I appreciate the concern that Fung, O'Rourke, and Sabel express for transparency, broad-based participation, yardstick competition, independent monitoring, and coordinated action in the current proliferation of piece-meal, divergent, and arbitrary standards in labor regulations and practices in the international marketplace. I also applaud their sincerity of purpose and good intentions. But I find the purpose rather narrow, and the emphasis partly misplaced—and potentially even harmful to the world's poorest people.

Fung, O'Rourke, and Sabel make a good argument for flexible, locally variable, context-specific labor standards, which is a welcome change from the frequent demand from some activists for pious universal (i.e. rich-country) standards. For example, the authors want to judge the labor practices in a transnational enterprise in Vietnam by comparing them with practices in a similar enterprise in Indonesia, not with those in Europe or North America, at least not initially. But it is not clear why they do not extend the argument for comparison or context-specificity to the rest of the economy in Vietnam, or for that matter in any such poor country. The transnational enterprise in a poor country, for all the horror stories about its working conditions and wage levels that outrage consumers in rich countries, is usually, though not always, an island of relatively "decent work" (to use a phrase popular in ILO documents) in an ocean of "indecent" and brutal work conditions in the rest of the economy. If those islands are the "sweatshops" (and they no doubt are by some lofty standards), we have to keep in mind the reality that many of the millions of workers in the rest of the economy are usually banging at the gates of the "sweatshops" for a chance of entry.

Suppose that, with participation of firms, activists, and workers in specific firms, labor standards are "ratcheted up" in line with similar enterprises elsewhere. The result may be improvements for workers inside those enterprises, but one should be wary of the consequences for the much larger number of workers left out. In fact, sometimes the improvements in the tiny "island" may be at the expense of those left floating (or sinking) in the "ocean." Relatively high wages or better working conditions in the transnational enterprise are likely to raise the latter's cost of hiring labor, and thus depress labor demand and job prospects for the rest of the workers.

The workers thus left out are more likely to crowd the inferior job markets, depressing wages and working conditions. For example, child labor banned from internationally certified workplaces may end up in much inferior jobs and occupations (including child prostitution), as casual empiricism as well as the findings of an OXFAM study of displaced child labor in Bangladesh suggest. In India, only 5 percent of child labor are in the international sector; the rest is employed in various parts of the domestic economy. When you abolish child labor in this international sector through commonly adopted comprehensive policies like Ratcheting Labor Standards (RLS), without at the same time doing anything about the displaced children, the glow of moral comfort enjoyed by the consumer in rich countries who scrupulously looks for the "no-child-labor-used" label or certification is largely illusory. Similar is the case in the "no-sweatshop" certification, where a "living wage" and better work conditions for a few workers in the monitored international sector may mean non-living or sub-human wages and work conditions for many elsewhere in the economy.

These criticisms do not imply that RLS is a bad thing. But the approach—in general, any regulatory approach that does not take into consideration what economists call opportunity costs and the enormous side-effects for the vast majority of the working poor outside of the regulated facility—as well as the need for compensation to those who are actually or potentially displaced by the regulations—is very narrow and, without these simultaneous other measures, potentially harmful. If the "ethical consumer" of the rich country is prepared to pay more for products where she can feel more comfortable about the process of their production, RLS should be combined with coordinated action to collect the premium paid by the consumer to finance the compensatory measures. The banned child worker should be given adequate scholarships to pay not just for schooling, but for the income foregone by the child's desperately poor family. (A recent program for the rural poor in Mexico, called Progresa, offering scholarships for children, paid to a designated woman in the household, conditional on the children's school attendance, has already been quite successful in reducing child labor.)

In addition, some funds will need to be earmarked for improving the school facilities and making them more attractive or accessible for children. (In India, the number of children of school age who are neither in work nor in school is estimated to be more than three times the number of them in child labor.) An example of the possibility of coordinated action in this general area (regulation plus monitoring plus scholarships and amenities) is provided by the international soccer ball producers' agreement in the mid-1990s, whereby in the city of Sialkot in Pakistan (which is one of the world's largest producers of soccer balls, and used to employ many children for stitching the balls) the transnational sporting goods companies agreed to provide scholarships to the children who lost their jobs, some NGOs were to act as monitoring agencies, and the local government was to provide the schooling facilities. In the case of adult sweatshop labor, RLS has to be combined with some investments in the local community that help not just the workers in the improved facility but those who are outside. This is no doubt much more expensive than implementing a simple RLS, but massaging the conscience of the affluent consumers in an overwhelmingly poor world cannot be that cheap.

One should, however, distinguish between labor practices that involve forced or prison labor and hazardous and unsafe work conditions on the one hand, and other kinds of child labor or sweatshop work or substandard wage rates on the other. (The examples cited by Fung, O'Rourke, and Sabel combine the two kinds.) In the former case, I am all in favor of a legal or regulatory approach with appropriate monitoring and enforcement. (In the context of prison labor, while one is legitimately outraged by stories about China, let me note in passing that UNICOR, a corporation owned by the United States federal government, operates about one hundred factories with prison labor, sells over 150 products, with total sales of about $500 million dollars in 1995.)

But in the case of uncoerced child labor, sweatshops, and substandard wage rates and working conditions, I'd like the regulatory enthusiasts to take into account some of the side effects (and necessary compensation measures) that I have indicated above. Otherwise they may end up hurting the intended beneficiaries. •

Pranab Bardhan is professor of economics at the University of California–Berkeley and the chief editor of the Journal of Development Economics.

Read all eight responses to "Realizing Labor Standards," by Archon Fung, Dara O'Rourke, and Charles Sabel.