The Ethical and Economic Case Against Sweatshop Labor: A Critical Assessment

Maitland (1996), emphasis added.

For example, on the left, see Krugman (1997). On the right, see Williams (2004).

Miller (2003).

See Hayek (1968).

In the business ethics literature, the most significant defenses have been presented in Maitland (1996) and Zwolinski (2007). In the popular media, Krugman’s (1997) defense is still frequently cited, as are several articles by Kristof and Wudunn (2000) and Kristof (2009).

See, for instance, Peter Singer’s discussion of the living standards of the world’s poor today compared to 20 years ago (Singer 2009).

(Arnold and Hartman 2005, p. 208).

Arnold and Hartman (2005, p. 209).

See Hayek (1945).

Arnold (2010, p. 635).

Zwolinski (2007, pp. 691–693).

“Voluntary,” at least, in the sense that their choice is not coerced. We discuss the concept of coercion and its application to sweatshop labor in the section on “Coercion and Exploitation”.

Arnold and Hartman (2005, p. 209).

See Stringham (2010).

A point recognized by Arnold himself in his discussion of moral imagination (see Arnold 2003, p. 79).

Arnold (2010, p. 637).

See Hayek (1968).

Powell (2006) praises them on exactly this point.

Arnold and Hartman (2005).

Arnold (2010, p. 651).

Economists refer to this as contestable markets theory. There is also a large experimental economics literature that shows small numbers of buyers and sellers achieve results that approximate what a perfectly competitive market is supposed to achieve.

See Arnold and Hartman (2005, 2006). Arnold and Bowie (2007) equivocate on this issue. They grant their critics that the effect of efficiency wages on worker productivity is indeterminate “for the sake of argument” (p. 142).

Pollin et al. (2004, p. 156).

Pollin et al. (2004) also list improvements in product quality, marking, and overall expansion of product market demand. In most cases, these are likely exogenous and thus violate the ceteris paribus clause and thus do nothing to undermine the economic theory that predicts mandating higher wages reduces employment from what it otherwise would have been.

Arnold and Hartman (2005, p. 217).

Arnold and Hartman (2006, n. 46).

Arnold and Hartman (2005, p. 218).

Powell and Skarbek (2006) show that sweatshop jobs pay wages substantially above the living standards in the countries where they exist. Thus, loss of a sweatshop job already imposes a severe penalty on a worker. An efficiency wage is not necessary to create a downside to job loss.

Arnold and Bowie (2003, p. 238).

Their response is simply to reassert that it is the fact the case and then point to a single study from the El Salvadoran Ministry of Labor that found companies using North American productivity standards without accounting for either different nutritional conditions or technical capabilities of local workers. This latter does nothing to show that efficiency wages would be justified and the former is just a bald unsubstantiated assertion.

See Arnold and Hartman (2006) in which the authors offer three other mechanisms for how increases wages might not unemploy workers: passing on costs to consumers, cost cutting in other areas, and accepting a lower return. See Powell (2006) for a critical review of these mechanisms.

For instance, Bama Athreya, from the International Labor Rights Fund, admitted that wages in sweatshops were typically higher than in domestic industry but argued that it is the working conditions that need to be addressed in a public debate with Benjamin Powell at Grand Valley State December 1, 2008. Another example can be found in Arnold and Hartman (2006).

Arnold and Bowie (2003, p. 253).

Arnold and Hartman (2006, p. 8).

Sollars and Englander (2007).

Nicotex and Sam Bridge were surveyed because they were identified as sweatshops and protested by the National Labor Committee. Powell and Clark (2010).

Skarbek et al. (2011).

Pollin et al. (2011, p. 155).

Curiously, Harrison and Scorse (2010, p. 263) go to great lengths to cast anti-sweatshop activism in the most favorable light. Most of the article emphasizes how they do not find any additional unemployment effects from anti-sweatshop activism beyond the unemployment effect of the minimum wage. Yet it is anti-sweatshop activism that was in large part responsible for increasing the minimum wage.

Arnold (2010, p. 639).

Arnold (2010, pp. 641–642).

