Subscribe now

By subscribing, you’ll join over 80,000 people who read Apricitas!

If people are mining for gold, it’s good to sell shovels. And if there’s a gold rush, it’s good if 30% of your nation’s economy is based on shovel manufacturing. Taiwanese GDP rose at a mammoth 23.6% annualized rate in the last quarter of 2025, the fastest pace since the nation’s initial recovery from the 2008 recession, as ravenous demand for AI chips from Taiwan Semiconductor Manufacturing Company (TSMC) buoys the nation’s economy. Growth in net exports has driven the vast majority of this boom, as data center builders in the US, China, and beyond try to get their hands on as much compute as possible.

Overall, the Taiwanese economy has grown more than 12% over the last year, the fastest annual growth rate in nearly 50 years, and GDP has increased almost 23% since the launch of ChatGPT in late 2022. Industrial production and export orders have both hit record highs. Gross exports exceed $63B per month, more than double the pre-COVID average. Manufacturing now makes up nearly 40% of the nation’s economy, the highest since the late 1980s. The global rush for artificial intelligence has been a massive boon to the island nation, pushing its economy back to the growth rates that characterized its post-WWII economic miracle.

Yet Taiwan’s still caught in a difficult position as the world’s largest producer of cutting-edge chips and the primary manufacturer for AI chip designer NVIDIA. On one side of it is China, Taiwan’s chief geopolitical foe, the world’s second-largest consumer of computing power, and a country that has repeatedly shaken traditional AI behemoths with models like Qwen and Deepseek. On the other side is the United States—dependent on Taiwan for chips, desperate to contain China’s AI ambitions, and paradoxically hoping to keep China dependent on American companies’ hardware and software. Taiwan’s exports both across the strait and across the Pacific have skyrocketed over the last year, rising more than 70% as the US and China invest hundreds of billions in developing smarter AI models.

Taiwan needs to maintain this “silicon shield” of high-end chip manufacturing to counter geopolitical threats—as long as the PRC is dependent on Taiwan’s semiconductors, it is much less likely to attack the island, and as long as the US is dependent, it is much more likely to rush to Taiwan’s aid. For the same reason, both China and the US hope to pull semiconductor manufacturing away from the island to reduce their security dependence on Taiwan. China has undergone a long period of steadily indigenizing semiconductor production through national industrial policy, while the US built heavily subsidized semiconductor fabs with the 2022 CHIPS Act and is now using tariff threats to push TSMC to move more production to America.

Yet both countries also have to balance their desire to ease security dependencies with their massive current need for AI chips that only Taiwan can provide. Working to keep as many advanced chips as possible for itself and hoping to hamstring Chinese companies’ AI development, the US government has barred the export of NVIDIA’s flagship B300 chip or the slightly pared-down B30A model that the company made specifically in hopes it would be more amenable to export control officials. Instead, the US has greenlit the export of the less-powerful NVIDIA H200s to China, provided 25% of revenues are kicked back to the US government. China originally banned imports of those H200s, negotiating for better chips and hoping to push its domestic semiconductor industry, but is now rumored to have given the green light for ByteDance, Alibaba, & Tencent to import up to 400k of them to train their own AI models.

Taiwan’s economic boom is expected to continue into the near future, with the national statistics agency projecting that GDP growth will approach 8% in 2026, the third-highest growth rate this millennium. Export orders, a tight predictor for future monthly exports, hit a record high of more than $76B, up 44% over the last year. America has long been Taiwan’s top foreign customer, even surpassing China as a destination for computer parts exports in 2022, but it has rapidly expanded its lead and now dwarfs all other trading partners. Taiwanese computer and electronics export orders bound for the USA currently exceed $20B per month, more than double the next-closest major region.

In fact, America now imports more directly from Taiwan than from China for the first time since 1992—the trade war has cut US direct trade with China, while the AI boom has caused a surge in spending on Taiwanese-made semiconductors. The US obviously imports many Chinese goods indirectly, sometimes transshipped through third countries, but much more frequently just as components of finished manufactured goods produced elsewhere. Yet the same can be said of Taiwanese chips, many of which are assembled into more complicated electronics in countries like Mexico before making their way to the US. The massive rise of Taiwanese-US trade is a genuine sea change in where America’s overseas supply-chain dependencies lie.

Yet there are still significant economic risks amidst Taiwan’s modern miracle. The massive success of TSMC reinforces the island’s two-track economy, where chipworkers prosper and the rest of the economy muddles along. The country has made major financial moves to prevent the export boom from driving currency appreciation, including deregulating many of the life insurance companies that hold foreign-currency assets. A cheap currency makes it harder for consumers to afford imports, especially food and energy, which are largely produced off the island. The country is increasingly dependent on imported LNG as it struggles to meet the power demands of semiconductor manufacturing. And of course, any reversal in the fortunes of AI companies would now hit Taiwan extra hard. The country will have to stay at the forefront of AI chip manufacturing, navigate one of the world’s tensest geopolitical rivalries, and manage a delicate macroeconomic situation to keep its economic boom going.

Breaking Down Taiwan’s Boom

Read more

Subscribe now

By subscribing, you’ll join over 75,000 people who read Apricitas!

Yesterday, the Supreme Court struck down the majority of Trump’s tariffs in the largest legal blow to Trump’s 2nd-term economic policy. “The Framers gave ‘Congress alone’ the power to impose tariffs during peacetime,” said Chief Justice Roberts, “and the foreign affairs implications of tariffs do not make it any more likely that Congress would relinquish its tariff power through vague language, or without careful limits. Accordingly, the President must ‘point to clear congressional authorization’ to justify his extraordinary assertion of that power. He cannot.”

The court thus ruled that the International Emergency Economic Powers Act (IEEPA) did not give the president the authority to impose tariffs, meaning all of the country-level tariffs—50% on Brazilian goods, 15% on imports from the EU, etc—have been rendered null and void. The sectoral tariffs on steel, aluminum, cars, furniture, and other goods used the more legally robust Section 232 national security tariff authority from the 1962 Trade Expansion Act and will thus stay in place unchanged, as will the Section 301 tariffs used primarily against China that predate 2025. Yet the country-level tariffs made up the vast majority of Trump’s overall trade war, representing roughly 73% of the total new tariff revenue since he returned to office.

However, the White House and Supreme Court both knew that “many current federal laws continue to grant the President expansive tariff authority” outside of IEEPA, although these pathways come with more time and procedural regulations. Trump responded to the court ruling by immediately invoking one of those other federal laws—Section 122 of the 1974 Trade Act—to impose a 10% baseline tariff to replace the IEEPA tariffs.

Section 122 can only be used for flat tariffs of up to 15%, with no between-country variation allowed. The administration has thus settled on a 10% tariff—the lowest prior country-level tariff rate—while the important tariff exemptions for computers, energy, pharmaceuticals, smartphones, certain foodstuffs, and most USMCA-compliant goods all remained. The net effect is that average US tariff rates will dip roughly 3 percentage points in the wake of this ruling, and no imported goods will see a net increase in tariffs.

Yet Section 122 only allows tariffs to persist for 150 days without congressional authorization, which remains unlikely. It also rests on some shaky legal footing of its own, having been written to empower the President to prevent international payments problems within the fixed exchange rate system that no longer exists. There will certainly be legal challenges to this new Section 122 executive order in time, but before its 150 days are up, the administration could reconstruct much of the pre-ruling tariff regime using the more legally secure sector-level Section 232 “national security” investigations and country-level Section 301 “unfair foreign trade practices” investigations.

Still, this is the most significant restriction on Trump’s trade war capabilities since he retook office. Without IEEPA, the President can no longer wake up in the morning and decide to raise tariffs on Indonesian goods before lunch. The Section 232 and 301 processes are much more bureaucratic, taking months to put together the proper paperwork and giving businesses more ability to directly challenge the specific “national security” or “unfair foreign trade practices” justifications in court. Plus, since the US collected more than $150B in taxes under executive orders now deemed retrospectively illegal, we could see a rather massive refund to importers. It’s an important win in placing some constraints on executive authority and ensuring Congress retains the power of the purse.

Plus, while US tariffs remain extremely high compared to their pre-Trump equilibrium, they have meaningfully eased even before this week. In November, the US finally exempted coffee, bananas, and other foodstuffs that climatologically cannot be produced at a sufficient scale in the US. The Trump administration had also eased tariffs on Chinese goods down to the lowest level since the start of his second term after negotiations in Beijing. Tariffs on many Indian goods were reduced by 32% in a deal signed in early February. Tariff revenue was running at roughly $400B annualized at its peak, but has slid closer to $350B annualized in the last two months of data, and will likely fall more in the January and February numbers. That’s still much higher than in 2024, but a meaningful retreat in tariff intensity.

The Trump administration could use this ruling as an excuse to continue that partial trade war retreat. Yet in reality, the next few months are more likely to reverse the small amount of trade reliberalization America saw this winter. Lawyers will be working tirelessly to recreate the tariff levels that existed in January. Donald Trump will go back to making ever-changing tariff threats (as of writing, he now claims the 10% 122 tariffs will be raised to 15% “in the next short number of months”, though no formal executive order exists yet, so it’s unclear if this will actually happen or what the details would look like). Even though the Supreme Court has reined in the worst of Trump’s trade war, tariffs and their associated chaos are not going away any time soon.

Subscribe now

How Have the Tariffs Been Going?

More than a year into the second Trump administration, it’s worth pausing to ask if all these tariffs are even achieving their desired policy outcomes. The ostensible goal was to reshore manufacturing production, create good-paying blue-collar jobs, and close the trade deficit. On the jobs front, the US manufacturing sector continues to lose jobs, now joined by declining employment in transportation and warehousing, while growth in construction has fallen dramatically. For an administration rhetorically tying itself to the fortunes of blue-collar workers, this is an especially dismal outcome.

Meanwhile, the massive increase in US tariffs is also driving up prices. Goods prices are up more than 1.7% over the last year, with goods excluding food & energy up 1.9% and durable goods up 2.1%. Those numbers are low compared to the worst of the 2021-2022 inflation surge, when prices were rising more than 6% per year, but they are still meaningfully higher than normal. In fact, manufactured products tend to experience rapid price deflation, so a 2% growth rate is 3-4% above normal levels. Durable goods inflation, for example, is currently running faster than at any point between 1994 and 2020. Americans, not foreigners, and American consumers especially, have been paying the vast majority of the tariffs.

Are the tariffs driving a supposed reindustrialization investment boom? No, real construction of factories has continually declined throughout Trump’s term as CHIPS Act & IRA projects are completed, while tariffs weigh on other sectors. The only place where America undoubtedly has an investment boom is in computers and data centers, which remain completely exempt from Trump’s tariffs.

Finally, did the “Liberation Day” tariffs achieve their ostensible purpose of “fixing” the US trade deficit? No, annual net imports of goods & services hit another all-time high in 2025, still well over $1T even after adjusting for inflation. The reduction of net imports in the last three quarters of the year did not make up for the fact that imports surged at the beginning of 2025 as businesses stockpiled ahead of the tariffs. Much of the downward swing in trade is just reduced imports of pharmaceuticals, which remain exempt from tariffs but looked to be under threat at the beginning of last year. Combined, these are not the results of a working policy regime torn apart by one court ruling, but a regressive tax struggling to achieve its supposed economic aims.

Subscribe now

Conclusions

Perhaps worst of all, none of the uncertainty that characterized early 2025 trade policy has actually been resolved amidst all this. One year ago, Trump was repeatedly imposing and retracting ruinous tariffs on products from Mexico and Canada, our closest neighbors and some of our largest trading partners, before the President largely backed off in March and just exempted the vast majority of Canadian and Mexican goods. Besides some tweaks to car and metals tariffs, American trade policy with the two countries has been fundamentally unchanged since then. There still is no “deal” with either Mark Carney or Claudia Sheinbaum, even the more vacuous kind signed with the EU or Japan, and there is very little honest effort in the way of negotiations despite how important these countries’ goods are to the US economy.

Instead, Trump is regularly threatening to ground the Canadian-made regional jets that form the backbone of regional carriers, block the Gordie Howe International Bridge built to connect Detroit and Windsor, or impose 100% tariffs on Canadian goods. Many of these threats barely register as news stories in the United States because of their sheer volume and how infrequently Trump actually follows through. Yet he sometimes does, like with the 50% tariffs on Brazilian or Indian goods, making all his threats impossible to completely write off.

How should businesses in countries like the UK, which have supposed “deals” with Trump to keep the tariffs on their goods at only 10%, feel about today’s announcement that all tariffs will eventually go to 15%? Should they ignore it and hope that nobody drafts the executive order, like what happened with the 100% tariffs threatened against Canada? Should they lobby for even more nebulous concessions to keep the tariff rate low, like Switzerland had to do? What should they tell their US business partners or customers? Why should they sign the next “deal” if Trump so quickly reneges on this one?

This ruling should have forced some introspection from the Trump administration. Why did we rest our signature economic policy plans on shaky legal theories? Why has it not achieved its desired policy aims? Why are we forced to exempt the vast majority of imports from tariffs to keep the economy functioning? Why can’t we decide on the right tariff rate and keep it there for an extended period of time? Instead, they’re likely to retreat back into the tariff chaos that has dragged down the economy throughout the last year.

Subscribe now

By subscribing, you’ll join over 75,000 people who read Apricitas!

The United States is undertaking a historically unprecedented investment boom to build the computers, data centers, and other physical infrastructure needed to train and deploy Artificial Intelligence. Hundreds of billions of dollars have already been spent by hyperscalers racing to build smarter AI systems, and investment from major tech companies is set to shatter all previous records again this year. Amidst this boom, spending on data center construction has hit a new record high, now exceeding a $42B annualized pace, a more than 300% increase since the launch of ChatGPT in late 2022. While growth has slowed over the last six months, investment is still up more than 18% over the last year alone.

Yet that figure only reflects the costs to build data center facilities themselves, not the much larger costs of the expensive GPUs, TPUs, and other electronics housed within. Real US fixed investment in those computers and related peripheral equipment has surged to a record high of more than $270B annualized, up nearly 50% over the last year and up 77% since ChatGPT’s launch.

Likewise, spending on computers is dwarfed by investment in software development, which now exceeds $750B annualized, with an increasingly large share focused on training, deploying, and integrating AI systems. Combined, spending on data center, computer, & software investment now exceeds $1T annualized and is roughly equivalent to 3.5% of US GDP, another record high. This is also likely an undercount of the true value of AI-related fixed investment, as official statistics have a difficult time accounting for spending on computer parts or properly capturing all software development. Still, spending on data center facilities now rivals total office construction, spending on computers now exceeds all factory building, and spending on software investment now dwarfs new home construction. Across nearly all dimensions, investment in AI systems continues breaking records.

That investment is also poised to break more records throughout 2026—physical capital expenditures among the hyperscalers of Meta, Alphabet, Amazon, and Microsoft alone will comfortably exceed $600B this year, up from $370B last year. Alphabet alone expects its spending to reach $175B-$185B in 2026, with Meta spending more than $115B and Amazon a mammoth $200B. Yet this investment boom is an almost uniquely American phenomenon—computer investment in Canada and the UK are still below their 2022 highs, while EU and Japanese investment show no rapid acceleration. Even Chinese software and information technology fixed investment, the closest competitor to the United States, is up “only” 20% over the last year, and from a much lower baseline level.

In other words, America is making a globally and historically unprecedented bet on the success of Artificial Intelligence. As a share of the economy, that AI boom is already one of the largest investments in American history—dwarfing the peak of the broadband, electricity, or interstate highway buildouts and vastly exceeding the Manhattan or Apollo projects. And yet, US tech companies are doubling down, raising the stakes on their $1T gamble that AI models will continue their exponential capabilities growth and eventually become valuable enough to repay such a colossal investment.

Subscribe now

Breaking Down the AI Boom

Since late 2022, US tech companies have shifted from a rapid-hiring asset-light growth model to a no-hiring, capital-intense one. Over the last year, these tech companies have cut roughly 20k jobs while their net holdings of property, plants, and equipment have risen by more than $215B, the fastest growth on record. In total, these companies now actively hold nearly $700B in physical assets, more than double their pre-ChatGPT levels. Given the rapid depreciation of electronics hardware, these numbers represent significantly more gross investment than the net values suggest.

Yet the vast majority of this investment comes in the form of data center computers made outside the United States and thus have to be imported from countries like Taiwan, Mexico, and Malaysia. Real imports of these computers, plus the related parts and peripherals, have skyrocketed to a record high of nearly $400B annualized in recent quarters. Imports of large computers, including those that populate AI data centers, are up more than 220% over the last year and 300% since the launch of ChatGPT.

Those AI-related imports remain by far the largest exemption to Trump’s trade war. In November, the most recent month for which we have official trade data, nearly $42B in computer & parts imports were brought into the US completely tariff-free. The computer & parts exemption alone represented more than 16% of total US goods imports that month, roughly equivalent to America’s total imports from the entire Eurozone.

The vast majority of those AI-related computer and parts imports come from three countries—Taiwan, home of TSMC, Mexico, a leading center for computer assembly, and Malaysia, a significant chip & computer parts supplier. Imports from Taiwan in particular have skyrocketed, rising 670% over the last three years alone. In fact, the US now imports roughly as much directly from Taiwan ($20.3B in November) as from China ($21B). Likewise, AI-related imports from Mexico are up more than 200%, and desktops have now displaced passenger vehicles & pickup trucks as America’s largest import from its southern neighbor. However, those Mexican computers are also mostly made with Taiwanese chips, as Mexico’s imports from the island nation peaked at an $80B/year pace in October, up 315% over the last year.

Meanwhile, real economic output has hit new record highs for each of the three main subindustries representing the tech sector—software publishing, computing infrastructure, and computer systems design services. Computing infrastructure output, which most directly represents the data centers backing AI training & inference, has risen by more than 15% over the last year and contributed 0.57% to the 4.4% annualized pace of growth in the 3rd quarter alone.

The AI boom, plus the completion of many semiconductor fabricators subsidized by the CHIPS Act, has helped push US semiconductor industrial production to record highs, up more than 15% over the last year and 30% since the launch of ChatGPT. American fabricator capacity has grown even faster, up more than 20% and 55% respectively over those same time frames. As more chip fabs come online, US production should increase further in 2026—even though US factory construction is declining, electrical & electronics factories are still being built at a rate of $95B per year.

These data centers and chip fabs all consume electricity at extremely large scales, meaning electricity sales to the commercial sector—which includes data centers—have risen by more than 10TWh since 2019, significantly exceeding growth in both residential and industrial uses. That growth is not expected to slow anytime soon, as the US Energy Information Administration projects commercial power consumption will rise another 2.6% this year and 4.3% in 2027. It is also hyperconcentrated in the electricity grids of Texas’s ERCOT and the mid-Atlantic’s PJM, America’s two primary epicenters of data center construction. Data center demand is thus also contributing to the growing real investment in US electricity generation, which is hovering near a record high of $110B annualized, and growing commercial electricity prices, which are up nearly 7% over the last year.

