TY - JOUR
AU1 - Friedman,, Nick
AB - Abstract In this article, I critically review the economic theory of corporate liability design, focusing on the allocation of liability between a corporation and its individual human agents. I apply this theory to transnational commercial contexts where human rights abuses occur and assess the likely efficacy of some putative liability regimes, including regimes requiring corporations to undertake human rights due diligence throughout their global supply chains. I advance a set of general considerations justifying the efficacy of due diligence in relation to alternative liability regimes. I argue, however, that due diligence regimes will likely under-deter severe human rights abuses unless they are supported by substantial entity-level sanctions and, in at least some cases, by supplementary liability for individual executives. The analysis has significant policy implications for current national and international efforts to enforce human rights norms against corporations. 1. Introduction Recent international initiatives seek enhanced accountability for human rights abuses by corporations. The most significant instrument so far is the United Nations Guiding Principles on Business and Human Rights (UNGPs).1 Notably, the UNGPs do not recognise an international law duty on corporations to respect human rights wherever they operate. Instead, they emphasise each state’s duty to ‘prevent, investigate, punish and redress’ human rights abuses by corporations in its jurisdiction ‘through effective policies, legislation, regulations and adjudication’.2 More recently, the UN working group on human rights and transnational corporations released a draft treaty (the Zero Draft) to enforce human rights norms against corporations.3 Like the UNGPs, the Zero Draft treats such enforcement as a matter of domestic regulation. It obligates states to establish a ‘comprehensive and adequate system of legal liability’ for their corporations’ local and foreign conduct, including the use of ‘effective, proportionate, and dissuasive sanctions’.4 Some states have already begun to implement this obligation; more are expected to follow.5 This obligation raises foundational questions about what effective liability would look like in this context. What is the nature and cause of the various human rights abuses committed by corporations, at home and abroad? Given their nature and causes, what sanctions would be effective at preventing and redressing them? When should sanctions target the corporation, when its individual human agents? When should liability be civil or criminal, strict or fault-based? Which sanctions would be proportionate, in the sense that they exact redress from all and only those corporate actors who are morally responsible for such harms, from each to the extent of their wrongdoing? There are inadequate resources in the existing ‘business and human rights’ literature to address these questions.6 Helpfully, corporate law scholarship provides a framework that can be adapted for thinking through the options. However, the nature of human rights and the circumstances under which corporations violate them require careful attention to core assumptions underpinning this economics-centred scholarship. In this article, I seek to bridge this gap between human rights and corporate law, concentrating in particular on the allocation of liability between a corporation and its individual agents. This is only one element of corporate liability design, but it is among the most fundamental. The considerations I discuss apply to many kinds of human rights obligation that might be imposed on corporations or their agents, but I focus on the emerging obligation of ‘due diligence’. Scholars and activists now see due diligence as the key legal mechanism by which to prevent transnational human rights abuses by corporations, but there has been little theoretical or empirical work to justify its perceived efficacy relative to other liability regimes. My first aim is to supply one such justification. I also note that due diligence is widely treated as an obligation exclusive to the corporate entity. My second aim is to show that this entity-level obligation may under-deter certain abuses unless it is supplemented by liability for individual executives.7 If these conclusions are sound, they have significant policy implications for current national and international efforts to enforce human rights against corporations. My argument is as follows. Section 2 synthesises disparate threads of corporate law scholarship to offer human rights scholars and policy makers a single point of entry into the economic theory of corporate liability design, concentrating on the allocation of individual and entity liability. The argument develops and critiques this theory in certain respects, so may interest corporate lawyers as well.8 Section 3 clarifies relevant aspects of the due diligence obligation. Section 4 applies the economic theory to some possible liability regimes, advancing a general argument for a due diligence obligation supported, in at least some cases, by both entity- and individual-level sanctions. I conclude with examples of the form that individual liability might take. Three preliminary notes are in order. First, the forthcoming analysis is necessarily general and tentative. Precisely which forms of liability are appropriate in a given context turns on complex empirical matters.9 These vary with the human rights in question, the characteristics of the target corporations, the degree of their involvement with abuses, the likelihood with which certain corporate investments might prevent those abuses and the enforcement machinery in the relevant jurisdictions. Accordingly, I make no concrete policy proposals. I do, however, advance some generally applicable criteria for assessing the need for due diligence and for supporting that obligation with individual liability. These criteria are plausibly met in many instances of human rights abuse by corporations. At the least, my argument provides a framework in which these empirical investigations can be undertaken and the answers acted on. Secondly, corporations can be involved with different kinds of human rights abuse. I am mainly concerned with ‘transnational’ abuses committed by a corporation in a low-income ‘host state’ that is controlled by a large, well-resourced corporation domiciled in a high-income ‘home state’.10 (I use parent-subsidiary terminology for ease of expression, but the analysis accommodates other modes of control, including control of ‘independent’ suppliers.) I am also concerned with abuses that are ‘severe’, which the UNGPs delineate by ‘their scale, scope and irremediable character’.11 I take such abuses to include serious labour and environmental wrongs that have an acute impact on the health or well-being of a large number of people, as well as acts that would qualify as ‘gross’ human rights abuses under various international instruments (such as extrajudicial killings, slavery or torture).12 The nature of corporate involvement with such abuses varies widely. Corporate actors—intentionally or negligently, by act or omission—may commit primary abuses themselves, or they may contribute to primary abuses by external actors (like suppliers or governmental agents).13 Each liability regime must make complex decisions about which primary abuses and which contributory acts and omissions to prohibit. I do not address such matters here. Finally, economic analysis is not the only way to approach liability design.14 I rely on it because it underpins the scholarship that pays most attention to the allocation of liability. Moreover, economic arguments are sometimes perceived to justify lenient regimes of corporate accountability for human rights abuses. In fact, I show that they may justify enhanced regimes of both entity- and individual-level sanctions. In any event, the considerations I discuss—how sanctions impact individual and corporate behaviour, how best to use enforcement resources and so on—are important whether or not one fully subscribes to the economic analysis of law. 2. Corporate Liability Design In this section I introduce the theory of corporate liability design, focusing in particular on the allocation of entity and individual liability.15 By ‘liability’ I mean subjection to a legal obligation to perform or refrain from some act, the breach of which is a condition for the imposition of a sanction that is administered, directly or indirectly, by the state. By ‘sanction’ I mean broadly to include criminal and administrative fines, civil damages and, in the case of individuals, imprisonment. The obligation/sanction pair may fall on the corporate entity (which I have called ‘entity liability’) or the corporation’s agents (which I have called ‘individual liability’).16 Corporate law scholars treat human agents as causally responsible for all corporate acts and omissions.17 Indeed, entity liability ordinarily exists within a de jure regime of dual liability in which the corporation is vicariously (and, in tort, jointly and severally) liable for the wrongs of its agents. In practice, however, these agents contractually shift their risk to the corporation through indemnification or insurance, creating a de facto, unitary regime of entity liability.18 Thus, the distinction between regimes of entity and individual liability depends on prohibiting this risk shifting, such that the individual bears her liability costs herself. In what follows, I assume that it will be so prohibited.19 This section begins with the dynamics of entity liability and explains why it is the default strategy for deterring corporate misconduct, before discussing the conditions under which it must be supplemented by individual liability. I then introduce a key distinction between harm- and act-based liability. Finally, I address some objections to the economic analysis of human rights liability. A. The Dynamics and Benefits of Entity Liability Profit-maximising corporations try to minimise their costs. Here, I take these to consist of the corporation’s production costs (including wages, materials, technologies and so on) and its expected liability costs for accidents arising probabilistically from the acts or omissions of its agents.20 The expected liability costs for some act or omission are a function of the nominal sanction imposed on the corporation upon occurrence of an accident,21 discounted by two probabilities: the chance that some act or omission will cause an accident and, if it does, the chance that the accident will be detected and successfully sanctioned.22 Table 1 demonstrates how a nominal fine of £1 million, after discounting for the chances of an accident and sanction, translates to expected liability costs which are a fraction of that amount.23 Table 1. Entity Liability   Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Expected Liability Costsa . Total Costsb . None £0 20% 10% £1,000,000 £20,000 £20,000 Good £2000 15% 10% £1,000,000 £15,000 £17,000 Best £5000 13% 10% £1,000,000 £13,000 £18,000 Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Expected Liability Costsa . Total Costsb . None £0 20% 10% £1,000,000 £20,000 £20,000 Good £2000 15% 10% £1,000,000 £15,000 £17,000 Best £5000 13% 10% £1,000,000 £13,000 £18,000 a (b) × (c) × (d). b (a) + (e). Open in new tab Table 1. Entity Liability   Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Expected Liability Costsa . Total Costsb . None £0 20% 10% £1,000,000 £20,000 £20,000 Good £2000 15% 10% £1,000,000 £15,000 £17,000 Best £5000 13% 10% £1,000,000 £13,000 £18,000 Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Expected Liability Costsa . Total Costsb . None £0 20% 10% £1,000,000 £20,000 £20,000 Good £2000 15% 10% £1,000,000 £15,000 £17,000 Best £5000 13% 10% £1,000,000 £13,000 £18,000 a (b) × (c) × (d). b (a) + (e). Open in new tab There is an important relationship between expected liability costs and production costs: the corporation can decrease the former by making accident-preventing investments in what I will call ‘production safety’—better equipment, new technology, cleaner materials and so on. A profit-maximising corporation will make such investments, but only to a point. Initial investments in production safety will yield relatively large marginal decreases in expected liability costs. In Table 1, for example, the corporation must choose whether to install one or neither of two safety devices. The good one costs only £2000, but reduces expected liability by £5000. Installing this device maximises profits because it saves money overall. However, such investments will generally yield diminishing marginal returns, so that continued decreases in expected liability require increasingly large investments in production safety. In Table 1, the best device costs another £3000, but further reduces expected liability by only £2000. A profit-maximising corporation will skip that device. In general, it will invest in production safety only until its marginal increase in production costs is offset by the marginal reduction in its expected liability.24 At that equilibrium point, the corporation cannot adjust its production costs or expected liability costs without increasing its costs overall. In Table 1, that point is the good, not the best, device. Depending on the type of accident under consideration, the good