American Manufacturing Is Surging–Despite Tariffs, Not Because of Them

One of the more annoying hazards of the wonkery biz is commenting on available-but-lagging information that’s subsequently reversed or clarified by future data releases. In January, I wrote that the American manufacturing renaissance promised by the White House was “missing in action” because manufacturing sentiment was persistently depressed last year and actual performance had noticeably sagged in the fall. Since then, however, both have turned up: Industrial output has gained in each of the last seven months, and manufacturers report increased optimism—a marked shift from the dour reports of 2025. Perfect timing.

The Trump administration has definitely noticed this reversal, recently trumpeting the Institute for Supply Management’s (ISM) latest manufacturing purchasing managers index (PMI), which, in their words, showed that “U.S. manufacturing activity surged in July at the strongest rate in more than four years — driven by soaring demand, record production, and a wave of new hiring.” Now, the vice president is gloating, White House trade adviser Peter Navarro is calling it a “robust tariff- and tax cut-induced renaissance,” and prominent protectionists are snarkily citing the PMI as proof anti-tariff “think tankers, economists, and columnists” don’t know what they’re (ahem, we’re) talking about.

The new narrative from the pro-tariff crowd—which rarely differentiates correlation from causation—is that the ongoing manufacturing upswing is a clear Trump trade policy success story.

Yet there are several reasons to doubt that spin. The U.S. manufacturing sector’s recent growth is real but has been exaggerated in recent surveys. More importantly, the hard data we now have show that the nation’s industrial upswing is being driven by non-tariff forces—ones that have more than offset a clear tariff headwind. In fact, American factories would likely be doing even better without the tariffs, an inconvenient reality that brand new research confirms. Let’s dig in.

Debunking the narrative that protectionism is working.

Sentiment isn’t output, and it started in a deep hole.

The first problem with the protectionists’ spin is wonky but important: Much of the recent tariff triumphalism rests on PMI surveys that are useful for gauging short-term industry sentiment and forecasting future trends but can misrepresent what’s happening nationwide and over the long term. As economist Dave Hebert just detailed, the PMI records the share of purchasing managers in various U.S. manufacturing industries who report an increase, decrease, or no change in their orders, employment, prices, and other categories of business activity. The index documents only the direction of the change, not its magnitude. Thus, Hebert explains, a big company like Ford could lay off 1,000 workers while two other small automotive firms each hire five, and the index would read 66.7, signaling a robust “expansion” even though actual employment fell by 990.

The PMI’s reporting of rising inventories and slower deliveries (which can signal backlogs caused by high demand) can also be misleading: Both are treated as unambiguously positive developments by the ISM even though in the real world they could signal deteriorating external factors (e.g., Iran-related stockpiling and supply chain bottlenecks). Thus, part of the PMI’s recent “strength” probably reflects costly frictions created by the Iran conflict and tariffs—frictions not worth crowing about.

The other issue with the PMI surveys is that, as Bloomberg’s Shawn Donnan just explained, they track only “how purchasing managers feel about this month versus the one before,” meaning that “if this month is marginally better than a lousy one before it, the index points up.” In January, I highlighted 2025’s long string of contractionary PMI readings because they continued the “lousy” situation from previous months—persistently pessimistic manufacturing sentiment that directly countered the administration’s promises of a tariff-driven boom. That sentiment has now changed, but—as Donnan notes—“what the ISM is really showing so far this year is a short-term bounce in sentiment from low levels.” Put another way, it’s a few steps up after dozens of steps down.

Finally, the manufacturing surveys report a broadly negative view of tariffs. The ISM PMI’s latest prices paid index, for example, shows that “raw materials prices increased for the 22nd straight month” thanks to tariffs and Iran, while many tariffed materials (especially metals) are persistently in short supply. Meanwhile, the commentary accompanying the PMI reports—actual statements from people working at surveyed U.S. manufacturers—has consistently painted tariffs as a headwind, not a tailwind. As The Economist reported in March:

The most recent ISM reports continue this trend and are echoed by other surveys. As I noted in Bloomberg a few weeks ago, for example, the National Association of Manufacturers has consistently found that trade-related uncertainty was members’ top challenge since early 2025 (only edged out now by Iran). American manufacturers’ message on tariffs is clear: They’re a problem to be managed, not a solution.

The claims ignore powerful non-tariff forces driving the recent improvement.

Despite these problems, real-world data do show an uptick in the U.S. manufacturing sector, especially in 2026. According to the Federal Reserve’s industrial production index, domestic manufacturing output has been on a decent run since Trump took office, outside of that multimonth dip in the second half of last year.

That’s good news for the sector, but there’s little reason to think it’s owed to Trump’s tariffs. More likely, the growth is happening despite them.