In fact, US spending on AI is so ravenous that it is breaking one of the iron laws of computing—throughout history, electronics almost always get cheaper over time as manufacturing processes improve and more power is squeezed onto smaller and smaller chips. Yet prices are now rising at the second-fastest pace on record, only trailing the worst of the COVID-era semiconductor shortages. That was a period of much more generalized price spikes as well, meaning that computer prices are actually increasing faster relative to aggregate inflation today than during the worst of the chip shortage. Some subcomponents have absolutely skyrocketed in price, with computer memory prices more than tripling over the last year—America’s AI boom is so massive that it is currently straining the capacity of global electronics manufacturers.

Subscribe now

Conclusions

The share of companies using AI directly in their production processes has been continually rising over the last few years, with roughly 18% of firms now saying they use AI directly in their business functions. Among America’s largest firms and in some sectors like Information or Professional Services, that number can reach as high as 30%, and that’s before even accounting for the extremely common indirect use of AI among individual employees.

Yet despite rising AI usage across the board, most tech companies are not seeing a massive reacceleration of their gross revenue. Make no mistake, tech companies’ earnings have rebounded from their 2022-2023 lows and are now back in healthy territory; they are just vastly outmatched by the unprecedented growth in AI investment. OpenAI, Anthropic, Google, and Meta all hope for this trend to be reversed in the coming years, for revenues to eventually catch up with the rapid growth in spending as is necessary for their investment gambles to pay off.

Yet in the meantime, the impact of this investment has been less pronounced from a national accounting point of view than the eye-popping headline figures would suggest. The tech industry has been a notable driver of GDP growth over the last year, but since the vast majority of computer fixed investment comes directly from imports, much of the short-term GDP boost from AI investment accrues to foreign electronics manufacturers (Taiwanese GDP is up 12.4% since late 2024). In the most recent data, AI investment net of imports only contributed 0.3% to the 4.4% annualized pace of US GDP growth.

In other words, the US economy is buoyed by the AI boom but is nowhere near entirely driven by it—meaning these investments have to pay off both as useful productivity tools for end users and a positive use of expensive compute for infrastructure providers if they hope to accelerate American growth. Even if the technology succeeds, losing companies—whether they be Microsoft, Google, or Anthropic—could be burned if their customers end up moving en masse to superior models developed by their competitors.

Still, recently released LLMs can now semi-reliably complete software development tasks that would take a human 6 hours, and capabilities measured in task-length have been doubling roughly every 7 months. Meanwhile, most AI companies still have massive spending capabilities from their other profitable digital ventures, and they’re easily able to tap investors for additional money at extremely high valuations. For now, America’s AI boom shows no signs of slowing down.

Subscribe now

By subscribing, you’ll join over 75,000 people who read Apricitas!

Just one year ago, Elon Musk rushed into the White House as the head of the newly formed “Department of Government Efficiency”, or DOGE, aiming to radically shrink the federal government. Without approval from Congress, DOGE attempted to functionally dissolve the Department of Education, the Consumer Financial Protection Bureau, the United States Agency for International Development, the US Institute of Peace, and several other agencies. The civil servants in charge of the Bureau of the Fiscal Service, the Internal Revenue Service, and the Social Security Administration were all bumped aside as Elon aimed to control key spending processes in the federal government. Grants for universities, charities, businesses, and scientists were temporarily withheld or cancelled. Musk fired tens of thousands of federal employees and pushed all those remaining to take the opportunity to resign.

After a chaotic six months, DOGE itself was functionally shuttered, and Elon was out as part of a then-acrimonious breakup over his opposition to the increased deficit spending in the “One Big Beautiful Bill.” Yet it took until today before we could finally examine the macroeconomic impacts of DOGE, as all the federal employees on deferred resignation programs were only formally laid off near the end of 2025.

DOGE’s legacy is officially the single-largest annual decline in the federal workforce in 75 years—with total federal employment down by roughly 277k, or more than 9%, since Trump’s inauguration. That’s as large as the drop in employment seen at the end of the Korean war or during the post-Cold War “Peace Dividend,” except concentrated over a period of less than a year and focused more heavily on civilian instead of military employment. It’s also a stark contrast from Trump’s first term, which saw total federal employment rise by 70k. The share of the US workforce employed by the federal government, already hovering near historical lows before DOGE, was knocked down to the lowest level in modern history.

However, DOGE was not able to come close to Elon’s promised trillions in cuts to overall federal spending, which is dominated not by administrative expenses but by Social Security, public health insurance, income redistribution programs, interest, and the military. The supposed “hundreds of billions in fraud” in those major spending programs were never found because fraud is nowhere near that common, contrary to the claims that Musk continues to make. Total gross federal expenditures in 2025 were roughly $7.8T, up from $7.4T in 2024, with spending up across basically all major functions.

Yet even though DOGE was wholly unable to reduce the federal deficit, it still had important lasting effects throughout the economy. Washington, DC, is now in a localized recession, with the city losing 4.2% of all its jobs over the last year—the district’s fastest pace of decline in 75 years outside of the early pandemic. In nearby Maryland and Virginia, which also have significant federal presences, employment is down 0.5% and up a meagre 0.2%, respectively. Yet federal jobs were lost in every single US state and territory, with 80% of the drop in federal employment occurring outside of the broader DC area, boosting overall unemployment across the country. Many of the government’s most important science and health agencies saw severe employment losses that it will be difficult to recover from—and of course, the heavy cuts to the extremely small share of the budget spent on effective international aid programs will harm long-run global economic growth.

Subscribe now

Who Got DOGE’d?

By and large, federal workers can be split into four categories. First, there are the 1.5M bureaucrats, administrators, lawyers, analysts, social workers, researchers, and support staff for civilian agencies. Second, there are the 600k delivery, sorting, and management workers of the US Postal Service, America’s largest state-owned enterprise. Third, there are the 560k defence employees at the Pentagon and other US military installations (though active-duty soldiers are not counted as employees). Finally, there are the 370k doctors and nurses at government-run hospitals for the Department of Veterans Affairs and the Indian Health Service. In the months since DOGE began its work, about 12.4% of nondefense civil servants were out of a job, an identical 12.4% of military employees, about 3.7% of federal hospital employees, and 1.8% of USPS employees. That means the majority of job losses, 182k of the 277k total, occurred at civilian bureaucratic agencies.

Numerically, firings at the Veterans Health Administration (VHA) were the largest source of job cuts, with the public healthcare system down more than 23k employees. Yet those job losses were only a small share of the agency’s massive 400k-strong workforce—firings at the Internal Revenue Service (IRS) were proportionately much larger, with the tax-collection agency down more than 20k of its 100k workers over the last year. IRS employment is now at the lowest level since pre-COVID, amidst ongoing administrative chaos at the agency (There have been 5 separate IRS commissioners since Trump took office, and Treasury Secretary Scott Bessent has been acting commissioner since August). From a pure fiscal deficit perspective, this is certainly the most counterproductive source of job cuts, as IRS employees raise revenue via increased enforcement by amounts that far exceed their salaries. The Food & Drug Administration, National Institutes of Health, Forest Service, and Social Security Administration are other major agencies that have seen significant employment losses.

Yet USAID was by far the single-largest reduction in percentage terms, with 92% of the agency’s nearly 5k employees fired over the last year, leaving just a skeleton crew finishing up dismantling the agency. Only tiny agencies like the Institute of Education Sciences (-87% of their 191 employees), the National Endowment for the Humanities (-73% of 216 employees), and the Agency for Healthcare Research and Quality (-64% of 297 employees) have seen firings of similar magnitudes. Even the semi-major agencies with deep cuts—like USDA Rural Development (-37%), the Bureau of the Fiscal Service (-34%), and the Small Business Administration (-31%)—don’t even come close.

The only major agency that added jobs was Immigration & Customs Enforcement (ICE), which hired roughly 6k new workers over the last year amidst Trump’s mass deportations push, while the Department of Homeland Security is the only major department that added jobs. ICE claims hiring has accelerated further in recent months and that the agency has actually added another 6k jobs since November, the last month for which official data is available. In total, this means that while DOGE did not have macroeconomically significant budget impacts, it did have important effects on the US job market.

The Labor Market Impacts of DOGE

Of the net 277k federal employees who lost their jobs over the last year, roughly 40% still remain out of work amidst America’s weakening job market, pushing the unemployment rate for former feds to the highest level since early COVID. Given how many employees simply moved into retirement rather than look for other jobs, it’s plausible that an outright majority of working-age former feds are still unemployed. This is large enough to be a measurable driver of rising aggregate unemployment—total unemployment levels rose by roughly 660k over the last year, of which the rise in former federal unemployment contributed roughly 110k.

Read more

Subscribe now

By subscribing, you’ll join over 75,000 people who read Apricitas!

The US job market slowed down significantly in the latter half of 2025, with the country adding functionally zero jobs over the last 5 months as the unemployment rate ticked up to some of the highest levels since early 2021. Averaged across the entirety of 2025, the US added only 44k jobs per month, the lowest since 2020 and lower than any year during the 2010s. It’s a dramatic reversal for a labor market that was setting modern hiring records only a few years ago.

In 2021 and 2022, the US economy experienced the “Great Resignation” (or perhaps more accurately the “Great Reshuffling”), where tens of millions of workers quit their jobs and were hired into better-paid positions in other companies. Then, between 2023 and 2024, the US economy settled into a “low hire, low fire” equilibrium as employers were not onboarding many new workers, but were also laying off fewer workers than pre-pandemic. That has now given way to a “no-hire” economy—layoffs have inched up towards their pre-pandemic average, but hiring has sunk all the way down to 2013 levels.

Throughout this period, the unemployment rate has crept up from a low of 3.4% in late 2023 to a recent high of 4.5% set in November 2025. Looking at broader measures of underemployment—including people who want a job but aren’t in the labor force or who are part-time but want a full-time job—the job market is in nearly the worst place since the early pandemic, with roughly 10.8% of the broader labor force looking for more work.

Today’s labor market is perhaps most reminiscent of the “jobless recovery” that characterized the early 1990s or mid-2000s, where employment rates remained stagnant even as overall economic growth had already rebounded. The nearly three-year period from February 2023 to the present is the longest time since those previous “jobless recoveries” where employment rates for prime-age workers have stayed within a one percentage point range, never declining below 80.4% and never rising above 80.9%—despite a more-than-7% increase in US GDP over the same timeframe.

That does not reflect an economy at the apex of full employment, both because America has seen higher employment rates in its own past and because many countries (like Canada, Australia, Japan, and the vast majority of EU member states) all currently have noticeably higher employment rates than the US. But it does reflect a labor market continually stagnating in a “no-hire” equilibrium.

Subscribe now

Breaking Down the Job Slowdown

The US has added only 525k jobs over the last year, the slowest pace of job growth since 2020, and a number less than one-third of any period in the late 2010s. That slowdown has been spread across nearly all major industries, with government, professional & business services, goods-producing industries, and the trade, transportation, & utilities super-sectors all losing jobs. The main industry adding jobs is healthcare, which is on a long-run growth trend amidst the rising medical needs of an aging population, and thus is almost always growing faster than the labor market as a whole.

Digging further, the contraction in blue-collar jobs continues to intensify, with their employment levels shrinking by more than 145k over the last year. Job growth in construction has functionally zeroed out, while manufacturing and logistics are each losing tens of thousands of jobs. Warehouses alone have shrunk their workforce by more than 50k over the last year, putting the sector 150k jobs below its early-2022 peak.

Meanwhile, the US public sector lost more than 157k jobs over the last year, primarily thanks to the DOGE cuts at the federal government. Federal employees represent only 13% of all public-sector jobs, yet the 277k jobs eliminated at the national level, plus the 45k losses at the state level, easily overwhelmed the slow growth at the local level. That local government job growth is roughly evenly split between public K-12 schools and all other local government functions, while state government job losses are concentrated entirely within public Colleges, Universities, & other educational institutions.

Job growth in professional services industries also remains negative, with employment shrinking by roughly 62k over the last year. The high-paid white-collar sectors like law, accounting, architecture, and management consulting are still adding jobs, just at a much slower rate than pre-COVID, but employment levels at temp agencies and administrative support firms have been continually declining for years now.

Finally, the tech sector labor market remains extremely weak, with employment shrinking by more than 19k over the last year and down roughly 100k since late 2023. That’s not as bad as the worst of the 2024 tech job market, but it is much worse than the 2022 boom or any period during the 2010s. At least during the 2008 financial crisis, it took only just over a year before tech employment began growing again—today, employment continues falling nearly three years after the beginning of the tech-cession. If this decline persists through 2026, it will beat the dot-com bust for the longest period of tech job market losses in American history.

Subscribe now

The Victims of the No-Hire Economy

Who has been most affected by this shift to the “no-hire” economy? Established workers with existing jobs can afford to wait out the current labor market slowdown, but early-career workers are finding it harder to land their first jobs, making young people among the hardest hit by today’s job market weakness.

Over the last year, the unemployment rate for teenagers is up by 2.9 percentage points, the unemployment rate for people in their early 20s is up by 0.7 percentage points, but the unemployment rate for those above 25 is up by only 0.2 percentage points. This is also an increasingly bad labor market for recent graduates, as the unemployment rate for young college educated workers is up by a full 1.6 percentage points over the last year. More broadly, the unemployment rate for new labor market entrants (like young adults looking for their first paycheck) and labor market reentrants (like stay-at-home parents going back to office work) are both hovering near their highest levels since 2016.

The distribution of wage growth has also radically changed since amidst the “no-hire” economy. From 2021-2022, wage growth for the lowest-paid workers accelerated to the fastest pace in a generation, amidst the Great Reshuffling of those employees into higher-paid fields and companies, and the US economy experienced a rapid decrease in wage inequality as pay growth for the bottom quintile of workers significantly exceeded growth for the top quintile.

Yet wage growth for these lowest-paid workers has ground to a halt over the last two years, and they are now seeing much lower raises than their higher-paid counterparts. In fact, this is now the largest gap in wage growth favoring high earners since the early 2010s, showing just how hard today’s “no-hire” labor market has hit low-income workers.

Subscribe now

Conclusions

The 2025 slowdown in job growth is clearly not just a story about the US population shrinking amidst a rapid decline in immigration. We still don’t have good data on how many foreign-born workers have left the US since Trump’s immigration crackdown, but we do know that employment rates for native-born Americans are slowly falling. Nor is this entirely a story about AI-driven disemployment—sectors extremely insulated from computer technology are some of those with the fastest job declines.

Instead, it’s a combination of factors—tighter monetary policy and contractionary fiscal impulses, higher tariffs and lower immigration, technological shifts and rising uncertainty—all chipping away at job growth from different angles. This leaves the labor market on an uneasy precipice, since employment will begin to decline outright unless hiring rates pick up in the near future. In other words, today’s jobless recovery risks hardening into an outright labor market decline.

Subscribe now

By subscribing, you’ll join over 75,000 people who read Apricitas!

America is losing jobs in blue-collar industries, something that last occurred during the initial shock of the early pandemic and the depths of the Great Recession. The country is down 65k industrial jobs over the last year, a dramatic reversal from 2024, when the US added a lower-than-usual but still respectable 250k jobs. A major slowdown has hit all blue-collar sectors this year, including construction, mining, and utilities—though manufacturing and transportation are driving the vast majority of US job losses.

In total, employment across trades and industry is now down 123k from the all-time peak reached in early 2025 and has been declining nearly every month since February. This is likely underselling the damage as well, since preliminary estimates for upcoming annual jobs data revisions suggest an additional loss of 100k manufacturing jobs and 30k construction jobs.

Manufacturing jobs have been suffering the longest, with employment declining for more than two years straight at this point as the sector’s share of the total US workforce shrinks to a record low of less than 8%. Transportation equipment (mostly car manufacturing) and electronics lead the losses, but the decline has been remarkably broad-based, with all major industries except metals & nonmetallic minerals bleeding jobs. In total, the US has now lost more than 200k manufacturing jobs compared to the recent peak in early 2023.

Yet the biggest shift this year has been the rapid decline in construction sector job growth, which has only added 52k jobs over the last twelve months, compared to 191k the twelve months prior. Residential contractors—including plumbers, electricians, roofers, and other specialty workers—have seen by far the most dramatic shift, losing nearly 55k jobs over the last year alone.

The causes of this blue-collar downturn are multifaceted—manufacturing remains in structural decline amidst the slowdown in demand for durable goods and consumer electronics. The early-COVID homebuilding boom has ended as builders finish work on the large number of projects that began in 2021/2022. Employment in oil & gas extraction continues to drop as crude prices sink to the lowest level since 2021. Transportation and warehousing jobs are declining as the trucking sector struggles.

Yet recent federal policy moves have been counterproductive. Tariffs are hurting blue-collar employment by raising the costs of manufacturing inputs. Immigration raids are disproportionately hurting the construction sector. Cuts to industrial policy subsidies have helped push factory construction down more than 8% over the last year. The administration’s desired “blue-collar boom” is not happening; quite the opposite.

Subscribe now

A Detailed Look at The Manufacturing Malaise

Read more

Subscribe now

By subscribing, you’ll join over 75,000 people who read Apricitas!

Today, roughly one-in-five American workers are immigrants. They contribute massively to the US economy in sectors as wide-ranging as healthcare, construction, education, agriculture, finance, manufacturing, and tech. That means the question “How many people have immigrated to the US this year?” is increasingly important to the American economy, making it all the worse that we currently have no good answer.

In a country like Canada, the vast majority of immigrants come via airports or the few land border crossings with the United States, making it comparatively easy to produce quarterly inflow statistics. The national census is done every five years, ensuring immigration counts are benchmarked to comprehensive data more frequently. Plus, the Canadian government maintains a series of specialized surveys dedicated explicitly to tracking immigrants’ education, labor market, housing, and other outcomes. US immigration statistics are comparatively sparse, with aggregate population estimates only produced once per year and comprehensive censuses only done once a decade.

Previously, this wasn’t much of a problem—America is a much larger country with a much lower immigration rate than Canada, and population flows tended to be stable enough that shifts in immigration policy were not macroeconomically significant. That changed during the pandemic, as COVID-related travel restrictions caused an unprecedented slowdown in immigration flows at a time when regular data collection was extremely difficult. Immigration subsequently rebounded to the highest level in decades, but that 2021-2023 immigration surge did not show up in official data for years because so many of the immigrants went through asylum or refugee systems that were not well-tracked by population statistics.

Now, the same problem is happening in the opposite direction—the Trump administration’s crackdown on immigration is not well-tracked by public statistics, so it is totally unknown how many fewer immigrants have moved to the US this year or how many residents have actually been deported. The last official population estimate was released before the Trump administration began stripping many immigrants of legal status, deliberately shrinking legal immigration flows, and implementing mass deportations of undocumented immigrants. Answering simple questions like “what is a normal amount of job growth, given the immigration downturn” is now so difficult that forecasters and think-tanks are coming up with wildly different answers.

