It was such an honor for me to be invited back to the China Forum again this year, this time to join a panel discussion alongside Dr. Adrian Zenz (who also moderated our discussion), Brian Hoxie, Director of the Forced Labor Division at U.S. Customs and Border Protection, and Louisa Greve, Director of Global Advocacy for the Uyghur Human Rights Project.
For the uninitiated, Dr. Zenz is the German researcher who effectively broke the story1 on China’s campaign of persecution against the Uyghurs. Ms. Greve is a longtime advocate for Uyghurs whose work was instrumental in leading to the creation of the UFLPA. And of course, Mr. Hoxie is directing CBP’s forced labor trade enforcement. There is truly no one in the U.S. government more equipped to talk about how this law is being enforced than Mr. Hoxie.
It is humbling to be invited to address this community that is so rightly proud of the work that they have done to produce (and operate) what is arguably the marquis U.S. human rights law. If I’m honest, it is also quite challenging, because there is so much about this law that is still, in my view, so widely misunderstood. I’m publishing my prepared remarks, as this was my best attempt to illuminate what doesn’t yet work well about this law, and to make the case for why it’s worth improving.
I spoke with a reporter this week who asked if I’m optimistic that this law can achieve its intended effect of improving the lives of Uyghurs. I told her, unequivocally yes. Don’t get me wrong. On any given day I can be as discouraged by anyone by the lack of focus and risk of severe adverse commercial results to non-wrongdoers. But it’s MLK’s arc of the moral universe. So gird your heart against cynicism, and look around
As a result of this law, every company I know is presently working to dig deeper into their supply chains than they have ever previously had a legal (let alone commercial) imperative to do so. In other words, as far as I can tell, the business community is holding up its end of the bargain.
Yes, the initial foray of this law has been (profoundly) messy. But such is life (and legislating) in a democracy. That’s a feature, not a bug. As Ms. Greve pointed out during the panel discussion, for the law to achieve its full intended effect, there is a future state where multiple countries are aligned in the same effort.
But between here and there, someone is going to have to sweat the details of how this is supposed to work, and get into the weeds. I hope this discussion adds value to that journey
It is a great honor for me once again to join the China Forum. I’d like to extend my thanks to Dr. Zenz and Amb. Bremberg for the invitation.
I am a partner in the international trade practice at Kelley Drye & Warren, in Washington DC. My practice consists of advising corporations on how to ensure that they remain compliant with U.S. trade laws, particularly those governing the importation of goods.
Sometimes, folks have described my practice as being “in the weeds” of international trade. That used to bother me, until I learned to embrace it. Because, by virtue of my line of work, I do spend all day, every day, curating my expertise from within the rules-based international system.
It is from this unique vantage that I’d like to describe for you the Uyghur Forced Labor Prevention Act. Even if you think you understand what this law is, or how it operates, I’d ask you to suspend your preconceived notions, and hear it from me anew.
A few weeks ago, in the newsletter I publish on forced labor trade enforcement, I described UFLPA enforcement as unfolding “according to an extra-legal process, governed neither by statute nor regulation, and so far without the benefit of input from the judiciary.”
This provoked strong disagreement from some folks who believe the UFLPA is a good and important law, who felt it sent a wrong, or at the very least unfair message. And while I agree that the UFLPA is good and important, I must be clear: that was not an expression of opinion; it was a statement of fact.
The UFLPA is being enforced according to a process not contemplated by the law itself. While I believe that CBP is doing its level best to enforce the law vigorously and fairly, my audacious claim holds water.
Before I unpack that, I will highlight two critical features of the UFLPA that I agree are good and important. I will then explain how the rebuttable presumption of the UFLPA works, how CBP picks targets for detention activity, and what happens after CBP detains merchandise. This is the process that I describe as “extra-legal”. I’ll conclude with some thoughts on why this all matters, and should concern an audience of policymakers and lawmakers, even those in favor of “vigorous enforcement”.
The Goodness and Importance of the UFLPA
First, by enacting the UFLPA, Congress answered a question that only it could answer: namely, that so far as the United States is concerned, the labor component of China’s state-sponsored programs targeting Uyghurs constitute forced labor. This question might have proven difficult for CBP to resolve on its own just, simply in the course of enforcing the forced labor import ban.
If one major world power wants to spin the compelled labor dimensions of an ethnic elimination campaign with euphemisms like “labor transfer” and “poverty alleviation”, it is incumbent for other world powers—and especially those with commercial ties to such practices—to clarify what is actually happening. Ergo, the determination of Congress that such conduct constitutes forced labor for purposes of U.S. law, was essential.
Second, as a result of the state-sponsored nature of this forced labor, and the pervasive quality of the ethnic elimination campaign throughout Xinjiang, it was necessary to define the scope of the law’s coverage to include all of the affected geography (i.e., all of Xinjiang). Moreover, due to the labor transfer programs, which move forced Uyghur labor throughout the country, the adoption of the UFLPA Entity List, particularly to help identify the recipients and beneficiaries of such labor, was an important achievement.
