Securing human rights through risk-management methods: Breakthrough or misalignment?

Abstract This is a study of three authoritative instruments that promote a common idea: economic activities and development should be conducted with respect for human rights. The World Bank Framework, the International Financial Corporation Performance Standards and the UN Guiding Principles on business and human rights are examined to get clarity on how human rights risk management differs from more conventional management approaches. The focus here is on prevention of human rights impacts. Do the three instruments employ approaches adequate for handling human rights risks? To understand prevention, one needs to reflect on what makes human rights a particular type of impact and account for the regulatory context of protecting human rights transnationally. The analysis identifies four ‘offsets’ through which economic decision-makers can distort their human rights performance and place causal observers at a disadvantage. Prevention becomes an issue of how to relate to ‘residual impacts’ on human rights. This article finds that the ‘hierarchy or mitigation’ and even ‘human rights due diligence’ under illuminate the challenge. The proposal here is to add ‘reduction at source’ as a parameter of human rights risk management. The sources for this analysis are the three instruments, and the practice of implementing organizations, particularly IFC projects, CAO cases, impact assessments, and CSR reports. In conclusion, the potential for cross-fertilization among instruments is genuine. Increased clarity on prevention of human rights impacts should assist economic decision-makers in their risk management task and casual observers in assessing their performance.


Introduction
The World Bank's Environmental and Social Framework (ESF), 1 the International Finance Corporation's (IFC) Performance Standards (PSs), 2 and the UN Guiding principles on business and human rights (UNGPs) 3 are three authoritative instruments that promote a common idea: economic activities and development should be conducted with respect for human rights. In their current form, the instruments are relatively new. The UNGPs were endorsed unanimously in the UN Human Rights Council in 2011. The IFC revised its PSs in 2012 and aligned them with the UNGPs on the idea of corporate respect for human rights. 4 The World Bank redesigned its policies in 2017, resulting in the new ESF with a clearer provision on human rights. 5 This signifies a convergence on human rights in the economic area not witnessed before, even if merely a matter of international soft law. On the one hand, the World Bank was shying away from human rights as interfering with its development promotion mandate, 6 and on the other hand, the UN failed to deliver a legal framework for the social responsibilities of multinational enterprises despite commencing efforts in the 1970s. 7 Human rights have traditionally been the purview of public international law laying down obligations of states to ensure human rights, and of national constitutions laying down fundamental rights at the top of domestic legal orders. Financial institutions 8 and multinationals 9 have long been scrutinized for their negative impacts on societies and environment, especially as global economic integration gathered speed. 10 Now human rights have become more accepted as relevant in the decision-making processes of financial institutions (WB Group) 11 and business enterprises (UNGPs). In this new paradigm, economic decision-makers are expected to operationalize human rights protections through risk management methods. Indeed, the UNGPs ask companies to undertake 'human rights due diligence' 12 while the WB Group commits to do its own due diligence when granting loans 13 and asks its clients to implement management plans. 14 That entails businesses and development projects have to take both preventive and corrective measures to identify and address risks and impacts.
This article is concerned with the prevention of human rights impacts. It seeks to understand what preventive measures are appropriate in order to say that human rights are respected. The focus on prevention of human rights impacts can be justified in several ways. The six-year UN mandate of John Ruggie leading to the UNGPs was driven by a desire to shift the emphasis in human rights protection towards prevention 15 and not ending up relying disproportionately on judicial procedures dealing with impacts ex post. 16 The WB Group also indicates that prevention is preferred to reparations through its 'hierarchy of mitigation'. 17 More broadly, maybe prevention should be especially emphasized when the risk of harm concerns human dignity which is what human rights seek to protect 18 and which seems to make such affront inherently difficult to repair or 'remediate'. Indeed the human rights system as a whole has evolved to emphasize 4 IFC, PS 1, supra note 2, para. 3. 5 A Vision for Sustainable Development, in ESF, supra note 1, para. 3. prevention more strongly. 19 Finally, there is also something to be said about the particular context where rights are discussed: holding transnational economic decision-makers accountable is often impossible or at least extremely difficult under available legal frameworks. 20 In such a context, prevention is at risk of becoming flexible and optional when the threat of liability is not there to incentivize preventive action and deter excessive risk-taking.
Thus, getting clarity on prevention of human rights impacts acquires urgency in this regulatory context and for this specific type of harm. The questions asked herein are: Is the prevention of human rights impacts explained compellingly in the three instruments? Are the risk management approaches employed in the three instruments able to handle the human rights impacts, with their particularities and regulatory context outlined herein? Is there potential for cross-fertilization between WB Group and UNGPs approaches to risk management? Increased clarity on prevention of human rights impacts matters in two main ways: it shapes the risk management choices available to decision-makers and it enables observers to assess performance and thus hold decision-makers accountable.
The article places emphasis on primary sources such as the three instruments and explanatory materials, and the practice of implementing organizations (particularly the IFC). More specifically, the article draws on a sample of Compliance Advisor Ombudsman (CAO) cases and IFC projects, human rights impact assessments (HRIAs) of companies and several guides for conducting such assessments, and corporate social responsibility (CSR) reports. 21 In terms of structure, the article continues with a second section that explains the problem of prevention in a specific human rights and regulatory context, and how the WB Group and UNGPs struggle to offer a compelling answer to the problem flagged here. The third section proposes to add a new parameter to human rights due diligence and offers a different way to think of prevention and avoidance of human rights impacts. The conclusions spell out implications for the WB group and the UNGPs and propose that cross-fertilization between the two systems is possible.
2. An analytical framework for the problem of prevention By putting the spotlight on several elements that shape human rights risk management, this section outlines an analytical framework for the study of prevention of human rights impacts. Two main themes are pursued to explain what is specific about the regulatory context where risk management measures to protect human rights will be taken, and what is the puzzle around the prevention of human rights impacts. The WB Group and UNGPs appear to face different challenges when it comes to the management of human rights impacts. On the one hand, the WB Group is focused on social impacts but might not fully grasp that human rights impacts are not a mere species of social impacts. On the other hand, the UNGPs rely on many stakeholders to assess corporate performance and press for improvements. However, stakeholders are at a disadvantage and can be overwhelmed, confused, and mislead by corporate risk management information. In short, as the two systems face different problems traceable to insufficient reflection on this specific context, these problems are most difficult at the prevention stage. There is, however, some potential for mutual enrichment on the issue of risk management, especially prevention, that this article seeks to establish.