Even answering how many immigrants have left the United States amidst ICE crackdowns is functionally impossible. The US Department of Homeland Security claims that “two million illegal aliens have been removed or self-deported since January 20th,” but these estimates come entirely from the Current Population Survey (CPS) data used to calculate national unemployment statistics and are fundamentally inconsistent with other official data sources. Indeed, because the CPS data has a comparatively small sample and is extremely vulnerable to shifting response rates, the Census Bureau itself explicitly cautions against this kind of use “when other data are available.” The problem is that, for the moment, we have no better data sources to rely on. America is left flying blind.

Subscribe now

The Challenge of Immigration Data Collection

Imagine you are regularly surveying a small town of 1,000 people. Every month, you randomly knock on 100 neighbors’ houses and ask them if they are unemployed. Usually, you get roughly 50 men and 50 women to answer, because the town is half women and half men. Yet some months, by pure chance, you end up with 60 men and 40 women or vice versa. If men are structurally more likely to be unemployed than women, a monthly sample with more men than women will make it misleadingly appear as though unemployment is rising! What should you do about this?

To mitigate this effect, official US jobs data deploys population weighting. To oversimplify, instead of taking the raw unemployment rate in a sample of 60 men and 40 women, they would instead measure the unemployment rate for men and women separately and then weigh them each by half. The Current Population Survey (CPS), which measures US unemployment, deploys population weights for age, race, sex, and Hispanic ethnicity, but not for immigration status.

Returning to our small town example, imagine that 40% of the population is Hispanic. Half of the Hispanic population are immigrants, so immigrants make up 20% of the town’s total population. If a sample randomly comes back 20 Hispanic, 10 immigrant, and 80 non-Hispanic, this will be corrected by population weighting to instead match the real demographics of the town. It’s important to note that even though the immigrant share of the population was not directly controlled for, controlling for Hispanic ethnicity indirectly corrected for the undercounting of immigrants in this case.

Now imagine the government announces a major crackdown on immigration. Border controls are increased to reduce undocumented immigration, laws are rewritten to reduce documented immigration, legal status is stripped from large numbers of existing immigrants, and interior enforcement is amped up to increase deportations. Some immigrants leave your town and some of those who remain become fearful that you will refer them to other parts of the government for harassment or deportation if they disclose their immigration status. The town now has a lower population, immigrants have become less likely to answer your surveys, and those who do answer are more likely to lie about their citizenship status.

Samples start regularly coming back 10 Hispanic, 5 immigrant, and 90 non-Hispanic. The population-weighting for ethnicity corrects for the undersampling of Hispanics, but since there’s no extra population-weighting for immigrants, it looks as though the immigrant share of the population has dipped from 20% to 10%.

That’s a microcosm of what’s been happening in official US data this year—the weighted share of Hispanic CPS respondents claiming to be immigrants has dipped by more than three percentage points, hitting the lowest level in decades. The share of Black respondents claiming to be foreign-born has dropped by roughly two percentage points, and the share of Asian respondents has dropped one percentage point.

Only looking at the CPS data, it’s impossible to determine how much of this downturn in immigrant identification is due to deportations versus how much is just immigrants within the US becoming less likely to respond to surveys. This introduces an important second problem—measuring total population levels. Just as in the small town example, we know that some immigrants left after the immigration crackdown, but these leavers will obviously never be captured in surveys of the town itself. Given this, how can we possibly know the population of the town?

US employment data gets around this problem by projecting forward the population at the beginning of the year and simply never straying from those numbers. In January, they projected the over-16 US population would hit 274M in August, and August unemployment levels were calculated by multiplying the unemployed share of survey respondents to that population level. The same procedure is done to calculate the level of immigrants and non-immigrants in the US, which is where the Department of Homeland Security’s “two million fewer immigrants” figure comes from.

Yet because aggregate population figures are unchangeable constants in between their yearly benchmarkings, this means the same DHS figures that show a decrease in immigrants also show a massive increase in native-born adult populations. That increase is flatly impossible, as we have great records of US birth/death rates and know that 6M extra full-grown adults could not have appeared in the last few months. Thus, when the White House brags that “Over the past year, more than 2.5 million native-born Americans gained employment”, this is simply not the case—the share of native-born Americans with jobs has not budged, the US just erroneously thinks there are more native-born Americans because immigrants aren’t answering surveys. To return to our small town example, it’s as if you determined the population was 1000 at the start of the year and saw the immigrant count drop by 100 and thus presumed 100 more native-born residents appeared out of nowhere.

Beyond the unrealistic native-born counts, there are more reasons to doubt the accuracy of current CPS immigration data. If a net two million immigrants had actually left the United States since the beginning of the year, the country’s population would be shrinking at the fastest rate on record. Given that the vast majority of these immigrants are working-age employees, the US workforce would also be shrinking at the fastest pace ever, and payrolls would likely be falling by more than 100k per month. Actual payroll data are definitely weak, but are not shrinking anywhere near that fast, meaning the CPS is likely vastly overstating the decline in immigrant populations.

Yet answering how much CPS is overstating the immigrant decline is nearly impossible at this point. While the lack of massive payroll declines is enough evidence to know the CPS immigrant counts are inaccurate, it’s also impossible to calculate the true decline in immigration by working backwards purely from payroll data. The majority of recent immigrants are legal asylees with work permits, and stripping those asylees of legal status will show up as declines in payroll data—yet many legal and undocumented immigrants are self-employed, agricultural workers, contributing to family businesses, or get paid under-the-table in ways that make payroll data unable to fully count them. However, even looking at CPS Household Survey data adjusted to exclude those types of workers and therby match nonfarm payroll concepts shows a persistent, significant gap between the two measures that did not exist pre-COVID. That means either official population counts are (still) underestimating the size of recent immigration flows, payroll counts are underestimating the speed of job losses, or some combination of the two.

The next key data will be the census annual population estimates released in December, built off the more comprehensive American Community Survey results, followed by the updated CPS population controls in January. At that point, we will know much more about the size of the immigration downturn, but even still we will be relying on six-month-old data that fails to capture the full extent of the shifts in US immigration flows.

Subscribe now

Conclusions

Given all that we do not know about current US immigration flows, what data can we rely on? First are the employment rates among immigrants and non-immigrants. While measured employment levels are heavily distorted by how many immigrants answer government surveys, the share of those respondents who report having a job should be much less affected. Obviously, to the extent that immigrants are pretending to be native-born in surveys they will distort the native-born data, and to the extent they avoid surveys they will cause the sampled foreign-born population to become unrepresentative. Yet the distortions to employment rates caused by immigrant response issues should be comparatively much, much smaller than the distortions caused to employment levels, especially for the data on native-born Americans. This represents the best information we have on the recent effects of Trump’s immigration crackdown on the US labor market.

From that data, it’s extremely clear that working-age native-born Americans have not seen any substantial increase in employment rates over the last year—in fact, the share of them with jobs has actually inched down compared to 2024. Employment rates among immigrants have decreased slightly more than for native-born Americans, but the decline is also only marginal.

On the flip side, unemployment rates for native-born Americans have been rising steadily and are now up 0.5 percentage points compared to this time last year. Unemployment rates for immigrants have actually fallen 0.1 percentage points over the same timeframe, meaning US-born workers are more likely to be unemployed than foreign-born workers today.

We may lack plenty of clarity on how immigration is affecting the aggregate US labor market, but the data we do have still unequivocally rejects the idea that mass deportations are delivering big wins for American-born workers. This should not be surprising given the decades of empirical economic research showing that immigration does not drive down wages or increase unemployment, but it highlights the core flaw with the administration’s strategy of touting deportations as an economic panacea. We may still lack the data for a proper accounting of the economic costs of Trump’s immigration crackdown, but we can already say it’s not delivering the promised benefits.

Subscribe now

By subscribing, you’ll join over 72,000 people who read Apricitas!

America’s tech industry is in the midst of a historic boom driven by the race for artificial intelligence. Hyperscalers are spending hundreds of billions of dollars to build data centers, AI investment is driving record shares of overall GDP growth, and AI systems themselves are being increasingly deployed throughout the economy. Real economic output in the three major AI-related subindustries—software publishers, data infrastructure providers, and computer systems designers—have all skyrocketed over the last few years.

The companies behind this tech boom are heavily concentrated in California—the seven largest companies in the world are all AI-related, and five of them (NVIDIA, Apple, Alphabet, Meta, & Broadcom) are headquartered in the Bay Area, while the remaining two (Microsoft & Amazon) are Seattle companies with major operations in San Francisco. The information industry, which covers most tech companies, contributed a full 1.7 percentage points to the 4.3% annualized GDP growth California saw last quarter.

Yet amidst the AI boom, the US continues bleeding tech jobs—tech employment is down 90k from its 2023 peak and has declined by more than 10k over the last year. That is a massive slowdown from the 300k-per-year job growth pace seen in the 2022 tech boom or the 150k-per-year pace regularly sustained in the years before COVID. Unemployment rates for computer and math-related occupations have risen roughly 1% since summer 2023.

Plus, the tech job slowdown is hitting the California job market doubly hard, as the state is a shrinking share of US tech employment even while national tech employment is declining. Before COVID, nearly 19% of all US tech jobs were located in the Golden State; today that share is down to nearly 16%. Over the last year alone, California's share of tech employment has declined by another 0.4%.

It’s hard to quantify the precise effect of the AI boom on overall tech industry employment—new employees are required for the design and development of AI models, but these models often go on to displace coding jobs in other tech companies. The massive downturn in tech employment preceded the deployment of ChatGPT and has decelerated, not accelerated, since 2022.

Yet at this point, it is increasingly hard to deny some negative impact of AI on software and computer science jobs. This is now the largest and longest sustained drawdown in tech employment since the dot-com bust, significantly worse than either COVID-19 or the 2008 recession. Despite tech stocks rebounding, interest rates falling, and fixed investment surging, nothing has caused job growth to return to anywhere near normal levels.

Meanwhile, California’s high cost of living and lack of infrastructure investment continue to drive tech jobs to other states. Prices for housing and energy remain some of the highest in the nation, yet the state builds too little of either, making it hard to accommodate newcomers in its growing economy. These problems affect tech acutely but are mirrored across all sectors—California’s aggregate job growth is still 0.5 percentage points slower than the rest of the nation’s. That means Silicon Valley remains in an investment-heavy, hiring-light posture, and California remains at the epicenter of the tech industry’s weak job market.

Subscribe now

Silicon Valley Slump

Tech job growth dynamics between California and the rest of the US saw a massive structural shift right after the pandemic hit. Both continue to follow a relatively similar cyclical boom-and-bust pattern; it’s just that California was perpetually growing 3-4% faster than the rest of the nation through 2019 and has been growing 3-4% slower since 2021. Where are these new tech jobs going instead? While we don’t have extremely detailed data for most states, what data we do have suggests the largest destinations for tech workers are Florida, Texas, and Washington.

In total, California’s tech-sector employment grew by 60k between the beginning of 2020 and mid-2022 before declining by 71k in the three years since. Web search portals (like Google) and computer systems designers (like IBM) have led the decline, while computer infrastructure providers (like Amazon Web Services) have held up the strongest.

Those job losses mean that California’s share of US software publishing jobs has fallen to the lowest levels in the 35 years for which data is available. Meanwhile, its share of computer system design jobs has fallen to the lowest level since 2010, its share of web search jobs has fallen to the lowest level since 2011, and its share of computing infrastructure jobs has fallen to the lowest level since 2019.

The slowdown in the tech job market means that San Francisco has seen the largest job losses of any metro area in the US, with employment dropping by 11.6k over the last year (though it is certain to be overtaken by DC later this year as the full effects of DOGE show up in official data). San Francisco has also seen the largest employment drop of any major metro area since COVID, with its employment declining by 3.8% even as national employment has risen by nearly 5% and 84% of major metro areas have recovered to 2019 job levels. Nearby San Jose MSA has also suffered, though not quite as harshly, as their employment is down 1.4% since the beginning of 2020 and they’ve lost 2k jobs over the last year.

Looking narrowly at the information sector, both the San Francisco and San Jose metro areas have been steadily losing jobs since mid-2022. The pace of loss has eased up in both MSAs, and San Jose is doing much better overall than San Francisco, but at no point has the bleeding completely stopped.

The good news for Californians is that while tech-sector jobs continue to leave the state, the jobs that remain are disproportionately high-paying. The share of information sector compensation going to Californians is actually up slightly from its pre-pandemic peak even as the share of information-sector employment has shrunk dramatically. Nearly one-third of all tech-related wages and benefits still go to residents of the Golden State.

That’s been an important alleviating factor for California’s state government, which is extremely dependent on Silicon Valley to fund public services. The state’s income and corporate taxes are paid disproportionately by the high profit margins, large salary packages, and cyclical capital gains of the tech industry. The AI boom has allowed California tax receipts to rebound from their 2023 lows, though they still remain well below their 2022 peaks.

With the worst of the tech-cession over, the Golden State’s economy is once again growing faster than the national average, even though it is no longer comfortably blowing past the competition as it was pre-pandemic. Indeed, even amidst its current rebound California still trails both Texas and Florida in GDP growth over the last year, and the state’s economy is growing at a rate more than 25% slower than it did in the five years pre-pandemic. This is a partial recovery but not a true return to form.

Subscribe now

Conclusions

Several factors will continue to push against California’s economy in the years to come. The state has the nation’s most immigrant-heavy economy, and tech is one of its most immigrant-heavy industries—the Trump administration’s drastically tightened restrictions on legal immigration are thus certain to weigh heavily on their economy. California’s high cost of energy also means that despite the tech industry’s concentration in the Bay Area, few data centers are planned in the state itself, pushing more of the AI ecosystem to places like Virginia, Oregon, and Texas.

Yet there are also important tailwinds for California. The state has just passed SB79, a bill that allows more housing construction near transit stops throughout the state, representing one of the largest upzonings in the modern United States. That will help significantly alleviate California’s chronic housing shortage that pushes so many workers and businesses out of state. Telework rates, previously a major driver of California outmigration, have also stabilized or declined in the last couple of years. These look likely to slow the rate of Silicon Valley job losses, but on their own they will not be enough to completely stop the steady relocation of tech jobs out of California.

The more pressing question is the more difficult one: what will happen to the tech job market as AI continues to be deployed throughout the economy? Will they be more like journalists, whose jobs were decimated by the internet age, or like radiologists, whose jobs have surprisingly thrived thanks to AI? So far, they’ve held up much better than artistic industries like graphic design services (where jobs are down 10% since late 2022), but are obviously still stuck in their worst job market in years. Only time will tell how much AI will replace or complement existing programming jobs as the technology continues to develop.

Subscribe now

By subscribing, you’ll join over 72,000 people who read Apricitas!

During his second term, Donald Trump has pursued the largest trade war in modern American history, sending tariffs to the highest level since the infamous Smoot-Hawley Act of 1930. Yet at the same time, roughly half of American imports remain completely exempt from his trade war—billions in goods are carved out because of their country of origin or product category. Mexican and Canadian goods that comply with Trump’s first-term USMCA deal remain tariff-free, as do energy products like crude oil, precious metals like gold, and products slated for future tariffs like pharmaceuticals. The single largest of these exemptions, covering an astonishing $34B of imports per month, is for computers and parts—an exemption that AI companies are now completely reliant on for their record-breaking investment push.

American tech companies have been in a frenzy to develop increasingly advanced AI models ever since the release of ChatGPT in late 2022. Training and running these models require processing unprecedented levels of information, which in turn requires some of the largest data centers ever built, which themselves require thousands of advanced computers to operate. However, the modern electronics used in data centers have the most complex supply chains in human history, and America makes up only a small part of the direct manufacturing involved. Thus, imports of the large computers commonly found in data centers now exceed $235B annualized, up 227% compared to before the launch of ChatGPT, while imports of computer parts exceed $67B annualized, up more than 100%. Without this exemption, importers would have paid an extra roughly $8.9B so far this year (naively applying the 10% baseline tariff rate that has been in place since April), or roughly $19.2B (applying the country-level tariff rates in place as of today). The current AI boom would simply be impossible if tech companies had to pay the same tariffs that car manufacturers or homebuilders currently face.

Meanwhile, US economic growth is increasingly driven by today’s AI buildout. In Q2, computer & software investment officially contributed roughly 0.71% to the 3.8% annualized pace of growth, which is almost certainly an underestimate given how official data struggles to capture investment in electronic parts. Large parts of the US economy are now totally dependent on this tariff exception for computer imports.

On one hand, this is a victory for free-traders: even the world’s most pro-tariff administration cannot fully escape economic reality and remains unwilling to mess with its most important supply chains. This massive tariff carveout is much of the reason why, despite the trade war, America’s stock market keeps surging to new highs, and its economic growth has beaten the dire forecasts from early April. It serves as a tacit admission that a full-scale trade war would kill America’s golden geese.

On the other hand, the pick-and-choose style of tariff exemptions means the Trump administration is punishing the rest of the economy to the benefit of tech companies. Tariffs are making it more expensive to build or buy anything unrelated to computing, so consumers and businesses are effectively being incentivized into the AI ecosystem. The White House is making a giant gamble on artificial intelligence as the future of the world’s economy, and if they’re wrong America will have wasted tens of billions of dollars that could have gone to more valuable investments.

More broadly, the AI boom raises a fundamental challenge to Trump’s entire trade strategy. If free trade is delivering such amazing results for the one sector still able to enjoy it, why are we subjecting farmers, manufacturers, and families to intense protectionism? In other words, if data centers get to be exempt from tariffs, shouldn’t we all?

Subscribe now

Breaking Down the AI Investment Boom

The planet’s six largest companies—NVIDIA, Microsoft, Apple, Alphabet, Amazon, and Meta—are all American tech giants deeply enmeshed in the AI ecosystem. Together, they have a valuation exceeding $15T and are investing hundreds of billions in the physical infrastructure underpinning AI development. In the United States, construction of data centers now exceeds a $41B/year pace, rising more than 200% over the last three years. Today, America is spending nearly as much on data center construction as it spends on office buildings.

Yet those eye-popping numbers only reflect the facilities themselves, excluding the computers and other electronics housed inside. The brains of AI data centers are much more valuable than their skeletons, and thus US investment in computers & electronics far outstrips the amount spent on data centers alone. Fixed investment in computers & related equipment has risen to nearly $250B after adjusting for inflation, a record high and a nearly-$100B increase over the last year.

It’s hard to know precisely how much of that investment is directly attributable to AI—after all, computers & data centers are used for a wide variety of digital tasks, including video streaming, cloud storage, data management, etc—but we can get a better idea by looking at the spending of tech companies specifically. Businesses in the information technology sector have increased their holdings of property, plants, & equipment by more than $180B over the last year, a record high that is more than triple the 2023 pace.

Subscribe now

Breaking Down the AI Import Boom

That rising data center investment requires record amounts of foreign inputs since advanced computer manufacturing is a highly complex and extremely globalized industry. The AI investment boom is dependent on imports of GPUs, computers, parts, and other electronic equipment from countries like Taiwan, Mexico, and Vietnam. Those imports are now skyrocketing, reaching more than $370B annualized after adjusting for inflation, and show little signs of slowing down.

