The Dutch Banking Sector Agreement on Human Rights: An Exercise in Regulation, Experimentation or Advocacy?

The role of banks in projects which result in adverse human rights impacts has been brought to the fore in recent years. However, there are serious obstacles to regulate the (often extraterritorial) financing activities of banks under national law. The UN Guiding Principles on Business and Human Rights and OECD Guidelines for Multinational Enterprises attempt to respond to this ‘governance gap’, stipulating that all business enterprises have a responsibility to respect human rights. However, banks’ compliance with such standards has been frustrated by a failure to understand how these standards apply to them.<br><br>In 2016, the Dutch Government collaborated with the Dutch banking sector and civil society to create the Dutch Banking Sector Agreement: a multistakeholder initiative initiated to improve adhering banks’ performance with respect to the UN Guiding Principles on Business and Human Rights and the OECD Guidelines on Multinational Enterprises. This article reviews how the actors involved drafted the Agreement in light of prevalent divergences in understandings over how human rights apply to banks’ financing activities. It then looks to scholarship on transnational private regulation, experimentalism governance, and social constructivism in mapping three roles the Dutch Banking Sector Agreement could play: regulate, experimentation, and advocacy.


85
The Dutch Banking Sector Agreement on Human Rights: An Exercise in Regulation, Experimentation or Advocacy?
Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses but also focussed on multistakeholder strategies that could create possibilities for taking action that would otherwise be unavailable to individual businesses. 7 In October 2016, 'the Dutch Banking Sector Agreement on international responsible business conduct regarding human rights' (the DBA) 8 was announced: a collaboration between the banking sector, the Government, trade unions, and civil society organisations (CSOs), all based within the Netherlands. This initiative is the first initiative representing banks, CSOs, and Government in relation to finance and human rights. It was described by the then Minister of Foreign Trade and Development Cooperation Lilianne Ploumen as 'setting an example for the world to follow, and that is something to be proud of'. 9 The DBA focusses on banks' responsibility to respect human rights, as stipulated in the UNGPs and Guidelines, 10 within their corporate lending and project finance activities. 11 The content of the Agreement broadly follows the requirements of the second pillar of the UNGPs: it requires banks to have a policy commitment to respect human rights, 12 human rights due diligence procedures, 13 to enable remediation, 14 and to be transparent about what steps they take to do so. 15 The Agreement thus represents a unique opportunity for banks to improve their human rights due diligence processes in partnership with both Government and civil society, whose objective is to ensure that the adhering banks meet their international human rights responsibilities as stipulated under the UNGPs and OECD Guidelines. 16 It also aims to increase their capacity to deal with problematic human rights issues they cannot resolve on their own. 17 It was signed in December 2016 and has had limited public outputs at the time of writing. It would therefore be premature to make predilections as to what the likely impact of the Agreement will be. However, given the Dutch Government's ambitions to replicate similar models across different financial sectors in the near future, it is appropriate at this juncture to discuss the potential contribution the Agreement can make towards the human rights performance of banks.
In doing so, this article will first describe how banks' existing procedures and perspectives have affected how they understand how the UNGPs/Guidelines apply to them, focussing on the divergences from authoritative guidance (Section 2). Section 3 will then introduce the key features of the Agreement: its governance structures, the actors involved and its content. Section 4 will then analyse the extent to which divergences between banks' existing procedures/perspectives and the logics of UNGPs/Guidelines appear to have played a role in the drafting of the DBA. Section 5 will conclude with where the added value of the DBA might lie, concluding there are three roles the DBA can play to further banks' human rights performance: regulation, experimentation and advocacy.
Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses 86 role this plays in their understanding of their relationship to human rights and to the extent of the UNGPs/ Guidelines relevance to them, identifying five sources of misalignment (Section 2.2). It will then look at the main industry-led attempt to provide guidance on how human rights responsibilities apply to banks -the Thun Group papers of 2013 (Section 2.3) and 2017 (Section 2.4) -examining how these misalignments have played a prominent role in the drafting thereof.

The application of human rights to banks
The UNGPs set out what is required of business enterprises to meet their responsibility to respect: a human rights policy; human rights due diligence procedures which identify potential adverse human rights impacts, take measures to prevent and mitigate them, track the effectiveness of such measures, and communicate these actions externally; and the remediation of any adverse human rights impacts business enterprises cause or contribute to. At the same time, they do not provide for a precise means to discharge this responsibility: 18 'they are not intended to be a toolkit, its components simply to be taken off the shelf and plugged in'. 19 The UNGPs/Guidelines state that a business enterprise's responsibility to respect human rights requires that they 'avoid causing or contributing to adverse human rights impacts through their own activities, and address such impacts when they occur [and] seek to prevent or mitigate adverse human rights impacts that are directly linked to their operations, products or services by their business relationships, even if they have not contributed to those impacts'. 20 For a business enterprise to be 'directly linked', no causal connection between the activities of the business and the adverse human rights impact is necessary. A business enterprise is directly linked to any adverse human rights impacts 1) caused by any entity with which it has a business relationship and 2) linked to its business operations, products or services. 'Business relationships' extend not only to business partners but to entities (non-State based and State based) in a business enterprise's value/supply chains. 21 The Guidelines further state that direct linkage does not 'shift responsibility from the entity causing an adverse human rights impact to the enterprise with which it has a business relationship'. 22 This phrase was later also adopted by the UN Office of the High Commissioner for Human Rights (OHCHR). 23 There is no express definition of what it means to cause or contribute to an adverse impact apart from the fact that it must arise from a business's 'own activities' which may include both actions and omissions. 24 The Guidelines further define contribution as 'a substantial contribution, meaning an activity that causes, facilitates or incentivises another entity to cause an adverse impact and does not include minor or trivial contributions'. 25 The OECD has cited the following as examples of contribution in a finance context: a bank wielding control over the client in an investment relationship, 26 a bank setting up 'an unrealistic timetable for a construction firm to build offices for the bank, resulting in labour abuses', 27 and a bank 'lending money to a company to construct a large processing plant to be built on a community land that results in the displacement of affected populations without meaningful stakeholder engagement'. 28 In this latter example, the reason that the bank is deemed contributing to eventual harm is neither that it funded the harm nor that actions undertaken under the business relationship made the harm more likely. It is because the bank failed

87
The Dutch Banking Sector Agreement on Human Rights: An Exercise in Regulation, Experimentation or Advocacy?
Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses to have appropriate mitigation measures in place to prevent the harm. If the bank had advance agreements with the client that they carry out meaningful stakeholder engagement, it would be directly linked. 29 The OHCHR has clarified that direct linkage does not intend to differentiate 'direct' linkage from 'indirect' linkage; even relatively minor relationships are considered to result in direct linkage, and, while a business can be directly linked to a harm regardless of the number of intermediary relationships between them and the entity causing the harm, there must be a connection to its products, services and operations. 30 The OECD has followed these criteria. 31 This connection to products, services and operations is the principle means by which 'reasonable limits' are placed on the extent of a business enterprise's responsibility to respect in directly linked situations. 32 The example used is where a business is sourcing clothes from a factory which makes handbags on a separate production line; the business enterprise would not be directly linked to any harms resulting from the production line on which handbags are made as it is only purchasing the clothes. 33 In the context of finance, the OHCHR proffered the example of a bank lending money to a client for use in a particular project then it will not be linked to harms resulting from another project carried out by the same client. 34 A further means of limiting the scope of actions to be taken is through the size, sector, operational context, ownership and structure of the business enterprise, with larger companies expected to do more to meet their responsibility to respect human rights, 35 and the severity of the impacts, which is to be judged by the scale, scope and irremediable character of the impacts themselves. 36 The OECD has endorsed prioritisation as a means by which banks can take action in relation to the most severe human rights impacts they are linked to. 37 There are two main differentiations of responsibility based on these categories of involvement: what actions are expected from a business with respect to their human rights due diligence and what role the business should play with respect to remedy. While a business enterprise's due diligence processes are expected to cover adverse impacts that arise under all three categories of involvement, 38 the extent to which the business enterprise should take action will vary depending on whether it is causing, contributing or directly linked: 39 'Where a business enterprise causes or may cause an adverse human rights impact, it should take the necessary steps to cease or prevent the impact', 40 whereas in cases where a business enterprise is directly linked it is not expected to provide remedy but to still exercise leverage over the entity causing the impact(s) to prevent or mitigate it/them. 41 Where they do not have sufficient leverage to end the impacts, they should consider adopting means to increase that leverage to end the impact or consider ending the relationship with entity causing the impact. This will depend on whether that relationship is 'crucial' to the business enterprise and what potential adverse human rights impacts could result from terminating it. 42 With respect to remedy, a business enterprise should provide for or cooperate in the remediation of impacts which they have caused or contributed to 43 and may provide remedy for those impacts they are directly linked to. 44 While not stipulated in either the UNGPs or the Guidelines, there has been a recent inclination amongst the OHCHR and OECD to expect business enterprises to exercise 29