Particularities of context
The context of protecting human rights in economic operations will be discussed in two ways. First, the context is of a global economy that has increasingly integrated through international production and distribution systems 22 but remains wrapped in a porous legal framework; the latter struggles to match the mobility and complexity of global value chains and secure human rights effectively. Second, there is the particular institutional context of the UNGPs and WB Group analysed herein that shapes the different ways in which they can incentivize responsible conduct and protect human rights.

Global economic and regulatory context
During and after his UN mandate, Ruggie provided evidence-based and insightful analysis about recent dynamics in the global economy bearing on how to enhance safeguards of human rights. 23 The EU emphasized the rapidly-changing global economy landscape as global value chains have become a dominant form of production and distribution. 24 This realization is backed by data and analysis on new patterns of trade and investment supplied by the UNCTAD. 25 The International Labour Organization also dedicated its 2016 annual conference to the issue of labour rights in international supply chains. 26 The eagerness of developing and developed states to integrate in the global economy transpires from the growing number of trade agreements and investment agreements they conclude. The rise of the Chinese economy and of other developing countries is transforming received ways of thinking about development and human rights. These shifts influence regulatory options and require new understandings on how to secure and promote human rights in the global economy and hold economic decision-makers accountable.
The Ruggie mandate (2005-2011) was the frontline for debates around these issues. The search for effective ways to protect human rights in the global economy transpires from Ruggie's break with established wisdom and his trademark approach of 'principled pragmatism' 27 and 'polycentric governance' 28 as a systemic way to shape transnational business conduct. The search for innovations is also discernable in human rights law attempts to develop the extraterritorial obligations of states. 29 The World Bank's shift towards a greater reliance on the local frameworks of borrowing states denotes a similar concern for adaptation to new realities. 30 The IFC also revised its safeguards to make direct references to human rights and the UNGPs and bring itself in line with the latest developments in CSR and supply chain management. 31 The bottom line is that a strong, legal framework is yet to emerge to hold economic decisionmakers accountable for infringing human rights. Indeed the three instruments here are international soft law, although the WB Group has genuine leverage over its clients and its standards become legally binding upon entering a contractual relationship. 32 The efforts to devise a legally binding instrument for transnational companies continue today 33 due to successive failures to reach agreement in the 1970s and early 2000. 34 But the difficulty of matching global value chains with statist regulation seems daunting. Studies have drawn attention to the fragmentation of international law. 35 This places doubt on the feasibility of aiming for an encompassing human rights legal framework at the apex of the international economic decision-making. Instead increased interest goes to 'transnational law' treatments 36 that account for the variety of rules that shape economic decision-making. 37 Regarding business operations, 'regulatory governance' scholarship explains this diversity of drivers and methodological ways forward to shape a new research agenda. 38 Procedural regulations such as mandatory impact assessments point to a potential technique originally developed in environmental protection, 39 but that spread to cover social assessments in development finance, 40 and even human rights assessments in trade agreements 41 and corporate practice. Indeed, a broader procedural emphasis on due diligence as a comprehensive procedural approach is recognizable in many legal fields 42 and has a long history. 43 Not accidentally, both the UNGPs and the WB Group refer to (human rights) due diligence as an important operational concept 44 and their explanations about the expected procedural steps are strikingly similar. 45 However, there are concerns about these de-centralized 46 or process-oriented regulatory solutions. 47 Overly-complex arrangements with many moving parts enable businesses to exploit power differentials at different junctures. Over-reliance on procedural regulations create opportunities for symbolic compliance. Along these lines, this article flags the risk that such an opportunity exists at the stage of prevention of impacts and can be exploited particularly in a context where regulations are not stringent enough. The argument here contributes to the 'polycentric governance' scholarship by showing how insufficient clarity around prevention and avoidance of impacts maybe inhibit the evolution of the regulatory regime that the UNGPs meant to facilitate.

Institutional context of WB Group and UNGPs
Some features of the instruments that are important for the current analysis appear clearly when they are compared side by side. The WB Group has detailed standards through which it can specify more precisely the expected conduct. The UNGPs are confined to 10,000 words and identify key elements of responsible conduct in a rather generic fashion.
The WB Group has its own implementation infrastructure, an elaborated one, in place that can conduct ex ante and ex post evaluations of risks and impacts. 48 The UNGPs have no infrastructure of their ownno monitoring, no adjudicating, and no enforcement mechanismsexcept the UN Working Group of five experts entrusted with a primarily dissemination and capacity-building mandate. 49 The WB Group has the leverage of a financial institution with US$252 billion in subscribed capital, while the UNGPs lack any such financial and contractual leverage. As a result, the WB Group can concentrate on tweaking its own legal order, as it did when it revised its policies expressed in the ESF or the PSs, to make sure its lending achieves the stated developmental objectives. The WB Group does not depend on actors in its environment to exercise pressure for its safeguards to be observed: it only has to remain mindful of competition from other financial institutions and objections from its clients. In contrast, as they lack direct leverage on businesses and states, the UNGPs cannot separate themselves from their governance environment but have to plunge in that environment in order to mobilize sources of leverage from all possible public and private actors.
To protect human rights the UNGPs count on a 'polycentric governance' system encompassing various forms of regulationpublic, private and hybridspread from the global to the local levels. Indeed, the UNGPs count on facilitating developments in several governance systems which in turn are expected to pick up the baton, further specify the expected conduct, and apply their leverage on businesses to improve their performance. By necessity, the UNGPs have to facilitate a multi-layered, de-centralized regulatory order on which they depend for their success. The UNGPs rely on persuasion and clarity as the only leverage they have to secure a wide degree of stakeholder support while the WB Group can rely on itselfits financial leverage, expertise, and infrastructure. Therefore, it is essential for the UNGPs to deliver conceptual clarity on issues of human rights risk management; that, however, requires closer attention to how human rights impacts might differ from social impacts.

Particularities of preventing human rights impacts
The current explanations of prevention suffer because what makes human rights impacts a special category has not been sufficiently explored. On the one hand, human rights impacts can be seen as a species of social impacts. Prevention of human rights impacts appears, then, rather self-explanatory and would benefit little from further elaboration of risk management in abstract; the prize would be on further specification, guidance and mechanisms to secure application in practice. The WB Group, and even the UNGPs to some extent, follow this approach. On the other hand, there might be something different about human rights impacts that problematizes risk management choices and has wider ramifications that have been insufficiently discussed. 48 The WB Group has two complaint mechanisms: the Inspection Panel for the Bank and the Compliance Advisor Ombudsman (CAO) for the IFC. Both conduct independent investigations in response of allegations of harm from affected stakeholders and issue recommendations to the management of the WB Group. 49 Human Rights Council, Human rights and transnational corporations and other business enterprises, A/HRC/17/L.17/ Rev.1, 2011.