Similarly, just because Cadillac can profitably put leather seats in its cars does not prove that all firms could profit more by putting leather seats in their cars.

Vogel (2005).

Some critics of sweatshops advocate global rather than national regulation to avoid so-called “races to the bottom.” We lump global and national regulation together because the standard economic defense of sweatshops is the same in both cases. In either case, the automation of production and shift from low-productivity to high-productivity countries leads to adverse consequences for poorer workers.

Or as explained in the previous section, changing the mix of compensation away from that desired by the employees.

Many health and safety regulations and wage mandates in the first world are very close to what the market would produce anyway so they have little impact. For instance, 97.5% of all workers earned more than the US Federal minimum wage in 2005 (http://​www.​bls.​gov/​cps/​minwage2005.​htm). Hall and Leeson compare the per capita income in the United States when it adopted various wage and safety regulations to the incomes in third-world countries today and find that sweatshop-intensive developing countries are 35–100 years from reaching the level of development the United States was at when it adopted similar policies. See Hall and Leeson (2007).

Market forces continue to be the main cause of increased safety even in countries with strong regulatory standards. For example, in the United States, the expected penalty per worker for OSHA violations amounts to only $1.34 per worker while market forces, through compensating wage differentials and workers’ compensation premiums imposed costs in excess of $1,234 per worker. See Viscusi (2005, p. 851). Also see their discussion on pp. 854–860 of how safety was improved through economic growth, and the creation of OSHA had no impact on increasing the trend.

Elliot and Freeman (2004).

Elliott and Freeman (2004, p. 86).

http://​www.​sweatfree.​org/​shopping_​suppliercriteria​.

Both Powell and Zwolinski have commended Arnold and Hartman for their work in documenting some voluntary actions firms have taken (see Powell 2006; Zwolinski 2007). We, also, believe that making these actions more widely known to other companies can help speed the discovery of improvements that can take place without harming workers.

Arnold and Hartman (2003, p. 427).

Varley (1998).

Arnold (2010, p. 632).

Sollars and Englander (2007, pp. 121–122).

Arnold and Bowie (2007, p. 139).

It is, of course, possible that their goal is the welfare of the particular workers who happened to be employed by sweatshops, and not the welfare of workers in general. We shall address this possibility in “Coercion and Exploitation” section.

Arnold (2010, pp. 645–646).

Arnold (2010, p. 642).

Actually, what Sollars and Englande (2007, p. 115, emphasis added), actually say is that “MNEs or their managers have duties not to tolerate or encourage violations of the rule of law.”. Arnold and Bowie assume that violations of the law are tantamount to violations of the rule of law. We will allow this assumption for the sake of the present exposition, but will return to criticize it later.

Arnold and Bowie (2007, p. 139).

Ibid.

Arnold (2010, p. 638).

See Powell (2010, Chapter 2) for a survey of the demands of the various groups in the anti-sweatshop movement. Draft Manuscript. Available on request, 2011.

Arnold (2010, p. 639).

Powell (2006).

Wolff (1970).

Green (2010).

This idea is not mere idle philosophical speculation. As one referee pointed out, the Sullivan Principles of corporate conduct made violating Apartheid-era laws a condition of doing business in South Africa.

Forced labor is inconsistent with both the autonomy-based and welfarist justifications of sweatshop labor given in Zwolinski (2007, pp. 691–665). Obviously, a worker who is physically compelled to work in a sweatshop cannot be said to autonomously choose sweatshop employment. Nor can it be inferred that sweatshop labor is his or her most preferred alternative (and hence likely his or her most welfare-enhancing alternative) if his or her employment is forced, and not chosen.

See, for instance, Arnold and Bowie (2003) “the intentional violation of the legal rights of workers in the interest of economic efficiency is fundamentally incompatible with the duty of MNEs to respect workers” (p. 228), Arnold and Hartman (2005) on the need to move the sweatshop debate “beyond the entrenched, polarized, political narrative of economic efficiency versus increased regulatory protection for workers’ rights” (p. 212), Arnold and Hartman (2006) “those who are genuinely interested in the welfare of the citizens of developing nations ought to demand that MNCs and their contractors respect local labor laws, rather than excusing those MNCs that violate local laws in the name of economic efficiency” (p. 690).

