Have labour practices and human rights disclosures enhanced corporate accountability? The case of the GRI framework

Highlights • Companies over-claim adherence to GRI reporting guidelines while failing to report detailed information on their workforce.Fig. 1 Compliance with GRI Disclosure Guidelines.• Companies fail to provide material information.• Limited evidence of companies acknowledging impediments they encounter when reporting on their workforce.• Companies pay more attention to their internal (as opposed to their external) workforce.• GRI fails to achieve enhanced comparability, transparency and accountability goals. Highlights • Companies over-claim adherence to GRI reporting guidelines while failing to report detailed information on their workforce.• Companies fail to provide material information.• Limited evidence of companies acknowledging impediments they encounter when reporting on their workforce.• Companies pay more attention to their internal (as opposed to their external) workforce.• GRI fails to achieve enhanced comparability, transparency and accountability goals. Abstract This paper critically evaluates Transnational Corporations’ (TNCs) claimed adherence to the Global Reporting Initiative (GRI)’s ‘labour’ and ‘human rights’ reporting guidelines and examines how successful the GRI has been in enhancing comparability and transparency. We found limited evidence of TNCs discharging their accountability to their workforce and, rather, we found evidence to suggest that disclosure was motivated more by enhancing their legitimacy. TNCs failed to adhere to the guidelines, which meant that material information items were often missing, rendering comparability of information meaningless. Instead, TNCs reported large volumes of generic/anecdotal information without acknowledging the impediments they faced in practice.

border nature of such relationships, bringing additional legal, institutional and socio-cultural dynamics into play. So TNCs may have reasons to find reporting on the external workforce challenging.
Most of the human rights endeavours of civil society activities relating to the external workforce are not enforceable by international institutional benchmarks such as UN frameworks (Sikka, 2011) or by national-level government (Cooper, Coulson, & Taylor, 2011). At a deeper level, promoting the ethical treatment of supply-chain workers is challenged by the global institutional environment: the preeminent neoliberal hegemony on corporate governance not only privileges the interests of capital over labour, it actively promotes the marginalisation of any interests competing with the pursuit of shareholder value (Sikka, 2010).
Focusing on the accountability of TNCs to their workforce in the complex context of global value chains, this paper provides evidence on how TNCs, across sectors and countries of origin, treat information disclosure. It also comments on the extent to which such reporting is part of a process of incremental progress or merely a platform to gain/maintain corporate legitimacy on the issue of labour rights in the face of popular concerns. One distinguishing feature of this paper is the way we closely examine the detailed information that TNCs are expected to have disclosed once they claim disclosure on their (internal and external) labour practices in their GRI Index table. As part of the reporting process, the GRI index table enables TNCs to explicitly declareand cross-referencethe extent of their compliance to the reporting guidelines and thus illustrate the materiality of the information reported. The examination of index tables allowed us to reveal how important it was for TNCs to illustrate the level of adherence to this internationally respected set of guidelines (the GRI) even where within the contents of the full report, some TNCs were prepared to declare the 'immateriality' of such disclosure on a variety of justifications. In the absence of mandatory audit or any meaningful assurance provision, TNCs use the GRI Index table to gain and maintain legitimacy. We argue that, in doing so, TNCs are applying Lindblom's (1993) second, third and fourth legitimation strategies with little concern for the risk of being 'found out' or rebuked at a later stage. This undermines the materiality and hence the comparability of the information that TNCs report.
The paper is structured as follows. The next section presents a review of the existing literature on reporting practices on labour issues and how this intersects with legitimacy theory. In the methods section, it is explained how the GRI enabled the examination of the extent and quality of information that TNCs had disclosed. The results are then discussed in two parts; the first concentrating on information found on TNCs that had disclosed on their labour and human rights performance indicators and the second on the extent to which TNCs had over-claimed the true extent of disclosure on each indicator. Conclusions and discussions are presented in the final section.

Accountability to the workforce: a global challenge
A company's workforce is regarded as a 'primary stakeholder' in most models (Donaldson & Preston, 1995;Freeman, 1984). Companies that claim a high degree of ethical behaviour toward their workers are likely to be perceived as more socially responsible in general (Cohen, Taylor, & Muller-Camen, 2012) and therefore to be perceived more favourably as a result (Crane, Matten, & Spence, 2008).
A company's concern for workers' welfare can be traced back to the philanthropists of the nineteenth century, when the impulse to improve employees' welfare was deemed to be mutually beneficial (Cannon, 1994). In the twentieth century, such philanthropic approaches were superseded by more instrumentalist concerns about satisfying institutionalised forms of employee interests through collective bargaining. However, due to the decline in unionization globally (Wailes, Bamber, & Lansbury, 2011) and increased labour market deregulation it has become more difficult to promote standardised company employment practices. In the developed world, for example, companies adhere to a range of national and supranational regulatory regimes regarding employment, making generalised policies in companies transcending borders difficult to standardise. Hence, corporations tend not to consider employee issues as a major component of social responsibility, leaving little scope for worker-related CSR initiatives beyond national-level legal compliance. In developing countries, however, the levels of regulation and the enforcement mechanisms that protect workers' rights are considerably lower than those in developed countries (Crane & Matten, 2010). In such settings, employees tend to be viewed at best in a paternalist vein and at worst in a unitarist vein (Fox, 1966) wherein labour-as-cost is the primary orientation for human resource policy.
The concept of accountability to the workforce is rather complex for TNCs. Traditionally, TNCs have been attracted to low-cost workforces and relaxed labour standards in developing countries (Dicken, 2007). However, more recently greater attention has been paid to global value chains (Gereffi & Lee, 2012), a perspective which considers the international workforce in a more nuanced way than some of the more unilateral notions of the international division of labour. Edwards and Kuruvilla (2005) focus on the internal and external power dynamics within global value chains and show how their interdependencies can vary between TNCs, resulting in different implications for the workforce in each case.
TNCs that organise, manage and govern the global value chains have a significant impact on the gains achieved by suppliers and hence workers (Gereffi, Humphrey, Kaplinsky, & Sturgeon, 2001). However, the shareholder-value orientation of (in particular Anglo-Saxon) TNCs which dominates their governance structures (Ezzamel, Willmott, & Worthington, 2008;McSweeney, 2009) is regarded as a major impediment. Sikka (2011) argues that while civil and political rights are manifestly the imperative by-products of economic growth, corporations view economic growth in terms of financial and contractual obligations whereby social, cultural and political rights (which provide a conducive setting for human rights) are ignored and excluded. More broadly, Sikka (2010) alludes to how, under neoliberalism, nation statesparticularly in developing countriesare incentivised to compete to attract capital (from TNCs) by making inducements and concessions (at the expense of, for example, welfare of the workforce). Based on this ideology, contemporary accounting practices are designed to promote shareholder supremacy by reducing costs such as wages and other labour related expenses (Sikka, 2015). Thus, it should not be surprising that, in national settings with an emphasis on shareholders, TNCs can treat reporting on their employees as a tool to serve the interest of their shareholders rather than those of their workforce.
Considering the complex nature of global value chains, one of the challenges is to introduce some form of effective enforcement mechanism to help mitigate labour rights abuses below those directly employed in TNCs (Stephen, 2002). While the dominant legal structure and governance mechanisms are shareholder-oriented, particularly in the Anglo-Saxon world, respecting and protecting labour rights could still become an integral part of decision making (Muchlinski, 2012;Scherer & Palazzo, 2011). In such settings, governments' endeavours to promote CSR reporting often come to no avail as there is an absence of the necessary institutional infrastructures in place to ensure their effective implementation and functioning (e.g., Lauwo, Otusanya, & Bakre, 2016).
Due to the increasing pressures for global convergence of CSR standards and the manner in which TNCs are expected to illustrate their involvement in CSR (Davis, Whitman, & Zald, 2010, p. 39), many TNCs present an account of their social and environmental performance (Fortanier, Kolk, & Pinkse, 2011) by reporting large volumes of social and environmental information in an attempt to illustrate their accountability to their wider stakeholders. In doing so, they are also expected to identify and understand which of their economic impacts matter, for example, to the workforce and how they can define and manage their responsibilities 2 and subsequently engage with the relevant stakeholders to enhance their accountability (Campbell, 2005). However, TNCs are often not actively engaged with local communities and CSR practices are often not adopted by subsidiaries or suppliers in developing countries, contrary to rising expectations (Christmann & Taylor, 2001). All this highlights the complexity of value chains where there can be discrepancies between what TNCs may intend in relation to their workforce and what is achievable in reality. In occasions when '… there is a conflict about the objectives of social activity … accounting information has an ideological function in that it is used to legitimize particular activities or rationalize past [corporate] behaviour.' (Cooper & Sherer, 1984, p. 223). Under such circumstances, TNCs report information in an attempt to illustrate their transparency and hence elevate their accountability.