First challenge: Human rights impacts distinguished from social impacts
The OHCHR defines adverse human rights impacts as impacts that occur 'when an action removes or reduces the ability of an individual to enjoy her or his human rights'. 50 Notably, the UNGPs refer specifically to human rights impacts, while the WB Group speaks of social impacts. Interestingly, the ESF does not define social impacts but chooses to illustrate them with an enumeration, 51 while the IFC defines them rather self-referentially as impacts on communities and workers. 52 In this enumeration and throughout the WB Group instruments there are references to, or overlaps with, human rights standards such as child labour, forced labour, health and safety, security issues, and land dispossession. 53 Actually, the IFC pondered the relation between social impacts and human rights and deemed that human rights risks 'can be effectively addressed through social and environmental considerations'. 54 All three instruments envisage proactive and reactive actions to manage risks and impacts. 55 In sum, a conceptual explanation of social and human rights impacts is absent from WB Group instruments; furthermore, in both the UNGPs and WB Group systems attention quickly moves to operational issues of risk management.
Prevention of human rights impacts, to stay with the OHCHR definition, means 'actions taken to ensure such impact does not occur'. 56 Also related to prevention, the Guide adds another conceptmitigationas it speaks of mitigating adverse impact which 'refers to actions taken to reduce its extent, with any residual impact then requiring remediation'. 57 By comparison, through its mitigation hierarchy the WB Group asks its clients to avoid impacts and 'where avoidance is not possible, minimize or reduce risks and impacts to acceptable levels' and compensate residual impacts. 58 The message is to 'favor the avoidance of impacts over minimization'. 59 But further specificity on preventive measures is necessary to operationalize these rather abstract concepts and general goals. In this respect, the WB Group instruments offer detailed requirements and extensive guidance on various specific measures a client should put in place to prevent impacts and manage risks diligently. By contrast, the UNGPs remain at a generic level and count that further specification of expected conduct (e.g., by issue, industry, geography, and profession) will emerge as the polycentric governance system evolves.
This boils down to a picture infused with common sense saying that negative impacts should be avoided and risk of harm be minimized through reasonable preventive measures suited to the particularities of the operational context. 60 This article asks whether social impacts and human rights impacts should be conceptually separated more. 61 Clearly a project can have a negative effect on people by introducing a change in their life: they have to move to another place, the environment is degraded with wide-ranging impacts on people's livelihoods, their way of living changes with the presence of the project, and many other nuisances. 62  among various categories of negative impacts, not all individual and social interests are protected as human rights and suffering a harm is not automatically an infringement of human rights. Human rights are derived from human dignity 63 and they are commonly understood to be minimum levels 64 of treatment to live a life with dignity. 65 Tolerating residual impacts is conceivable for social impacts, as the WB Group expressly indicates. 66 IFC explained residual impacts as 'the impact that is predicted to remain once mitigation measures have been designed into the intended activity'. 67 With human rights such toleration means going under a minimum threshold of human dignity which would be inherently problematic; therefore the OHCHR Guide says that 'any residual impact' requires remediation. 68 So a key difference between social and human rights impacts regards tolerance of residual impact. That has direct implications on remediation (ex post stage) as the OHCHR indicate, but also has an insufficiently understood bearing on preventive measures (ex ante stage). More questioning is warranted regarding residual impact at preventive stage in order to ensure that preventive strategy and effort are adequate, i.e., commensurate with the specific nature of human rights impacts as opposed to social impacts.
Insisting that human rights set minimum levels of treatment does not mean that a certain minimum level of human rights enjoyment can be established in abstract 69 and thus facilitate mechanical determination of (ir-)responsibility. 70 Nor do minimum levels mean that human rights solely depict the bare minimum safeguards to which individuals are entitled 71 with no bearing on wider societal problems (i.e., societal distribution of risks, inequality, poverty, environmental degradation) 72 that can, indeed, be discussed and contested in human rights terms. 73 Rather, insisting on minimum standards of treatment for a life with dignity as elaborated in international human rights law signifies that human rights impacts are not a mere specie of social impacts; instead they do benefit from an increased institutional level of agreement among states and therefore require particular attention in order not to be watered down through inadequate protective measures. Beitz, supra note 64, at 29-30. 74 Scheinin analyses various approaches to 'identifying a core within the normative framework of human rights law' that should be inviolable. He endorses views about the 'essential core' of a human right and the related approach of core obligations of the state regarding human rights, and presents methodologies for how to operationalize these 'core' approaches to human rights protection. M. Scheinin, 'Core Rights and Obligations', in D. Shelton (ed.) The Oxford Handbook of International Human Rights Law (2013), 527.
The WB Group and the UNGPs state that negative impacts should be avoided. There is, thus, full clarity at the general level of principle. However, the explanation of risk management, which operationalizes the avoidance aim, seems incomplete. The message is: Avoid infringements of rights and breach of obligations, seek to eliminate negative impacts ex ante, work diligently to reduce/minimize risk of harm, and ensure you repair residual harm ex post (at least for human rights if not for social impacts). But if human rights are a particular type of impact, is it enough to say reduction of risk to human rights? Is it sufficient to ask for remediation for the impact that was not prevented? Can more be said about what preventive effort is adequate? The thesis here is that there might be a looseness in how prevention and avoidance of harm are explained in the three systems. Because human rights have this particularity of non-tolerance to residual impact, further elaboration of prevention and risk management is warranted.
Just because no preventive system is fool-proof and capable of avoiding all impacts, the result cannot be that some residual impact should be acceptable ex ante as a fact of life, a matter of feasibility and costs, as a constraint that real life puts on high aspirations. An explanatory gap remains. Thus, on the one hand, the UNGPs made clear no residual impact can be tolerated ex post (remediation stage) but are silent on residual impact ex ante (prevention stage). On the other hand, the WB Group, with their focus on social impacts, is not close to recognizing the problem either; it seems to tolerate even significant social impacts and even allow for no remediation ex post. The puzzle, then, is how to explain non-tolerance of residual impact ex ante without setting bright lines that lack feasibility in practice, e.g., full elimination of risk? Indeed, explanation of prevention should not collapse into unfeasible or unrealistic expectations that all human rights risks can and must be eliminated ex ante through due diligence because human rights are different from 'mere' social impacts and reflect 'minimum' standards of treatment.

