allows them to
deduce the quality of published accounting figures
(Klein, 2002) and make more accurate predictions of
future performance (Bhat et al. 2006).
The multinational status of the company makes
this disclosure more useful for investors in view of
the multitude of governance standards and principles
to which MNCs are subject and the difficulty of
assessing these companies’ risks and actual and future
performances (Luo, 2005 b; Radebaugh et al. 2006).
Governance disclosure is also important for
companies since raising capital at lower cost in
international financial markets depends increasingly
on efficiency and reliability guarantees that the
governance system can provide3. This largely explains
why companies increasingly provide to their
shareholders and other interested parties a description
of their situation in terms of governance. However,
despite the global consensus on the content of these
disclosures4, reporting on governance varies widely
around the world (Bushman et al. 2004; Vander
Bauwhede and Willekens, 2008; Ben Othman and
Zeghal, 2008).
2.1 Theoretical framework of the determinants of information disclosure on governance
According to the agency theory, greater corporate
governance disclosure would reduce agency costs
resulting either from separating ownership and control
or from debt (Vander Bauwhede and Willekens,
2008). This incentive is particularly relevant for
MNCs because the complexity of their global
operation aggravates the informational gap and the
difficulties of observing efforts made by managers to
optimally use funds of geographically scattered
shareholders or creditors (Luo, 2005 b). From a
political perspective, managers can disseminate
information about the level of their compliance with
good governance principles. They thus avoid the risk
of implementing more costly regulations for the
company in the governance and governance disclosure
fields5. The desire to mitigate the adverse effects of
political visibility on the international scale is greater
for large MNCs that attract more attention from the
media, politicians and the general public (Watts and
Zimmerman, 1978). From a signaling perspective,
well-governed MNCs would be interested in
increasing their governance disclosure level in order to
signal their governance quality to investors and have a
more advantageous cost of capital. Beyond the desire
to reduce agency and political and capital costs,
governance disclosure could serve as a tool to
legitimize a company’s activities. This is especially
true for MNCs which need to appear legitimate in
terms of the integrity of their governance system and
3 ASX Corporate Governance Council (2003), Principles of
Good Corporate Governance and Best Practice Recommendations, March 2003 4 UNCTAD, 2009 Review of the implementation status of
corporate governance disclosures: an inventory of disclosure requirements in 24 emerging markets, Twenty-sixth session, TD/B/C.II/ISAR/CRP.8 5 The Sarbanes Oxley Act 2002 and the European directive
2006/46/CE enacted as a reaction to a number of major corporate and accounting scandals of several large American and European companies set new and more severe reforms in the field of corporate governance and the information governance dissemination.
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
672
to avoid incurring costs resulting from non-legitimacy,
especially after the accounting scandals involving
several of these companies (legitimacy theory).
At an international level, accounting models
suggest that a number of factors are at play to explain
variation of governance disclosure practices. These
factors include the political and legal environment
(Belkaoui, 1983; La Porta et al. 1998; Jaggi and Low,
2000), the national culture (Gray, 1988; Zarzeski,
1996), the economic development level (Belkaoui,
1983; Adhikari and Tondkar, 1992), and capital
market factors such as size and activity level
(Adhikari and Tondkar, 1992; Doupnik and Salter,
1995).
2.2 Previous Studies on determinants of governance information disclosure
Many studies have been conducted on the
determinants of governance disclosure at a national
level (Labelle, 2002; Bujaki and McConomy, 2002;
Parum, 2005; Bhuiyan and Biswas, 2007; Parsa et al.
2007; Webb et al. 2008 b) and also internationally
(Bushman et al. 2004; Vander Bauwhede and
Willekens, 2008; Ben Othman and Zeghal, 2008;
Khan, 2009). Generally, results of these studies
showed that theories discussed above have a strong
explanatory power with respect to disclosure behavior
for the governance of the studied companies. Within a
national framework, variations in governance
disclosure levels are explained by factors representing
4.5 Measurement and descriptive analysis of the explanatory variables 4.5.1 Measurement of the explanatory variables
Measurements of explanatory and control variables
are consistent with national and international
empirical research and are summarized in table 5.
