On a more serious note:

I’m skeptical that the Trump Administration’s new economic pressure campaign, which Bessent has promised will cut off virtually all of Iran’s trade and put the country on a path to “complete global isolation and a subsistence economy,” will meaningfully move the needle against Iran, particularly from a strategic perspective.

In part this is because the Trump Administration has *already* been imposing the most impactful economic measure it can against Iran, the naval blockade of Iranian ports. As I argued in an essay back in December, by the end of last year U.S. sanctions on oil exports by rogue regimes had become comparatively ineffective, at least as long as President Trump remained unwilling to blow up the U.S.-China economic relationship over China’s oil purchases. Despite hundreds of Iran-related designations last year, including against shipping companies and Chinese “teapot” refineries, Iran in 2025 exported between 1.5 million and 2 million barrels per day of crude oil, condensates, and fuel oil, almost all of it to China. That meant that Iran’s oil exports were running at more than 75% of their volumes in 2016 and 2017, when the “JCPOA” nuclear deal was in place and most U.S. sanctions on Iran had been lifted. While U.S. sanctions on the teapot refiners have affected their operations, it is also clear that sanctioned Chinese teapot refiners (which have few direct connections to the U.S.) have survived and continue processing Iranian crude.

Yes, Trump could threaten to sanction large Chinese banks or other internationally integrated Chinese firms unless China stops importing Iranian oil. But as we saw during the trade war last spring, China also has leverage in the U.S.-China relationship, and I don’t see Trump wanting to blow up that relationship (or still-vulnerable U.S. supply chains) over Iran.

The naval blockade has cut off China’s oil trade with Iran in a more diplomatically palatable manner: data analysts assess that Iran’s current oil exports are near zero. Iran does have tens of millions of barrels sitting in floating storage in Asia, a legacy of exports from earlier this year, that provide some ongoing revenue (and sanctions plus diplomacy with the Chinese might somewhat reduce these sales)—but this is a declining asset and there are already reports Iran is making fewer offers for fall sales as it runs out of non-blockaded oil to sell. Given that Iran has historically relied on oil exports for between a quarter and half of its government budget, a sustained loss of oil revenues will put intense fiscal pressure on the government over time.

That is not to say the planned sanctions will have no value. Treasury can play a game of sanctions whack-a-mole with Iran’s cross-border truck trade with Iraq, Pakistan, and Turkey, all of which continue to carry everything from machinery to consumer electronics. A decision by the UAE to follow up on its recent announcement that it will sever economic ties with Iran would deprive Iran of a potential (comparatively small) oil smuggling route and also force Iran to find other supply routes for the 30% of its imports that currently move through the UAE. The UAE has also historically been a major center for Iran’s illicit financial flows. (That said, a Wall Street Journal reporter this week found little evidence that the UAE is, in fact, cutting off its economic ties with Iran, despite the pledge and the new Trump Administration sanctions campaign, illustrating how cutting economic ties often proves harder in practice than in theory).

The U.S. has already seized approximately $1 billion in Iranian crypto. An ongoing sanctions campaign against crypto exchanges, as well as against the exchange houses and other financial institutions in Iraq, Turkey, and the UAE that still do business with Iran, will keep making it harder for Iran to move its money. Trump should lean on both the Pakistani government and Pakistani port operators to make sure that Iran cannot follow through on its plans to use Pakistani ports to circumvent the U.S. blockade.

Collectively, these measures might knock a couple of additional percentage points off Iran’s GDP. But they won’t be as economically significant as the naval blockade Trump has already put in place.

The more important question—from a strategic perspective—is what even a successful economic pressure campaign might actually accomplish.

Judging by World Bank data, past periods of intense sanctions on Iran, including the 2010-2014 period and the 2018-2020 period, have driven Iran’s GDP growth negative. Iran’s GDP sank by more than 3.5% in 2012, for example, after growing strongly in 2010 (and despite the fact that global oil prices that year were quite elevated). Trump’s first-term reimposition of sanctions drove Iran’s GDP growth rate from +3% in 2017 to -3.7% in 2018. Given the IMF is already expecting Iran’s economy to contract by more than 5% this year because of the war, any additional hit from the sanctions will be painful.

But these past numbers also tell another story: Iran’s government has weathered serious sanctions-induced economic declines in the past. I see no immediate reason to expect this year’s recession to be different, especially after Iran demonstrated its brutal willingness to murder thousands of its own people last winter.

Moreover, Iran’s past sanctions-induced GDP declines pale in comparison to the GDP declines that other authoritarian regimes have weathered in recent decades. The Syrian economy fell by more than 25% in each of 2012 and 2013 during Syria’s civil war with a total peak-to-trough economic decline of more than 80%. While Assad did eventually fall, it took more than a decade and a successful military campaign by Syrian rebels in 2024 to topple him. In Venezuela, the Maduro government oversaw a GDP collapse of nearly 90% between 2010 and 2020. The economic situation was so dire that the average Venezuelan lost more than 20 pounds due to food shortages. And yet it took a midnight snatch-and-grab operation by U.S. Special Forces on January 3 of this year to finally depose him.

This history suggests that even if Operation Economic Outcast is successful, Trump will have to keep dealing with Iran’s mullahs—or at least, given the Syria and Venezuela examples, that the mullahs will outlast his remaining 2.5 years in office. He recently appears to have shifted toward a posture of strategic patience in the negotiations—letting the economic pain play out, rather than rushing back into the June ceasefire deal that later collapsed. It is always possible that economic pain combined with somewhat less apparent American eagerness to strike a deal will convince the Iranian government to back away from some of its recent maximalist demands. But few of the Iran experts I have spoken with expect that the Iranian government will fundamentally change which issues it is prepared to compromise on.

Which means Trump, who has never been known for having a particularly long attention span, may well continue to be stuck dealing with various unpalatable options on Iran.

What I’m thinking about:

Visa comes to Syria: One thing Trump does right with his sanctions policy is to provide comprehensive sanctions removal when circumstances change. Many U.S. officials like to salami-slice sanctions lifting, offering complex, limited relief while keeping layers of sanctions in place. I understand the instinct, and limited sanctions relief can make sense as a temporary confidence-building measure. But history and experience make it pretty clear that comprehensive relief tends to be both more workable and more durable over the long term—it provides more certainty to businesses, delivers clearer benefits to the (former) sanctions target, and builds confidence between the U.S. and former adversaries.

Against this background, I respect Trump’s approach to Syria, where, this last week, the U.S. government ended its last remaining major sanctions impacting the country: the “State Sponsor of Terrorism” (SST) designation and the terrorist designation of the al-Nusrah Front, the former al-Qaeda spinoff that evolved into the rebel group that toppled Assad in late 2024 and now runs the country. And we are already seeing results: Shortly after Trump removed the SST designation, Syrian President Ahmed al-Sharaa apparently used a Visa card to pay for a meal in a Damascus restaurant. This is the type of hopeful message that can make the prospect of sanctions relief a useful bargaining chip.

That Canada trade war: When I wrote last week’s Substack, the U.S. and Canada appeared to be on the cusp of resolving their latest trade war. But talks collapsed late last Friday. The U.S. has now moved forward with imposing 50% tariffs on $20 billion of imports from Canada, while Canada has announced retaliatory tariffs on $20 billion of goods it imports from the U.S., slated to come into force Sept. 8. Discussions on social media have turned acrimonious, both sides are threatening still more tariffs, and Trump has now renamed Lake Ontario “Lake America” due to his social media spat with Ontario Premier Doug Ford. It all seems pretty grim. That said, I am going to remain optimistic—at least for now—that the U.S. and Canada will be able to get back to constructive negotiations and reach some kind of deal later this year or early next year that will keep the U.S. trade relationship with its second-largest trading partner from further deteriorating.

Trump (potentially) bans Chinese equipment from the U.S. electrical grid: On August 26 Trump issued an interesting Executive Order authorizing the Energy Department to set up a new regime to restrict the use of foreign products in the U.S. power grid. The E.O. reminds me of Trump’s 2019 “ICTS E.O.” that authorized the Commerce Department to begin restricting the use of Chinese telecommunications kit in U.S. networks. Combined with the FCC’s ongoing campaign against Chinese internet-connected devices, the E.O. signals the Trump Administration’s plans to continue limiting the U.S. use of Chinese products it deems high-risk, despite Trump’s overall desire for a somewhat stable and predictable U.S.-China economic relationship. The new E.O. requires determinations and rulemaking before it comes into effect, and I expect the process may take some time.

