New U.S. Trade Representative Katherine Tai will testify in front of the Senate Finance Committee today and the House Ways and Means Committee tomorrow on the Biden administration’s trade policy agenda. Ambassador Tai has given several recent speeches on new trade policy goals, including a recent interview where she was asked about trade rules that may encourage a race-to-the-bottom, and how to flip the paradigm to incentivize a race-to-the-top. Here is her response (emphasis added):

Tai: I hear you. But I think that is the way the global economy needs to evolve. We need to build in incentives. And when we talk about incentives, we’re talking about carrots and sticks. So in some ways, what we need to do is figure out how we need to re-devise those carrots, maybe the shape of the carrots, when they’re applied. We need to look at the sticks in terms of how to enforce. We also want to be taking a look at rules that were devised when our world, frankly, looked and operated differently, and maybe think about where those rules need to be tightened or loosened.

Though the question was not about GSP, it is hard to think of a policy for which the answer is more timely:

  • Congress created the GSP program to promote development nearly 50 years ago, in a very different world, and many of its key features have not changed in decades (if ever)
  • GSP expired on December 31, and Congress is considering possible changes to the program as part of any renewal legislation
  • New eligibility criteria (i.e., “sticks”) that countries must meet to retain GSP have been proposed, but there have been limited discussions on how to “re-devise those carrots, maybe the shape of the carrots” to incentivize compliance with new obligations

So what are the old rules and possible new or re-devised carrots in the GSP context? We examine several decades-old rules that reduce GSP’s effectiveness as a development and policy tool, and suggest a few minor changes that could have a large, positive impact on GSP in the 2020s and beyond.

The elephant in the room: apparel and footwear

Sadly, Members of Congress believe expanding GSP product coverage to include apparel and footwear, which have been statutorily excluded since its creation, is too controversial. But facts are facts and here they are: in 2020, the United States imported $14.8 billion in apparel and footwear from GSP countries that were subject to duties and ineligible for GSP. Excluding oil, that was more than twice as much as the $7 billion in imports subject to duties/ineligible for GSP of all other products combined. Amazingly that understates the impact, as apparel and footwear accounted for $2.4 billion of the $2.8 billion (85%) in tariffs faced for products wholly excluded from the GSP eligibility list.

While the exclusion perhaps made sense in the beginning, we’ve heard domestic production accounted for about half of U.S. apparel and footwear sales in 1974, today over 98% of U.S. clothes and shoes are imported. Given they were excluded from the start, GSP of course had nothing to do with the shift, but program coverage is still hamstrung by “rules that were devised when our world, frankly, looked and operated differently.”

The not-quite-elephant in the room: competitive need limitations (or CNLs)

Most people that follow GSP know of CNLs, but few now that CNLs eliminated as much as 1/3 of all potential GSP benefits in 2020! Here are the numbers: in 2020, imports claiming GSP totaled about $16 billion. By comparison, the value of imports on the nominal GSP list but excluded for an individual country was $8 billion. (This figure does not include imports of products removed as part of a country practice review, since they are covered by separate rules.)

How did we get here? Again, it starts with “rules that were devised when our world, frankly, looked and operated differently.” CNL rules have existed from the beginning, but the last major update was in 1996, at which time:

  • China wasn’t in the WTO yet and imports from China were 55% lower than imports from Japan
  • Malaysia was the top GSP import source
  • 13 current U.S. FTA partners were part of the GSP program
  • 13 current EU members were part of the GSP
  • 21 of the top 32 import sources under GSP are no longer beneficiaries

In short, the GSP program, U.S. trade policy, and global trade patterns were all very different.

How rule designs themselves were problematic: for calendar year 1996, CNLs were set at $75 million or 50% of U.S. imports from the world. The GSP statute set the de minimis level – under which the President may disregard the 50% threshold – at $13 million in imports from the world (not the GSP country). The regular CNL increases by $5 million annually and the de minimis increases by $500,000 annually.

And there’s the catch: by setting a value change instead of a rate, the real annual increase has fallen significantly over time. For example, the CNL change from $75 million in 1996 to $80 million in 1997 was a 6.7% increase, but 25 years later the increase from $195 million in 2020 to $200 million in 2021 is just a 2.6% increase. The value of CNL increases diminishes with each passing year that Congress does not update the 1996 rules.