Second challenge: Four offsets and the casual observer
By accounting for the regulatory context and contrasting the UNGPs and WB Group systems, the argument so far has been that there is insufficient clarity on prevention which, in turn, hinders 'casual observers' from performing their crucial role in the polycentric governance regime envisaged by the UNGPs. Casual observers refer here to the various stakeholders that track the performance of economic decision-makers; thus, these observers are not adjudication mechanisms but more-or-less informed stakeholders struggling to assess whether businesses do their due diligence and respect human rights. The casual observer might be consumers, investors, NGOs, public authorities contracting with business, the media, all sorts of intermediaries and analysts including academia; even businesses tracking the performance of peers and competitors. The casual observer tracks the performance of businesses (pillar 2 in the UNGPs) and of states and international governmental organizations like the WB Group (pillar 1).
During his mandate, Ruggie sought improvements to a global human rights order marred by governance gaps and subject to major tensions. Sceptical about the feasibility of straightforward regulatory solutions, such as a treaty to be ratified and implemented by states under the oversight of the UN, Ruggie placed faith in 'polycentric governance'. It is a more complex system due to many moving parts, relying on multiple actors to exercise leverage over businesses. The casual observer is, thus, at the heart of the polycentric governance system and essential for the UNGPs ambition to facilitate the evolution to more institutionalized forms for protecting human rights. In contrast with the UNGPs, the WB Group system does not suffer from casual observer problems: it can rely on its own leverage and infrastructure to ensure its requirements are observed.
While understanding how to manage social impacts does benefit from the detailed requirements laid down by the WB system, how to manage human rights impacts is a new area for all stakeholders. Guidance has emerged and is moving human rights management from generalities and aspirations to the level of specificities and implementation. However, a business describing its human rights due diligence efforts can quickly overwhelm and confuse the casual observer through a combination of high detail and patchy information. The casual observer might not have sufficient information, the right type of data, or data presented in an accessible form to comprehend the company's impacts and measures taken. To judge levels of commitment and risk management effort, the observers have to keep track of process requirementsdetailed and technicaland acquire specific knowledge of operational contexts, industries, countries, and human rights issues. Furthermore, companies seek to improve their image, including by downplaying weak performance and highlighting instead areas where they do better, or where they can communicate accomplishments more easily. Companies are, thus, in a position to overwhelm observers and they can 'offset' an irresponsible behaviour with a commendable one, that is, compensate harms in one area (e.g., environmental degradation) with good deeds in other areas (e.g., efforts to tackle child labour). These factors make the task of evaluating the management measures companies have taken quite challenging. A way to cut through the above difficulties is to remain mindful of potential 'offsets'.
Four such offsets that are specific to human rights management are identified herein. They show the complications accompanying the management of human rights risks and how casual observers could be outmanoeuvred by deliberately or unintentionally misleading corporate narratives. First, good deeds might offset negative impacts. Contributions to a country's development, economic and social benefits for stakeholders, philanthropic acts, and any other commendable deeds are important positive impacts of business activity, but can 'focus the attention of the general public' the wrong way. 75 The UNGPs explained this as an unacceptable offset: businesses activities 'may support and promote human rights, which may contribute to the enjoyment of rights. But this does not offset a failure to respect human rights throughout their operations'. 76 Second, remediation might offset prevention. But deficient preventive effort cannot be excused through reparatory measures, should the risk of harm materialize. Companies should not recklessly take human rights risks and make insufficient efforts at the prevention stage. In other words, the ex post stage of impact management could displace the ex ante measures. This offset is signalled by the emphasis the WB Group and UNGPs put on prevention. Indeed, the WB Group's hierarchy of mitigation prioritizes avoidance and reduction of risk over remediation 77 and the UNGPs' due diligence is a proactive approach meant to prevent harm from occurring in the first place. Given the imperfections of transnational and local regulatory frameworks, the weakness in redress mechanisms 78 reinforces reckless risk-taking on top of not offering remedies for harms. The UNGPs are explicit that all impacts must be remediated or else a right has been infringed; all residual impact must be remediated, the OHCHR Guide says. 79 In contrast, the WB Group is equally explicit that it allows the offset of negative impacts by positive ones and it can even tolerate residual social impacts as some impacts might not need to be remedied; this is the lowest layer of the mitigation hierarchy. 80 This contrast between the UNGPs and the WB system points to insufficiently explored differences between social and human rights impacts.
Third, there could be an offset at the prevention stage, and it regards the reduction of risks to human rights. This offset is acknowledged in the OHCHR, which states that mitigating human rights risks requires reduction in likelihood. 81  OHCHR Guide, supra note 50, at 7 ('The mitigation of human rights risks refers to actions taken to reduce the likelihood of a certain adverse impact occurring.'). a reduction in harm (magnitude of damage) while probability of harm (likelihood) remained the same is not consistent with human rights due diligence. This would be an offset between magnitude and probability, which the OHCHR precludes. However, such offset is accepted in conventional risk management as a way to reduce risks in either dimension 82 and used by the IFC to manage social risks. 83 The explanation for the OHCHR stance derives from the specific character of human rights impacts: they do not accommodate residual impact. Indeed, allowing constant probability of impact while only the consequences were reduced through mitigation measures would have actually meant acceptability of some residual impact as the normal cost of doing business. With this the OHCHR Guide takes a step forward on human rights risk management by hinting that residual impact is intolerable not only ex post, but also ex ante. However, this clarification arguably still allows too much flexibility on prevention.
Fourth, there is another offset that so far has been overlooked in UNGPs materials and it could also happen at the prevention stage. This has to do with the capacity of preventive measure to eliminate risks as not all preventive measures are the same. Some could only reduce risk while other measures might be capable of eliminating it altogether. Some measures may tackle deeper causes of human rights impacts while others only address symptoms or more superficial causes of impacts. It is important to insist that superficial mitigation measures should not offset feasible measures able to tackle deeper causes of impacts. The capacity of preventive measures to eliminate rather than simply reduce the risk should be tracked. There should not be an offset between reduction and elimination of risk. This is the offset highlighted in this article.
The UNGPs recognized the first three offsets and with this began clarifying the particularities of human rights impacts and their risk management. The fourth offset has not received attention in human rights scholarship, but somehow, surprisingly, it could be in the WB system that one could find useful insights. However, the UNGPs inadvertently produce some confusion through the way they explain due diligence, which is the operational concept designed to handle these offsets.

The UNGPs treatment of prevention: An unnecessary complication
The UNGPs put forward the responsibility to respect that says a company has to 'avoid infringing' human rights and 'address adverse human rights impacts'. 84 In fact, the UNGPs are aligned with the WB Group as they both share the aim of avoiding negative impacts. The WB Group speaks of the 'mitigation hierarchy' which places avoidance on the top layer as clients are asked to avoid-reduce-compensate harms. 85 Thus, the UNGPs are conceptually geared towards avoidance of impacts through the principle in the responsibility to respect. The operational concept of due diligence says that businesses should 'prevent' or 'mitigate' adverse impacts 86 and thus introduces two concepts relevant ex ante: prevention and mitigation. The OHCHR Guide defines prevention as 'actions taken to ensure such impact does not occur' and mitigation of an impact means 'actions taken to reduce its extent'. 87 Prevention is non-occurrence and mitigation is reduction of risk.