The U.S. government’s growing equity portfolio does not get enough attention.

Since Trump’s second inauguration, the government has announced plans to take equity stakes in more than 35 companies, at a cost to the budget north of $27 billion. Thanks in large part to Intel’s strong performance, the market value of these stakes is likely somewhere on the order of $60 billion. On the current trajectory, I generally expect that by January 2029 the U.S. government will own an equity portfolio of perhaps 100–200 stakes with a market value of $100–$200 billion. That would not crack the top ten of global sovereign wealth funds, but it is not to be sneezed at.

My recently graduated Georgetown RA Arian Zand and I wrote a paper about the equity stakes announced through mid-May, and while the number of stakes has grown since, the basic contours of Trump Administration equity stakes were already pretty clear. (Jonathan Hillman at CFR maintains an excellent living tracker of the equity deals.)

First, there is a clear sectoral prioritization. There is a large cluster of deals that are intended to boost critical minerals mining and processing and another large cluster focused on computer chip innovation, including Intel, a semiconductor laser company, and multiple quantum computing firms. The U.S. International Development Finance Corporation (DFC), which Congress explicitly authorized to make minority equity investments in 2018, has made several portfolio investments to serve development purposes. The DFC has also been active in the critical minerals file. Other deals appear to be one-offs, including a single defense-sector deal involving rocket motors and a nuclear energy deal in which the government equity is contingent on certain performance metrics. (NB: My sense is that the defense sector is worried about the government extracting more equity, given the leverage the government has over the sector.)

Second, in many cases government equity comes alongside other forms of government support. Many of the critical minerals deals, for example, also come with promises of loans, grants, offtake agreements, and/or permitting assistance. Having talked to people involved in a couple of the deals, I suspect that in at least some cases it is the promise of these other forms of support that makes the equity attractive, rather than the equity being attractive on its own merits (at least when compared to private investor alternatives).

There is also the question of what “soft” support the government might provide once it has equity, even if the support isn’t formally part of the equity deal. The Commerce Department, for example, has reportedly gone to bat for Intel, helping it find customers now that Commerce owns tens of billions of dollars of Intel stock. My sense is that mining companies may think that government ownership will speed permitting even if permitting commitments aren’t formally part of a deal. And then we saw the U.S. government help USA Rare Earth, a rare-earths mining company, acquire a Brazilian mining asset a couple of months after the government bought into the company, with DFC support for USA Rare Earth’s acquisition of the asset. That was presumably good for USA Rare Earth, and it also shows how the U.S. can use its new part-SOEs to acquire overseas assets that the government itself views as strategically interesting.

Third, the structure of the deals generally appears to be bespoke. (It is almost as if a bunch of private-equity guys went into government and just started treating the government as a giant PE shop, negotiating bespoke terms that make sense for a given deal…). Take corporate governance and the government’s shareholder rights. At one end of the spectrum is the government’s “Class G” golden share in U.S. Steel, taken as part of granting Japan’s Nippon Steel CFIUS approval to buy U.S. Steel. It provides the government a number of specific governance rights but has no economic value. At the other end of the spectrum is the government’s 9%+ stake in Intel, in which the government has obliged itself to vote however Intel’s board recommends. (This is a pretty sweet deal for Intel’s board.) Some, but not all, of the deals have either lockup periods before the government can sell its shares or limits on how it can dispose of them.

Fourth, most of the deals are built on a tenuous legal foundation. Equity stakes currently appear to be held by four agencies—the Department of War (Defense), Commerce, Energy, and the DFC—and each agency relies on a different statutory basis. The DFC is the only one of these where Congress clearly granted the authority to take equity stakes (and subjected those stakes to limits, such as requiring that they remain minority stakes). Commerce has simply expansively interpreted the CHIPS Act to authorize equity on the ground that the Act authorizes the Department to “enter into agreements, including contracts, grants and cooperative agreements, and other transactions as may be necessary and on such terms as the Secretary considers appropriate.” The Defense Department is reinterpreting Title III of the Defense Production Act (DPA) to authorize equity and, it appears, also relying on certain industrial base authorities. An Intel shareholder has challenged the legality of the Intel deal in court, and it will be interesting to see how that lawsuit plays out. Congress, meanwhile, may explicitly grant an equity authority as part of the DPA reauthorization now winding its way through Congress, while subjecting DPA equity to certain guardrails. Energy appears to be relying on its expansive “other transaction authority.” (Energy also so far appears only to have taken warrants, not direct equity.)

Turning to a normative perspective, I see a place for U.S. government equity investments, for example in sectors that are strategically important but commercially low-return, or to help seed new technologies that have national security utility but where the commercial purpose is unclear. (As an aside, In-Q-Tel, the CIA-backed VC fund established in the 1990s, is structured as a government-funded but independent nonprofit, so the government doesn’t actually own shares in those companies.)

The history of U.S. government equity investments is more complicated than commonly appreciated. The U.S. has on multiple occasions taken equity stakes in companies as part of government bailouts, such as the 2008–2009 financial and auto industry bailouts, and in the 1960s and 1970s bought up railroads to stabilize the failing industry (Amtrak is a legacy of this). The U.S. had the right to appoint board members to the Union Pacific when it was building the transcontinental railroad, though it did not own an economic stake in the firm. Several U.S. states, including Alaska and New Mexico, have sovereign wealth funds, mostly funded by natural resource revenues—not to mention the publicly owned utilities that exist in many states. North Dakota also owns one of the largest U.S. flour mills, a legacy of the state’s farmers not wanting to be dependent on mills further east. During the 1930s the Reconstruction Finance Corporation (RFC) took stakes in thousands of failing banks.

I’m quite skeptical, however, about many aspects of the Trump Administration’s approach.

One concern is favoritism: Is the government going to favor the companies it owns stakes in, in ways that disadvantage private competitors? Deals that are linked to permitting, for example, raise the question of whether purely private competitors will also be able to get permits. And if the private competitors can’t get permits, or at least not on similar terms, without giving the government equity, will private money continue to flow into the sector or will we see a steady expansion of government ownership and a withering of private investment?

I’m also concerned about politicization and discipline. Will the government start leaning on companies it owns stakes in to serve political objectives beyond the objectives that led the government to take a stake? Think pressure to site a new facility in a politically important district, even if the location makes little commercial sense, or to do business with a politically influential constituency. And if a company the government owns a stake in starts failing in the face of better private-sector competitors, will the government have the discipline to let it fail, or will the government start bailing out the firms it owns part of in order to avoid political consequences? (A bailout need not be a transparent transfer of cash—the government could, for example, steer an overpriced Pentagon contract to one of its companies.)

Then there is the concern that has created the most political controversy: potential corruption. I have no basis to assess whether any of the current equity deals have in fact been improper, but there is already political controversy over the fact that a handful of the deals appear to involve companies linked to relatives of Administration officials.

Companies considering government equity investments need to be thoughtful about what they are signing up for. The government has already shown that when it has governance rights, it is prepared to use them: the U.S. government last fall stopped U.S. Steel from halting steel production at a plant in Illinois. Even when the government has given up formal governance rights, the practical reality is that it will hold substantial sway over its investees. Intel’s CEO appeared to feel it necessary to call the Commerce Department before finalizing a recent private-sector share issuance, even though in theory the government will just support Intel’s board recommendations. The next president could, for example, put quite a bit of pressure on companies in which the government holds equity stakes over issues such as executive compensation and hiring preferences.

In an ideal world, Congress would step in to clean up the mess, for example, by establishing a formal U.S. strategic investment fund to make investments important to national and economic security while managing them in a way that reduces the risks of favoritism and politicization. Or, at a minimum, it would amend the statutes the government is relying on for its equity binge to set up guardrails. Barring that, the government could at least start publishing details of the deals and/or deal documents, something that it does not currently do beyond press releases. (For deals involving public companies, SEC filings often reveal some details, but almost nothing is known about deals involving privately held companies.) Companies, too, could voluntarily disclose more, though I recognize that many will be reluctant to do so.