The majority of America’s AI-related imports come from Taiwan, the headquarters of world-leading semiconductor manufacturing firm TSMC. US imports from the island nation have skyrocketed from a roughly $25B/yr pace in late 2023 to a $160B/year pace today. Imports from Mexico have also increased significantly, more than tripling over the last two years, while imports from smaller electronics manufacturing hubs like Vietnam and Thailand have risen at a more modest pace. Only computer imports from China have declined this year, as the 20% tariff on all Chinese goods leaves no exemptions for electronics like the tariffs on Vietnam, Mexico, or Taiwan.

Breaking down the trade with Taiwan, the vast majority of imports come in the form of larger computer systems designed to slot into the server racks of American data centers. Those imports have nearly quintupled since the beginning of the year as companies rush to bring as many computers as possible into the country. Data centers also rely on a complex array of parts for their regular operation, and imports of the specialized parts made by companies like TSMC have risen to roughly $30B annualized. Imports of GPUs themselves have also increased significantly, although they make up only a small share of direct US-Taiwan trade—most GPUs are already assembled into computers before boarding ships for America.

Subscribe now

The Notes You Don’t Play

Heavy tariffs are a form of industrial policy, allowing the government to intervene directly in the economy to support certain sectors at the expense of others. That primarily happens through supply channels—American aluminum manufacturers currently enjoy massive government support via the 50% tariffs on aluminum while the downstream airplane manufacturers reliant on aluminum suffer higher input costs. Yet tariffs and their exemptions can also shape investment decisions by changing the relative price of capital goods—right now, it is much more expensive to build an aerospace manufacturing plant relative to a data center because only the plant’s equipment will be hit with tariffs. Real US investment in structures has sunk since the beginning of this year while investment in industrial equipment grew tepidly; only tariff-exempt electronics investment is booming.

The other key issue is that data centers are not just dependent on electronics themselves, they are also dependent on electrical infrastructure for the industrial-scale power required to run modern AI models. Electricity sales to commercial users (which includes data centers) have risen by roughly 7% since 2020, making it the largest contributor to overall US load growth ahead of residential and industrial sales. Yet in specific submarkets, the impact of data centers has been much, much larger—annual commercial electricity consumption in Virginia has risen by roughly 45% (24 TWh) since 2019 as the megacluster “Data Center Alley” continues to grow in the DC exurbs. Oregon has seen a similar 47% (8 TWh) spike in commercial power consumption, Texas has also seen a more than 16% (21 TWh) increase, and tiny North Dakota has seen a 65% increase (4.6 TWh) after the construction of a handful of data centers.

However, Trump has not exempted electricity generation infrastructure from tariffs like he has with data center computers. Transformers, batteries, switches, and more all face substantial tariffs, making it harder for power companies to meet the electricity demands of the AI boom. In fact, Trump is specifically targeting certain electrical equipment for extra-high tariffs—copper wiring faces a 50% tariff, generators face elevated tariffs based on their steel content, and the Department of Commerce has started an investigation into wind turbine imports that sets the stage for future tariffs. At the same time, the administration has been trying to forcibly cancel many clean energy projects by revoking their funding or permits, exemplified by the recent stalling of the nearly-complete Revolution Wind project in Rhode Island.

The obvious result of a policy package that taxes electricity investment but exempts data center investment just exacerbates the problems caused by a power grid already struggling to keep up with rising demand. US construction of powerplants and electrical infrastructure has stagnated in recent years, sliding down from their late-2023 highs and holding near $150B/year even as the AI boom accelerates. Since the start of the year, the US Energy Information Administration has cut its 2025 renewable capacity growth forecast even as it has raised its coal & natural gas capacity growth forecasts. All of this is part of the reason US electricity prices are up more than 6% when compared to last year and are projected to rise even further.

Subscribe now

How the Electronics Exemption Could End

This tariff exemption for computer imports has always been billed as temporary, even though the Trump administration has thus far proven unwilling to close it. Back in April, Commerce Secretary Howard Lutnick said that electronics tariffs would be coming “in a month or two”, and then in early August, President Trump said they would be signed in the next two weeks, only for both deadlines to pass uneventfully. This would not be the first time a “temporary” exemption was just made functionally permanent (the 30-day pause for most tariffs on Canada/Mexico is now more than six months old) but it’s still worth examining what possible upcoming semiconductor tariffs could look like.

The current rumors of chip tariffs come in two broad shapes. Most likely, Trump will make an announcement similar to the one he recently made for pharmaceuticals—insanely high tariffs, but with a massive exemption for any company that “invests in the US.” Reportedly, some versions of this plan require a 1:1 ratio of imports to promised future semiconductor production, forcing the US into eventual aggregate “self-sufficiency” without strictly taxing every chip that crosses the border. In practice, most multinational electronics manufacturers (like TSMC) already have some large-scale investments within the US that they can point to, and so the administration will functionally exempt everything important and only impose tariffs on a small number of Chinese electronics manufacturers.

The administration could also attempt to place tariffs on electronics based on the value of the semiconductors within. This would be the most administratively complex by far, requiring manufacturers to give a price estimate for every single chip and pressing customs agents into enforcing accurate appraisals. This approach would be prohibitively annoying for manufacturers of cars, medical instruments, or other complex machinery, all of which can contain hundreds of different semiconductors that would need to be individually tariffed. The advantage this approach has over standard tariffs is that it would theoretically avoid punishing companies for imports of the many non-semiconductor parts in their finished electronics. Its advantage over the “exemption for investment” proposal is that it would actually set up a concrete incentive for domestic chip production, instead of just settling for promises.

Subscribe now

Conclusions

If tariffs truly were such an unalloyed good, the White House would be chomping at the bit to close the largest remaining exemption. By their own logic, electronics tariffs would bring in tens of billions of dollars in revenue and revitalize domestic advanced semiconductor production! Instead, the administration is scared to rock the boat lest a poorly designed tariff hurt tech companies’ stock valuation and halt the domestic data center buildout.

Paradoxically, the White House is unwilling to use tariffs on the technologically valuable industries because those tariffs come with much larger short-term pain. Domestic electronics production is obviously more important to US national security than furniture production, but America is currently imposing 50% tariffs on kitchen cabinets and 0% tariffs on data center computers because the electronics tariffs would kill the AI boom and furniture tariffs will only raise prices for consumers.

In other words, the Trump administration is increasingly gambling the future of the American economy on AI. If they’re right, the US will capture the dual rewards of both coding the tools of the next economic revolution and owning the infrastructure that runs it. If they’re wrong, the US will be left with a bunch of distressed assets and will have dramatically underinvested in the projects that actually mattered. The large tariff exemption for computers is distorting the American economy in ways that could reshape the long-run future of the country.

Yet in the short run, it’s undeniable that the carveout has achieved its desired effect of accelerating America’s AI boom. The US is at the technological forefront of a rapidly expanding industry because of the free trade policies that this administration ostensibly despises. It’s just tragic that this privilege is not extended to businesses outside Silicon Valley.

Subscribe now

By subscribing, you’ll join over 68,000 people who read Apricitas weekly!

Since the start of his second term, Donald Trump has begun the largest trade war in modern American history—at peak, tariffs had risen to more than eleven times their 2024 levels, and today they remain more than seven times higher than last year even as Trump’s 115% China tariffs and much of his April 2nd tariffs remain paused. Trump has also been constantly manipulating those tariff rates since coming into office, unable to go more than 28 days without another major change. That pattern continued this month, when he suddenly announced a doubling of tariffs on steel & aluminum, and it is almost certain to continue next month as the 90-day pause on April 2nd country-level tariffs expires.

Americans are now paying the costs of those tariffs, in record amounts—Treasury statements indicate that Customs and Border Protection collected $23B in tariffs during May, nearly triple what they collected last year. By revenue, this represents the largest US tax hike in more than 30 years, and it is all concentrated in the haphazard imposition of one of the most economically destructive taxes possible.

This historic increase in trade restrictions, coupled with high volatility and uncertainty, has wreaked havoc on US trade flows over the last four months. Between November and March, the US goods trade deficit surged 53%, before dropping 46% in April as the largest of Trump’s tariffs began to take effect. This is historically unprecedented trade volatility, with the March-to-April change in the trade deficit being the largest on record by an incredibly wide margin. That combination of high trade restrictiveness and extreme policy uncertainty is now pushing American trade to the breaking point.

Subscribe now

The Largest Contributors to the Import Surge

Although virtually every part of US trade has been affected by these tariffs, the shifts in the overall trade deficit have been dominated by only two types of goods: pharmaceuticals & gold. US imports of gold spiked to record highs early this year as uncertainty spiked gold prices and fears that gold would not be exempted from tariffs caused a rift between American and European prices. The end result was a massive surge in gold imports, especially from Switzerland, as panicked consumers searched for safety and financial firms worked to arbitrage the American price premium. That surge in imports rapidly dissipated when it became clear that gold bullion would be exempt from the tariffs, leading to a historic drop in imports during April. Yet this is a great example of the costs of tariff uncertainty, as exempting gold imports from the get-go would have avoided much of this panic buying.

The second major contributor to recent historic shifts in the trade deficit has been pharmaceuticals, where imports peaked at 142% above their 2024 levels before dropping. Drug imports have been mentioned as a key tariff target for months, and while they have so far remained exempt from almost all of Trump’s tariffs, pharmaceutical companies are not taking any chances and stocking up as much inventory as possible. The vast majority of that import surge has come from the European Union, home of research and manufacturing hubs that have long been America’s number one foreign source of medicine—and they’ve been particularly concentrated in Ireland, where many drug multinationals manufacture for tax avoidance purposes.

Subscribe now

America’s Collapsing Trade With China

The largest tariffs America has implemented are those on China—at peak, tariffs on most Chinese goods were a staggering 145%, and even now they sit at a historically unprecedented high of 30% for most goods. To put the intensity into perspective, half of all April tariff revenue was collected from Chinese imports. Those tariffs have dramatically reduced the amount of direct trade between the world’s two largest economies, with US imports from China dropping more than 40% between January and April, reaching the lowest levels since March 2020.

That drop in Chinese imports was led by the country’s large consumer electronics sector, which has long been the world’s primary source of technological devices. US imports of Chinese computers are down 70% compared to last year, with laptop imports dropping to the lowest level in more than 20 years. Imports from Taiwan, Mexico, Vietnam, and a few other countries (like Thailand) have largely made up the slack but are still collectively struggling to meet America’s ravenous demand for computers.

Likewise, US smartphone imports dramatically tanked as tariffs on China ramped up, with imports also dropping 70% compared to last April. America instead bought more phones from India and Vietnam, with both countries surpassing China as a source of imports in the last two months. This largely represents the decision-making of Apple, the country’s dominant phone brand, which has long worked to partially diversify its supply chains out of China and is now putting in overtime to assemble most American phones in its Indian & Vietnamese factories. Yet right now, those non-Chinese-made phones are still nowhere near enough to make up for the shortfall caused by the China tariffs. That’s despite those non-Chinese phones remaining exempt from all tariffs, as Trump has made no moves to implement his long-promised semiconductor and electronics tariffs.

Subscribe now

Looking at Other Import Categories

Yet Trump has imposed other universal sectoral tariffs, and those are already having significant impacts on US trade flows. First is the 25% tariff on all foreign cars, which went into effect on April 2nd and affects roughly $250B in American imports. Those tariffs caused a rapid surge in vehicle imports through March as consumers/dealerships rushed to buy cars before the tariffs went into effect, and then imports dropped significantly in April when the tariffs officially began. Indeed, in the most recent data, total US car imports were down 33% compared to last year. That drop in imports was seen across almost all major US trading partners but was concentrated in cars from Mexico, the EU, Canada, and South Korea.

Then there is the 25% tariff on all steel and aluminum imports that began in March and doubled to 50% in June. The data we have through April shows a significant decrease in aluminum imports after some frantic tariff frontrunning in March, with similar but more muted effects for US steel imports. The imposition of 50% tariffs will likely further reduce metal imports, but at an even higher cost to the wider population of downstream manufacturers, construction companies, and consumers that rely on affordable materials.

Subscribe now

The Pain in Exports

In addition to damaging American imports, the trade war has also had deleterious effects on US exports. Usually, this happens through rising exchange rates, as tariffs cause the dollar to appreciate and make US exports less competitive in the short term, but since April the dollar has actually depreciated heavily as part of the broader “sell America” trade.

The drop in US exports has instead come primarily from retaliatory tariffs and hardening foreign public opinion to US-made goods. Indeed, overall US exports have held steady amidst the trade war; it’s exports to a key subset of trading partners that have tanked—gross exports to Canada have dropped by nearly $50B annualized, and gross exports to both China and Mexico have dropped by more than $25B.

By product, the largest hit to US exports came in the vehicle sector, which saw a more-than-$20B-annualized drop compared to last year. For context, that is the single-largest drop in US car exports outside of the initial onset of COVID-19. Foreigners are proving less likely to want American vehicles, and US auto manufacturers are looking to redirect as much production as possible towards domestic demand in order to minimize their need to import. Expect even more volatility in America’s vehicle trade soon, as the effects of car parts tariffs will begin showing up next month, and Trump has recently been mulling over further increases to US auto tariffs.

Subscribe now

Conclusions

Today, no other wealthy country has tariffs as high as America, no country has experienced a comparably rapid decrease in trade openness, and no country has ever changed policy course so frequently. Trump’s first-term trade war, which looks masterfully planned and incredibly minor in comparison, still imposed significant costs on the US economy. It is no wonder that his second-term trade war has proven so destructive.

Over the last two months, Trump has at least managed to avoid any tariff escalations on the scale of early April, but he has also not solidified any major tariff reductions. More than two-thirds of the way through Trump’s 90-day tariff pause for countries besides China, the US has only been able to ink one small nonbinding trade “deal” with the UK. The next few weeks will provide crucial signals for how Trump wants to operate moving forward—will he kick the can down the road for another 90 days, make some face-saving reductions in overall tariff rates, or reimpose much of his April 2nd tariff threats? Either way, Trump has already done unprecedented damage to US trade relationships, and a high-tariff America remains the new normal.

Subscribe now

By subscribing, you’ll join over 65,000 people who read Apricitas weekly!

In any other administration, the announcement of 25% tariffs on cars & parts would be the single-largest economic story of the year—they currently hit more than $353B in US imports, having a larger economic effect than all of the tariffs implemented during Trump’s first term combined. These tariffs primarily affect imports from close American allies like the EU, Japan, & South Korea, who supply the majority of foreign-made cars to the United States. Yet the President won’t even spare the highly integrated North American supply chain, as tariffs currently apply to the non-US content in Mexican and Canadian-made vehicles.

Tariffs will also eventually be applied to the non-US content in Mexican/Canadian-made parts, which would increase the amount of imports affected by tariffs by another $68B when implemented. Plus, the Trump administration is looking to expand the trade war even further with tariffs on foreign-made heavy trucks and their parts, which could increase the amount of affected vehicle imports by roughly another $20B.

All of this was implemented in the maximally chaotic Trumpian way—only days before the tariffs on car parts were set to go into effect, the White House announced that they would not “stack” with other existing tariffs on steel, aluminum, China, or non-USMCA-compliant Canadian/Mexican goods. A steel car part from China could have faced additional tariffs of 70% (25% for being a car part, 25% for its steel content, and 20% for being from China) above the high preexisting tariffs from Trump’s first-term trade war, but now instead faces “only” a 25% increase.

At the same time, the administration announced a complex system of rebates for domestic carmakers to partially offset the tariffs on their input parts in the near term. Any company that manufactures vehicles in the US can claim a rebate worth up to 3.75% of the price of their cars, which is enough to offset the tariffs on components worth up to 15% of the car’s value. Next year, that rebate shrinks to 2.5%, and afterwards it totally disappears.

Finally, even though both executive orders mentioned that the US would eventually be applying tariffs to the Mexican & Canadian content of USMCA-compliant parts, no process to assess those tariffs or schedule for their implementation has been announced. A fact sheet from the White House seems to suggest the new rebate is designed to offset the cost of tariffs on non-USMCA-compliant auto parts, implying the tariffs on USMCA-compliant parts won’t be implemented for at least two years—but this is no hard guarantee and the administration could easily change their minds again at any point. The USMCA is set to be renegotiated next year anyway, so tariffs on parts from America’s neighbors could still change substantially.

Then, Trump signed his first preliminary trade “deal” with the United Kingdom, where the largest US tariff reduction came from a new tariff-rate-quota system for British vehicles. All UK cars still face tariffs, but the first 100k exports face only a 10% tariff while any additional exports face the full 25%. This is not a consequential change to vehicle tariffs, as the UK only makes up roughly 4% of American car imports and is heavily concentrated in the luxury segment, but this “deal” will likely serve as a model for negotiations with larger trade partners.

Still, roughly half of the cars bought in America are foreign-assembled vehicles that will be directly affected by the tariffs, and it’ll be impossible for even US-made vehicles to completely avoid being impacted. That’s purely because there is no such thing as an “all-American” car—among the most popular models last year, none are more than 70% American when you account for the embedded value of foreign parts, and most contain even more imported components than that. Bestselling iconic vehicles like the Ford F-150 are less than a third American. Even this data overstates it—the American and Canadian car industries are so tightly intertwined that the US government doesn’t even try to parse their contributions apart, yet Trump still plows forward with tariffs against his northern neighbor.

That disruption will also hit at a particularly difficult time for the US auto industry. Prices for new and used vehicles are still stubbornly high years after the inflation surge of 2021. Auto loan delinquencies continue rising as consumers struggle to keep up with payments amidst the high cost of living. Domestic vehicle production remains weak, and automakers have shed more than twenty thousand jobs over the last year. People who could afford to were already rushing to buy up cars before tariffs went into effect, reducing dealerships’ existing inventory. Plus, American industry has fallen technologically behind the rapidly growing Chinese EV makers from whom it already enjoys steep trade protection, and Trump is set on dismantling the Inflation Reduction Act policies that were at least attempting to catch them up. Companies like Ford and Subaru are already being forced to withdraw their annual guidance and hike prices for their foreign-made models. At this difficult moment, tariffs will make things worse by adding $6400 to the average price of a new car, increasing the costs of assembling vehicles in the US, and encouraging technological complacency among domestic automakers.

Breaking Down the Tariffs on Vehicles

Read more

Subscribe now

By subscribing, you’ll join over 65,000 people who read Apricitas weekly!

Last week, Donald Trump made the single-largest walkback of his trade war so far by agreeing to lower the tariffs on most Chinese goods by 115% as part of a 90-day “pause” for negotiations. Accounting for exceptions and sector-specific tariffs, the effective US tariff rate on China has dipped 74 percentage points, essentially bringing US policy back to the status quo of April 5th. Chinese imports now “only” face the 20% China-specific tariffs implemented in February/March, plus the 10% April 2nd tariffs imposed on nearly-all overseas imports, plus the Section 301 China tariffs that have persisted since Trump’s first term, plus sectoral tariffs on all imports of steel/aluminum/car parts/etc. In return, the US got no major concessions from the Chinese—their tariffs on US exports will come down, but remain 10% higher than pre-April, and Xi Jinping didn’t even have to symbolically lower tariffs on beef and ethanol like the UK did.