Arnold and Hartman (2006, p. 690).

We also believe that laws that unjustly violate the autonomy of workers and potential workers should sometimes be disobeyed. Most of what we say in the following regarding violations justified by appeals to welfare, however, will also apply to violations justified by appeals to autonomy; so for the sake of brevity, we will focus here only on welfarist arguments.

The form of problem manifests itself elsewhere in Arnold’s argument, as discussed in Zwolinski (2007, pp. 698–700). Jeremy Snyder, a philosopher generally sympathetic to Arnold and Bowie’s critique of sweatshops, makes a similar point in Snyder (2008, pp. 394–395).

Arnold (2010, p. 632).

Arnold (2010, p. 632).

See, for an overview, Anderson (2006).

See Arnold and Hartman (2005) “No one in this debate advocates forced labor” (p. 679, n. 5), Sollars and Englander (2007) “We agree…that workers should not be physically coerced” (p. 122), Zwolinski (2007). “The truth of premise 1 [in the argument for the moral impermissibility of interfering with sweatshop labor] hinges on whether people do in fact choose to work in sweatshops, and fails in cases of genuinely forced labor” (p. 696)

The only person of whom we are aware who does make this claim is Miller (2003, p. 97). But he does not defend the claim at any length, and has not (as far as we are aware) repeated it since.

Arnold and Bowie (2003, p. 229).

Arnold and Bowie (2003, p. 229).

Arnold and Bowie (2003, p. 229). See also Arnold and Bowie (2007).

Arnold and Bowie (2003, pp. 229–231).

Arnold and Bowie (2003, p. 229).

Arnold and Bowie (2003, p. 230).

See Sollars and Englander (2007, pp. 122–123) and Zwolinski (2007).

Arnold (2010).

Arnold and Bowie (2007, pp. 140–142).

Arnold (2001, p. 56).

Arnold (2001, p. 54).

See Morse (2000, pp. 1054–1063).

Arnold and Bowie (2003, p. 229).

Sollars and Englander (2007, p. 123). We are not convinced that the distinction between “conditions of employment” and “informal practices” suggested by Sollars and Englander (p. 123), and picked up explicitly by Arnold and Bowie (2007, p. 141) is a helpful one. For, in the standard case, one of the conditions of employment will simply be that one comply with the informal practices of the workplace and the occasional job-related special requests of one’s supervisor.

Sollars and Englander (2007, p. 123) and Arnold and Bowie (2003, p. 229).

These examples were originally presented in Arnold and Bowie (2003) but are reproduced in their entirety in Arnold and Bowie (2007, pp. 140–141).

We leave this as an open question, because it is not clear from Arnold and Bowie’s presentation which of the three possible types of “compulsion” are supposed to be at work in them. Threatening to fire an employee might qualify as “rational compulsion,” but it is doubtful that it could be interpreted as either psychological or physical compulsion.

See, for instance, Arnold and Bowie (2003), Mayer (2007), Meyers (2004), Snyder (2010), and Young (2006).

Some of the most influential accounts include Goodin (1987), Wood (1995), Wertheimer (1996), Sample (2003), Snyder (2008), Mayer (2007), and Valdman (2008, 2009). Wertheimer (2008) provides an overview of most of the main philosophical accounts. Snyder (2010) provides another overview with specific focus on the application of such accounts to the issue of sweatshop labor.

This particular version of the claim is taken from Meyers (2004, p. 331).

Eisenberg (2009, vols. 15–16).

We defend workers’ freedom to organize collectively voluntarily which is distinct from laws that allow labor unions to organize workers where a subset of all workers has the legal right to collectively bargain for all workers even when some workers would rather bargain individually.

See Pogge (2005, p. 7).

Pogge (2005, p. 6).

See Zimmerman (1981).