My hunch is that some form of government equity investing will be here to stay: it is something that a number of experts and policymakers on both sides of the aisle have advocated for in recent years, and the government will find equity, alongside other kinds of support, a useful tool in some circumstances. This is not a terribly high-confidence bet: I could also see scenarios in which Congress or the next president decides to get out of the equity business and sell off stakes, much as Congress in the late 1940s required the government to sell off the stakes in banks and other companies that the RFC had acquired during the Great Depression. But if government equity is here to stay, I hope that we can bring a more “regular order” and less potentially politicized approach to it.

What I am thinking about:

How can Trump actually ramp up economic warfare on Iran? Trump is now promising to intensify economic warfare on Iran, a decision that seems driven largely by frustration over his inability to score a clear military victory over the country. Iran has been sanctioned for decades, and fairly intensively since 2010—so what more can be done?

Trump has already implemented the most important part of an economic warfare strategy, which is the ongoing naval blockade of Iran’s ports. That action appears to have reduced Iran’s oil exports from about 2 million barrels per day pre-war (almost all China-bound) to under 500,000, and should also be disrupting Iran’s non-oil trade. Helping the UAE enforce its recently announced cutoff of trade with Iran should also be useful, given the UAE’s role as a transshipment hub and financial center for Iran. Addressing Iran’s cross-border trade with Iraq and Turkey will be an ongoing game of sanctions whack-a-mole, and Trump will need to engage with Pakistan to cut off Iran’s newly proposed plans to use Pakistani ports and overland shipments to circumvent the blockade. Still, even if everything goes perfectly, Iran has weathered staggering declines in GDP in the past without collapsing—more than 30% between 2012 and 2015, and nearly 50% between 2017 and 2020. It will be interesting to see how Iran’s economic pain tolerance compares to Trump’s attention span.

I guess I’ll have to wait for litigation over “Section 338” tariffs: At the time of writing, it *looks like* Washington and Ottawa have reached a deal over the “Section 338” tariffs that Trump threatened last month purportedly over Canada’s discrimination against U.S. autos, booze, and dairy products, though the deal has not been finalized as of this morning. If finalized, the deal reportedly will include Canadian concessions over those three items, as well as U.S. cuts in steel and aluminum tariffs on Canada, and generally strikes me as good economics and good politics all around (certainly compared to the alternative of further straining the U.S.-Canada relationship). It also helps set up more constructive U.S.-Canadian negotiations over the USMCA trade deal.

The downside, for us legal folks, is that this likely means we won’t get a lawsuit to decide what Section 338 of the Smoot-Hawley Tariff Act actually means, at least until Trump decides to deploy it against someone else.

Beef: It’s what’s for (expensive) dinner: After going back and forth on whether to slash beef import tariffs in May, Trump finally announced this morning that he would suspend tariffs that affect ground beef for 90 days. Current tariffs on beef imports run north of 26%, so eliminating those should partially deliver on Trump’s statement that imported beef will be sold at 25% below current market prices. From what I have read this morning, it isn’t clear whether there is some other element to that commitment, e.g., particular distributors who have made specific commitments.

Second, a promise. My Substack really will mostly be about national security and economics. But as someone who uses AI essentially every day, I can’t help but wade into the debate that has been raging online and in the legacy press about the use of AI in writing.

My first big, and rather contrarian, take on AI is that it can make people *more* creative rather than less.

Take my own recreational use of AI. I’ve always enjoyed political cartoons, and long had ideas in my head about them. But I am a terrible draftsman. (Truly terrible. My 9-year-old draws better than I do). And so I have never been able to create a political cartoon, not because I lack the idea for one, but because I lack the ability to execute. Today, Gemini makes it easy for me to generate and edit expressive cartoons that were previously stuck inside my head. (No, you’re not getting these. I share these with a few friends on text chains. I don’t post them online).

This ability to use AI to help with creative expression is something that applies to most people. I believe that people are fundamentally creative. A person might have an idea for a book—plot, character, scene—but just isn’t good at drafting prose. AI can help express those ideas. Or take video. Sure, much AI video content that actually exists is thoughtless slop. But we are going to see artists who previously couldn’t render their actually-creative ideas into video find that now they can.

The issue of whether AI makes us humans more creative or replaces our creativity depends on how we humans use it. If we fall into the habit of simply telling AI to come up with ideas whole cloth then execute them, like much of the slop we see on the internet, sure, we might turn into those caricatures of humans floating in armchairs depicted in the Pixar movie Wall-E. But we can instead choose to use AI to help us be creative and retain a steady hand as idea creators and idea iterators, while relying on tools to cut out the AI slop we see online that is, well, just slop.

I recognize that for some—maybe many—creators, drafting and editing are themselves part of the creative process, and that creativity is not simply about idea generation. The nature of creativity, however, is a complex philosophical topic, and while drafting and editing are crucial for some creators, for many they can be a barrier to executing the creative process.

We humans absolutely need to keep thinking. But the issue isn’t the AI, the issue is us.

This brings me to one of my current pet peeves about the use of AI, which is disclosure policies. The typical “use of AI” disclosure policy makes no intellectual sense. Earlier this summer, I published an essay on the future of American economic power for the Carnegie Endowment for International Peace. I used AI to copy-edit an initial draft and to help with creating some of the charts, and Carnegie required me to disclose both uses. I am fully supportive of that. But as I understand Carnegie’s AI policy, had I used AI to *come up with the underlying thesis of the essay* but then drafted and edited the paper myself, I would *not* have had to disclose that I used AI in creating the paper.

This is absurd: I have to disclose the fact that I used AI to make minor textual edits to ideas I came up with (as friends know, I am truly terrible with typos), but if I served as a human amanuensis for the machine, scribbling down the ideas that AI came up with, no disclosure would be required. AI disclosure policies should have intellectual coherence. My hunch is that AI disclosure policies that focus on the origins of text rather than the origins of the underlying ideas are in reality less about their stated purpose of intellectual accountability and more about basic IP protection—if AI writes the words, in general they can’t be copyrighted.

As an aside, I very much appreciate how the two schools my children attend are integrating AI into their curriculum. My daughter’s (public) school has AI give her the type of detailed paragraph-by-paragraph feedback on writing assignments that no teacher juggling 20 fourth graders is going to give. My son last year, in a classic rite of American middle school, read Lord of the Flies. One of his assignments was to get AI to analyze a part of the book, and then he had to critique what AI got right and wrong in its analysis. Seeing these uses of AI in school makes me somewhat more optimistic that AI will strengthen education, at least if we can figure out how to convince students to stop using it simply to cheat.

Turning to the professional front, I expect AI will put me out of a job within a few years. Or at least, that AI will run me out of the work that clients currently pay me to do. Earlier this summer a client of mine told me that it had a handful of somewhat similar issues that it wanted legal advice on, and hired several human lawyers (including me) to offer views on some of them, but also “hired” AI to do the work on one of the issues. The client reported afterwards that it was happy with the “artisanal” legal analysis, but that AI did a pretty good job too. (AI, I am sure, was much, much, much cheaper).

Legal work strikes me as vulnerable to AI, which can already draft sanctions reps and warranties, license applications, and compliance manuals, and analyze documents for due diligence and the like, reasonably well. I think I can still do a better job, and AI gets it wrong often enough that any AI legal work needs a close review. (Looking at you, Sullivan & Cromwell). But we are still in the early days of AI. Given the pace of improvement, I have to be realistic that within a few years AI will be able to do 90% of what I do.

Now, maybe as AI takes over more and more actual legal work, I can spend more time just schmoozing with clients over lunch, or adapt and use AI to offer more products. For example, AI already lets me offer my clients some policy-oriented data analysis of how tariff changes are actually impacting trade flows, something I was not previously well positioned to offer. (Look for some public takes on this in a future Substack).

But the clients might well find that their $20/month Claude subscription is good enough and that they just need me much less. In that world, we lawyers will have to figure out something else to do to pay the bills. Or maybe Aaron Bastani will prove prescient: AI will let us all live in a society of fully automated luxury communism. I’ll be able to spend my days at a house in the mountains, enjoying long hikes and time with the kids, while keeping my creative instincts sharp by occasionally having AI help me create political cartoons.

