Corporate political connection and corporate social responsibility disclosures

Corporate Political Connection and Corporate Social Responsibility Disclosures: A Neo-Pluralist Hypothesis and Empirical Evidence Mohammad Badrul Muttakin Dessalegn Getie Mihret Arifur K

Corporate Political Connection and Corporate Social Responsibility Disclosures: A Neo-Pluralist Hypothesis and Empirical Evidence

Mohammad Badrul Muttakin Dessalegn Getie Mihret Arifur Khan Department of Accounting, Deakin University, Melbourne, Australia

Forthcoming in Accounting, Auditing and and Accountability Journal

Corporate Political Connection and Corporate Social Responsibility Disclosures: A

Neo-Pluralist Hypothesis and Empirical Evidence Abstract

Purpose – This study examines the association of corporate political connection with the

level of voluntary corporate social responsibility (CSR) disclosures to determine how the relationships between the state and the corporate sector influence CSR engagement

Design/methodology/approach – Based on a neo-pluralist view of legitimacy theory, which

conceptualises the state as a concentration of power amenable to exploitation by the corporate sector, the study develops and empirically tests a hypothesis that CSR disclosures are inversely associated with political connection A sample of 936 firm-year observations is used with data collected from annual reports of companies listed on the Dhaka Stock Exchange in Bangladesh from 2005 to 2013

Findings – Results indicate that corporate political connection is associated with reduced

CSR disclosures This finding suggests that the perceived need for CSR disclosures as a legitimation strategy diminishes for politically connected firms The finding supports a neo-pluralist argument that political connection could enable firms to eschew stakeholder pressure associated with potential legitimacy threats originating from poor CSR performance This conclusion challenges the pluralist view of legitimacy theory that considers the state as a neutral arbiter resolving conflict among stakeholder groups in society

Originality/value – The study makes a significant contribution to the literature by

developing a neo-pluralist theorisation of voluntary CSR disclosures within legitimacy theory and empirically testing it Because prior empirical CSR disclosure research is largely underpinned by the pluralistic conception of society, examining this phenomenon from a neo-pluralist perspective enables a more complete understanding of CSR disclosure behaviours of firms

Keywords – Corporate Social Responsibility, Legitimacy theory, Neo-pluralist view,

Stakeholder theory, Bangladesh

Paper type – Research paper

1 Introduction

Voluntary corporate social responsibility (CSR) disclosures have attracted significant research attention in recent decades due to heightened societal interest in CSR One of the focus areas in this respect has been understanding the rationales for firms’ CSR engagement (Parker, 2014) and interpreting ensuing corporate behaviour Prior research indicates that one possible motivation behind voluntary CSR disclosures is to portray business practices as

congruent with societal expectations (see, for example, Cho et al., 2012; Deegan and Rankin,

1996; Milne and Patten, 2002; O’Donovan, 2002; Suchman, 1995) This view maintains that CSR disclosures are driven by the interest in legitimising the firm’s existence by placating key stakeholders but not necessarily discharging corporate accountability obligations (Killian and O'Regan, 2016) pertaining to the environmental and social impact of business on society (Cho

et al., 2015; Deegan and Rankin, 1996, p 60; Milne and Gray, 2013) The need for legitimacy

is underpinned by the concept of an underlying social contract between corporations and society To maintain the social contract, firms attempt to influence legitimacy by attending to stakeholders’ expectations—including those of the government, consumers (Deegan and Rankin, 1996), shareholders and special interest groups (Roberts, 1992), for example, through voluntary CSR disclosures Legitimacy theory has been widely adopted to explain voluntary CSR disclosure behaviours of firms as a strategy to communicate firms’ actual CSR performance or to manipulate stakeholder perceptions about such performance This mainstream articulation of legitimacy theory is problematic because it is premised on the pluralist notion of society, which conceptualises the state as a neutral arbiter for freely competing interests in society