Theoretical framework: accountability or legitimacy?
Companies claiming to treat their workforce ethically (Gray, Adams, & Owen, 1996) would be expected to provide an account of their actions by reporting on workforce-related conduct and procedures. However, in the absence of mandatory reporting requirements, social and environmental reporting may not necessarily lead to enhanced transparency. Instead, such voluntary reporting can be used by TNCs as a means of depicting the appearance of being transparent with the purpose of enhancing legitimacy. In effect, such reporting can act '… as a corporate veil that simultaneously [provides] a new face to the outside world while protecting the inner workings of the organisation from external view.' (Hopwood, 2009, p. 437).
Because transparency can never be achieved, acknowledging the limits within which accountability 3 can be achieved is the first step towards any meaningful reporting. In addition, the recognition of, and engagement with, legitimate stakeholders is an essential step towards meaningful social reporting (Cooper & Owen, 2007) and of discharging accountability. Roberts (2009) regards this as 'intelligent' accountability, calling for an active enquiry and engagement with stakeholders (i.e., listening, asking questions and observing evidence as how corporate claims are put into action) over time. This would ultimately engender trust between TNCs and their stakeholders, and hence the discharge of accountability.
In essence, TNCs' relationship with their stakeholders is key to securing legitimacy and continuing survival. Under legitimacy theory, a legitimate organisation constantly ensures the congruence of its value system with that of the wider society (Lindblom, 1993) and gains its legitimacy only when its relevant public assesses its value system and perceives it as acceptable (Patten, 1992). It is only then that a company is given its licence to operate and gains access to its vital resources such as the labour market (O'Donovan, 2002). The relationships between organisations and their stakeholders are complex and are shaped by implicit and explicit terms of the social contract, asserting the expectations that stakeholders have of the TNCs (e.g., Deegan & Unerman, 2011;Gray et al., 1996).
For TNCs whose value chains operate in different geographical locationsoften developing countries with varying social norms and valuesthe nature of relationships with their extended workforce varies. It is not unknown for there to be discrepancies between what TNCs do and what is expected of them at any point in time (a legitimacy gap) or for them to report information as a response to this discrepancy (O'Donovan, 2002) to either maintain their legitimacy or to mitigate pressures on them (Milne & Patten, 2002). In situations when there is a threat to TNCs' legitimacy (because they are not discharging their accountability, O'Dwyer, 2002), reporting is used as a legitimation tool rather than a mechanism to discharge accountability (Gray & Bebbington, 2000).
TNCs can report on their workforce either to discharge accountability to their stakeholders and gain legitimacy (if reporting is intended to discharge accountabilityi.e., intelligent accountability as argued by Roberts, 2009) or to maintain/repair their legitimacy. In this paper, we consider that in the former instances, companies are expected to closely adhere to the GRI reporting guidelines as an endeavour to discharge their accountability to their workforce if they have invested in techniques and procedures that could result in a high level of conformance with the GRI guidelines. For the latter group, companies can declare compliance with the guidelines and fail (or decide not) to report detailed information (as per guidelines) to gain or maintain their legitimacy. For this group of TNCs, 'declaring' compliance with the guidelines is more likely to be an attempt to gain or maintain legitimacy of their human resource management policies that is in line with the international guidelines. In doing so, companies' reporting can be in line 2 At the same time, it is important to note that CSR is perceived differently in each individual country and depending on the local cultures and customs, the expectations of what CSR is going to deliver can vary considerably from one country to another (Katz, Swanson, & Nelson, 2001). 3 The conscious acknowledgement of the impossibility to achieve full transparency could be the basis for a very different sense of responsibility that in turn might allow for a very different enactment of accountability (Roberts, 2009, p. 958).
with the second, third or fourth strategies (as adopted from Lindblom, 1993) discussed below: 1. to change the standard for the treatment of their workforce in conformance with the GRI guidelines, and subsequently inform their international audience of the changes; 2. not to change the way they treat their workforce, but to demonstrate and justify the appropriateness of how they treat their workforce and are therefore in compliance with the GRI guidelines; 3. endeavour to alter the perceptions of their relevant public of the key issues by associating themselves with symbols that have a high legitimacy status (e.g., claiming adherence to the GRI reporting guidelines) while not reporting as per GRI specifications; 4. endeavour to alter societal expectations by aligning them with the way they report on the treatment of their workforce and claim conformance with the GRI guidelines.
The existing literature predominantly discusses and presents evidence in support of companies reporting social and environmental information to gain, maintain or repair their legitimacy (e.g., Campbell, Craven, & Shrives, 2003;Deegan, Rankin, & Tobin, 2002;Kilian & Hennigs, 2014;Liesen, Hoepner, Patten, & Figge, 2015;Milne & Patten, 2002;Nègre, Verdier, Cho, & Patten, 2017;Tilling & Tilt, 2010;Patten & Zhao, 2014). A smaller number of studies have examined reporting on workforce issues and have supported legitimacy theory. For example, Kamal and Deegan (2013) examined a sample of Australian textile and garment companies operating in Bangladesh and found that, irrespective of international concerns about working conditions and safety at workplace, companies' related disclosures fell short of the expected international standards and most disclosures were aimed at maintaining legitimacy and meeting the expectations of the local communities, resulting in limited accountability and transparency. In another study, Williams and Adams (2013) found that a bank's reporting on its employees between 1980 and 1995 was more influenced by legitimacy considerations in a broader social-organisational context than by the motivation to enhance transparency and accountability to/for their workforce. Another relevant study is that of Boiral (2013) who closely examined the quality of information reported by 23 companies in energy and mining sectors using the GRI framework and found evidence in support of legitimacy theory. Along the same lines but in more detail, Diouf and Boiral (2017) sought the perception of socially responsible investment practitioners in Canada. They found that while GRI is viewed favourably as a work-in-progress, companies are perceived to select, adapt or modify GRI indicators to project a favourable image among their stakeholders (while avoiding negative aspects). Lack of comparability, timeliness and reliability of information were the main weaknesses highlighted.