What I’m thinking about:

What will new U.S. trade rules of origin actually look like? The Trump Administration seems to be getting more seriousabout actually establishing new, anti-China rules of origin, judging from White House trade advisor Peter Navarro’s paper out this week assessing alleged tariff circumvention and the growing number of reports that USTR is pressing on the issue in the USMCA renegotiation. The basic issue is this: under current U.S. trade law, in many cases a product made in a third country, like Mexico, will count as “Mexican” for U.S. tariff purposes (which may be 0% tariffs) as long as it is assembled in Mexico, even if many of the parts come from China (which might face much higher tariffs). This both creates a competitive advantage for third-country manufacturers that use Chinese parts relative to their American competitors, who typically have to pay tariffs on the cheap Chinese parts, and keeps the U.S. indirectly dependent on Chinese supply chains.

Trump’s “ART” trade deals, the deals he has struck with various U.S. trade partners, have placeholders for rules of origin: provisions that allow the U.S. in the future to, for example, say that a product will only count as “Malaysian” if the product is both made in Malaysia and has below X% Chinese content. (Or, conversely, if the product has at least Y% Malaysian content). My sense is that the Administration wants to work out how new rules of origin could actually work in the USMCA context, and then try to deploy those standards elsewhere. But what the rules look like—will they turn on percentages of content, the origin of “critical components,” or something else—remains TBD.

Whither U.S.-E.U. trade? There continues to be lots of rhetorical tension in the U.S.-E.U. trade relationship, with ongoing U.S. criticism of European tech regulation, complaints about slow implementation of the trade framework agreement agreed to last year, and the like. On the other hand, the U.S. continues to take an approach to actually implementing its Trump trade commitments to the E.U. in a way that benefits the E.U.—for example, the U.S.’s 10% “Section 301” tariff on the E.U. includes underlying MFN tariffs, whereas most (but not all) other countries’ 301 tariffs stack the 10% (or 12.5%) on top of the underlying MFN tariff, which to me says that maybe things continue to go ok beneath the public rhetoric. The E.U. is also taking more and more measures against its own trade with China, which would seem to create some opportunities for cooperation.

Will the Section 338 tariffs on Canada come into force next week? A few weeks ago Trump dusted off Section 338 of the Smoot-Hawley Tariff Act to threaten 50% tariffs on $20 billion of U.S. imports from Canada, which are slated to come into force August 19. (Section 338 requires 30 days’ notice before tariffs can come into effect). The Trump Administration’s goal is pretty clearly to put pressure on Canada to make more USMCA-related concessions, and Canada is reportedly looking for offramps, but it isn’t clear to me if there will be a zone of agreement before the tariffs actually come into effect. If the tariffs do come into effect, I am looking forward to the fascinating (for trade lawyers) legal challenges that will certainly follow.

It is time to unpack what is going on here.

For more than a decade, a chorus of bipartisan policymakers in Washington has been raising concerns about U.S. reliance on Chinese tech products. Concern initially centered on the risk that China could use Chinese-made telecommunications equipment from companies such as Huawei and ZTE to engage in espionage, but in recent years policymakers have grown wary of a much broader range of products: The Biden Administration, for example, in early 2025 announced regulations on Chinese internet connected cars, which the Trump Administration recently enforced to ban EV company Polestar. And of course there was the well-publicized saga over TikTok, which eventually resulted in TikTok’s Chinese parent sort of divesting itself of its U.S. operations.

I wrote a paper a few years back describing the concerns, and they generally fall into three categories: That China can use technology to conduct espionage against Americans (that humanoid robot can see everything in your house); that it can engage in covert propaganda campaigns (this was a particular concern with TikTok); and that, in a time of serious conflict, it could take control of a device to cause harm in the U.S. (crashing your car into a wall, or, even more ominously, Terminator robots).

Ok, clear enough on the policy concerns. But why is the Federal Communications Commission involved in these bans? Doesn’t it mostly sell wireless spectrum and deal with broadcast media consolidation? And isn’t it run by a guy with controversial views about media oversight?

The FCC derives its authority to ban high-risk products from the Secure and Trusted Communications Networks Act of 2019, which directed the FCC to establish and maintain a list, the “Covered List,” of any “communications equipment or service” that, pursuant to a process laid out in the Act, “poses an unacceptable risk to the national security of the United States…”

The Covered List initially had no real consequences, but the Secure Equipment Act of 2021 required the FCC to begin denying FCC certifications to companies and products placed on the Covered List. At the time, Congress mostly understood this as a tool to ban U.S. telecom carriers from using network equipment manufactured by Huawei and ZTE, and several members of Congress also used the authority to begin targeting remote access security cameras made by Chinese companies. But the FCC’s regulatory authority actually sweeps quite broadly: Under longstanding U.S. law, the FCC has to certify essentially any device that connects to the internet, a requirement that initially existed largely to ensure that the device does not interfere with U.S. communications networks by, for example, emitting too much electromagnetic radiation. (If you look closely at the box that phone or computer came in, or pretty much any other device you own that uses WiFi or cellular networks, you’ll see a small FCC mark that indicates the device has a license).

Companies typically obtain FCC certifications for electronic devices via a self-certification process, and tens if not hundreds of thousands of devices currently have them. But if a company cannot obtain an FCC certification for its products, the company’s products effectively cannot connect to the internet in the U.S. And for a tech product, that pretty much leaves them DOA in the U.S. market.

The Biden Administration deployed this authority comparatively narrowly: In addition to targeting Huawei, ZTE, and the Chinese security cameras targeted by Congress, the Biden Administration used its authority to restrict several Chinese companies from offering telecoms service in the U.S. and, in 2024, to ban Kaspersky Lab’s cybersecurity software.

The Trump Administration, however, has put the Covered List into overdrive. It has applied bans not just to products made by specific named companies (e.g., Hikvision), but to entire categories of products, e.g., drones. And the Trump FCC has applied the bans on a global basis—to all foreign-made products—rather than solely to Chinese products, while creating a “white list” approach for U.S. companies that manufacture abroad and allied country companies to get carveouts from the ban.

Part of the Biden Administration’s limited use of the FCC to ban products reflected the fact that Biden preferred another tool, the Commerce Department’s “ICTS Rule,” which Trump established during his first term to give the Commerce Department an authority to ban high-risk internet-connected devices. (This was at a time when the Covered List was still being developed). The Biden Commerce Department staffed up an office to implement the ICTS rule, used it for the Chinese EV car ban, and was beginning to investigate restrictions on a range of other technologies. Under Trump, however, the ICTS office seems to have faded away with the FCC stepping into the breach. I’m not sure if that simply reflects agency leadership personalities, legal considerations, or other factors. Chinese dronemaker DJI has also filed a so-far unsuccessful lawsuit against the FCC’s drone ban — a suit that, if it proves successful, could shift energy back to Commerce. But the shift to the FCC is clear.

Of course, whether the Commerce Department or FCC implements a ban, a ban on Chinese products without concurrent efforts to build up U.S. and/or allied production risks leaving Americans without access to important devices and technologies. The Biden Administration, for example, imposed its ban on Chinese EVs after Congress enacted an array of subsidies for the U.S. EV industry (many of which have since been repealed) to make the U.S. industry more competitive. In 2024, when the Biden Administration used an unrelated but conceptually similar Coast Guard authority to restrict new foreign-made cargo cranes at U.S. ports (citing the risk that the cameras could surveil U.S. imports and exports), Biden also found some infrastructure money to encourage a Japanese company to start building cargo cranes in the U.S.

The Trump Administration understands this reality and has also tried to boost U.S. production of products in parallel with the FCC bans. An FCC factsheet released last month argued that restrictions on drone imports had spurred more than $4 billion in U.S. drone investment. The Pentagon has stepped up its own investments in the U.S. drone supply chain. Given China’s current lead on at least the mechanical side of robotics, the U.S. will almost certainly also need to support U.S. robotics if we want the ban to help U.S. industry rather than just seeing the U.S. lose access to advanced robots. I also think the Trump Administration should be more amenable to imports of robots from allied countries, like Japan and Korea, which are innovators in the field, rather than focusing so strongly on solely U.S. production, though I realize that is not really the ethos of a number of Trump officials.

