Lessons from the First U.S. Win in a Trade and Labor Dispute: The USMCA RRM Atento Decision

For the first time, an arbitral panel has sided with the United States in a trade and labor dispute. On August 21, the United States and Mexico announced the publication of the second panel decision under the USMCA/T-MEC/CUSMA Facility-Specific Rapid Response Labor Mechanism (RRM). The panel found that there “had been a Denial of Rights” at Atento Services, a company with call center facilities in Hidalgo, Mexico. For background on the case, the primary arguments of the parties, and the panel hearing, you can check out this post from May.

As I noted in my earlier post, this dispute is special for several reasons, but most importantly now, the Atento decision is special because it is the first in which a panel has found a Denial of Rights under the RRM, and therefore the only case in which the United States has been successful in a labor proceeding under an FTA.

This post will very briefly recap the panel’s principal conclusions. The panel rejected Mexico’s arguments (1) that the facilities in question were not “Covered Facilit[ies]” under the USMCA, and (2) that the Mexican government’s intervention was sufficient remedy for any Denial of Rights that had occurred at the facilities. In reaching these conclusions, the decision reinforces issues raised in the context of the first RRM panel about the purpose of the RRM, the role of panels under the Mechanism, and the relevance of international law.

Interpreting the Treaty Through a “Labor Law Perspective”

As anticipated, the first topic covered in depth by the panel in the decision is the definition of “Covered Facility.” This is not a new issue. The first RRM panel (dealing with the San Martín mine situation) confronted a similar defensive argument from the Mexican government, and in fact the Atento panel relied heavily on the San Martín panel’s conclusions in reaching its result.

Let’s start with the definition in the agreement. There are two pathways to show that a facility is a “Covered Facility”. The USMCA defines a “Covered Facility” as:

a facility in the territory of a Party that: (i) produces a good or supplies a service traded between the Parties; or (ii) produces a good or supplies a service that competes in the territory of a Party with a good or a service of the other Party, and is a facility in a Priority Sector.

I will say up front that I think Mexico likely had a losing argument here, unlike in the San Martín case, which was a closer call. If we took a survey of people on the street – in either country – I think we would find most people would conclude that a call center in Mexico is a Covered Facility, especially under the second definition, if we gave them a few more facts about call centers in the United States and Mexico. The panel also reached that conclusion, but the panel’s reasoning is rather difficult to parse.

With respect to the first definition, the panel chooses to adopt Mexico’s interpretation that the agreement requires a party to demonstrate that the good or service of the facility is “effectively traded” (the better translation may be “actually traded”; the passages I quote are from the official English translation of the original Spanish.) and not just of a kind that is traded. The panel reached this conclusion after a somewhat novel application of the Vienna Convention on the Law of Treaties. For example, in considering the “structural context” of the definition of “Covered Facility”, the panel invokes a 2010 law journal article in which the author “cautioned that labor law has historically not been subject to extraterritorial application” as a reason for interpreting the text broadly. On another occasion, the panel refers to the same law journal special issue to emphasize this point about the “historical lack of extraterritoriality”. Following this discussion, the panel finds that Atento’s services are not actually traded between the two countries.

With respect to (ii), the panel adopts the interpretation that the first RRM panel (the San Martín panel) developed, although the Atento panel makes clear that the prior panel’s conclusions were not binding on it. In getting there, the panel again takes several detours and turns. I will flag just two.

First, the panel appears to question, without saying expressly, that a sub-issue before it is the level of specificity of the service, so that it can figure out what services compete against each other. The panel seems to be asking: what is it that the facility does and, therefore, what are its comparators? How general or specific ought one be? Here, the panel turns to industrial categories of services as a means of classification but then does not rely on those categories to reach its conclusion.

Second, the panel states at least twice that it is interpreting the language of the RRM from a labor perspective. For example, the panel states that:

A second contextual element in interpreting “competes” is that panelists on the lists are selected based on their expertise in labor law. ... [so] the terms should primarily be interpreted from a labor law perspective rather than from a trade law perspective.

The panel also decides that: “given the nature and purposes of the RRLM, the burden of proof should not require absolute certainty.” (The panel doubles down on these ideas in its discussion of whether there is a Denial of Rights, saying that it will apply norms and rights recognized by the International Labor Organization instead of the WTO, including because the RRM is “a special labor accord” which means it must be governed by its own special rules.) And elsewhere, the panel discusses the “priority sectors” as defined in the Agreement and implies that the parties could have been more efficient in describing the sectors that are not covered rather than those that are covered.

These sorts of statements are, I fear, unhelpful. They do not align with the rules of interpretation under international law and they raise doubts about the approach taken by panels under the Mechanism, hearkening back to what I think is a fundamental challenge of the RRM and the role of panels under it: is this a dispute settlement proceeding or is this a fact-finding process?

Trade Rules of Evidence

After confirming that the Atento facilities are covered, the panel finds that there “were” denials (see my point below about timing) in three steps.

First, the panel recounts many statements made by workers and others close to the situation at Atento, most of which are redacted in the report, but which convey significant irregularities in the unionization process at the worksites. The panel then concludes that this evidence proves that the company wrongfully interfered with and discriminated against the union.