device might be acceptable from society’s perspective. If not, the state must adjust the corporation’s expected liability costs by raising the chances of sanction (which requires increased enforcement spending) or raising the nominal sanction amount (which costs the state nothing). Either approach upsets the corporation’s equilibrium and incentivises it to make still further investments in safety from which previously it may have refrained. In Table 2, for example, raising the fine to £2 million makes the best safety device the most cost-effective for the corporation. To this extent, expected liability costs are the interface between the corporation and the state. They translate the strength of the state’s deterrence objectives into proportional financial incentives for the corporation to curtail the misconduct. Table 2. Entity Liability with Increased Fine   Safety Device . Cost . Chance of Accident . Chance of Sanction . Fine . Expected Liability Costs . Total Costs . None £0 20% 10% £2,000,000 £40,000 £40,000 Good £2000 15% 10% £2,000,000 £30,000 £32,000 Best £5000 13% 10% £2,000,000 £26,000 £31,000 Safety Device . Cost . Chance of Accident . Chance of Sanction . Fine . Expected Liability Costs . Total Costs . None £0 20% 10% £2,000,000 £40,000 £40,000 Good £2000 15% 10% £2,000,000 £30,000 £32,000 Best £5000 13% 10% £2,000,000 £26,000 £31,000 Open in new tab Table 2. Entity Liability with Increased Fine   Safety Device . Cost . Chance of Accident . Chance of Sanction . Fine . Expected Liability Costs . Total Costs . None £0 20% 10% £2,000,000 £40,000 £40,000 Good £2000 15% 10% £2,000,000 £30,000 £32,000 Best £5000 13% 10% £2,000,000 £26,000 £31,000 Safety Device . Cost . Chance of Accident . Chance of Sanction . Fine . Expected Liability Costs . Total Costs . None £0 20% 10% £2,000,000 £40,000 £40,000 Good £2000 15% 10% £2,000,000 £30,000 £32,000 Best £5000 13% 10% £2,000,000 £26,000 £31,000 Open in new tab The appeal of entity liability is apparent. For many types of accidents, the state must simply decide what level of deterrence it wants to achieve, and then, relying only on the corporation’s profit-seeking motive, raise the corporation’s expected liability costs until that level is reached. Typically, it is a low-intervention strategy: where accidents can be prevented by several means, entity liability lets each corporation pursue whichever means it prefers.25 To fully demonstrate the appeal of entity liability, one must compare it to a pure regime of individual liability.26 I do not undertake that comparison here. It is enough to note three reasons why entity liability is typically essential to a corporate liability regime. First, a corporation’s agents will often be judgment-proof, ie they will lack the assets to compensate society for the accidents they cause. Since they do not bear the full costs of the harm they cause, agents will be under-incentivised to prevent them.27 Wealthier corporations are better placed to bear those costs. Secondly, where entity and individual liability can achieve the same level of deterrence, entity liability is frequently cheaper for the corporation. The corporation bears risk more efficiently than its agents, who may demand a substantial risk premium on their compensation to induce them to work under threat of individual liability.28 Thirdly, entity liability is normally cheaper from an enforcement perspective because corporations can usually detect and punish internal wrongdoing more easily than the state.29 For these reasons, in the likely circumstances where culpable agents are judgment-proof, risk averse and more visible to the corporation than the state, entity liability is the default choice for a corporate liability regime, and usually preferable to a regime of pure individual liability. Below, I address the conditions under which entity liability alone is inadequate. B. When Entity Liability Fails With entity liability, what matters ultimately is whether (i) expected liability costs can be set adequately to incentivise the corporation and (ii) whether the corporation can transmit those incentives to its agents. Sometimes it may be impossible to do one or both. I take each in turn. (i) Sanction insufficiency Sometimes sanctions cannot be made high enough to induce the desired level of deterrence. Borrowing a term from Kraakman, I call this ‘sanction insufficiency’.30 It arises in three overlapping circumstances. First, it arises where a corporation is undercapitalised (perhaps deliberately) relative to its risk profile, such that the sanction exceeds the corporation’s ability to pay.31 The corporation’s assets represent the maximum sanction that it can afford. Through the doctrine of limited liability, its assets cap its expected liability costs and so, too, the investments it will make to reduce them.32 This result is demonstrated in Table 3, which retains the figures from Table 2 but introduces a constraint of available assets (totalling £300,000).33 Instead of some chance of paying the full amount of the nominal sanction, the corporation merely faces some chance of handing over the sum of its limited assets. The effect of the sanction on the corporation’s expected liability is now dampened, such that the safety devices reduce those costs less than before. In fact, both devices now cost more than they save; the corporation’s costs are lowest when it installs neither. Since expected liability costs are determined by its assets, not the sanction, continuing to increase the sanction will achieve no additional deterrence. And, of course, the fine would bankrupt this corporation if it were actually enforced.34 Table 3. Liability of Undercapitalised Entity   Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Assets (after device)a . Expected Liability Costs . Total Costsd . Unlimited Assetsb . (f) Actual Assetsc . None £0 20% 10% £2,000,000 £300,000 £40,000 £6,000 £6,000 Good £2000 15% 10% £2,000,000 £298,000 £30,000 £4,470 £6,470 Best £5000 13% 10% £2,000,000 £295,000 £26,000 £3,835 £8,835 Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Assets (after device)a . Expected Liability Costs . Total Costsd . Unlimited Assetsb . (f) Actual Assetsc . None £0 20% 10% £2,000,000 £300,000 £40,000 £6,000 £6,000 Good £2000 15% 10% £2,000,000 £298,000 £30,000 £4,470 £6,470 Best £5000 13% 10% £2,000,000 £295,000 £26,000 £3,835 £8,835 a £300,000 – (a). b (b) × (c) × (d). c (b) × (c) × (e). d (a) + (f). Open in new tab Table 3. Liability of Undercapitalised Entity   Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Assets (after device)a . Expected Liability Costs . Total Costsd . Unlimited Assetsb . (f) Actual Assetsc . None £0 20% 10% £2,000,000 £300,000 £40,000 £6,000 £6,000 Good £2000 15% 10% £2,000,000 £298,000 £30,000 £4,470 £6,470 Best £5000 13% 10% £2,000,000 £295,000 £26,000 £3,835 £8,835 Safety Device . (a) Cost . (b) Chance of Accident . (c) Chance of Sanction . (d) Fine . (e) Assets (after device)a . Expected Liability Costs . Total Costsd . Unlimited Assetsb . (f) Actual Assetsc . None £0 20% 10% £2,000,000 £300,000 £40,000 £6,000 £6,000 Good £2000 15% 10% £2,000,000 £298,000 £30,000 £4,470 £6,470 Best £5000 13% 10% £2,000,000 £295,000 £26,000 £3,835 £8,835 a £300,000 – (a). b (b) × (c) × (d). c (b) × (c) × (e). d (a) + (f). Open in new tab Secondly, sanction insufficiency occurs where the social costs of the corporation’s harmful activity are extreme, as in industries that pose grave or systemic threats to human well-being or the environment. These costs may be high in a monetary sense, but they may also be high in a moral sense, as they are for egregious harms like slavery, torture and the loss of life. A liability regime should incentivise corporations to spend heavily to reduce the occurrence of such harms. However, the nominal sanctions necessary to induce the desired expenditure may be beyond the ability of even large corporations to pay, to say nothing of the many smaller corporations that are capable of the same type and scale of harm. There may be other practical limits on the severity of sanctions.35 Thirdly, where the probability of detection and sanction is very low—because state institutions are weak or the misconduct is difficult to prove—or is (perhaps mistakenly) perceived as very low by executives, even a massive fine translates to low expected liability costs.36 The corporation will then spend little to offset them with production safety.37 Sanction insufficiency can be mitigated in one of two ways: with supplementary individual liability (which I discuss below) or with increased enforcement spending.38 The latter may not be politically or financially feasible. Depending on the corporation’s assets, the costs of production safety and the extent to which enforcement spending boosts detection and sanction, it may also be ineffective.39 Where both solutions are feasible and effective, the choice turns on which achieves the desired deterrence more cheaply. There are at least two reasons to think supplementary individual liability will often be more cost-effective for the state. First, corporations generally have more resources to dedicate to their legal defence. As entity-level sanctions go up, so too will corporate defence spending and, in turn, the public enforcement resources needed to overcome it. Secondly, past a certain point, continued reductions in deterrence at the entity level will likely require increasingly large enforcement expenditures by the state. However, as Kraakman argues, individual liability ‘threatens a new population of potential offenders’ and so ‘should deter cheaply, even at modest levels of severity’.40 Moreover, ‘since the same resources already expended for detection and prosecution of offending firms also generally suffice against their managers, enforcement officials can purchase deterrence at little additional cost’.41 (ii) Internal disciplinary failure Sanction insufficiency is not the only impediment to the success of entity liability. Another impediment concerns what I will call ‘internal disciplinary failures’. Where such failures occur, even sufficient sanctions may not prevent accidents. Sometimes a corporation can prevent accidents by installing new equipment, using safer materials and so on. Frequently, however, it needs its employees to do something: double-check safety valves, stay awake on the night shift, carefully design a bribery-prevention system. In such cases, even where sanctions sufficiently incentivise a corporation to prevent accidents, the success of entity liability requires well-functioning disciplinary mechanisms by which the corporation can convey this incentive to its agents. The relationship between the corporation and its agents is a case of principal–agent conflict.42 One feature of this conflict is germane. It concerns the care and effort with which an agent performs her job. Under entity liability, careless agents increase the corporation’s exposure to sanction. All else being equal, it wants them to take more care. But from the agent’s perspective, care requires additional time and effort, sometimes denoted as her ‘costs of care’.43 If wages are fixed, she will prefer to take less care, not more, and may seek out alternative employment that requires less care for the same wage.44 Thus, she will not take care unless the corporation pays her to; that is, the corporation must customise rewards and penalties in her wage contract such that her expected wage when she takes care (including compensation for her increased costs of care) exceeds her wage when she does not.45 Then, having agreed to pay more for careful behaviour, the corporation must monitor that she does indeed take care and enforce the penalty provisions if she does not. The negotiation, monitoring and enforcement of such contracts are the corporation’s key internal disciplinary mechanisms. These mechanisms may fail, leading to suboptimal deterrence of accidents even under sufficient sanctions. For example, labour market conditions or the employee’s lack of personal wealth may limit the nature or extent of the rewards and penalties by which a corporation can induce care by the agent.46 Even a severe penalty—docking an agent’s full annual wage—may not, under plausible assumptions, induce proper care.47 Alternatively, it may be too difficult or expensive for the corporation to monitor its agent’s behaviour, such that her wage is effectively independent of the care she takes.48 This may be the case in large, decentralised organisations, or where harm results from the complex activities of multiple employees interacting over long time horizons. Or (importantly for the forthcoming analysis) it may be that greater care is required from executives. Executives are the ones who decide for and on behalf of the corporation—whether to invest in production safety, discipline careless agents and so on—but they are chiefly motivated to maximise their own utility.49 Preventing certain accidents requires executives to put more effort into their management and oversight responsibilities. But for these executives, effort is costly. Under a system of sufficient entity sanctions, it may be in the corporation’s profit-maximising interest for them to put more effort into these responsibilities. However, for reasons of self-interest, they may be disinclined to do this.50 For example, they might reject contractual provisions that make their salaries depend on their