Asian Journal of Finance Accounting
ISSN 1946-052X 2012, Vol. 4, No. 1
6 www.macrothink.orgajfa
With regard to block ownership, David and Kochhar 1996 and Kim and Lee 2003 state that if an individual has a substantial amount of interest in a particular company usually
measured at 5, he or she will be more interested in the company and hence the block holders is an important player to have higher disclosure since they have the voting power that
could be used as a tool to monitor the agents. From the agency theory, it could be inferred that block holders have the interests in the firms; most likely they might put the pressure on
the management to disclose all the material information. Therefore, it could be expected that there is a positive relationship between block holders and disclosure. Chau and Gray 2002,
Luo, Courtenay and Hossain 20061, Huafang and Jianguo 2007 and Norita and Shamsul Nahar 2004 find that extent of outside block ownership is positively associated with
voluntary disclosures and hence their finding is in line with theoretical expectation. However, Eng and Mak 2003 find that block holder ownership is not related to disclosure. The finding
of Lakhal 2005 seems not to be in line with theoretical expectation.
Finally, regarding institutional investors, Salleh and Mallin 2002, Le et al. 2006, Langnan, Steven and Weibin 2007 and Ramzi 2008 collectively agree on the important role of
institutional shareholders in the monitoring of firms because of the voting power, the potential benefits from their activism, the existence of lesser ability to liquidate the shares
without affecting the share price, the fiduciary responsibility towards the ultimate owners, and ability to monitor executives since they are professionals. Institutional ownership seems
to be an important player to have higher disclosure since their voting power can be used as a tool to monitor the agents David and Kochhar, 1996. Therefore, it could be expected that
there is a positive relationship between institutional investors and disclosure. The findings of Eng and Mak 2003, and Lakhal 2005 are in line with theoretical expectation. Whereby,
Eng and Mak 2003 use government ownership as a proxy for institutional shareholder and Lakhal 2005 use the foreign institutional investors ownership as the proxy. However,
Huafang and Jianguo 2007 find that there is state ownership and legal ownership are not related to disclosure and hence they suggest that the Chinese regulators should gradually
encourage multi-ownership.
3. Development of Hypotheses and Research Design
3.1 Development of Hypotheses
Disclosing the material information of the firms reduces the information asymmetry between the management and the owners, and it will also reduce the agency conflicts between them.
According to Patel et al. 2002 and United Nation 2003, disclosure is an integral part of the corporate governance because it shows the extent of how good corporate governance is.
Leong 2005 also mentions that disclosure and transparency are partners of good corporate governance. Moreover, Beekes and Brown 2006 study 250 Australian firms rated in the
2002 Horwath Corporate Governance Report and find that better-governed firms do make more informative disclosure. Hence, the researcher is interested to examine whether
corporate governance variables could affect the social and environmental information disclosure and the following hypotheses are developed.
1
The existence of outside block ownership significantly decreases managers’ ability to limit voluntary disclosure.
Asian Journal of Finance Accounting
ISSN 1946-052X 2012, Vol. 4, No. 1
7 www.macrothink.orgajfa
3.1.1 Board Leadership Structure There should be a control mechanism that could put pressure on the management to disclose
the material information. One of the mechanisms, derived from the agency theory is board independence from the management Fama Jensen, 1983. Corporate governance literature
suggests that board independence could be achieved from separate leadership structure. This suggestion is also supported by empirical findings of prior researchers such as Gul and Leung
2004, Lakhal 2005, Byard et al. 2006 and Huafang and Jianguo 2007 since they find that there is a positive relationship between separate leadership structure and disclosure.
Therefore, it is expected that separate leadership structure results in higher social and environmental disclosure and the following hypothesis, in an alternative form, is developed.
Ha1: Social and Environmental disclosure is positively related to separate leadership structure.
3.1.2 Board Composition Several researchers based on the agency theory highlight the importance of board
independence. It has been well recognized by the corporate governance literature that higher proportion of independent non-executive directors makes the board independent from the
management Choe Lee, 2003; Yatim, Kent Clarkson, 2006. Furthermore, the findings of prior researchers are in line with the expectation of agency theory, for instance, Chen and
Jaggi 2000, Gul and Leung 2004, Byard et al. 2006, and Cheng and Courtenay 2006 and Norita and Shamsul Nahar 2004. They find that higher proportion of independent
directors has positive impact on disclosure. Therefore, the researcher expects that higher proportion of independent non-executive directors will contribute to higher disclosure and the
following hypothesis, in an alternative form, is developed.
Ha2: Social and environmental disclosure is positively related to proportion of independent non-executive directors on the board.