Rather than introducing clarity this raises the question of when it is acceptable for a company to move from the prevention mode (non-occurrence of impacts) to mitigation mode (reduction)? This is particularly problematic given that UNGPs have no review mechanism and no detailed process requirements, as the WB Group system has. The UNGPs, importantly, outline four 82 Ibid. ('In traditional risk assessment, risk factors in both the consequences of an event (its severity) and its probability.'). 83 IFC, PS 1, supra note 2, para. 1 (fn. 2). 84 UNGPs, supra note 3, Principle 11. 85 See supra notes 58-9. IFC Guidance Note 1, supra note 2, GN 61. 86 UNGPs, supra note 3, Principle 17. 87 OHCHR Guide, supra note 50, at 7. components of due diligence but these still remain generic steps of risk management at a higher level of generality. Furthermore, the UNGPs identify the types of involvement of a company with adverse impactscausation, contribution, or linkagesand specify the expected conduct for each of them: cease harmful conduct, exercise leverage over third parties, and offer remediation. 88 Again, such prescriptions are important and have changed the responsibility paradigm in supply chain management. 89 However, they are rather general conceptual categories and inherently cannot give precise guidance on choosing between specific preventive measures. The problem is that outright avoidance strategies are available in a minority of situations; for the majority of situations, elimination of impact morphs into reduction of risk. At this moment an offset at the prevention stage is appearing. The problem is not the generality of human rights due diligence treatment in the UNGPs and OHCHR Guide, but that they do not seem to recognize this offset and offer further conceptual guidance.
As a result, the UNGPs under-illuminate an offset at the prevention stage. The particularity of human rights impacts among other types of negative impacts is under-explained. Despite speaking of avoidance and non-occurrence, the UNGPs invite problems as they move from defining corporate responsibility in principle to its operationalization through due diligence. Factoring in the institutional setting in which they are embedded, the UNGPs offer generic risk management guidance with no infrastructure to fall back on to further specify and compel desirable conduct and thus, in effect, offer a concept of eliminating impacts that morphs into reduction of risk. The UNGPs give in to a conventional narrative on risk management that can throw the responsibility to respect and human rights due diligence out of alignment. This diminishes clarity on prevention and reduces the persuasive power of the UNGPs, which is a key source of its leverage as a soft law instrument in a polycentric governance system.
The search for clarity could go further to see whether other instruments inspired by the UNGPs add the necessary interpretation. By design, HRIAs appear as an important tool to counter the relativization power of companies in debates with casual observers. However, one does find in HRIAs neither clarifications on what makes human rights a particular sort of social impact, nor reflections on the offset at the preventive stage. For example, a comprehensive HRIA draws on the WB's 'mitigation hierarchy' which 'fits well' with the UNGPs as it 'emphasises prevention of negative human rights impacts first and foremost, with mitigation where that is appropriate, and remediation where impacts have not been prevented'. 90 A closer look at HRIAs that follow the UNGPs and the environmental impact assessments (EIAs) generated by the WB system reveals that both types of assessments rely little on the general and abstract notions of due diligence and mitigation hierarchy. Their task is to go impact-by-impact and identify the adequate preventive and corrective measures. Just like the EIAs, HRIAs jump directly to specific mitigation measures for specific rights and can easily bypass conceptual aspects of human rights risk management. The determinant element in these impact assessments is not conceptual clarity on prevention, but the checklists 91 that deliver high precision in assessment and mitigation measures. In sum, HRIAswhether the assessments themselves or the guidance on how to undertake themdo not clarify the puzzle raised in this article. Does the WB system fare better than the UNGPs in dealing conceptually with the offset explained herein? No. The WB system bypasses the general discussion on prevention but has the infrastructure that might solve directly the offset difficulty when specific mitigation measures are decided. Indeed, the IFC has multiple layers of ex ante and ex post review to back up the 88 UNGPs, supra note 3, Principle 19 Commentary. 89 The Wilmar case revealed the IFC had an outdated view on supply chain management. See 'Indonesia / Wilmar Group-03/Jambi', filed 9 November 2011, available at www.cao-ombudsman.org/cases/case_detail.aspx?id=177. 90 Myanmar Centre for Responsible Business et al., Tourism Sector-Wide Impact Assessment (2015). 91 The WB Group has a checklist of 70 pages (IFC) and 100 pages (World Bank) of dense and detailed PSs. The HRIAs employ a checklist of human rights standards developed in international law. mitigation hierarchy and compliance with PSs. 92 There is a further no to the above question: the WB Group system is built around social impacts. It does not yet have sensitivity to how human rights impacts might differ from social impacts. There are complications resulting from this focus on social impacts. Thus, the IFC tolerates residual social impacts even ex post, they can be offset through positive impacts or ignored altogether if not significant enough. At the latest revision of its standards, the IFC undertook a gap analysis that 'confirmed that the multiple dimensions of rights' are 'well addressed in the Standards'. 93 This section showed there are complexities and specificities in managing human rights risks that should be accounted for. First, human rights impacts should be distinguished from social impacts; residual impact is one concept proposed herein to separate them. Second, the difficult position of observers tracking corporate performance should be acknowledged; the four offsets identified herein show there is a possibility that observers get a distorted picture of a company's performance as they witness an elaborate display of corporate policies and specific data. Third and finally, the insufficiently theorized ramifications of managing human rights risks in a less institutionalized regime hinders casual observers and even well-meaning corporate staff and pushes them towards unfeasible and bright-line explanations of prevention that might misrepresent the risk management task at hand. Such shortcomings matter from a regulatory perspective, as they throw sand in the wheels of polycentric governance and could stunt the evolutionary dynamics the UNGPs count on.

Prevention as reduction at source
The UNGPs struck a new chord with their human rights due diligence concept, which identified generic steps without which a corporate response is flawed, and further clarified the nature of human rights risk management by identifying useful 'parameters' and the three offsets. This article suggests that to deal with the fourth offset another parameter should be added as a response to the problem of 'residual impact'.
The missing factor becomes visible if one takes a root cause approach to infringements of human rights. 94 Risk management has to be capable of identifying and addressing deeper factors of risk. A company needs to adopt a package of mitigation measures that is not oblivious of root causes of infringements 95 and does not end up as a superficial due diligence process but addresses the problem at source. With reduction at source as a parameter, observers see whether the package of preventive measures is incomplete as the company is silent on a specific sort of action. This parameter helps spot more easily inadequate preventive measures that otherwise would strike one as appropriate and 'reasonable' conduct. By adding the 'reduction at source' parameter, the UNGPs would clarify that non-tolerance to residual impact applies not only ex post (remediation), but also ex ante (prevention).