Labour standards and the Global Reporting Initiative
The International Labour Organisation (ILO) labour standards provide a logical benchmark to assess labour practices around the globe. Core labour standards were adapted in 1998 with a view to establish a balance between procedural justice and social rights (Alston, 2005). This is the 'decent work' agenda (ILO, 2014). The standards also require a process and some have questioned the shift from ILO standards constituting a mandatory to that of a voluntary code and the development of the notion of 'soft regulation' as a means of governance on labour standards (Kuruvilla & Verma, 2006;Sobczak, 2006;Wells, 2007). Voluntary initiatives have been criticized for numerous reasons: (a) they are dependent on the 'spotlight of brand activism' which is both 'roving and random', (b) they deflect from the activities of sub-contractors working directly for TNCs and where the worst abuses occur, and (c) corporate codes are 'only as good as those that wrote them' (Royle, 2010, p. 263). However, while these criticisms may be true of TNCs complying to their own voluntary codes, they may not necessarily hold true, on all three points, where they are subject to external standards. It is argued that external standards can be of some assistance in reducing labour rights abuses down the value chain (Nadvi, 2008).
It is with these points in mind that the GRI may address some of these gaps. GRI was established in the late 1990s to introduce a framework that could serve the interests of investors and civil society organisations in benchmarking companies' activities. The first set of guidelines was published in 2000. 4 The GRI itself is an independent not-for-profit organisation based in Amsterdam, governed by a 16 member board, advised by a 100 member 'stakeholder council, both drawn from and to represent the interests of business (via appointments from TNCs), labour (through appointments from global union federations) and civil society (via appointments from non-governmental organisations).
GRI has been the most popular framework worldwide 5 and established a momentum of legitimacy among key stakeholders in its early stages (Brown, de Jong, & Levy, 2009). The guidelines are devised based on consultations with a range of stakeholder groups from all over the world so that a broad range of issues are reflected in the guidelines, making them applicable to a wide range of organisations including small enterprises, NGOs, 6 large TNCs and public sector organisations (GRI, 2011, p. 2). The main intention of the GRI is to harmonise sustainability reporting practices so that, as with financial reporting practices, consistency and hence comparability of the reports can be achieved for companies around the globe (Willis, 2003). Even though the growth and maturity of the GRI has resulted in the overall improvement of sustainability reports, there are doubts about its favourable impact on the 4 The guidelines go through an intensive revision process by a group of stakeholders in every couple of years to implement the latest changes. After the full set was published in 2000, the second generation of guidelines, G2, were published in 2002, GRI G3 in 2006, GRI G3.1 in 2011 and G4 in 2013. The differences between G3 and its successors (G3.1 and G4) are minor. For example, the principle of materiality remains the same in G4 as it was in G3. G4 makes more explicit links between materiality and the management and performance information organisations should report on. 5 Adopted by 78% of the largest 250 global companies in 2011; 80% in 2013, 74% in 2015 (KPMG, 2013(KPMG, , 2015. 6 Non-Government Organisations. comparability of reports, of the overall transparency of companies (Toppinen & Korhonen-Kurki, 2013) and of its ability to encourage the inclusion of sustainability into decision making processes (Adams, 2015). Despite the guidelines representing a means of enhancing corporate accountability (e.g., Bowen, 2000;Willis, 2003), evidence suggests that the momentum for adoption has been thwartedpossibly due to the perception that stakeholder needs are not being sufficiently met (Levy, Brown, & De Jong, 2010). Further criticisms laid against the GRI guidelines highlight that many organisations adopting the guidelines do not necessarily witness improved performance and can often manipulate the guidelines just to appear to be more transparent (Moneva, Archel, & Correa, 2006) while in truth it is business as usual for most companies (Milne & Gray, 2013).
In this paper, we analyse reports written to GRI G3 reporting guidelines. The guidelines emphasise the attention TNCs should pay to the materiality 7 of information as well as identifying and illustrating engagement with their stakeholder when reporting information. In line with the KPMG's (2013, p. 8) definition of materiality, we take the view that companies provide 'material' information when they: (1) identify the key issues that have the greatest potential impacts on their workforce (whether internal or external) and (2) make clear the process they have used to assess materiality; this clarity should incorporate the involvement of their workforce as well as how they have responded and provided feedback to their workforce. In this paper, we have focused on the GRI workforce related guidelines that outline specifications on how material information on different aspects of workforce (e.g., identification of their key issues and active engagement with the workforce) can be reported.
Our choice of the GRI guidelines as a benchmark for employee-related disclosure is based on a number of rationales. First, GRI provides the most comprehensive set of reporting guidelines available for the comparison of reporting by companies (Sutantoputra, 2009;Vigneau, Humphreys, & Moon, 2015). Second, a wide range of stakeholder groups from different continents took part in the development of the GRI guidelines and its performance indicators. The international steering committee that oversees the activities of the GRI has representatives from different stakeholder groups, corporations and the UN (Reynolds & Yuthas, 2008, p. 53). Third, the requirement to cross-reference and to index and the existence of specific guidelines for compliance on each category, reduces the need to develop a coding framewhich would inevitably be subject to normative judgement or interpretation. 8 The guidelines provide a systematic approach to the preparation and presentation of information with an emphasis on stakeholder engagement. 9 Hence, companies that adopt the guidelines need to have certain systems or procedures in place before they can fully adhere to them. In this paper, the guidelines enable us to examine information characteristics (e.g., completeness, and comparability) of information.
Using the GRI G3 guidelines, we examine how companies report on their workforce by concentrating on: (1) their labour practices and (2) the employment-related aspects of the human rights category. While the former covers worker issues internal to the organisation, the latter covers issues relating to the workforce outside the organisation and often in their value chains in developing countries (Ehnert, Parsa, Roper, Wagner, & Muller-Camen, 2016). In order to examine the two sets of workforce-related issues, we use the only two GRI categories that relate to the workforce. 10 : (i) 'Labor Practices & Decent Work' (LAs) and aspects of (ii) 'Human Rights' (HRs). The LA indicators are specifically aligned to ILO labour standards and the ILO decent work agenda 11 There are nine core LA indicators and five 'additional' ones. The HR indicators are based on generally accepted norms and principles such as the UN Declaration of Human Rights and the core conventions of the ILO (GRI, 2013). Of the nine HR indicators, six are core and three are 'additional.' In this paper, we concentrate on GRI core indicators (and not on additional indicators) as they are designed to cover all material information that can relate to the workforce (GRI, 2011, p. 7). 12 For each performance indicator under the LA and HR categories, the GRI lists a number of necessary compilations, allowing us to examine the completeness and hence comparability of information.

Labour practices and decent work (LA)
The extant labour and decent work literature tends to be based on studies focusing on single-country cases rather than looking into global trends. For example, there have been a limited number of equality-related studies over the years, among them the work of Adams and Harte (1998) on the changing perception of women in UK banking and retail sectors; Adams, Hill and Roberts (1995) on corporate equal opportunities (non) disclosure; Adams and Harte (2000) on the potential for social accounting on making discrimination visible; and Grosser and Moon (2005) on the workplace reporting of workplace gender mainstreaming. Earlier studies show that disclosures on employment issues were most likely to come from European or South African companies (Roberts, 1990 13 ). In another study, Adams et al. (1995) 14 note that although there were substantial disclosures on employee-related issues, mostly on 7 According to the GRI guidelines (GRI, 2011), the '… materiality is commonly thought of as a threshold for influencing the economic decisions of those using an organisation's financial statements, investors in particular' (p. 8).
8 As Gray and Gray (2011) point out, for example, an endeavour to define Human Rights or putting the issue of human rights (as per defined by individual rights) would be complicated. 9 In doing so, the GRI has adopted many of the conceptual reporting principles of the Financial Accounting Concepts (FASB, 1996) and provides guidance to prepare information with similar qualitative characteristics as those for the financial information (Reynolds & Yuthas, 2008). 10 There are six categories of indicators: (1) Economic (ECs), (2) Environmental (ENs), (3) Labour Practices & Decent Work (LAs), (4) Human Rights (HRs), (5) Society (SOs) and (6) Product Responsibility (PRs). 11 According to ILO: 'Productive employment and decent work are key elements to achieving a fair globalization and poverty reduction. The ILO has developed an agenda for the community of work looking at job creation, rights at work, social protection and social dialogue, with gender equality as a crosscutting objective'.
(http://www.ilo.org/global/topics/decent-work/lang-en/index.htm, (Accessed 18 March 2015)). 12 Additional information may be material to some of the companies and not to all. 13 Examined the reports of 200 multinational companies. 14 Examined social disclosures of 150 companies from six European countries. descriptive aspects (e.g., policies, description of activities, quantitative information such as number of employees and payroll costs), very little information had been reported on more sensitive issues such as those required by the GRI reporting guidelines. A significant lack of transparency and accountability to employees is prevalent and companies should take strides towards genuine accountability (intelligent accountability) to their employees by identifying and actively engaging with their employees (i.e., considering and responding to their employees' needs and concerns in their decisions, policies and practices) (Williams & Adams, 2013). While the whole nature of the relationship between companies and their employees tends to be influenced by a range of social, economic and political aspects (Williams & Adams, 2013), in a global context, the nature of such relationships can vary substantially from one region to another and hence so can the information that companies report on their employees. If TNCs report information in an endeavour to discharge their accountability to their internal workforce, the identification of the possible key stakeholders in relation to the employees (e.g., employee representatives or other officials in their branches and subsidiaries based abroad) and engaging with them actively to maintain a certain level of responsiveness to their issues will be an inevitable feature of the reporting process. This is essential if companies are going to present material information that would allow comparability of their relationships in different regions.