This is the largest easing of trade restrictions since Trump first began his China trade war more than seven years ago, bringing tariffs on the PRC down from “near-total embargo” levels to merely “extremely prohibitive” levels. Indeed, tariffs are now low enough to allow a significant amount of US-China trade to flow, especially given that companies will certainly take advantage of the pause by stocking up on necessary imports. Still, even after this announcement, Trump has imposed more tariffs on China than during his entire first term, which “only” saw an effective 16 percentage point increase in tariff rates compared to the 28 percentage points added through May.

In total, this still leaves overall US tariff rates roughly 8 times higher than before his Presidency, and relative US protectionism has only declined to “highest since the Great Depression” territory. That’s before accounting for the bevy of sector-specific tariffs slated to come online over the next few months, any possible unpausing of the April 2nd country-level tariffs, and any further re-escalation with China. Indeed, Treasury Secretary Scott Bessent described current China tariffs as a “floor” and the 54% tariffs announced through April 2nd as a “ceiling,” implying that tariffs on China can also only go up from here. In other words, America continues to face its worst trade shock in a generation, even if Trump’s tariff retreat significantly reduces the risk of an imminent self-inflicted recession.

The most obvious question, then, is “what was even the point of all this?”. Right now, US tariffs are mostly where they would have been if Trump’s April 2nd announcement “only” imposed his campaign promise of 10% tariffs on almost all imports. Instead, we got a week of total chaos as importers scrambled to plan around tariffs as high as 50%, followed by a month of total chaos as it looked like the administration was trying to abruptly hard-decouple from China. Millions of businesses and consumers were wrecked because they had to pay a 145% tariff that was subsequently lifted with just as little warning as it was imposed.

Bessent retroactively framed this episode as just “an unfortunate escalation”, but at every step of the process, it was the White House choosing to instigate and escalate. The “fire, aim, ready” strategy turned already-bad policy into a completely untenable disaster, and the shock-and-awe approach only served to cloak the fact that the administration still lacks coherent trade policy goals. Indeed, Trump has been completely unable to keep tariffs stationary for more than a few weeks at a time, and given his unwillingness to commit to anything but repeated “pauses”, it looks like that trend will continue over the near future. Tariffs on pharmaceuticals are supposedly slated for the next few weeks, the White House has promised to provide “revised” versions of the April 2nd tariff rates to most countries before July, and negotiations with China will likely result in even more trade restrictions between the world’s two largest economies. Trump is far from done trying to wreck America’s position in the international trade system.

Subscribe now

What Scared Scott Bessent

If the US chose to rush headlong into this trade war with 145% tariffs on China, what pushed them to retreat so quickly? Core to the underlying rationale behind the trade war was the theory that China is so export-dependent and currently has such weak consumer demand that they, not Americans, would have to bare most of the costs of US tariffs. The US is by far the largest consumer market in the world and a massive net importer—the administration thus believed American tariffs would force Chinese manufacturers to lower their prices to maintain market access given the lack of alternative export markets of comparable size and wealth.

Yet the evidence has shown the exact opposite—Americans were bearing almost the entire price increases caused by Trump’s China tariffs. In February and March, as tariffs on all Chinese goods rose by 20%, the pre-tariff cost of Chinese imports only dipped by 0.6%. In April, when the effective tariff rate on all Chinese goods rose by an effective 74%, pre-tariff prices only fell by 0.1%. In other words, Americans were paying nearly double for Chinese imports on average—and in key sectors like chemicals, furniture, plastics, and more, they were absorbing a full 145% increase. Chinese exporters were barely lowering their prices at all to “accommodate” Trump’s tariffs.

The other scary data point was on trade volumes themselves, which were shrinking dramatically in response to the tariffs. In April, Chinese exports to the US dropped more than 20% compared to the same time last year, even as overall Chinese exports continued to grow rapidly. Without some form of tariff reduction, businesses and consumers would’ve faced acute shortages in just a couple of months.

Subscribe now

Where the China Pause Leaves Us

Indeed, for nearly $200B worth of goods, China represents more than 70% of US imports, meaning hundreds of items would be practically unobtainable for Americans if 145% tariffs persisted. Key products like smartphones and computers had been previously exempted from those ultra-high tariffs after lobbying from firms like Apple & NVIDIA, but many goods like batteries, game consoles, microphones, and more still faced 145% tariff rates. The 90-day pause brings effective tariffs down precipitously for those key products where the US is most dependent on Chinese imports, such as toys, plastic housewares, electric heaters, and much more.

Still, overall tariffs remain extremely high for many US imports. Cars face a 19% effective tariff, pickup/delivery trucks a 10% tariff, vehicle parts a 10% tariff, and rubber tires an 18% tariff from the various motor vehicle tariffs imposed by the Trump administration. Batteries, toys, phones, computers, and other goods commonly imported from China still face heavy tariffs of up to 30%. Plus, there are still the 25% tariffs for steel & aluminum imports and the 10% baseline tariff on all overseas imports.

Broken down by country, Mexico and Canada are the major trade partners currently facing the lowest increase in US tariffs, since most of their goods are tariff-free under an exemption for exports compliant with the USMCA trade deal. That arrangement was announced as just a one-month pause back in March, but the Trump administration has still not removed the exemption a full two-and-a-half months later, and has functionally stopped talking about the issue. Meanwhile, the only countries facing similarly low tariffs are African, Middle Eastern, and South American countries who mainly export raw minerals exempt from the April 2nd tariffs. The bulk of countries instead face new tariffs near the 10% baseline level, but key trading partners like Germany, Japan, & South Korea face even higher tariffs given their large amount of car exports. Then, of course, China still faces the highest effective tariff increase of any nation by far at 28 percentage points.

Subscribe now

Conclusions

The 90-day pause is perhaps further attempts to negotiate a “soft-decoupling” between the US & China rather than a “hard-decoupling.” Originally, American officials rushed to soften tariffs by exempting major consumer goods like iPhones & laptops as well as key electronics components from tariffs, while the Chinese exempted the high-end US semiconductors and equipment used for its AI/chipmaking industries—expect more of these sector-specific tariff changes as part of negotiations. If American officials are dead-set on ending economic “reliance” on imports of “strategic necessities” from China, the next 90 days will be spent hammering out exactly which goods America will re-hit with heavy tariffs. Of course, this could have just as easily been done at any point before April 2nd, but alas.

What’s the end goal? US representatives keep reporting that they’re looking at a multi-year project to get China to an agreement modeled on the “phase one” deal from the end of Trump’s first term, essentially asking Beijing for more commitments to buy US goods in exchange for normalized trade relations. This strategy did not work last time, as China reneged on their promises rather quickly amidst the economic shock of the COVID pandemic, and the US never lowered tariffs on Chinese goods by much. It’s hard to understand how an even more slapdash version of the same strategy will achieve results this time.

It’s even harder to understand what sort of arrangement the Trump administration wants with countries that are not China. In recent days, officials have been talking less about individual country “deals” and more about regional negotiations and another possible round of unilateral changes in tariff rates. That’s a far cry from the “90 deals in 90 days” rhetoric previously coming from the administration, and is a tacit admission that the April 2nd tariff levels are not a rational baseline for policy. For the few major partners the US still wants to negotiate with, the UK “deal” at least provides some template: the vast majority of tariffs will stay, but some tariffs can be reduced in exchange for agreements to restrain exports, nebulous purchase commitments, and symbolic reductions of tariffs on US goods. Still, we’re nearly halfway to when the 90-day pause on April 2nd tariffs is up, and the only “deal” struck has been that extremely tiny one with the UK.

Any one of Trump’s current trade policies—a 30% tariff on China, a 25% tariff on cars, a 10% tariff on the European Union—on their own would have been considered the most brutal trade war in a generation. Instead, Trump has pursued all of them, combined, all at once, and in the most chaotic way possible. Even if he walked his trade war completely back to “just” the 10% baseline tariff that is apparently non-negotiable, it would leave tariff rates at the highest level since the 1940s and would make America the world’s only high-income nation with double-digit tariff rates. Trump may be easing his trade war, but he is far from ending it.

Subscribe now

By subscribing you’ll join over 65,000 people who read Apricitas weekly!

Last week, British negotiators signed a historic, comprehensive trade agreement with a major global economy—in India. They also inked a tiny, slapdash, informal farce of a trade “deal” with the United States.

The US-UK “deal” is the first and only one the US has negotiated since Trump paused tariffs above 10% on countries besides China for 90 days. In the short text of the agreement, the only actual tariff reduction America formally commits to is reducing the 25% car tariff to 10% for the first 100k vehicles exported from Great Britain. Washington promises to create a similar system for British steel & aluminum, but provided exactly zero details on what this actually means. In the press conference, US Commerce Secretary Howard Lutnick also promised that aircraft engines & parts would be exempted from US tariffs, but this was a lie—the official document does not even mention planes. In “exchange”, the US gets some tiny tariff-free quota for beef and ethanol exports. In practice, this means trade policy between the two countries is still several times more restrictive than before Trump took office.

To say this “deal” is tiny would be an understatement—it is microscopic. The United Kingdom is a relatively small US trading partner, only representing about 2% of American goods imports (that’s less than Vietnam, Taiwan, Ireland, India, or Italy), and this deal only reduced tariffs on about 14% of those British goods. The overall impact on US trade restrictiveness is functionally unnoticeable—the Trump administration’s quiet decision to indefinitely delay tariffs on Mexican/Canadian auto parts was roughly 5-6x more impactful. It would not be inaccurate to characterize this entire episode as basically a PR tour for slightly lower tariffs on Aston Martins & Bentleys.

If this is all that can be achieved with the UK, a close NATO ally who runs a trade deficit with the US and has long been desperate to sign a real free trade agreement post-Brexit, it does not portend well for other trade negotiations. Even assuming all non-China countries can achieve the same proportional reduction in their April 2nd tariffs as the UK, the average US tariff rate would still be 34% when the 90-day pause lapses in July. Keep in mind that the US would have to “negotiate” more than a dozen separate “deals” over the next sixty days to approximate this result, even if they only focused on major trading partners. In the meantime, the US keeps drawing up more tariffs—just in the last month, it has begun plans for taxes on imported trucks & aircraft worth a combined ~$60B.

Still, the contours of this UK-US “deal” are at least instructive for how the US will be approaching trade policy in the near future. The administration is clearly dead-set on keeping tariffs of at least 10% for most goods, hence why they refused to exempt British cars from tariffs entirely. Yet there is some more flexibility on the sector-specific tariffs than initially indicated—and after seeing tariff reduction for British cars, other major vehicle exporters will certainly be demanding similar concessions. The US is also clearly trying to force countries to join an anti-China trade bloc in key sectors like steel & aluminum to get relief from US tariffs, hoping to push them to impose their own restrictions on Chinese exports. And it is focusing on securing cheap & easy symbolic “wins” for agricultural & energy exporters. Still, overall US tariffs are not being lowered in any meaningful way.

Subscribe now

Details on the Deal

The only concrete tariff reduction America agreed to in this “deal” with the UK was a tariff-rate-quota system for British cars. This means the first 100k vehicles exported to the US will face “only” a 10% tariff, and all subsequent imports will face the full 25% tariffs previously imposed. That would lower tariffs on the vast majority of UK car exports, as Americans only bought 110k British vehicles in 2024, but it is also clearly designed to guarantee exports remain below their current relatively low levels.

Indeed, the United Kingdom only makes up about 4% of US vehicle imports, behind the EU, Japan, South Korea, Mexico, and Canada, and is heavily concentrated in the high-end segment—the main vehicles imported are Land Rovers, Bentleys, and BMWs, plus some Mini Coopers. The idea of these foreign luxury cars permanently facing lower tariffs than the more middle-class and American-made offerings of GM, Kia, Honda, etc is obviously untenable both to American car companies and to other foreign governments that Trump wants to negotiate with. The UK “deal” should thus be understood as an example of the White House’s end goal in vehicle tariff negotiations—a lower but nonzero tariff rate, as long as vehicle exports to the US are forced to decrease.

Beyond the reduction in car tariffs, the closest tariff reduction from the US side was a promise to develop a tariff-rate-quota system for British steel & aluminum once the UK meets American demands for “the security of the supply chains” and “ownership of relevant production facilities.” This is a clear broadside at China, using tariffs as negotiating tools to shut them out of other export markets and force them to divest from foreign steel & aluminum manufacturing assets.

More broadly, the agreement emphasizes that both countries should aggressively enforce rules-of-origin to “prevent non-participants from using our bilateral arrangement to circumvent tariffs” in another clear effort to prevent Chinese goods from being shipped through the UK to the US. Like with the car tariffs, this should be interpreted as an example of the administration’s end goal in other steel & aluminum trade negotiations—“we’ll lower tariffs on key metals, so long as you agree to shut out China.”

It’s unclear how comprehensive this steel and aluminum tariff-rate-quota system would be, but Downing Street interpreted it as a 0% tariff on all British exports. Given how small the British metals industry is and the fact that total tariff-affected imports are less than $1B, it would be fairly trivial for the US to give the UK a quota large enough to functionally exempt it from tariffs. However, again, the document did not give any concrete details on how this agreement will actually take shape.

Finally, the US also makes nebulous promises that the UK will be first in line to negotiate when America imposes long-threatened tariffs on major sectors like electronics, lumber, semiconductors, and more. Particularly important is the White House’s promise “to promptly negotiate significantly preferential treatment outcomes on pharmaceuticals”, as drugs are the UK’s second-largest US-bound export behind cars. Still, pharmaceuticals are currently exempt from all tariffs, so even if the UK receives preferential treatment after tariffs are imposed, its drug industry will still face much higher trade barriers than it does today. This entire “deal” offers no pathway to actually normalize trade relations between the US and one of its closest allies.

Subscribe now

Conclusions

At the current pace, Trump will only be able to sign three of these “deals” before the 90-day tariff pause ends in early July. If the White House tried hard, they could likely accelerate this pace by cobbling together similar slapdash “deals” with small trading partners like India, Australia, or Israel. However, this is clearly not a tenable long-term strategy, as there seems to be no coherent plan for negotiations with America’s major trading partners like the EU, South Korea, or Japan—let alone China, who still faces tariffs of 145% on most goods.

Yet this whole episode is most reminiscent of Trump’s “agreements” with Mexico and Canada from the start of this year. At first, there was an attempt at the theater of negotiations (remember when Canada appointed a fentanyl czar to secure a one-month delay in tariffs?), but even this theater was dropped rather quickly, and Trump just began making unilateral increases and decreases in tariffs. If the President is going to back down from the massive tariffs set to be re-imposed in 60 days, it will likely be via face-saving unilateral retreat, not an avalanche of “deals” like this—and no matter what, tariffs will remain significantly higher than when he first took office.

Subscribe now

By subscribing you’ll join over 60,000 people who read Apricitas!

The US economy shrank at the beginning of 2025 for the first time since the recession scares of early 2022—GDP contracted at a 0.3% annualized rate, a major downgrade from the 2.4% growth registered at the end of last year. The culprit was the massive economic drag from trade wars, with companies and consumers forgoing business-as-usual to stock up on foreign goods before Trump’s massive tariffs took effect. The trade deficit expanded to record highs in Q1, with investment jumping as American businesses stored or installed their imported goods.

In fact, the trade deficit expanded at such a rapid pace that net exports notched their largest quarterly decline since records began in 1947. Now, it’s important to clarify that this import surge does not itself shrink the economy—GDP is an estimate of domestic production, and is not affected by purchases from abroad. Yet domestic production is difficult to measure directly, so GDP is estimated by summing consumption, investment, government output, and exports while subtracting imports only to prevent double-counting. Thus, consumers rushing out to buy foreign-made clothes will show up as an increase in imports (a negative “contribution” to GDP) and an increase in consumption (an equal-and-opposite positive “contribution” to GDP) and have no net effect on economic growth.

In this case, however, the surge in imports primarily manifested as a surge in business inventories—it was corporations, not consumers, doing most of the stockpiling. The positive contribution from inventory growth was the highest since late 2021, “offsetting” a bit less than half the jump in the trade deficit. Fixed investment of foreign equipment also helped significantly “offset” the increase in imports at the start of this year.

Yet it’s also important to say that the costs and uncertainty of trade policy chaos can have real negative growth effects. If a construction company rushed to buy a bunch of wood from Canada ahead of tariffs, that would show up as an increase in imports and an increase in inventories which cancel each other out. But if that company slowed down construction so it could afford to stockpile inputs, that would show up as a hidden drag on GDP in the form of lower investment. In other words, tariff front-running can indirectly slow the economy, and that’s exactly what we saw at the start of this year.

Subscribe now

Breaking Down the Import Surge

Real imports of goods shot up to a record high at the start of 2025, rising nearly 11% from the end of 2024. The only recent trade shifts of a similar scale are the COVID pandemic and the 2008 recession, yet while those events caused unprecedented contractions in trade, the threat of tariffs has done the exact inverse and caused trade to surge at a speed not seen in 50 years.

Computers & pharmaceuticals were the two sectors behind the vast majority of that record import surge—outside those industries, the jump in imports was noticeable but not record-breaking. Pharmaceuticals, an industry Trump has long promised to hit with tariffs on but has exempted from nearly all trade policy actions so far, saw their real imports jump 60% from their previous record high. On the other hand, computers were not specifically targeted by Trump in the same way as pharmaceuticals during Q1, but they were heavily dependent on a Chinese & broader East-Asian supply chain that risked heavy disruption from tariffs. Real imports of laptops, desktops, and other computers thus rose by more than 50% in Q1 2025 as companies tried to prep for disruptions to their complex supply chains. Where did all those drugs and electronics go?

The pharmaceuticals were mostly stockpiled by medical companies—drug wholesalers’ inventories rose by the highest amount ever recorded, beating the previous record-high by a factor of 4. In fact, the rise in pharmaceutical inventories was so large that it was single-handedly responsible for nearly 40% of the total increase in inventories across the entire US economy. This is downstream of both genuine stockpiling of the range of medicines made only outside America and a rush by drug multinationals who use transfer pricing schemes to book as much profit as possible in Ireland, Switzerland, Singapore, and other low-tax jurisdictions before the tariffs kick in. This record pharmaceutical stockpiling is also far from over, as preliminary data from the EU countries responsible for most US drug imports shows even larger increases on the near horizon.

By contrast, the record-high imports of computers were mostly driven by rising investment at AI labs and other tech companies, not by business inventory growth or consumer electronics spending. Fixed investment in computers & peripheral equipment rose 21% in Q1 to a new record high as firms continue investing more and more in computing capabilities. In fact, computers & broader information processing equipment were responsible for more than 70% of the increase in total fixed investment this quarter. This import surge is also unlikely to slow anytime soon—even though Trump has promised tariffs on computers, semiconductors, and electronics at some indeterminate point in the future, at the moment most computers used for AI training come from either Taiwan or Mexico and remain totally exempt from tariffs.

