More good news for American industry over the past couple of days. Today’s jobs report featured gains in construction and manufacturing. The economy has added 72,000 manufacturing jobs since December, and employment in the sector is more than 300,000 jobs above where it would have been on the trendline from 2024, when we were losing close to 20,000 per month.
Yesterday’s PMIs for September were likewise strong:
S&P Global: “Production and new orders both increase as domestic demand strengthens; workforce numbers expanded to greatest degree in over five years.”
Institute for Supply Management: “New Orders Growing; Production Growing; Employment Growing.”
ICYMI: Last week’s newsletter, Baghdad Bobs of American Industry, discussed these PMI trends and the folks who continue to proclaim the demise of American manufacturing regardless of the data. Unfortunately, the examples offered such a dense set of links that spam filters were triggered. So if you didn’t see it then, read it now, and remind your inbox that this fine weekly product is not spam!
LEADING THE WEEK: TREASURY (MAN) BITES HEDGE FUND (DOG)
On Monday, the U.S. Treasury Department announced it would begin aggressive scrutiny of various tax-avoidance maneuvers that Wall Street has cooked up in recent years. Secretary Scott Bessent: “Treasury is serious about cracking down on transactions designed to dodge taxes or exploit our federal tax code.” Suffice to say, a Republican administration placing Wall Street in the crosshairs is an uncommon and very encouraging sign.
It also underscores how far over the line the money managers have gone. The issue was thrust into the spotlight over the summer when Treasury first began making noise on the issue and Bloomberg published a feature story on the nearly $200 billion pouring into “tax-aware long-short” funds that in some cases help the rich avoid taxes even on their wage income.
The leading practitioner, AQR, declares proudly on its website that “taxes can create a gap between how your investments perform and how your wealth grows. Our tax-aware capabilities are designed to help close this gap.” Quite the euphemism!
See, taxes create a gap between your earnings and your growing wealth because some share of your earnings has to go toward supporting your nation and providing public goods to you and your fellow citizens. If everyone tries to “close this gap,” the country falls apart. I wrote more about this in my Financial Times column this week: Do US lawmakers finally have capital in the crosshairs?
A BAD ANALYSIS ROUND-UP
Hard to keep up with it all, but some favorites from the past week or two:
When is a work requirement not a work requirement? From the Cato Institute, The New Right Discovers Welfare Without Work Requirements. Cato characterizes the Trump administration’s draft proposal to open child care subsidies to families with a stay-at-home parent as “welfare without work requirements” even though the benefit would only be available to families with earned income from a working parent. This framing makes sense only if one thinks of a work requirement’s premise as “every able-bodied person must be in the formal labor market” as opposed to “public support to a family should be premised on, and create incentives for, that family working to support itself.”
The fundamental libertarian error here is to treat the individual rather than the family as the benefit recipient. But if we are going to treat the individual as the recipient, why is it the working parent in one case and the stay-at-home parent in the other? Why not make it the working parent in both cases? Or the child, for that matter? The sterile libertarian mindset is simply incompatible with the way human beings live and societies function.
One might go so far as to suggest that the Cato Institute’s lament that making families with a stay-at-home-parent eligible means “taking childcare subsidies away from working parents” is pretextual, seeing as their preferred policy is “to shrink welfare programs, including childcare subsidies.”
The data is beside the point. From the Wall Street Journal editorial board:
The Tepid Trump Economy (June 30). Growth comes in low, “last year’s tax bill and deregulation would be driving faster growth if not for Mr. Trump’s border taxes that raise costs and uncertainty for business.”
The U.S. Economy Is Accelerating (September 30). Growth estimates revised upward! Tariffs bad anyway.
Justin Wolfers, Conservative. It’s been too long since we’ve heard from Justin Wolfers, who had a lovely post this week on why My Libertarian Friends Were Right. Whereas he never noticed this during, say, the Obama or Biden administrations, he has now discovered that “concentration of power can be abused” and that’s led him “to become more small-c conservative. The old conservative idea was that we have rules of the game that have served us well for generations. We should be slow to change them.” Welcome, Justin! We trust he will be a vocal proponent for this perspective during the next Democratic administration.
