NAFTA pullout: What would really change? Not much. A chapter by chapter analysis

After last month’s G-20 meetings in Buenos Aires, President Trump renewed his threat to unilaterally pull out of the North American Free Trade Agreement (NAFTA) as a way to jam Congress into approving his NAFTA 2.0. The chatter risks undermining the legislative strategy of Trump’s trade adviser Robert Lighthizer, who wants to get broad bipartisan support and reportedly sees the threat as counterproductive.

But besides the political optics, how much would actually change? The New Democrat Coalition in Congress has stated that pullout would “cause economic chaos.” The Chamber of Commerce said that “ This is a risky gambit when you consider that withdrawal from NAFTA without a successor agreement could jeopardize 1.8 million American jobs.”

The fears may be overblown. As Timothy Meyer and I argued earlier in the fall in Politico, Trump controls NAFTA membership, but Congress controls NAFTA tariffs. This is a crucial check and balance that limits the damage that Trump can do on his own. The worst he could do, we argued, was produce a Zombie NAFTA — where the U.S. was no longer a NAFTA member as a matter of international law (after giving the requisite six months’ notice) but would continue to apply NAFTA duties to Canadian and Mexican exports. This could trigger all sorts of quirky responses, including that the European Union might demand the same free-ride deal that our North American partners get. (A paper earlier this week suggests the EU Parliament is already mulling advising a WTO challenge to NAFTA 2.0.) A dispute of this kind would keep lawyers busy, but it needn’t lead to immediate economic crisis or even noticeable economic change.[UPDATE: I’ve added an addendum below to address the president’s termination/proclamation powers, a point only passingly dealt with in the Politico piece and here.]

So, what changes would a Zombie NAFTA entail? NAFTA is a 22 chapter, 393 page document (not counting the lengthy tariff schedules), so it’s difficult to know with precision what unexpected changes could be triggered. However, at a high level, we can say that Zombie NAFTA could be an outcome that the pact’s critics on the left and right prefer to both NAFTA 1.0 and 2.0, while not necessarily unleashing the doom that the pact’s fans fear. In short, there wouldn’t be any necessary economic fallout of a pullout, and the governance changes might actually make many people happy or wouldn’t make any material difference because of the backstop of WTO rules.

Here is my quick and dirty take, chapter by chapter. Happy to correct anything I’ve gotten wrong or overlooked.

Chapter One: Objectives
. NO CHANGE. These are procedural matters that won’t make major changes in peoples’ lives.
Chapter Two: General Definitions. NO CHANGE. These are procedural matters that won’t make major changes in peoples’ lives.
Chapter Three: National Treatment and Market Access for Goods.
NO CHANGE, UNLESS CONGRESS AND/OR ALLIES LOSE THEIR MINDS. Congress authorized specific changes to U.S. tariffs on Mexican and Canadian goods in Section 201 of NAFTA’s implementing legislation, and there’s nothing there that suggests Trump can unilaterally roll them back. (In fact, Clinton’s Statement of Administrative Action on the bill only references an increase in tariffs if one of the other countries withdraw, not if the US does). I’ve seen no indication that the incoming Congress would proactively enable a regional trade war with our closest allies. Since the US would continue to treat Canada’s and Mexico’s products in a favorable manner, it’s unclear why they would raise their tariffs on US products (though they would be within their international legal rights to do so, at least up to WTO levels — which are higher on average than ours). So a story that Zombie NAFTA would lead to a trade war requires an assumption that Canada or Mexico would fire the first shot, rather than de-escalate and assuage(their preferred method the last two years).

Chapter Four: Rules of Origin. DITTO. Congress authorized these changes in Section 202 of the implementing legislation, and there’s nothing there that suggests Trump can unilaterally roll them back.

Chapter Five: Customs Procedures. DITTO. Sections 205–209 and 611–624 of the implementing legislation make various changes to US law that Trump can’t unilaterally reverse.

Chapter Six: Energy and Basic Petrochemicals. DITTO. According to the Clinton administration’s statement, this chapter “Essentially duplicates” what was already in the 1988 US-Canada trade pact, so no change unless Trump pulls out of that too.

Chapter Seven: Agriculture and Sanitary and Phytosanitary Measures. DITTO. This chapter required numerous small tweaks to US agricultural laws, none of which would be affected by NAFTA pullout. For instance, Mexican meat would still benefit from preferential treatment. All three countries also have significant WTO obligations in this area that will provide a floor on treatment.