From the pluralist perspective, power—the capacity to influence others (see Max Weber’s definition of power in Gerth and Mills, 1948, p 180)—is conceived as widely

distributed among stakeholders in society, albeit not necessarily evenly This notion of power

suggests that multiple stakeholders can influence state policymaking and that no particular

stakeholder can systematically dominate others (Gray et al., 1996, p 16) Nevertheless, power

asymmetries abound, especially in non-pluralistic societies, and the state may serve as an institution with a concentration of power amenable to exploitation by dominant interest groups

in society, including corporations (Archel et al., 2011) as the state may align itself with

interests This perspective challenges the concept of neutrality of the state that underpins the pluralist view of society By recognising the limitations of the pluralist assumptions, a critical strand of the legitimacy theory literature interprets the theory from a neo-pluralist perspective

(Soobaroyen and Mahadeo, 2016; Guthrie and Parker, 1989; Gray et al., 1996), which

recognises the unequal distribution of power in society and questions neutrality of the state

(Archel et al., 2009) This alternative theoretical perspective warrants empirical investigation

because CSR disclosure issues are gaining in importance in the era of globalisation, when stakeholders may be interested in firms operating in environments atypical of a pluralist society Despite repeated calls to refine CSR research within legitimacy theory (Guthrie and

Parker, 1989; Gray et al., 1996), empirical research has largely continued to adopt the pluralist

view of the theory In particular, empirical studies based on the neo-pluralist view of society are yet to emerge

This study responds to these calls by developing and empirically testing a neo-pluralist hypothesis on a possible inverse association between corporate political connection and the level of voluntary CSR disclosures The study is based on data collected from annual reports of non-financial companies listed on the Dhaka Stock Exchange (DSE) in Bangladesh Our findings support the neo-pluralist hypothesis that politically connected firms perceive a reduced need for CSR disclosures as a legitimation strategy We argue that firms may use political connection to fend off potential risk of losing the social contract associated with perceived poor CSR performance The empirical context of Bangladesh provides an ideal setting to empirically test such a neo-pluralist hypothesis on CSR disclosures, because the country is characterised by

a tenuous institutional environment, which implies limitations on mechanisms of enforcing

desirable corporate behaviour in the event of poor CSR performance In such contexts, firms could employ political connection rather than legitimating through voluntary CSR disclosures

to avoid possible legitimacy challenge regardless of whether the firms meet CSR performance expectations

This study contributes to research and policy as follows First, it extends the literature

on voluntary CSR disclosures within legitimacy theory by developing and empirically testing a neo-pluralist hypothesis The study provides a timely response to Richardson’s (2015) call to challenge mainstream views in accounting research through quantitative critical research and Patten’s (2015) call for methodological diversity to achieve greater understanding of research phenomena within the critical stream of research In this respect, the study enriches the existing critical theorisation regarding drivers of voluntary CSR disclosure by generating an additional factor, namely, corporate political connection that influences voluntary CSR disclosures Because legitimacy theory research has largely adopted a pluralistic view of society, our study

in a setting where key assumptions of a pluralistic society are absent enables testing of the theory’s explanatory power Second, by exposing corporate political connection as a key restraining factor that could limit CSR performance, results could inform stakeholders such as

non-governmental organisations (NGOs) (Belal et al., 2015) and other civil society groups (see

Apostol, 2015) in any effort to advocate for policy changes directed at ensuring corporate accountability to society Furthermore, given the increasing integration of developing economies into the global supply chain and the tendency of CSR reporting practices to vary by country (Chen and Bouvain, 2009; Kolk and Pinkse, 2010; Young and Marais, 2012), the findings of this study may inform international stakeholders interested in exploring CSR issues pertinent to emerging economies as an input for making business decisions

The remainder of this paper is organised as follows In section 2, we further develop the motivation for the study by highlighting the empirical context of Bangladesh This is followed

by the theoretical framework and hypothesis formulation in section 3 Section 4 outlines the

research design, and section 5 presents our data analysis, which is followed by a conclusion in section 6