Banks' approaches to human rights
Banks have to a large extent already incorporated many similar practices to those required under human rights due diligence across some of their activities, in their existing environmental and social (E&S) due diligence procedures, which they have adapted to include human rights risks. However, there are five key differences between the logics of E&S due diligence and human rights due diligence as understood under the UNGPs/Guidelines. First, E&S due diligence is not always understood as a matter of responsibility but instead of good business or 'opportunity'. One asset manager anonymously stated, 'We have an opportunity to drive E&S performance, but not necessarily the responsibility. Our fund managers would definitely not agree with the "responsibility".' 46 The contexts in which banks take action is often understood by the extent to which they have the opportunity to enact change in a client's behaviour, rather than their responsibility for an impact. 47 Thus, banks have traditionally taken actions to prevent adverse impacts, but, in cases where they fail to do so, this is more likely be deemed as a missed capitalisation on a particular opportunity to affect change, rather than an omission potentially giving rise to increased responsibility, as understood in the context of the UNGPs. 48 This difference in understanding will be labelled as the opportunity to affect/ responsibility to respect divergence.
Second, E&S due diligence is principally concerned with risks that could present a liability to the financial institution, often understood as a loss of revenue, costly litigation or a decline in reputational image. Human rights risks are considered risks to the bank rather than risks to people, although these risks may coincide and there is evidence for a change of attitudes in this regard. 49 This market-based paradigm depends on adverse impacts on a bank's bottom line as the key incentive for a bank to carry out due diligence, and it also acts as a disincentive for banks to take action where doing so might deprive them of a business opportunity. 50 This difference between E&S due diligence and human rights due diligence will be referred to as the risks to bank/risks to people divergence.
Third, E&S due diligence is commonly understood to primarily cover those actions taken by banks up to the point where the transaction takes place as opposed to the full duration of the business relationship as under the UNGPs. 51 The types of measures expected will depend on the type of financial product offered and the type of client. For instance, micro-financing may only require transaction screening -which may largely only involve the commercial team which initially reviews the proposed transaction -while corporate lending and project finance will likely require the involvement of E&S risk assessment and compliance teams, with escalation to higher decision-making bodies in cases where the proposals are deemed high risk, such as credit committees, reputational risk committees, and the Board. In the latter case, a much broader array of checks and standards may be expected to be carried out, including desk reviews of public and commercially provided information, client assessments, and site visits back the bank's risk assessment teams. This means

89
The Dutch Banking Sector Agreement on Human Rights: An Exercise in Regulation, Experimentation or Advocacy?
Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses that the focus of banks is often toward larger corporates and clients in less regulated countries, which may thus fail to capture severe human rights risks which lie with SMEs or in well-regulated countries. 52 The reasons banks feel that actions to be taken should be determined by the type of transaction has two justifications: one in opportunity and one in policy. Banks focus on transactions because they tend to view their level of influence over the client as at its highest during their decision as to whether or not to provide finance, decreasing significantly once finance has been provided (e.g., over the course of the relationship). 53 Also, as the implementers of the actions to be taken, banks will favour department specific policies, whereby the actions to be taken are already framed by the products offered, providing clear an implementable guidance for employees. Thus E&S risk-based due diligence refers to processes undertaken prior to making a decision relating to the provision of financial services to a client, 54 rather than over the full duration of the relationship as expected in the UNGPs/Guidelines. This is labelled the transactional due diligence/human rights due diligence divergence.
The fourth, is simply that banks limit themselves primarily to engaging with their client and not external stakeholders, as understood in the UNGPs. 55 This will be referred to as the client-only engagement/stakeholder engagement divergence. The fifth and final divergence is that banks do not necessarily see the need to 'account for' what actions they take publicly, sometimes citing regulatory rules or client confidentiality as a reason for not doing so. 56 In the UNGPs, there is an understanding that the relationship between civil society and business will move from 'naming and shaming' -whereby CSOs shame businesses for human rights impacts -to 'knowing and showing' where businesses understand and deal with their human rights impacts, showing what actions they are taking. 57 In drawing attention to the above divergences, I am not attempting to suggest that there are not justifiable reasons for why banks may view their responsibility in a particular way or that there is a uniform understanding of human rights across all banks and departments. Neither am I suggesting that the above divergent categories cannot overlap with each other: the avoidance of human rights controversies that might adversely affect the bank will, to some extent, include considerations of risks to people. Instead, I am describing these divergences in order to illuminate why there are continued deliberate or accidental misunderstandings amongst banks as to how the UNGPs/Guidelines apply to them, including key concepts such as responsibility, business relationships, remediation, termination, and leverage.
When banks have considered their involvement in human rights impacts, they often did not consider themselves to be in even the lowest category of involvement (directly linked) in relation to human rights harm, preferring to consider themselves as not involved or 'indirectly linked'. 58 They have also sometimes understood terms within the UNGPs/Guidelines narrowly, considering 'business relationships' to not include even their immediate clients and value/supply chains as limited to their own suppliers. 59 There has also been a stilted recognition of the role of termination of relationships or remediation within banks' activities. Few banks have demonstrated actively engaging in remediation. 60 Some development banks have followed the lead of international financial institutions such as the International Finance Corporation and European Investment Bank in setting up non-judicial grievance mechanisms. 61 To the knowledge of the author, no commercial bank has yet followed suit. In a 2013 survey of financial institutions conducted by the OECD, banks were asked to identify factors which would affect what leverage they had over a client in relation to an E&S risk. Banks cited the potential for an E&S issue to have a negative impact on company performance, for the E&S issue to have a negative impact on the bank, or for the client/investment to be publicly linked to the E&S issue as three of the four main factors as to whether they had leverage under the Guidelines in corporate lending situations. 62 These factors arguably relate more to risks to the bank than the leverage a bank has to influence the situation. Furthermore, banks' understanding of risks is still driven by a risks to bank paradigm. Commercial providers of E&S risk information for use in E&S due diligence, such as Reprisk, still view environmental and social risks as those risks to the bank: to a bank's reputation, to banks complying with national law and to a bank's finances. 63