Subscribe now

The Other Bits of GDP

Outside of the rapid shifts in trade itself, consumption growth also changed considerably at the start of this year. Spending on motor vehicles and parts had their worst quarter since mid-2021, mostly as a come-down from the rapid growth at the end of 2024. This was one area where tariff front-running was extremely early, with consumers rushing to buy cars before Trump even took office. This led to muted car sales in January/February, only partially offset by the rapid increase in sales after Trump officially announced auto tariffs in March. Meanwhile, spending on pet supplies, toys, cosmetics, clothes, and other miscellaneous non-durables kept growth in that category relatively normal. Finally, services consumption had the smallest contribution to GDP growth in a year and a half thanks to slowdowns in healthcare and food services spending.

DOGE cuts to federal employment and grantmaking also shrank public-sector output in Q1 for the first time since mid-2022, with both major categories of federal activity contracting. Defense output had its worst quarter in years amidst cuts to spending on civilian employees and miscellaneous military supplies, while nondefense output also shrank due to employment cuts. Finally, cuts to federal education and health grants also dragged heavily on state & local activity.

Yet even amidst all of this chaos, “core” American economic growth has held up relatively normally. Real final sales to private domestic purchasers—a mouthful of a term that just represents the sum of private-sector consumption & fixed investment—grew at a standard rate of 3%. These metrics are not wholly untainted by the effects of trade policy (the increase in fixed investment is mostly from tariff-driven computer imports, for example), but provide evidence that the economy has not yet entered a recessionary spiral. Going forward, volatility in overall GDP will be extra high given the wild swings in imports and inventories, so keeping an eye on the health of these “core” metrics will be critical.

Subscribe now

Conclusions

This data does not predate all of Trump’s tariffs—taxes on Chinese imports were already 20% by the end of Q1, tariffs on non-USMCA-compliant imports from Mexico/Canada were 25%, and the steel & aluminum tariffs had also gone into effect—but it does predate the entirety of the massive April 2nd tariff hikes. In other words, the economic shifts seen in these numbers are mostly from the threat and uncertainty surrounding tariffs, not from the actual tariffs themselves. Trade policy shifts will continue dominating the US economy in the short run, and the tariffs will serve as an even larger drag on growth as they start to truly bite.

Subscribe now

By subscribing you’ll join over 60,000 people who read Apricitas!

When Donald Trump imposed sweeping tariffs on almost all overseas imports earlier this month, he also left detailed clues on his next trade war moves.

The country-level tariffs announced on April 2nd exempted automobiles, car parts, steel, and aluminum because those goods were already hit with 25% universal tariffs, but the White House also published a large 37-page list of other exemptions for everything from sawdust to gold bullion to potassium chloride. The administration categorized those imports as “copper, pharmaceuticals, semiconductors, lumber articles, certain critical minerals, and energy and energy products”—and followed up days later to add specific exemptions for phones, computers, TVs, and chipmaking machinery. Those goods (with the exception of energy) map directly onto the categories of products Trump has mentioned as targets for separate 25% universal tariffs.

In effect, they have revealed the precise list of goods they plan to hit with future tariffs.

These tariffs will likely be authorized under Section 232 of the 1962 Trade Expansion Act, just like the automobile, steel, and aluminum tariffs already in place. This law allows the President to restrict imports of select goods if they impair US national security, yet the President is only granted these powers after the Department of Commerce conducts an investigation and determines that these imports actually do threaten national security. All the Section 232 tariffs imposed by Trump this year have cited investigations done during his first term, so he would need Commerce to complete new investigations before imposing tariffs on, say, critical minerals. That’s why the administration has already kick-started its investigations into the “threats” posed by imports of lumber, copper, pharmaceuticals, and semiconductors.

However, the fact that Trump has openly posted the list of goods to be hit with tariffs is a stunning revelation. In effect, it admits that the Department of Commerce will be conducting sham investigations to arrive at predetermined results that say exactly what the President wants them to say. Presidents have always had lots of influence over the 232 process—after all, they’re usually the ones who call for investigations—but experts at Commerce’s Bureau of Industry and Security are supposed to determine whether imports actually pose a national security threat while recommending exactly which goods should be hit with tariffs and at what levels. Given that the Trump administration has already publicly listed all the goods they want to hit with tariffs and repeatedly centered a tariff rate of 25%, there will be no legitimate investigation. It’s hard to take seriously the claim that imports of children’s picture books, the artificial sweetener sucralose, and dried coral impair US national security, yet all of these goods were exempted from the April 2nd tariffs and thus will likely be deemed threats by Commerce so they can be hit by Section 232 tariffs.

Combined, the targeted products represent roughly $670B in imports in 2024, and if Trump imposed his preferred 25% tariff rate on all of them, it would be equivalent to a roughly 5 percentage point increase in the overall US tariff rate. That is larger, in total impact, than the entirety of Trump’s first-term trade war or the 10% tariffs currently in place against most countries besides China.

Plus, they could go even further—Trump claims he wants to tariff the “whole electronics supply chain” in addition to the previously mentioned semiconductor tariffs, so he could hit a wider range of devices in addition to computers & phones. The April 2nd tariff executive order also exempts “all articles that may become subject to duties pursuant to future actions under Section 232” in addition to the specific copper, semiconductor, pharmaceutical, and lumber exemptions, which implies there are even more universal tariffs under consideration. Trump is far from done with his global trade war.

Breaking Down Trump’s Upcoming Tariffs

Read more

Subscribe now

By subscribing you’ll join over 60,000 people who read Apricitas!

Last week, Donald Trump announced the largest tariff package in modern American history—a 10% universal baseline tariff on imports from every country except Canada & Mexico plus additional country-specific tariffs that go as high as 40%. The 10% tariffs went into effect last Saturday, making for the largest single-day tariff hike in modern US history and bringing effective tariff rates to the highest level in nearly a century. Those additional country-level tariffs came into force three days ago, completely shattering the record set just a few days before, but were immediately paused when financial markets puked in response to their implementation.

Trump claimed those now-paused country-level tariffs were “reciprocal” and that the US is only hitting back at other nations’ tariffs and trade barriers, but that was a blatant lie—the numbers justifying these tariffs were fabricated from calculations based on the US bilateral goods trade deficit. Free-trade zealot Switzerland got a higher “reciprocal” tariff than heavy-tariff-user Argentina simply because America’s trade deficit with Switzerland is larger than with Argentina. More importantly, several of the largest American trading partners got hit with massive tariffs—Vietnam got 46%, Taiwan got 32%, South Korea got 25%, Japan got 24%, and the EU got 20%—before the pause went into effect. Imports from China, already hit by 20% universal tariffs plus the preexisting tariffs from Trump’s first-term trade war, got an additional 34% tariff.

That wasn’t the end of it—tariffs on China went up an additional 91% beyond the April 2nd tariff announcement after two successive rounds of retaliation, reaching 145% on most goods. This is the end of most direct trade between the world’s two largest economies—outside of the categories facing only 20% tariffs, it would almost always be more economical to source from any other country, produce domestically, or simply not buy goods outright than try to pay tariffs this high. In total, America has just blockaded itself from all overseas trade and completely severed the majority of US-China economic links in the single largest negative shock to global supply chains since the early pandemic.

Subscribe now

Contextualizing the Tariff

It should go without saying that this is an insane way to run the world’s largest economy. Imagine being a US solar panel importer, panicking because most of your goods come from Southeast Asian nations like Vietnam and Thailand threatened with insane tariffs. After April 2nd, you rush out to import as much as you can before a 36% effective tariff increase takes effect April 9th, only to have the tariffs reduced by 26% the next day. Or imagine being an electronics retailer, already facing tariffs of 16% on imported game consoles that mostly come from China. You brace for an expected 32% tariff increase based on the April 2nd tariff rates, instead receive a 61% tariff hike because China tariffs were increased further, and then receive another 27% tariff hike the next day even when most other imports saw tariff relief. Or imagine being that same electronics retailer, panicking about the effects of tariffs on your cell phone imports, watching tariffs accelerate up to 65% amid the China trade war, only for the White House to quietly publish an addendum yesterday exempting smartphones, computers, and parts from all of the tariffs imposed since April 2nd.

Yet it’s also just worth mentioning the sheer size of Trump’s tariff hikes in fiscal terms—they are among the largest peacetime tax hikes in American history. Analysis from the Tax Foundation estimates that total tariffs will increase federal revenues by about 0.56% of US GDP if sustained, and the April 2nd announcement alone would have increased revenues by 0.45%. For context, the passage of the Affordable Care Act in 2010 caused a smaller net increase in tax revenues than Trump did last week. If a more sensible tax like a VAT or income tax was implemented with similar scale, suddenness, and volatility, it would still be a major shock to the macroeconomy—but Trump has manifested this entirely through an unprecedented increase in some of the most economically distortive taxes in existence. Plus, these revenue estimates understate the impact of the tariff hikes—taxes on Chinese imports are now so large that the US will collect vanishingly little revenue from them given the likely total collapse in trade volumes.

The cumulative impacts of all of Trump’s tariff moves thus far has been an unprecedented rise in the cost for many foreign-made goods. Motor vehicles, America’s single-largest manufactured import, currently face a 21% effective tariff rate once exemptions for US content are taken into account. Electronics like batteries, game consoles, audio equipment, and more now face massive tariff rates, given their concentration in China. In fact, many goods—toys, plastic kitchenware, party supplies, etc—are so concentrated in China that they now suddenly face functional tariffs in excess of 100%, which will cause chaos for their supply chains.

Meanwhile, the 10% tariffs hit bananas, cacao, palm oil, and dozens of other raw agricultural products that simply don’t grow in American climates, hurting food manufacturers and grocery shoppers. They hit dozens of raw inputs, parts, and components that feed US factories, hurting American manufacturing output. They hit most of the capital goods, construction materials, and equipment necessary for reindustrialization, making it harder to build factories, data centers, or housing in America. The unending tariff uncertainty continues to worsen consumer confidence and business expectations.

This whole episode is a continuing, unnecessary economic disaster—why use a fake formula? Why hit small countries like Vanuatu and Lesotho? Why exempt gold and laptops but not tropical agricultural staples like coffee and bananas? Why give only a few days’ warning between announcement and implementation? Why unilaterally break all of America’s free trade agreements? Why pursue tariffs this large to even begin with? US financial markets have fallen in response to both the direct negative economic effects of these taxes and the now-obvious degradation of US governing institutions.

Trump, for his part, is far from done raising tariffs—he is already lining up more universal tariffs on pharmaceuticals, lumber, semiconductors, critical minerals, and additional goods that were exempted from this round of tariffs. After 90 days, he could reimplement some or all of the now-paused tariffs just as he did against Canada and Mexico earlier this year. Plus, he could accelerate the trade war against the EU and Canada by responding to their retaliatory tariffs. The day after reducing tariffs on most countries, he was back to threatening Mexico with tariffs again. America is already at the highest tariff rate in more than a century, and the chaos is unlikely to end.

Subscribe now

Subscribe now

By subscribing you’ll join over 60,000 people who read Apricitas!

This afternoon, Donald Trump is set to implement possibly the single-largest tariff hike in modern American history—and we know functionally zero details about it. Reporting indicates the White House team was still cooking up possible options for the president as late as yesterday, with basic information like the tariff rates and countries affected still up in the air. These tariffs will supposedly go into effect “immediately”, likely having the same midnight-on-April 3rd implementation date as Trump’s automobile tariffs, thus leaving businesses zero time to digest the information and formally prepare. This chaotic approach has sent measures of trade policy uncertainty skyrocketing to record high levels—and that alone imposes significant economic costs by making spending, investment, hiring, and transport decisions harder.

Indeed, unclear and belligerent trade policy has already terrified American consumers, who are increasingly scared about tariffs’ impact on prices and the wider economy. Inflation expectations have risen dramatically in the University of Michigan’s national survey, with long-run expectations jumping to the highest levels in more than 30 years. Variance in inflation expectations has also skyrocketed, underscoring the profound uncertainty most people have about the near future. A record high share of respondents are complaining about government economic policy, mostly citing the negative impact of tariffs, and a record high share are also expecting business conditions to get worse over the next year. Survey data from The Conference Board showed households’ forward-looking economic expectations at a 12-year low, and most Americans are now planning to cut back on vacations and other leisure spending.

The one place consumers aren’t cutting back is in durable goods—in fact, it is here where Americans are rushing out to buy things before tariffs go into place. The share of consumers saying you should make large purchases now because prices will go up soon has jumped since the election, and those with the means to do are trying to stockpile what they can. The car industry has been one of the sectors most affected by the rush to get ahead of tariffs, with US vehicle sales jumping 12.7% in March as Trump announced universal 25% tariffs on motor vehicles & parts.

Consumers and businesses alike have also rushed to import goods before Trump imposes more tariffs, sending the US goods trade deficit skyrocketing. Much of that deficit surge is composed of a massive increase in gold imports from financiers, but it also includes a significant rise in imports of consumer goods (up 24% from this time last year) and capital goods (up 18%). In January, imports of computers were up 57% from last year and imports of phones were up 33%. This panic buying certainly continued through March, and given the continued tariff back-and-forth will also likely extend into April.

Behind closed doors in Congress, Treasury Secretary Bessent reportedly claimed today’s tariffs are a “cap” and that rates would only go lower, not higher, from there. Yet that is a cold comfort given how high tariff rates are likely to go—and also difficult to believe given that Trump likes escalatory tariff retaliation cycles, has promised further tariffs on key sectors like lumber, copper, pharmaceuticals, or semiconductors, and keeps threatening more tariffs like those on buyers of Russian, Iranian, or Venezuelan oil. Even if it is a cap, elevated uncertainty could persist for an extended period as exclusions and reductions are added on a country-by-country or product-by-product basis. This maximum-attention and maximum-uncertainty approach is likely to continue—and it is certain to amplify the negative costs of the tariffs Trump is implementing today.

Subscribe now

The Business End of Uncertainty

Economic optimism among businesses both big and small has been falling for months as Trump’s tariff ambitions grow. Among CFOs surveyed by the Richmond Federal Reserve, 30.5% named tariffs as a pressing concern in Q1, up from only 8.3% at the end of last year, and concerns about uncertainty doubled over the same timeframe. The shift in trade policy is also by far-and-away Trump’s least popular move among businesses, with a sizeable share of CFOs now saying changes in tariffs have caused them to cut their hiring and investment plans. Small businesses, too, are growing increasingly worried, with 64% of them now expecting to be negatively impacted by tariffs and only 8% expecting to be positively impacted.

One of the most robust measures of business inflation expectations also shows rising uncertainty and cost concerns among American companies. The Atlanta Fed asks companies to project their own unit costs forward, and since businesses intimately know their input prices and make decisions based on them, this tends to be a more reliable way of forecasting prices than just asking people for their general inflation expectations. Through February, the Atlanta Fed’s data shows inflation expectations and uncertainty had both risen as tariff threats escalated. Perhaps most importantly, inflation expectations are diverging from experienced inflation for the first time since the early pandemic—the threat of tariffs has loomed so large that companies see current price relief as illusory and expect costs to reaccelerate soon.

The Dallas Fed’s survey of Texas manufacturers gives a more up-to-date glimpse at similar dynamics. These businesses, deeply intertwined with foreign supply chains in Mexico and beyond, have become incredibly frustrated by the administration’s constantly-shifting trade policy. They now expect faster cost inflation, worse business conditions, and slower investment growth than at the start of the year. More than half of the comments left on the survey vented about the negative impacts of tariffs.

Business uncertainty among Texas manufacturers is also rising at the highest rate since early COVID, with companies growing increasingly unsure of even their more-pessimistic economic outlook. One executive at a Texas computer manufacturing company put it succinctly: “how can you do business planning with all of this uncertainty and the daily changes in direction made by the Trump administration?”

Subscribe now

Conclusions

Today’s tariffs will be bad enough on their own, but had they been passed as a legislative package you could at least rest easy knowing the matter was settled for the time being. Instead, they’ll remain continually subject to change by the President while setting off a chain of retaliation and renegotiation that will keep global trade policy in flux for weeks or months. Any eventual resolution will also stand on shaky ground—Trump has repeatedly shown his willingness to violate formal free trade agreements, including those he has personally signed, so businesses and foreign countries will never 100% trust that his trade war ambitions are complete. The pervasive uncertainty that has undermined the economy over the last few months will increase before it decreases.

Subscribe now

By subscribing you’ll join over 58,000 people who read Apricitas!

For two months now, Donald Trump has been repeatedly threatening a massive trade war against both of America’s neighbors. In early February he signed an executive order placing 25% tariffs on all goods from Mexico and Canada, but backed down before the implementation date and agreed to a one-month pause after receiving token concessions. In early March he let those tariffs come into force for a few days, only to partially backtrack by announcing the tariffs would exclude all goods imported under the USMCA trade deal’s rules of origin until early April. In the meantime, he implemented 25% tariffs on all steel and aluminum entering the United States, products where Mexico and Canada are both key sources of imports, and he continually promises universal tariffs will actually go into effect starting next week alongside a suite of tariffs on most of America’s largest trading partners. Plus, he’s continually expanding the trade war using additional actions against major Canadian and Mexican exports like the automobile tariffs signed yesterday and upcoming tariffs on lumber, agricultural products, and more.

These tariffs’ economic costs will be disproportionately borne by American workers, companies, and businesses, but the US is an extremely large diversified economy with room to absorb shocks that would thoroughly devastate most other countries. By contrast, Canada and Mexico are much smaller by both population & GDP—even though the cost of the trade war is mostly absorbed by Americans, the share that falls on Canada and Mexico is larger relative to their economies. US trade with Canada and Mexico is 7% the size of its GDP, meanwhile Mexico and Canada’s trade with the US is 45% and 34% the size of their GDP, respectively, making them by far the most exposed countries to the effects of a US-led trade war. The Peterson Institute for International Economics thus estimates that the trade war would reduce Mexican and Canadian GDP by 4-5x as much as US GDP and would increase inflation by 3-4x as much. Officials at all levels of the Mexican and Canadian governments are drawing up plans to brace for a possible trade-war-driven recession.

The cruel irony is that both countries have spent the last decade-plus trying to make themselves the centerpiece of US “friend-shoring” efforts to move supply chains towards its close allies. When Trump’s first-term trade war against China erupted, it was factories in Mexico or Canada that worked to replace many Chinese-made goods. When Biden passed the Inflation Reduction Act, North American EVs and batteries could qualify for large tax credits not extended to the Europeans or Japanese. When he signed the CHIPS Act, Canada and Mexico rushed to find ways to partner on the semiconductor supply chain buildout. When the US used tariffs to functionally ban imports of Chinese EVs, it pressured Mexico and Canada to follow suit with tariffs of their own. Now, both countries are instead being treated as economic enemies, significantly worse than dozens of dictatorships where the US still has normal trade relations.

Unlike the first term, this is a Trump administration that believes in on-shoring or nothing, sees its neighbors’ manufacturing industries as undermining rather than assisting their own, and views tariffs as a means to cajole allies just as much as a means to punish enemies. Expanding the trade war like this is going to amplify its costs, undermine US growth, increase business uncertainty, hurt US allies, and make the administration’s stated goal of “strategic decoupling” from China even more difficult. If Trump continues down this path, it’ll be the death knell of the “friend-shoring” era and an incredibly costly move for economies throughout North America.