Human rights (HR)
With TNCs being increasingly confronted with the question of labour rights abuses in their value chainsnot covered under the LA codes, abovethe GRI category of 'human rights' contains some significant sub-categories that directly relate to labour rights affecting workers in supplier organisations and those not directly employed by the TNC. There have been significant developments in human rights policies and their implementation (McPhail & Ferguson, 2016). While the literature on human rights reporting is even more limited than that on employee reporting, there have been calls in recent years for more scholarly debate about human rights. For example, Frankental (2011) calls for research on different aspects of human rights including whether companies illustrate that they have '… proper management and reporting systems in place to anticipate and address [human rights issues]' (p. 764). Cooper et al. (2011) focus on the right to work in a safe environment and present evidence that, in spite of the claims made by the Scottish government to ensure workplace safety, in reality the social and political system in place failed to do so. More recently and in a broader context, Posner (2016) highlights the emerging business and human rights agenda and discusses how a number of key areas, such as value chains and labour rights, must be focused on for further development.
While TNCs exert economic veto power over developing countries, they are at the same time under pressure from international activist groups to abide by human rights norms (Stephen, 2002). One of the key features of the UN Framework on Human Rights and Business is the 'moral' responsibility of corporations to respect human rights irrespective of whether doing so is required by (local) law. Arguably, if TNCs regard human rights as their moral responsibility, it is likely that the distinction between the responsibility to respect and the duty to protect will begin to break down (Cragg, 2012, p. 32). Once the distinction between the two is blurred, TNCs are more likelyin the absence of any international law and law enforcement mechanismto take human rights issues more seriously. 15 At this stage, information disclosure can play a significant role in promoting human rights and discharging corporate accountability to employees and the workforce in companies' value chains (Gallhofer, Haslam, & van der Walt, 2011). Islam and Jain (2013) focus on major Australian garment and retail companies and found a low level of human rights disclosures. They explain that soft disclosure regulations and multiple failed attempts to introduce hard regulations provide these companies with grounds to avoid improvements in their human rights reporting. In a case study of a Tanzanian gold mine, Lauwo and Otusanya (2014) similarly conclude that human rights obligations should become mandatory rather than being regarded as a duty of care.
While the concept of human rights can be viewed from different perspectives (Muchlinski, 2012), the GRI reporting guidelines focuses on whether or not human rights issues: '…are considered in investment and supplier/contractor selection practices. …[and whether] … the indicators cover employee and security forces training on human rights as well as non-discrimination, freedom of association, child labor, indigenous rights, and forced and compulsory labour' 16 (GRI, 2011, p. 32). As Frankental (2011, p. 763) notes, the GRI provides '… a widely recognised set of human rights benchmarks for companies, addressing wide-ranging issues …' The GRI guidelines on human rights reporting requires companies to have systems and procedures in place to screen human rights breaches and identify the course of action they need to take to rectify troublesome situations. Similar to the LA categories, we view the human rights issues from a moral perspective whereby TNCs should acknowledge their responsibility but should also be able to identify the workforce in their value chains and illustrate their endeavours to actively engage with the parties whose involvements are essential to ensuring that breaches are avoided. This would require a systematic approach taking into account geographical variation and the context in which global value chains operate. Similar arguments are presented by Gallhofer et al. (2011) who draw attention to the importance of governance in promoting and protecting human rights and emphasise the fact that accounting (i.e., information disclosure) can have a significant role in this process. More recently, McPhail and Ferguson (2016) consider the momentum in regulatory progress as the way forward for human rights reporting. 15 Gallhofer et al. (2011) point out that states and corporations have different responsibilities in promoting and safeguarding human rights (p. 769). There is no international law for human rights and it is only up-to the national law of the countries to have and then to enforce human right laws. One disadvantage of this is the multinational companies can always move away from a country with strict human rights law to countries with less strict laws (Gallhofer et al., 2011). 16 The GRI

Sample selection and data collection
For the purpose of this paper, a version of content analysis combined with more detailed text analysis of corporate sustainability reports is used to measure the extent and nature of information reported on labour and human rights-related issues. Content analysis has been referred to as a systematic and replicable technique that allows the examination of text using explicit coding rules (Krippendorff, 2012). However, rather than attempting to identify word patterns distributed within reports, the adapted form of content analysis used here is based on the analysis of the index tables used on each report. As discussed earlier, the coding rules used in this paper have been developed based on the GRI guidelines for the 'Labour and Decent Work' (LA) and 'Human Rights' (HR) performance indicators. This analysis is supplemented by a second stage textual thematic analysis whereby the text in the relevant cross-referenced sections of the reports are read and compared to the GRI guidelines to which they claim to be adhering.
The sample of TNCs selected was taken from the Forbes 250 largest companies list, ranked by size of assets, regardless of country of origin or sector. Between June and October 2011, we collected the latest sustainability report from every company listed on the Forbes 250 index. Most companies provided stand-alone reports, but for some companies the sustainability information was included (generally as a chapter or two) in the financial reports (so-called integrated reports) and in other cases the report was web-based only. In our final sample, we included only those companies that had formally adopted GRI as a benchmark and had also included a GRI index table in their reports. Of the original Forbes 250, 215 had some form of sustainability report, 166 had claimed to have adopted the GRI and of these, 131 had provided a GRI index table. It is the 131 GRI-indexed sustainability reports that form the basis of our analysis.
In the GRI index table, companies are required to note the reporting status for each GRI indicator and indicate where, in the full report, the relevant information is provided. For the first step of our categorisation of company performance against GRI, we analysed whether companies claimed to have reported either 'full', 'partial' or 'no' disclosure against each GRI LA and HR sub-category. For the second stage, we compared the information presented in the full report, in relation to the status or reporting categorised in the index.
The coding scheme used for examining the information was based on the GRI guidelines for LA and HR indicators. Fig. 1 illustrates different scenarios where a company's disclosure can be recoded as 'non-', 'partial' or 'full' disclosure. Having examined the actual information that companies had reported against the information they claimed to have reported (stage 2), a number of scenarios could emerge where companies were considered to have had claimed their disclosures incorrectly. Scenarios S2, S4 and S5 indicate over-claiming when a company could have, for example, claimed to have full disclosure for a certain indicator (in its GRI index table) but having examined the item, it became evident that the company had either over-claimed and reported no information (i.e., S2) or only partially disclosed the item (i.e., S4).
If companies had claimed disclosure but did not make explicit whether this was 'full' or 'partial' disclosure, we considered that they had claimed full disclosure. We also did not check whether companies had not claimed but reported the information somewhere else in their reports (or on their webpages) (S2 and S3). We felt that was outside the scope of the paper. We did check for S7 and found only three companies had 'full' disclosure while claiming 'partial' disclosure. In Fig. 1, all the possibilities that we examined are marked by solid lines (S4-S9).