For starters, three powerful factors coincided with the 2025–26 tariffs and are the most likely drivers of U.S. factory output. On the supply side, the One Big Beautiful Bill Act restored and made permanent provisions that allow U.S. businesses to immediately deduct spending on equipment, machinery, and research and development (R&D), and the law temporarily allowed U.S. manufacturers to do the same for spending on structures, effective January 2026. As the Tax Foundation explains at the link above, research shows that cutting the after-tax cost of these business inputs boosts investment and growth, and they estimate that the OBBBA’s permanent expensing provisions will boost long-run GDP by a significant amount (0.7 percent). As I and others have explained for years now, these neutral, free-market reforms are particularly beneficial for large, capital-intensive manufacturers, and their timing aligns with the current U.S. manufacturing acceleration. (The temporary expensing provisions might induce a sugar high but won’t affect long-run growth; they should instead be made permanent.) The White House, for what it’s worth, seems to agree: Its official press release credits business tax cuts, not tariffs, for the current manufacturing “boom.”

On the demand side, the U.S. artificial intelligence buildout—now totaling trillions of dollars in new spending that dwarfs previous U.S. infrastructure booms—has become a big and direct source of orders for American manufacturers that equip both the stuff needed to build a data center (bulldozers, trucks, structural steel, etc.) and what goes inside them (electrical equipment, server racks, semiconductors, etc.) To grok the mind-boggling scale of this demand, consider an April report that GE Vernova’s electrification unit “sold more grid equipment, including substations and transformers, to data center customers in the first quarter than all of last year.” They’re certainly not alone.

Third, the Iran war has both boosted domestic output (and prices) of energy and energy-related goods and set off a short-term stockpiling rush among companies looking to get ahead of possible supply crunches later this year. The latter trend, Bloomberg reports, was what drove a significant manufacturing expansion in the spring. Unlike tax reform and (maybe?) AI, this is a short-term sugar high—extra purchases now mean fewer purchases later—but still important to acknowledge in any honest assessment of the current state of U.S. manufacturing. Tariff champions rarely do.

Look under the hood, and the data also rebut the pro-tariff spin.

The hard data paint an even starker picture. For one thing, the recent growth has been historically modest: The Fed’s manufacturing index in July 2026 was 99.3—just a hair above where it was in February 2020 and around 2.5 percent below mid-2018, when Trump’s first tariffs really got rolling. The pace of the current expansion also isn’t stellar: Output is up roughly 1.3 points over the last 12 months, against a 2001–2019 median of 1.4. This recovery is fine and good, but it’s hardly a renaissance.

A deeper dive creates even bigger problems for the pro-tariff case. Most basically, the manufacturing sector’s rebound—and the growth of the industries driving that rebound—began months before Trump’s new tariffs took effect in March 2025:

(Several other industries not shown here are still shrinking—more on that in sec.) Given these trends and similar ones for manufacturing productivity, the 2025–26 “boom” is less a “Trump phenomenon” than it is the continuation of a rebound that began months earlier, likely owed to the AI buildout and the business cycle.

The next hit to the tariff success story comes from import volumes over the recent “boom” period. In a protectionist framework, imports and domestic production are substitutes: Restrict the former with tariffs or other measures, and the latter zooms ahead. Yet this is decidedly not the case for goods imports overall and for industries doing the heavy lifting in the Fed’s manufacturing data. Most importantly, data from the Bureau of Economic Analysis show that real imports of capital goods, excluding autos, were up 39 percent between the fourth quarter of 2024 and the second quarter of 2026—the largest sustained import increase outside the post-pandemic rebound and one that coincided with a major increase in domestic production of similar goods:

Oddly, the White House press release and Council of Economic Advisers Chair Kevin Hassett acknowledge this trend, boasting that the “historic peak” in capital goods imports is evidence of a major U.S. factory expansion. The more likely culprit, however, is the AI buildout: booming domestic production of data-center inputs alongside booming imports of the same (especially from Mexico), thanks to a generational demand shock—not tariffs knocking out foreign competition and encouraging new factories.

Indeed, the hard data show that U.S. tariffs and manufacturing performance have actually run in opposite directions since Trump took office. Using the U.S. International Trade Commission DataWeb, we can calculate the tariffs collected on goods falling into the Fed’s industry groups over the last few years and thus the actual level of protection those industries have been getting. The pattern we see is definitely not what a pro-tariff story would predict.

First, the top two U.S. industries powering most of the recent U.S. manufacturing expansion—computers/electronics and aerospace—are the ones facing the lowest (and declining!) tariff protection, while the industries enjoying the highest levels of protection are contributing little or even shrinking outright:

The trend in production capacity is even worse for the pro-tariff case, with lightly protected industries expanding and highly protected ones contracting.

Even an apparent tariff winner—the moderately protected U.S. automotive industry–is less than it seems. Yes, motor vehicles and parts output have expanded since 2024, but overall production capacity (see above) and fixed investment are basically flat, while jobs are down—trends possibly reflecting the U.S. industry’s cannibalistic business model and automakers’ cutthroat lobbying (against each other!). Even more damning, however, is the fact that domestic automakers enjoyed a huge tariff-related subsidy—the Section 232 import adjustment offset—that let companies assembling vehicles here claim a credit (3.75 percent of the cars’ MSRP) to offset duties paid on imported parts through 2030. Given that this credit’s potential size was actually larger last year than all duties assessed on imported car parts, the industry’s modest growth may be owed to the simple fact that politically powerful U.S. automakers won a tariff exemption as large as their tariff bill. (And now they get big tariff refunds too.) The charts above won’t reflect the tariff credit. Include it, and the “pro-tariff automotive story” gets even weaker.