Subscribe now

How a Trade War Could Break Mexico & Canada

Right now, Trump’s tariffs on Canadian & Mexican goods only affect steel, aluminum, and products that do not enter under the rules required for preferential treatment under the USMCA deal signed during Trump’s first term. Roughly 62% of imports from Canada and 51% of imports from Mexico didn’t enter under those zero-tariff USMCA provisions in 2024, either because baseline US tariffs were already zero in that category or because USMCA compliance was too economically or bureaucratically onerous. That means that, at most, tariffs are currently affecting roughly $257B in imports from Canada and $259B from Mexico, bringing the total amount of imports currently affected by Trump’s tariffs on Canada, China, and Mexico down to a maximum of $952B.

Last year, the single-largest non-USMCA-compliant import was Canadian crude oil, followed by Mexican cars & parts, Canadian aluminum, and then mostly-Canadian industrial/aircraft turbines. Various baked goods, silver, and vehicle trailers could also be hit hard. In practice, however, the large majority of those imports likely just need to fill out paperwork to become USMCA-compliant—the Mexican Secretary of Economy said that 85-90% of total exports could qualify for the exclusion while the Canadian trade minister said that “virtually all” could be eligible. Yet it remains extremely unclear precisely how much is being exempted, and those estimates might be sugar-coated given that the USMCA does not have systems built out to accept many goods or that tens of thousands of corporations would have to move quickly to get a reprieve they will ostensibly lose in only a few days.

For Canada and Mexico, even these smaller tariffs are the largest trade shock since the pandemic, and the perpetual threat of universal tariffs looms extremely large. Much of the Canadian economy is directly or indirectly built around exports to the United States—as of the latest data available, Canadian value-add in US-bound exports made up roughly 17% of the country’s economy, boosted by the large southern-oriented manufacturing and energy sectors. Within manufacturing, the primary metals, motor vehicles, and wood product sectors were the three largest contributors—all three of which have or are expected to be hit with even higher additional tariffs.

For virtually all Canadian provinces, exports to the United States rival or exceeds exports to the rest of Canada combined. As a share of GDP, the oil-producing provinces of Alberta and Saskatchewan are most dependent on exports to the US, but even the largest Canadian provinces like Ontario or Quebec have much of their economy wrapped up in southbound trade. Indeed, they would almost certainly be hit harder given the lower tariff on oil exports and inflexibility of pipeline trade compared to sectors like auto manufacturing.

Even the threat of US tariffs is already having negative economic effects by amplifying fear and uncertainty. In surveys done by the Bank of Canada, Canucks are already bracing for the impact of trade tensions—consumers are trying to save more and spend less, businesses are cutting hiring and slowing investment, and both have raised their inflation expectations for the year. Employees in trade-dependent sectors like oil, gas, mining, and manufacturing are increasingly worried about losing their jobs.

However, Mexico would likely be hit even harder than Canada given that country’s economy is substantially smaller, more industrial, and more connected to the United States. Roughly one-in-six Mexican workers are in manufacturing, compared to one-in-eleven Canadian workers or one-in-twelve American workers. Plus, roughly 30% of all Mexican manufacturing employees work in export-oriented factories (aka Maquiladoras), the vast majority of which are built around shipping goods to the United States. Those workers are heavily concentrated in the auto sector, but are also spread through industries like computers, plastics, appliances, metals, and more.

Even now, the tariffs on non-USMCA-compliant goods are certainly hitting Mexico harder than Canada—while both countries exported roughly $250B to the US outside the trade deal’s rules in 2024, more of Mexico’s exports were manufactured products where all non-USMCA goods are tariffed and thus definitively can’t become eligible just by filling out paperwork. In 2024, $39B of US imports from Mexico were in these non-energy goods categories subject to tariffs compared to $10B in total imports from Canada. Those Mexican imports include $10.8B in vehicle parts, $7.4B in cars, and $1.6B in ICE engines, meaning the country’s large export-oriented auto sector has already been hardest hit even before the specific car-related tariffs kick in.

States along Mexico’s northern border are unsurprisingly going to be hit the hardest by any further trade war escalation—Chihuahua, Coahuila, and Baja California have the highest US-export-to-GDP ratio of any state in Mexico. Those northern states had also seen some of the fastest production and investment growth over the last few years as US-Mexico trade continued to grow.

Subscribe now

No Friends Ashore

The cruel irony of Trump’s tariffs threats is that both Mexico and Canada worked hard to support US industry amidst recent “nearshoring” and “friend-shoring” pushes—including during Trump’s first term. In the modern industrial policy era, the US has undertaken multiple escalating government interventions in trade and manufacturing sectors to support domestic industry but has generally accepted that no modern country can or should go it alone. Instead, it worked to onshore key sectors like semiconductors while mostly limiting itself to the already-arduous task of partial decoupling from China. Even though neither Canada nor Mexico were the largest beneficiaries of tariff-driven “trade-diversion” away from China, they did both play massively important roles in that reorientation of supply chains.

Take the computer manufacturing industry as an example—ten years ago, China was the dominant source of US computer imports by a wide margin, sending three times as much as Mexico and exceeding all other countries combined. During Trump’s first-term trade war, he put large tariffs on Chinese desktops, storage, and other components, only sparing laptops (until the universal tariffs last month). Yet those 1st-term tariffs were high enough to shift large subsections of the US computer supply chain—desktop imports from China peaked just as the trade war began in 2018, and overall computer imports from China peaked in 2022. It was factories in Mexico that covered those lost Chinese imports starting in 2018, it was factories in Mexico that first plugged into the resurgent US semiconductor supply chain, and it is factories in Mexico that are now supporting the US data center buildout amidst the AI boom. Last year, Mexico dethroned China as the single largest source of US computer imports even as the large east-west laptop trade continued, and computers are on the verge of dethroning cars as Mexico’s single-largest northbound export.

Or take Canada’s aluminum industry, which has grown to replace large chunks of what the US would once import from its geopolitical rivals. The trade war, sanctions, and the evolving aluminum tariff system from Trump’s first term all placed heavy restrictions on US imports from China and Russia, causing aluminum imports from both countries to peak in 2017. Canadian aluminum has stepped in to fill that gap, especially after Trump exempted them from the tariffs imposed during his first term. Canada has natural advantages in the sector due to its surplus of hydropower useful for the energy-intensive smelting processes, and the country’s aluminum industry has been a reliable partner for so long that it’s a key supplier for US arms manufacturers with the Department of Defense long considering it a national security asset that America could rely on even if the country was cut off from international shipping.

Both Canada and Mexico were also betting on further spillover benefits from the large industrial policy push of America’s CHIPS/IRA spending and the continuation of friend-shoring trends. Industrial fixed investment in Mexico has skyrocketed over the last four years amidst a massive government-backed infrastructure buildout and increasing private-sector industrial development. At its peak, nonresidential construction rose more than 50% above pre-COVID levels as the construction of transportation infrastructure, refineries, factories, and more grew to meet expected demand growth. Investment in the machinery and equipment needed to fill those new factories has also been continually increasing, reaching a new record high just before the end of 2024.

Mexico has also seen rapid employment growth within the “maquiladora” factories devoted to export industries—in fact, they’ve been the source of the entirety of Mexico’s post-pandemic formal manufacturing job gains. These near-shored jobs represent much more than simply re-packaging Chinese components—the share of Mexican value-add in its exports is currently near the highest levels on record.

Meanwhile, more than 2.6M Canadian jobs are now embodied in exports to the US, a record high. That includes 1.2M Canadian manufacturing jobs—530k directly and 640k indirectly—numbers that have nearly entirely recovered from the impact of the Great Recession. Both countries were betting big on the continued growth of their mutually-beneficial economic partnerships with America, only for the US to try pulling the rug out from under them via massive trade wars.

Subscribe now

Conclusions

The current US administration has wide-ranging, unclear, and shifting demands of its major trading partners—depending on the time of day, the complaint could be about immigration, economics, insufficient defense spending, drugs, or annexation. Vanishingly few of these demands are even internally consistent—basically no fentanyl comes through the northern border, Mexico is already cooperating with America’s more restrictive immigration policies, and Trump is reducing Canadians’ desire to buy US military equipment by repeatedly threatening to annex the country. Even accepting Trump’s economic arguments on the supposed destructive power of bilateral US trade deficits, tariffs on Canada don’t make sense—the country buys more manufactured goods from the US than vice-versa, with the deficit entirely explained by pipeline oil imports that Trump himself supports. The administration does not have a thought-out negotiating strategy either—“cooperation” on border and drug enforcement stalled the tariffs against Canada and Mexico for one month, but no progress satisfied Trump enough to prevent the implementation of tariffs this month, and then no subsequent concession was required for him to partially backtrack after implementation. There’s good reason why the Wall Street Journal editorial board called these tariffs “the dumbest trade war in history.”

The easy answer, then, is for the US to just stop pursuing them. Indeed, there is a decent chance Trump will once again only implement the universal tariffs on Canada and Mexico for a couple of days in April before coming up with another excuse to delay them. That would have large economic costs from chaos and uncertainty, but it would at least avoid the most catastrophic impacts of an extended trade war. Yet if he keeps his promises and implements tariffs for an extended period of time, what can Mexico and Canada do? Given its size, the gravity of the US economy is simply impossible to escape—yet there are still some options available to minimize the harm.

For one, Mexican and Canadian leaders have an extremely strong mandate from voters to stand up to US tariffs. Mexican President Sheinbaum is perhaps the most popular democratically-elected leader on the planet right now, and Canadian figures like Justin Trudeau, Mark Carney, or Doug Ford have seen their popularity jump after taking tougher stances on the trade war. Trump’s position, by contrast, is likely to cost him popularity over time—already, his approval ratings are slipping as most voters oppose his Canada/Mexico tariffs and say he’s not prioritizing addressing the high cost of living. The two countries can possibly force Trump to backtrack by outlasting him politically even if they cannot economically—and making the fight even more politically costly by targeting Trump-aligned constituencies for retaliation has already been core to their strategy.

For two, they could use the economic crisis generated by tariffs as a way to push through much-needed economic reforms. One of the best ways to stop Canadian steelworkers and lumberjacks from losing jobs would be reforms to address the country’s chronic underproduction of housing. Mexico could seize the moment to push for a stronger response to the pre-tariff economic slowdowns or make reforms to the nation’s struggling oil sector. In an extended trade war, both countries are likely to double down on more public infrastructure investment.

For three, Mexico and Canada will not be alone in their battle—universal tariffs on key goods plus Trump’s promised sector-specific tariffs are slated for April 2nd and will start trade spats with most of the world’s major economies. That move will make it more difficult for Americans to just swap Mexican-made computers for Japanese-made ones, thus increasing the share of trade war harms absorbed by the US. Conversely, it will likely make it easier for Mexico and Canada to pursue deeper overseas trade ties—something both countries are actively encouraging.

Vanishingly few details on those April 2nd tariffs have been released even as the date rapidly approaches, but they promise new trade wars on other close US allies like the European Union, Japan, South Korea, India, Australia, and more. Although none of those countries would be hit nearly as hard as Mexico or Canada, the combined effect on the US and global economy could be massive—perhaps marking the bitter end of the nearshoring policy movement once and for all.

Subscribe now

By subscribing you’ll join over 57,000 people who read Apricitas!

What was once unthinkable is now reality. At 12:01 AM this morning, Donald Trump imposed the largest tariffs in modern US history—a 25% tariff on all Mexican imports, a 10% tariff on Canadian energy products, a 25% tariff on all other Canadian goods, and a doubling of the 10% tariffs on China imposed 30 days ago. The costs of these actions are enormous, covering $1.3T in US imports or roughly 42% of all goods brought into the United States. If households and companies tried simply to import the same goods from the same countries as last year, they would now be forced to pay an 11.5% effective tariff or more than $370B in total. Sustained, any one of these actions would exceed the total impact of Trump’s entire first-term trade wars, but combined they represent the single-largest tariff hike since the infamous Smoot-Hawley Act of nearly a century ago.

For as long as they last, these tariffs will drive up US prices for key raw materials like gasoline, fertilizers, steel, aluminum, wood, plastic, and more. Groceries, especially fresh fruits and vegetables from Mexico, will become harder to find. Manufacturing industries reliant on complex integrated North American supply chains—vehicles, computers, chemicals, airplanes, and more—could grind to a halt if those links are forcibly severed. Costs could spike for phones, laptops, and appliances where production is particularly concentrated in China and Mexico.

Exporters will be hurt by increased costs for raw materials, currency appreciation, and upcoming retaliatory tariffs—all of which will cut into US economic activity. In fact, the Canadian government has already retaliated, immediately hitting $21B in US exports, and the Chinese government has scheduled tariffs on $22B in US exports for early next week. Even more retaliation is coming—Mexico is prepping to unveil its response package, Canada is threatening another $86B in tariffs, and provincial Canadian leaders are promising their own reprisals. American export-oriented farmers, energy producers, and automakers are already in the crosshairs.

Fear of these tariffs was already having profound effects on consumer sentiment and business behavior, mostly by inducing panic-buying and hoarding throughout the previous few months. Preliminary data for January shows the trade deficit jumping by more than 25% as Americans rushed to buy foreign-made goods before Trump took office. That trend likely accelerated after the early February tariff scare, continuing to undermine US growth, consumer confidence, and inflation expectations.

Today’s tariffs, especially the ones on Canada and Mexico, are a self-inflicted wound with no good reasoning behind them. The legal justification that this is about fentanyl and undocumented immigration makes no sense given Trump did not even try negotiating with Canadian or Mexican authorities for more “concessions” before imposing tariffs. The lack of exclusions for any goods—even energy, raw minerals, basic foodstuffs, etc—and the public-facing posture of “maximum uncertainty” makes the tariffs counterproductive as industrial policy. Finally, from a geopolitical perspective, the decision to put brutal tariffs on Canada and Mexico only serves to breed animosity and push once-close allies away.

With the imposition of these tariffs, it is as though Trump has self-imposed what many foreign adversaries have only dreamed of: a partial blockade of the United States of America. This is the single-largest obstruction of US trade in post-WWII history, perhaps only behind the initial shock of COVID. As long as these tariffs remain in place, they will drive up prices, drag down US growth, and continually risk pushing the smaller Canadian or Mexican economies into recessions.

Even if tariffs are quickly rescinded, imposing them will have only undermined faith in future negotiations—after all, if Trump is so easily willing to break the USMCA trade deal he personally signed, why would you sign more trade deals with him? Decades of mutual exchange and market integration throughout North America were unilaterally reversed in one fell swoop this morning—that damage to trust is enormous and difficult to repair.

Subscribe now

A Deeper Breakdown of Trump’s New Tariffs

Trump’s universal tariffs on America’s 3 largest trading partners cover such a breadth of goods that virtually every industry will be affected, from pharmaceuticals to textiles to aerospace. Yet the hardest-hit major products are concentrated in a few key sectors—vehicles, electronics, and energy. Delivery and pickup trucks, the vast majority of which are come from Mexico, are among the most impacted—as are the seats, parts, and other components built throughout North America to serve the continent’s sprawling auto supply chain. Electronics (like game consoles, TVs, and computers) are also hit hard by this trade war since they’re mostly manufactured in China and to a lesser extent Mexico. Finally, there’s the large amount of US oil and gas imports that mostly originate from Canada. Key pipelines powering American refineries, power plants, businesses, and households may have been spared the full 25% tariffs, but a 10% price hike is still significant.

Indeed, Midwestern US refineries source the vast majority of their crude oil from Canada, and have vanishingly little ability to replace it given the pipeline networks and how much they’ve adapted to specifically servicing Canadian crude. Midwestern states are thus likely to see noticeable increases in gasoline prices going forward now that the 10% tariff has officially taken hold. That will also be true of other northern-border communities dependent on Canadian electricity or gas.

Fresh fruits and vegetables are another key set of goods highly affected by the tariff push, as Mexico is the dominant source of imported American produce. Alongside Canada, it represents the vast majority of fresh berries, limes, cucumbers, peppers, and tomatoes consumed in the United States. That’s especially true in mid-winter, when many beloved foodstuffs can’t easily grow in American climates—March saw the 2nd highest import volumes for Mexican fruits & vegetables last year. Alongside a variety of other foodstuff, Mexico is also a major source for imported soft and hard drinks, while Canada is the number one source for the (small) number of eggs the US regularly buys from abroad.

The auto sector is also key, as America is the world’s largest vehicle importer and much of North American trade is thus related to car production. The US does roughly $100B in cross-border automotive trade with Canada and another $150B with Mexico—tariffs will undermine US manufacturing, considering how intertwined trade is within the continent. Indeed, they would instead benefit the subset of European, Korean, or Japanese carmakers who do almost all manufacturing work abroad and simply ship the final product to the US. This problem is especially pronounced in the EV market, where Mexico has surged to become America’s number one source of overall EV imports while simultaneously being a major provider of EV batteries and parts to American factories.

Indeed, the size of the nearby car industry is one of the most important factors in determining local trade war risks—hence why the state of Michigan may have the most to lose right now. Its proximity to Canada and abundance of auto manufacturers means that the state’s total trade volume with Canada, Mexico, and China is 1/4 the size of its GDP. Several other states stand out for their high exposure—especially Mexico’s neighbor Texas, refinery hub Illinous, and other manufacturing states near the Great Lakes.

Subscribe now

Conclusions

Trump’s trade war continues to escalate as countries retaliate at to massive tariffs he’s imposed. Xi Jinping had already hit roughly $12B in energy, tractors, trucks, and agricultural machinery when Trump imposed 10% tariffs on China last month. In response to Trump doubling the tariff on all Chinese imports to 20%, China has now hit $22B in US farm exports with retaliatory tariffs while also placing restrictions on the operations of several US multinationals. The largest US export hit was soybeans—one of the top victims of Trump’s first-term trade war—alongside cotton, sorghum, meats, poultry, and more. In targeting lots of farm products, Xi is still unsubtly trying to concentrate fire on Trump-supporting constituencies to minimize the political benefit of pursuing the trade war.

Indeed, the Canadian government is pursuing a similar strategy with its first tranche of retaliatory tariffs, while drawing up plans for an even larger response if the trade war persists. In a sign of escalations that might be coming, Ontario Premier Doug Ford was planning to cut off power to sections of bordering US states, but has seemingly moderated to “only” implementing a 25% export tax. For her part, Mexican President Claudia Sheinbaum is holding off on official retaliatory measures until the week. Yet Trump seemingly shows no interest in backing down, instead only wanting to push this self-destructive battle even further. Just a few weeks into his 2nd presidency, Trump’s tariff blockade is already escalating well beyond the trade wars of his 1st term.

Subscribe now

By subscribing you’ll join over 57,000 people who read Apricitas!

At the start of this month, Donald Trump signed the largest US tariff hike in a generation—a 25% universal tax on all imports from Canada and Mexico, plus a 10% tax on goods from China. These 3 countries are America’s largest trading partners, representing $1.3T worth of imports or roughly 43% of all goods brought into the United States. Trump backed down on the Mexican and Canadian tariffs just hours before they were scheduled to go into effect, announcing a 30-day pause after achieving negligible concessions, but the tariffs on China were allowed to come into full force—just weeks into office, the 2nd Trump Trade War had already begun.