Scores
We developed a checklist where we included all the compilations for each LA and HR indicators, thereby facilitating the close examination of non-disclosure, partial disclosure or full disclosure (for more details see Tables 2 and 4). When examining the information, we considered an indicator to be fully disclosed only if the company had reported information on all its compilations. If information had been provided on some of the compilations (i.e., at least one of the compilations), this was counted as partial disclosure. Different performance indicators have different numbers of compilations ranging from 10 for LA7 to only one for LA4, LA5, LA10 and LA14. All the HR indicators have two compilations.

Findings
In this section, we examine the extent of labour and human rights related issues reported by our selected companies in compliance with the GRI reporting guidelines. This provides us with insight into the actual status of compliance with the guidelines and whether full disclosure (as claimed by many companies) is indeed full disclosure.
First we examined the information which companies reported on each LA and HR indicators, classifying the tested information into three categories of: 'no disclosure' (no relevant information disclosed), 'partial disclosure' (only some of the information relevant to the indicator disclosed) and 'full disclosure' (all information required by GRI disclosed) (see Table 1). In the pursuit of additional insight, we then examined the core indicators by looking at the 'compilations' for each indicator (see Table 2). We then examined whether reports contained the information companies claimed to have reported in their GRI index tables (using the scenarios presented in Fig. 1).
6.1. Tested disclosures 6.1.1. Labour indicators Table 1 presents the results of our examination of the information companies had reported on each LA and HR indicator. While a number of labour indicators allow the identification of key issues relating to the internal workforce (i.e., LA1, LA2, LA7, LA13, LA14), there are indicators focusing on engagement with their workforce in the form of dialogue and the support they provide (i.e., LA4, LA5, LA8 and LA10). 17 For the full disclosure column, the three lowest scores were for the turnover by age/gender/region (LA2), composition of governance bodies (LA13) and rates of injury/absenteeism by region (LA7) indicators. Of the three indicators, LA2 had the highest non-disclosure.
LA2 with 6 compilations requires a breakdown of information into age group, gender and region (see Table 2). A close examination of the compilations shows that more companies (15% or more) had reported information by region and gender rather than by age group (about 9% or less). For LA13, while most companies had reported by gender, a much smaller percentage of companies had reported by minority group. In the case of LA7, many detailed disclosures are outlined in the guidelines. More specifically, the guidelines require companies to report on injury rate, occupational disease rate, lost day rate, absentee rate, and absolute number of fatalities, each factor broken down for region and contractors within region. Only 2 of the compilations had been reported by 11% and 15% of companies while 5 were reported by less than 5%, suggesting that most companies failed to provide the extent of detail outlined under this indicator. The highest overall disclosures (i.e. partial plus full disclosure) were observed for LA8 (82.5%), LA7 (at 75.6%) and LA13 (71%). For all three indicators, most had been reported 'partially', rather than 'fully'. Once again, it emerged that more specific information and detailed breakdown was generally lacking.
The least well reported indicators were: salary ratio: men/women (LA14 with 87% non-disclosure), minimum notice periods (LA5 with 80.2% non-disclosure) and turnover by age, gender and region (LA2 with 71.8% non-disclosure). The inclusion of category LA14 was relatively recent to GRI, which could, to some extent, explain the unavailability of the information in many reports. At the same time, it has only one compilation (see Table 2)meaning that companies had only to decide whether to disclose or not disclose this one ratio. LA5 also has only one compilation and the same logic applies. In the case of LA2, a small number of companies had provided a breakdown of information by gender and a smaller number by age.
Overall, the most frequently disclosed items were quite generic. When more specific issues were to be identified and reported on, the percentage of companies reporting dropped considerably. For instance, companies were weak in reporting on their workforce in different regions. As an example, the compilations for LA7 specify that companies should provide a breakdown of rates of injuries and absenteeism by regions. Our findings show that hardly any company had reported on the lost day rate, occupational disease rate, or absentee rate for independent contractors in different regions. Disclosures were also low when companies had to present statistics for different categories such as minority groups, gender and age categories.

Human rights indicators
Of the six human rights indicators, investment and procurement practices (HR1) and suppliers screened for human rights (HR2) were the least disclosed indicators (with 92% non-disclosure for HR1 and 81% for HR2see Table 3). These two indicators also had the lowest full (and overall) disclosures (HR1 at 3.1% for full disclosure and 7.7% for overall disclosure and HR2 at 3.8% and 19.1%, respectively).
The compilations for HR1 (see Table 4) reveal that slightly more companies (8%) had reported on the number of significant investment agreements that included human rights clauses than on the significant investments agreements that had undergone human rights

LA5 Minimum notice periods
The minimum number of weeks' notice typically provided to employees and their elected representatives prior to the implementation of significant operational changes that could substantially affect them.

15%
LA7 Rates of injury/absenteeism by region Injury rate (IR) for total workforce by region 8% IR for contractors by regions 5% Occupational diseases rate (ODR) for total workforce by region 6% ODR for contractors by region 3% Lost day rate (LDR) for total workforce by region 6% (LDR) for independent contractors by region 4% Absentee rate (AR) for total workforce by region. 5% AR for independent contractors by region 1% Absolute (Abs) number of fatalities for total workforce by region 15% Abs number of fatalities for contractors by region 11% LA8 Education/training workforce and families regarding serious illnesses/diseases Education/training program(s) to assist workforce members, their families or community members 31% Counselling program(s) to assist workforce members, their families or community members 16% Prevention/Risk control program(s) to assist work force members, their families or community members 24% Treatment program(s) to assist workforce members, their families or community members 21% screening (4%). In the case of HR2, more companies (9%) reported on the actions they took as a result of their human rights screening than outlined their human rights screening criteria on their suppliers and contractors (about 5%). The inclusion of human rights clauses or screening of contract would have required systematic and concerted efforts to actively engage with suppliers so that the external workforce likely to be affected by the TNC's operations could be protected from human rights breaches. The most reported indicators were child labour (HR6 at 26 + 31.3 = 57%see Table 3), forced labour (HR7 at 26 + 29 = 55%) and risk to freedom of association (HR5 at 24.4 + 26.7 = 51.1%). Looking at their compilations, a much smaller proportion of TNCs had reported that they had procedures to identify those operations subject to significant risk of either child labour, forced labour or freedom of association (31%, 30% and 27%, respectively). While companies were paying more attention to child labour, forced labour and risks to freedom of association (HR6, HR7 and HR5), less attention was being paid to broader human rights issues such as investment and procurement practices and supplier screening (HR1, HR2). Overall, TNCs hesitance to report on procedures used to identify breaches of human rights issues (e.g., the first compilation for: HR1, HR2, HR5-HR7) is indicative of how little attention they are paying to systematically approaching human rights aspects of their external workforces.

Over-claimed disclosures
The close examination of reports revealed that, overall, companies were exaggerating, in their index tables, the level of information they were claiming to be disclosing in the reports. In the following sub-sections, we analyse the extent of reporting, and over-claiming, for the Labour and Human Rights indicators.

Labour indicators
The results (see Table 5) suggest that companies over-claimed their disclosures for all LA categories in their index tables. The highest over-claiming was for equality (LA13) at 73%, education and training regarding serious illnesses and diseases (LA8 at 63%), employment (LA1 at 64%) and rates of injury/absenteeism by region (LA7 at 56%). Looking at the compilations for each of the three indicators (Table 2), we see that the scores for most compilations are very low.  HR5 Risks to freedom of association Operations identified in which the right to exercise freedom or association and collective bargaining may be at significant risk.