There also were two premises that turned out to be false: 1) that once products reached CNL levels they could sustain trade without GSP benefits; and 2) if trade did shift, it would go to lesser developed GSP countries and lead therefore promote further growth in GSP countries. Empirical research in 2011 by Dr. Shushanik Hakobyan (then at the University of Virginia, now at the International Monetary Fund) found that the opposite was true:

  • “By the third year of exclusion, imports from affected countries decrease by more than 70 percent relative to the pre-exclusion average…These findings imply that CNLs may not be targeting the “super competitive” exporters, rather these country-product pairs may need the preferential treatment to access the US market.”
  • “By the third year of exclusion, the share of other GSP eligible countries increases by 8 percentage points, whereas that of non-GSP countries expands by 22 percentage points…CNLs seem to provide a competitive edge to non-GSP countries in capturing the US market.”

We’re unaware of any more recent academic research on CNL impacts, but the fact that approximately 95% of products excluded in the past were below the CNL (or de minimis) thresholds in 2020 strongly suggests that post-GSP loss import declines continue – and exports rarely recover.

Congress’ self-correction mechanism also has failed: the GSP statute explicitly allows products that fall below CNL thresholds after losing GSP to be reinstated. The statute does not even appear to require a formal review! Yet in practice, it almost never happens, even when stakeholders attempt to navigate the GSP Annual Review process.

Case in point: wind turbine parts from Brazil

This seems like a good example, since the preceding question to Ambassador Tai’s response asked how to align trade and environmental policy “in a global trade environment where all the incentives are cheaper, faster and not necessarily environmentally friendly.” The wind turbine parts lost GSP on July 1, 2006 based on 2005 imports of $143 million (CNL threshold: $120 million). Imports from Brazil continued to grow in 2006 and 2007, then collapsed just as Dr. Hakobyan’s research showed was the norm:

  • From 2007 to 2008, U.S. imports from Brazil fell from $305 million to $65 million (-79%) and the share of U.S. imports from Brazil fell to 4.8%
  • From 2008 to 2009, U.S. imports from Brazil fell from $64 million to $27 million (-58% more) and the share of U.S. imports from Brazil fell to 2.8%
  • In 2008, China became the top supplier, nearly tripling its import value (from $80 million to $230 million) and more than doubling its import share (from 8% to 17%) since the last full year of GSP benefits for Brazil in 2005
  • China remained the #1 supplier for wind turbine parts every year except one between 2008 and 2019, accounting for 35% of U.S. imports in 2017 before Section 301 tariffs were imposed it 2018

Shouldn’t someone have petitioned for restored GSP based on the clear statutory authority to restore GSP for products that fall below the CNL threshold? GE Energy did. Twice. As part of both the 2009 and 2010 GSP Annual Reviews.

Shouldn’t they have noted the import collapse and how non-GSP countries were benefiting? Of course, and they did in their 2009 petition: “In 2008, Brazil’s exports of HTS 8503.00.95 declined by 78.7% by value while at the same time, the developed nations of China, Japan, the United Kingdom and France increased exports by 49%, 24%, 25% and 37% respectively, by value.” And again in their 2010 petition: “Of the eight other countries supplying more product to the U.S. market none are BDC’s.“

Shouldn’t they have argued that restored GSP would promote President Obama’s goal to increase U.S. wind energy production? Yup, and the did that too!

But both petitions were denied. And no one has petitioned since despite the fact the GSP imports from Brazil have never gotten close to CNL thresholds in subsequent years. Fool me once and what not. On several levels, GE Energy is far from alone:

  • It appears that fewer than 10 product redesignations were granted between 2007 and 2020, of which 4 were never formally requested (e.g., 3 products for Ukraine were restored after Russia invaded Crimea)
  • Stakeholders petitioned to restore GSP for 22 products as part of the 2020 GSP Annual Review, and USTR denied all of them without review (i.e., requests were denied on the front end as opposed to on the merits following hearings, testimony, etc.)
  • We know multiple companies in multiple sectors importing from multiple countries that filed multiple petitions, all rejected, and eventually just gave up.

In summary, 25-year old rules based on flawed premises and longstanding norms not to use the authority granted by statute now eliminate 1/3 of GSP benefits annually.

GSP priorities for the 2020s and beyond

It seems clear that anyone creating the GSP program from scratch in 2021 would design a very different program than the one that exists today, which represents a mix of primarily 1970s and 1990s trade priorities. GSP has eligibility criteria that prohibit countries “dominated or controlled by international communism” or that participate in cartels blocking exports to the United States (e.g., OPEC), but no criteria related the environment.

The key question: beyond “development” in a broad sense, what goals do U.S. policymakers hope to achieve through the GSP program today? Three broad themes that were/are priorities for Congress and both the Trump and Biden administrations:

  1. Using the eligibility criteria as leverage in negotiations with GSP countries (though priority areas differ depending on the policymaker)
  2. Encouraging U.S. companies to source less from China
  3. Providing tariff relief and promoting job growth for constituents and other U.S. companies

Ignoring that elephant (wearing imported clothes and shoes), the current GSP rules regarding CNLs and norms regarding redesignation undermine all of these priorities.