2010 (30%) The growth in foreign direct investment (FDI) has also been remarkable during the last two decades, as a number of more-developed Asian countries have outsourced their factory production, mainly textiles, to Bangladesh However, rapid industrialisation and investments by overseas investors has generated concerns about corporate accountability regarding employee, environmental and ethical issues, which has led to increasing demands for enhanced transparency in business practices (Belal and Owen, 2007) Furthermore, in recent years, some high-profile environmental and safety-related corporate disasters that claimed hundreds of lives in the garment industrial sector have resulted in significant pressure from foreign buyers on firms to demonstrate socially responsible business practices Well-publicised

incidents include the collapse of ‘Rana Plaza’ (Siddiqui and Uddin, 2016; Sinkovics et al.,

2016.), a factory building, that killed over 1200 garment workers on April 24, 2013 (Washington Post, 2013) Similarly, 112 people were burned alive or jumped to their deaths when the Tazreen Fashions Ltd building caught fire on November 24, 2012 (The Daily Star, 2012) As a result, U.S President Barack Obama announced the suspension of U.S trade privileges for Bangladesh in June 2013 over concerns regarding labour rights and workers’ safety (Washington Post, 2013) These incidents attracted international attention to Bangladeshi politics, labour rights, and workers’ safety and resulted in international buyers

facing significant pressure to demand socially responsible business practices from Bangladeshi suppliers

The incidents also reveal the politics-business nexus in Bangladesh and its implications for socially responsible corporate behaviours For example, no legal actions were taken against Delwar Hossein, the owner of Tazreen Fashions, because the local industry body, the Bangladesh Garments Manufacturers and Exporters Association (BGMEA), a powerful lobby group whose members account for approximately 10 percent of parliament, provided legal assistance to him The group defended him on the grounds that he was not on the premises at the time the fire broke out (Allchin, 2013) Similarly, Sohel Rana, owner of ‘Rana Plaza’, was

a local political leader of the ruling party Bangladeshi Awami League (BAL) (The Wall Street Journal, 2013) Following the Rana Plaza incident, ‘the Bangladeshi Prime Minister Sheikh Hasina publicly denied Rana’s affiliation with her party although the media exposed—shortly afterwards—his continued political activities as a member of the ruling party (Gomes, 2013)

Politically connected businesses are commonplace around the world (Faccio et al.,

2006), although they are less common in countries that apply stringent regulations against political conflicts of interest (Faccio, 2006) Particularly, corporate political connection is pervasive in countries with weak regulatory environments and high levels of corruption

Bangladesh is characterised by a poor regulatory environment (Farooque et al., 2007) and

according to Transparency International’s index, the country has a high level of corruption [1]

Furthermore, Muttakin et al (2015) highlight that political connection is extensive in

Bangladesh Business owners can easily obtain party nominations for parliamentary elections

by making large donations, thereby placing them in a position to use the government system to maintain, defend and advance business interests Bangladeshi business owners such as those in the garment industry are some of the most powerful politicians in the country Workers and their unions shun expressing opposition to the businessmen because the financial and political power of the businessmen portends excessive costs of legal and political action against any

pressure group For example, Yardley (2012) notes that Bangladesh factory owners are major political donors, and some of them own newspapers and television stations, which feature relatively positive news about businesses while ignoring issues such as labour rights He also notes that in Bangladesh’s parliament, roughly two-thirds of the members belong to the country’s three largest business associations [2]

Political connection in the Bangladeshi corporate governance context takes two forms First, government officials could serve as members of the boards of directors at numerous listed companies, or retired politicians may be assigned as directors arguably in exchange for favours (see, for example, Sobhan and Werner, 2003) Second, many industrial elites in Bangladesh either become part of the political process or develop strong personal and/or business relationships with political leaders and their families and make financial contributions

to the major political parties (Chowdhury, 2009) The success of such businesses tends to depend upon the connections they maintain with political parties In recent years, a large number of businessmen in Bangladesh have mainly been tied with two major political parties, i.e., Bangladeshi Awami League (BAL) and the Bangladesh National Party (BNP) In the ninth parliament, 59% of the elected Members of Parliament were businessmen, with 44% having assets worth at least one crore taka (US$10 million) (Chowdhury, 2009) As argued later in this paper, these forms of political connection can serve as a strategy to manage potential legitimacy gaps and associated regulatory or political action