Chapter Eight: Emergency Action. DITTO. These provisions “largely track the current U.S. global safeguard law,” according to the Clinton administration, so pull out of NAFTA alone would not lead to noticeable changes.
Chapter Nine: Standards-Related Measures.
NO CHANGE. The WTO agreements signed a year later established deeper requirements than NAFTA. Even if NAFTA goes, Mexico, Canada, and the US could continue to bring cases against each other’s consumer and regulatory standards, as they have in the past.
Chapter Ten: Government Procurement. SOME POSSIBLE CHANGE VIS A VIS MEXICO. In Section 361 of NAFTA’s implementing legislation, Congress authorized Mexico and Canada to be waived from Buy America requirements. Moreover, Canada enjoys similar waivers under WTO laws, so there would be little change in the bilateral procurement relationship. The story is more complicated for Mexico, which is not covered by the WTO rules. Trump could conceivably un-waive Mexican firms from engaging in the procurement process, but it’s unclear to me from the statute what if any say Congress would have over that. Experts on procurement law feel free to weigh in and I’ll correct or expand this.
Chapter Eleven: Investment.
BIGGEST CHANGE. This would be the biggest change under a Zombie NAFTA scenario: it would be tantamount to a withdrawal of US consent to be sued under investor-state dispute settlement. Moreover, it would limit US investors’ ability to sue Canada and Mexico. For critics of the system, then, pullout is preferable to both NAFTA 1.0 and NAFTA 2.0 (which eliminates ISDS rights for most firms except those in favored industries like the Koch Brothers). (It’s worth noting that under longstanding US practice, Congress is not the body that provides the consent to be sued in ISDS — it’s the executive that does so on it’s foreign policy powers. Thus why there’s nothing specific in NAFTA implementing legislation about the investor rules. If Congress somehow disapproved formally of ISDS but the executive branch wanted to keep it going, arbitrators would likely get to decide whether the US’ granting of consent was still valid. In some cases involving Russia, arbitrators have second-guessed whether non-ratification was a bar to bringing a case, and have produced some counterintuitive and controversial results.)

Chapter Twelve: Cross-Border Trade in Services. NO CHANGE. The US has more substantial commitments under WTO rules than NAFTA rules, so NAFTA pullout wouldn’t change anything.

Chapter Thirteen: Telecommunications. CHANGE? I’m not an expert in these rules, so I would hesitate to say much here. However, there are WTO rules on telecom than may provide some protection.

Chapter Fourteen: Financial Services. NO CHANGE. The US has more substantial commitments under WTO rules than NAFTA rules, so NAFTA pullout wouldn’t change anything.

Chapter Fifteen: Competition Policy, Monopolies and State Enterprises. NO CHANGE. These rules aren’t particularly strong relative to what is now the standard pro-monopoly template in NAFTA 2.0 and other recent deals, so Americans are unlikely to notice any change if they go away.

Chapter Sixteen: Temporary Entry for Business Persons. NO CHANGE. Congress controls immigration policy, so Trump can’t change anything unilaterally here.
Chapter Seventeen: Intellectual Property. NO
CHANGE. The US has more substantial commitments under WTO rules than NAFTA rules, so NAFTA pullout wouldn’t change anything.
Chapter Eighteen: Publication, Notification and Administration of Laws.
These are procedural matters that won’t make major changes in peoples’ lives.

Chapter Nineteen: Review and Dispute Settlement in Antidumping/ Countervailing Duty Matters. CHANGE? Congress has a substantial role in setting the contours of anti-dumping policy, so Trump could not make major changes here without Congress. That said, on specific anti-dumping decisions, current statute gives the executive a lot of latitude. As the Peterson Institute has warned, this is where we could see some mischief — none of it productive to building a legislative consensus.

Chapter Twenty: Institutional Arrangements and Dispute Settlement Procedures. BIG CHANGE ON PAPER, LESS IN PRACTICE. Canada and Mexico won’t be able to sue the US (or vice versa), which will probably make NAFTA’s critics on the left and the right happier than they would be with either NAFTA 1.0 or 2.0. That said, they’ll still be able to challenge each other at the WTO, which is where they’ve tended to sue each other anyway.
Chapter Twenty-One: Exceptions.
Because the US won’t be able to be sued, the loss of these defense provisions won’t matter. Critics often derided the strength of these exceptions anyway.