Goalposts, moved. From the American Action Forum: “It is time to stop treating the outcome of tariff policy as a hypothetical. The experiment was run in 2001 and failed.” Employment in the steel industry has declined, argues Douglas Holtz-Eakin, thus proving steel tariffs do not work. But as Nick explains:
Tariffs are the only reason why the United States still has a steel industry of any size. China’s excess capacity alone is sufficient to satisfy 100% of U.S. steel demand. If we let them, they would. Free traders frequently set the bar for tariff effectiveness at whether they triggered a “boom” in the year or two after implementation. This is ridiculous, because the point of tariffs is to trigger a structural readjustment in demand that requires investment, construction, etc. over longer time horizons. The whole point is that you’re building capacity that didn’t previously exist. If it did, you wouldn’t need tariffs.
…
For a test of whether the current round of steel tariffs can drive new investment, see e.g. Mesabi Metallic’s announcement yesterday of the largest investment in the history of American steel, which followed two other massive capital commitments (Hyundai and Nippon). Output is up across the sector, and new capacity coming online brings the price down. That’s how it’s always worked since the days of my favorite tariff “experiment,” which is the creation of the American industrial base behind a 25-40% tariff wall during the Industrial Revolution.
That’s a good segue to…
THE REINDUSTRIALIZATION MARCHES ON
Trump to Unveil Planned $15 Billion Iowa Steel Project (Wall Street Journal). “Mesabi’s proposed mill would be one of the largest in the U.S. with an initial annual production capacity of 7.5 million tons. Later additions could increase the plant’s capacity to about 10 million tons a year.”
The Journal also has a good overview of the deep hole the U.S. dug itself on critical minerals (if you’ll pardon the pun), the Trump administration’s aggressive dealmaking to begin climbing out, and why it will still take a long time to recover. See How the U.S. Is Attacking China’s Control of Critical Minerals. Meanwhile, with the U.S.-China summit in the rearview mirror, President Trump seems fed up with China continuing to squeeze the critical minerals supply chain: US slows aircraft-part exports to China as Trump seeks leverage in trade negotiations, sources say (Reuters).
And on the Journal’s op-ed page, a remarkable commentary from JPMorganChase CEO Jamie Dimon: A Plan for the Western World’s Revival:
I will focus my attention on strategic foreign economic policy. Since World War II, America has fostered economic and military relations, trading with more than 100 nations and forming more than 50 military alliances. The levers to accomplish our foreign-policy goals include trade policy, through tariffs, quotas, investment policies, regulatory barriers and other instruments. Another lever is industrial policy, which encompasses subsidized loans, tax rules, export credits, immigration policies, government investment and development finance.
Perhaps some day the Journal’s editorial board will start reading the rest of its paper.
JPMorganChase putting some money where its mouth is. “Ford Motor Co., JPMorgan Chase and the state of Michigan are establishing a $3 billion program that addresses manufacturing and supply chain challenges by providing companies with financing, workforce support and other resources to scale production” (Manufacturing Dive).
Congress, moving. IFP’s Aidan Mackenzie says, “WE HAVE A PERMITTING DEAL” and it’s a good one, making progress on transmission infrastructure, NEPA litigation, and various types of energy projects. If this gets across the finish line, following on the housing legislation from mid-year, that would be not-insignificant progress on a couple of long-stalled priorities.
NOT BOOMING: HIGHER EDUCATION
Per Gallup, Record-Low 31% Say College Education Is ‘Very Important’. That’s “four percentage points below last year’s reading, the previous low among the seven times Gallup has asked this question since 1978. As recently as 2013, more than twice as many U.S. adults (70%) said college was very important.”
Not helping matters, Cornell. The university is making headlines for other reasons this week, but its Report of the Committee on the Future of the American University is worth a look. In particular, the assessment of shortcomings in its undergraduate program focuses heavily on a failure to equip graduates with the soft skills necessary to succeed in the world. As NYU’s Jay Van Bavel summarizes the findings, “companies have a ‘jarring’ skepticism of higher ed thanks to a new generation of degree holders who lack key skills like handling uncertainty, compromising, and absorbing feedback. Faculty also report that incoming students, despite four years of undergraduate education, are ‘underdeveloped’ in these relational or interpersonal skills.”