2.2 Bangladesh’s institutional background and CSR disclosure

The regulations that govern the business practices of Bangladeshi firms include the Companies Act, 1994, the Income Tax Ordinance, 1969, and the Securities and Exchange rules, 1987 Nevertheless, neither any legislative pronouncement nor the Securities and Exchange rules, 1987, mandate CSR disclosures for public limited companies CSR disclosure levels of Bangladeshi companies tend to be low and of a descriptive nature, reporting mostly

positive news (Belal, 1999; Imam, 2000) The disclosures also tend to focus on

employee-related matters, as compared to CSR information on a more comprehensive set of issues (Belal, 2001) Disclosure of information on important social issues such as child labour, equal opportunities and poverty alleviation is limited due to a focus on profits and the absence of legal requirements (Belal and Cooper, 2011) While a mandatory requirement for CSR disclosures is regarded as an option to enhance corporate accountability on CSR matters, a lack

of political will and prohibitive costs involved may hinder such a move (Belal et al., 2015) By

contrast, the desire to satisfy key stakeholder groups such as international buyers, tends to generate motivation for CSR reporting in Bangladesh (Belal and Owen, 2015; Belal and Owen 2007)

3 Theoretical framework and hypothesis development

3.1 A neo-pluralist view of legitimacy theory and CSR disclosures

Legitimacy theory is widely adopted to explain voluntary CSR disclosures (e.g., Deegan, 2002; Deegan, Rankin and Tobin, 2002; O’Donovan, 2002) Guthrie and Parker (1989) questioned the explanatory power of the theory and argued that political economic considerations determine corporate decisions to disclose CSR information Similarly, recent

critical literature (Archel et al., 2009; Gray et al., 1995) problematises the pluralist version of

legitimacy theory by interrogating the key assumptions underpinning this version of the theory This dominant articulation of legitimacy theory is grounded in the concept of a ‘social contract’, which is envisaged to provide the basis for identifying the rights society grants to firms and the societal expectations that ensue in return From this perspective, firms will have the legitimacy to operate if the values and actions of the firms are consistent with societal values and expectations Such congruence enables the firm to garner community acceptance and good corporate image necessary for the firm’s continued existence (Deegan and Rankin, 1996) On the other hand, disparity between the two phenomena produces a legitimacy gap that might prompt economic, legal, social, or political sanctions on organisations (Deegan and

Rankin, 1996; Dowling and Pfeffer, 1975) That is, the pluralist view envisages that the implicit and explicit terms of the social contract would be discontinued if companies fail to operate within the bounds of societal expectations, i.e., when corporate legitimacy is lost This view maintains that to avoid these consequences, firms employ strategies to justify their

existence by inter alia providing voluntary CSR reports aimed at portraying the firm as a

socially responsible business (Deegan and Rankin, 1996)

Organisations influence the perceived legitimacy society confers on them (Dowling and Pfeffer, 1975) through managing legitimacy threats using a range of strategies, including CSR disclosures (O’Donovan, 2002) This conception draws on the liberal idea of individual freedom to act and exercise choice through the market for economic matters or through political action for non-economic issues Because the state provides mechanisms for political action, the liberal notion of CSR is premised on the neutrality of the state in relation to arbitrating competing interests Neutrality of the state enables stakeholders to exert pressure to

“enforce” the social contract in the event that corporations fail to meet CSR obligations For example, legal requirements place accountability obligations on firms (Gray et al, 1996, p 39-

40; 43), and CSR reporting could serve as a communication channel in this respect (Gray et al.,

1996, p 44) Such reports may serve to communicate substantive CSR engagements of firms, represent a symbolic act without an actual commitment to CSR, or a combination of these two possibilities (Ashforth and Gibbs, 1990) That is, voluntary CSR disclosure could be provided

to legitimate the existence of the corporation (Deegan, 2010, p 334) by disclosing information

on actual CSR activities or by merely altering the perceptions of stakeholders about CSR activities of the firm The pluralist view maintains that pressure on companies to engage in socially responsible activities originates from stakeholders, such as the state, consumers (Deegan and Rankin, 1996), special interest groups, shareholders (Roberts, 1992), and NGOs