3.1.3 Board Size As explained earlier, agency theory foresees the problems that could arise due to the
separation of ownership and management. In order to reduce the problems between them, the board, as a middle person, should be independent from the management so that the board
could be able to monitor the management effectively which is in line with the interest of the shareholders. Several researchers such as Mak and Li 2001, Yoshikawa and Phan 2003,
Yatim et al. 2006 and Khanchel 2007 collectively agree that in order for the board to be free from the management, especially from the CEO, and to effectively control the CEO,
board size should be small due to difficulties in organizing and coordinating large group of directors. The finding of Byard et al. 2006 is in line with theoretical expectation since their
study of 1279 firms over the years 2000 to 2002 shows that analyst forecast accuracy based on the financial disclosure decreases with board size. Therefore, it could be expected that
smaller board size will result in better disclosure and the following hypothesis, in an alternative form, is developed.
Asian Journal of Finance Accounting
ISSN 1946-052X 2012, Vol. 4, No. 1
8 www.macrothink.orgajfa
Ha3: Social and environmental disclosure is negatively related to board size. 3.1.4 Ownership
Corporate governance literature highlights the important role of ownership on the disclosure. If the directors hold the shares, they might not want to disclose all the material information to
the outsiders since they would like to channel the benefits of the firms for themselves, not to the outside shareholders and they might conceal some fraud transactions. Hence, it could be
expected that higher proportion of director ownership results in lower level of disclosure. The above expectation is supported by Chau and Gray 2002, Eng and Mak 2003, and Leung
and Horwitz 2004 since they find that there is a negative relationship between director ownership and disclosure. Therefore, the following hypothesis, in an alternative form, is
developed.
Ha4: Social and environmental disclosure is negatively related to director ownership. On the contrary, in the case of institutional and block ownership, they might put the pressure
on the management to disclose all the material information. It is because disclosure is one of the best ways to find out whether the performance of the management meets their
expectations, for example, in terms of profitability and risk management. From the agency theory, it also could be derived that since institutional shareholders and block holders have
higher proportion of ownership interest, they might be more interested to monitor the firms for better disclosure in order for them to make the economic decisions as well as to discipline
poorly performing management Kang Sorensen, 1999; Kim Nofsinger, 2004; Birt, Bilson, Smith and Whaley, 2006. This theoretical expectation, i.e. higher institutional or
block ownership has a positive impact on disclosure, is supported by the findings of Eng and Mak 2003, and Lakhal 2005, whereby Eng and Mak 2003 use government ownership as
a proxy for institutional shareholder and Lakhal 2005 uses the foreign institutional investors ownership as the proxy. In the case of block ownership, Chau and Gray 2002,
Luo et al. 2006, Huafang and Jianguo 2007 and Norita and Shamsul Nahar 2004 find that extent of outside block ownership is positively associated with voluntary disclosures and
hence their finding is in line with theoretical expectation. Therefore, the following hypotheses, in an alternative form, are developed. They are:
Ha5: Social and environmental disclosure is positively related to block ownership. Ha6: Social and environmental disclosure is positively related to institutional ownership.
3.2 Research Design
3.2.1 Variables and Empirical Model Dependent Variable
Weighted social and environmental information disclosure score is used as a dependent variable and questionnaire Refer to Appendix is developed to obtain views on the
importance of each disclosure item from financial analysts and accountants. Before the actual questionnaire is sent, pilot test has been conducted and the findings show that alpha value is
Asian Journal of Finance Accounting
ISSN 1946-052X 2012, Vol. 4, No. 1
9 www.macrothink.orgajfa
0.94 and so it has been concluded that the questionnaire is reliable. In addition, pilot test results show that the overall mean score of comprehensiveness of the questionnaire is 4.05,
understandability of the questions is 4.10 and understandability of the instruction is 4.62. Therefore, it can be concluded that the pilot test questionnaire is good enough to be used as
an actual questionnaire. List of social and environmental disclosure items and pilot test results can be referred to Table 1 and Table 2 respectively.
Table 1: List of Social and Environmental Information Disclosure Share option schemes-Policy Quantitative
Share option schemes-Policy Qualitative Profit sharing schemes-Policy Quantitative
Profit sharing schemes-Policy Qualitative Amount spent on training
Policy on training Welfare information
Recruitment policy Charitable donations Quantitative
Charitable donations Qualitative Community programs Quantitative
Community programs Qualitative Environmental protection policy Quantitative
Environmental protection policy Qualitative Effect on environment Quantitative
Effect on environment Qualitative
Table 2. Mean Score for Comprehensiveness and Understandability of the Questions and the Instruction Pilot Test
Mean No.