Table 5. Measurements of the explanatory and control variables
Variable Expected
Sign
Measurement
National factors
ORIGIN Positive Dummy variable that takes 1 if the country has a legal system that originates in ‘common
law’ practices and 0 otherwise.
INVEST
Positive Investor protection is measured by the ‘anti-director rights’ index developed by La Porta et al
(1998) and revised par Djankov et al. (2006). This index is obtained through adding 1 if : 1-
The country allows shareholders to send their vote by proxy ; 2- Shareholders are not
required to deposit their shares before voting in a general assembly ; 3- The Cumulative
voting or the proportional representation of minors in the board is authorized ; 4- There is a
legal device allowing an aggrieved minority to initiate legal actions against managers or
board members.; 5- The percentage of the required capital to obtain an AGO organization is
less than 10% and 6- There is an obligation to consult shareholders before deleting their
preferential subscription rights.
LAW Positive Application quality of laws measured by means of the following variables: regulatory
quality, ‘rule of law’ and corruption control (Kauffmann et al. 2010).
INDIV
Positive Individualism index indicating a preference for a loosely knit social framework in society
wherein individuals are supposed to take care of themselves and their immediate families
only. (Hofstede, 1980).
DIS
Negative Hierarchical distance index measuring the degree of acceptance of the unequal power
distribution within institutions and organizations by members of the society (Hofstede,
1980).
IA Negative Uncertainty avoidance index measuring the propensity of a society feeling threatened by
uncertain or unknown situations (Hofstede, 1980)..
MAS Positive Masculinity index measuring how far the dominant values of a society are preferred with
respect to achievement, assertiveness and material success (Hofstede, 1980)..
ECODEV Positive Economic development level measured by the gross domestic product per person in $US of
2010. World development indicators 2010
KMDEV
Positive Capital market development level measured by the average of the following variables 1-
Average stock market capitalization in $US adjusted to the country’s GDP in $US over the
years 2007-2010 ; 2- Total value of traded shares (in GDP percentage) over the years 2007-
2010. World development indicators 2010
POLI Positive Level of civil liberties in a country measured by the ‘Voice and Accountability’ indicator.
(Kauffmann et al. 2010).
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
681
Governance factors
BSIZE Positive Board size measured by the number of the directors.
BINDEP Positive Board independence measured by the ratio: Number of independent members/ Board size.
BDUAL Negative Dichotomous variable that takes the value of 1 if the positions of Chief Executive Officer
and Chairman are combined and 0 otherwise.
BDIVER Positive Degree of board diversity measured by the percentage of women on the board of directors.
ACINDEP Positive Audit committee independence measured by the ratio of independent members over the total
number of members of the audit committee.
GCOM Positive Binary variable that takes 1 if the board has a governance committee and 0 otherwise.
Operational factors
BLOC
Negative Ownership concentration measured by the percentage of shares held by blockholders having
5% or more of the capital.
INTERLIST Positive Binary variable that takes the value of 1 if the company is listed in a foreign stock market
and zero otherwise.
MULTIN
Positive Degree of multinationalisation measured by the average of the following three ratios :
foreign sales/ total sales in $US;workforce abroad/ total work force and assets abroad/total
assets in $US. This measure is based on the transnationality index (TNI) calculated by the
UNCTAD to evaluate the multinationalisation degree of non-financial MNCs.
SIZE Positive Company size measured by the total assets logarithm in $US.
DEBT Positive Company debt measured by the ratio Total debt/ total assets.
PERF Positive Company performance measured by the return on equity.
4.5.2 Descriptive analysis of explanatory variables
In table 6, the average index of investor protection
(INVEST) in countries of our sample is 3.264. China
has the lowest protection level with 1 as the index
compared to the United Kingdom, Spain, Brazil and
India that have the highest index of 5. The index
reflecting the quality of applying laws (LAW) varies
between a minimum of -0.75 (in Russia) and a
maximum of 2.05 (in Denmark). A country’s degree
of individualism in our sample (INDIV) varies
between a minimum index of 17 in Taiwan and a
maximum index of 91 in the United States. The
average index of acceptance for hierarchical
inequities (DIS) is 48.07. The degree of uncertainty
avoidance (IA) varies between a minimal value of 23
in Denmark and a maximum value of 94 in Belgium.