level of care. Alternatively, where their contracts include such provisions, they may decline to enforce them against themselves. Where internal disciplinary mechanisms fail, even a corporation that is well motivated to avoid a large fine might fail to induce the necessary care in its agents. C. Supplementary Individual Liability These shortcomings of entity liability—stemming from sanction insufficiency or internal disciplinary failure—can be mitigated by supplementary liability for the corporation’s agents. The agent may be subjected to monetary sanctions or, if the agent is likely to be judgment-proof, to non-monetary sanctions like imprisonment.51 Either way, the agent’s own expected liability costs from an accident are raised beyond what the corporation, ex hypothesi, is willing or able to do. In much the same way that the corporation responds to entity liability, the agent will try to reduce her liability costs and may, in the process, improve production safety. Supplementary individual liability mitigates sanction insufficiency when it is aimed at executives. As noted above, executives decide on the corporation’s safety investments, but what they decide is determined by their own interests. When they are made personally liable for corporate accidents, they will seek to minimise their own liability costs by directing the corporation to invest in production safety, even when such investment is not profit-maximising.52 This result is demonstrated in Table 4, which builds on the example of sanction insufficiency from Table 3. Assume that the corporation has one executive who is subject to a state-imposed fine whenever an accident occurs.53 She is also responsible for overseeing the safety device and incurs costs of care commensurate to its complexity. Her nominal salary is fixed: it excludes additional compensation for her costs of care (since, under sanction insufficiency, the corporation prefers that she take no care); also assume that the corporation is prohibited from insuring, indemnifying or otherwise compensating away her liability costs. Taking into account the fine and her costs of care, her expected take-home pay is highest with the best device installed. Although that device remains the least profitable for the corporation, the self-interested executive will cause the corporation to install it.54 The result is greater accident prevention than would occur under entity liability alone. Table 4. Supplementary Individual Liability   Safety Device . (a) Chance of Accident . (b) Chance of Sanction . Individual Executive . Corporation . (c) Nominal Salary . (d) Fine . (e) Expected Liability Costsa . (f) Costs of Care . (g) Expected Salaryb . (h) Cost of Device . (i) Expected Liability Costsc . Total Costsd . None 20% 10% £20,000 £8000 £160 £0 £19,840 £0 £5600 £25,600 Good 15% 10% £20,000 £8000 £120 £10 £19,870 £2000 £4170 £26,170 Best 13% 10% £20,000 £8000 £104 £20 £19,876 £5000 £3575 £28,575 Safety Device . (a) Chance of Accident . (b) Chance of Sanction . Individual Executive . Corporation . (c) Nominal Salary . (d) Fine . (e) Expected Liability Costsa . (f) Costs of Care . (g) Expected Salaryb . (h) Cost of Device . (i) Expected Liability Costsc . Total Costsd . None 20% 10% £20,000 £8000 £160 £0 £19,840 £0 £5600 £25,600 Good 15% 10% £20,000 £8000 £120 £10 £19,870 £2000 £4170 £26,170 Best 13% 10% £20,000 £8000 £104 £20 £19,876 £5000 £3575 £28,575 a (a) × (b) × (d). b (c) – (e) – (f). c (a) × (b) × (£300,000 – (c) – (h)). d (c) + (h) + (i). These total costs are based on Table 3 but now reflect (i) the executive’s salary and (ii) expected liability costs adjusted downwards (since the salary further limits the corporation’s assets). Open in new tab Table 4. Supplementary Individual Liability   Safety Device . (a) Chance of Accident . (b) Chance of Sanction . Individual Executive . Corporation . (c) Nominal Salary . (d) Fine . (e) Expected Liability Costsa . (f) Costs of Care . (g) Expected Salaryb . (h) Cost of Device . (i) Expected Liability Costsc . Total Costsd . None 20% 10% £20,000 £8000 £160 £0 £19,840 £0 £5600 £25,600 Good 15% 10% £20,000 £8000 £120 £10 £19,870 £2000 £4170 £26,170 Best 13% 10% £20,000 £8000 £104 £20 £19,876 £5000 £3575 £28,575 Safety Device . (a) Chance of Accident . (b) Chance of Sanction . Individual Executive . Corporation . (c) Nominal Salary . (d) Fine . (e) Expected Liability Costsa . (f) Costs of Care . (g) Expected Salaryb . (h) Cost of Device . (i) Expected Liability Costsc . Total Costsd . None 20% 10% £20,000 £8000 £160 £0 £19,840 £0 £5600 £25,600 Good 15% 10% £20,000 £8000 £120 £10 £19,870 £2000 £4170 £26,170 Best 13% 10% £20,000 £8000 £104 £20 £19,876 £5000 £3575 £28,575 a (a) × (b) × (d). b (c) – (e) – (f). c (a) × (b) × (£300,000 – (c) – (h)). d (c) + (h) + (i). These total costs are based on Table 3 but now reflect (i) the executive’s salary and (ii) expected liability costs adjusted downwards (since the salary further limits the corporation’s assets). Open in new tab Two further aspects of the example bear noting. First, without individual liability, executive and corporate interests would be aligned here: the corporation neither rewards her for preventing accidents nor punishes her when they occur (since it has no profit-maximising interest in doing so); because care is costly, her expected pay would be highest with no safety device to oversee. Secondly, it is apparent that risk-shifting mechanisms like insurance or indemnification, by reducing her expected liability to zero, would eliminate the executive’s incentive to install the safety devices. The example thus underscores the importance of prohibiting risk shifting under such circumstances.55 One worry with such prohibitions is that the agent may then seek other comparably paid employment that requires less care and less exposure to liability, or may simply prefer not to work at all. However, this worry has less force in the case of well-paid executives who are subjected to individual liability regimes of wide application (covering all large corporations, an entire industry, etc). Executives are often compensated well beyond their disutility of work, such that they will prefer modest liability costs to not working at all; and if all comparable jobs carry the same burden, there will be few viable alternatives offering the same nominal wage free of these costs. Supplementary individual liability cures or mitigates internal disciplinary failures when it is aimed at agents whom the corporation cannot control on its own.56 The corporation’s disciplinary mechanisms are limited to contractual rewards and penalties. The state, however, can apply a broader range of monetary and non-monetary sanctions to the corporation’s agents. By doing so, it raises their liability costs beyond what the corporation could do on its own, inducing them to invest in levels of care and effort from which previously they may have refrained. The success of individual liability as a cure for internal disciplinary failures depends on several factors. It is less likely to change the behaviour of lower-level agents undertaking complex joint activities in a large, decentralised organisation. If it is difficult for the corporation itself to hold them responsible under these conditions, it may, on account of the information asymmetries it faces, be even more onerous for the state to do so. The threat of individual liability under such conditions is remote: the agent will apply a substantial discount factor to any nominal sanction she faces, dampening the incentive for her to take care. Moreover, lower-level agents usually have limited wealth, which constrains the state’s options for monetary penalties as much as the corporation’s. This problem can be mitigated by the threat of imprisonment, but that would be inappropriate in many cases. The strategy may be more successful, however, when aimed at relatively visible forms of misconduct by wealthier executives. Here, the principal disciplinary failure is executive reluctance to enforce standards of care on themselves. Individual liability solves this problem by shifting the enforcement decision from executives to the state.57 The foregoing analysis ignores two important tools by which the corporation aligns executive interests with its own. First, executive compensation frequently includes stock options that tie the executive’s interests to the corporation’s profit. Stock options therefore exacerbate sanction insufficiency by inclining executives to support the corporation’s underinvestment in production safety. However, they mitigate internal disciplinary failures by inclining executives to support safety investments that are in the corporation’s profit-maximising interest. Thus, the effect of stock options on executive decision making may require individual sanctions to be made more or less severe, as the case may be.58 Secondly, directors and officers have fiduciary duties to act in the corporate interest.59 These duties may be more or less demanding between jurisdictions and may be subject to exculpatory clauses in corporate charters.60 Where effective, they may exacerbate sanction insufficiency and mitigate disciplinary failures; individual sanctions should be adjusted accordingly.61 Even accounting for these complications, it is clear that supplementary individual liability may secure additional deterrence that entity liability alone may lack. D. Harm- and Act-Based Liability One final aspect of corporate liability design is relevant to the ensuing analysis: the distinction between harm- and act-based liability.62 Under a harm-based regime, which I have assumed in my examples thus far, liability is imposed upon the occurrence of harm that the corporation has caused to some person or thing (as exemplified by the rules of tort or the criminal law), leaving the corporation to decide whether and how to prevent that harm. Under an act-based regime—frequently found in health and safety, environmental and financial regulation—the corporation is required to perform certain acts that the state has itself determined will prevent harm. Liability is triggered by the corporation’s failure to perform those acts, regardless of whether that failure has yet caused any harm.63 For example, rather than giving the corporation a choice of which safety device to install, the state might simply mandate that it install the best one on pain of sanction. Act-based liability offers three advantages over harm-based counterparts. First, it is typically effective at lower sanction levels than harm-based liability.64 Under act-based liability, the corporation no longer discounts the nominal penalty by the chances that its failure to perform some act will cause an accident (since the occurrence of an accident is irrelevant). The corporation discounts the penalty only once, by the chance that its failure will be detected and sanctioned. The result of this single rather than double discounting is that a lower nominal sanction yields the same expected liability cost that the corporation would face under an otherwise equivalent harm-based regime.65 For this reason, act-based liability may be attractive in cases of sanction insufficiency. Secondly, whereas harm-based liability defers the state’s intervention until harm has actually occurred, act-based liability allows the state to enforce harm-preventing measures ex ante.66 For this reason, the former is better suited to redressing harms which are wholly or largely economic in nature. As long as the victims are made financially whole after the fact, society may be indifferent between prevention or monetary compensation. Act-based liability, by contrast, helps prevent the occurrence of serious and/or essentially non-compensable harms.67 Finally, since act-based liability is not conditioned on the occurrence of accidents, it avoids complex inquiries into whether such accidents are caused by the corporation’s failure to perform the required act. E. Objections to Economic Analysis of Human Rights Liability A common objection to the kind of economic analysis I have offered is that it puts a price tag on harm; yet some harms (of which human rights abuses are a paradigm example) cannot or should not be priced. The objection is valid but misplaced. The only sanctions that can be imposed on corporations are, at bottom, financial. The critical question for a corporate liability regime is not whether this pricing of harm is appropriate (since it cannot be otherwise), but rather which principles should be used to price it. There are essentially two approaches by which one might determine the appropriate sanction. The first—which I have assumed thus far—begins by determining the desired level of deterrence (on moral, economic or other grounds) and then reverse-engineers a sanction that will achieve it. The second begins by determining the appropriate amount of the sanction itself. Whatever level of deterrence results from that sanction is, necessarily, optimal. Much of the corporate law scholarship on which I have relied follows this second approach, arguing, in particular, that sanction amounts should be determined by economic principles of social welfare maximisation.68 As I presently discuss, this leads to some more forceful objections, some of which are well founded in a human rights context. I should emphasise, therefore, that the dynamics of corporate liability design—as I have described them above—in