I’m generally supportive of restrictions on Chinese tech devices as both a security measure and as an element of U.S. industrial policy. That said, I am concerned that from a security focus, emphasizing bans on Chinese devices without also working on cybersecurity more generally simply encourages the Chinese to find other vectors into the U.S. tech ecosystem. Indeed, my take is that this is essentially what happened with respect to U.S. telecommunications networks. The U.S. succeeded in banning Huawei and ZTE equipment, but the Chinese responded with sophisticated hacks on western kit and have gained remarkable levels of access to U.S. communications networks, to the point that the federal government, which long opposed encrypted communications because it likes to spy on us, started recommending we encrypt our calls and messages. The ban was a tactical success but if the goal was to secure U.S. communications against espionage, we suffered a strategic failure. The lesson? It is important to strengthen the entire defensive infrastructure.

Going forward the FCC, and the Trump Administration more broadly, will need to figure out how to deal with Chinese retaliation against FCC measures. China did not meaningfully retaliate earlier this year when the FCC initially restricted drones and routers, but last week Beijing hit back with sanctions on several U.S. companies and new export control measures, as well as investigation on U.S. printers and office equipment. This retaliation has much less practical impact than did China’s sweeping rare earths export controls last year, but it does signal China’s displeasure with the FCC actions and potentially creates a risk of more extensive retaliation if FCC measures continue to expand.

One final point: The FCC’s current focus on national security risks may seem new, but is consistent with the agency’s history. When Congress set up the FCC in 1934, one of the reasons was “for the purpose of the national defense.” The FCC for decades protected U.S. telecoms networks by restricting foreign ownership (Rupert Murdoch is an American citizen today because he had to get naturalized to buy Fox in the 1980s) and reviewed applications to land telecommunications cables and interconnect with foreign telecoms networks. What it is doing today is new, but not without precedent.

What is on my mind:

A new week, a new Section 232 tariff: Trump yesterday announced new tariffs on imports of polysilicon and derivative products (i.e., solar cells). The most interesting part of the new program is that it sets a “minimum import price” (MIP) for imports of certain products, rather than simply applying a tariff of a certain percentage on the value of imports. With the Administration talking about price floors for critical minerals, it strikes me that this might not be the last MIP….

A new ART with Jordan: Sure, if you Google “Jordan ART” what you get is a lot of artistic images of basketball great Michael Jordan. But my interest in the phrase is less the debate over who is basketball’s GOAT than the “agreement on reciprocal trade” that Jordan and the Trump Administration agreed to last month, which contains terms broadly similar to the other ARTs that Trump has finalized.

Seven more government equity investments: The Commerce Department’s equity binge seems to be undeterred by the ongoing lawsuit over the legality of the Department’s investment in Intel last year, with the Department recently announcing plans to take equity stakes in 7 semiconductor related companies in order to support U.S. semiconductor R&D.

Lord, I was born a tariff man
Tryin’ to tax the foreigners and doin’ the best I can
And if the courts strike ’em down
I hope you’ll understand
That I’ll remain a tariff man

This week President Trump moved to the second phase of the “fallback tariffs” his Administration has been planning since February 20, when the Supreme Court ruled in Learning Resources v. Trump that the tariffs Trump imposed last year under IEEPA, a 1970s emergency powers statute, were unlawful. The U.S. Trade Representative announced late yesterday afternoon that it would, as expected, impose 10% or 12.5% tariffs on a total of 60 economies (59 countries plus the E.U. as an entity) purportedly over their failure to ban or adequately enforce a ban on their own imports of products made by forced labor. These tariffs, under Section 301 of the Trade Act of 1974, replace the “Section 122” tariffs that Trump put in place immediately after his Supreme Court loss and that were limited by statute to 150 days.

A couple of thoughts on the new Section 301 tariffs:

• The final “forced labor” 301 tariffs are similar to the original proposal that USTR published in early June, and not too dissimilar from the Section 122 rates that were in place from February through yesterday.

• There were several changes from the proposal. Notably, a handful of countries, including India, Sri Lanka, and Honduras, got shifted from a proposed 12.5% rate to a 10% rate, which USTR attributed to those countries’ progress on forced labor issues since the original proposal was published.

• For the E.U., Taiwan, Japan, Korea, and Switzerland, the new 301 rates generally include underlying MFN tariff rates, whereas the 122 tariffs were on top of the underlying MFN rates. For other countries, the Section 301 tariffs stack on top of underlying MFN rates. This means the E.U. and Taiwan get a bit of a break. (Japan, Korea, and Switzerland’s break is offset by their rate moving from a 10% Section 122 tariff to a 12.5% Section 301 tariff.)

• The list of excluded products grew a bit from the proposal. USTR added exclusions for some agricultural products, antiques and collectibles, certain metal-related products, and certain semiconductor manufacturing equipment, among others. Canada and Mexico continue to benefit from 0% tariffs for many “USMCA” products, though Canada now faces other levies discussed below.

Importantly, the overall U.S. tariff rate remains below what it was on February 19, before the Supreme Court rendered its decision in Learning Resources. Although the Trump Administration has indicated that it generally plans to restore the “IEEPA Rates,” including through a forthcoming separate Section 301 action on foreign “industrial overcapacity,” the U.S. tariff rate is going to remain a bit lower for a bit longer. The Financial Times reports that the effective U.S. tariff rate in light of the action is holding steady at 10.8% compared to 15.8% before the ruling (albeit still well above the 3% or so it was before Trump’s second inauguration).

New, separate 25% Section 301 tariffs on Brazil also came into force this week, after being announced last week. These tariffs are ostensibly over a whole range of Brazilian trade and economic malfeasance, but they basically boil down to the fact that Trump doesn’t like Brazilian President Luiz Inácio Lula da Silva (“Lula”), and that Lula and the Brazilian courts have gone after both former Brazilian President Jair Bolsonaro and Elon Musk’s social media platform. The Brazil 301 tariffs, however, exempt a large majority of U.S. imports from Brazil, and so their practical impact is smaller than the headline rate suggests. (These tariffs will be combined with the 12.5% rate Brazil faces under the “forced labor” Section 301, so some products from Brazil will face a 37.5% rate).

More ominous than the 301 tariffs were Trump’s new, legally and conceptually distinct “Section 338” tariffs on Canada, which he announced on July 20. These tariffs get their name from the fact that Congress enacted the provision as Section 338 of the Tariff Act of 1930, better known as the Smoot-Hawley Tariff Act. (Here’s to you, Ben Stein). The actual macroeconomic impact of these tariffs is not monstrous—yes, there is a 50% headline rate, which stings, but they cover only about 5% of U.S. imports from Canada. In my view the major goal of the tariffs, aside from Trump’s ongoing intense dislike of the country, is to put pressure on Canada to become more pliant in the ongoing USMCA negotiations, which the Trump Administration clearly thinks are going better with Mexico than with Canada. But boy, have the 338 tariffs triggered fierce political backlash across Canada, and I do worry they might simply force Canada to dig in.

The reason I say the Section 338 tariffs are ominous is not so much their economic impact as that Trump has now shown he is willing to deploy this never-before-used provision of Smoot-Hawley. And, having shown he will use it on Canada, we have to start wondering who is next. Section 338, if upheld by the courts, could become a quite flexible new tariff weapon for Trump to wield in his various economic and geopolitical spats.

Of course, all of these tariffs are going to be challenged in court, consistent with Trump’s ongoing commitment to ensuring that trade lawyers, at least, win from his trade policies. (Thank you Trump!)

Section 338 might have been implicitly repealed by Congress over the years. It might require fact-finding by the International Trade Commission. The rate might require a better justification than the one Trump offered this past week. We will see. It will certainly make for interesting litigation, unless, of course, Trump and Canada settle in the coming weeks, before the tariffs take effect.

As a statute, Section 301 rests on a much firmer legal basis, and it has been used to threaten or impose tariffs something like 130 times since the 1970s. It was, for example, the basis for Trump’s Term 1 tariffs on China.