(Belal et al., 2015) To address potential or actual pressure in this regard, firms may employ legitimation strategies by attending to the stakeholders’ expectations (Gray et al., 1995;

O’Donovan, 2002) Along this line of thought, Deegan et al (2002) find that environmental

disclosures are motivated by management’s interest in legitimising corporate activities In pluralistic societies, firms identify key stakeholders and devise legitimation strategies to

address stakeholder expectations regarding CSR (Bailey et al., 2000; Buhr, 2002; Nasi et al.,

1997) For instance, organisations attend to the expectations of the state to reduce the possibility of issuance of new legislation detrimental to the operation of the firm (Deegan and Rankin, 1996; Roberts, 1992)

The explanatory power of legitimacy theory regarding this commonly held view on voluntary CSR disclosures has been criticised (see, for example, Guthrie and Parker, 1989),

and calls have been made to extend legitimacy theory into a more critical domain (Archel et

al., 2009) The critical view challenges the claim that stakeholder groups can exert pressure on

firms to discharge corporate accountability pertaining to CSR matters (see, for example, Deegan and Gordon, 1996) and that the state provides neutral institutional mechanisms to adjudicate conflicts that involve multiple stakeholders The pluralist view on these issues is based on a problematic assumption that power is widely distributed among freely competing groups in society and that the state is free of ideological alignment with particular stakeholders

in society Contrary to this conception, the state may align its interests with the corporate sector, as, for example, advancing corporate profit also contributes to achieving macro-economic growth goals of the state (Miller, 1991) Furthermore, power asymmetries are commonplace in practice, and not all societies exhibit wide distribution of power

The Marxist literature advances a radical view that the state is a mechanism dedicated to preserve the interests of the corporate sector While this extreme view could be arguable, the neo-pluralist view of society maintains that the state could “willingly or under pressure” align

with corporate interests (Gray et al 1996, p 19; see also Tinker, 1984) This critical view of

legitimacy theory bears particular relevance for emerging economies, where the neo-pluralist conception of society applies and thus associated implications for voluntary CSR disclosures

cannot be ruled out While companies could face “political, social and economic” pressure if corporate social performance falls short of societal expectations, firms’ choice of strategies to address such challenges proactively is likely to depend on the political setting and the ensuing role of the state In non-pluralist societies where the state could align with corporate interests, firms could employ political connection as a more crucial strategy for their survival than legitimating through CSR disclosures Against this background, we develop a neo-pluralist hypothesis on the association between CSR disclosures and political connection

3.2 Political connections and CSR disclosure: a neo-pluralist hypothesis

Conceptualising the state as potentially aligned with the interests of some groups in society permits exploring empirical possibilities regarding CSR disclosure behaviours of firms This view calls for a departure from the pluralist assumption that power is distributed among various members of society and that the state lacks ideological alignment and material connection with some groups in society Political connection could be one form in which the ideological alignment of the state with the corporate sector manifests itself or an instrumental connection between the two develops This connection could, in turn, interfere with the role of the state as a neutral adjudicator in society with respect to stakeholder pressure on firms for increased CSR engagement Legitimacy threats originating from poor CSR performance may eventuate when the state provides neutral institutional and legal mechanisms of enforcement to address stakeholder concerns, for example, by introducing a new legislation (O'dwyer, 2002) Critical theorising of the role of the state permits a new way of interpreting implications of political connection for CSR activities and disclosures

In addition to other corporate attributes such as firm size, board characteristics, industry type, and export orientation that influence voluntary CSR disclosure practices (Patten, 1992; Roberts, 1992), political connection could impact on CSR disclosure behaviours of firms Firms may be interested in maintaining relations with politicians because people holding key

government positions might help the firms, for example, through relaxed regulatory oversight

on CSR matters (Hillman, 2005; Niessen and Ruenzi, 2010) In some institutional settings, a

reciprocal relationship exists between firms and government systems (Aronson et al 2005) in

which corporate social activities serve firms to pay back the government for support and to gain

continuous political benefits in the future (Li et al., 2015)