no way depend on which of these two approaches one adopts, nor do they depend on the moral, economic or other grounds by which one determines the appropriate amount of deterrence or sanction. A profit-maximising corporation will respond to sanctions as described. The basic idea underlying the welfare-maximising approach is that commercial activity generates costs for third parties (in the form of the harm it causes them) that are external to the corporation’s profit-maximisation function.69 A liability regime, so the argument goes, should assign a cost to these harms and then calculate a sanction that internalises those costs within the corporation.70 The corporation’s prices and output will then reflect the total social costs of its activity. By trying to minimise those costs and thus maximise its profit, the corporation will simultaneously maximise social welfare. Welfare maximisation entails the possibility of ‘overdeterrence’, a situation in which the expected economic costs of increased deterrence outweigh its expected economic benefits. As I showed earlier, the point at which a corporation minimises costs is not always the point of maximum production safety: additional safety may be possible but too costly from the corporation’s perspective.71 It may also be too costly from a welfare perspective. Society bears the cost of the corporation’s harm, but also bears (in the form of higher prices and lower output) the corporation’s costs of abating it.72 At a certain point, society may prefer to stomach some residual harm rather than pay to reduce it still further. Additional investments in production safety beyond that point would, on the welfare-maximising approach, constitute overdeterrence. This approach attracts strong objections concerning the difficulty of accurately assigning costs to harms and developing sufficiently complex welfare measures, the incommensurability of value, and the moral implications of a utilitarian balancing of rights and interests. I cannot engage these objections here but, at least in a human rights context, they caution against privileging the goal of welfare maximisation. One worry in particular is that severe human rights abuses constitute egregious moral wrongs; it is both impossible and perverse to assign an economic cost to the occurrence of slavery, torture or death, to be balanced against consumer welfare. Relatedly, one cannot be agnostic, ex ante, about the ‘optimal’ level of abuse that would result from this cost assignment. One must begin, instead, from the premise that the optimal occurrence of severe human rights abuses is zero. While it is impossible for any liability regime to prevent these abuses entirely, it should incentivise corporations to spend heavily to eradicate them, more or less regardless of the negative effect on consumers. Hence, I assume that corporate liability regimes for human rights abuses should lean towards the first approach I described, adopting as a goal the elimination of severe abuses and then setting sanctions that will achieve it.73 This is not to say that something like overdeterrence is irrelevant: there are good pragmatic reasons (including feasibility and the opportunity cost of limited resources) why society may resign itself to some minimal occurrence of abuse. I only assume, on account of the nature of such abuses, that concerns about overdeterrence should be discounted. In what follows, I still occasionally refer to the notions of externalities and internalisation, which are helpful in several respects. Severe human rights abuses do impose both moral and financial costs on victims; following the principle of proportionality, those costs should be reflected in the severity of sanctions (even accepting that they cannot be precisely calculated). Moreover, these notions highlight the way that corporations have achieved artificially low prices and high output for home state consumers by externalising the true costs that human rights compliance would impose on their businesses. Sanctions must be set at a level that impounds these costs in the corporation’s profit function to the greatest possible extent. *** These, then, are the general considerations bearing on the allocation of individual and entity liability when the state seeks to induce corporate investment in production safety. In section 4, I apply these considerations to transnational commercial contexts where human rights abuses occur and assess the likely efficacy of some putative liability regimes. First, however, I discuss one particular investment in production safety that is of special importance to the forthcoming assessment. 3. Due Diligence Due diligence came to prominence as a component of the UNGPs.74 It has since been taken up by international organisations, advocacy groups, scholars and, to some extent, governments and corporations.75 A recent UN report notes this ‘convergence around the due diligence concept’, describing it as the ‘primary expectation of behaviour for any business with respect to its [human rights] responsibilities’.76 While the UNGPs stopped short of asserting a legal obligation on corporations to undertake due diligence, the Zero Draft marks a shift towards ‘mandatory’ due diligence. Article 6.1 obligates states to ‘take all necessary legal and policy measures to ensure that ... business enterprises that undertake business activities of a transnational character … prevent and mitigate human rights abuses throughout their operations’. States are increasingly imposing due diligence obligations on their corporations.77 Despite the prominent role of due diligence in the business and human rights discourse, some aspects of its nature and content are unclear. The first issue concerns the way that due diligence should generate, or relate to, liability. One approach situates due diligence in a scheme of tort or criminal liability. Liability here is harm-based, but the corporation can escape or reduce it by showing that it implemented due diligence.78 Another approach sees due diligence as part of a regulatory scheme requiring corporations to implement it. Liability here is act-based, triggered by a mere failure of due diligence regardless of whether a human rights abuse has yet occurred.79 This approach is akin to the French and Dutch legislation discussed below.80 It is consistent with a conception of due diligence as ‘mandatory’,81 and with the frequently expressed view that ‘The most important purpose [of due diligence] is to prevent adverse impacts’, rather than merely ensuring adequate compensation for harm ex post.82 Here, I focus on due diligence in this act-based, regulatory mode.83 A second issue concerns the specific measures that due diligence requires. The UNGPs describe these in general terms, requiring that corporations (i) identify actual or potential human rights abuses; (ii) take steps to cease or prevent such abuses; and (iii) track the implementation of those steps.84 However, the recently released OECD Due Diligence Guidance for Responsible Business Conduct (OECD Guidance) provides concrete examples of what might be required.85 For instance, at the stage of identifying human rights risks—say, potential labour or environmental abuses—a corporation might conduct on-site inspections and worker interviews.86 More generally, the identification stage requires impact assessments, inspections, audits and consultations.87 At the prevention stage—say, to mitigate potential health risks to local communities—a corporation might reroute an oil pipeline or change its production processes to eliminate toxic substances.88 More generally, prevention includes training for employees and investing in new machinery and technology.89 At the tracking stage—for example, measuring the reduction of child labour in its supply chain—a corporation might once again conduct on-site visits.90 More generally, tracking requires third-party audits, the establishment of local grievance mechanisms and consultation with local communities, and will need to be done frequently in contexts posing a high risk of severe harm.91 As these examples make clear, and as the UN working group itself recognises, ‘Managing human rights risks and impacts in supply chains can be extremely complex’.92 The ‘due diligence’ label invokes the risk assessments that corporations conduct before concluding certain commercial transactions.93 In my view, it is closer to what corporations understand as internal control. The benchmark standard on internal control defines it as a ‘process … designed to provide reasonable assurance regarding the achievement of objectives relating to operations, reporting, and compliance’.94 This process involves: identifying company objectives and risks to achieving them; implementing preventive and detective policies, procedures and technologies to mitigate those risks; and evaluative measures to determine whether the risks are indeed being mitigated.95 This is remarkably similar to due diligence as described above. In what follows, I refer to due diligence as a system of internal controls to detect, prevent and redress human rights abuses.96 Some further points about these controls are germane. First, they must be implemented throughout a corporation’s supply chain, not just within the parent and its subsidiaries.97 Secondly, their design and implementation is an ongoing, iterative process of adaptation to changing circumstances.98 Thirdly, implementation is a matter of degree. As the OECD Guidance notes, ‘When the likelihood and severity of an adverse impact is high, then due diligence should be more extensive’.99 Finally, like the design and implementation of any system of internal controls, due diligence requires significant effort from executives. Human rights are implicated in some of the most central activities of boards and senior management, including starting or ending business relations in certain places or with certain partners; designing, overseeing and evaluating organisation-wide systems; deciding how products should be built and services provided; assessing legal compliance; and managing risk exposure. Indeed, the OECD Guidance notes that a corporation should ‘Assign oversight and responsibility for due diligence to relevant senior management’, including ‘responsibility for ensuring that activities that cause or contribute to adverse impacts cease, and for preventing activities that may cause or contribute to adverse impacts in the future’.100 The need for board engagement is recognised by corporations themselves.101 One can fairly assume, then, that due diligence—depending on the severity of the abuse and the extent to which the state seeks to eliminate it—will impose substantial start-up and ongoing implementation costs on corporations relative to their size. This includes not only the compliance costs of implementing internal controls, but also, perhaps more significantly, the costs of preventing and remediating the potential or actual human rights abuses that these measures reveal: paying workers properly, upgrading factories, investing in cleaner processes and so on. I emphasise cost not as an objection to due diligence (though it will unquestionably raise prices and lower output across the board); rather, my point is to underscore how reticent corporations will be to implement serious due diligence, and thus to emphasise the severity of sanction needed to incentivise it. 4. Corporate Liability for Human Rights Abuses: Assessing Some Options The efficacy of a liability regime is best considered in comparison to alternatives. These alternatives are determined by several variables. I will discuss the choice between pure entity liability and supplementary individual liability, between harm- and act-based liability, and between implementing the regime in the host or the home state. I ignore the choice between strict and fault-based liability, between civil and criminal liability, and between private and public enforcement.102 Table 5 provides a partial categorisation based on these variables. I cannot assess each possible regime here; I confine myself to some of the most plausible. I advance some general considerations in favour of home state due diligence regimes backed by entity-level sanctions and, in at least some cases, supported also by individual executive liability. Table 5. Possible Liability Regimes   Jurisdiction . Liability Trigger . Liability Target . . Host state Harm Host state enterprise A + Host state individuals B Act Host state enterprise C + Host state individuals D Home state Harm Host state enterprise E + Host state individuals F Home state enterprise G + Home state individuals H Act Home state enterprise I + Home state individuals J Jurisdiction . Liability Trigger . Liability Target . . Host state Harm Host state enterprise A + Host state individuals B Act Host state enterprise C + Host state individuals D Home state Harm Host state enterprise E + Host state individuals F Home state enterprise G + Home state individuals H Act Home state enterprise I + Home state individuals J Open in new tab Table 5. Possible Liability Regimes   Jurisdiction . Liability Trigger . Liability Target . . Host state Harm Host state enterprise A + Host state individuals B Act Host state enterprise C + Host state individuals D Home state Harm Host state enterprise E + Host state individuals F Home state enterprise G + Home state individuals H Act Home state enterprise