While there may be some commercial enterprises within Xinjiang that might not be active participants in forced Uyghur labor, the Chinese Communist Party’s commitment not only to perpetuate abuses, but to lie about it, cover it up, feign shutting it down, and then declare the overall approach “totally correct”, all while running a surveillance state that makes free movement by outside actors and independent assessment of the facts on the ground impossible, rendered the region-wide approach of the UFLPA a necessity.
UFLPA Rebuttable Presumption
It is common for proponents of the UFLPA to talk about the law’s important role in “shifting the burden of proof to companies to prove their shipments aren’t tainted with forced labor.” Technically, that is a true statement. But it elides a distinction that is critical to understanding how the law works.
The rebuttable presumption established by the UFLPA is a presumption about the condition of labor used to produce certain goods—namely, those made wholly or in part in Xinjiang, or by a UFLPA Listed Entity. These are presumed to have been made with forced labor, and so are inadmissible under the U.S. forced labor import ban.
But there is a conceptually prior question. Which goods are made wholly or in part in Xinjiang, or by a UFLPA Listed Entity? On this question, the UFLPA is silent.
In some instances, the answer might seem fairly obvious. For example, goods might be shipped to the U.S. directly from Xinjiang. Or, CBP might open a container at the port and discover boxes and boxes of red dates helpfully labeled “Bingtuan red dates” (which is the transliterated name for the XPCC). But even in these instances, where it’s relatively straightforward to “spot the Xinjiang goods”, CBP found it necessary to make modifications to its data collection from importers in order not to miss the low hanging fruit.
More often than not—in fact, in almost all cases—Xinjiang content isn’t being shipped directly to the United States. And if you open a container at the port, goods made with Xinjiang content look exactly the same as goods made with admissible content.
How CBP Picks Detention Targets
This presents a quandary for CBP, because the UFLPA, and the forced labor import ban on which it is built, are enforced against specific imported goods. For better or worse, this is the power (and liability) of forced labor trade laws, as enforced in the United States. It is therefore incumbent on someone to identify goods that are made with Xinjiang content. Goods to which the UFLPA applies.
At present, the law doesn’t even provide for an importer to make a voluntary representation about the UFLPA compliance of its imported merchandise, if it wish to do so. Therefore, CBP alone has to make the call. And because CBP does not know, a priori, which goods have been mined, produced, or manufactured in Xinjiang, it has to make the best guess possible.
Now, to say that CBP has to guess which imports to detain is not to say that CBP doesn’t make an educated guess. It most certainly does. CBP has access to some information related to import activity, and where it lacks data—for example, on the depths of global supply chains beyond the foreign producer—it has purchased additional data from external software providers, and is stitching such information into the enforcement algorithms.
But data alone—even purchased data—only gets CBP so far. As I wrote a few weeks ago, informal inferential reasoning is the only true precondition to enforcement. It all starts here. CBP has to infer which goods, from which final manufacturers are most likely to contain Xinjiang content. That is why academic research outfits—NGOs, and the like—have taken on such outsize importance in the UFLPA universe. They serve up juicy, well-documented enforcement targets for breakfast.
When CBP picks targets that might contain Xinjiang content, importers respond by trying to disprove that geographic assertion. They’re not rebutting the forced labor presumption of the UFLPA, mind you. They’re just trying to prove that their goods aren’t subject to the UFLPA in the first place. This describes approximately 99.95%+ of all UFLPA enforcement to date. And according to CBP’s own stats, importers have been successful in proving that they imported compliant merchandise somewhere between 45 and 70% of the time.
In other words, almost all UFLPA detentions are being handled under a process not provided for by the statute. And a generous majority (or slight minority) of all goods stopped by CBP have proven to be the wrong goods.
What Happens Post-Detention
Now, I’m going to circle back to my audacious claim in just a moment, but before I do, let me describe very briefly what follows after CBP picks what it believes to a good target, based on available data and informal inferential reasoning. A quick tour of the weeds within the weeds, if you will.
When CBP decides to bring the enforcement storm against a particular target, it does so with great intensity. Customs will stop every shipment of goods from a given target.
Depending on the commodity or type of product CBP is targeting – cotton, polysilicon, PVC, renewable batteries – it will demand the production of traceability documentation. Traceability documentation is produced by the importer, but it is collected from the finished goods supplier, who in turn collects it from other parties throughout the supply chain. There might be 3 parties in a polysilicon supply chain, or 30 parties in a cotton supply chain. All of them have to play ball.