Finally, there’s the timing of the manufacturing sector’s 2026 acceleration. The effective duty rate on all manufactured imports peaked last October at 12.1 percent, and output stagnated. Following the Supreme Court’s February ruling against Trump’s “emergency” tariffs, however, the effective rate fell below 8 percent and has hovered there ever since. And as the average tariff rate fell (and trade policy uncertainty waned), manufacturing output zoomed ahead:

Correlation isn’t causation, of course, and I’d wager that the recent surge has more to do with AI, tax reform, and Iran than with declining tariffs. But it’s still no less true that the manufacturing “boom” protectionists now celebrate arrived as the President’s signature tariffs—and, importantly, his power to chaotically impose them—waned.

The standard response to these and other figures would doubtless be that tariff-fueled investments take time to show up in the Fed’s output and capacity figures. That’s true (and one reason why I won’t yet claim total vindication), but it runs into another real-world data problem: Future manufacturing growth needs new factories, yet real private spending on manufacturing structures is down substantially since its 2024 peak. (The White House’s notes on factory construction jobs, meanwhile, are simply bogus.)

Tariffs supposedly work by changing the return on domestic investment versus importing, and new investment would show up in the construction and capacity figures. Neither is telling an optimistic tariff story so far, despite some flashy (and isolated) headlines about promised factory spending.

Summing it all up.

Unlike the politicians and pundits, I’m not declaring victory today because it’s still too early for definitive conclusions about Trump’s wild—and still ongoing—tariff experiment. But the initial evidence runs strongly against the “tariff boom” narrative and is consistent with both an AI-fueled alternative and the country’s long history of both protectionist stagnation and support for dying legacy industries instead of thriving new ones. (On the last point, consider today’s sky-high tariffs on apparel, furniture, and footwear contrasted with low ones on semiconductors, aerospace, and data-center equipment.) The burden rests on tariff advocates to demonstrate otherwise.

More rigorous research will likely make their job even harder. Just this week, for example, a new paper from four Federal Reserve economists examined the 2025 tariffs and found that—thanks mainly to broad exemptions for the capital goods fueling the data-center buildout—the AI investment boom was strong enough to keep imports and output growing, despite the tariffs that were applied (particularly on consumer goods). Their counterfactual: Absent the AI investment boom, Trump’s tariffs would’ve caused imports to fall by 10 percent, GDP to contract by 0.7 percent, and prices to rise. Apply the tariffs evenly (no carveouts), and growth shrinks even more (-1.0 percent).

Another challenge will come from a forthcoming paper in the Journal of Supply Chain Management, “Paying More and Doing Less,” by Michigan State economist Jason W. Miller and colleagues. It examines five Federal Reserve regional manufacturing surveys and finds “compelling evidence that the 2025 tariffs compressed U.S. manufacturers’ gross margins because prices paid for inputs increased more than prices received for outputs.” The authors add that “[t]he 2025 tariffs also had a negative effect on U.S. manufacturers’ new orders, employment, and capital investment,” with effects that were “significantly more negative on employment and capital investment than new orders.” The tariffs’ costs, in other words, outweighed their benefits.

Both papers arrived in my inbox after I started writing this newsletter. More will probably follow. And in each, the following story emerges: Faced with a large and unstable tariff shock and an even larger AI demand shock, American manufacturers didn’t build new factories and hire new workers en masse but instead deployed tariff mitigation strategies and boosted production at the factories they already had. Manufacturing growth was in advanced industries aided by tariff exemptions and accompanied by rising imports, while old, politically connected industries won heavy tariff protection yet still kept declining, perversely insulated from competitive pressures that might have motivated improvement. The negative correlation between tariffs and manufacturing growth isn’t concrete proof that tariffs caused industrial stagnation, but it’s a direct hit to the common protectionist claim of a tariff-fueled boom.

The best thing tariff advocates can say today is that their chaotic and costly taxes didn’t kill the AI golden goose or offset gains from last year’s corporate tax reforms (a genuine free market victory, by the way). But the AI boom’s capital expenditure wave might not last, and U.S. manufacturers facing tariff-inflated costs across their supply mix—along with U.S. laws that ensure future uncertainty—may be unwilling to invest in new factories on a wager that the AI gravy train keeps rolling (especially given other economic headwinds like an aging population and declining immigration). At that point, protectionists might not even have correlation to brag about.

Markets FTW

The beautiful website Ordinary Abundance pairs items you’d find in any modest U.S. apartment today—a stereo, lamp, refrigerator, and much more—with famous historical statements on those same goods, demonstrating how the “normal” stuff we enjoy today was considered a sci-fi miracle in the past. Scroll through it in a few minutes—and then stop taking your modern living standards for granted.
 

Chart of the Week

If only there were some kind of sign …

China’s lost 18 million manufacturing jobs—12 percent of its factory workforce—since 2013