I + Home state individuals J Jurisdiction . Liability Trigger . Liability Target . . Host state Harm Host state enterprise A + Host state individuals B Act Host state enterprise C + Host state individuals D Home state Harm Host state enterprise E + Host state individuals F Home state enterprise G + Home state individuals H Act Home state enterprise I + Home state individuals J Open in new tab A. Host State Liability In this section I assess the likely efficacy of host state liability regimes, focusing first on entity and then on individual liability. As a general matter, host state human rights enforcement against corporations presents a central case of sanction insufficiency. First, the moral cost of severe human rights abuses is acute. Egregious harms like slavery, torture and the loss of life, for example, are essentially not compensable. Nevertheless, the state must set a sanction that reflects the gravity of such abuses and incentivises corporations to invest heavily to eliminate them. Such penalties would frequently be massive. Severe human rights abuses can be financially costly, too. Damages from the Bhopal chemical leak, for example, were estimated in the billions of dollars.103 Taking into account the chances of an accident and sanction, penalties high enough to internalise such costs would be beyond the ability of even large host state corporations to pay, to say nothing of the smaller ones capable of the same scale of disaster. Secondly, host states generally have weak enforcement environments, so there are inadequate prospects that human rights abuses will be detected and sanctioned. Host states are less likely than home states to enact robust liability regimes in the first place. Even when domestic law protects human rights, the existence of weak institutions, limited budgets and corruption means that host states or private litigants are often unable or unwilling to enforce it.104 Host state governments are particularly unlikely to pursue corporations that are accused (as they often are) of assisting primary abuses by those governments themselves.105 Abuses may also go undetected: victims tend to be politically and economically marginalised, they may live or work in isolated geographic areas, and the media and civil society may not have the freedom or resources to expose wrongdoing. On account of these enforcement obstacles, the host state corporation will greatly discount the nominal amount of entity-level sanctions it faces, dampening their impact on its expected liability costs. To increase deterrence, already-inflated penalty levels must go even higher.106 Thirdly, host state corporations with significant human rights exposure may be undercapitalised relative to their risk profiles. Limited liability incentivises corporations to shift risk disproportionately to underfunded subsidiaries.107 There is empirical evidence that corporations in some of the most worrisome industries act on this incentive.108 Indeed, in human rights litigation against corporations, victims frequently allege that host state subsidiaries are undercapitalised. These limited assets cap their expected liability costs, such that continuing to raise sanctions achieves no additional deterrence among them and bankrupts those that are actually sanctioned. These problems are most forcefully present in cases of harm-based liability,109 in which the corporation discounts nominal sanctions by both the chance of an accident and the chance of sanction. However, they will limit the efficacy of act-based due diligence regimes, too.110 Although such regimes may require lower sanction levels, incentivising costly investments in due diligence will still require hefty sanctions relative to a corporation’s size. The impact of these sanctions will still be dampened by the limited assets of host state corporations. And, even assuming a host state’s willingness to adopt a due diligence regime, an institutionally weak and under-resourced enforcement environment means that successful sanction will remain unlikely. These problems of sanction insufficiency will likely be compounded by internal disciplinary failures. Where human rights abuses are the product of acts and omissions by many agents undertaking complex, decentralised activities over time, it may be difficult or prohibitively expensive for the corporation to determine whether a given agent has taken care to prevent the abuse. Moreover, host state agents will often be relatively poor, such that the corporation has limited options for imposing contractual penalties for carelessness. Under these conditions, even a well-motivated corporation may struggle to incentivise rights-respecting behaviour by its agents. In general, then, under a regime of entity liability—whether harm- or act-based—it is likely that a host state corporation’s expected liability costs will substantially understate the human rights costs of its commercial activity. It will invest too little in production safety and too many abuses will occur. Supplementary individual liability is unlikely to meaningfully improve the efficacy of host state regimes.111 If enforcement is weak against corporations, it will be at least as weak against their agents; weaker, perhaps, because of the information asymmetries faced by the state or injured parties as they try to pinpoint culpable individuals in a large organisation. Furthermore, if those agents are poor, that limits the state’s punitive options as much as the corporation’s. Imprisonment may mitigate the problem of judgment-proof agents, but this would often be inappropriate, and is perhaps especially so as a means for enforcing human rights.112 The extent to which the foregoing factors will undermine the success of a given host state liability regime depends, as I have noted, on the specific human rights to be protected, the capacity for enforcement and the availability of corporate assets (among other contextual matters), which will vary between host states and industries. In general, however, these considerations suggest that host state liability regimes will substantially under-deter severe human rights abuses. B. Harm-Based, Home State Liability In light of these challenges, it is small wonder that human rights lawyers have pinned their hopes on home states. Two strategies are current. The first (discussed in this subsection) pursues harm-based tort liability in home state courts. The second (discussed in the next subsection) advocates home state legislation of act-based due diligence regimes. Tort actions in home states take two forms. The first involves suing the host state subsidiary (and potentially its agents) in the courts of its home state parent.113 One difficulty with this approach is that it leaves most causes of sanction insufficiency untouched. The costs of severe human rights abuses remain high; the asset pool of the host state corporation and its agents remains limited; any internal disciplinary failures within the host state corporation persist. In theory, this strategy alleviates sanction insufficiency in one respect: it improves enforcement prospects by abandoning weak institutional environments for stronger ones. However, this gain is offset by new enforcement challenges. These include the complex jurisdictional hurdles that such suits entail on account of their ‘foreign-cubed’ nature: the home state court must adjudicate allegations by a foreign plaintiff of foreign conduct by a foreign defendant.114 Basic principles of sovereignty and international comity often militate against exercising jurisdiction under these circumstances.115 The second form of home state tort suit involves joining the home state parent (and potentially its home state agents) as a defendant.116 If it works, this strategy alleviates two causes of sanction insufficiency: by introducing a home state defendant, it clears the jurisdictional hurdle just described and opens the path to robust enforcement by well-functioning home state courts; and it facilitates access to the parent’s larger asset pool. Again, however, the strategy introduces a new and fundamental enforcement challenge: the doctrine of separate legal personality prevents the attribution of the harm, which occurs in the host state, to the home state parent.117 The success of such suits thus requires attributing some wrongdoing by the subsidiary (or, even more challengingly, by a supplier) to the parent itself. This is tricky: not only must the victims prove that the home state parent owed a duty of care to host state victims, but also that the breach of that duty caused their injuries.118 The introduction of parent liability for host state conduct may also exacerbate internal disciplinary failures: even a properly motivated parent may have difficulty monitoring and enforcing careful behaviour by agents doing complex tasks in foreign countries.119 Thus, without far-reaching reforms to the laws of jurisdiction, separate personality and/or tort liability, home state regimes of harm-based liability present limited chances of success. Even huge sanctions will often translate to low expected liability costs for corporations and their agents, yielding insufficient investments in production safety. Given the widespread persistence of severe human rights abuses in transnational commercial contexts, it is clear that the existing availability of these tort-based strategies has done little to deter them. C. Act-Based, Home State Liability for Due Diligence High-income home states are increasingly enacting legislation aimed at incentivising their domestic corporations to undertake human rights due diligence throughout their global operations. Initial efforts were restricted to modern slavery and imposed only reporting obligations, requiring corporations to disclose the extent of the diligence, if any, that they undertake to eliminate slavery in their supply chains.120 More recent legislation goes further, mandating corporations to undertake due diligence—not just within their subsidiaries, but throughout their supply chains—to prevent certain human rights abuses. It is this kind of act-based due diligence regime—already adopted in France and the Netherlands, and now being considered by the European Commission and several states—that I address here.121 In comparison to the options considered above, a home state regime of act-based due diligence offers greater prospects of deterring severe human rights abuses. This is so even if the regime is supported by entity sanctions alone. To begin with, such regimes surmount the foregoing enforcement challenges. Due diligence would be imposed by the home state on its own corporations. Much of the conduct it would require of them (principally, designing and implementing internal controls) would happen within the home state corporation on home state soil. The obligation is thus suited to the jurisdiction and supervision of home state courts and regulators. A home state due diligence regime has relatively strong enforcement prospects. First, it avoids the difficulties of attributing host state wrongdoing to the home state parent. Secondly, there is no complex inquiry into causal links between non-compliance and harm. The principal question is whether the corporation has implemented the required controls.122 Thirdly, many of the controls that due diligence requires—organisation-wide systems for identifying and tracking human rights risks, and operational plans for preventing and remediating them—would be relatively visible to well-resourced home state regulators. Thus, a home state due diligence regime creates not only a home-grown obligation that is enforceable against the parent, but one that carries reasonable prospects of enforcement. Precisely for this reason, it alleviates another cause of sanction insufficiency: it succeeds in putting the parent’s assets in play. Unlike underfunded host state subsidiaries, the substantial sanctions that may be needed to induce effective due diligence are less likely to exceed the assets of wealthier parent corporations, and are thus more likely to incentivise the parent to invest heavily in production safety throughout its operations. A home state due diligence regime has advantages beyond alleviating sanction insufficiency. First, if it induces substantial investment in internal controls, due diligence will mitigate some internal disciplinary failures. Since many controls would need to detect potentially harmful conduct by the corporation’s agents, they would assist the corporation in monitoring and enforcing standards of care. Secondly, home state due diligence regimes mitigate the deregulatory race to the bottom among host states. If a home state requires its corporations to implement due diligence throughout their operations, it removes some incentives for those corporations to shop around for the least regulated host. Finally, and perhaps most importantly, an act-based due diligence regime allows home state regulators to proactively investigate and enforce corporate compliance with the required internal controls, rather than deferring their involvement until a human rights abuse actually occurs. Corporations that do not comply—because, for example, they have incompetent managers or underestimate the likelihood of sanctions—can be caught and corrected prior to the occurrence of an abuse, thereby securing ex ante prevention as a priority over ex post compensation. For these