WB Group
It is perhaps surprising that is within the WB Group system policies and practice that this parameter of reduction at source features most prominently. The concept of 'reduction at source' 92 The oversight layers consist of reviews of clients by the IFC management, by expert-driven EIAs, multiple layers of CAO review, and internal IFC review (self-review). The Wilmar case is an example where all these layers of review were exhausted. IFC, supra note 89. 93 Ibid., at 2. appears in IFC-triggered EIAs. The IFC expects reasonable efforts to minimize impacts to 'as low as reasonably practicable' levels (the ALARP principle in EIAs). ALARP envisages measures such as design modifications and other reduction at source methods. 96 The IFC pursues minimization at source which is achieved both through EIAs ex ante and ex post reviews.
First, EIAs offer detailed information on mitigation measures. Some EIAs show graphically and comprehensively the significance of impacts pre and post mitigation. 97 Such clear and structured presentations help demonstrate that the client didor failed toidentify comprehensive packages of mitigation measures, including minimization at source. The EIAs are a condition for disbursement and the WB Group has the leverage to insist on an independent and in-depth impact assessment prior to supporting the project. 98 Second, some EIAs explain reduction at source in the context of pollution. This is an application of PS 3 which explains that '"pollution prevention" does not mean absolute elimination of emissions, but the avoidance at source whenever possible, and, if not possible, then subsequent minimization of pollution : : : '. 99 One company's key environmental goal is to minimize emissions and disturbance to the environment. Its EIA explains this 'shall be achieved primarily through minimizing pollution ("reduction at source")'. 100 Third, the WB system can deliver in-depth review of performance, both of the client's and the WB management's performance. In the IFC system, the CAO is at the centre of the ex post review system. Through the years it has delivered highly critical evaluations of IFC performance. These CAO reviews can take an explicit root cause orientation. This appears clearly in the Wilmar case where CAO's Terms of Reference indicated 'developing an understanding of the immediate and underlying causes for any noncompliance identified by the CAO' as the IFC failed in this palm oil case to observe its own PSs. 101 Finally, the IFC guide on human rights risk management draws attention to root causes. In their assessments, companies should consider the 'source of the risks and impacts [that is] identify the root causes, trigger points, and key actors of the risks and impacts (e.g. the business activity itself, a possible contractor, supplier and/or government involvement etc)'. 102 Furthermore, through grievance mechanisms, 'companies can attempt to address the root cause early on, and forestall a minor issue from escalating into a major problem'. 103 The above demonstrate the IFC is aware of reduction at source as an operational way to pursue to avoidance goal; furthermore, the review infrastructure offer the opportunity and leverage to specify those preventive and corrective measures that can go deep towards root causes. In contrast the mitigation hierarchy the places 'avoidance' on its top layer seems to play a minor role once the discussion has moved to the concrete level of appropriate mitigation measures in a specific context. The IFC review body, the CAO, hardly mentions the hierarchy in its assessments. The IFC management also does not place expected management measures on the levels of a hierarchy when assessing a client. Instead, in their reviews, both the CAO and IFC track compliance with the specific PS requirements and suggest detailed mitigation measures appropriate to the specific context.
An illustration of the IFC infrastructure in action comes from a Mongolian minerals project. 104 Locals complained that a river was diverted and wanted the diversion stopped, and claimed 96 Oil Terminal EIA, supra note 67, at 6-4 (see graphic). 97 Ibid., at xliii-liii. Also SRK Consulting, TERRA Industrial Farming Complex -Environmental and Social Impact Assessment (2014) 168-89 (hereinafter 'Terra EIA'). violations of human rights from the diversion. 105 In their request to experts, herders wanted to know about alternatives and mitigation measures to avoid or reduce impacts; through the CAO mediation process, an expert study was commissioned. 106 Based on that credible information, herders negotiated with the company, under the CAO umbrella, measures that both created more access to water and compensated the herders affected. 107 This shows the diversity of measuresmainly restoration and compensation but also some measures on avoidanceavailable when there is an expert study and a negotiation process. Those two aspects determined the right balance of elimination and reduction of impacts acceptable to stakeholders in a specific context. Although the initial request was for full avoidance of impacts through stopping the project, the acceptable result was a mitigation package exhibiting all levels of the mitigation hierarchy.
The WB system carries insights for the UNGPs. Deprived of leverage and supporting infrastructure and having a blind spot on offset and the reduction of source parameter of human rights due diligence, the UNGPs risk relying disproportionately on the general and principled definition of the responsibility to respect as avoidance of adverse impacts. If one wonders what is the relative weight that such abstract notion can carry in practice the answer can be gleaned from the similar conceptthe mitigation hierarchyused by the IFC: rather little. It is therefore important that the operational notion of due diligence used by the UNGPs is not skipping out of alignment with the corporate responsibility to respect. The suggested way to achieve that is to emphasize the parameter of reduction at source: it is the answer to the fourth offset and embodies the root cause orientation that human rights risk management should have.

UNGPs
The UNGPs display an implicit root cause orientation. This is discernible in the way the UNGPs explained responsibilities depending on the involvement in abuse: by causation, by contributions, or through linkages. 108 Such root causes of human rights impacts are not solely within the control of the company, as when a company infringes rights in its workplace or employs polluting technologies. Indeed, there are more complex settings characterized by harmful practices of third parties and/or far-reaching impacts of a company's own decisions. Even in these settings, where a business does not have control, it retains a responsibility to exercise human rights due diligence and thus identify and address deeper factors of risk, whether in its operating environment or throughout the supply chain.
First, a company must cease its harmful conduct with ripple effects, for example, a buyers' purchasing practices that squeeze suppliers through contractual terms on price, delivery time, and quality to an extent that suppliers decide to cut corners and infringe worker rights. Such purchasing decisions are a root cause within the control of the buyer company, a deep factor of risk that the company introduces into the supply chain.