27%
Measures taken by the organisation to support these rights. 50%

HR6 Risk of child labour
Operations identified as having significant risk for incidents of child labour 31% Measures taken to contribute to the elimination of child labour 58%

HR7 Risk of forced labour
Operations identified as having significant risk for incidents of compulsory labour 30% Measures taken to contribute to the elimination of forced labour 54% Evidence for LA13 revealed that companies had reported on the gender composition (80% and 70%see Table 2) but not on their minorities (15%). Nonetheless, most of these companies declared themselves as full disclosures while they had reported only on some of the compilations. Companies' provision of general statements about their policies and procedures rather than the provision of specific statistics as outlined in the guidelines is another illustration of companies recognising the importance attached to diversity in their governance structures. Below is an instance where a company claimed full disclosure but included none of the relevant statistics to supports its claims: 'We realigned the governance of our global diversity activities …. As a result, Diversity Operating Committees led by the four regional CEOs and senior managers are now responsible for implementing our global strategy in line with regional requirements. We have also reviewed our global diversity strategy and are now drawing on the support of members of the Executive Board as well as the Board of Directors in this area. … Initial progress has already been achieved in Asia Pacific and Europe.' (Credit Suisse, 2008, p. 31) Another similar example from AEGON states: ' … that greater diversity within its workforce brings clear benefits to its business: By creating a wider, more diverse pool of talent. By improving the company's understanding of its customer base and broadening its appeal to different customer segments. By further strengthening AEGON's brand and reputation.' (AEGON, 2010, p. 60) Other examples include Rosenft and China Mobile, which provided unrelated statistics (e.g., staff training or the number of managers at different levels) but failed to include breakdowns by minorities or age group.
For LA8, full reporting requires companies to list assistance programmes offered to their employees and their families in the form of education, counselling, prevention/risk and treatment/health for serious illnesses and diseases. The guidelines do not require companies to report on (or indeed have) all of these in order to comply with 'full disclosure', but where such programmes are offered, it must be specified whether they are offered to workers, workers' families or communities. Some companies stated that the community's welfare is not seen as a priority for them as 'Education, training, counselling, prevention, and risk-control programmes to assist other people or community members regarding serious diseases are usually a governmental responsibility' (Adecco, 2008, p. 62). Depending on the location, this may be legitimate. It could also be an attempt to deflect attention from what is deemed to be unfavourable information. Another relevant example relates to the ICBC, a Chinese Bank which claimed to have fully disclosed information in adherence to LA8 guidelines while the actual examination of their report revealed full compliance with only 2 of the 8 compilations. They stated: 'The Bank has continued to step up investment in employee training, extended avenues of training, gradually establishing a competency-based training system that instils a working culture of career development among employees. The "Six Libraries" … were established to continue to enrich the training resources of the Bank, and new training methods … were introduced.' (ICBC, 2010, pp. 89-90) The bank reported no information about counselling programmes for the families of the workforce. Another example is from China Mobile, which claimed full disclosure but scored zero for all compilations. The company discussed issues related to LA8 but failed to provide any relevant statistics, stating that: 'In order to effectively guarantee employee health and safety, we have continuously strengthened our workplace health and safety management systems and policies by establishing health records for employees and providing regular medical examinations and health related information and training, and have achieved good results.' (China Mobile, 2009, p. 17) A slightly different point emerges under the guidelines set out for LA1. Here, companies are expected to report on four major aspects: total workforce, employment contract (whether permanent or fixed contract), employment type (whether full-time or parttime) and their workforce in different regions. While the majority of companies in our sample had reported on their total workforce, Rates of injury/absenteeism by region 56 LA8 Education/training workforce and families 63 LA10 Training hours per year, by employment category 51 LA13 Composition of governance bodies 73 LA14 Salary ratio: men/women 28 most companies failed to report on the employment contract, employment type or their workforce in different regions. The failure to disclose one of the four aspects meant that the majority of companies which had declared 'full disclosure' were reclassified as only 'partially disclosing'. For example, JP Morgan Chase had declared disclosure by stating that they employ '…approximately 240,000 men and women in more than 60 countries.' (JP Morgan Chase, 2010, p. 65), failing to provide a breakdown of the total workforce in each region and in terms of employment contract and employment type. Employment contract and employment type were the two categories least well reported by companies, which suggests a high degree of caution in disclosing statistics related to workers in precarious employment. While the over-claiming indicates that companies recognise the importance attached to the key aspects of employment contracts, the failure to provide all the statistics set out in the GRI guidelines could be due to the absence of a systematic approach to gathering data. Some of the compilations for LA1 were not reported by companies from certain countries. As shown in Table 2, the breakdown of information on regions, type of contracts, type of employment and supervised workforce was low. For example, PetroChina had claimed full adherence to LA1 by providing the following general statement with no relevant statistics: 'We strictly adhere to the Labor Law of the People's Republic of China, Labor Contract Law of the People's Republic of China and other relevant regulations of jurisdictions in which our shares are listed and we rigorously fulfil international conventions endorsed by the Chinese government. We have established a sound employment management system covering labor contracts, remunerations, insurance and benefits, performance evaluation, rewards and penalties and professional training.' (PetroChina, 2008, p. 35) (Emphasis added) Interestingly, this company is explicitly declaring its full compliance with those international conventions that are endorsed by the Chinese government. In other words, their non-disclosures could be those which are not endorsed by the government and hence not reported. Nonetheless, the company decided to claim full disclosure in their GRI index table, most likely to maintain or increase their legitimacy, certainly not as a sincere attempt to identify the key issues as set out in the guidelines. This is consistent with the findings of Diouf and Boiral (2017) that companies select, adapt or modify GRI indicators to cater a favourable image among their stakeholders.
The indicators which attracted the best levels of accurate reporting (the lowest levels of over-claiming) were percentage of workforce covered by collective bargaining (LA4 at 26%) and salary ratio: men/women (LA14 at 28%). Looking at Table 2, both LA4 and LA14 have only one compilation. For LA4, the overall disclosure (5.3 + 38.2 = 43.5%,see Table 1) was not particularly high. For LA14, see above, the indicator is relatively recent and hence the possible unavailability of the data.
In summary, we observed that in many cases of over-claiming, detailed information was either not deemed 'relevant' to the company or not required by national legislation. For example, many companies did not provide regional information for their minorities. In some countries, regional data or data on minorities are considered politically sensitive to report. In such cases, companies provided no relevant explanation yet still claimed 'full' disclosure. While some compilations may not necessarily be of interest to their internal workforce, they may be so for their global investors, customers and suppliers. What is interesting is that many companies have decided to claim 'full' disclosure irrespective of the alleged 'irrelevance', to them, of some of the compilations. In the case of non-relevance (and hence no materiality to their internal workforce), companies are expected to explain non-disclosure if they report information with the intention of raising their transparency rather than attaining and maintaining legitimacy.

Human rights indicators
The indicators for investment and procurement practices (HR1) and suppliers screened for human rights (HR2) had the highest overclaiming at 54% and 60% (see Table 6). For these two indicators, most companies provided general statements rather than the specific statistics required by the GRI guidelines and only 9% or less had complied with the 4 compilations (see Table 4), indicating that companies hardly reported any human rights issues relating to their workforce in their value chains. Nonetheless, they clearly realised the importance attached to these issues and may have considered non-reporting as a threat to their legitimacy or to be regarded as less legitimate.
For HR1, the following two examples are common for the majority of companies that claimed full disclosure but had in effect reported none: 'ABB maintains and regularly reviews a list of sensitive countries where it has, or considers engaging in, business operations. Human rights, as well as legal, financial and security criteria, are included in risk assessments, and are among the factors in deciding whether ABB does business in a particular country. Based partly or wholly on human rights considerations, ABB has not Table 6 Over-claimed Disclosures for Human Rights Indicators.

