Ratcheting Labor Standards (RLS) is a proposal about how to achieve worker rights. It relies on methods of monitoring and enforcement that seem dubious from the standpoint of centralized rule-making. So it is perhaps no surprise that many of our critics have responded skeptically. Nevertheless, we maintain that advancing workers' rights, broadly conceived, requires enlisting the energies of consumers, NGOs, auditing firms, workers, and individual corporations.

RLS, like other labor standards strategies, assumes first that workers have the right to decent treatment—for the same reasons they have human rights. Second, more contentiously, RLS assumes that little is achieved, and much may be jeopardized, by defining human rights of labor too narrowly, and then entrusting the enforcement of "core" rights to existing international organizations. Third, its distinctive hope is that something like the protections that a traditional rights regime promises, but seldom delivers, can be reached through a new route: a system of decentralized but linked efforts to improve labor standards that in effect reveals what we mean by human rights at work as we learn how to implement them in different settings. Thus, RLS aims to fill the gap between the promise and performance of the rights agenda, not to substitute for it.

Some of our critics, rejecting the second assumption, overlook the defects of traditional rights views. David Moberg and Robin Broad are the most ardent advocates of a clear, universal set of minimum human rights at work that apply to all nations and firms, because, as Moberg writes, "some uniform standards are legitimate, whatever the level of development." As Broad and Moberg argue, the effectiveness of such codes depends on building wide consensus around a small set of what are deemed core rights. But to achieve consensus on rights, sharp lines must be drawn between covered and uncovered areas. Within areas covered by rights, there are then crisp distinctions between acceptable and unacceptable conduct. But these distinctions are arbitrary except from the point of view of the political battles to achieve consensus. Moreover, determined firms can easily escape the discipline of its code. And nothing in the process of reaching consensus, or enforcing a code of rules, fosters movement to improve labor standards themselves.

Thus, as Heather White observes, health and safety problems are widespread, very important to workers, and often relatively easy to fix. Yet health and safety seldom appears on lists of "core" rights. Discrimination, which is not manifestly more (or less) widespread, onerous to employees, or difficult to fix is. Moberg himself is aware of this sort of problem: He would like living, or at least minimum, wages to be part of a consensus on basic labor rights, but the idea is too controversial to survive the exclusionary razor of international consensus. The same arbitrariness plagues such basic rights as freedom of association. Very generally speaking, OECD countries are better at respecting this basic right than developing countries. But there are important differences within the "advanced" group. Indeed, Human Rights Watch¹ last September reported widespread violations of the rights to organize in the United States. Among developing countries the differences are at least as great as between the United States and, say, Austria. Countries like China and El Salvador severely restrict this freedom, while India is more respectful of it. If a bright line separates countries that respect the right to freedom of association from those that don't—and without such a line there can not be a code of core rights—on which side does the United States fall? India? The three best factories in China? Its three worst?

Such ambiguities invite wrongdoers to make a sham of compliance. To silence international outcry, a dictator might stop assassinating labor leaders (a plain violation of freedom of association) and instead intimidate workers by organizing them into official unions controlled by his son in law (a practice not so easy to categorize). A company that doesn't want to be denounced for discrimination can recover what it spends on, say, treating men and women equally, by cutting back expenditures on health and safety.

Finally, at their very worst, bright-line rules sometimes "solve" labor problems while imperiling the people the rights were meant to protect. This is the result, as we know from the work of Kaushik Basu and others, in cases such as child labor, where uniform prohibitions and sanctions can push children out of factories into prostitution, and deny their families crucial incomes.

These observations on the limits of minimal rights are hardly original to us. Several of the respondents, in fact, recognize the weaknesses of strategies based on bright-line rights and uniform standards. But the remedies they offer seem almost a fig-leaf for despair, rather than the foundations for a robust alternative.

Guy Standing, for example, worries that the declaration of core standards by the International Labor Organization (ILO) will create a ceiling rather than a floor for the development of worker rights. He is aware of the mismatch between sharp lines and complex reality. (The elaborate system of classification of labor conditions of which he speaks, however, is a research project, not an operating system; and it is more likely to encourage debate of the kind that RLS aims to provoke and systematize than to resolve matters once and for all.) We share, as should be obvious from the design of RLS itself, his concern for the possibility of collusion between monitors and those being monitored.² But we simply don't follow his conclusion that a system based solely on incentives will do better than our carrot-and-stick public-accountability approach. Wouldn't the honest application of incentives also require social monitoring?

Ian Ayres is also aware of the complexity of the regulatory problem. Indeed, he worries that any system, RLS included, that tries to take this complexity fully into account will suffer from information-overload. To reduce institutional burdens to manageable proportions he suggests monetizing labor standards by basing them on wages and benefits. The great danger, as he acknowledges, is that crucial areas such as health and safety are neglected because costs and benefits are hard to monetize. For this reason it is unlikely that his solution will find much support, except as a very chastened second-best. RLS is a way to develop metrics that incorporate additional dimensions of worker welfare, monetizable and not. As to the information-overload problem in general: experiences in other regulatory areas noted in our piece suggest that RLS-type systems both need and supply much more information than traditional ones and often develop the institutional machinery necessary to effectively process that information.

Pranab Bardhan argues that RLS has troubles dealing with unintended consequences, and especially the problem of boundaries and spillover effects between firms in the "informal" and "formal" sectors. We agree that this is a large difficulty for RLS. However, we have reason to think that RLS will respond better to these problems than other regulatory strategies. First, we believe that the same principles of transparency, comparison, and competition would be useful in continuously improving labor standards in the informal sector, though the institutions would obviously differ from those designed to regulate multinationals and their suppliers. Second, RLS is an open system in which parties are encouraged to raise new concerns as relevant dimensions of social performance. We would expect NGOs and other groups to raise objections to allegedly successful programs and regulations in RLS on account of their spillover effects on the informal sector or other unintended consequences. Once raised, successive innovations would have to account for these no longer unintended impacts in order to earn high marks.

So for example, the ILO and many others now recognize that developing a humane, and sane, approach to realizing children's rights require careful, coordinated, pro-active experimentation with strategies such as compensatory measures and integrated programs. Unlike bright-line rights and their enforcement institutions, RLS aims to encourage, evaluate, and diffuse just such constructive exploration.

Because the most successful programs are likely to come from many different kinds of actors, it would be a mistake to assign some a privileged role in developing these solutions and standards. That is why we reject the suggestion of Levinson, Moberg, and Broad that organized labor and traditional regulators can be the only real champions of labor standards. A central aim of RLS is to devise methods to increase space for free association, in part because we expect that worker organizations will often prove to be effective monitors and innovators. But improvements can also come from other quarters—NGOs, informal workers, or even firms themselves. A strategy that recognizes and encourages this diversity does not threaten the cause of labor or the project of regulation, but rather strengthens and complements it. Similarly, we reject Basu's suggestion that labor standards ought to be determined by a "consensus from the tropics." We whole-heartedly endorse developing nations as full participants along with other parties, but disagree that the tropics (especially where governments are not responsive to their citizens) should possess a monopoly on labor standards.

RLS is speculative and incomplete in its formulation—in keeping with its own design principles. We cannot pretend to have answered all of the questions; answers must arise from collaboration between practitioners and observers of regulation. Our original contribution grows out of such collaboration, and it is our hope that this exchange will foster it further. •

2 We ourselves have written some of the strongest critiques on record of monitoring practices. See O'Rourke, "Monitoring the Monitors," http://web.mit.edu/dorourke/www/.