Chapter Twenty-Two: Final Provisions. Trump will be making use of one of these rules to withdraw US participation. After that, they won’t matter.


On Twitter, Joel Trachtman raises the reasonable point: “ Do you believe that 125(c) of the Trade Act of 1974 does not give the President authority to raise tariffs on withdrawal from NAFTA?” For the uninitiated, this is language that is “read into” NAFTA through a chain of cross-references.

  1. Section 101 of NAFTA’s implementing legislation references the Fast Track procedures and criteria of section 1103 of the Omnibus Trade and Competitiveness Act of 1988 (19 U.S.C. 2903).
  2. Section 1105 of the 1988 Act reads in termination provisions of the 1974 Trade Act’s Section 125.
  3. The result: all three statutes have to be considered when looking at the president’s power to proclaim tariff reductions in or out upon various NAFTA termination scenarios.

These proclamation powers seem broad. They have a few elements:

  • Section 125(a) gives congressional sanction to agreements that have end dates. NAFTA has no expiration date, so that’s not relevant.
  • Jumping ahead, Section 125(c) gives the president the right to (within a band) raise tariffs if doing so further its rights or obligations under a trade agreement. This would allow the president to respond to winning a case under NAFTA that requires tariff revision. In a pullout scenario, Trump would neither be furthering the pro-liberalization rights or obligations.
  • Section 125(d) gives the president the power to adjust tariffs when another country does something wrong — again, not applicable here in a Trump unilateralist scenario.
  • Section 125(f) requires the president to consult with the public before taking any action, unless national security demands a quicker response. Whatever Trump’s criticisms of NAFTA, any problems are hardly new — unlike some of the genuine changes and challenges to the steel industry that gave Trump a colorable argument there.
  • Jumping back, Section 125(b) is the broadest of them all, spelling out, “The President may at any time terminate, in whole or in part, any proclamation made under this Act.”

Because Section 201 of NAFTA’s implementing act gives the president to proclaim NAFTA’s tariff concessions into existence, this patchwork of rules might suggest he can reverse the clock.

I think this is wrong, for several reasons.

First, these rules exist in a context. Why would Congress go to the trouble to spell out subsections a, c, and d (that limit proclamation modifications to specific scenarios) if the president could proclaim modifications for any reason at all? Because that would be absurd and deplete a,c, and d of meaning, some other meaning for b must be construed.

Luckily, the legislative history provides some guidance. U.S. trade law has had some version of this language — “The President may at any time terminate, in whole or in part, any proclamation” carried out to lower tariffs under a trade agreement — since 1934. It always followed some statement of congressional purpose.

The original bill’s supporters pointed to the importance of the objectives Congress laid out, namely “expanding foreign markets for the products of the United States (as a means of assisting in the present emergency in restoring the American standard of living in overcoming domestic unemployment and the present economic depression, in increasing the purchasing power of the American public, and in establishing and maintaining a better relationship among various branches of American agriculture, industry, mining and commerce).”

Later acts all had their own version of this limiting language — and all made clear that liberalization and cooperation were the default, protectionism the exception, and inter-branch checks and balances were the order of the day. For instance, the 1962 Act had as its purpose,

“The purposes of this Act are, through trade agreements affording mutual trade benefits — (1) to stimulate the economic growth of the United States and maintain and enlarge foreign markets for the products of United States agriculture, industry, mining, and commerce; (2) to strengthen economic relations with foreign countries through the development of open and nondiscriminatory trading in the free world; and (3) to prevent Communist economic penetration.”

The 1974 Act listed the following:

“The purposes of this chapter are, through trade agreements affording mutual benefits — (1) to foster the economic growth of and full employment in the United States and to strengthen economic relations between the United States and foreign countries through open and nondiscriminatory world trade; (2) to harmonize, reduce, and eliminate barriers to trade on a basis which assures substantially equivalent competitive opportunities for the commerce of the United States; (3) to establish fairness and equity in international trading relations, including reform of the General Agreement on Tariffs and Trade; (4) to provide adequate procedures to safeguard American industry and labor against unfair or injurious import competition, and to assist industries, firm,[1] workers, and communities to adjust to changes in international trade flows; (5) to open up market opportunities for United States commerce in nonmarket economies; and (6) to provide fair and reasonable access to products of less developed countries in the United States market…”

There’s a variety of ways to read these various legislative mandates, but none suggests a Trumpian anti-liberalization outcome in trading relationships that are basically harmonious.