Addressing this root causea company's own decision or business modelcan be illustrated with Arla's HRIA. The Danish dairy company Arla critically reflects on its business model in a HRIA for its new operations in Senegal. It explains it uses 'a business model based on repackaging and distribution of imported European farmer milk'. 109 Local stakeholders are concerned that Arla does not rely on the local supply chain and will likely have disruptive effects on the domestic dairy sector and its small-scale farmers. 110 In terms of root causes, the HRIA identifies structural obstacles (not less than 20) that the Senegalese diary sector faces, but deeper causes on Arla's side 105 Ibid., Letter of complaint (3 February 2013). 106 Ibid., Herders' complaint resolution agreement (2017) at 2. 107 Ibid., 6-11. 108 UNGPs, supra note 3, Principles 13, 19. 109 Arla, Human Rights Assessment in Senegal, Report November 2015, at 29, available at www.arla.com/globalassets/arlaglobal/company-overview/responsibility/human-rights/arla-human-rights-assessment-in-senegal-november-2015_1.pdf. 110 Ibid., at 3. are also highlighted, going as far as regulatory frameworks in the EU. 111 The company acknowledges it might 'become an integrated part of the complex systemic factors, which is linked to the continuous underdevelopment of the dairy sector by amplifying the current difficulties'. 112 Regarding mitigation measures, the company considersand discardswithdrawal (outright avoidance of impacts) and seeks to devise a package of measures which are identified and analysed through consultations with local stakeholders. 113 With these the company seeks to design 'business models that facilitate a beneficial co-existence between Arla's business and the development of the dairy sector'. 114 Outright elimination of impacts by not entering the market or fully relying on the domestic diary sector appears unfeasible, but the company identified root causes and seemed prepared to adopt 'reduction at source' measures.
Second, the company not causing or contributing to adverse impacts also has to take action regarding third parties with which it has a business relationship when they infringe human rights, such as land dispossession or worker rights abuses. Here, the company could recognize and tackle root causes outside its control if it seeks a more sustainable and systemic solution to abuses. Under the UNGPs, the company is expected to first exercise leverage over the third party and actors in its ecosystem and if no progress is made, to consider terminating the relationship. 115 The EIA Terra submitted to the IFC offers an example of reduction at source as the company addresses deep factors of risk in its own conduct and third-party conduct. The DRC-based agri-company Terra explains how it manages one major social impact, that is, population influx triggered by the economic opportunities the project creates. 116 The company does not pass the responsibility to local authorities and the burden to local communities for such 'indirect impacts'. 117 Instead, the company adopted an Influx Management Plan which sets out its 'contribution to the mitigation of the negative impacts associated with uncontrolled influx'. 118 The EIA identified five social impacts generated by the influx. 119 To address them, the company plans to adopt measures within its control (e.g., managing expectations through clear policies on recruitment policy, and on compensation and livelihood restoration) and exercise leverage over third parties (e.g., co-operation and support for local government and NGOs). This is an illustration of root cause thinking to avoid infringing human rights: the company explores ways to reduce the influx itself (a root cause) as well as to minimize the harmful effects of the influx that, if left unmanaged, can end up being infringements of human rights.

Prevention and avoidance re-explained
This factor of reduction (or minimization) at source deals with the problem of residual impact ex ante. It helps to explain, differently, what avoidance of adverse impacts means in a human rights context. Avoidance is neither about guaranteeing non-occurrence of harm (elimination of risk can probably be achieved only exceptionally), nor about aiming for elimination of impacts but falling short in real life (which opens the door for tacit acceptance of 'residual impact'). It is about systematically striving to identify and address root causes as an obligation of conduct and an essential feature of respecting human rights. 111 Ibid., at 28-9. 112 Ibid., at 3. 113 Ibid., at 60-1. 114 Ibid., at 31. 115 UNGPs, supra note 3, Principle 19 Commentary. 116 Terra EIA, supra note 97, at 202. 117 IFC, Guidance note 1, supra note 2, GN 39 (regarding 'secondary or induced social impacts'). 118 Supra note 97, at 202. 119 The five impacts are consolidated in a table at ibid., 175-6. They are: spontaneous settlement due to perceived employment opportunities, increase of social pathologies and communicable diseases, and increased pressure on natural resources and social services. This article proposes that when outright elimination of risk is unfeasible, avoidance should be understood as reduction at source. This introduces a new parameter in human rights due diligence to emphasize that root causes must be identified and addressed. At a general level, the direction of due diligence if not the amount of effort can be clarified conceptually. Under favourable conditions, this approach to risk management can potentially eliminate a risk of harm altogether. Under less favourable conditions, by tackling root causes under and outside its control, a business achieves minimization of risk to the maximum extent feasible. What are acceptable levels of risks is not exclusively for the company to determine, but results from observing applicable laws and international human rights standards, and participatory process as explained in the UNGPs. There are no guarantees that the impact will be prevented but measures that address root causes at least have the inherent capacity to eliminate an impact under favourable conditions. Tackling deeper factors of risk is evidence that a company does not accept residual impacts ex ante. In sum, the aim of the UNGPs to avoid infringement (elimination of impact) can be achieved with either outright avoidance or reduction of risk measures as long as the root causes are tackled.
This view on prevention can improve on current explanations of avoidance of impacts that appear incomplete or offer problematic clarifications. First, the OHCHR Guide defined prevention as non-occurrence of an impact. Such elimination of an impact clearly could be achieved only in a limited number of situations. An activity could be avoided altogether: not investing in a high risk setting, or severing links to an abusive supplier. Or avoidance could entail redesigning a project or activity to eliminate completely a source of risk. 120 However a majority of situations are not amenable to such straightforward risk management treatments. 121 Thus, for most situations, avoidance of impact collapses into reduction of risk, which is fine as long as specific standards and procedures, and a strong review mechanism, are in place.
Second, there are explanations that couple avoidance with severe impacts. The IFC accounts for the 'severity and magnitude' of human rights risk and requires measures at the higher levels of the mitigation hierarchy, such as the avoidance level. 122 Following the UNGPs, severe impacts should be prioritized and more stringent measures should be taken to deal with such impacts. 123 A guide issued by a respected organization also explains avoidance by reference to severe impacts. 124 With such explanations, other 'less severe' human rights impacts are relegated implicitly to reduction; avoidance is made for them optional. This explanation creates inconsistencies as it introduces a distinction that is not present in the UNGPs.
Third and finally, avoidance of impact ex ante could be pursued from another angle: strengthening the ex post stage. While not an explanation of avoidance, a rule imposing strict liability would set a bright linejust as OHCHR did with non-occurrence of impactand change the prevention equation. A strong liability regime would not require the 'casual observer' to get immersed in the specifics of management systems and whether they address root causes or not. Higher level of prevention could be expected from building up deterrence through heavy penalties, imposed ex post. However, as with non-occurrence, this bright line approach appears unfeasible as legislators will support it but for a small minority of the worst abuses at best. 125 120 'In many cases, avoiding involves "designing" the project so that the feature causing an impact is modified or excluded (e.g., changing a process to eliminate a pollutant, rerouting a pipeline, relocating a facility).' IPIECA and Danish Institute for Human Rights, Integrating human rights into environmental, social and health impact assessments: A practical guide for the oil and gas industry (2013). The 'severity and magnitude' of human rights risk 'will determine which type of mitigating approach to pursue. Human rights impacts that pose significant and immediate risk to the health, safety and the lives of stakeholders will benefit from avoidance, reduction and/or restoration measures'. IFC, supra note 102, at 49. 123 UNGPs, supra note 3, Principles 14, 24. 124 MCRB, supra note 90 ('For particularly serious human rights abuses, avoidance or prevention is the only appropriate response'). 125 Supra note 33.