The Thun Group paper 2013
These divergences have revealed themselves when banks have collaborated to provide industry guidance for the implementation of the UNGPs/Guidelines. The most prominent industry guidance proffered by banks are the Thun Group papers. The Thun Group is an informal gathering of a number of international banks in Thun, Switzerland, which includes Barclays, BBVA, BNP Paribas, Credit Suisse AG, Deutsche Bank, RBS, Standard Chartered, UBS Group AG, UniCredit, J.P. Morgan, and ING -the latter of which is also a party to the DBA.
The first Thun Group paper was published in 2013 and aimed to provide clarity on how the UNGPs applied to banks. 64 It was notably selective in which guiding principles it analysed. It covered the UNGPs' operational principles regarding the responsibility to respect (how a business should meet its responsibility to respect human rights) without discussing the UNGPs' foundational principles (what a business's responsibility to respect human rights is). 65 This thus allowed it to focus on a discussion of possible practices to improve human rights situations while refraining from discussing when a bank can become involved in a human rights impact and which category of involvement it might find itself in. With regards to leverage, it did hint that where there is a lack of leverage to prevent human rights impacts this may prompt an in principle decision to decline the transaction in advance. However, where a decision to enter such a transaction is taken, it did not discuss the possible need to remediate the impact (as required in contribution situations) and/or terminate the business relationship (as required in some directly linked situations), where leverage fails. 66 Instead, it simply concluded that leverage is a factor in considering the potential for human rights impact and risk mitigation. This thus omits many of the consequences of viewing which actions should be taken under the responsibility to respect, favouring an opportunity to affect paradigm.
This paradigm also affected understandings of the scope of due diligence required. The definition of value/supply chains was once again relegated to a bank's own suppliers, not entities that constituted part of a client's supply or value chain, and was deemed outside of the scope of the report. However, unlike in some responses to the OECD 2013 survey, clients were deemed part of a bank's business relationships, and the human rights impacts of clients were deemed within the scope of banks' human rights due diligence. Because categories of involvement under Guiding Principle 13 were not within the scope of the report, Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses the paper did not differentiate between what actions would be required under due diligence on the basis of whether they were causing, contributing or directly linked to the impact. Instead, the extent of what is required under due diligence is to be determined using 'a practical and sensible perspective, taking into account the significance of the impact and the degree of linkage to the business activity. It will not be possible to evaluate every impact of every business decision regardless of proximity'. 67 This idea of 'proximity' was never defined but was introduced as a new factor which, alongside severity of impact, determined what action a bank should take. 68 The risks to bank lens had a more mediated role in defining how human rights risks should be understood. The paper cited both banks' desire to manage their impacts on society responsibly and the need to minimise related risks to the bank as a basis for a bank's decision to follow the UNGPs. 69 The document was aimed at mapping and analysing banks' 'potential adverse impacts on human rights and related risks for their own operations, including reputational, legal, operational and financial risks'. 70 Thus the report expressly orientated itself under a 'risks to people' approach while recognising a 'risks to bank' perspective as an important incentive for doing so. This had a notable effect on the content of certain statements the report made. For instance, it stated consideration should be given to a form of risk management which identifies and assesses potential impacts on rights holders, not simply risks to the bank itself. 71 It was never stated that risks to people were only to be recognised where they simultaneously presented a risk to the bank. At the same time, the report discussed 'human rights risks' to the bank throughout its various sections and sometimes referred to human rights impacts as 'controversies', illustrating that public perception and other risks to the bank still framed the authors' understanding of these issues.
The transactional due diligence approach also played a mediated role. On the one hand, an understanding of due diligence based purely on the steps taken to determine whether and how a transaction should take place was rejected: 'due diligence is an ongoing process, not something to be completed once and not revisited.' 72 At the same time, classification of what types of due diligence processes were necessary were based on a risk profile of the product and client, and the report divided the types of due diligence processes by retail and private banking, corporate and investment banking and asset management. 73 The client-only engagement approach was prevalent, with all discussion of banks' engagement being relegated to internal engagement and engagement with the client. Engagement with external stakeholders was the responsibility of the client, not the bank: 'Commercial banks have neither the expertise nor legitimacy to engage with clients' stakeholders.' 74 The knowing and withholding paradigm also played a role. Banks were expected to publish information on their policies and processes, as well as where possible their performance. 75 However, it was made clear that no such transparency should ever affect (pose a risk to) the 'legitimate requirements of client confidentiality'. 76 At the time of its release, key criticisms were posited around its failure to consider categories of involvement, 77 its failure to address the issue of remedy, 78 and its failure to endorse effective engagement with victims. 79 However, the 2013 paper was generally warmly received, 80 particularly for being the first 67 2013 Thun paper, supra note 64, p. 9. 68 Ibid., pp. 9-10. 69

Thun Group paper 2017
The second Thun Group paper expressly dealt with banks' perspectives of the scope of a bank's responsibility to respect human rights for the first time: rather than simply enumerating what types of action a bank can take, it focused on what actions a bank should take and, perhaps most controversially, what steps were not necessary for banks to take. It looked at Guiding Principle 13, which specifies the categories of involvement. The paper ruled out any question that a bank might cause or contribute to an adverse human rights impact through its financing or investment activities on the basis that these impacts arise from its clients' operations and are not occurring as part of the bank's own activities. 82 With regard to direct linkage, it reiterated that value chains were a bank's own suppliers and not the value chains of bank's financing activities. 83 It did, however, embrace the view that a bank can be directly linked to every impact caused by the activities of any company it provides corporate lending to, including the subsidiaries of that company. 84 The paper attempted to manage the consequences of this conclusion by using a new concept, 'proximity', as a means to delineate between 'high proximity' direct linkage and 'low proximity' direct linkage, based entirely on the corporate structuring of a corporate group rather than any UNGPs concepts. 85 The paper also stated that there was never any need for a bank to provide, or otherwise engage in, remediation in relation to impacts caused by clients they finance, as they cannot contribute to human rights impacts caused by their clients. 86 There was no discussion of when termination of relationships would be required under Guiding Principle 19 or when remediation would be required.
The opportunity to affect paradigm is clearly apparent in the 2017 paper, with John Ruggie (the author of the UNGPs) noting that, while the paper misreads the UNGPs, minimising the opportunity for banks to be seen as involved in human rights impacts, the examples the banks put forward of their conduct in the annexed case studies are much closer to what would be expected under the UNGPs. 87 The message is 'our responsibility under the UNGPs is low, but our actions exceed our responsibility', an ethos that banks can and do take positive actions even though they need not do so. Where a bank fails to take action where such an opportunity exists, the paper implicitly understands such situations as missed opportunities as opposed to a potential basis by which a bank should take greater responsibility for the impact: insufficiency of due diligence has no consequences for the responsibility of the bank. 88 The paper neither expressly defines human rights in terms of risks to people nor risks to banks as in the 2013 paper; it varies between discussing human rights impacts and (material) human rights risks. It also looks to E&S due diligence as the means by which the human rights risks can be mitigated. This contrasts with the 2013 OECD survey where banks had questions over the extent to which human rights due diligence could be incorporated into existing E&S approaches 89 and the 2013 Thun Group paper which stated that consideration should be paid to the development of a risk management system that goes beyond traditional parameters, 90 accepting the possibility that it might be possible that the UNGPs could not be met solely by amendment of existing procedures. 91 The transactional due diligence approach is codified into the 2017 paper: involvement is determined by the type of transaction and at the timing of the transaction. This is entrenched in another new concept introduced by the paper, 'unit of analysis', which defines the focus of due diligence by the type of financial transaction. 92 Both of the new concepts introduced to define what action should be taken appear to be understood at the point of transaction -'the type of financial product or service offered to the client determines this proximity upfront and thus the appropriate level of due diligence and mitigation measures' 93 -regardless of any developments in the situation post transaction. 94 There is no recognition in the 2017 papers of the need to exercise leverage after the transaction or of the continuum between directly linked and contribution whereby (in)action after the point of transaction can affect a bank's involvement: all of the case studies of direct linkage in the report end their discussion of due diligence at or before the point of transaction. The client-only engagement approach was also prevalent with any discussion of engagement limited to the client's responsibility. 95 There was no discussion of the need to account for what steps the bank takes publicly, reflective of the knowing and withholding attitude.
The paper received strong criticism from several corners including Ruggie, the UNWG, and a coalition of CSOs and individuals for failing to recognise that banks can contribute to adverse human rights impacts, 96 for not recognising that banks have a role in remediation, 97 for the addition of proximity as a means to dilute banks' responsibilities, 98 and for determining categories of involvement on an a priori basis: deciding what forms of due diligence should be carried out on the basis of the type of financing involved. 99 The overall perception of the paper was that elements of it were 'a cause for concern' potentially causing 'unnecessary confusion', 100 that it was 'deeply troubling' and took such liberties with the interpretation of central concepts in the UNGPs as to be comparable to 'alternative facts', 101 and that its misinterpretations were a potentially intentional means by which member banks sought to avoid their human rights responsibilities. 102 The Thun Group did commit to 'clarifying misunderstandings' related to the paper 103 and published a revised 2017 paper which largely ignored or relegated the above issues to being outside of its scope. 104 The above controversies prompted the OHCHR to release further guidance clarifying which factors influence how a bank is involved with an adverse human rights impact and what a bank's responsibility to remediate looks like in such situations. 105 They clarified that the severity of (potential) adverse human rights impacts is the most important criterion in determining the scope of due diligence required, while recognising that, due to the complexity and number of transactions a bank is engaged in, banks can prioritise in-depth due diligence for high-risk clients or transactions. 106 Where potential impacts are foreseeable in the Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses run-up to a transaction, mitigation measures within the transaction itself are the primary means by which a bank can ensure that it is directly linked to any resulting impacts instead of contributing, 107 but that even if a relationship with impacts is one of direct linkage at the point at which the transaction occurs, banks can be deemed contributing if they do not exercise reasonable leverage following the transaction to attempt to prevent and mitigate further impacts. 108 In this way, 'directly linked' and 'contribution' are deemed to be on a continuum.
The paper also gave additional guidance on how banks can remediate impacts, emphasising it is often not necessary for banks to engage in parallel remediation processes where their client is providing remediation but they should exercise leverage over their client to ensure they do so and/or engage or co-operate in remediation processes which at least remedy the extent of their own contribution. 109 Banks should set up their own grievance mechanism or engage in a mechanism set up under the course of an industry or multistakeholder initiative. 110 The interpretive guidance also highlighted the need to consider termination of relationships in situations where the exercise of leverage is insufficient to prevent adverse human rights impacts, 111 that direct linkage extends beyond client relationships, 112 and consultation with stakeholders is necessary both in the course of due diligence and when providing remedy. 113 It is against this backdrop of significant misalignments between the industry and the UNGPs/Guidelines that the DBA comes into play.

