Sweatshops, Structural Injustice, and the Wrong of Exploitation: Why Multinational Corporations Have Positive Duties to the Global Poor

For views according to which exploitative transactions are wrong because they are unfair, see Wertheimer (1996, Ch. 7), Meyers (2004, pp. 320–321, 324), Mayer (2007a, pp. 137–138, 141–142, b, p. 608), Barnes (2013, p. 31), Dänzer (2014), Ferguson (2016, pp. 953, 955, 966–967), Sollars and Englander (2018, pp. 23–27), and Kates (2019, pp. 33–34, 44–45). For the view that exploitative transactions are wrong because they are disrespectful or degrading, see Wood (1995) and Sample (2003). For a respect-based account of the wrong of employing workers in sweatshop conditions that is not framed in terms of exploitation, see Arnold and Bowie (2003) and (2007).

For the claim that taking advantage of a person is an essential feature of exploitation on virtually all plausible views, see Zwolinski (2012, p. 156), Barnes (2013, p. 28) and Vrousalis (2018, p. 2).

At least most critics of sweatshops hold that employers are guilty of wrongful exploitation of workers only if they could employ them under more favorable conditions (e.g., by paying higher wages or improving workplace safety). See, for example, Meyers (2004, p. 329), Mayer (2007b), Snyder (2008, pp. 390, 398, 400–401, 404), Ferguson (2016), and Kates (2019, p. 44). Those who believe that at least much sweatshop employment is not wrongfully exploitative sometimes appeal to the claim that firms simply could not employ their sweatshop workers in conditions that their opponents would regard as non-exploitative. See, for example, Sollars and Englander (2018, p. 20).

I assume, for the purposes of this article, that they are at least roughly correct. I cannot, however, offer a defense of them here.

For the claim that the fact that sweatshop workers are disadvantaged by background or structural injustice is relevant to their status as wrongfully exploited by their employers, see Meyers (2004, p. 328), Snyder (2008, p. 392), Kates (2019, pp. 36–37).

The qualifier “wrongful” is important here because, as I understand it, the Nonworseness Claim does not present any particular threat to the claim that a transaction is exploitative, since there are plausible accounts of what exploitation is on which it is at least possible for an action or transaction to be both exploitative and permissible. Rather, the Claim presents a challenge only to the view that a transaction that is consensual and mutually beneficial could be impermissible, despite the fact that refraining from transacting is permissible. I am grateful to an anonymous reviewer for encouraging me to clarify this.

For endorsement of this claim, see Meyers (2004), Zwolinski (2007, p. 699) and (2012, p. 169), Barnes (2013, p. 38), Kates (2019, pp. 27, 34), and Preiss (2018, pp. 5–6, 10).

The claim that sweatshop employment is at least often voluntary and beneficial to workers is defended by Zwolinski (2007, pp. 691–695). For acceptance of both components of this claim by sweatshop critics, see Arnold and Bowie (2003, pp. 229, 231), Meyers (2004), Mayer (2007a, pp. 141–142, b, p. 605), Snyder (2008, p. 390), Kates (2019), and Miklós (2019, p. 61).

When a person is wronged by another agent, she has a complaint on her own behalf against that agent’s wrongful conduct that others who were not wronged do not have, and, at least typically, she is entitled to make claims on the wrongdoer that others are not entitled to make (e.g., for compensation, or for an apology). My argument in this article requires only this widely accepted claim about what follows from the fact that one agent has wronged another. I take no particular position on what makes it the case that a person is wronged by particular forms of wrongful conduct, or, to put it in the terms in which it is often discussed, what makes it the case that agents’ duties are directed toward particular others, such that those others would be wronged if the duties are not complied with. For a general discussion of these issues, see May (2015). For competing views, see Thompson (2004), Darwall (2006), Sreenivasan (2010), May (2012), Cruft (2013), Hedahl (2013), and Steiner (2015).

Practices such as these are characterized as uncontroversially morally unacceptable in Meyers (2004, p. 319), Zwolinski (2007, pp. 710–712), and Snyder (2008, p. 389).

Although this narrow focus on cases that do not involve practices that are uncontroversially wrong is common in the philosophical literature on the ethics of sweatshop employment, and is necessary for my philosophical purposes in this article, it is important that we do not lose sight of the fact that at least many actual sweatshop employees do face harassment of various kinds, non-payment of wages, deception about working conditions, and other widely acknowledged types of morally unacceptable treatment. This fact is plausibly relevant to whether and under what conditions we ought to support the introduction of facilities that might turn out to be sweatshops (in either the narrow sense that I adopt for the purposes of the paper or the broader, more colloquial sense) into communities in the developing world. Nonetheless, focusing, at least initially, on cases that do not involve abuse, deception, or other uncontroversial wrongs is an essential starting point for developing a comprehensive view of the conditions that are relevant to an overall assessment of the ethical status of a range of different employment arrangements and conditions that are found in the actual world. I am grateful to an anonymous referee for encouraging me to consider this issue.