Second, the panel queries whether the wrongful interference and discrimination constitute a Denial of Rights under the Agreement. Interestingly, it relies on determinations by the ILO Committee on Freedom of Association and finds that a Denial of Rights “can occur not only due to legal absence but also when material conditions such as interference, reprisals or power imbalances prevent the free and effective exercise of freedom of association and collective bargaining.” After extensive discussion of international labor law norms, the panel ultimately concludes that (1) Atento denied its workers the exercise of their rights, and (2) Atento violated Mexican labor law in its multiple acts of interference and arbitrary dismissals. 

Third, the panel decides that the government of Mexico’s intervention did not remediate the Denial of Rights, which I will discuss further below.

Throughout this discussion, the decision is quite brief in linking the evidence to the outcome, which may make it difficult for members of the public and other consumers of RRM decisions to follow. About eight years ago, Mark Wu and I wrote a paper on the topic of evidence in trade disputes in which we sought to tackle some of the challenges faced by state-to-state litigators where they seek to prove complicated policies or practices of their trading partner governments. We identified labor as one such area where evidentiary presentation is challenging and we discussed how that difficulty can make it hard for a general audience to understand the fullness of the circumstances as well as the panel’s reasoning. Evidence was a major issue in the U.S.-Guatemala labor dispute, for example. Here, in addition to some of its unusual elaborations of the thresholds for showing a Denial of Rights, the Atento decision seems to have lost the thread at some points between the presentation of worker statements and relevant legal conclusions.

Timing Is Everything?

With respect to the final step of the panel’s analysis, Mexico had claimed that the panel did not have “jurisdiction” to consider actions that the Mexican government had taken with respect to the facilities in question. This argument gets complicated quickly in part because, as I have discussed in relation to the San Martín situation, the text of the RRM does not mention “jurisdiction” nor does it treat any of the concepts as jurisdictional on its face. It came up here because the Mexican government had taken some action against the company during its review and exchange with the United States before the establishment of the panel.

Why, according to Mexico, should the panel not have reviewed the Mexican Labor Ministry’s actions? Mexico emphasized that the RRM is a tool for reviewing company actions, not government actions. In that argument, Mexico is no doubt right. But, the United States claimed, the panel needed to evaluate whether the actions of the Mexican Ministry remedied the problem at Atento.

One difficulty with sorting these arguments is again the language of the treaty. First, the RRM text states that parties can convene a panel to determine whether there “has been a Denial of Rights” and to “verify compliance” with Mexican labor law. Elsewhere, it states that the panel shall decide whether “there is a Denial of Rights”. And still elsewhere, the Annex notes that the purpose of the RRM in general is to establish whether “workers are being denied” their rights. So, some of the agreement is framed in the present perfect test (“has been”) and other parts of the treaty are framed in the present: “is” and “are being”.

This ambiguity prompts us to ask: what moment of time matters for purposes of the panel? Is the panel asked to verify compliance at the moment it goes to visit the facilities? What if it finds there was a Denial at some point in the past? What if circumstances change between its hearing and its decision?

Second, a provision in the RRM Annex notes a panel may be called upon to evaluate the respondent’s remediation efforts, but it is not obvious that the conditions precedent for such an evaluation were triggered here.

Perhaps considering the RRM’s textual silence, the panel does not readily explain against what metric the panel assesses the government’s intervention apart from concluding that it “did not change the labor climate”:

In view of the gravity, duration and structural nature of the denial, the Panel notes that the remedies invoked by the respondent Party to be inadequate in counteracting the effects of the arbitrary dismissals and the company’s interference with the workers.

What Comes Next

What this decision means for the Atento company – or, more importantly, its workers – is not entirely clear. The treaty provides:

After receipt of a determination by a panel that there has been a Denial of Rights, the complainant Party may impose remedies after providing written notice to the respondent Party at least 5 business days in advance. A respondent Party can request that consultations be held during that 5 day period.

We do not know whether the United States has taken these steps and what remedies, if any, it is considering. USTR did not discuss its next steps in its press release. Given that Atento is a call center, it is hard to see how tariffs can be useful here, even if we assume that the corporate leadership is to blame. Again, the treaty states that the United States:

may impose remedies that are the most appropriate to remedy the Denial of Rights. The complainant Party shall select a remedy . . . that is proportional to the severity of the Denial of Rights and shall take the panel’s views on the severity of the Denial of Rights into account when selecting such remedies. (emphasis added)

The treaty suggests that remedies may include “the imposition of penalties on . . . services provided.” As the first services case, the selection of remedy here, if any, will be informative. There is also U.S. law guiding the USTR, but it does not do much to provide either ideas or guardrails. First, the Trade Representative must consult with the “appropriate congressional committees” and it “may . . . apply other remedies that are appropriate and available” under the agreement. The governments are required to stay in touch about the remediation and may re-convene the panel, it seems, if they disagree.

More generally, perhaps this outcome will lead more companies, and the Mexican government, to try to settle their differences with the United States under the Mechanism. Perhaps it will embolden the United States to continue to maintain course on the RRM investigations, spending down more of the $30 million that Congress appropriated to the Office of the U.S. Trade Representative for this purpose, and even to seek to maintain the tool as part of the USMCA review. While there may not be prominent advocacy specific to the Mechanism in this Administration, there is plenty of support for enforcement.