That said, Trump is now fundamentally reinterpreting the statute. Congress designed Section 301 as a tool Presidents could use to generate leverage on discrete trade issues with individual trading partners; Trump is turning it into something else entirely: a permanent tariff regime covering the world. And he is doing so on the basis of an investigation into purported forced labor issues that, simply as an investigatory document, is far weaker than most past Section 301 investigations—and an investigation that, in my view, does not meet the statute’s requirements.

We’ll see if the courts stand for it.

What’s on my mind?

Candidly, not a lot aside from Tariff Week 2026.

Banning Chinese AI models—can he do that? The Trump Administration substantially increased pressure, at least in terms of public rhetoric, on Chinese AI labs this week, with White House official Michael Kratsios, State Department official Jacob Helberg, and Treasury Secretary Scott Bessent putting out coordinated messages accusing Chinese AI company Kimi and others of stealing U.S. IP for their new models. That inspired me to write a post on LinkedIn looking at what the government’s options could be, if in fact they do want to ban or restrict U.S. use of Chinese AI models. We will see if the Trump Administration actually does anything here, which could complicate Trump’s plans to see Chinese President Xi Jinping in September.

The FCC continues its emergence as a national security agency: The FCC has been on fire over the last year in the national security space, announcing restrictions on foreign drones, foreign Wi-Fi routers, submarine cables, and, last week, certain high-risk components embedded in other devices. It has been fascinating to see the agency step into this role.

I feel something similar about the lurch in this country towards the worst of Chinese state capitalism.

In 2021, the Chinese government launched a sweeping crackdown on its big tech platforms, bringing a range of vague cases against Chinese technology companies in a bid to generally assert state control over the sector. Sure, some of the individual cases might have had merit. But the point wasn’t really the cases; the point was the control.

This approach has now jumped the Pacific and landed in Washington. When the Commerce Department wanted to transform Intel’s CHIPS Act grant into an equity investment, it made Intel an offer it couldn’t refuse. According to board minutes released as part of a lawsuit challenging the deal, a senior Commerce Department official “‘sent a clear message’ in a call to Intel CFO Zinsner that Intel “should not make the ‘mistake of thinking that this is a negotiation[]’ and ‘the President is expecting to announce a deal on these terms on Friday…[the Commerce official] acknowledged that the Board could choose not to accept the offer, but noted that rejection will have consequences that [Commerce] cannot predict and that [the B]oard needs to consider benefits of having a friend in this administration and what the administration can do to help Intel (including with customers).”

Or take the recent on-again, off-again export controls against AI company Anthropic’s most advanced models. There may have been genuine security risks from Mythos and Fable—I’m not a tech expert. (Though the fact that a Chinese company is releasing a fast-follower open-source model suggests that maybe U.S. and western companies would have been better off having quicker access to the Anthropic models to patch our own cyber vulnerabilities). But the Trump Administration’s action was clearly in large part about bringing Anthropic to heel and making sure that the AI industry is pliant. And it worked. Sam Altman’s OpenAI is discussing giving 5% of itself to the government as a payoff to avoid a similar fate or other regulation. (As the Commerce Department told Intel, “consider the benefits of having a friend in the administration.”)

Tariffs, too, have given the Administration substantial leverage over the corporate sector, with lobbying on tariff issues tripling during Trump’s first year back in office, and Trump succeeding in using the threat of tariffs as a sort of cudgel to negotiate opaque deals with drug companies to onshore production and to sell drugs through the government’s “TrumpRx” platform.

And don’t even get me started about the government contracts and other deals that appear to involve relatives of senior officials.

The Trump Administration has perfected this kind of lawless thuggery, but it didn’t originate it. In the Obama Administration, officials involved in sanctions policy—including yours truly—would travel around the world raising vaguely defined but definitely ominous-sounding “serious sanctions risks” about doing business in Iran, even with products that didn’t, strictly speaking, violate U.S. sanctions. The hope was that by failing to provide clarity on what was and what was not allowed, you could get business to just dry up—chilling economic activity.

At the time we—and I—rationalized this Chinese-type behavior. Seen through the 20/20 vision of hindsight, it was wrong then and is indefensible now.

I believe in an activist American government. I support the government making targeted equity investments in private companies, when the law allows and there is a compelling policy reason to deploy government equity as opposed to grants, loans, or other tools. I am entirely supportive of the use of novel tools like price floors and offtake agreements, which we have seen the Pentagon deploy to help build the U.S. critical minerals industrial base, as well as the ExIm Bank’s new “Project Vault” public-private partnership for critical minerals. I support the robust use of sanctions and export controls to protect and expand America’s competitive advantage in key technologies like AI. (Indeed, I have spent parts of my career working on those tools.) I am in favor of tough regulations related to AI and other tech products to reduce the risks of societal and existential harms. I support using tariffs strategically to boost the U.S. industrial base.

But what we see today is not just an activist government using its lawful tools aggressively, but the type of thuggishness that has long defined the Chinese government’s relationship with its private sector.

America is a country founded on the idea of the rule of law. Rules set by democratic process, subject to checks and balances, with the American people and companies entitled to a clear understanding of what is and is not lawful. We can choose to break the law and face the consequences. But we should always know what the law actually is.

If we are going to have the U.S. government make equity investments in private companies, the government should set up a proper sovereign wealth fund to do so, rather than relying on convoluted interpretations of statutes that have never before been understood to authorize equity. Export controls should be published in the Federal Register pursuant to regular-order rulemaking processes. Congress should pass laws to regulate the AI industry. If the U.S. wants to run Venezuela as some kind of odd colony, taking control of its oil revenues, we should have a transparent, auditable, and public process to oversee them. Congress, which the Constitution charges with responsibility over taxes and tariffs, should be involved in setting tariff rates.

The rule of law is what guarantees our freedoms. It’s what gives investors the certainty they need to deploy capital to support economic growth. It has been fundamental to our success for the last 250 years.

Admittedly, Congress has become the Supine Branch of the American government in which many (though not all!) members seem to prefer to spend their time appearing on cable news, making Instagram Reels, or having affairs with staffers rather than actually legislating. And there are times, when the Executive Branch needs to push the boundaries of the law to address an immediate crisis. But we cannot and should not pivot away from the foundations of the rule of law that have been so essential to our long-term success.

In today’s geopolitical world, there may be ways in which the U.S. needs to become more like China in order to beat China. We need an activist industrial policy if we are going to succeed in reversing decades of decline in the U.S. share of global semiconductor manufacturing. We probably need a bit more protection for critical industries. We certainly could do a better job than we have in recent decades of helping ensure U.S. influence over strategic materials and strategic assets outside the United States.

But the worst of Chinese state capitalism is something we should leave to Beijing.

What is on my mind:

Europe’s trade policy is getting a bit more American: The European Commission filed three interesting papers with the WTO as part of the ongoing consultation about the WTO’s future. The EU continues to explain its partial pivot away from the concept of MFN, which the EU first started talking about earlier this year, and also explains the challenges it sees to today’s multilateral trading order, including an imbalance between rights and responsibilities on the part of some countries and the need for more policy space for economic security considerations. These are all views that echo perspectives widely shared in Washington, and should be the basis for some points of agreement between Washington and Brussels. But of course, there remain the questions of geopolitical tensions and whether leaders on both sides of the Atlantic actually want to work together.

What will U.S. tariffs look like on July 24? The Section 122 tariffs that President Trump announced on February 20, 2026, hours after losing the IEEPA tariff Supreme Court case are slated to sunset after July 23. I have been expecting—and continue to expect—that the U.S. Trade Rep will finalize its proposed Section 301 tariffs over the alleged failure of U.S. trading partners to prohibit imports of products made by forced labor in the coming days, and those tariffs—either 10% or 12.5% depending on the country—will replace the 122 tariffs. That said, we are getting closer and closer to July 23 without the 301 being finalized, and I suppose there is some chance that, if USTR cannot get the 301 finalized, it might do something like try to deploy a second Section 122. Also, I’d note that USTR has yet to finalize its 301 investigation into industrial overcapacity. As a result, it appears that the Administration will NOT fully recreate the IEEPA tariff rates (which were generally above 12.5%) until at least later this year.