Finally, the degree of masculinity dominance in the
sampled countries (MAS) varies between a minimum
of 5 in Sweden and a maximum of 95 in Japan
suggesting that Sweden is the country with the highest
level of feminist values whereas masculinity values
are highest in Japan. On average, the degree of
economic development (ECODEV) of countries is $39
087.35 . The average degree of capital markets
development (KMDEV) is 138.276. The average score
on the level of civil liberties (POLI) in countries from
our sample is 0.929. There is a maximum value of
1.62 for Switzerland and a minimum value of -1.65 in
China. Table 7 shows that 40.3% of MNCs from our
sample are from common law countries (ORIGIN)
while 59.7% of MNCs are based in civil law
countries.
On average, the MNCs’ boards in our sample
are composed of 13 members (BSIZE). The smallest
boards have only 6 members while larger ones include
up to 27 members. On average, 60% of board
members of our sampled MNCs are not part of the
management team and are independent (BINDEP).
Concerning the variable BDUAL, in 34% of our
sampled MNCs the position of both CEO and
Chairman of the Board is held by one person . On
average, 12.5% of board members of the sampled
MNCs are women (BDIVER). Approximately 48% of
our MNCs have a governance committee set up by the
board of directors (GCOM). Majority shareholders
(BLOC) hold an average of 27.588% of MNCs
capital. The average multinationalisation percentage
(MULTIN) of our MNCs is 42.8%. Furthermore, 48%
of the MNCs are inter-listed on one or more foreign
stock markets (INTERLIST). The total assets
logarithm that indicates a company’s size (SIZE)
varied between a minimum of 9.241 and a maximum
of 13.529. On average, MNCs in our sample have an
ROE of 16.215% (PERF). The minimum level of
return on equity is -62% while the maximum is
established at 119%. The average debt ratio of the
sampled MNCs (DEBT) is 59.21% that indicates that
MNCs in our sample are highly leveraged.
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
682
Table 6. Descriptive statistics of explanatory variables
N Minimum Maximum Mean Standard deviation
Panel A : National factors
INVEST 159 1 5 3.264 0.889
LAW 159 -0.75 2.05 1.192 0.650
INDIV 159 17 91 67.14 23.5
DIS 159 18 81 48.07 12.637
IA 159 23 94 62.28 21.320
MAS 159 5 95 61.87 19.734
ECODEV 158 1 410 85 389 39 087.35 14 654.523
KMDEV 158 22.81 243.73 138.276 67.072
POLI 159 -1.65 1.62 0.929 0.7655
Panel B : Governance factors
BSIZE 159 6 27 12.685 3.489
BINDEP 153 0 1 60% 0.273
BDIVER 156 0 45% 12.48% 0.110
ACINDEP 153 0.17 100% 84% 0.234
Panel C: Operational factors
BLOC 151 1.30 98.08 27.588 24.229
MULTIN 149 012 1 42.8 25.6
SIZE 159 9.241 13.529 11.209 0.794
PERF 159 -62.0 119 16.215 15.951
DEBT 158 0.05 1.00 59.2 18.2
Variable definitions : INVEST= the score of investors protection level; LAW = average of the three variables: regulatory
quality , rule of law and corruption control; INDIV= Individualism score of the country; DIS= Hierarchical distance score;
IA= Uncertainty avoidance score ; MAS=Masculinity score ; ECODEV=gross national product per person; KMDEV=
average of two variables: average market capitalization of the country adjusted to GDP over the last four years 2007-2010;
2- Total value of traded shares (in percent of GDP) over the past four years from 2007 to 2010; POLI=Civil liberties score
‘Voice and Accountability’; BSIZE= number of directors on Board ; BINDEP= percentage of independent directors on the
Board; BDIVER : Percentage of women on the board of directors; ACINDEP = percentage of independent members of the
Audit Committee; BLOC= percentage of shares held by blockholders with 5% or more of the capital; MULTIN= average
of three ratios: foreign sales / total sales –workforce abroad/ total employment and foreign assets / total assets ; SIZE=total
assets logarithm; PERF= return on equity ; DEBT=total debt/ total assets.