That failure is rapidly becoming common knowledge. What’s notable about the Cornell report, though, is its conclusion that the solution is for universities to lean heavily into actively teaching those skills. On one hand, it would be a very good thing for universities to rediscover their roles as formative institutions rather than amusement park entitlements. On the other hand, the kinds of challenges the report describes are ones that are upstream of the university, in the way our public schools and communities are failing to prepare young people for adulthood at all. Making universities the site of remedial math and reading is bad enough, making them the site of remedial citizenship is unlikely to work. They should only be admitting students who can handle “the basics,” and we should keep in mind that most young people do not attend or succeed at college, which means that we have no choice but to ensure the basics are learned elsewhere.
Some better, or at least not-bad, news: A new paper from UCLA’s Robert Fairlie and Jane Wu, The Early Impacts of AI on Employment among Recent College Graduates, looks at the summer 2026 labor market and finds that, if anything, the most recent college graduates seem to have found employment especially quickly.
WHAT ELSE IS NEW IN THE WORLD OF ARTIFICIAL INTELLIGENCE?
Paying attention and paying interest. Apollo Global Management chief economist Torsten Slok warns that ordinary people using AI agents to help manage their finances may begin sweeping their savings en masse from low-interest checking accounts into higher-interest options, causing bank runs (via Andrew Curran). I’m not sure about that one. Seems like a long-winded way of saying that today’s enormous bank profits rely in part on high switching costs keeping customers in bad products. One solution to a possible run might be for banks to offer competitive interest rates on liquid deposits.
But Curran makes a larger, more important point: Widespread reliance on just one or a few AI models will lead to all sorts of weird, herd-like behaviors that are mitigated in normal life by the varying preferences and priorities of billions of human beings. I would make the point even more broadly, which is that life’s vitality and a civilization’s progress depend on interaction and conflict between different sets of values and ways of looking at the world.
The approach manifested by Anthropic, in its training of Claude on a “constitution” and a single definition of what it means to be “good,” runs the risk of overhomogenizing the approach to decision making that might govern various parts of our economy or society. Would we be better served by some sort of modified alignment training that sends numerous different versions of advanced models into the world, all with slightly different ways of thinking?
Speaking of thinking about things the same way… Did Anthropic’s A.I. Really Make a Scientific Discovery on Its Own? asks the New York Times. “An expert at the University of Copenhagen said his team had been sharing its research with the company’s A.I. model, Claude, and that its new finding matched their work.” There doesn’t seem to be much evidence that the Copenhagen team’s work was directly misappropriated by the model, but the case does underscore the importance of distinguishing between AI-driven scientific breakthroughs that are very much along the lines of what scientists were already discovering and ones that portend an entirely new trajectory for our frontiers.
One group still thinking differently? Chipmakers. Lawmakers plan to cut off China from AI chips. Nvidia, AMD want Trump to intervene. Every time these companies step up their efforts to undermine the democratic process and the interests of the nation that has made their success possible, they make the case for aggressive regulation a little bit stronger. As I testified before the House Foreign Affairs Committee in January, “under our Constitution, citizens have the right to petition their government, no matter how unwise their ideas. But also under our Constitution, it is the people’s representatives, not multinational corporations, that set national policy, to ensure that it advances the national interest.”
We just have to remember that if business leaders make it exceptionally clear that market incentives are misaligned with the public interest, then the case for a light touch on the free market simply does not apply.
Somehow even worse: China’s Tencent leases 100,000 chips from Oracle to accelerate AI push (Financial Times). It’s genuinely unclear how we could possibly still be allowing Chinese firms to access U.S.-controlled data centers. My only explanation is that the idea of blocking sales of our most advanced chips but then leasing them instead is just so absurd that people assume it’s not really happening.
AS FOR OTHER TRADE WITH CHINA
Tough beat for Ford CEO Jim Farley in the “Key Points” summary on this CNBC article: Ford CEO says it’s ‘too late’ for Europe to fend off Chinese automakers, but not for U.S.
Key Points:
Ford CEO Jim Farley said it’s “too late” for Europe to fend off an influx of Chinese automakers in the region, but the U.S. still has time to “be considerate” about its decision.
Farley’s comments come as Ford tries to compete against Chinese automakers in Europe, while also partnering with some Chinese companies on technology and to fill plants.