See Snyder (2008, 2010). Snyder also appeals, in the first of these articles, to Iris Marion Young’s account of political responsibility. It is not clear, however, exactly what work Young’s account is supposed to do in an account of exploitation. After all, Young is explicit in saying that her account of responsibility is distinct from what she calls the “liability” account, and that it will generally be inappropriate to blame individuals or groups who bear only a “political” responsibility for injustice. This is because, on Young’s account, political responsibility is borne by persons who are not “guilty or at fault for having caused a harm without valid excuses” (Young 2006, p. 119). So Young’s account cannot explain why sweatshops or MNEs deserve more blame than other parties for the fate of sweatshop workers. Nor, really, does it do much to explain why they should be seen as bearing a greater responsibility to rectify those injustices, even in her own special sense of political responsibility. She argues that this responsibility arises from social connection. But the special status of social connections is never defended in an entirely clear way. Young states that the kind of social connection that generates political responsibility is “participation in the diverse institutional processes that produce structural injustice.” And this might make sense if we thought that agents who participated in such structures were wrongly benefitting from them at the expense of victims of injustice. But this cannot be what Young is saying. For, if what grounds responsibility is not participation in unjust structures as such but wrongful participation in unjust structures, then Young’s account would seem to collapse into a standard liability model of responsibility. One might think instead that participation matters because it provides ones with opportunities to fight the injustice. But there is no reason to think that all who participate in unjust structures will have such opportunities, nor that all non-participants will lack them. It is hard to see, then, how “social connection” could be specified in any way that would non-arbitrarily assign any kind of special responsibility to sweatshops or MNEs for remedying structural injustice. Those who own or operate sweatshops or MNEs might have the power or opportunity to work to remedy certain kinds of structural injustice, but if this power and opportunity is neither universally nor uniquely present among them, it will not generate any universal or unique obligation on their part. For a helpful discussion of these issues, see Silvermint (2011). See also Zwolinski (2012).

Snyder (2008, p. 396).

Snyder (2008, p. 390).

Snyder (2008, p. 396).

So long as they discharge their imperfect duty of beneficence in some other way. The point here is really a specific application of what has been called, in the literature on exploitation, the “non-worseness claim,” which holds that it cannot be morally worse for A to interact with B than it is for A not to interact with B when the interaction is mutually beneficial, consensual, and free from negative externalities. See, for a discussion, Zwolinski (2007, pp. 708–710; 2008, pp. 357–360; 2009), Snyder (2009), and Wertheimer (2011, chapter 6). It is beyond this article to set out a full defense of the non-worseness claim (though see Zwolinski, “Exploitation and Neglect” for an attempt to do this). Instead, the discussion that follows attempts to press the intuitive force of the non-worseness claim in this specific context, without fully defending it as a general principle.

Alternatively, Snyder could hold that Company A is guilty of exploitation, whereas Company B is not, but that Company B is guilty of some other and perhaps more serious form of moral offense. This would save Snyder’s account from having to embrace the counterintuitive claim that Company A is acting in a worse way than Company B, but only at the price of reducing the moral significance of exploitation.

Actually, Snyder does not quite hold that it is “sufficient.” Several other conditions must be met for the employer to have this duty, but as they do not affect the present argument these need not concern us here.

Zwolinski (2007, 2008, 2009).

Snyder (2009).

Snyder (2009, p. 305).

Snyder (2009, p. 305).

We thank an anonymous referee for raising these questions.

In one extremely interesting experiment, Christopher Freiman and Shaun Nichols presented subjects with either an abstract or a concrete description of a situation involving a distribution of resources. The abstract version asks subjects to “suppose that some people make more money than other solely because they have genetic advantages,” while the concrete version asks them to “suppose Amy and Beth both want to be professional jazz singers. They both practice singing equally hard. Although jazz singing is the greatest natural talent of both Amy and Beth, Beth’s vocal range and articulation is naturally better than Amy’s because of differences in their genetics. Solely as a result of this genetic advantage, Beth’s singing is much more impressive. As a result, Beth attracts bigger audiences and hence gets more money than Amy.” Subjects were then asked whether the fact that the genetically advantaged individuals make more money is fair. Surprisingly, subjects who were given the concrete version of the case were significantly more likely to say that it is fair for the genetically advantaged individuals to make more money than those who were given the abstract version of the case. See Freiman and Nichols (2011).