Table 7. Descriptive statistics of binary explanatory variables
N Variable=1 Variable =0
Frequency Percentage Frequency Percentage
Panel A : National factors
ORIGIN 159 64 40.3% 95 59.7%
Panel B : Governance factors
BDUAL 159 54 34% 105 66%
GCOM 159 76 47.8% 83 52.2%
Panel C: Operational factors
INTERLIST 159 76 47.8% 83 52.2%
Variable definitions : ORIGIN : 1 if the country has a common law legal system and 0 otherwise; BDUAL: 1 if both
Chief Executive Officer and Chairman of the Board positions are held by one person and 0 otherwise ; GCOM: 1 if among
the specialized committees of the board there is a governance committee and 0 otherwise; INTERLIST: 1 if the firm is
listed on one or more foreign stock markets and 0 otherwise.
12
Five American MNCs whose business scope covers only the fifty states of the United States. These MNCs are classified by Global Fortune magazine as part of the MNCs group.
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
independent directors of the Board ; BDUAL=1 if the
positions of Chief Executive Officer and Chairman are
combined and 0 otherwise; BDIVER=percentage of
women among board members; ACINDEP=
percentage of independent directors on the audit
committee; GCOM= 1 if there is a governance
committee and 0 otherwise; BLOC= percentage of
shares held by blockholders with 5% or more of the
capital; MULTIN= average of three ratios: foreign
sales / total sales – foreign effective / total
employment and foreign assets / total assets;
INTERLIST =1 if the company is listed on one or
more foreign stock markets and 0 otherwise;
SIZE=Total assets logarithm; DEBT=total debt/ total
assets;PERF= accounting income before interest /
equity; εit= residual term.
The correlation matrix between explanatory
variables presented in Table 8 shows that more than
0.5 correlations exist among explanatory variables,
especially national ones, which leads to a
multicollinearity problem. Therefore, we have tested
the general multicollinearity level by calculating the
VIF (variance inflation factor). We noticed that the
VIF of many of our variables exceeds the threshold of
10 which indicates a generalized and widespread
multicollinearity in our model (Neter et al. 1989 cited
in Williams, 2004).
In order to solve this problem, we have
conducted our multivariate tests through alternately
implementing partially correlated variables in four
different multivariate models presented in Table 9. We
have tested these models, on the one hand, on level
indices LGD and quality indices QGD of information
disclosure on governance calculated through the
annual print mediums and on the other hand, on those
obtained via websites. Since the results obtained for
both sets of tests are substantially similar we have
decided to only present results related to disclosure
level LGD and to report, any differences, if applicable.
(3)
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
684
Table 8. Correlation matrix between explanatory variables
Table 9. Explanatory variables distribution in the regression models
Model 1 Governance variables + Operational variables
Model 2 Legal variables + Governance variables except BINDEP and ACINDEP + Operational variables
Model 3 Cultural variables + Governance variables except BINDEP and ACINDEP + Operational variables
Model 4 Economic, financial and political variables 13 + Governance variables except BINDEP, ACINDEP and GCOM + Operational variables
13
Although they are highly correlated, ECODEV and POLI variables were included in the same model (model 4) because the results obtained when separating them in two different models were identical to those obtained when they were included in the same model.
ORIGI INVEST LAW INDIV DIS CI MAS DEVE KMDEV POLI BSIZE INDEP BDUAL DIVER ACINDEP GCOM BLOC INTER MULTI TAIIL PERF DEBT
Legend : **Significant correlations at 1% . * Significant correlations at 5%. Correlations above 0.5 are indicated in bold.