We just have to remember that if business leaders make it exceptionally clear that market incentives are misaligned with the public interest, then the case for a light touch on the free market simply does not apply. (I repeat myself.)
Decision Time for Europe. EU steel exports hit by high energy costs, tariffs and China oversupply (Financial Times). “European steel exports have fallen by a fifth with production at historic lows as the industry contends with soaring energy costs, US tariffs and the knock-on impacts of vast oversupply from China.”
South China Morning Post’s Finbar Bermingham has a good primer: EU weighs sweeping new trade powers against China before make-or-break October. “The mood in Brussels is hardening and Europe’s actions in coming weeks could set the course of EU-China relations for years to come.” He also highlights a new paper from the Federation of German Industries (BDI), which argues that “even if de-risking entails costs, including in the form of counter-reactions, passivity would in the long term be economically considerably more costly and irresponsible from a security-policy perspective.”
The French are talking tough! Macron: “China today is attacking our domestic market for machine tools and motor vehicles. She already did it in chemistry. And so it is destroying our industrial base. We Europeans were naive.”
But, at the Hudson Institute, Ulrich Speck lays out his scenarios, and is not optimistic:
Scenario 1: China pretends to make concessions or makes minor concessions — the EU, led by Germany, plays along and doesn’t use its trade defense instruments. Likelihood 50%.
Scenario 2: China pretends to make concessions or makes minor concessions—the EU uses its trade defense instruments, we move towards a trade conflict. Likelihood 30%.
Scenario 3: China understands that the EU is united and serious, and makes serious concessions. Likelihood 20%.
My take is based on two assumptions: (1) Germany (Merz) is still very nervous about Chinese retaliation against German companies in China, and (2) China does not believe that the EU is a strong and united player, and it believes it has the upper hand.
As I’ve argued for a while, we cannot make this decision for the Europeans, but we do both them and ourselves a favor by putting them to the choice, and we can hope they choose well.
Already choosing well, Japan. Japanese companies exit China in record numbers (Financial Times):
A record number of Japanese companies have closed their China businesses under the triple pressure of slowing momentum in the Chinese economy, US tariffs and a diplomatic freeze between Tokyo and Beijing. As of June 2026, a total of 10,118 Japanese companies had an established presence in mainland China, according to Teikoku Databank, Japan’s largest corporate credit research firm. That figure, published on Wednesday, was the lowest number since the group began compiling the statistic in 2010. It represented a 22 per cent fall from 2024 and a 30 per cent drop from a 2012 peak for Japanese corporate presence in China.
One recalls the laments that the U.S. was pushing allies into China’s arms. But then, one recalls many things the Blob told us in recent years that proved less than prescient.
Axios gets the last word: “The bottom line: Trump-era tariffs are divisive, but the president’s case for rethinking the global trading system is proving less so.”
WHICH BRINGS US TO YOUR WEEKLY CANADA GOOSE…
…and it’s still Prime Minister Mark Carney. While Carney postures about turning away from the United States toward a closer partnership with the EU, in an effort to protect Canada’s “sovereignty,” EU urges UK to raise tariffs on Chinese cars to avoid ‘made in Europe’ barriers (Financial Times). According to the FT: “Brussels has warned Andy Burnham that Britain would need to raise tariffs on Chinese cars and align more closely with EU trade policy if it wants to avoid ‘made in Europe’ barriers that would hit key exports.” Wait a minute, that’s… that’s Donald Trump’s music.
Enjoy the weekend!
"The fundamental libertarian error here is to treat the individual rather than the family as the benefit recipient."
This is by no means unique to libertarians. Our entire society is structured around the individual, which is really weird considering that the first, most important, job of literally every human society is to produce and raise the next generation. Seems like successful societies would encourage family formation instead of maximal individual autonomy at all costs.
This is a problem that cultural Darwinism will take care of. But that's a bad solution, at least for us.
If the ploys used by the hedge funds, etc are legal or gray area stuff, the fault lies with Congress and/or the people at Treasury that write the regulations. Probably gray area stuff as vagueness is endemic in that world. So do the reviews, rewrite the regulations or statutes and perhaps fire some Treasury employees. Hedge funds are a plague but it is government that created the opportunity.