The Dutch Banking Agreement: A new response
The DBA represents one of five of the Dutch Government's 'IMVO covenanten', multistakeholder initiatives between corporations acting in particular industries, civil society organisations, trade unions and the Dutch Government: textiles, banking, plant proteins, sustainable forest management, and gold. 114 Two more such initiatives are expected to be negotiated in the financial sector, one in the insurance sector 115 and one in the pensions sector. 116 The Dutch Government has chosen these sectoral agreements as a principle means to discharge its own human rights due diligence in coherence with UNGPs stipulations regarding the State duty to protect, 117 contrasting with prominent innovations carried out by other governments in the realm of business and human rights, including an extraterritorial duty of care law 118 and legally binding reporting obligations. 119 While legally binding measures are not mutually exclusive with these sectoral agreements, the perceived adequacy of the DBA, and agreements like it, to improve businesses' behaviour may well have important implications for the Dutch Government's perception of the added value of any such legally binding measures.
The DBA seeks to apply the Dutch poldermodel of industrial relations -whereby labour issues are resolved through tri-partite cooperation, on equal terms, between employer's organisations, trade unions, and the Government -to the realm of banks and human rights, whereby the Dutch banking sector's industry association, Dutch trade unions and Dutch CSOs, and the Government attempt to co-operate to improve Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses Dutch banks' human rights performance. It requires adhering banks to meet the standards of the UNGPs/ Guidelines. Section 3.1 will describe the scope and governance of the DBA, Section 3.2 further examines the actors involved in the DBA and Section 3.3 then reviews the content of the DBA.

Scope and governance
The DBA has a limited material and temporal scope. Its material scope is limited to project finance and corporate lending, with the potential to broaden its scope to asset management. 120 Its temporal scope is a limited time span of three years during which adhering banks are required to implement the DBA. A distinction is thus made in the DBA between the regulators (the parties) and the regulated (the adhering banks). However, as the adhering banks are all members of one of the parties, the Dutch Banking Association, they do effectively have an opportunity to have their voices heard in decisions as to how they are regulated, to the extent that the association's internal processes properly represent member banks. A final actor is the Dutch Social and Economic Council (SEC), an organisation financed by industry, which is independent from the Government. The SEC traditionally facilitates discussions between employers' organisations and trade unions under the poldermodel.
The DBA creates three governance bodies to supervise its implementation: a Steering Committee, a Monitoring Committee, and a Dispute Resolution Mechanism. These governance bodies are then complemented by task-specific working groups whose aim is to implement key parts of the DBA. The Steering Committee oversees the day-to-day governance of the DBA and is composed of representatives of each type of participating actor (banks, Government, CSOs, and unions), as well as an independent chair. 122 All of its decisions are made unanimously on the basis of a balanced contribution between the parties. 123 The Monitoring Body is composed of individuals who are independent from the parties and oversees the parties' implementation of the DBA. This takes the form of an annual monitoring report which is delivered to the Steering Committee, which then publishes a summary. 124 The Dispute Resolution Mechanism is a mechanism open to parties who have a dispute about another party's performance, where complaints are decided upon by the parties (in the absence of those parties involved in the dispute). The working groups are set up with specific tasks, e.g., writing a report about how banks can increase their leverage, whose content is then published as part of the DBA.
The division of the three governance bodies can be understood as a loose division between standard setting, monitoring and enforcement. In the case of the DBA, standard setting can be seen in three locations: the DBA itself, the creation of future content which adhering banks are expected to comply with and the creation of performance indicators by which banks' compliance with the DBA is measured. The Steering Committee is responsible for standard setting. 125 It is composed of the parties who negotiated the DBA. Monitoring takes place through adhering banks individually reporting on their steps with respect to human rights compliance followed by annual valuations of banks' performance by the Monitoring Committee, based on the DBA and performance indicators. 126 The DBA also requires banks to publicly publish their human rights policy and due diligence procedures 127 and publicly report under the UN Guiding Principles Reporting Framework, an initiative based on the provisions of the UNGPs. 128 Enforcement can be divided into both 'ex post enforcement mechanisms, such as remedies and sanctions, but also ex ante instruments in form of monitoring and supervision'. 129 The DBA has an ex post enforcement mechanism in the form of the Dispute Resolution Mechanism which can expel a bank from the initiative for non-compliance. 130 However, the Monitoring Committee can also be said to have a role in enforcement as the information in its reports can be utilised by actors outside the DBA to hold adhering banks to account. Some analysts maintain that such ex ante instruments cannot truly be considered as enforcement, although the distinction between the two may sometimes blur in practice. 131 Others point towards the fact that stakeholders such as consumers, lenders, CSOs etcetera are the key drivers for businesses to meet their human rights responsibilities and therefore access to information by which these stakeholders can make decisions is essential in ensuring these initiatives are effective. 132 Some authors argue that structural separation is necessary to ensure that governance structures operate effectively. A research project which analysed numerous transnational private regulatory initiatives concluded that there was the need for a clear separation between standard setting, monitoring, and enforcement, with different actors being mandated with these different tasks. 133 Otherwise, conflicts of interests within these initiatives are likely to frustrate their effectiveness.
In the case of the DBA, there is a partial separation of actors involved across standard setting, monitoring and enforcement. The individuals involved in the Monitoring Committee are independent from both the standard setting (the 'parties') and the enforcement. One potential shortfall is that the conclusions of the Monitoring Committee are not made public directly. Instead, the Steering Committee publishes its 'summary of the analysis and the advice of the independent Monitoring Committee as part of its annual progress report'. 134 Hence, there is still the possibility of vested interests becoming involved prior to the information being made public. The Dispute Mechanism contains the same parties as the Steering Committee. There is therefore no functional separation between the devising of norms (standard setting) and the enforcement of those norms. While the DBA does stipulate that the party making the complaint and the party who acts as recipient of the complaint are not involved in determining the outcome, the fact that the same parties are involved in both standard setting and enforcement results in potential conflicts of interest and could have an adverse effect on the ability of the DBA to enforce its provisions. 135