As it relates to leverage in bilateral talks, it doesn’t matter if the goal is better market access or improved labor laws or stricter enforcement of environmental treaties – each and every stakeholder in beneficiary countries that cares about retaining GSP benefits strengthens the the United States’ negotiating hand. By eliminating GSP for both a large share of stakeholders, and in many case the biggest and most successful stakeholders in a country, CNLs weaken the United States’ hand in bilateral negotiations significantly. That makes it considerably less likely that GSP countries will undertake major reforms to maintain GSP, no matter how much policymakers think they should.

As it relates to encouraging U.S. companies to source less from China, the CNLs affect appears self-evident: companies will not shift sourcing from China to GSP countries if rising imports trigger CNLs and lead to higher tariffs. This is not a new argument against CNLs, only the most timely. Back in the mid-2000s, a joint submission from Carnegie Endowment for International Peace, German Marshall Fund, Oxfam America, and the Women’s Edge Coalition recommended eliminating CNLs that create “a glass ceiling.” From a purely development perspective, they argued CNLs have the unintended effect of chilling investment because “Investors appear reluctant to invest in certain sectors in marginal countries because they believe that as soon as their investment succeeds, they will no longer receive the preference.” That China is the country that companies are moving from – as opposed to flocking to – makes the argument no less relevant than when raised 15 years ago.

As it relates to tariff relief for constituents, the affect once again are clear: U.S. companies paid as much as $340 million in tariffs in 2020 for individually excluded products. Cumulatively over the years, it is billions of dollars in tariffs paid – after the sharp declines in imports that typically accompany GSP loss – and generally for products that do not even meet the current statutory definition of “competitive.”

Taking it back to the beginning, it’s hard to see how any of these outcomes – fewer incentives to meet eligibility criteria, extra incentives to go to/stay in China, or higher U.S. tariffs/lower U.S. imports – help workers in developing countries or further GSP’s original development goal.

Re-devising those carrots (or at least the shape of the carrots)

Development NGOs argued for eliminating CNLs altogether, and it would be the cleanest mechanism to eliminate issues that have arisen over the years due to them. But a few minor changes could make a big difference.

Reset the benchmarks: if the regular CNL threshold grew at the same annual rate as it in the beginning, the limit would have been a little more than $350 million in 2020. Maintaining the original CNL-to-de minimis ratio would put that in the $50-$60 million range. Those seems like good ballpark figures, then CNLs should be set to increase by a growth rate as opposed to set value Congress doesn’t have to worry about diminished benefits again. Increasing that threshold also provides a lot more room to run for companies and sectors seeking alternatives to China.

Make reinstated GSP for products below CNLs the norm: what does keeping non-competitive products out of GSP forever achieve? It hurts development, hurts the U.S. companies that import the product, and reduces the Administration’s leverage in bilateral negotiations. Importantly, CNLs do not protect domestic producers. Any U.S. company or industry that feels it is being harmed can request GSP removal of a given product regardless of import levels. Simply changing the GSP statute to say the President “should” restore GSP, instead of “may,” would provide clear congressional intent while retaining Administrations’ discretion in potentially sensitive reviews.

Make reinstated GSP for products above CNLs an option: the current statute prohibits reinstated GSP if imports from a country exceeded the CNLs in the prior year. There is an exemption to the 50% rule for products not made in the United States, but no possible exemption to the value threshold even for products that aren’t made domestically. Changing the GSP statute to say the President “may” restore GSP would give the Administration new tools to promote development and increase negotiating leverage for truly non-sensitive but high value products. If past history is any indication, statutory authority saying products “may” be redesignated is likely to be used sparingly.

Base CNL calculations only on products claiming GSP: products have lost GSP even though the value of imports claiming GSP was well below CNL thresholds. This generally happens when imports that can’t meet the rule of origin increase, but the result is punishing companies with significant local content (as intended) while having no impact on the companies only doing final assembly (that rules of origin are meant to discourage). It may be more likely to happen if companies shift production from China but need imported parts for several years as operations get up and running. The GSP statute’s “super-CNL” bases its calculations on GSP-claiming imports, and Congress could adopt that language for all CNLs.

Such changes are far from radical. All sticks and no carrots, as Congress appears to discussing, likely will continue to whittle away at the benefits provided by GSP product by product, country by country. But there’s a better way and a few small tweaks would greatly improve GSP’s effectiveness as a development tool, a leverage point, and an alternative to China in the years and decades to come.