This is consistent with the focus of nearly all discussions of the permissibility of sweatshop employment. This should be unsurprising, since it is only in cases involving terms of employment that are voluntarily accepted and genuinely beneficial to employees that plausible defenses of the permissibility of the terms of employment are possible.

The fact that most sweatshops are not operated directly by MNCs is widely noted in the literature on the ethics of sweatshop employment. See, for example, Arnold and Bowie (2003, pp. 225–226), Meyers (2004, p. 329), Young (2004, pp. 366–367), Sollars and Englander (2007, p. 116), Snyder (2008, p. 399) and Zwolinski (2012, p. 162).

Some ways in which the distinction between the two types of case, and the ways in which both MNCs and local firms that operate sweatshops are embedded in prevailing forms of structural injustice, might affect our assessments of responsibility for the exploitation of sweatshop workers will be discussed in “The Significance of Structural Injustice” section.

The view that firms are agents that can themselves be the bearers of obligations has, in recent years, become the mainstream view. For an influential early defense, see French (1979). More recent defenses include Silver (2005), Arnold (2006), Pettit (2007), and List and Pettit (2011). For an early argument against the possibility of corporate moral agency, see Velasquez (1983). More recent arguments include Velasquez (2003), Rönnegard (2013), and (Rönnegard 2015).

Both Wertheimer (1996, p. 289) and Jeremy Snyder (2008, p. 390) formulate the Claim in roughly this way, but do not explicitly limit it to voluntary transactions; see also Bailey (2011, p. 238) and Barnes (2013, p. 28). Although those who formulate it in this way appear to assume that proponents of the Claim would not necessarily take it to justify non-voluntary yet mutually beneficial transactions, it is, in my view, important to include the restriction to voluntary transactions in the formulation of the Claim itself. Common formulations of the Claim also do not explicitly state that the relevant transactions must not have any harmful effects on third parties, although this restriction is implicitly acknowledged in discussions of its implications. Since the restriction is widely understood and less important for the purposes of my argument, I do not explicitly include it in my formulation of the Claim. For formulations that are otherwise similar to those of Wertheimer, Snyder, and Bailey that both restrict the Claim to voluntary transactions and build in the requirement that no third parties are harmed, see Ferguson (2016, p. 956) and Malmqvist (2017, p. 478).

For this line of argument, see Zwolinski (Zwolinski 2007, pp. 699–700, 707–708, 2012, pp. 167–169), Sollars and Englander (2007, p. 119), and Powell and Zwolinski (2012, p. 460–470).

For similar cases, see Zwolinski (2012, p. 156) and Vrousalis (2018, p. 2).

To say that it is implausible that B’s complaint is stronger than those of the others is not necessarily to say that it must be weaker either (though I think that it is plausible that it is). We might think that some complaints cannot be evaluated as simply stronger or weaker than others, given the plurality of morally important considerations that might ground different complaints and the plausibility that these considerations cannot all be aggregated and weighed against each other within a single metric. It is enough for the purposes of my argument that the complaints of those exploited are, at least in some cases, clearly not stronger than the complaints of others. I am grateful to an anonymous reviewer for helpful comments on this issue.

Further details could be added to the case to make the implausibility of holding that those hired are wronged, while those not hired are not wronged, even clearer. Imagine, for example, that potential workers in each of the towns find out that their town is being considered for the site, and organize a campaign to send letters to M’s CEO pleading for their town to be chosen. Moved by the accurate descriptions of how they would benefit provided by the residents of Z, the CEO chooses to locate the site there.

As Michael Kates puts it, “firms are not morally required to offshore production, even when workers in developing countries would considerably benefit from these jobs. In fact, not only does virtually no one defend the view the multinational corporations have an obligation to offshore production, but many believe that keeping production in the developed world is something to be praised” (2019, p. 27).

For a similar claim, made in the context of a discussion of whether sweatshop regulation is justified, see Sollars and Englander (2018, pp. 21–22). They are responding primarily to the argument in defense of regulation made in Preiss (2014). For Preiss’s response, see his (2018).

