Cheap garment factories: keep or cut, and who decides?

A small Vietnamese fashion brand (hypothetical; I am not describing any company) places 30% of its orders with a subcontract factory abroad, at a price about 12% below the alternative. The latest audit reports three things: workers are threatened with dismissal if they refuse overtime, pregnant women are moved onto repeated short-term contracts, and managers obstruct anyone trying to form a union. The workers say they need the jobs. Cut the order or keep it, and who decides?

What does "the workers freely accept" prove?

Zwolinski builds the defence of cheap factories into six steps. Most workers choose to accept the conditions, though from a severely constrained set of options. That they chose is strong evidence it is their most-preferred option. Removing it would probably harm them, and it is plausible that the choice is autonomous enough that removing it would violate their autonomy. It is wrong, all else being equal, to harm people or violate their autonomy; so it is wrong to take that option away (Zwolinski, 2007, p. 695).

The conclusion is narrower than people often assume. The argument says only that workers have a claim not to be interfered with; it does not attempt to show that the owner's treatment of them is right or even permissible. Zwolinski says its success depends on empirical facts, and that the first premise fails in cases of genuinely forced labour (Zwolinski, 2007, p. 696). He concludes that workers' consent shows sweatshop work is probably their best option, but not that the people who run the factory are virtuous (Zwolinski, 2007, p. 712). Kates (2015) argues that the choice argument "fails on its own terms" even if its supporters' assumptions are granted; I read only the abstract (Kates, 2015, p. 191).

Which four gates apply to your factory?

Exploitation and valid-consent test asks four questions. Gate V: is the choice free, with no coercion or threat? If not, consent is invalid; stop and consider the duty of assistance. Gate R: do workers have enough information and a meaningful alternative? If not, consent is weak. Gate I: does the other party gain from the vulnerability gap, and does that gain violate rights? If so, it is a sign of exploitation. Gate N: does each person's acceptance harm the group's ability to bargain together? If so, regulation may respect autonomy rather than violate it. If no gate catches anything, there is not enough basis to conclude exploitation, but you still check the collective effect. This is the site's own grouping, not four steps from either author's paper.

Applied to the three hypothetical findings, plus a fourth about wages:

On wages, Zwolinski says that pay above the alternatives still does not refute a charge of exploitation, because it could still be an unfair split of the surplus; settling it would need a comparison of the profit rates of outsourcing firms with those of similar industries (Zwolinski, 2007, p. 707). But his own conclusion is that claims of exploitative sweatshop wages are implausible; the stronger rights concern is threats to a worker's job (Zwolinski, 2007, pp. 710-711).

Two more points are worth reading. On freedom: he holds that "the coercion of poverty" does not license third parties to take away workers' options, but that lack of information can be remedied through disclosure and education, so it is at least possible that the right response is to cure the involuntariness rather than remove the job, though he concedes that some, perhaps many, workers choose ignorantly (Zwolinski, 2007, pp. 701-702). On unions: if a factory suppresses unionisation both inside and outside the plant, the argument as stated does not succeed; he adds that this criticism will not often apply, and that blocking unions only inside the plant leaves the offer independent on this dimension (Zwolinski, 2007, p. 704).

What does fieldwork say about "voluntary"?

The report dates from 2015, so its figures are from that time. Human Rights Watch interviewed more than 340 people, including 270 garment workers from 73 factories in Phnom Penh and nearby provinces (Human Rights Watch, 2015). On overtime, HRW discussed concerns with workers in 48 factories: they generally preferred some overtime, but complained of threats of non-renewal or dismissal if they sought exemption, and HRW documented recent retaliation in at least 14 factories (Human Rights Watch, 2015). The report also records short-term contracts repeated beyond the two-year limit to control workers and avoid benefits. Better Factories Cambodia found that compliance with the two-year rule fell from 76% in 2011 to 67% in 2013-2014, and that nearly a third of factories used such contracts to avoid maternity and seniority benefits (Human Rights Watch, 2015).

Two things for a buyer. First, some of the worst conditions are in small unlicensed factories subcontracting to larger export factories, and Better Factories Cambodia's monitoring reaches them only where brands and the main factory name them and pay (Human Rights Watch, 2015). Second, in my reading, a human rights organisation's report on one sector, one country, one year is enough to defeat the "voluntary" premise where there is retaliation; it cannot measure what would happen to your factory's workers if your orders left.

Who should act first?

Zwolinski says the reasons to respond depend on who you are, because groups differ in what they know, what responsibility they bear and what they can change (Zwolinski, 2007, p. 712). Factory managers probably have the strongest reason, because they know most, bear near-primary responsibility and can change conditions. Multinationals probably come next, often best placed to know what needs fixing (Zwolinski, 2007, pp. 712-713). These firms may have the greatest power to do good, though they also have very close to the greatest power to harm (Zwolinski, 2007, p. 713). The role of governments is less clear, because a ban might push factories to cheaper places and leave workers worse off (Zwolinski, 2007, pp. 713-714).

Vogel lists Nike among the firms that monitor supplier factories and argues that CSR is a niche rather than a generic strategy (Vogel, 2005, pp. 1, 3). As I read it, a small brand is watched less closely than Nike, so for it an internal rule matters more than public opinion.

Keep on conditions, or pull out?

This is my reading, not a conclusion of Zwolinski, Kates, Vogel or HRW. There is no single answer for the whole factory: Zwolinski says it is hard to judge a sweatshop as such, because the details of each activity decide (Zwolinski, 2007, p. 712). I split by the three audit findings plus the wage question:

If the factory agrees to fix things by specific dates, keep the order. If it refuses or will not allow checks, withdraw in an orderly way: give notice, keep orders in progress until there is a replacement, and say why. Withdrawing immediately remains the owner's right, but its cost may fall on the workers, so withdrawal is not automatically clean. The owner of the brand decides their own part: whether to order again, at what price, on what terms. They do not decide for the workers whether to work there.

Going deeper: what did I read, and what did I not?

I read Zwolinski (2007) in full and use the printed page numbers (689 to 727). For Kates (2015) I read only the abstract because the paper is not open access. For Vogel (2005) I read the opening chapter in Brookings's sample. For the 140-page HRW (2015) report I read the summary and the overtime section on HRW's web page.

Three limits. Zwolinski's is a philosophy paper, and he says plainly that he is not positioned to assess the soundness of the economic data he cites (Zwolinski, 2007, p. 696). The four gates are the site's own grouping; I use them as a list of questions, not as an algorithm. The brand, the factory and the response table are mine and only illustrate. In practice you must also consider the law where the factory sits, your contract with it, and what workers say when managers are not present.