Human Rights Indicators
Over-claimed (Disclosed less than claimed) HR1 Investment and procurement practices 54 HR2 Suppliers screened for human rights 60 HR4 Incidents of discrimination 17 HR5 Risks to freedom of association 30 HR6 Risk of child labour 34 HR7 Risk of forced labour 35 taken any business with Myanmar or North Korea for several years. ABB completed its withdrawal from Sudan in June 2009, having taken no new business in the country since January 2007.' (ABB, 2010, p. 25) 'Bayer pays special attention to respect for human rights. Our Supplier Code of Conduct is based on the principles of the Global Compact and also takes up the sustainability principles and our Human Rights Position.' (Bayer, 2010, p. 33) And for HR2: 'We also advocate the observation of employee rights even outside of our own company. Our procurement principles, which are applicable worldwide, require that our suppliers strictly comply with the respective national legislation and recognize equality of opportunity as well as the right to collective bargaining and adequate minimum wages and benefits. If any of our business partners fundamentally violates one of those principles, we terminate the business relationship with them.' (Deutsche Bank, 2010, p. 68) 'We will work with all our suppliers to ensure they meet the ten principles of the Global Compact to guarantee an ethical supply chain. We want to ensure that our supply chain is "ethical" … we want our suppliers and contractors to work towards the same high levels of human rights, labour standards and environmental management that we aspire to. … we are a signatory to the UN Global Compact, and we are committed to promoting responsible business practice within our sphere of influence, so we have begun with an internal improvement project to re-assess our suppliers, and to revise our contracts. Any new supplier for EDF Energy will have to be GC compliant.' (EDF, 2008, p. 26) In all four examples, companies provide general information, occasionally interesting relevant information, but without addressing the specific information items (as per GRI guidelines). Bayer and EDF, for example, made reference to the adoption of the Global Compact by their suppliers, ignoring the GRI guidelines on screening suppliers for breaches of human rights. The two companies are typical examples of companies reporting relevant and interesting information to either deviate attention from key issues (Lindblom's second strategy) or change their audience's perception of what the key issues are by associating themselves to symbols (Global Compact, GRI) (Lindblom's third or fourth strategy). In contrast, Deutsche Bank alludes to national legislation on human rights, stating that they encourage their suppliers to respect the local national legislation rather than endeavouring to set up their own screening procedures for identifying risks to human rights. The GRI has set out specific guidelines as to how companies should approach human rights within their value chains. Under the guidelines, it is for the TNC to take courses of action, systematically screen and tackle human rights risks associated with their operations, rather than passing on the responsibilities to their local suppliers so that they can adhere with the local rules and regulations, leaving the question of what if some of the local norms, rules and regulations are either ambivalent to or in contradiction with the definition of human rights as set out by the GRI guidelines?
In contrast to the above, the lowest level of over-claiming is for incidents of discrimination (HR4 at 17%). Many companies stated that they had fully adhered to the guidelines and had no incidents of discriminations to report on (e.g., Lukoil, Repsol-YPF, Iberdrola). While on average 30% of companies had reported HR4, about 38% (i.e., 26% + 11.5%see Table 3) had either 'partially' or 'fully'reported on incidents of discrimination. A closer examination shows that the two compilations of HR4 are reported by more or less the same number of companies. An example of a company claiming 'full disclosure' but in fact only 'partially' disclosing is ABB and an example of a company claiming to report partial but providing no disclosure is Deutsche Post: 'All countries in ABB's sustainability management program are asked to report any incidents of discrimination. Six cases of discrimination and 18 of harassment were reported in 2010, resulting in a range of disciplinary measures.' (ABB, 2010, p. 26) 'Although Deutsche Post DHL is committed to managing diversity professionally and to creating a working environment that is free from discrimination, we do not publish data for this aspect. We do not report on the total number of incidents of discrimination or actions taken.' (Deutsche Post DHL, 2010, p. 221) Another example for HR4 is Rosneft, claiming full disclosure by stating its full adherence with the Russian legislation on human rights (not with those of the GRI) without stating what they are and how effectively they are enforced. This is similar to the Deutsch Bank example discussed earlier: 'In its activities, the Company adheres to the existing legislation prohibiting any form of human rights violation. Due to the nature of Company' activities and the existing Russian legislation, there is no significant risk of human rights violation by suppliers and contractors. In addition, in the process of defining report content this issue has not been found material.' (Rosneft, 2010, p.90) The most consistent reporting with the highest total disclosures (partial plus full disclosure) were for HR6 (at 57.3%see Table 3) and HR7 (at 55%see Table 3). The following examples of HR6 and HR7 are from: (1) KEPCO claiming full disclosure for HR6 while reporting no relevant information and (2) Fiat claiming full disclosure for HR7 but only partially reporting the information: 'KEPCO is strictly complying with the Labor Standard Act of Korea, ILO convention (No. 105 concerning the abolition of forced labor) and the labor standards of the 10 Principles of the Global Compact. (KEPCO, 2009, p. 57) '… there is no use of child or forced labor at [our] plants or plants of [our] suppliers. Every two years, the Group conducts a review to determine whether child labor is being used. On the basis of the latest analysis of 90% of employees outside Italy conducted in 2009, no company was found to have personnel under the minimum legal age to begin work or an apprenticeship under local legislation or, in any event, under fifteen, even in countries where the minimum legal age is lower.' (Fiat, 2010, p. 32) The example from KEPCO is another case of a company using its adherence to its national legislation as a justification for claiming full adherence with the GRI guidelines without doing so and without stating which specific aspects of the Korean legislation are in line with those of the GRI guidelines. In other words, in the absence of any mandatory external checks to ensure that companies reported on the specific aspects of the GRI as claimed in their GRI index tables, they can easily use their index tables as a platform to adopt different legitimacy strategies. In cases where companies pass on the responsibility for human rights to their suppliers, they claim full disclosure and provide an explanation, attempting to alter the societal expectations by aligning them with the way they report on how the key issues are dealt with in practice (as opposed to how as per GRI guidelines) (i.e., the fourth strategy).
For HR1 and HR2, companies recognise the importance of their value chains as over half of them over-claimed that they had reported on the two key indicators related to their value chains in their GRI index tables. This is further supported by the fact that nearly half of the companies reported on their child labour and forced labour practices. More specifically, companies reported on the measures they had taken to prevent the incidents of child labour and forced labour but did not report on such incidents to begin with. The two latter indicators (i.e., HR6 and HR7) were the second most popular indicators that companies over-claimed. Considering that value chains and the issues pertinent to them had recently become more topical (Gold, Seuring, & Beske, 2010) prior to the time when we collected data and hence subject to more scrutiny (e.g., Porter & Kramer, 2011;Tate, Ellram, & Kirchoff, 2010), our evidence can be one of the first indicating a low level of exposure by companies. However, with the increasing attention paid to sustainability issues in value chains and the realisation of a sustainability-related value chain risk management (Hofmann, Busse, Bode, & Henke, 2013), TNCs are bound to pay more attention to their labour and human rights issues at national and international levels.

