I’m generally pretty skeptical of sweatshop regulation. I buy the argument that for most workers at low-wage factories in developing countries, these jobs are their best option relative to alternatives such as working in agriculture or in the underground economy. I think many Western efforts to regulate or boycott arguments are misguided and harm people who they think they’re helping. But some libertarians make really bad arguments against sweatshop regulation, and intellectual honesty requires me to call them out. The existence of bad arguments for X, even when there are also good arguments, tends to undermine support for X.
One prominent recent book that defends sweatshop jobs is Out of Poverty: Sweatshops in the 21st Century by Ben Powell. It’s an interesting book with lots to think about, but I want to focus on one excerpt where Powell discusses workplace harassment. Powell objects to criticisms by anti-sweatshop groups of sexual harassment on the job:
If the risk of sexual harassment is part of the working conditions at a factory, then firms have to offer higher wages to attract workers…. In short, the analysis of sexual harassment on the job is much the same as the analysis of other working conditions. Laws that effectively eliminate sexual harassment would lower wages. If employees desired this, then market forces would remix the compensation package to minimize harassment and lower wages. (p. 184)
The mechanism here is the idea of “compensating differentials” for labor. If a job is unattractive in some dimension, such in the level of abuse workers endure, in a competitive market the employer will have to pay higher wages to compensate for that unattractive feature. Powell cites as evidence for this mechanism a 2010 study of U.S. markets by Joni Hersch, which finds that women do in fact enjoy a compensating differential of about $0.25.
The first thing to note, though, is that Hersch’s analysis cannot simply be applied uncritically to markets outside the US. There are at least two reasons for this. First, markets for sweatshop labor may be much less competitive than US labor markets. This is an important condition for the positive effect of compensating differentials on wages
Second, even if the compensating differentials mechanism has an effect, sexual harassment may have other negative effects on wages such as higher turnover, lower employee morale, or inefficient efforts to avoid harassment, which leads to lower productivity. Because we cannot decide from the armchair which effect dominates (the productivity effect or the effect of compensating differentials), we need to do careful empirical analysis of the labor markets in sweatshop countries before we conclude that workers are paid higher wages as compensation for harassment.
But let’s assume that workers are, in fact, paid higher wages as compensation for harassment. Powell argues that we should infer from the fact that workers take these sweatshop jobs that workers prefer the money to a harassment-free workplace, since if workers had different preferences, the employer would pay a lower wage and reduce the level of harassment. Because workers agree to jobs where they sexual harassment, anti-sweatshop activists shouldn’t oppose it.
This argument is, to put it charitably, not very good. In the first place, it doesn’t at all follow from the fact that workers accept jobs with sexual harassment that we shouldn’t object to those terms of employment. Sometimes we shouldn’t make offers to people even though they accept it, particularly the offer reflects some objectionable preference (such as the desire to grope and physically violate women). We should do what we can to change those preferences (or at least the behavior associated with them) rather than take them as a fixed cost, which is the position of most anti-sweatshop activists. Even if workers accept those terms of employment, for example, they may prefer that employers didn’t have the desire to sexually harass them. They may prefer not be part of a collective enterprise where anyone’s harassed. This possibility, raised by Amartya Sen in a critique of the revealed preference concept, means we cannot necessarily infer welfare from choice or behavior.
There are good arguments against some regulations of sweatshop labor, but those arguments are generally sensitive to the costs and benefits of different regulations or the voluntary nature of the worker’s choice. In this case of Powell’s argument against sweatshop regulation, unfortunately, neither of these desirable conditions obtain.