Will Congress re-empower Trump with a tariff tool? I completely understand why Congress both wants to pass a new Russia sanctions bill and why it wants to do so in honor of the late Sen. Lindsey Graham, a long-time hawk on Russia. However, I am opposed to the tariff provisions of the bill, which strike me as ripe for the Administration to use pretextually as the legal basis to impose tariffs on Europe, Japan, or other U.S. allies that continue to import comparatively small quantities of Russian energy. Congress should strip the tariff provisions out altogether and instead focus on the type of sanctions that will both be more effective at reducing Russian energy revenues and less prone to possible abuse.

Of course, I’m not an oracle and cannot know what post-Trump trade policy will be. It will depend on economics, politics, and the views of the next President, not to mention the perspectives of the foreign governments he or she will have to negotiate with.

But I am sure that one of the fundamental issues for Trump’s successor will be whether to simply adjust the U.S. trade architecture that Trump is building, or whether to set off to do something fundamentally different. If I had a Kalshi account and wanted to gamble, I’d bet on adjusting Trump’s architecture.

Trump’s trade architecture has three major components:

First, there is an ideological component. Trump has recast the objectives of U.S. trade policy away from promoting lower costs and towards raising U.S. working class wages, protecting (or, as he would frame it, promoting) favored U.S. industries, and raising revenue for the public fisc. Trump also, like his predecessors, sees trade as a geopolitical tool, but where his predecessors saw access to the U.S. market as a tool to strengthen alliances, Trump views the threat of even higher tariffs as a cudgel to compel foreigners to adopt U.S. policies.

Second, Trump’s trade architecture has a tariff component. Provided that his current and planned tariff regime holds up in court, Trump will have raised U.S. tariff rates by some 300%-400% from where they were on January 20, 2025, putting U.S. rates much closer to India than to the E.U. But the sheer rate increase is only one element of his tariffs. As important as the rates, Trump is trying to abandon the concept of MFN—a generally uniform tariff rate for a product regardless of its country of origin—in favor of a tiered approach to tariff rates: Mexico and Canada appear to have the most preferential rate (still 0 for many products, given the USMCA exemption); a number of countries Trump perceives as allies (U.K., Australia, Argentina, Guatemala, etc.) are in the next tier, at 10%; traditional U.S. allies including the E.U., Japan, and Korea are at 15%; and much of the rest of the world is at 19%-20%, with a few countries higher still. Of course, this tiered approach will only hold if Trump succeeds in legally re-creating the rates he initially created under IEEPA—the current legal authority he is relying on, for example, imposes a flat 10% rate on most countries. (Also, Trump has exempted many price-sensitive products from the tariffs in order to minimize consumer harm).

Finally, Trump’s trade architecture has a deal component, with nine final texts for “Agreements on Reciprocal Trade” and preliminary “frameworks” with many of America’s larger trading partners. These deals generally cap U.S. rates, get foreign countries to lower their rates on U.S. tariffs, get foreign countries to agree to remove a range of non-tariff barriers, such as recognizing U.S. motor vehicle and pharmaceutical standards, and require the countries to sign up for anti-China commitments. The deals might also, over time, create new “rules of origin” that would specify how much Chinese content can be incorporated into a product made in one of the trade partner countries, such as Malaysia, and still count as a “Malaysian” product. Trump is also developing sectoral deals, most notably a critical minerals deal, though not much substance has been publicly released.

In my base case scenario, I expect the next President to keep much of this architecture while reforming it to soften the rough edges and put his or her own stamp on it.

For example, I expect the next President—Democrat or a Republican—to bring U.S. tariff rates down. Trump’s tariffs have proven unpopular, much of the cost of the tariffs appears to be borne by U.S. companies and consumers, and the economic record of the tariffs in terms of job creation has so far been at best mixed. That said, I do NOT expect the next President to simply get rid of all Trump’s tariffs: tariffs take on a political logic of their own, develop domestic constituencies, and I expect that the next President will want to continue using tariffs as an industrial policy tool, albeit in a more targeted way. To vastly over-simplify my view, I expect that if a President inherits an average effective U.S. tariff rate of, e.g., 12%, he or she will bring it down to maybe 6 or 8% within a year or two, but not the 2.5% that prevailed before Trump’s second term.

I also expect the next President to want to keep much of the substance of Trump’s trade deals, albeit with fewer U.S. tariffs and maybe with the deals appearing somewhat less hegemonic. There is strong bipartisan support for many of the provisions that Trump has secured, at least on paper, in his trade deals: the foreign tariff reductions and removal of non-tariff barriers are broadly popular, and I suspect most Presidents will like having trade partners who are signed up to join U.S. economic campaigns against China. A critical minerals trade deal is also a concept with support on both sides of the political aisle. One area I could see the next President backing off of is digital, where Trump has used trade to fight against foreign regulation of U.S. tech companies. Moreover, I assume that many of America’s trade partners will be interested in renegotiating parts of the deals, which they agreed to only under the duress of threats of even higher U.S. tariffs.

Of course, the next President will also almost certainly want to put his or her own mark on the agenda and on the deals. E.g., a progressive Democratic President might want to recast some of the tariffs as carbon-based tariffs and amend the deals to make them more “green.” The next President might also try to stitch the deals, which are currently a sort of hub-and-spoke system, into something more multilateral. But I see these as amendments to the trade order Trump is creating, not an effort to start anew on something entirely different.

Part of my rationale for the base case scenario is that the last President to succeed Trump, Joe Biden, essentially just adopted the trade policy it inherited from Trump, with a few modest tweaks: Biden ultimately left U.S. tariff rates almost entirely unchanged, keeping all of Trump’s tariffs, while, in 2024, increasing tariffs on about 4% of the stuff the U.S. imports from China. (Yes, Biden tried to create an IPEF in Asia and a TTC with Europe, but none of those moved rates or survived beyond his Administration). I also think that while Trump’s tariff rates are unpopular—and, as stated above, I expect the next President to bring them down—tariffs have a tendency to build constituencies, the next President will appreciate a bit of revenue, and no President will want to be seen as just giving something (tariffs) away for free (even if, economically, Americans pay most of the tariffs).

To be sure, if the next President adopts and reforms Trump’s trade order rather than upending it, the next President will have to confront the dubious domestic legality of Trump’s trade agenda—the tariffs might yet be scaled back by the courts!—and the (admittedly small) risk that the supine branch of American government, Congress, will miraculously decide it wants to participate in its constitutional role rather than abdicating responsibility in favor of chasing cable news and social media hits. But those are topics for a future post.

Before I close, just as a thought exercise, if the next President wanted to start from whole cloth rather than simply adapting Trump’s trade order, what might he or she do? I can come up with three possible options:

The alliance of democracies (or other trade alliance): Thinkers from Mathias Dopfner to Wally Adeyemo have all, in recent years, pitched some version of a deeply integrated economic alliance of like-minded states in which they agree to (a) adhere to some set of high standards on labor, environment, etc.; (b) eliminate most trade barriers among themselves; and (c) have a higher common tariff against everyone else. We may well move in this direction over the long term, but my instinct is that this idea is harder to do in practice than it is to write about (think about all the, e.g., fighting over agricultural details!) and so probably won’t happen in the next Presidential term. A lighter version of this would be a “climate club” of countries trying to promote green trade. This might be a longer-term vision, just as the WTO only emerged from the GATT over many years.

Reducing the trade deficit: Trump’s trade policy has neither in fact reduced the U.S. trade deficit nor is it designed to do so: 10% or 15% tariffs on products like TVs won’t really encourage onshoring, they function as a sort of sales tax. But one could imagine a trade policy that would actually close the trade deficit, either by enacting the old Warren Buffett proposal of requiring importers to obtain credits based on U.S. exports, imposing the sort of capital account restrictions that my Carnegie colleague Michael Pettis has repeatedly advocated for, or by ginning up some new Plaza Accord or taking other measures to devalue the U.S. dollar.

Moving to a trading order centered on firms and products, rather than states: Many countries, including the U.S., have already begun to put more emphasis on the specific origins of a product rather than just the country it is from. The USMCA deal, for example, set up a “rapid response mechanism” that can remove tariff exemptions for products made in factories that fail to respect labor rights. The E.U. is setting up due diligence standards for imports. CBAMs tariff imports based on how much carbon was emitted in the manufacture and transport of a specific product. The U.S. has long legally barred imports made by forced labor. Given AI and better data, it might be possible to start imposing low tariffs for products that can show they meet a variety of high standards while applying higher tariffs to products that cannot, even if the products all come from the same country.