Many who reject the Nonworseness Claim (in all of its versions) claim that sweatshop employment is wrongfully exploitative because companies acquire potentially extensive obligations to their employees in virtue of employing them, including, for example, obligations to pay wages and ensure workplace safety conditions that meet certain standards. In other words, they hold that the employment relationship, once in place, generates new obligations that are owed by employers to those whom they in fact employ, even in cases in which they had no independent obligation to employ or otherwise benefit a subset of people from an appropriately specified group. While I cannot here argue directly against all views of this type that have been defended, the arguments that I offer in defense of Nonworseness Claim (W), along with the case that I make for thinking that MNCs have potentially extensive positive duties to the global poor, provide reasons to be skeptical of them. I will, however, briefly note two additional reasons for skepticism here. The general concern that lies behind these reasons for skepticism is that it is difficult to see how any view with this structure could be thought to be grounded in the right kind of concern for the interests of the global poor. First, in order to avoid acting wrongly, these views imply that MNCs can either hire some of the global poor and provide sufficiently good wages, working conditions, etc., or simply refrain from hiring any of the global poor at all. This means that whenever hiring from among the global poor and meeting the necessary conditions for avoiding wrongful exploitation would be even slightly more costly than, for example, employing only people in developed countries, a company is able to avoid wrongdoing by avoiding incurring the slight extra cost, even if this comes at a very large opportunity cost for the badly off people who could have been hired. My view, on the other hand, will at least often imply that a company in these circumstances is simply obligated to accept the slightly greater costs in order to benefit the global poor. The second reason that we should be skeptical of views on which entering an employment relationship generates new duties with respect to wages, working conditions, etc. that are owed specifically to those employed is that there does not seem to be a way of combining this type of view with any plausible view about the ethics of ending employment relationships that does not generate implausible implications. For example, if we accept a view on which one of the duties generated by an employment relationship is a duty to continue to employ those who have been hired in the absence of, for example, unacceptable job performance or financial exigency, then it will often be impermissible for employers to let reasonably well off and adequately performing workers go and replace them with badly off workers who would perform even better. But this result seems quite implausible—again, no view that is grounded in the right kind of concern for the interests of the global poor could imply this. On the other hand, if we accept a view on which entering an employment relationship does not generate a duty to continue to employ those who have been hired, then companies would do nothing wrong by responding to charges that they are wrongfully exploiting their sweatshop employees by, for example, letting all of them go and replacing them with much better off workers in the developed world. Once again, a view with this implication cannot be grounded in the right kind of concern for the interests of the global poor. My view, on the other hand, will at least often imply that companies are obligated to shift to employing worse off people, or to continue to employ badly off people and improve their wages, working conditions, etc. I am grateful to an anonymous reviewer for encouraging me to discuss views on which entering employment relationships generates new duties.

Alternatively, it might be suggested that, in light of the argument that I go on to make in “The Significance of Structural Injustice” section, the relevant group should be understood to include all of those who are victims of structural (economic) injustice. And it might also be suggested that this group plausibly includes potential garment workers in the USA, in addition to those in less wealthy countries, so that MNCs should be understood as potentially fulfilling the duties that they have as beneficiaries of structural injustice by employing or otherwise benefiting potential workers in the USA. While I cannot discuss this possibility in detail here, I want to offer two initial reasons to think that we should be skeptical of the claim that at least most potential workers in the USA are among the group that MNCs have positive duties to benefit. First, it seems clear that the relevant criterion for membership in this group must be that one is, on net, disadvantaged by structural injustice. And while it is no doubt true that many Americans are disadvantaged by certain structural injustices considered in isolation, it is less clear that very many are disadvantaged on net (though certainly at least some are). Second, even if some Americans are, on net, disadvantaged by structural injustice, there are plausibly reasons to limit the group among whom MNCs are obligated to employ or otherwise benefit to those who are most disadvantaged, since there are general moral reasons to prioritize addressing the most serious needs and deepest injustices before attending to less morally urgent matters. I am grateful to an anonymous reviewer for encouraging me to consider this issue.

For a similar claim, see Powell and Zwolinski (2012, p. 470).

The obligation to prioritize providing employment opportunities to those who are worse off is limited, of course, by considerations of feasibility. Firms are not, for example, obligated to hire candidates for jobs who are incapable of performing the job functions, or to offshore jobs to impoverished countries when doing so would involve taking on a significant risk of undermining the firm’s competitiveness. But when considerations of competitiveness either support offshoring, or at least do not advise against it, moving jobs to places where they will benefit worse off people can, on the view that I am defending, be morally required.