The banks, the CSOs/unions, and the Government
There are fourteen adhering banks, and the interests they pursue within the DBA are by no means uniform; there are major differences between the banks that have connotations for how the DBA affects them. First of all, there is a significant difference in size between the banks. The so-called 'big three' -ING, Rabobank, and ABN AMRO -each sit in the top 100 banks in the world, 136 have combined total assets of around Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses 2 trillion euros 137 and play a dominant role domestically, controlling between 60 to 80% of the banking markets for mortgages, business loans and savings in the Netherlands. 138 There is a recent trend within the Dutch financial sector towards business models which emphasise activities in the domestic market. 139 Internationally, most Dutch banks are much less active, with the third biggest bank, ABN AMRO, having only around 10% of its activities outside of the Netherlands (largely in private banking), 140 and the second biggest, Rabobank, having around 25% largely focussed on foreign consumers and the agricultural sector. 141 ING is the exception: an international-facing bank with two-thirds of its assets related to foreign activities, mostly in other European countries. 142 This difference in lending portfolio is likely to lead to significant variations in the consequences of agreeing to more stringent lending practices under the course of the DBA between the banks involved. An illustrative example is project finance. ING is considered in the top ten banks in the world for project finance, while none of the other banks are in the top 25. 143 Other adhering banks have much less in terms of corporate lending or project finance activities, for instance Van Lanschot whose core activities are private banking, asset management and merchant banking services. Thus human rights requirements on lending activities are likely to have much greater connotations for some banks both in impact and relevance.
There is also a great deal of variation as to the extent to which the banks are placed under the spotlight when it comes to human rights concerns, and how they orientate themselves with respect to such issues. For instance, in BankTrack's 'Banking with Principles' report benchmarking different banks on their implementation of the UNGPs worldwide, only ING and Rabobank were large enough to warrant analysis. 144 In the Dutch Fair Finance Guide -an online tool designed by CSOs which aims to mobilise customers to influence financial institutions to make more sustainable investment choices -eight of the fourteen adhering banks are ranked by their human rights performance. 145 Two of them orientate themselves as sustainable, ethical banks -ASN Bank and Triodos -and would likely suffer more as a result of negative publicity. Likewise, one of the adhering banks is a part State-owned development bank (FMO) and thus has public facing mandate that many of the other banks lack. The variety of business interests and reputational incentives between different banks may play a role in what banks have been and will be willing to agree to.
The inclusion of CSOs is often deemed essential to ensure the effectiveness of non-State initiatives which aim to ensure that businesses respect human rights. 146 CSOs' independence and commitment to values (or normative expertise) can be a conducive factor to the DBA's effectiveness. 147 The inclusion of three CSOs who are all active in human rights advocacy, including the Dutch Fair Finance Guide, has the potential to result in higher expectations of conduct from banks than an industry-only initiative. 148 Most importantly, their inclusion helps prevent the DBA from being captured and misused by banks, as the three CSOs and the unions have the option to exit and criticise the initiative. 149 In addition to playing a role as 'social regulators', the CSOs/unions have their own obligations alongside banks, most notably in the provision of relevant Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses information on human rights concerns 150 and amplifying the voices of affected stakeholders on the ground. 151 This can further build the capacity of banks to identify and respond to human rights risks and impacts, and thus improve their human rights performance.
While the majority of multistakeholder initiatives around business and human right issues do not include governments, 152 the Dutch Corporate Social Responsibility (CSR) agreements all include a role for the relevant ministry(ies), including the DBA which includes both the Ministry of Finance and the Ministry of Foreign Affairs. Abbot and Snidal argue that a government can play an important role in a multistakeholder initiative through the 'orchestration' of such initiatives, 153 including enacting complementary policies which incentivise businesses to comply with them. 154 The argument is that orchestration by the State can help direct and support decision-making processes relative to the often decentralised and unguided decisionmaking processes in business-driven initiatives. 155 Abbot and Snidal argue for two potential roles of States in the orchestration of transnational private regulation: directive and facilitative. Directive facilitation includes the background threat of mandatory regulation which creates an incentive for businesses to comply with the initiative, as well as public procurement requirements, transparency requirements and trade concessions. 156 Facilitative orchestration requires softer, non-mandatory action such as 'convening, facilitating, legitimating, negotiating, publicizing, ratifying, supervising, partnering and otherwise interacting' 157 and the provision of secure finance.
The DBA contains elements of both directive and facilitative orchestration. The Dutch Government is committed under the DBA to incentivise 'responsible business conduct in general through, inter alia, public procurement and trade missions'. 158 However, the Government is not treated as solely a regulator/ incentiviser of positive human rights performance but in fact has its own obligations alongside banks, 159 most notably in the provision of relevant information on human rights concerns 160 and helping banks exercise leverage over other actors such as their clients. 161 Again, this can enhance the possibilities open to banks in terms of improving their own human rights performance. The Agreement is facilitated through the role of the SEC and by secure finance provisions. 162