Description Accountants
Financial Analysts Overall
1 The questionnaire comprehensively
covers the important disclosure items of the annual reports.
4.22 3.92
4.05
2 The contents of the questionnaire are
easy and simple for the respondents to understand.
4.22 4.00
4.10
3 The
instructions to
answer the
questionnaire are clear. 4.78
4.50 4.62
The annual reports of the sample companies are checked against disclosure index developed by the researcher. The researcher uses dichotomous score, i.e. one is given if the company
Asian Journal of Finance Accounting
ISSN 1946-052X 2012, Vol. 4, No. 1
10 www.macrothink.orgajfa
discloses the information, and zero for otherwise. Since the annual reports are checked against the disclosure index to provide the disclosure score, during this process, some of the
disclosures in the annual reports are not clear for the researcher to decide whether some parts of annual report disclosure represent the items from the disclosure index. Hence, for these
confusing items, questionnaire is constructed and sent to the ten accountants and six financial analysts in order to seek their opinions on whether these confusing disclosures in the annual
reports represent the items in the disclosure check list. It is found out that there is no significant difference between the score provided by the researcher and the answers provided
by the selected accountants and financial analysts. Finally, the weight for each disclosure item is calculated by the mean score of each disclosure item provided by the accountants and
financial analysts.
Independent Variables There are six independent variables, which comprise of three structural measures of corporate
governance i.e. board leadership structure, board composition and board size and three measures of ownership structure i.e. director ownership, institutional ownership, and block
ownership. Finally, the empirical model of the study also includes two control variables related to firm-specific characteristics i.e. firm size and leverage because most of the prior
researchers such as Ho and Wong 2005, Chau and Gray 2002, Haniffa and Cooke 2002, Hyytinen and Pajarinen 2005, Eng and Mak 2003, Leung and Horwitz 2004, Ballesta
and Garcia-Meca 2005, Lakhal 2005, Luo et al. 2006, Huafang and Jianguo 2007 and Norita and Shamsul Nahar 2004 have used the above two variables as control variables in
examining the relationship between corporate governance variables and disclosures. Firm size is controlled because larger and more profitable firms tend to have higher disclosure
because they are closely followed by financial intermediaries and need more comprehensive disclosure to minimize the political costs of noncompliance Aksu Kosedag, 2006. For
instance, Kee and Pillay 2003 find that firm size is significantly associated with voluntary disclosure level. In addition, leverage is used as a control variable because in highly
leveraged firms, there is a high demand for and supply of information required by the lenders Aksu Kosedag, 2006. According to the corporate governance literature, firm size is
commonly measured by total assets and leverage is measured by the ratio of total debt to equity. The complete empirical model is as follow.
Yit= βo + β1 x1it+β2 x2it − β3 x3it − β4 x4it+β5 x5it + β6 x6it + β7 x7it + β8 x8it + µit Where,
i = 1,2,3,4,5,6,7,8,9,10,11,12 t=1, 2, 3, 4, 5,6,7,8,9,10
Y= Weighted social and environmental information disclosure score x1= Board leadership structure BLS
x2= Proportion of independent non-executive directors on the board INE_BZ
Asian Journal of Finance Accounting
ISSN 1946-052X 2012, Vol. 4, No. 1
11 www.macrothink.orgajfa
x3= Board size BZ x4= Proportion of director ownership DOWN
x5= Proportion of institutional ownership IOWN x6= Proportion of block ownership BOWN
x7= Log of total assets TA x8= Leverage TD_TE
µ= Error term 3.2.2 Sample selection and Statistical Methods
Samples include the twelve listed companies whose main activity is banking from 1996 until 2005. The total number of observations is 120 observations. However, some of the
observations need to be dropped due to unavailability of data and some companies were not classified as banks in all the ten years’ period. It left the final observations to 108
observations. There were significant events incurred during the 10-years span of the sample period. First, the introduction of MCCG 2007 promotes better transparency of information
disclosure. Secondly, the financial crisis incurred towards the end of the 1990s. Thirdly, CSR framework is introduced by Bursa Malaysia in 2006 and this framework lays down regulation
on disclosure of social and environment information. Therefore, the period before 2006 will be better in capturing the nature of purely voluntary environment for social and
environmental disclosure. Data were collected either from the annual reports of the companies or from Bloomberg. The main statistical method used in this study is panel data
analysis generalized least square method. Generalized least square method is used because the sample data are not normally distributed and the data have, either, heteroskedasticity
problem, autocorrelation problem or both. According to Gujarati 2003, using generalized least square method will overcome all these problems.
4. Profile of the Respondents