reasons, an act-based due diligence regime backed by entity-level sanctions presents strikingly better deterrent prospects than the other liability regimes I have considered. In some contexts, this may be enough to incentivise substantial due diligence investments by corporations and thereby reduce the number and severity of human rights abuses. The question is whether entity-level sanctions alone will achieve the desired level of deterrence in all cases. There are at least two reasons to think they may not. First, there will be cases where entity-level sanctions result in substantial, but still incomplete, reductions in human rights abuse, despite significant due diligence expenditure. Depending on the nature of the abuse in question, the state might properly determine that those residual abuses are justified by and proportionately related to its pursuit of some other legitimate goals (including economic ones). However, in industries posing grave and systemic threats to the most fundamental aspects of human well-being (like life, bodily integrity or freedom from slavery), the state might decide that something closer to a complete elimination of such threats is required. It is plausible that the marginal costs of the extra due diligence needed to reach this state of affairs would exceed the marginal expected liability costs imposed by any realistic increase in entity-level sanctions, such that a profit-maximising corporation would not incur the former to offset the latter. Put differently, the financial and moral cost of certain harms may be so high, and the desire to eliminate them so strong, that no entity-level sanction could achieve the desired level of deterrence. To mitigate this situation of sanction insufficiency, the state must add qualitatively, not just quantitatively, different sanctions.123 Secondly, as I argued earlier, executives carry much of the initial and ongoing burden of implementing and monitoring due diligence efforts. Even when entity liability is such that the corporation’s profits would be maximised by substantial due diligence, the required effort may not be utility-maximising for executives. The corporation could induce their effort with contractual rewards and penalties, but the executives themselves are the ones who structure those contracts and monitor their own compliance with them. These executives will therefore have both the motive and the opportunity to shirk due diligence responsibilities that the corporation itself might prefer them to undertake. As discussed, the corporation’s ability to reduce those incentives with tools like stock-based compensation or fiduciary suits may be limited. Where either or both problems obtain, the solution is to impose supplementary individual liability for due diligence failures on some subset of executives. In cases of sanction insufficiency, targeting executives with individual sanctions will get them to invest corporate resources in due diligence beyond what would be profit-maximising for the corporation. In cases of internal disciplinary failure among executives, individual liability transfers from those executives to the state the decision about whether to enforce the required standards of care and effort. In either case, the result would be a greater investment in due diligence, and thus greater deterrence of severe human rights abuses, than would be achieved under a regime of entity sanctions alone. Indeed, the primary objection to supplementary individual liability is that it may generate too much deterrence.124 Depending on the amounts of the respective entity- and individual-level sanctions and the type of abuse to be prevented, the cost increase for corporations and consumers may exceed the perceived social gains of additional deterrence. This objection relies on the logic of social welfare maximisation. It has limited force where the state, on grounds of moral principle, seeks to deter some grave harm as completely as possible, even at the risk of imposing substantial costs on corporations and consumers. I have suggested that this ought to be the state’s approach to severe human rights abuses. However, even if one adopts a welfare-maximising approach, the concern about overdeterrence does not bite in situations of sanction insufficiency, where, by definition, corporations have internalised too few of their true costs of doing business. The objection fares better in relation to internal disciplinary failures, where it can be addressed simply by reducing the entity-level sanction appropriately.125 On account of the complex empirical matters on which it turns, it is difficult accurately to guess the extent of overdeterrence in advance. Where this is a concern, it counsels an incrementalist approach to sanction design for due diligence. The state might begin with a regime of moderate entity sanctions and then monitor corporate compliance and executive-level engagement. Over time, the state could increase those sanctions and/or add supplementary individual liability as required to reach the desired level of deterrence. 5. Conclusion There is a developing obligation in international law for states to design effective corporate liability regimes for transnational human rights abuses. To meet this obligation, states must answer many complex questions which turn on jurisdiction-specific matters of legal and empirical context. I have offered general and tentative considerations addressing only some of these questions. I have not considered, for example, the matter of strict versus fault-based liability. Here, I merely note that only strict entity liability forces corporations to internalise the full cost of the harm they cause,126 and also alleviates evidentiary burdens of proving corporate fault. I have also ignored civil versus criminal liability; elsewhere, I have discussed the weakness of the case for corporate criminal liability and offered further reasons against it.127 Nor have I discussed the relative merits of private and public enforcement,128 or the extent to which substantial liability may incentivise corporations to relocate to less demanding regimes.129 For regimes adopting individual liability, there are many forms this might take. The Dutch child labour law, for instance, threatens executives with jail time for repeated due diligence failures by their corporation.130 In a different context, John Armour and Jeffrey Gordon (relying partly on considerations of sanction insufficiency) have argued for court-imposed fiduciary duties requiring banking executives to manage systemic financial risk.131 Courts could develop such duties to encompass human rights risks, too. The Sarbanes-Oxley Act in the United States offers yet another example. It requires certain executives to sign a quarterly statement, under penalty of fraud, declaring that they have designed and recently evaluated an effective system of internal controls for financial reporting.132 Similar statements could be mandated for human rights due diligence. Regardless of the form that individual liability takes, the law, as I have noted, would typically have to prohibit executives from contractually shifting it to the corporation through insurance or indemnification.133 The law would likely also require that individual sanctions be pegged to an individual executive’s compensation, to prevent corporations from simply offsetting the executive’s increased liability costs with bonuses or other forms of pay.134 Despite its limited ambit, my analysis yields at least three important principles that might inform current and future reform efforts. First, absent significant changes to the general rules governing harm-based liability in home states, the home state adoption of an act-based due diligence regime is an essential tool for preventing severe human rights abuses. Secondly, due diligence for severe human rights abuses requires substantial corporate expenditure that only substantial entity-level sanctions can induce. Regimes with no or weak sanctions for non-compliance will be inadequate deterrents. Finally, the UNGPs, the Zero Draft and the French legislation address due diligence requirements exclusively of the corporate entity, while the Dutch legislation currently stands alone in providing for supplementary executive liability for due diligence failures.135 I have argued that the most egregious human rights abuses may be under-deterred without supplementary executive liability, and may thus offer some support for this aspect of the Dutch model. For helpful feedback on an earlier draft, I thank Ekaterina Aristova, Shreya Atrey, Miles Jackson, Christos Kypraios, Liora Lazarus, Stephen Meili, Russell Perkins, Ewan Smith, Stefan Theil and participants in a seminar at the Bonavero Institute of Human Rights at which this work was presented. I also thank the reviewers for their thoughtful comments. Errors are my own. Footnotes 1 UN Office of the High Commissioner for Human Rights, ‘Guiding Principles on Business and Human Rights: Implementing the “Protect, Respect and Remedy” Framework’, HR/PUB/11/04, 2011 <www.ohchr.org/documents/publications/GuidingprinciplesBusinesshr_eN.pdf> accessed 14 July 2020. 2 ibid principle I.A.1. For an introduction to the history and framework of the UNGPs, see John Ruggie, Just Business (WW Norton & Company 2013); David Bilchitz and Surya Deva, ‘The Human Rights Obligations of Business: A Critical Framework for the Future’ in Surya Deva and David Bilchitz (eds), Human Rights Obligations of Business (CUP 2013). 3 UN Open-Ended Intergovernmental Working Group on Transnational Corporations and Other Business Enterprises with Respect to Human Rights, ‘Legally Binding Instrument to Regulate, in International Human Rights Law, the Activities of Transnational Corporations and Other Business Enterprises’ (second revised draft, 6 August 2020) <https://www.ohchr.org/Documents/HRBodies/HRCouncil/WGTransCorp/Session6/OEIGWG_Chair-Rapporteur_second_revised_draft_LBI_on_TNCs_and_OBEs_with_respect_to_Human_Rights.pdf> accessed 5 November 2020. 4 Zero Draft (n 3) arts 8.1 and 8.4; see also art 6. 5 See Business and Human Rights Resource Centre, ‘National Movements for Mandatory Human Rights Due Diligence in European Countries’ <www.business-humanrights.org/en/national-movements-for-mandatory-human-rights-due-diligence-in-european-countries> accessed 14 July 2020. 6 A notable exception is Surya Deva, Regulating Corporate Human Rights Violations (Routledge 2012). Deva argues for an ‘integrated theory of regulation’, according to which corporations should be subject to a range of incentives and sanctions across institutional, national and international levels. He offers a typology of such incentives and sanctions, but does not advocate for any particular combination. 7 I use the term ‘executives’ broadly to denote directors, officers and senior managers of a corporation. 8 It may also interest scholars of corporate social responsibility (CSR), commonly understood as a corporation’s voluntary compliance with moral or social expectations, including, but not limited to, respect for human rights. Owing to its perceived inefficacy, business and human rights scholars increasingly reject a CSR approach, so conceived, in favour of legally binding obligations. See Justine Nolan, ‘With Power Comes Responsibility: Human Rights and Corporate Accountability’ (2005) 28 UNSWLJ 581, 582–3; Michael Addo, ‘Human Rights Perspectives of Corporate Groups’ (2005) 37 Conn L Rev 667, 672–82. My argument here supports this shift. 9 Alan Sykes, ‘Corporate Liability for Extraterritorial Torts under the Alien Tort Statute and Beyond: An Economic Analysis’ (2012) 100 Geo LJ 2161, 2164. 10 I focus on transnational abuses because of the significant obstacles to holding corporations accountable in this context. For a comprehensive review, see Gwynne Skinner and others, ‘The Third Pillar: Access to Judicial Remedies for Human Rights Violations by Transnational Business’ (December 2013) 18–63 <https://corporatejustice.org/documents/publications/eccj/the_third_pillar_-access_to_judicial_remedies_for_human_rights_violation.-1-2.pdf> accessed 14 July 2020. See also European Parliament, ‘Access to Legal Remedies for Victims of Corporate Human Rights Abuses in Third Countries’ (PE 603.475, February 2019) (EP Study) 15; Gwynne L Skinner, ‘Beyond Kiobel: Providing Access to Judicial Remedies for Violations of International Human Rights Norms by Transnational Business in a New (Post-Kiobel) World’ (2014) 46 Colum Hum Rts L Rev 158, 168–73. I focus on large, well-resourced controlling corporations because they are the typical defendants in human rights cases against corporations. EP Study 18. 11 UNGPs (n 1) principle II.A.14, commentary. 12 I have in mind the archetypal cases that fuel the business and human rights movement: Shell allegedly directing its Nigerian subsidiary to aid the Nigerian government in torturing and killing environmental protestors (see Kiobel v Royal Dutch Petroleum Co 569 US 108 (2013)); Myanmar soldiers, allegedly acting for and with the knowledge of Unocal and Total, subjecting villagers to forced labour, rape, torture and murder (see Doe v Unocal Corp, 248 F3d 915 (9th Cir 2001)); the 2013 garment factory collapse at Rana Plaza in Bangladesh that killed over 1000 workers; and the 1984 Union Carbide gas leak in Bhopal, India, which killed thousands in neighbouring communities and poisoned hundreds of thousands more. See Justine Nolan, ‘Refining the Rules of the Game: The Corporate Responsibility to Respect Human Rights’ (2014) 30 Utrecht Journal of International and European Law 7, 10–11 (describing the Bhopal and Rana Plaza disasters). 