The documentation for traceability packages is extensive, and by CBP’s own recognition, somewhat defy easy categorization. CBP lists exemplars, but cannot define the full extent of a traceability package. You’ll need purchase orders, invoices and proofs of payment, but also shipment records, inventory records, and production records that connect the dots at each node in the supply chain. Everything has to tick and tie back together. It needs to be summarized. Translated. Indexed. Technologically enhanced traceability is always subordinate to the paper. In CBP’s view, actual transaction documents are king.
A traceability package might consist of dozens of individual documents, and be between 600 and 1000 pages long. And the question presented to CBP is deceptively simple: does this traceability package depict an actual supply chain, or does it not? This is one of the most difficult factual and conceptual questions I’ve ever encountered in international trade law.
An importer might have to produce a traceability package for every type of product in an entry. There might be 12 types of products in an entry or 50. CBP might detain 50 entries from a single supplier, and has, in some instances, detained hundreds. No published standards indicate when it will move the storm on from one finished goods producer to a subsequent target.
The further away you get from the official “high-priority sectors” the more untethered the traceability demands have become. CBP started targeting imports that contain PVC after an NGO report documented PVC production in Xinjiang. But CBP’s traceability demand included not just PVC, but every raw material in the finished merchandise, including many with no alleged link to Xinjiang whatsoever. CBP’s traceability demand for renewable batteries contemplates traceability packages for 17 different components, many of which are not subject to any allegation of Xinjiang nexus, and inexplicably, 13 different machines used in the production process, among other demands.
An extra-legal process, governed neither by statute nor regulation, and without the benefit of guidance from the judiciary? I believe the characterization is charitable. From the selection of a detention target, to the scope of a traceability demand, to the rejection of an entry for lack of some obscure, never-before-defined piece of paperwork from a company in India, to the ultimate evaluation of a traceability package, to the decision of when to move enforcement on to another detention target—none of these decisions are governed by the UFLPA, or any other individual statute or regulation in the customs law canon.
The point is not that CBP is getting everything wrong, or even that it is always abusing its discretion (though sometimes it does get things quite wrong, and sometimes can be extremely unfair). It’s just that the entire UFLPA enforcement apparatus, top to bottom, is agency discretion. If you’re not familiar with these weeds of international trade law, then you’ll have to take my word for it. This is deeply weird.
The great irony to all of this is that my experience with the CBP civil servants who are charged with enforcing this law is that, to a person, they are are doing their level best to be efficient and tenacious to the facts, and to render consistent decisions, as they work their way from one traceability package to another. It’s good and important work, if ever there was some.
Why This Matters
I am honored to share these remarks to an audience that may be among the most optimistic in America regarding the goodness and importance of the UFLPA, and I address you all, as one who shares your conviction that preventing the importation and sale of goods made by forced Uyghur labor is not just one of the most important objectives of any U.S. trade law, but is actually doable.
And yet, I feel constrained to try and explain why interested observers in the policy space should care about these weeds. It is a deeply unsatisfying to be told, as a trade lawyer, that: “this is the point?” That the goal of this law is simply to decouple the U.S. and China economies, come what may.
If elected officials believe that is the right course of action, there are tools well suited to achieving it. The forced labor import ban just isn’t one of them. This law requires us to indulge a particular conceit: that it is possible to distinguish between goods tainted with forced labor, and goods not so tainted, the former being inadmissible, and the latter eligible to trade freely. So, three quick thoughts on why this should matter.
One, the rule of law should always mean something to an American audience, perhaps most of all in this arena where the marquis U.S. human rights law is on display. There are jurisdictions that pass deliberately opaque laws for the purpose of using such “laws” as a political joystick, subject to the whims of the ruling (usually single) party. The U.S. should relish its difference in this regard.
Two, if appeals to American exceptionalism and/or continuing to perfect the U.S. legacy of constructing the international rules-based order leave you unmoved, there is also this: UFLPA enforcement is breathtakingly inefficient. It is possible to burn hundreds of millions of dollars of taxpayer dollars chasing unending document production demands on a tiny set of imports from an even smaller set of foreign producers. But that is not the only way to UFLPA. I wrote an essay a few months back called A Better Way To UFLPA, containing some thoughts on how to be a touch more strategic in enforcement.
My final appeal is this: the forced labor import ban is the ultimate condition on market access, but it has yet to be wielded with any intentionality. A certain type of company ought to loom larger in the policymaker imagination: the company that has zero ambition to import goods with any supply chain link to forced Uyghur labor, and that is prepared to invest substantially in mapping and documenting its compliant supply chain to prove as much. Such companies—many of them my clients—only want to manage their trade in a compliant posture, and yet find it impossible to do so against this enforcement landscape.
The power to define the conditions of market access, and ensure that goods made with forced Uyghur labor actually do not enter the U.S. market remains firmly in the grasp of the U.S. government. It should be wielded wisely.
Thank you.
If you’re not familiar, this write-up in the WSJ from 2019, describing the research by this “born-again Christian anthropologist” is a must-read.