The treaty discussions currently taking place in the UN are testing this path towards responsible risk-taking as the focus is decidedly on remedies and multinational enterprises' liability. 126 The current approaches to avoidance and elimination of impacts show the need for a different explanation of prevention that (i) could make sense in a wide variety of operational contexts, (ii) would be feasible to adopt by policymakers and implement by businesses, and (iii) reflect the particularity of human rights impacts. Otherwise, avoidance of impact collapses into reduction of risk. This has ramifications where the regulatory regime is not strong in terms of defining specific measures and backing them with strong adjudication and enforcement mechanisms. The unfortunate result of current explanatory treatments is a de facto acceptance of 'residual impact' ex ante, increased power for businesses to relativize their observance of human rights, and the crippling of the 'casual observer' on which the current polycentric regime depends on.
Measures targeting root causes of impacts are not the exclusive type of measures that a company should take to avoid adverse impacts and thus fulfil its responsibility to respect human rights. Through management systems companies design a bundle of impact-minimizing and management measures (monitoring, grievance, stakeholder consultation, capacity building). 127 Not all such measures can readily be explained through root causes lenses. Therefore, the thesis here is that whatever package of preventive measures is devised it should not conspicuously miss measures capable of identifying and tackling deeper factors of risk. Not accepting residual impact ex ante means demonstrating to casual observers that 'reduction at source' measures were identified, considered, and included in the risk management package.
Reduction at source does not set impossible tasks for companies. First, a company does not have to eliminate a root cause or to do it alone; it has to identify and tackle it, possibly in collaboration with others. 128 Thus, considerations of technical and financial feasibility the WB Group mentions are not ignored. 129 Second, when a third party's harmful conduct is the root cause of infringements, preventive measures do not require, for example, encroaching on the prerogatives of public authorities, 130 or playing inappropriate roles regarding subsidiaries and contractors. 131 But in line with the UNGPs, leverage should be exercised over third parties to improve protections before terminating the business relationship; 132 to minimize at source when third parties are involved means to revise one's own conduct incentivizing third parties into noncompliance as well as to use leverage over them.

Conclusions
The 'reduction at source' parameter proposed herein adds a new indicator regarding the prevention of human rights impacts. It facilitates the explanation of what the management of human rights risks entails and how it differs from managing other negative impacts on society and environment. The analysis found that current explanations in both the UNGPs and the WB Group fall short but in different ways. They do not fully capture the particularities of human rights as a special type of social impact. This shortcoming is compounded when the local and transnational regulatory contextwith its weaknessesis factored in; the absence of liability changes the 126 UN Inter-Governmental Working Group, Legally Binding Instrument to Regulate, in International Human Rights Law, the Activities of Transnational Corporations and Other Business Enterprises ('Zero Draft') 16 July 2018. 127 prevention-remediation equation and results in excessive risk-taking and/or inadequate preventive measures.
The explanation offered herein helps in separating when adequate preventive effort was made from when it was not. 'Reduction at source' is not a bright line but it is still a rather straightforward parameter to track given that corporate narratives silent on deeper causes of risk stand out conspicuously. This is a procedural parameter as it measures activity directed towards root causes, but is not so overly technical and context specific as to overwhelm casual observers. It should assist economic decision-makers in their risk management task and casual observers in assessing their performance. Recognizing the new parameter helps explain why residual impact on human rights is problematic and cannot be tolerated ex ante not only ex post, and how risks should be managed without creating unfeasible requirements.
For the UNGPs this analysis contributes a new parameter to human rights due diligence. It helps UNGPs explain that residual impact is intolerable from a human rights perspective, not only at the ex post stage but also ex ante. This explanation of prevention indicates that under current readings of the UNGPs the avoidance of impact collapses into mere reduction of risk; this is undesirable and can be avoided. On the one hand, should that collapse be allowed to happen, companies acquire discretion to offset rigorous preventive measures with less demanding superficial ones. This enables companies to develop confusing and misleading narratives about their human rights performance that place casual observers at a disadvantage in their oversight efforts. The UNGPs already identified three offsets and this article continues in that direction by adding a fourth offset. On the other hand, the collapse of avoidance into reduction marks a threat to the UNGPs internal consistency. The corporate responsibility to respect sets the avoidance aim while due diligence pushes for a mere reduction of risk. One does not want to convey the impression that the responsibility to respect and the UNGPs are aspirational while due diligence and human rights risk management have to meet the test of practicality in the real world. A conventional and common-sense infused approach to human rights risk management pushes the two key conceptsresponsibility to respect and due diligenceout of alignment and thus reduces the explanatory power of the UNGPs. That diminishes the latter's only leveragepersuasion based on conceptual clarityat a moment when the business and human rights regime is less formalized, is struggling to evolve, and relies on diverse stakeholders to push for due diligence and more systemic change. By analysing the particularities of human rights impacts and the regulatory context, the article explained how a lack of clarity on prevention can shortcut the evolutionary dynamics the UNGPs seek to facilitate through their 'polycentric governance' approach.
For the WB Group, this explanation helps grasp how human rights impacts are different from social impacts. The WB Group has the risk management knowledge and also the implementation infrastructure that the UNGPs lack. It is able to specify detailed preventive measures and subject clients to rigorous reviews, and skip general discussions on prevention altogether. But so far it has not accounted for the specificities of human rights risks as it has concentrated on social impacts. With this blind spot the WB Group has more of an explanatory problem than an operational one. The WB Group does not have to rely on casual observers as the UNGPs do. The WB Group can rely on its own detailed requirements, infrastructure and leverage to address root causes of harms. For the WB Group, attention to the four offsets in human rights management can help to better differentiate between social and human rights impacts, improve its explanations on risk management, and demonstrate the WB Group fully accounts for its human rights impacts. In sum, the article argued that cross-fertilization between the WB Groups' and the UNGPs' conceptualizations is possible and necessary. Combining the insights produced by the WB Group and the UNGPs can deliver a clearer explanation of prevention and management of human rights impacts.