Second, this context would be given weight — especially in a scenario where courts would otherwise be forced to find a delegation of power unconstitutional. In March of 1934, Rep. John McCormack (D-Mass.) argued on the House floor that:

“in the event of this bill’s passing and being submitted to the Supreme Court on the question of its constitutionality, the Supreme Court, as it has in other cases, would, in part, look to the purposes which the passage of the legislation sought to bring about. So this has a legal significance. It is not merely inserting high-sounding words. This language sets forth the definite policy of Congress in the passing of this bill, and it also sets out the direction which the executive branch of the Government should and must take in carrying out the powers necessarily delegated.”

In a 1943 hearing on renewing the authorities, an administration official said “So far as I know, sir, the President can terminate proclamations under this section if and when he chooses, although he is directed earlier in the act to carry out the provisions of the trade agreements pursuant to proclamations.”

In 1974 hearings, legislators further expounded on their justification for Section 125:

“The purpose of [it] is to provide additional flexibility in existing law to enable the President to exercise United States rights and obligations as fully
as foreign countries under the GATT and other international trade agreements, so as to protect United States trading interests In the context of the procedures of GATT or other trade agreements.
Section [125] of this Act provides two basic authorities, however. First, as described above, it provides authority to withdraw or suspend concessions or
other trade obligations and to Increase or impose tariffs or other Import restrictions where the United States has the right to do so under International trade agreements. In other words, it enables the President to give domestic legal effect to the withdrawal, suspension, or termination of trade agreement concessions to any foreign country in the exercise of our International rights and obligations. The authority enables the President to react to actions by other countries and also to Implement the withdrawal of United States concessions under the renegotiation rights of the GATT…
The use of this authority will be limited to matters pertaining to our rights and
obligations under international trade agreements. It is not the intention to use
this authority either as a substitute or extension of other authorities under this or other Acts. It would not be used, for example, as an additional avenue to provide Import relief, or to Impose a surcharge.
Much of the authority contained in section [125]already exists in current law,
In the termination authority contained In section 255 of the Trade Expansion
Act and the implementing authority contained in section 210(a) (2) of that Act. Section 402 of the proposed act is explicit, however, on questions of partial withdrawal of concessions (setting intermediate roles between those presently In existence and those previously In existence) and terminating for a time, that is, suspending obligations or concessions. This explicit authority in section 402 Is necessary to clarify these technical issues which hinder flexible administration of the trade agreements program.”

Finally, in the 1974 Senate report describing the proclamation authorities, legislators wrote:

“Serious problems would be posed if a trade agreement to cut tariffs were terminated. The Committee was concerned both with the possible effects of a sudden return to higher rates of duty and with the possibility that the United States would take no action should other countries terminate their trade agreement obligation to the United States. Clearly, such a situation requires a continuing review by the Congress of future action in the trade agreements program. Thus the Committee has rejected any proposal which would have left the disposition of such terminated rates of duty to the discretion of the President”

So the limitations and the marching orders matter, even if the legislative drafting leaves something to be desired.

Courts have taken a similar minimalist approach. The U.S. Court of Customs in U.S. v. American Bitumuls (1957) determined that the virtually “unqualified power to terminate any proclamation” must be read in light of “the probable intent of Congress.” The court hypothesized that Congress only intended to leave the president with a remedy if another country terminated a trade deal. Notably, the early 1930s-1970s versions of the trade acts included only the equivalents of Section 125(a) and (b), not the additional limitations of c, d, and e. So if courts were arguing for limits even in the 1950s, the case for a limited reading is even stronger since Congress added additional limitations in 1974.

Indeed, the 1934 Act was broad in some ways but narrow in others. Namely, it envisioned time-limited delegation of power and time-limited trade deals. Members of Congress also emphasized the importance of the time limit. Rep. Emanuel Celler (D-N.Y.) said, “I am willing to trust the President of the United States with all the powers embraced in this bill, and they are indeed stupendous; gigantic, unprecedented powers, but I wish that the President shall be limited as to the time he may exercise these powers… If we are to grant the powers during an emergency, let it be for an emergency and not a permanent affair.” So Trump terminating a 25 year-old trade deal would seem to exceed what Congress was envisioning when it first crafted the trade powers.

In short, a Court that wanted to find the proclamation powers constitutional would have a variety of limiting principles at their disposal — all of which a norm-breaking, anti-liberalizing NAFTA pullout would violate. On the other hand, if a Court wanted to strike the authorities altogether, they could do that too.