This argument suggests that firms may use political connections to eschew regulatory action that could originate from stakeholder pressure Particularly, in a weak institutional environment such as that of Bangladesh, the management of politically connected firms could focus more on advancing economic interests than allocating resources to CSR activities From

a neo-pluralist perspective, it can be argued that a trade-off exists between political connection and CSR disclosures Siddiqui and Uddin (2016) examine the state-business nexus in response

to human rights violations in Bangladeshi businesses and document that the nexus exacerbates human rights disasters They also show that the Bangladeshi state, ruled by family-led political parties, is more inclined to protect businesses than to ensure the observance of human rights by businesses In a neo-pluralist society where power asymmetries abound and the state may be

prone to align with interests (Archel et al., 2009; Gray et al., 1995) of the corporate sector,

political connection may undermine the need to garner legitimacy through genuine engagement

in CSR activities Thus, we propose the following hypothesis

Hypothesis: There is an inverse association between corporate political connection and the extent of CSR disclosure

4 Research design

4.1 Sample and data

The sample selection procedure is reported in Table 1 Our sample consists of all 155 non-financial companies listed on the Dhaka Stock Exchange (DSE) in Bangladesh from 2005

to 2013 [3] [4] Of the 1,395 firm-year observations, we excluded 459 due to missing or

incomplete information The final sample comprises the remaining 936 firm-years The data for our analysis are obtained from multiple sources We collected the financial data from the annual reports of the sample companies CSR information was hand-collected from CSR disclosures, corporate governance disclosures, directors’ reports, Chairman’s statements, and notes to the financial statements contained in the companies’ annual reports Data on political connections were collected from a variety of sources We used national election data from the Bangladesh Election Commission to identify members of the boards of directors in our sample firms who contested or were elected in national parliamentary elections during the sample period (2005-2013) We also checked the names of committee members and advisory council members of political parties from party websites Finally, we consulted four major national

dailies (namely, The Daily Star, The Bangladesh Observer, The New Nation and The Financial

Express) to identify directors’ political affiliations Our process of identifying political

connections is consistent with that of other studies (Muttakin et al., 2015; Chaney et al., 2011; Faccio et al., 2006)

<Table 1 about here>

4.2 Model specification

We used OLS regression to test the hypothesised relationship between corporate social responsibility disclosure (CSRDI) and corporate political connections The corporate political connection (PCON) variable is measured in two ways: a dummy variable for presence of politically connected board members or major shareholders (PCONDUM) and proportion of politically connected board members to total board members (PCONPRO) To test our hypothesis, we use the following model:

CSRDI = α + β 1 PCON + β 2 LEV + β 3 FAGE + β 4 FSIZE + β 5 ROA+ β 6 YEAR DUMMIES + ε

Where:

CSRDI is the corporate social responsibility disclosure score/ index;

PCON is measured in two ways: (i) PCONDUM equals 1 for politically connected firms and 0 otherwise We consider a firm to be politically connected when at least one of its major shareholders or a member of its board or the CEO is or was a member of the parliament, a minister or closely associated with a political party or

a politician and (ii) PCONPRO is the proportion of politically connected directors

on the board;

LEV is the ratio of book value of total debt to total assets Companies with high leverage may have closer relationships with their creditors and use other means to

disclose social responsibility information (Purushothaman et al., 2000);

FAGE is the natural log of the number of years since the firm’s inception An older firm

is concerned about its reputation and provides more social responsibility disclosures (Roberts, 1992);

FISZE is the natural logarithm of total assets Larger firms receive more attention from the various groups in society and, therefore, would be under greater pressure to disclose more social activities to legitimise their businesses (Haniffa and Cooke 2005); and

ROA is the ratio of earnings before interest and taxes to total assets Profitable companies demonstrate their contribution to society’s well-being and legitimise their existence through disclosing more social information (Haniffa and Cook, 2005)

4.3 Dependent Variable- Corporate Social Responsibility disclosure index (CSRDI)

CSR disclosure index (CSRDI) represents the dependent variable in this study To assess the extent of CSR disclosure in annual reports, a checklist containing 20 items was constructed