No exemptions of any kind were made—not even for specialty agricultural goods or raw minerals found only in the targeted countries. The only carveout was a slightly lower 10% tariff for Canadian energy products, most of which consist of crude oil that travels by pipeline to key US refineries in the Midwest. The “de minimis” exemption that allows small packages of under $800 in value to enter the US tax-free was removed, which will primarily impact large-scale low-cost Chinese e-commerce shippers like Temu and Shein but would also heavily impact goods from Canadian & Mexican small businesses. Likewise, there are no “duty drawbacks” that would allow companies to recoup tariffs paid on goods that cross borders multiple times as part of the normal manufacturing process. On their own, each of these tariffs would represent some of the largest trade war actions in postwar American history, combined they represent the largest possible tariff hike in more than a century.

The tariffs on Canada and Mexico would impose massive costs on the US economy—driving up the price of gas and groceries, paralyzing key industries like auto manufacturing, and inspiring retaliation against American exporters. They would have even more devastating impacts on the Canadian and Mexican economies, where trade with the US is a much larger share of GDP than vice-versa. The pretextual justification that this was about “drug trafficking” and “migration” rang hollow as Trump repeatedly made economic arguments about America’s trade balance to justify the tariffs. The threat of brutal tariffs on Canada, whose contribution to US undocumented immigration and fentanyl smuggling are minuscule to say the least, highlighted the ridiculousness of the whole enterprise.

These tariffs were so near-universally detested that markets and businesses found it difficult to take the threats seriously even as Trump signed formal executive orders and repeatedly saber-rattled at Canadian and Mexican officials. Yet the possibility of tariffs alone has done large amounts of harm by increasing economic uncertainty while undermining confidence in the USMCA trade deal Trump himself signed in his first term. They damaged public opinion of the US among some of its closest allies while driving both Canadian and Mexican leaders to try diversifying away from the US. Plus, the nightmare scenario where tariffs actually come into force might yet occur—the pause leaves the sword of Damocles in place and sets the stage to repeat the last-minute crisis as the new March 4th deadline approaches.

However, the implemented tariffs against China alone already rival some of the largest moves of Trump’s 1st-term trade war—affecting $445B in regular imports and roughly $20-$40B in “de minimis” imports. There are no exemptions, even for politically salient consumer goods like laptops, cell phones, and toys that were spared from Trump’s first trade war. There was also remarkably little warning, as tariffs were officially signed late on February 1st and began at midnight on February 4th. This is a Trump administration that is less afraid of spooking markets or consumers than in its first term.

Yet the tariffs were clearly not part of a well-thought-out plan either. Take the sudden ending of the “de minimis” exemption for Chinese imports—US Customs & Border Protection (CBP) didn’t have the capacity to process the millions of “de minimus” packages through the new formal means required by the Trump Executive Order. The result was that USPS abruptly stopped accepting packages from China & Hong Kong altogether and millions of packages began piling up at US airports much faster than CBP could clear them. The White House relented after just a couple of days, amending the executive order to restore “de minimis” to Chinese imports until the Secretary of Commerce could affirm that adequate processing systems are in place. This was an extremely predictable problem, which is why the prior administration’s significant restrictions on Chinese “de minimis” imports were taking time to implement, but the new White House is instead shooting first and asking questions later.

Then, the Trump administration formally announced 25% tariffs on all steel and aluminum imports entering the country beginning March 12th, restarting a trade war he began in his first term while ending all the prior country exemptions and tariff exclusions he negotiated. These hit a much smaller dollar value than the tariffs on China, with US steel and aluminum imports tallying roughly $45B in 2024, but they mainly affect trade with close US allies like Canada, Mexico, South Korea, and the EU. Unlike the China, Canada, or Mexico tariffs, they also provide no way to ease the economic impacts by allowing companies to source imports from different countries—all steel and aluminum will have to be made in America or subject to high tariffs.

Finally, on February 13th Trump announced a vague but extremely wide-ranging policy of imposing “reciprocal” tariffs on all foreign nations. Ostensibly, this is just the US mass-retaliating against the existing tariffs imposed by places like India, Argentina, and the EU. Yet the proposal also claims authority to retaliate against policies that are explicitly not trade barriers (like value-added taxes) and gives the president a wide range to determine what other policies count as “unfair” trade barriers. Given that, and the enormous complexity of theoretically going country-by-country and tariff-by-tariff in retaliation, the proposal seems closer to a way to justify large blanket tariffs against any nation the same way fentanyl was used as an excuse for tariffs on imports from Mexico, Canada, and China.

Combined, these moves are setting the tone for Trump’s 2nd trade war—expect lots of proclamations, many of which will not be followed through on or will only happen after some can-kicking. There will be a ramping up of the bipartisan focus on China and Chinese businesses abroad, on top of prior tariffs and at a faster pace than his first term. There will be some global tariffs on amorphous categories of key goods—starting with metals and likely spreading to automobiles, semiconductors, pharmaceuticals, and more. Finally, there will be mixed attempts at watered-down versions of the universal tariffs Trump promised during his campaign through mechanisms like the “reciprocal” tariff proposals released last week.

Subscribe now

Looking at the New Tariffs Already in Place

Trump’s first-term trade policy was dominated by his long-running trade war with China, which began in 2018 and escalated for several years until roughly 2/3rds of all goods once imported from China were affected by tariffs. Washington and Beijing signed a phase-one deal to begin resolving the trade war in early 2020, yet that deal broke down shortly after implementation. The large China tariffs put in place by the Trump administration were thus mostly left in place by the Biden administration, who also added some additional tariffs on Chinese batteries, semiconductors, metals, and critical minerals. Amidst all of these tariffs, China went from representing 21.8% of all US goods imports pre-trade-war to 13.4% of US goods imports today—and Trump’s 10% tariff on all Chinese imports already rivals the most dramatic of all prior trade war actions.

These new universal tariffs hit all the goods previously spared from the trade war, most importantly the bevy of consumer electronics deemed too salient for tariffs by the first Trump administration. In 2024, the two largest US imports from China were phones ($50B) and computers ($36B), most of which represent smartphones and laptops heretofore exempt from tariffs. Game consoles and TVs/Monitors were the sixth and seventh largest imports, both at roughly $7B and also mostly exempt from tariffs until now. The US also still imports billions of dollars in computer parts from China, even though they were hit hard by the first-term trade war.

In most of these goods, there has already been some import diversification away from China since Trump first took office (the country represented 66% of US laptop imports in 2024, down from 93% in 2017, and 72% of US cell phone imports, down from 80% in 2017), but the US still remains heavily dependant on the PRC. Already, some electronics companies are passing on the tariff to consumers—computer giant Acer is hiking laptop prices by 10% and the CEO stated he expects many competitors will do the same.

Batteries are another major Chinese import hit by tariffs, and they have exploded in economic importance since the first trade war amidst the rise of electric vehicles and large-scale battery energy storage systems. China is the world’s largest producer of batteries by a wide margin, and the US imported more than $17B worth of batteries from them in 2024. Efforts to decouple the US from reliance on US batteries have been underway for years, especially in the wake of the Inflation Reduction Act’s subsidies for domestic production. Those have also been coupled with increased tariffs from the Biden administration, who hiked taxes on the $2.4B worth of Chinese EV battery imports from 7.5% to 25% as of September 2024 and planned the same hike for the $13.2B of Chinese non-EV battery imports for January 2026. The new 10% tariffs are additive on top of those increases, forcing EV makers and battery storage facilities to eat price hikes or find new battery sources even faster.

As mentioned earlier, the executive order also excludes China from America’s “de minimis” system for imports of small packages. The “de minimis” system is a legal exemption for packages under $800 per person per day designed so that minor purchases from individual consumers don’t have to go through the expensive and resource-intensive formal import process or pay tariffs. That exemption became an important tool for hyper-low-cost e-commerce giants like Temu and Shein—companies could ship directly to US consumers and sidestep the tariffs faced by large retail stores on imported clothes or goods from China. It’s impossible to get an exact number for the value of these imports precisely because they don’t go through the full formal import process, but data recorded from China’s side estimates it at $23B while US CBP estimates it at roughly $45B. Yet both agree that the volume of packages has hit record levels and has been surging for years, especially since the pandemic. Ending the “de minimis” exemption for Chinese imports, should it be formally implemented and survive legal challenges, would heavily impact these packages—simultaneously hitting them with higher processing costs, normal US tariff rates, preexisting trade war tariffs, and the new 10% China tariff.

Finally, remember that in addition to all these goods being hit with tariffs for the first time, there’s already a wide range of Chinese goods with high tariffs from the first trade war which will now see further 10% tariff hikes. Since these products are still being imported from China despite 5-7 years of continuous tariffs, it stands to reason that they represent categories where decoupling from China is more difficult and thus the economic cost of further tariffs is higher. In 2024 America collected $2.4B of tariffs on $9B worth of imported Chinese car parts—if suitable replacements for them were easy to find, they would have been found already.

In response to the 10% tariff hike, China has already implemented retaliatory measures against the US—however, the relative size of Chinese tariff increases has been small. Beijing only imposed tariffs on roughly $12B of US exports in contrast to the roughly $470B worth of Chinese exports affected by American tariffs. Chinese retaliation also primarily focused on commodity energy exports—crude oil, coal, and natural gas—where bilateral trade flows can shift more easily without major economic costs to either country. There were also roughly $5B worth of American trucks, tractors, and agricultural machinery exports hit by Chinese retaliatory tariffs, a narrower measure designed to hit Trump-supporting constituencies hardest. In addition this, China also took several other non-tariff measures to respond—export controls on key minerals, an antitrust investigation into Google, and restricting biotech company Illumina’s ability to operate in China.

Beijing appears content to wait out the initial wave of trade restrictions and is holding its ammo in case it’s needed for further rounds of tariffs (after all, Trump promised tariffs of up to 60% on China on the campaign trail). The country still holds significant leverage over US farmers—soybeans are the single-largest US export to China, totaling $12.7B in 2024 (roughly 10% of total exports to China), and they were hit so hard by the first trade war that they required a US government bailout. Beijing could turn the screws on American farmers again by targeting US soy—as well as other agricultural commodities like cotton, sorghum, and select meats. They could also go after Boeing (aircraft & related parts are America’s second-largest export to China) and US semiconductor manufacturers, who still export decent amounts to the country. Yet for now, the primary impacts of the China trade war escalation on the US economy come from rising import costs rather than any implemented retaliatory measures.

Subscribe now

The Tariffs Possibly Coming Soon

Before the flurry of tariffs on China, Canada, Mexico, steel, and aluminum, it briefly looked as though this Trump administration wouldn’t focus on trade policy right from the gun—In his first few days in office, the President didn’t take any immediate action to raise tariffs, only directing various agencies to prepare studies on the negative impacts of international trade by April. Those studies span the range from symbolic (drawing up an “external revenue service” to take over the responsibility of collecting tariffs) to vague (countering “currency manipulation”), to serious (reviewing the USMCA, looking at more ways to expand the current trade war with China, and raising the possibility of ending Permanent Normal Trading Relations with Beijing). Yet all of them share an aversion to strict policy formulation, instead posing a wide latitude of trade policies and leaving the executive branch discretion to either impose or negotiate them away.

The recent memo on “reciprocal” tariffs is a good example of the discretionary chaos—the proposal claims to restore “fair” trade by hitting other countries with tariffs in proportion to their existing tariffs on American exports, yet it’s missing even basic details of what this would mean in practice. The White House complains about India’s large tariffs on motorcycles—will they retaliate by putting large tariffs on the vanishingly few motorcycles imported to the US from India, by picking a different import of a similar size, or by hitting India with an across-the-board tariff on all imports? It’s completely unclear. This is also only framed as a ratcheting-upwards, as there will be no tariff reductions for countries with preexisting lower tariffs on US exports than vice-versa. It’s also another violation of basic World Trade Organization rules by the US—the most-favored-nation system was designed so that countries would set tariff rates on all WTO members equally by type of good, and could only negotiate them downwards on a country-by-country basis through trade agreements. The idea that America would hit India specifically with an extra tariff just because India tariffs motorcycles from all countries flies in the face of this principle, and if taken literally would turn the US tariff schedule into a Frankenstein’s monster of complexity with thousands of unique rates depending on the type of good and country of origin.

Yet the proposal also goes well beyond simply retaliating against foreign tariffs or trade barriers by claiming that foreign value added taxes (VATs) are “unfair” and “discriminatory” towards the US and thus equivalent to tariffs. This is patently ridiculous—Germany’s VAT hits all domestic consumption evenly, whether it’s an American car or a German one—but it provides a pretextual justification to slap basically any country with tariffs since nearly all of them have VATs. If that wasn’t enough, there’s another catch-all that allows retaliation against any practice the US Trade Representative deems “unfair”—in essence, the President is directing agencies to make up possible wide-ranging tariff hikes on all countries by April, with only a loose nod to the idea that they should somehow be “reciprocal”.

Chair of the White House Council of Economic Advisors Stephan Miran and Treasury Secretary Scott Bessent have mused about a system where countries would face differentiated tariff rates based on a variety of factors including their rate of tariffs against the US, whether they facilitate transshipment of Chinese goods to America, their geopolitical relationship to the US, how they treat American multinationals, and more. Importantly, they have also hinted at tariffs against close allies for not meeting defense spending targets and against neutral nations for voting the “wrong way” at the United Nations or “grandstanding” against the US. In practice, the vagueness of the reciprocal tariff memo can be read as the first steps required to implement such a system—even if countries have low tariff rates on US exports, the White House can point to their VATs or other policies to fabricate a justification for hitting enemies or noncompliant allies with tariffs.

It’s impossible to make any accurate predictions about what the effect of these “reciprocal” tariffs would be without more details, but a literal interpretation where America retaliates 1-1 against foreign tariffs and VATs would drive US tariff rates to the highest levels since the Great Recession’s infamous Smoot-Hawley Act. It’s doubtful Trump will actually implement tariffs that high, but he may threaten massive uniform tariffs against the EU, UK, Japan, India, Brazil, and others the way he already is against Canada and Mexico. In seeking public comment on the “reciprocal tariff” proposal, the Office of the United States Trade Representative has named the places it is “particularly interested” in—a list that consists of twenty countries plus the EU and thus composes the vast majority of overall US trade. Among those, they’re also focusing on countries where the US has a large bilateral goods trade deficit—that means China, the EU, Mexico, Vietnam, Canada, Japan, Taiwan, South Korea, India, and Thailand. The “reciprocal” tariff proposal will likely be used to slot foreign countries into several tariff bins which can then be threatened as part of negotiations or implemented to to raise overall tariff rates and reshape trade relationships.

In addition to these more general or country-specific tariffs, the White House has repeatedly discussed its desire to implement universal tariffs on specific goods, which began with the renewed tariffs on all imported steel and aluminum. It’s unclear precisely what other categories will be hit with tariffs, but Trump and other administration officials have repeatedly mentioned other metals, lumber, motor vehicles, pharmaceuticals, and semiconductors. These tariffs are much more straight-up mercantilist and even less geopolitically driven than China or “reciprocal” tariffs, would primarily hit close US allies, and would leave less room for the kind of nearshoring effects that eased the impacts of first-term tariffs.

Start with the largest of these sector-specific tariff threats: motor vehicles. The Trump administration already went through the motions of a section 232 investigation into vehicle and parts imports during his first term and threatened the same 25% universal tariffs he is threatening today, but ultimately backed down. The size of US vehicle trade meant large tariffs would have prohibitive economic costs—cars are the single-largest US import category, representing $224B on their own in 2024, and when combined with car parts and light trucks compose nearly 11% of all US imports. Yet now that Trump is back in office and less concerned about backlash, he’s resurrecting the large tariff proposal with renewed vigor.

Except for those heavily tariffed car parts from China, US vehicle-related imports almost universally come from close allies like Canada, Mexico, Japan, South Korea, and the EU. For the North American auto trade in particular, components and vehicles often cross borders multiple times before a consumer ends up behind the wheel, and deintegrating supply chains would be extremely disruptive. The auto industry is also still in a precarious position, as prices for new and used vehicles remain extremely elevated from pandemic-era disruptions while US vehicle manufacturing employment has dropped by 30k over the last year.

Next, there are the floated 25% tariffs on imported pharmaceutical products. In terms of details, there’s been even less provided than for steel or autos, yet the contours of policy are visible. Just looking at the likely targets of prepackaged medicines and immunological drugs, the US imports more than $200B worth of goods, the vast majority of which come from the European Union or elsewhere in Europe. The continent’s advanced drugmaking industry and the attractive tax-shifting implications for many pharmaceutical multinationals have caused US-EU pharma trade to boom over the last ten years.

Finally, the last repeated proposal is for universal tariffs on semiconductor imports. The US bought $38B worth of chips in 2024, with direct imports of high-end chips from countries like Taiwan skyrocketing amidst the AI boom. Yet the US also imports a wide variety of less-cutting-edge semiconductors as components for vehicles, consumer electronics, appliances, medical devices, and more, and tariffs would affect input costs for all of these industries. The implemented policy might even be more expansive than just hitting chips—tariffs would also have to hit computers and other electronics made with semiconductors to prevent companies from sidestepping them by simply assembling electronics outside the US. As a result of fabricator subsidies passed in the 2022 CHIPS Act, US production of semiconductors has been rising for years and will continue rising through the early Trump administration, but America remains far from wholly chip self-sufficient and tariffs here would still have wide-ranging negative impacts.

Subscribe now

Conclusions

Throughout all of this, American consumers are growing increasingly worried about the impact of tariffs. In December and January, Americans rushed to buy vehicles and other durables out of fear that tariffs would soon cause price hikes, then the February tariff executive orders against Canada, Mexico, and China caused consumer sentiment to plummet. Economic policy uncertainty has spiked significantly, making it even more difficult for businesses to make planning and investment decisions. Yet Trump remains largely undeterred, pursuing even more aggressive versions of the trade policy strategies that dominated his first term.

The 10% tariffs on China and universal tariffs on steel/aluminum are blatant reruns of first-term policies just with greater speed and at larger levels. The 25% tariff threats against Canada and Mexico over “fentanyl and immigration” are larger versions of the 5% tariffs against Mexico that Trump threatened in 2019 over migration but never followed through on. Trump placed universal tariffs on washing machines during his first term and is now expanding those policy ideas toward larger imports like vehicles, semiconductors, and pharmaceuticals. Even the “reciprocal tariff” memo is just a more serious attempt at policy frameworks that Trump was openly proposing as early as 2017.

Those strategies didn’t achieve their economic goals during Trump’s first term, with the academic consensus showing that the costs far outweighed the benefits. Doubling down in a second term is unlikely to change that. Yet the first-term tariffs may have achieved some of their political goals by signaling support for communities and workers negatively impacted by globalization even if they didn’t help them materially. A louder, brasher trade policy following an era of much larger inflation might have the opposite effect, costing the President political support among voters mostly concerned about their pocketbooks. For the self-described tariff man, backlash from Main Street and Wall Street might be the only thing keeping his second-term trade war ambitions in check.