It is an empirical question whether this accurately describes the behavior of most actual companies that operate or contract with sweatshops, and I cannot discuss the issue in detail here. My suspicion is that it does accurately describe the behavior of at least many companies, and that because of this the view that I defend implies that these companies are obligated to do significantly more than they are actually doing to benefit the global poor. A potentially useful way of beginning to assess whether any particular company is obligated to do more than it is in fact doing is to look at its profit margins, as well as salaries of high-level employees, and consider whether redirecting at least some resources in ways that would benefit sweatshop workers in its supply chain would be feasible. In many cases I suspect the answer will be that such redirection of resources would be possible. On the other hand, at least some evidence that a company is satisfying its obligations would be provided by a combination of lower than average profit margins and a practice of ensuring that even the lowest level workers in its supply chain are paid a living wage. I am grateful to anonymous reviewer for encouraging me to consider how we might attempt to assess whether companies are complying with the obligations that I defend.

See, for example, Singer (1972), Unger (1996), and Murphy (2000). Even those who think that factors in addition to capacity to help play a significant role in determining which agents are obligated to contribute to benefiting those in need, and how extensive their obligations are, tend to think that capacity to help is among the relevant factors. See, for example, Miller (2001).

Defenders of the view that it is permissible for firms to employ workers in sweatshop conditions often appeal to the long-term benefits to both sweatshop workers and the communities in which sweatshops are located in order to support their view; see, for example, Maitland (2008). As Chris Meyers points out, however, the evidence that introducing sweatshops into a country or community will have substantially beneficial long-term effects is mixed at best (2004, p. 322). However, since I am arguing that MNCs can be obligated to locate new employment opportunities in impoverished countries and provide benefits that exceed those provided by sweatshop jobs alone, by, for example, offering conditions of employment that are intuitively non-exploitative, there is, it seems to me, significantly less reason to be concerned that the new employment opportunities that would be provided if companies like M followed my recommendations would not significantly benefit both the workers and the broader communities in which they live.

For a powerful defense of the view that beneficiaries of global structural injustice have potentially extensive obligations to benefit those who are disadvantaged by such injustice, see Ashford (2018). Other sympathetic discussions include Woodruff (2018), McMahan (2018), and Hill (2018). I defend a version of the view, with particular application to duties to contribute to addressing the threat of climate change, in Berkey (2017).

In particular, it is plausible that they ought to aim to ensure, as much as possible, that workers in their supply chains are paid living wages, that safety-related risks are kept reasonably limited, and that (mandatory) work hours are at least not unreasonably long.

It is important to note that here Young has in mind responsibility in what she calls the “liability” sense, which involves backward-looking assessments of fault made in order to assign blame and compensatory obligations (2004, p. 368). She argues that an adequate overall account of responsibility for global labor justice, including the issue of sweatshop employment, requires an additional, “political” sense of responsibility to supplement the liability sense. And, as I have noted, she holds that a wide range of agents have political responsibilities to contribute to addressing the structural injustices that make employing people in sweatshop conditions both possible and potentially difficult to avoid. I am grateful to an anonymous reviewer for encouraging me to clarify this aspect of Young’s claim about the responsibility of local owners and managers.

This is especially true if we think that agents that have direct responsibility for an objectionable state of affairs have, all else equal, stronger obligations to contribute to remedying them. It is not obvious, however, that Young is herself committed to such a view.

Despite the fact that I have emphasized some key points of departure between my view and Young’s, it is worth noting that there is at least some overlap in our views as well. For example, Young’s account of political responsibility assigns duties to agents at least in part on the basis of capacity to contribute to addressing injustice. Indeed, as an anonymous reviewer suggests, her account of political responsibility can plausibly be interpreted as implying positive duties to the global poor of at least roughly the same kind that I defend. One of the most important features of my view, however, is that the content of these positive duties is to do whatever will best contribute to improving the conditions of the global poor, whether that involves attempting to reform structural conditions, providing higher wages and better working conditions, some combination of these things, or, in principle, anything else that would benefit those to whom the duties are owed. Young’s view, on the other hand, is that the content of the duties is limited to promoting structural reform. I see no principled reason for accepting this limitation. I am grateful to an anonymous reviewer for encouraging me to clarify these aspects of the relationship between my view and Young’s.

This is the case on any view on which firms have a conditional obligation to, for example, pay workers a living wage, or provide workplace safety conditions that meet a standard of adequacy (e.g., Arnold and Bowie 2003; Meyers 2004; Snyder 2008 and 2013; Kates 2019). On views with this structure, firms are under no obligation to hire or otherwise benefit impoverished people in developing countries; but if they do hire them, they are obligated to provide either a living wage (if possible), or a fair wage, where the standards for fairness are determined by an independent account that might be maximally favorable to workers (Kates 2019).