The content of the Agreement
The content of the Agreement roughly follows the following structure: the focus and scope of the Agreement (Articles 1 and 2), how the responsibility to respect human rights under the UNGPs/Guidelines can be met by banks (Articles 3, 5, 6 and 7), other opportunities to improve human rights (Articles 5, 8, 9 and 10), the role of actors and governance structures (Articles 11-13) and final provisions on how the Agreement should be implemented and how potential conflicts with national law should be dealt with (Article 14). This sub-section will introduce the key elements of these provisions before they are further analysed in Sections 4 and 5. It is worth noting at the outset that there is some imprecise language in the DBA, bordering on obfuscation, with further specification of key concepts such as leverage and remediation left to future collaborative action between the parties. Article 1 founds the basis of the text of the DBA on the UNGPs/Guidelines, stipulating that banks should implement them throughout their activities, 163 while making clear that the majority of the DBA's own provisions requiring implementation of the UNGPs/Guidelines apply only to project finance/corporate lending at the moment, 164 with possible extension to investment activities in the event that insurers, pension funds and other institutional investors fail to negotiate agreements on such activities themselves. 165 Article 3 requires banks have a human rights policy, which must 'include a commitment to respect human rights, in conformity with the OECD Guidelines and the UNGPs', 166 as well as black lists of activities banks will not finance, information on human rights due diligence procedures and sectoral and thematic policies, outlining parameters in which the bank conducts business in sectors deemed high-risk. 167 Article 4 is entitled 'Human rights due diligence and client engagement'. It requires banks to have due diligence procedures within two years of signing the declaration of adherence, which must 'take into account' the work of the OECD Working Party on Responsible Business Conduct. 168 Article 4(2) makes clear that human rights due diligence goes beyond risks to the bank and includes risks to rights holders.
Engagement, however, is clearly limited to client engagement, with banks required to 'ascertain' that client processes involve meaningful and effective consultation with potentially affected groups and other relevant stakeholders: 169 banks must require such consultation in project finance situations and other situations required by the International Finance Corporation's Performance Standards or Voluntary Guidelines on the Responsible Governance of Land Tenure and actively promote it in corporate loans where there is a fair possibility of land rights violations. 170 It requires banks to require their clients carry out Free Prior and Informed Consent procedures in situations required by the above standards and promote it in corporate loans in situations where there is a fair possibility of land rights. 171 It also requires that banks monitor whether the steps they take with clients are successful, look to means to increase leverage, and as a last resort, at the discretion of the bank, end a relationship with a client where they are not willing or able to comply with due diligence requirements. 172 Adhering banks are expected to collaborate with each other and the parties in improving their due diligence systems. 173 Article 6 deals with transparency, requiring banks to disclose their exposure to economic sectors and their lending portfolio, including their investment universe for asset management. 174 They must report to the Steering Committee on their commitments with respect to their development of and implementation of human rights due diligence procedures and human rights policies. 175 They are also required to report in line with the UN Guiding Principles Reporting Framework, a multistakeholder designed reporting initiative based on the UNGPs. 176 Article 7 refers to 'Enabling Remediation', restating the principle statements made by the UNGPs and Guidelines on remediation. 177 It stipulates that banks require their clients to have, or participate in, grievance mechanisms in cases of project finance or promote their clients to have, or participate in, grievance mechanisms in cases of corporate lending where there are severe human rights violations known to the bank. 178 The question of a bank's role in remediation is left to a working group which includes representatives of the parties and independent experts, including a representative from the OECD Working Group. 179 The Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses DBA specifies that a voluntary mechanism will be set up by the Dutch Banking Association to deal with 'notifications of breaches' of the Guidelines and 'eligible complaints'. 180 The outcome of this mechanism is advice to the bank as to how it can best handle selected cases of human rights impacts. 181 Articles 5 and 8-10 also look to steps banks must take, but the requirements appear to be more 'decoupled' from the concepts in the UNGPs as the requirements above. Instead, they appear more in line with an 'opportunity to affect' logic than one borne from the responsibility to respect. 182 Article 5 concerns the design of 'tools', one on human rights risk and impact identification -a human rights matrix/database developed through joint commission of a third party -and value chain mapping exercises -starting with an analysis of palm oil, cocoa, gold, and then oil and gas through a collective effort of banks facilitated by a third party. These mapping exercises are then followed by follow-up activities, 183 and the parties will collectively strive to make a difference in respect of human rights and labour rights (including freedom of association, collective bargaining and a living wage) 184 within the selected sectors. 185 Article 9 requires parties and adhering banks to work together and publish a study of increasing leverage which looks to creative ways to increase leverage through engagement processes, traditional commercial leverage, alongside business peers, leverage through bilateral engagement and leverage through multistakeholder engagement. 186 This study is to be used by banks for input to their continuous constructive efforts to increase their leverage, for example when addressing the findings of the value chain exercises. 187 Article 10 is a short 'Sustainable Development' clause which does not stipulate any action to be taken but states that the parties and adhering banks to the DBA 'are motivated to go beyond acting or supporting actions in conformity with the OECD Guidelines and the UNGPs ("do no harm") to include promoting human rights by promoting sustainable development ("do good")'.

The Dutch Banking Agreement: Achieving convergence or divergence with the UNGPs/ Guidelines
Having introduced the content of the DBA, Section 4 will look at the extent to which the five divergences identified in Section 2 -opportunity to affect/responsibility to respect, risks to banks/risks to people, transactional due diligence/human rights due diligence, client-only engagement/stakeholder engagement, knowing and withholding/knowing and showing -have played a role in the text of the DBA.

Opportunity to affect vs responsibility to respect
The opportunity to affect paradigm diverges from the responsibility to respect paradigm in two ways. First, it focusses on prioritising action based on where there is the most opportunity to affect change as opposed to the severity of potential/actual impacts. Second, it seeks to downplay a bank's human rights responsibilities and does not use categories of involvement under the UNGPs/Guidelines -'causing', 'contributing' and 'directly linked' -as a basis for what action should be taken. Consequently, many banks have traditionally failed to recognise banks' role in access to remedy, the need to terminate business relationships under the UNGPs/ Guidelines and that the need to address the extent to which banks' involvement in human rights impacts extends beyond their immediate clients. Both aspects of this paradigm can be seen in the drafting of the DBA. The former can be seen to some extent in the fact that many of the initiatives under the DBA either go beyond the requirements of the UNGPs/Guidelines, e.g., Article 10 on Sustainable Development, or may be a means to help banks further meet the UNGPs/Guidelines but are not expressly couched in terms of the

101
The Dutch Banking Sector Agreement on Human Rights: An Exercise in Regulation, Experimentation or Advocacy?
Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses responsibility to respect, such as the development matrix/database and the value chains mapping exercises. However, severity of impacts is a key basis for where such efforts have been prioritised: value chains were chosen on the basis that they all represent high risk sectors for banks 188 and the matrix/database is expected to start with mapping those sectors which are deemed high risk by the adhering banks 189 While the DBA does not expressly state the extent to which it is deemed that such actions are necessary, relevant to, or over-and-above what is required in the UNGPs/Guidelines, banks are bound by the DBA to take action in contributing to these initiatives.
The DBA covers termination of business relationships, value chains, and remediation. However, it does not necessarily couch these actions in terms of responsibility, and places them in a lower tier of necessary action than those actions which banks are more familiar with taking under E&S due diligence processes. Value chains are limited to those mapping exercises under the text and decided upon by the Steering Committee, 190 and identification of impacts in value chains is not considered as part-and-parcel of a bank's human rights due diligence processes under Article 4, which limits itself to actions to be taken with respect to the client. 191 Termination of relationships is mentioned as a possible action where a client is repeatedly not able or willing to comply with material due diligence requirements, but this is left entirely to the discretion of the bank. 192 Remediation is referred to under the article 'Enabling Remediation', which focusses on banks taking steps to ensure their clients provide remediation, not the banks themselves. This article also raises the question how impacts can be addressed or remediated based on the categories of a bank's involvement in human rights impacts, but omits any explanation of the categories of a bank's involvement, leaving this to further exploration by a working group, 193 which promises to incorporate guidance provided by the OECD. 194 The findings of the working group are to be integrated into the practices of the adhering banks. 195 Banks are required to have individual complaints mechanisms open to clients and third parties, but the DBA makes clear these are not grievance mechanisms as understood in the UNGPs/Guidelines; 196 as Brightwell observed, 'It looks like the banks are committing to develop complaints procedures, but only as long as they don't need to be effective.' 197 The DBA's governance structures arguably also place remediation as a lesser responsibility than the responsibility to develop a human rights policy and human rights due diligence procedures. Banks need only report on steps taken to prevent and mitigate impacts in their annual reports used by the Monitoring Committee, covering their implementation of Articles 3 (on policies) and 4 (on due diligence), but not Article 7 (on enabling remediation).