13 Andrew Clapham and Scott Jerbi, ‘Categories of Corporate Complicity in Human Rights Abuses’ (2001) 24 Hastings Int’l & Comp L Rev 339. See also UNGPs (n 1) principles II.A.13, 17 and 19. 14 Arguments drawn from morality and organisational behaviour theory are, of course, also relevant, though I do not explore them here. 15 The seminal contributions on this subject include: Guido Calabresi The Cost of Accidents (Yale UP 1970) ch 4; Christopher Stone, ‘The Place of Enterprise Liability in the Control of Corporate Conduct’ (1980) 90 Yale LJ 1; John C Coffee, ‘“No Soul to Damn, No Body to Kick”: An Unscandalized Inquiry into the Problem of Corporate Punishment’ (1981) 79 Mich L Rev 386; Lewis Kornhauser, ‘An Economic Analysis of the Choice between Enterprise and Personal Liability for Accidents’ (1982) 70 CLR 1345; Reinier Kraakman, ‘Corporate Liability Strategies and the Costs of Legal Controls’ (1984) 93 Yale LJ 857; Alan Sykes, ‘The Economics of Vicarious Liability’ (1984) 93 Yale LJ 1231; Jonathan Macey, ‘Agency Theory and the Criminal Liability of Organizations’ (1991) 71 BU L Rev 315; Kathleen Segerson and Tom Tietenberg, ‘The Structure of Penalties in Environmental Enforcement: An Economic Analysis’ (1992) 23 Journal of Environmental Economics and Management 179; Mitchell Polinsky and Steven Shavell, ‘Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?’ (1993) 13 International Review of Law & Economics 239. For general introductions to the economic analysis of law, see Steven Shavell, Foundations of Economic Analysis of Law (Harvard UP 2004); Richard Posner, Economic Analysis of Law (9th edn, Wolters Kluwer 2014). 16 Although the term ‘enterprise liability’ appears in earlier scholarship (see eg Stone (n 15); Kraakman (n 15)), it also refers to legal doctrines (like veil piercing) by which separate corporate entities are held jointly liable for some debt by virtue of a shared ownership structure. See Meredith Dearborn, ‘Enterprise Liability: Reviewing and Revitalizing Liability for Corporate Groups’ (2009) 97 Cal L Rev 195. I adopt ‘entity liability’ to avoid confusion. 17 Jennifer Arlen and Reinier Kraakman, ‘Controlling Corporate Misconduct: An Analysis of Corporate Liability Regimes’ (1997) 72 NYU L Rev 687, 695 (‘corporations don’t misbehave, people do’). See also Neil Campbell and John Armour, ‘Demystifying the Civil Liability of Corporate Agents’ (2003) 62 CLJ 290, 290. 18 Kraakman (n 15) 858–9. See also Stone (n 15) 8, 12. 19 See further 14–16 below. 20 I adopt the language of ‘accidents’, but the framework applies to a range of misconduct, including intentional wrongdoing, and is stated broadly enough to cover torts, crimes and regulatory offences. I tend also to refer to public regulatory enforcement, but the analysis holds for private tort actions by injured parties (in which case the state’s role is largely confined to supplying the relevant legal rules and imposing the proper measure of damages through its courts). See Steven Shavell, ‘Liability for Harm versus Regulation of Safety’ (1984) 13 JLS 357 (discussing the choice between tort liability and statutory regulation). 21 Below, I consider situations in which a sanction is imposed not upon the occurrence of an accident, but upon the occurrence of some precursor act or omission that causes accidents probabilistically. See nn 62–67 and accompanying text. 22 The concept of expected liability costs assumes that corporations choose between projects based on their net present value (ie by considering the nominal costs and benefits of a project and the likelihood with which they will materialise). In what follows, I assume that corporations are risk-neutral, such that they are indifferent between projects with different risk profiles as long as they bear the same expected pay-off. This simplifies but does not fundamentally alter the analysis. Shavell, Foundations (n 15) 177–8. See also Henry Hansmann and Reinier Kraakman, ‘Toward Unlimited Shareholder Liability for Corporate Torts’ (1991) 100 Yale LJ 1879, 1882 fn 6. 23 In what follows, I forgo formal mathematical models showing the precise value ranges over which my conclusions hold. (For examples of such models supporting some of my conclusions, see Segerson and Tietenberg (n 15) and Polinsky and Shavell (n 15).) Instead, I offer numerical examples to demonstrate the application and plausibility of my claims. The reader can test the rough limits of these claims through experimentation with alternative figures. 24 Shavell, Foundations (n 15) 178–9, 207–9. 25 Stone (n 15) 8, 12–13. See also John Armour and Jeffrey Gordon, ‘Systemic Harms and Shareholder Value’ (2014) 6 Journal of Legal Analysis 35, 47. 26 Either regime, under certain assumptions, yields the same deterrence at the same cost to the corporation and society. Kornhauser (n 15) 1347–60; Sykes, ‘Vicarious Liability’ (n 15) 1233–56; Segerson and Tietenberg (n 15) 182–8; Sykes, ‘Corporate Liability’ (n 9) 2183–4. In fact, some of these assumptions are implausible; relaxing them explains the need for entity liability. 27 Sykes, ‘Vicarious Liability’ (n 15) 1239–42; Steven Shavell, ‘The Judgment Proof Problem’ (1986) 6 International Review of Law & Economics 45; Segerson and Tietenberg (n 15) 191–2. See also Kornhauser (n 15) 1362–6. 28 Kraakman (n 15) 862–6; Armour and Gordon (n 25) 36, 50–1. For simplicity’s sake, I assume throughout that agents are risk-neutral; again, this does not fundamentally alter my conclusions. 29 Kornhauser (n 15) 1351; Jennifer Arlen, ‘The Potentially Perverse Effects of Corporate Criminal Liability’ (1994) 23 JLS 833, 834–5; Arlen and Kraakman (n 17) 687, 692, 695–6. 30 Kraakman (n 15) 867–8. See also Coffee, ‘No Soul’ (n 15) 389–90 (describing this phenomenon as the ‘deterrence trap’). The following discussion is consistent with Kraakman’s treatment, but departs from it in some respects. I refer to roughly the same three phenomena that undermine entity liability, though he does not label them all, as I do here, as causes of ‘sanction insufficiency’. More significantly, despite recognising that these phenomena are ‘overlapping’, Kraakman proposes that different groups of individuals should be targeted in different ways to solve each. I propose supplementary individual liability for executives as a general solution to sanction insufficiency in human rights contexts, whatever the cause. 31 cf Kraakman (n 15) 867–9 (calling this phenomenon ‘asset insufficiency’ and describing it as a form of sanction insufficiency). 32 Hansmann and Kraakman (n 22) 1879. See also Segerson and Tietenberg (n 15) 191–2; Armour and Gordon (n 25) 71; Coffee, ‘No Soul’ (n 15) 390. 33 By adjusting the amount or probability of the sanction, Table 3 can be adapted to demonstrate the results of the further causes of sanction insufficiency discussed below. 34 Corporate bankruptcy has significant knock-on effects (particularly for innocent lower-level employees and the families supported by their income) and ought to be treated with appropriate concern. Kraakman (n 15) 852; Coffee, ‘No Soul’ (n 15) 408. 35 cf Kraakman (n 15) 867–8, 876–8, 880–3. Kraakman discusses sanction insufficiency (as he uses the phrase) in quite general terms, focusing mainly on ‘practical limits on the severity of firm penalties’. ibid 880. He does not expressly discuss the case of extreme social costs, but it is consistent with his analysis. 36 cf ibid 888–9 (calling this phenomenon ‘enforcement insufficiency’ and describing its distinction from sanction insufficiency as one of degree). 37 In theory, this particular cause of sanction insufficiency could be solved by simply raising the nominal amount of the sanction. However, this once again raises the spectre of fines that exceed anyone’s ability to pay, reintroducing other causes of sanction insufficiency (and threats of bankruptcy) through the backdoor. 38 Some instances of sanction insufficiency (particularly those caused by insufficient assets) may also be mitigated by ensuring that the corporation is properly insured or capitalised. This may be appropriate where ex post compensation for harm is as good as preventing it (Kraakman (n 15) 870); it is less so for the kinds of human rights abuses I consider below, where prevention is paramount. 39 In Table 3, for example, increasing sanction prospects to 90% would cause the corporation to install the best safety device, but this may not occur if the other figures are changed. 40 Kraakman (n 15) 886. See also Daniel S Nagin and Greg Pogarsky, ‘Integrating Celerity, Impulsivity, and Extralegal Sanction Threats into a Model of General Deterrence: Theory and Evidence’ (2001) 39 Criminology 865. 41 Kraakman (n 15) 886. 42 Corporate law scholarship treats the corporation as a nexus of contracts among shareholders, managers, creditors and employees. These contracts give rise to a range of principal-agent conflicts. Reinier Kraakman and others, The Anatomy of Corporate Law (3rd edn, OUP 2017) 5–6, 29–31. See also Michael Jensen and William Meckling, ‘Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure’ (1976) 3 Journal of Financial Economics 305. For this reason, at least from the perspective of corporate law, an understanding of the economic theory of agency ‘is essential to any inquiry into the behavior of corporate actors’. Macey (n 15) 319. See also ibid 320–2; Sykes, ‘Vicarious Liability’ (n 15) 1233–9. 43 Shavell, Foundations (n 15) 179. 44 Kornhauser (n 15) 1352–5; Sykes, ‘Vicarious Liability’ (n 15) 1237; Segerson and Tietenberg (n 15) 184–5; Polinsky and Shavell (n 15) 242–3. 45 Kornhauser (n 15) 1352–60; Sykes, ‘Vicarious Liability’ (n 15) 1235–9; Polinsky and Shavell (n 15) 242–3. 46 Sykes, ‘Vicarious Liability’ (n 15) 1237; Segerson and Tietenberg (n 15) 180, 197–8. 47 Polinsky and Shavell (n 15) 242–3. 48 Sykes, ‘Vicarious Liability’ (n 15) 1236–7, 1247–55; Segerson and Tietenberg (n 15) 179–80. 49 Macey (n 15) 320. As Kraakman notes ((n 15) 863), ‘managers will manage with an eye to increasing their own expected utility by maximizing future compensation including salary, job tenure, promotion prospects, informal perquisites, and opportunities for consuming leisure and other goods on the job’. 50 Jennifer Arlen and William Carney, ‘Vicarious Liability for Fraud on Securities Markets: Theory and Evidence’ (1992) U Ill L Rev 691, 708–16. See also Coffee, ‘No Soul’ (n 15) 408 (arguing that ‘evidence of internal discipline within large corporations is conspicuously absent at senior corporate levels’); John Coffee, ‘Corporate Crime and Punishment: A Non-Chicago View of the Economics of Criminal Sanctions’ (1980) 17 Am Crim L Rev 419, 443. 51 Steven Shavell, ‘Criminal Law and the Optimal Use of Nonmonetary Sanctions as a Deterrent’ (1985) 85 Colum L Rev 1232. 52 Kraakman (n 15) 867, 878–80, 885; Armour and Gordon (n 25) 71. Furthermore, as Kraakman argues ((n 15) 880), a dual regime of entity and individual liability ‘provides two distinct pressure points and thus two chances for successful deterrence … Because firms and managers may be affected differently by the same kind of sanction—for example, punitive damages or fines—imposing these penalties on both firms and managers “diversifies” their impact, and lessens the risk that they were set too low for either of their prospective targets.’ With the addition of non-monetary sanctions against individuals, ‘dual liability can also diversify across qualitatively different kinds of sanctions’. ibid (emphasis in original). 53 Like the corporation, the executive’s expected liability costs are a function of the nominal sanction she faces discounted by the chances of an accident and successful sanction (which I assume here is the same for both the corporation and the executive). Imposing liability on the executive when an accident occurs (regardless of whether she is at fault) is, of course, a simple form of liability; more complex, fault-based forms are possible. 54 Note that the same circumstances giving rise to sanction insufficiency with respect to corporations may give rise to the same with respect to individuals. Where a sanction exceeds the individual’s ability to pay—because it is too high, or the agent’s assets or chances of being successfully sanctioned are too low—she will be under-incentivised to take care. Where this occurs in combination with a failure of entity liability, it may simply be impossible for the state to achieve the desired level of deterrence. 55 Again, insurance is less problematic when the liability regime’s primary concern is ensuring that victims are fully compensated for their losses rather than prioritising the prevention of accidents. 56 For a formal model demonstrating this result, see Polinsky and Shavell (n 15). See also Segerson and Tietenberg (n 15) 190–1, 198. 57 cf Arlen and Carney (n 50) 708–16 (arguing that individual liability for securities fraud shifts the enforcement decision from the board to injured plaintiffs who have greater incentives to pursue sanctions). 58 The effect of stock options on executive decision making is much debated. See Kraakman and others (n 42) 66 (noting that ‘managerial rewards can … be as much a strategy for controlling agency costs as a symptom of them’); Lucian Bebchuk and others, ‘Managerial Power and Rent Extraction in the Design of Executive Compensation’ (2002) 69 U Chi L Rev 751. 