We follow previous Bangladeshi studies (Khan et al., 2013; Muttakin and Khan, 2014) in

constructing this checklist These 20 items comprise five categories, i.e., community, environment, employee, product and service, and value-added information, which are relevant

to the Bangladeshi environment Community involvement covers three items: charitable donations and subscriptions, sponsorship and advertising and community programmes (health and education) The environment category covers disclosures regarding environmental policies The employee information covers nine aspects, i.e., number of employees, employee relations, employee welfare, employee education, employee training and development, employee profit sharing, managerial remuneration, workers’ occupational health and safety, child labour and related actions The product and service category covers six aspects: types of products disclosed, product research and development, product quality and safety, discussion of marketing networks, customer service and satisfaction, and customer awards Value added

information refers to corporate disclosure of a value-added statement To develop our CSR disclosure index, a dichotomous procedure is applied whereby a company is awarded 1 if an item included in the checklist is disclosed and 0 otherwise The CSR disclosure index is then derived by computing the ratio of actual scores awarded to the maximum score attainable (20)

by that company (Haniffa and Cooke 2005, Khan et al 2013)

We undertook a content analysis to develop our CSR disclosure index To ensure reliability and consistency, we read the annual reports thoroughly (Haniffa and Cooke, 2005) before any decision was made about scoring We also took a number of precautionary steps to address potential data reliability issues that could arise Two of the researchers independently conducted the coding for a sample of companies to identify the possibility of multiple interpretations The researchers verified each other’s coding sheets to ensure reliability of ratings and discussed possible issues to consider in performing the coding for the whole sample The two researchers divided the population in two and coded the data The two researchers continually discussed when they encountered issues not highlighted during the initial discussion

We used Cronbach’s coefficient alpha (Cronbach, 1951) to assess the internal consistency of CSRDI, i.e., the degree to which the items in a test measure the same construct Botosan (1997) and Gul and Leung (2004) use coefficient alpha as a reliability statistic useful for assessing the degree to which correlation among the information categories of the disclosure index is attenuated due to random error The coefficient alpha for the five information categories in our disclosure index is 0.79 This result provides good support that the set of items in the disclosure scoring index captures the same underlying construct

5 Results

Panel A of Table 2 provides descriptive statistics for the variables used in the study The average CSR disclosure (CSRDI) for the sample is 0.245 Fifty-six percent of the companies in

our sample are politically connected (PCONDUM) and 34.8% are export (EXPORT) firms The average firm size (FSIZE) of companies in the sample is 8.84 (natural log), the average firm age (FAGE) is approximately 25 years, and average level of profitability (ROA) is approximately 7% Panel B of Table 2 provides the average CSR disclosure (CSRDI) based on observations classified by industry We find that firms in the food (pharmaceutical) sector provide lowest (highest) level of CSR disclosures

<Table 2 about here>

Table 3 presents the correlation matrix among the variables in our regression model Political connection (PCONDUM) and proportion of politically connected directors on board (PCONPRO) are negatively correlated with CSR disclosure index (CSRDI) We also document that the CSR disclosure index (CSRDI) is positively correlated with board independence (BIND) and export-orientation of companies (EXPORT) The control variables such as leverage (LEV), firm age (FAGE), firm size (FSIZE) and profitability (ROA) are significantly correlated with CSR disclosure index (CSRDI).[5]

<Table 3 about here>

Table 4 reports the results of regressing the hypothesised variables on CSR disclosure (CSRDI) In model 1, we examine the impact of political connections (PCONDUM) on the CSR disclosure index (CSRDI) We find a negative and significant coefficient (β=-0.029, p<0.001) of the political connections (PCONDUM) variable This implies that a firm’s political connections result in a lower level of CSR disclosure, supporting our hypothesis Indeed, the strength of the political influence could vary with the number of politically-connected members

on the board It could be argued that the influence of a single politically-connected member could be less decisive than that of several connected directors Thus, we consider an alternative measure of political connections in model 2 of Table 4 that could capture this effect: the proportion of politically connected directors (PCONPRO) on the board Once again, we find a