Risks to bank vs risks to people
The risks to bank paradigm also plays two roles in deciding how a bank's relationship to human rights is constituted: it measures human rights risks by how human rights related controversies might impact a bank's bottom line but also pays close attention to any adverse impacts to a bank's bottom line that might result from increased human rights performance. Generally, the DBA looks to risks to people as the basis for what action should be taken: The adhering banks understand that human rights due diligence goes beyond assessing and addressing material risks to the bank itself, and seek effective ways to introduce inclusion of assessing and addressing human rights risks to the rights holders by their clients in the banks due diligence and/or engagement procedures, prioritising high-risk sectors and regions. 198 Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational

Corporations for Human Rights Abuses
With regard to the premise that existing E&S due diligence practices can subsume human rights due diligence, it recognises that the need to find effective ways to deal with risks to people goes beyond existing procedures. It does not go as far as the 2013 Thun Group paper, which stated that a gap analysis would be important to understand to what extent human rights due diligence can be subsumed within existing E&S due diligence and research on the extent to which new processes may be necessary, instead working on the assumption that human rights due diligence can be integrated into existing processes.
This reliance on human rights impacts on people is not limited to banks' human rights due diligence but also extends to banks' human rights reporting, with banks expected to report under the UNGPs Reporting Framework. This framework requires that business enterprises assess human rights risks in relation to people, expressly rejecting any assessment based on material risks to business. 199 There are two subtle ways in which risks to bank paradigms do enter the DBA's assessment of what action banks are required to take however. The first, is that the DBA does, at certain points, reference other standards which do incorporate, if they are not entirely based upon, a risks to bank (or 'materiality') lens as a means to decide what actions a bank should take, such as the IFC's Performance Standards 200 and the Equator Principles. 201 The second, in those situations where banks decide they should terminate a relationship on the basis of human rights grounds, the client's failure to meet due diligence requirements is understood as failure to meet 'material due diligence requirements'. 202 The second connotation of a risks to bank paradigm, that the costs to the bank of increased human rights performance be considered a prevalent concern, can also be considered to arise throughout the DBA, particularly in the references to the need to produce a 'level playing field' internationally. This is mentioned in the preamble, 203 in relation to the establishment of effective due diligence systems, 204 and in relation to remediation, 205 as well as in an obligation on the Dutch Government to contribute to a level playing field through encouraging banks from other EU and OECD Member States to also act in conformity with the UNGPs/ Guidelines. 206 This concern that privately set norms may place businesses committed to being compliant with them at a competitive disadvantage compared to businesses which are not required to meet similar standards of conduct is a familiar concern in relation to other initiatives, including the Equator Principles.
The Equator Principles is a global transnational private regulatory regime which covers 91 institutions in 37 countries and proclaims to cover 70% of project finance in emerging markets, although the accuracy of this letter figure has been questioned. 207 It dwarfs the DBA in terms of global market share. Yet, even under the Equator Principles, financial institutions have faced the challenges of competing with actors who are not members, particularly in emerging markets such as China, India and Russia where their position is not so dominant. 208 Meyerstein argues to depend on their ability to instil compliance on core financial institutions that cover the bulk of finance. 209 Others have highlighted that the Equator Principles' 'success' lies partly in its record of assisting financial institutions with handling human rights and environmental risks in relation to their financing activities, bringing about safer investments and less reputational risks. 210 The Dutch Government also has an essential role to play in ensuring that Dutch banks meet the requirements of the DBA, particularly through market mechanisms: it is required to incentivise 'responsible Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses business conduct in general through, inter alia, public procurement and trade missions'. 211 Through utilising the DBA as a means through which to evaluate Dutch banks for the purposes of procurement, trade and other possible advantages within the Dutch market, the Dutch Government can ensure that domestic markets incentivise positive human rights performance even when global markets are less obliging.
At no point is it suggested that banks' compliance with the DBA is contingent on them not being placed at a competitive disadvantage as a result. However, this need to reduce any adverse effects of compliance with the DBA on adhering banks' bottom lines is recognised in the need for parties to the DBA to promote certain aspects of it abroad. 212

Transactional due diligence vs human rights due diligence
As discussed above, a key criticism of banks' understanding of their responsibility to respect in the 2017 Thun Group papers was that the steps a bank should take cannot be decided on an a priori basis, for instance through specific policies delineated by the type of financing, but will depend on the context. 213 A second is that there is insufficient attention paid to the actions a bank should take following the transaction.
The DBA recognises this need for context sensitivity, acknowledging that 'there might not be an 'one-size-fits-all' approach for how individual banks engage with their clients on preventing and addressing adverse human rights impacts' and basing the kinds of actions to be taken more clearly on severity rather than the nature of the transaction. 214 The due diligence provisions do not state that actions to identify, prevent and mitigate human rights impacts are limited to the lead up to the transaction, and they expressly require the banks to monitor, track and assess the outcomes of their due diligence, taking action where relevant to increase their leverage and (at the bank's discretion) terminate a relationship where this fails. 215 It also stresses the need to look to creative means of exercising leverage beyond engagement with the client. 216 Thus, the DBA accepts the need for more than a transactional due diligence approach, adopting an understanding closer to the human rights due diligence approach in the UNGP/Guidelines.

Client-only engagement vs stakeholder engagement
Client-only engagement plays a dominant role in the DBA. All mention of meaningful consultations with stakeholders is the duty of the client, with banks expected to influence clients to do so, rather than also the responsibility of the bank. The DBA itself seeks to collaborate with the OECD, but does not expressly set up platforms for engagement with rightsholders. Inclusion of the CSOs and unions which are party to the DBA is itself seen as stakeholder engagement, and the CSOs and unions are expected to empower ways of interaction with affected stakeholders, 217 but this requirement is placed only on CSOs and there is no equivalent requirement on the adhering banks. Thus inclusion in the process from external stakeholders seems to depend on representation through these parties. There is no grievance mechanism set up under the DBA which is open to external stakeholders to complain about its implementation, despite this being required under the UNGPs. 218 As the DBA seeks to regulate activities abroad from the perspectives of Dutch actors, this raises serious concerns over its output legitimacy amongst the intended beneficiaries, and may be seen as a form of imperialism, enforcing culturally or economically inappropriate standards. 219 Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses

Knowing and withholding vs knowing and showing
Client confidentiality plays a dominant role in banks' reasons for not disclosing their human rights performance. 220 The DBA states that all of the commitments of the banks are subject to law 221 and disclosure of banks' actions in pursuance of the DBA's provisions is limited by client confidentiality and banks' internal policies. 222 It also states that banks should engage in being as transparent as possible and that the parties should discuss options for greater transparency. 223 The first and only public output of the DBA to date was legal advice on what forms of information banks could share with the CSOs and the public. It did not limit itself to law, including provisions of loan agreements that banks enter into and market practice. 224 It concluded that individual client information cannot be discussed unless there is the consent of the client or this information is released in compliance with a statutory obligation of disclosure, unless it is made anonymous. 225 This is likely to significantly impact what banks can 'know and show' in the course of their activities and in their reports under the DBA. The discussions under the DBA are to be made in accordance with a confidentiality protocol, and, at the time of writing, there have been limited public updates as to its progress, despite many deadlines for public output under the DBA outstanding. 226 This has important implications for how banks account for the actions they take under the DBA, how their actions are monitored and how the provisions of the DBA are enforced.