59 See eg Companies Act 2006, ss 172 (requiring directors to promote the company’s success, albeit after considering factors affecting a broad range of stakeholders) and 174 (requiring directors to exercise care). 60 Jennifer Hill and Matthew Conaglen, ‘Directors’ Duties and Legal Safe Harbours: A Comparative Analysis’ in Gordon Smith and Andrew Gold (eds), Research Handbook on Fiduciary Law (Edward Elgar 2018). Under Delaware law, for example, a company may exculpate directors for breach of the duty of care, even grossly negligent ones. See Del Code tit 8, § 102(b)(7). 61 Even where these tools operate effectively, ‘Monitoring and evaluating the performance of top managers are simply too difficult to permit anything like a total convergence of interest between managers and shareholders’. Kraakman (n 15) 864. 62 Stone (n 15) 11–19 (describing act-based liability regimes in terms of the imposition of ‘standards’); Shavell, Foundations (n 15) 478–9. See also Steven Shavell, ‘Liability for Harm versus Regulation of Safety’ (1984) 13 JLS 357 (discussing the choice between tort liability and statutory regulation), 478–9. 63 Of course, the state may also require the corporation to refrain from certain acts. For present purposes, nothing turns on this distinction between acts and omissions. 64 See Shavell, Foundations (n 15) 478 (explaining this feature of act-based liability from a welfare-maximising perspective). 65 Table 1, for example, depicts a regime of harm-based liability. There, a nominal fine of £1 million, after discounting for the 20% chance of an accident (with no safety device) and a 10% chance of sanction, translates to expected liability costs of £20,000. Under an act-based regime, where the nominal fine for failing to install a safety device is discounted only by the 10% chance of sanction, a fine of just £200,000 achieves the same expected liability costs for the corporation. 66 Stone (n 15) 17–18. 67 These benefits come at a price: act-based liability is informationally more demanding on the state than harm-based regimes, in part because the state must accurately estimate not only the cost and likelihood of the harm that the corporation causes, but also the efficacy of various means of harm prevention. See Shavell, Foundations (n 15) 478. In what follows, I will simply assume that the state can cheaply and accurately satisfy its informational needs (particularly with respect to the due diligence obligation), though in reality this would influence the choice of regime. 68 See eg Kornhauser (n 15); Sykes, ‘Vicarious Liability’ (n 15); Arlen and Kraakman (n 17); Polinsky and Shavell (n 15). 69 Shavell, Foundations (n 15) 178–82, 207–11. See also Kornhauser (n 15) 1354; Segerson and Tietenberg (n 15) 185; Polinsky and Shavell (n 15) 239–44. 70 Some social costs—like the corporation’s production costs—are already internal to the corporation’s profit-maximisation function. A sanction, on the welfare-maximising approach, should internalise those costs that remain external. Such a sanction should be set equal to the social cost of the harm, appropriately inflated for the chances of an accident and successful sanction, plus the public’s costs of enforcing the sanction. Mitchell Polinsky and Steven Shavell, ‘Enforcement Costs and the Optimal Magnitude and Probability of Fines’ (1992) 35 JLE 133; Arlen (n 29) 834–5. 71 See Table 1. 72 It also bears the costs of enforcing the liability regime. See n 70 above. 73 This inevitably associates a price with the abuse, but unlike the welfare-maximising approach, the price is arrived at indirectly and ex post, as a by-product of determining which sanction will achieve the desired deterrence. 74 UNGPs (n 1) principles II.A.15 and II.B.17–24. 75 United Nations General Assembly, ‘Report of the Working Group on the Issue of Human Rights and Transnational Corporations and Other Business Enterprises’ A/73/163 (16 July 2018) (Working Group Report 2018) paras III.A.20–3. 76 ibid paras I.B.5 and II.B.16. 77 See nn 120–121 below and accompanying text. 78 See eg Irene Pietropaoli and others, ‘A UK Failure to Prevent Mechanism for Corporate Human Rights Harms’ (BIICL 2017) 36–49; Doug Cassel, ‘Outlining the Case for a Common Law Duty of Care of Business to Exercise Human Rights Due Diligence’ (2016) 1 BHRJ 179; Cees Van Dam, ‘Tort Law and Human Rights: Brothers in Arms—On the Role of Tort Law in the Area of Business and Human Rights’ (2011) Journal of European Tort Law 221; Lucien J Dhooge, ‘Due Diligence as a Defense to Corporate Liability Pursuant to the Alien Tort Statute’ (2008) 22 Emory Int’l L Rev 455. 79 See eg Olivier De Schutter, ‘Towards a New Treaty on Business and Human Rights’ (2015) 1 BHRJ 41, 53; Skinner, ‘Beyond Kiobel’ (n 10) 261. 80 The French law also provides for harm-based liability: injured parties can sue for damages caused by a corporation’s non-compliance with its due diligence obligations. See European Parliament (n 10) 133. 81 By contrast, the extent to which harm-based schemes incentivise due diligence turns on the severity and prospects of the underlying tort or criminal liability (with respect to which due diligence operates as a defence). As discussed below, such incentives can vary greatly and may sometimes be minimal. 82 Companion Note II to the Working Group Report 2018 (n 75) 2; see also Working Group Report 2018 (ibid) para II.A.13; OECD, ‘OECD Due Diligence Guidance for Responsible Business Conduct’ (OECD Publishing 2018) 16. 83 These approaches to due diligence are distinct, but not mutually exclusive: a corporation could be liable for actual human rights abuses and liable for non-compliance with statutory due diligence requirements. However, to eliminate perverse incentives for corporations to forgo due diligence, compliance with the latter may (depending on the structure of the respective legal regimes) need to excuse or mitigate liability for the former. See Arlen (n 29) 836–7. 84 UNGPs (n 1) principles 17–21. 85 OECD Guidance (n 82). 86 ibid 67. 87 ibid 61–7. 88 ibid 75. 89 ibid 32, 75, 81. 90 ibid 83. 91 ibid 82–3. 92 Companion Note II to the Working Group Report 2018 (n 75) 13. 93 Jonathan Bonnitcha and Robert McCorquodale, ‘The Concept of “Due Diligence” in the UN Guiding Principles on Business and Human Rights’ (2017) 28 EJIL 899, 901–2. This language also links the concept to tort liability. ibid 902–3; Robert McCorquodale and others, ‘Human Rights Due Diligence in Law and Practice: Good Practices and Challenges for Business Enterprises’ (2017) 2 BHRJ 195, 198. 94 Committee of Sponsoring Organizations of the Treadway Commission, ‘Internal Control—Integrated Framework, Executive Summary’ (May 2013) 8 <www.coso.org/Documents/990025P-Executive-Summary-final-may20.pdf> accessed 14 July 2020. 95 ibid 4–5. Internal controls are most commonly thought of as means to ensure accurate financial reporting, but the framework is applicable to a broad range of business objectives, including the prevention of human rights abuses. 96 This should not imply that due diligence is exclusively a matter of internal controls, only that it is mainly so. 97 UNGPs (n 1) principles 13 (commentary) and 17. 98 UNGPs (n 1) principle 17; OECD Guidance (n 82) 17; Companion Note II to the Working Group Report 2018 (n 75) 2, 6. 99 OECD Guidance (n 82) 17. See also UNGPs (n 1) principle 17 (stating that due diligence will ‘vary in complexity with the size of the business enterprise, the risk of severe human rights impacts, and the nature and context of its operations’). 100 OECD Guidance (n 82) 23, 29. 101 McCorquodale and others (n 93) 212. 102 In what follows, I assume that harm-based regimes are enforced by injured parties while act-based regimes are enforced by the state. 103 Amnesty International, ‘Clouds of Injustice: Bhopal Disaster 20 Years On’ (Amnesty International Publications 2004) <www.amnesty.org/download/Documents/96000/asa200152004en.pdf> accessed 14 July 2020. 104 Skinner, ‘Beyond Kiobel’ (n 10) 168–73; Gwynne Skinner, ‘Rethinking Limited Liability of Parent Corporations for Foreign Subsidiaries’ Violations of International Human Rights Law’ (2015) 72 Wash & Lee L Rev 1769, 1799–803. 105 Sykes, ‘Corporate Liability’ (n 9) 2194. 106 I assume that there is little scope in most host states to improve rates of detection and sanction substantially through increased enforcement expenditure. 107 Hansmann and Kraakman (n 22) 1879. See also Stone (n 15) 70–6; Skinner, ‘Rethinking’ (n 104) 1803–6. 108 Hansmann and Kraakman (n 22) 1881 and the literature cited there. 109 See Table 5, option A. 110 See Table 5, option C. 111 See Table 5, options B and D. 112 See Liora Lazarus, ‘Positive Obligations and Criminal Justice: Duties to Protect or Coerce’ in Lucia Zedner and Julian Roberts (eds), Principled Approaches to Criminal Law and Criminal Justice (OUP 2012). Of course, more general arguments against imprisonment as an enforcement tool—including on grounds of economic and racial inequality—might also be made. 113 See Table 5, options E and F. 114 Morrison v Nat’l Australia Bank Ltd, 561 US 247, 283 fn 11 (2010) (Stevens J). For a discussion of these hurdles, see Skinner and others (n 10) 18–30. 115 See eg the discussion in Kiobel (n 12) 115–17 (Roberts CJ), 127–9, 132–3 (Breyer J). 116 See Table 5, options G and H. 117 Skinner, ‘Beyond Kiobel’ (n 10) 212–19. 118 A rare, recent and still tentative example of such success is the UK Supreme Court’s decision in Lungowe v Vedanta Resources plc [2019] UKSC 20, in which the Court accepted as a preliminary matter that the English parent company may owe a duty of care to Zambian victims with respect to pollution caused by its Zambian subsidiary. ibid paras 44, 61–2. Even so, the Court made clear that it would have declined jurisdiction had the applicants not demonstrated that they could not access justice in Zambia. ibid paras 87–102. The decision also turned crucially on a European Union regulation providing for the English courts’ jurisdiction. ibid paras 4, 16. The case remains ongoing. See also the similar preliminary ruling by the Canadian Supreme Court in Nevsun Resources Ltd v Araya 2020 SCC 5. In general, tort suits against home state parents face difficult evidentiary burdens. Skinner and others (n 10) 43–5. See also James Goudkamp, ‘Duty of Care between Actors in Supply Chains’ (2017) JIPL 205 (noting the difficulties of establishing a duty of care between actors in supply chains). 119 Sykes, ‘Corporate Liability’ (n 9) 2205. 120 See California Transparency in Supply Chains Act of 2010 (US), Cal Civ Code § 1714.43; Modern Slavery Act 2015 (UK), s 54; Modern Slavery Act 2018 (Australia), ss 5, 13. 121 See n 5 above. For an overview of the French and Dutch laws, see Chiara Macchi and Claire Bright, ‘Hardening Soft Law: The Implementation of Human Rights Due Diligence Requirements in Domestic Legislation’ in Martina Buscemi and others (eds), Legal Sources in Business and Human Rights (Brill 2020); European Parliament (n 10) 132–3. 122 This benefit would be realised most fully in strict liability regimes. Attributing fault to corporations is notoriously challenging. Nick Friedman, ‘Corporations as Moral Agents: Trade-Offs in Criminal Liability and Human Rights for Corporations’ (2020) 83 MLR 255, 258–61. 123 As noted earlier, a combination of entity and individual liability may also be cheaper, from a public enforcement perspective, than a unitary regime of massive entity-level sanctions. 124 Kraakman (n 15) 864–7, 883–7; Polinsky and Shavell (n 15) 243–4. 125 Polinsky and Shavell (n 15) 240–1, 243–4. 126 Shavell, Foundations (n 15) 179–81. See also Steven Shavell, ‘Strict Liability versus Negligence’ (1980) 9 JLS 1. I assume, on moral grounds, that individual liability should be strict only in exceptional cases (if any). 127 Friedman (n 122). 128 See eg Shavell, ‘Liability for Harm’ (n 20). 129 Sykes, ‘Corporate Liability’ (n 9) 2196–7. Solving this problem requires a coordinated approach (likely in the form of an international, or at least multilateral, treaty) among the high-income states most attractive to parent corporations. 130 Macchi and Bright (n 121). 131 Armour and Gordon (n 26). 132 18 USC § 1350 (providing for up to $5 million fines or 20 years’ imprisonment for false statements). 133 See nn 18–19 above and accompanying text. See also Stone (n 15) 12; Kraakman (n 15) 858–9; Armour and Gordon (n 25) 70. 134 Armour and Gordon (n 25) 69. cf Kraakman (n 15) 878–80 (arguing that such compensation packages are difficult to arrange). 135 The reporting obligations in the UK, Australia and California are also addressed exclusively to the corporate entity. See California Transparency in Supply Chains Act of 2010 (US), Cal Civ Code § 1714.43; Modern Slavery Act 2015 (UK), s 54; Modern Slavery Act 2018 (Australia), ss 5, 13. © The Author(s) 2020. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Corporate Liability Design for Human Rights Abuses: Individual and Entity Liability for Due Diligence
JF - Oxford Journal of Legal Studies
DO - 10.1093/ojls/gqaa052
DA - 0018-06-12
UR - https://www.deepdyve.com/lp/oxford-university-press/corporate-liability-design-for-human-rights-abuses-individual-and-7pS0LUf0yT
SP - 1
EP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -