Trade Agreements and Labor Standards

Theodore H. Moran
Theodore H. Moran Nonresident senior fellow - Peterson Institute

May 1, 2004

The National Academy of Sciences report, Monitoring International Labor Standards: Techniques and Sources of Information, shows that assessing compliance can be done thoroughly and transparently, allowing government officials, multilateral agencies, “socially responsible” investors, and nongovernmental organizations to identify where they differ on crucial aspects of the assessment process.

Although the report, prepared by a committee I chaired, carefully avoids any discussion of inserting labor standards into trade agreements, the analysis has important implications, in my opinion, for the trade and labor standards debate. It reveals how far the world is from agreement on determining compliance so as to instruct trade dispute panels and appellate bodies; it shows that the proposal to “let the ILO [International Labor Organization] determine innocence or guilt” could not possibly work except for simple cases in small, passive states; and it shows that formulating a multilateral jurisprudence for use in trade and labor cases would require fundamental, substantive changes to labor law in both developing and developed countries, including the United States.

This policy brief outlines the challenges in assessing compliance with the four ILO core labor standards, lays out a simple framework for “due diligence” in investigating compliance—which could apply to China, Mexico, Sweden, or the United States—and addresses the implications for trying to include labor standards in modern trade agreements.


Assessing Compliance and Non-Compliance

The ILO’s 1998 Declaration on Fundamental Principles and Rights at Work defines the four core labor standards as freedom of association and effective recognition of the right to collective bargaining; the elimination of all forms of forced or compulsory labor; the effective abolition of child labor; and the elimination of discrimination with respect to employment and occupation. The difficulties in assessing compliance begin with defining what the core labor standards mean and specifying the obligations assumed by countries that want to comply. Another challenge is the need to “operationalize” compliance, meaning to identify indicators that allow outside observers—other governments, international agencies, labor associations, church groups, student logo committees, college presidents—to judge whether a country is complying with ILO standards. Finally, assessing compliance requires determining how to use often incomplete, non-representative, non-comparable, or potentially biased sources of information to evaluate compliance or non-compliance, and to draw well-justified inferences from the available evidence.

Defining Country Obligations

Freedom of association and effective recognition of the right to collective bargaining constitute one of the oldest core labor standards, with a lengthy and detailed record of debate about what constitutes compliance. It might be comforting to think that decades of work by the ILO—the repository of multilateral investigation into allegations of labor standards violations for more than fifty years—would have left issues of definition and compliance assessment thoroughly settled. But the reality is otherwise.

There is broad international consensus, for example, that governments should refrain from punishing workers who back their negotiations with employers with strike threats, and that governments should have enforcement mechanisms in place to prevent employers from taking action against workers who strike. This would logically seem to imply that labor legislation that permits employers to hire permanent replacements for striking workers contravenes this core standard. But ILO jurisprudence comes to no such conclusion. Instead, the ILO has equivocated over the years, acknowledging that the ability of employers to fill the positions of striking workers with permanent replacements “poses a risk” to effective recognition of the right of collective bargaining but does not necessarily violate this standard unless it occurs on an unspecified “extensive” basis.

The ILO lists the United States along with Burkina Faso, Cape Verde, Central African Republic, Djibouti, Madagascar, and Niger as the principal countries around the globe with legislation that permits hiring replacements for striking workers.

Another problem area involves laws that require both a “closed shop”—which permit collective agreements requiring employers to recruit only workers who are trade union members and who must remain union members and pay their dues in order to keep their jobs—and the “right to work.”

Conventional labor market analysis considers “closed shop” requirements an infringement on the ability of workers who are not part of a trade union to engage in freedom of association. With “right to work” laws, in contrast, the state guarantees the right of workers to obtain jobs that receive the benefits of the union’s collective bargaining even if they neither participate in collective bargaining organized by unions nor pay union dues. This support for “free-riding,” in the analytics of social science, arguably constitutes a powerful, indirect constraint on the ability of trade unions to organize workers effectively. Here again, ILO jurisprudence has left the basic issues unresolved, allowing both “closed shops” and “right to work” laws, as long as states do not impose by statute a particular trade union monopoly.

Gaps in ILO Interpretations

There are gaps in the ILO’s treatment of freedom of association and right to collective bargaining. One of the more prominent is the possible control of unions by criminal elements. In some countries, government officials or gangsters may organize unions as a protection racket, with employers recognizing the unions so that their store windows will not be smashed on a regular basis. Here ILO jurisprudence is silent, imposing no anti-corruption standard of conduct for union leadership. As a result, investigators might consider high union density as evidence of a country’s respect for freedom of association and right to collective bargaining, whereas the reality is widespread control of unions for criminal purposes.

There are many other areas in which ILO jurisprudence is clear but country practices vary considerably. On the right to strike, for example, there are controversies about regulations forbidding certain kinds of strikes such as “sympathy strikes” or “protest strikes.” In ILO jurisprudence, trade unions should have recourse to protest strikes aimed at criticizing a government’s economic and social policies, and sympathy strikes should be lawful when the initial strike is lawful. There are controversies about whether a majority of the workers affected have to approve a strike, a requirement the ILO has said is “excessive.”

The problems associated with identifying criteria for country compliance applies to other core labor standards as well. Does respect for non-discrimination, for instance, require provision of subsidized legal services to help with grievance actions or to protect those who file a complaint against retaliation? The strong presumption is that the answer is yes, but a judgment about what would constitute an “adequate” level of services or “adequate” amount of public subsidy could vary greatly depending upon a country’s particular circumstances.

Does compliance with non-discrimination prevent the use of explicit quotas (by race, religion, nationality, tribe or ethnic group) for hiring? While many member states consider explicit quotas to be anathema, ILO jurisprudence does permit their use to achieve numerical targets.

Is compliance with the forced or compulsory labor standard incompatible with private work programs in prisons, with required participation in prison work programs as a condition of parole, or with privatization of prison systems? ILO jurisprudence considers employment of prison labor by private contractors to be impermissible, but many governments, including New Zealand, the United Kingdom, and the United States consider private contractors to be an integral part of the modern management of penal institutions.

Identifying Indicators of Compliance

The first steps for policymakers in assessing a country’s compliance are to outline what obligations a country assumes when it pledges to adhere to a core labor standard and then to identify indicators that determine whether a country is in compliance or not. These indicators might be found on at least three levels: the first involves appraising a country’s legal framework relating to the core labor standard; that is, for example, whether laws and regulations protect freedom of association, right to collective bargaining, and right to strike within the assessor’s understanding of the country’s obligations.

The second level involves appraising the government’s performance in implementing the standard. Evaluators must examine both effort and effectiveness, including whether the government is devoting enough attention to protecting organizing, bargaining, and striking, and whether in so doing the government is generating an acceptable level of results to qualify as being “in compliance.”

Along both measures of government performance—effort and effectiveness—the resulting evaluation will depend upon the resources available to the government, and the urgency of competing claims on those resources (e.g., to deal with HIV/AIDS or to provide potable water). Since the level of development and the government’s competing needs will limit what might be devoted to enforcing compliance with core labor standards, the evaluation of compliance will have to involve a decision about whether and how much to discount the inputs and outcomes to account for these factors. In the contemporary world, the degree to which a country can be “forgiven” for low levels of public sector inputs or poor public sector results due to the country’s poverty and/or competing need for expenditures elsewhere—whether for freedom of association, child labor, forced labor, or discrimination—would be almost entirely subjective.

The third level of investigation deals with labor market outcomes that are not under the direct control of the government, such as the number of strikes that take place. As I discuss later, however, similar outcomes (e.g., a large number of strikes) could often be interpreted in opposite ways, either as evidence of robust compliance or deficient compliance.

Quality of Information Sources

The sources of information that government officials, NGOs, multilateral lending agencies, and others might use to evaluate compliance with each of the core labor standards involve both qualitative reports and quantitative data. The quality of both kinds of information is uneven. With regard to freedom of association and right to collective bargaining, in particular, the available sources of evidence may be limited to conditions at relatively large firms within the formal sector that comprise a tiny proportion of all firms within a country. This information does not constitute a representative sample of how workers are treated in the overall economy.

Qualitative sources may provide field reporting on worker treatment. In an ideal world, the evidence would come from experts who were independent of the parties (governments, employers, workers) under investigation. In practice, however, qualitative sources most often reflect the perspective of the reporters and may be skewed. Many reports are complaint-driven, leading to selection bias. Others are screened by committee or softened for political reasons, such as to avoid offending the country under investigation.

Quantitative Sources

The relative importance of the data collected using quantitative sources of information to measure compliance may vary with the particular standard under investigation. For cases involving freedom of association and the right to collective bargaining, some statistics, including the number of strikes and the number of striking workers, will be important, but most of the crucial evidence about compliance is likely to be qualitative rather than quantitative. On cases involving discrimination, however, some quantitative data sets are likely to be central to the investigation—differences in literacy rates or school enrollment rates between men and women, or among different ethnic, tribal, racial, religious, or other groups, for example, are a key target for assessors to evaluate. Similarly, an important indicator in judging whether a country has effectively abolished child labor or is making sufficient progress toward that goal is survey data recording changes in the number and proportion of children ages five to seventeen who are economically active and the number of hours they work.

In an ideal world, there are rigorous standards that survey samples must meet in order to be considered an accurate portrayal of conditions within a given economy, or among countries. In the real world of collecting data on labor markets, however, these standards are seldom met. The data are often neither comprehensive nor comparable within countries over time, and are unlikely to be comparable across countries.

As I noted earlier, drawing appropriate inferences from the information available to investigators requires great care. In the case of freedom of association and right to collective bargaining, for example, evaluating information about the frequency of strikes, the length of strikes, and person-days devoted to striking is particularly tricky. Low numbers might mean that labor and management have a good, productive relationship. But low numbers might also mean that the system of labor relations is repressive, and workers fear that if they strike they will be replaced or suffer other reprisals.

The puzzle of what to make of ostensibly objective data applies to the other core labor standards as well.

What conclusions about discrimination can be derived from relative wage levels? Do lower wage rates for women indicate that women are less productive, that they receive less training, or that they are less assertive because they have inadequate protection against retaliation when they complain about pay?

A Framework for Investigation

Given the difficulties in defining what core labor standards mean in specific circumstances, in specifying obligations assumed by governments, in operationalizing compliance and non-compliance so that observers can recognize one or the other, in finding reliable sources of information, and in making defensible inferences, the report proposes a simple matrix framework within which assessors and counter-assessors can carry out a comparably thorough exercise in investigating compliance with each core labor standard, while allowing all participants to identify differences and challenge contrasting interpretations.

Assessors are invited to evaluate a reasonably complete array of indicators pertaining to legal framework, government performance (effort and outcome), and overall labor market outcomes for a given standard, and to make a decision about how to characterize the extent of compliance and whether compliance is deteriorating or improving. The report does not attempt to delineate what kinds and levels of compliance would determine whether a country was doing an “adequate” or an “inadequate” job of compliance. The report argues that it is not possible to establish objective cutoff points—say, the proportion of the national budget devoted to labor inspections—that are appropriate for every country and every situation. Ultimately, the decision about adequate or inadequate compliance depends upon the problematic issues examined earlier, such as the assessor’s interpretation of a country’s obligations, hierarchy in weighting the components of compliance, and inference from the available evidence.

Rather than pretending that subjective and contingent elements can be eliminated from the assessment process, the matrix framework simply provides a roadmap to “due diligence” in investigating compliance in a manner that allows any assessor to contest the appraisals of others, thereby forcing all parties to defend their evaluations openly and in detail.

What are the implications of this analysis for proposals to make compliance with core labor standards a condition for countries to enjoy trade benefits, and to make non-compliance a condition to deny a country trade benefits on a bilateral, regional, or multilateral basis?

My own conclusion is that the analysis presented here has rather striking implications for the feasibility of conditioning the provision of trade advantages upon verdicts of guilt or innocence in complying with core labor standards. (For the debate about the desirability or undesirability of including labor standards in trade agreements, see my recent Brookings book, Beyond Sweatshops.)

Even assuming heroic efforts on the part of investigators in carrying out “due diligence,” with thorough and dispassionate evaluation, it becomes clear that the world is far removed from having the consensus that could serve as the basis for a multilateral jurisprudence—path-breaking past work of the ILO notwithstanding—to decide cases that could be backed by sanctions including trade benefits or fines. It is hard to imagine how to instruct members of trade dispute settlement panels—or subsequent appellate bodies—so that they could render consistent verdicts of guilt or innocence in any but the most widely accepted, clear cut, and egregious cases of violations of a core labor standard.

This does not mean that asymmetrical bargaining power in negotiating bilateral preferential trade agreements cannot perhaps be used to improve laws and government performance in the smaller, weaker participant. That would help bring the country toward compliance with those aspects of the core labor standards that are relatively settled, especially where it is unlikely that the state bringing the complaints faces any chance of being challenged on its own interpretations and practices. For example, under the U.S.-Cambodian trade agreement, the United States may challenge labor practices in Cambodia, but the possibility that Cambodia will challenge labor practices in the United States is almost nonexistent. But this does not mean that the international system has created a basis for multilateral jurisprudence with regard to labor standards.

In criticizing the adequacy of labor legislation in the countries in the Central American Free Trade Agreement (CAFTA), three members of the House Ways and Means Committee—Democrats Sander M. Levin (Calif.), Charles B. Rangel (New York), and Xavier Becerra (Calif.)—objected to national laws in some of the Central American states. These laws do not limit voting on a strike to union voters, and do not automatically confer accelerated judicial review upon cases of workers dismissed for union activities or award them more than double severance pay.

The House members objected to the absence of federal legislation allowing agricultural workers in general, and migrant agricultural workers in particular to unionize, and to elect foreign nationals as their leaders (e.g., that Nicaraguan migrant agricultural workers in Costa Rica would be organized and led by Nicaraguan nationals). They objected that some national laws—in Honduras, for example—required that a firm have at least thirty workers to form a trade union. National laws do not prevent an employer from threatening to close or move a plant when faced with an organizing campaign.

A genuine multilateral dispute settlement mechanism would have to give comparable scrutiny to all member states, including the United States. But in the United States, non-union workers have a right to vote about striking (in states with “right-to-work” provisions); there is no expedited handling of anti-union dismissal cases; the National Labor Relations Board normally orders no more than reinstatement of the employee with back pay; and firms are allowed to hire permanent replacements for striking workers, and to allow those replacements to vote in an election decertifying the original union. Agricultural workers are exempt from U.S. federal labor statutes (although some states, like California and Florida, provide for the right to organize and for collective bargaining under state law). There is a dollar volume of business necessary to establish federal jurisdiction under the Commerce clause. Employers threaten to move or close their plants when faced with union organizing campaigns, according to international labor union sources, in more than 60 percent of the cases.

Appealing to ILO interpretations of what constitutes acceptable legislation—let alone acceptable levels of performance—would only reveal how far CAFTA participants are from having in place a consistent basis for judging each other’s compliance status. What is really involved in the inclusion of worker standards in U.S. free trade agreements is not an attempt to expand multilateral jurisprudence but an attempt to use unilateral arm-twisting to push other countries in a U.S.-determined righteous direction.

Whether hypothetical negotiations to move the members of the WTO toward congruence on the precise interpretation of what constitutes compliance would constitute a race to the top (tightening standards worldwide), or a race to the bottom (loosening standards worldwide) is unknowable, but certainly America’s right to work laws and provisions on the right to hire permanent replacement workers would be central to the debate. A successful outcome to such hypothetical negotiations on labor standards would doubtless elicit the same surprise among many in the United States that has accompanied NAFTA Chapter 11 cases, which cover the imposition of environmental and other regulatory policies on international firms.

Daniel Price, a principal architect of NAFTA Chapter 11, points out that “all of a sudden, from the perspective of the United States, NAFTA says you can be a defendant. The United States can be a defendant. My only advice is, get over it. It is true.”


There is much work to be done to generate agreement on what constitutes a country’s innocence or guilt in compliance with core labor standards: what constitutes an adequate foundation in law, appropriate level of effort, and necessary outcome in performance to be in compliance. The NAS report provides some fundamental guidelines and caveats—all of which can surely be improved with use—to push assessors toward ever more thorough and rigorous investigation of how workers are being treated around the world, to sharpen their ability to identify where they disagree, and to stimulate them to explore why. Americans are seeking assurances that their prosperity does not come at the expense of workers in the developing world. This report shows them how to begin to find an answer.