I’m sure that many of you have other Please share them, either in the comments or email me, harrell@peterharrelllaw.com. And please do subscribe to this Substack, as I’m trying to build the audience!

*A note to my non-American readers: I write from an American perspective and like most Americans pay woefully little attention to the preferences of foreigners. That said, it is clear that while Trump is rewriting the U.S. trade architecture, the rest of the world is increasingly writing its own trade architecture, including major tariff-reducing E.U. deals with India and Mercosur and Mark Carney’s efforts to create some sort of middle-power trading bloc. While the next U.S. President may well seek ways to work with some of the deals, in my base case scenario I expect that we may see some continued divergence, for at least the mid-term, between the way the U.S. orders its relationship with its major trading partners and the way those trading partners order their trade among themselves.

What is on my mind:

USMCA renewal: I don’t envy the Mexican or Canadian officials charged with getting the USMCA renegotiation over the finish line. I posted recently about what I am watching with the negotiations.

The great middle class wage stagnation debate: Noah Smith posted a chart last week pointing out that, despite lots of popular commentary to the contrary, in fact real middle income wages have increased in recent years. AEI recently arguedthat the middle class is shrinking not because more people are falling behind, but because the upper middle class is growing. It is certainly true that a lot of elite discourse on wage stagnation is still shaped by academic research from the 2010s that focused on the great stagnation of the 2000-2012/2014 period, whereas median incomes have generally risen over the past decade. On the other hand a number of specific demographic groups, like non-college educated men, really have seen their incomes continue to stagnate or decline.

Europe (and Canada’s) potential new defense industrial bank: I’m a sucker for interesting industrial policy, and so am interested to see how the new, Canadian-proposed “Defense, Security and Resilience Bank” plays out. I certainly think that a new low-cost debt financing vehicle is an important part of industrial policy. This might be a topic of a future essay.

There’s no doubt that America is in a gloomy mood on its 250^th, with polling showing that most Americans think the country is on the wrong track and that our best days are behind us. We seem to be unable to figure out how to keep a pond of water on the National Mall from filling with algae, and a former Fox News commentator is now prosecuting a former Olympian for touching a torn piece of the pool’s liner. As others have pointed out, however, gloom around a major anniversary is something of a national tradition at this point. We were gloomy in 1976, in the wake of losing the Vietnam War, President Nixon’s criminality, high inflation, and a burgeoning national debt. In August 1925, a year before the 150^th, 40,000 Klansmen paraded down Pennsylvania Avenue. In short, we’ve been here before.

Aside from my law practice, I spend most of my mental energy thinking about international economics and geopolitics. With that in mind, it strikes me that today is like 1976 not only in the malaise in the national zeitgeist, but also in the manner in which my country is going through an intellectual and policy paradigm shift, as we turn the geopolitical page on America’s “unipolar moment” that emerged after the collapse of the Soviet empire and the economic page on the neoliberal consensus that shaped economic thinking for essentially my entire life.

It is fair to see Reagan’s election in 1980 as the dawn of the neoliberal era and Bill Clinton’s 1996 SOTU comment that “the era of big government is over” as its consolidation. But the intellectual foundation for the paradigm really began to take hold in the 1970s. Edward Crane, Charles Koch, and Murray Rothbard founded the Cato Institute in January 1977. Milton and Rose Friedman published Free to Choose, which brought many of their economic ideas to a popular audience, in January 1980. Congress deregulated the airlines in 1978. (Yes, I know some progressives today rue that decision). Free-market economics was moving rapidly from a fringe theory into the political mainstream.

The same thing is happening today, albeit potentially on an even faster timeline due to President Trump’s bull-in-a-China-shop personality, the collapse of intellectual elites in both parties, and the energy for new ideas from a restive American public.

In the last 18 months, our economic and national security establishment went from complaining about China’s state capitalism to emulating it, with announced plans to take government equity stakes in more than 20 companies so far, and apparent plans to take shares in the AI behemoths as well. At this pace, America won’t just have a sovereign wealth fund at the end of Trump’s term, it’ll have a sovereign wealth fund with hundreds of holdings worth hundreds of billions of dollars.

Twenty-seven years after Bruce Ackerman and Anne Alstott published The Stakeholder Society, we have Trump Accounts, in which the government, philanthropists, and companies currying favor with the Trump Administration all contribute to financial nest eggs for kids. (Regardless of your politics, if you have kids, do sign them up for a Trump Account. It’ll add up for them over time).

I remember sitting with Brian Deese in his office early in the Biden Administration, along with a number of colleagues, debating whether the Biden Administration should embrace the phrase “industrial policy,” or whether it was too politically toxic. A few years later, after the CHIPS Act, the Inflation Reduction Act’s green subsidies, the Pentagon’s Office of Strategic Capital, Trump’s “Section 232” tariffs, and growing policy support for industrial policy for sectors from shipbuilding to pharmaceuticals, we all seem to have become industrial policy-ists now.

I don’t know how the intellectual framework for this ferment will sort itself out. Presumably the Hewlett Foundation and others will continue to fund the smart people in the Academy and elsewhere to write the books to distill it all, while other voices will continue to defend a more traditional free-market approach to the U.S. economy. Meanwhile, Zohran Mamdani will continue to charm us all: whatever you think of his policy ideas, who can’t help but smile at the images of him, in a full suit, jumping into an NYC municipal pool to beat the sweltering heat? That heat risks making our celebrations tomorrow a little sweatier than we’d ideally like them to be.

What I’m thinking about:

What is the future of American economic power? I have a new piece out for the Carnegie Endowment on the future of American economic power. I argue that Trump has the most ambitious agenda for American economic power and hegemony of any president in decades, but that his clear interest in hegemony (both economically and with respect to places like Greenland and Venezuela) has also triggered a backlash that may ultimately undermine U.S. economic power. I also trace the history of how the U.S. thought about its economic power and shifted, post-9/11, from economic system building to weaponizing economic interdependence. It is an analytic think-tank piece and does not offer policy prescriptions (much less legal advice) but hopefully some number of you find it interesting. And if you aren’t interested in economics, the essay is just one of a Carnegie compendium of pieces on the future of American power across multiple domains.

Count me skeptical of the U.S. government owning stakes in the big AI companies. There is a place for government equity investments, for example in low-return strategic supply chains that the U.S. has a security interest in on-shoring or friend-shoring (think critical minerals), or in emerging national security technologies that don’t have as much commercial application. And we definitely need to regulate and tax the AI companies and their customers. But gimmicky proposals to give 5% of an AI company to the government strike me as a transparent bid to avoid taxation and regulation—while taxation and regulation would ultimately raise more money for the government and result in more, and less political, oversight of AI than would equity. Plus, don’t get me started on the risks of Presidents appointing their cronies to the boards of AI companies, who will then prod the companies to develop AI that spits out politically favored results.

Trump’s new 301 tariffs will likely kick off later this month: The current 10% Section 122 tariffs sunset after July 23^rd, and my sense is that the U.S. Trade Representative is working hard to finalize the 301 “Forced Labor” tariffs to be ready to come into force, at a 10% or 12.5% rate (depending on the country), by July 24^th. Interestingly, given regulatory timelines, at this point those 301 tariffs will likely be the only tariffs in place later this month, since at this point USTR almost certainly cannot impose its planned “industrial overcapacity” tariffs as well until sometime in mid-August. (USTR has not yet released its report for that investigation, and it would need to subject any proposed tariffs to notice-and-comment processes). As a result, we probably will not see the full recreation of the President’s “IEEPA” tariff rates until later this year.

Of course, the 301 tariffs will, like Trump’s previous IEEPA tariffs and current Section 122 tariffs, be subject to litigation. While 301 is definitely a tariff statute and has held up well in the courts in the past, Trump is using it in a fundamentally new way that differs from Congress’s intended purpose, and this will be interesting litigation to watch.

Finally, let me know what you think. This is a new endeavor for me, and I’d welcome feedback. (And by all means unsubscribe if you have no interest).

Wishing you a very happy Fourth of July.