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
685
5 Checking the Assumptions of the Study 5.1 Determinants of governance disclosure policy of MNCs in the annual print mediums Table 10 shows that the four models are significant at p< 0.001 and respectively account for nearly 42%, 50% , 66% and 32% of the dependent variables’ variance. The p value of the Breusch-Pagan test presents a statistic of non-significant χ² indicating the absence of the heteroscedasticity problem of residues. The averages of VIF that do not strongly exceed 1, show that there is no multicolinearity problem. We can also notice that implementing cultural dimensions contributes most in improving the explanation of variation occuring in the LGD variable (Model 3: ∆R2 = 0.328***). The results of multivariate tests (Model 2) show that in accordance with our expectations, ORIGIN and LAW variables have a positive and significant impact on the LGD variable. The INVEST variable also has a significant impact on the LGD variable but in the opposite direction (β = -0.289 ***) which probably indicates the need of MNCs from countries with weak investor protection to disclose more information with respect to their governance in order to overcome concerns of international investors concerning weak laws in these countries (Webb et al, 2008 a). It would also enable them to report on the quality of their governance practices (Patel and Dallas, 2002). As expected, model 3 reveals that LGD increases in the most individualistic cultures INDIV (more competitive and less confidential) and decreases in cultures with high risk aversion IA (where companies fear that disclosing their governance could lead to competitive disadvantages). The coefficient of the cultural variable DIS also appears to be significant but not with the expected sign. This result is, however, consistent with Zarzeski’s results (1996) and also with Jaggi and Low (2000), Hope (2003) and Archambault and Archambault (2003) who note that MNCs more dependent on international resources, are likely to deviate from their discretionary original culture and provide better disclosure in order to show the quality of their public operations. The inverse relationship found between the cultural dimension MAS and LGD may be due to the ambiguity of the relationship between this dimension and disclosure advanced by several researchers (Jaggi and Low, 2000 and Archambault and Archambault, 2003) and raised by Gray (1988) himself. Multivariate analysis (Model 4) confirms the positive impact of the variable KMDEV on the LGD of MNCs (β = 0.342 ***) but not the impact of the degree of economic development
ECODEV and of the level of civil liberties and freedom of expression, POLI (Model 4).
As expected, the variables BSIZE, BINDEP and
BDIVER (models 1, 2, 3 and 4) and MULTIN (model
2, 3, and 4) have a significant and positive impact on
the variable LGD.
The lack of consistency in the results of the
variables BDUAL and BLOC in addition to the low
levels of their significance do not allow for drawing
any conclusions about their actual impact.
Furthermore, the variables ACINDEP, GCOM and
INTERLIST, SIZE, PERF, DEBT appear to be non-
significant in all models.
5.2 Determinants of the governance disclosure policy of MNCs via websites
According to table 11, the adjusted R2 of the four
models (15%, 18%, 21% and 25%) are lower than
those obtained through the annual print mediums. It
also turns out that implementing national factors does
not significantly increase the explanatory power of the
models. Only the implementation of economic,
financial and political variables (particularly the
KMDEV variable) allows us to significantly enhance
R2 (Model 4: Δ R2 = 0.075 ***). Results of the
regression (model 2, 3 and 4) show that the legal
variables (ORIGIN, INVEST and LAW), the variables
ECODEV and POLI, in addition to two of the four
cultural dimensions, INDIV and DIS, are not relevant
to an explanation of the disclosure policy of MNCs’
governance. Furthermore, the results highlight a
significant relationship (at 10% and 5 %) but in the
direction opposite to the one predicted between the
cultural values of uncertainty avoidance IA and
masculinity MAS and LGD. This can possibly be
explained by the multinational nature of our
companies that may drive them to depart from the
usual secretive direction of their countries of origin in
order to compete in international capital markets. In
this context, Zarzeski (1996, p.20) mentions that : «
When a firm does business in a global market, it is
operating in a different 'culture' and therefore may
need to have different practices. High levels of
financial disclosures may be necessary for
international survival because disclosure of quality
operations should result in lower resource costs. When
enterprises from more secretive countries perceive
economic gain from increasing their financial
disclosures, cultural borrowing may occur. The culture
being borrowed will be a ‘global market culture’
rather than a specific country culture».
Corporate Ownership & Control / Volume 12, Issue 4, Summer 2015, Continued – 6
686
Table 10. Multiple regressions under ordinary least squares of the relationship between the governance disclosure policy of MNCs in annual print mediums and its determinants