The potential of the Dutch Banking Agreement: Regulation, experimentation or advocacy?
The DBA could be argued to serve multiple objectives. First, it can be seen as regulation: it seeks to ensure adhering banks meet their responsibility to respect human rights, with a focus on their project finance and corporate lending activities, through the creation of new rules. 227 Second, it can be seen as a form of experimentalism: creating new and creative means for banks to address problems that they cannot address alone, which have not necessarily been tried before. Third, it may delve into advocacy: attempting to pursue a normative agenda to be taken up at the international level. It is argued in this section that, while it is too early to tell how the DBA will perform in practice, it could pursue all three of these agendas.
Regulation through rules is a key means of transnational private regulation. 228 The DBA does create some rules in respect of actions similar to those many banks already undertake under their existing E&S practices, often referencing standards in project finance which the banks with the most project finance exposure are already committed to meeting. 229 The DBA's ability to enforce these rules may be affected in practice by the failure to ensure there is structural separation between standard setting, monitoring, and enforcement in its governance structures (see Section 3.1). Furthermore, client confidentiality may present an obstacle with respect to the DBA monitoring, and the responding to, specific cases of non-compliance (see Section 4.5).
Where the DBA does not enforce rules, it looks to experimentalism. Whereas the development of rules depends on convergence and well definable benchmarks, experimentalism is a response to the opposite: 'a situation in which actors do not know their precise goals or how best to achieve them ex ante, but Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses must discover both in the course of problem-solving.' 230 Experimentalism seeks to adapt general goals to specific contexts rather than imposing one-size-fits-all solutions through co-ordinated learning from local experimentation. 231 Goals and the means for achieving them are conceived as provisional and are subject to revision in the light of experience. 232 In its most ideal form, experimentalism seeks to identify broadly defined goals and metrics which are only provisionally established, with the actors under the framework (in this case the adhering banks) being offered autonomy in how they pursue these goals, but at the same time reporting regularly on their performance and participating in a continual peer review of the actions they have taken. 233 While the actions taken are initially reviewed against the preliminary goals and metrics, these are then periodically revised in light of the response to the problems and new possibilities revealed through the reviews of earlier efforts. 234 This process is cyclical, with new provisional goals and metrics then designed for the next iterative cycle. 235 The inclusion of multiple stakeholders, including CSOs, as agenda setters and/or problem solvers has been seen as crucial to the success of such projects, mitigating the chances of the initiative being captured by the banking sector's interests. 236 While effective private initiatives have often been perceived to depend on a broad convergence of interests, values and beliefs among the parties, experimentalist regimes are often employed where there are prevalent interest and value conflicts between the parties. 237 There are many analogous traits between the DBA and experimentalism, including its iterative nature, broad framework goals, deliberative decisionmaking and emphasis on learning. The DBA itself is not expected to exist in perpetuity, ending after three years of implementation. 238 It can be renewed based on input from the final report from the DBA's Monitoring Committee, which will review its achievements, obstacles for implementation and any solutions to these obstacles. 239 The Steering Committee can set up new projects, 240 and the parties are expected to engage in the process of constructing new agreements following the first report by the Monitoring Committee, based on those topics under which such additional agreements would be most effective. 241 The DBA thus expects a reflexivity also expected under experimental governance.
Where the DBA is not enforcing rules, it is often pursuing broad framework goals, either based on concepts under the UNGPs such as remediation or increasing leverage or based on even more general goals such as sustainable development or having a positive impact on people's lives. 242 The means to implement these goals is either through adhering banks exchanging and learning from best practices 243 or new tools and studies to assist banks with their human rights performance. 244 One key area where there may not even be a broader framework goal is with respect to remediation, where there is still no express recognition that banks themselves have a role in directly providing remediation. Instead of proposing that studies or experiments be carried out as to how banks can remediate impacts in practice, the DBA creates a working group to further explore how the UNGPs/Guidelines apply in this area. 245 To the extent that the working group is based on deliberation between the parties and experts, this will 'force those around the table to be articulate in their choice of policies and ways to obtain their goals' Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses thus creating some form of checks and balances in terms of the output reached. 246 At the same time, if the parties are unable to agree on at least a broadly defined goal that banks should remediate impacts in certain instances, the DBA will fail to improve banks' performance with respect to direct remediation.
The DBA itself also has opportunities for advocating for better normative standards in other global initiatives. It is not unknown for Dutch banks to advocate for (positive) change with respect to human rights initiatives and regulations. ASN Bank and Rabobank signed onto a letter urging the Dutch Government to pass mandatory extraterritorial laws to challenge child labour abroad. 247 Four of the five Equator Principles banks adhering to the DBA -Rabobank, ABN AMRO, FMO, and NIBC -signed a letter urging more universal, geographical coverage of those activities regulated by the Equator Principles. 248 Adhering banks have also been involved in creating new standards aimed at reducing adverse human rights impacts, calling on other banks to endorse them. 249 The DBA itself has provisions requiring parties to advocate for the uptake of similar norms at the international level. For instance, the outcomes of the remedy working group and the exchange of best due diligence practices is expected to feed into the OECD processes. 250 The Dutch Government and Dutch Banking Association are expected to advocate for norms in the DBA to be adopted abroad. 251 The OHCHR cited the DBA in its response to the Thun Group and is seen by some as having the potential to play a much more promising role than the Thun Group in bringing about better approaches to the UNGPs in the banking sector. 252 Ruggie described the social constructivist view of the development of norms as the basis for norm generation in business and human rights. 253 He argues that for norms to be adopted by relevant institutions (and then implemented), there is a need for norm entrepreneurship. Those proposing a new norm (the DBA parties) should not only ensure that that norm is created with the appropriate information and expertise but also ensure that the norms are promoted through public advocacy, including demonstrating the appropriateness and effectiveness of the norms in light of the relevant institutions' missions and objectives. 254 Finding alignments between the three key types of actors in banks and human rights discourse -banks, civil society organisations and Government -could lead to an opportunity to produce credible and influential norms on the international stage. The DBA's success across these paradigms will depend on the potential for its governance structures to enforce those aspects of it which are enforceable; the extent to which parties are initially willing to invest in the experimentalist aspects of the DBA and then the extent to which these initiatives display a potential of success from the differing perspectives and interests of the banks, Government, unions and CSOs and the extent to which the DBA's normative deliberations are successful in improving international discourse surrounding the implication of human rights standards for banks.

Conclusion
The Dutch Banking Sector Agreement represents a novel experiment in regard to ensuring increased human rights performance of banks abroad. Rather than extraterritorial enforcement via hard law, such as reporting requirements, extraterritorial jurisdiction or the creation of a corporate duty of care over events abroad, the Utrecht Law Review | Volume 14 | Issue 2, 2018 | Special Issue: Accountability of Multinational Corporations for Human Rights Abuses Netherlands has favoured a softer approach, rooted in multistakeholder collaboration. The DBA represents the first multistakeholder initiative in banking and human rights and benefits from a stronger alignment with the UNGPs/Guidelines than previous industry understandings of the relationship between banks and human rights. In particular it recognises a greater responsibility for banks for human rights impacts related to their financing activities, that human rights impact prevention is about risks to people not risks to the bank, and that actions a bank should take in their due diligence processes covers much more than the events leading up the transaction, as is the focus of environmental and social due diligence. At the same time, it is still not fully coherent with the UNGPs/Guidelines, particularly with respect to remediation and termination of business relationships, it does not properly recognise the need for banks to engage in stakeholder engagement in how they take actions to respect human rights, and it will most likely be adversely impacted by laws and banks' practices with respect to client confidentiality.
The DBA can play three roles. First, it sets certain standards which regulates adhering banks' behaviour. These standards are prevalent where there is greater convergence between banks' existing E&S due diligence procedures and the UNGPs/Guidelines. However, there are questions as to the role that conflicts of interest between the different governance bodies and functions of the DBA, as well as the fact that banks are unlikely to share information around particular cases of adverse human rights impacts, will adversely affect the enforcement of the DBA. Second, for aspects of the UNGPs/Guidelines which are less likely to be integrated across banks' procedures (e.g., remediation or exercising leverage over value chains) or where a one-size-fits-all approach is not appropriate, the DBA favours experimentalism, focussing on exploring new means of meeting these more broadly defined goals. The success of this experimentalism will depend on the achievements of the different projects set up under the DBA, and whether they are successful from the perspectives of the different parties. Third, the DBA may engage in advocacy, introducing a multistakeholder perspective on the applicability of the UNGPs/Guidelines to banks, and encouraging other banks to adopt a more ambitious understanding of their role with respect to human rights.
The future success of the DBA will depend on the extent to which banks are able to transform their human rights policies, due diligence processes, and procedures with respect to remediation to further integrate the UNGPs/Guidelines. Ultimately, only time will tell whether this initiative will improve the human rights performance of banks in practice.