Discussion and concluding remarks
With awareness of breaches of labour rights and human rights increasing, more attention is paid to the fair treatment of the workforce in global value chains. TNCs are increasingly under pressure to provide an account of how fairly and ethically they treat their workforce. In this paper, we focus on the largest TNCs and closely examine the information they reported on their external as well as their internal workforce in their value chains, having adopted the GRI reporting guidelines on the labour practices and human rights and declaring the status of their compliance with each specific aspect in an index table. While TNCs in our sample generally realise the significance attached to their external workforce, evidence of over-claiming on all indicators suggests that TNCs fail to pay close attention to the reporting guidelines.
The GRI core indicators are generally 'material' to most companies (GRI, 2011, p. 7). However, in those cases where some information may not be material to some companies, the specific cases can be identified and their non-materiality declaredwith reasonsin the reports. The GRI sets out clear guidelines on the materiality of the information which require companies not only to identify the key issues but also to actively engage with their stakeholders. While the absence of materiality could be an issue on some GRI categories (e.g., major oil spillages under EN GRI categories, not under scrutiny, here), it is difficult to see how any of the labour or human rights categories, defined under the GRI guidelines, could allow any company to exempt themselves. However, attempting to justify material exemption should at least be transparent. Instead, we found numerous cases of TNCs reporting large volumes of irrelevant information (as also observed by Fortanier et al., 2011) as a means of deviating from the key information disclosure required. There appeared to be a tendency among some TNCs to claim full disclosure in their GRI index tables (S8, S6 and S4see Fig. 1) even though, in the report itself, the absence of core details were explained away as being irrelevant to home country government regulatory requirements. In cases of over-claiming, we found hardly any evidence of companies flagging how they dealt with difficult situations or the impediments they encountered in practice. While some companies could have highlighted inherent discrepancies, which to a large extent, are expected across diverse locations, they instead chose to claim 'full disclosure' and failed to provide the required information. At best, some over-claiming companies acknowledged that they operated in full adherence to 'local' rules and regulations without any further discussion of what these rules were, how they differed from the recommended guidelines, or how they impeded them from full or partial compliance with the GRI guidelines. One of the essential aspects of TNCs reporting information as an effort to discharge their accountability, is the acknowledgement of impediments they face or the weaknesses in their systems; issues which, if not acknowledged and discussed, can never be addressed or remedied. Our examination of the reports revealed none of these. Most TNCs were complacent about the guidelines at a practical level and to the question of whether they could implement them (reflecting, somewhat the findings of Barrientos & Smith, 2007). The geographical dispersion of TNCs' operations is one of the main features of global value chains that can result in considerable discrepancies in how TNCs treat their workforce (Edwards & Kuruvilla, 2005). This was not reflected or acknowledged in the majority of the reports we examined and TNCs were selective in adopting the guidelines (Diouf & Boiral, 2017;Lauwo et al., 2016).
In essence, having the option to claim high compliance with the GRI guidelines was treated by TNCs as a tool to convey that they were 'doing well' in terms of their workforce. While TNCs reported more on high profile aspects related to their internal workforce (such as 'employment type', 'rates of injuries and absenteeism by region', 'education and training of the workforce and their families regarding serious illnesses' and 'the composition of the boards'), they paid considerably less attention to human rights aspects (such as 'risk of child labour', 'risk of forced labour' and 'risks to freedom of association'). There were notable weaknesses in providing detailed statistics, breaking down information on internal workforce (e.g., statistics per: region, age groups, gender and for minority groups) as well as presenting detailed information on external workforce in value chains. This could be associated with TNCs' failure to either invest in systematic data gathering or recognise the importance attached to such information or identify and acknowledge material differences. Instead, we came across spurious statistics coupled with many statements of intent.
While the GRI has made a valuable attempt to introduce some form of standardisation in order to elevate the quality of information by focusing on a number of characteristics including transparency, materiality and comparability of information (e.g., GRI Standards Glossary 2016, p. 11; Reynolds & Yuthas, 2008;Sutantoputra, 2009), a lot more needs to be done. Our evidence suggests that those leading global companies that had explicitly claimed to have adopted the GRI guidelines on Labour Practices and Human Rights did not necessarily present material information that would result in enhanced transparency and would allow meaningful comparability. Indeed, there appears to be a (deliberate) misinterpretation of the guidelines by TNCs, whereby TNCs easily overclaimed their adherence to the guidelines without being found out or any independent external body flagging out what TNCs did not report and got away with. Our evidence that TNCs may have adopted a number of different legitimacy strategies is just a small indication of how the absence of any meaningful (mandatory) external audit has led the GRI guidelines to serve purposes other than those originally intended. Some of the guidelines (e.g., for HR1 and HR2) are devised to encourage stakeholder engagement in developing countries. Again, this was almost non-existent when we examined the reports. The identification of key issues and acknowledgement of material weaknesses (as suggested by Roberts, 2009) and active engagement with (the local) stakeholders are important facets of corporate accountability (Campbell, 2005;Cooper & Owen, 2007).
The overwhelming evidence of over-claiming could be a manifestation of 'organisational hypocrisy' (Brunsson, 1989) about their workforces in response to reputational competitive pressures and the desire to project a legitimate image (e.g., see Brunsson, 2003, p. 221). More generally, it reflects the predominance of neoliberalism (Lauwo et al., 2016;Sikka, 2010Sikka, , 2015 whereby all the systems and measures including accounting (actors, techniques and conventions) (Chiapello, 2017) are designed to privilege financial capital without paying meaningful attention to the rights and welfare of labour. TNCs' failure and, to a large extent, ambivalence to reporting material issues not only reflects TNC's unsystematic approach and lax attitude towards reporting but highlights a much more poignant problem that in the absence of any systematic scrutiny and pressures that would highlight TNCs' hypocrisies and hold them to account, there is unlikely to be any changes in corporate behaviours and, more importantly, to their organisational culture.
With the GRI reporting guidelines being replaced by the new GRI Sustainability Reporting Standards 18 from 1 July 2018, 'materiality' has remained central to reporting standards. As a standard setter, the GRI does not consider its role to check the content of the reports to ensure that TNCs have reported in accordance with its standards (see GRI 101, p. 26), and it remains for the TNCs to seek assurance provisions. Given the observed level of exaggeration and that many TNCs in our sample disingenuously stated that some indicators do not apply to them, there is a need to independently check the 'completeness' of stated disclosures and openly publish the results. While this independent examination can be funded by companies that wish to receive accreditation on what they have reported, there is a range of key issues that need to be considered and addressed here. For example, under whose jurisdiction might this independent examination take place? How would the independence of such examination be ensured? How well-informed would the examiners be about workforce related issues within different national borders, especially in developing countries and emerging markets where civil society institutions are often weak or non-existent? In our view, the presence of strong and functional civil society institutions (e.g., NGOs, trade unions, free media and other supportive mechanisms) within different national borders are essential if TNCs are to be monitored for the treatment of their workforce in their value chains. This cannot be achieved unless national governments in developing countries and emerging markets legally acknowledge the importance of social activist groups and grant them the freedom (as well as the protection needed) to operate freely and independently. At the same time, national governments need to devise the necessary legal and regulatory requirements with effective enforcement mechanisms to protect their workforce.
Our recommendations resonate with similar calls by recent studies on developing countries for substantial regulatory, policy and institutional reforms (e.g., Belal, Cooper, & Khan, 2015;Lauwo et al., 2016). Belal and Roberts (2010), for example, recommend that the core aspects of CSR reporting should be made mandatory in Bangladesh but doing so would require the effective implementation of law enforcement agencies to be strengthened and adequately resourced. Similarly, Lauwo et al. (2016) raises concerns over the adequacy and effectiveness of legal and regulatory requirements and calls for a more proactive role by government. A more supportive role by government can have defining impacts on CSR initiatives (as found by Rahman, 2017).
Another recommendation would be for TNCs to include representation of their internal workforce in their governance structures. Future research could compare TNCs that are headquartered in institutional environments with opposing approaches to their workforce in their governance structures (e.g., Anglo-Saxon versus stakeholder approach governance mechanisms) to explore any notable differences in the extent and nature of how the two groups of TNCs report on their internal and external workforce. Germanbased TNCs, being from a co-ordinated market economy (CME), have employees systematically involved in their governance structures and in various aspects of business activity (Hiss, 2009). By contrast, British TNCs, being from a liberal market economy (LME), have no employee representation in their governance structures and are characterised with weak employee voice and operate within a minimalist legislative employment framework (e.g., Muller-Camen, Croucher, Flynn, & Schroder, 2011). How might such differences influence the treatment of, and reporting on, the external workforce? While there are significant differences in certain aspects of human resource management practices of companies from LMEs and CMEs (e.g., Farndale, Brewster, & Poutsma, 2008), future research can broaden our understanding of how differences in governance mechanisms of these archetype TNCs (e.g., Bottenberg, Tuschke, & Flickinger, 2017) set the remits for their workforce reporting in view of their internal structural contexts and support mechanisms and how differently they respond to prevailing neoliberal forces at play in their national and global environments (as suggested by Chiapello, 2017).