2026-06-15 17:17:45
The photo above is not from China; it’s from Japan. In the 1970s, Daiei was Japan’s top retailer. But after Japan’s asset bubble burst around 1990, it became Japan’s most famous “zombie” company — staggering along unprofitably, kept afloat by a constant stream of below-market-rate loans from UFJ Bank and other big Japanese banks. Eventually the company was acquired by Aeon, a more successful retailer, and its once-storied brand is slated to be retired for good in the next few years.
I tend to be very skeptical of comparisons between post-1990 Japan and post-2021 China, because there are just so many differences between the two economies (and between the global economic environments at the time). Their industrial policies are different, their trading relationships are different, their bubbles and busts happened for very different reasons, and so on. But in the case of “zombie” companies, there may be some important parallels.
What’s important about Daiei is not how it failed, but why it didn’t fail much sooner. Caballero, Hoshi, and Kashyap wrote a paper in 2008 arguing that “zombie” companies like Daiei held the Japanese economy back during the 1990s (and, in some cases, even beyond the 1990s).
The basic story is that after 1990, the Japanese economy slowed down, and lots of companies that used to be profitable — especially in the construction, retail, and trading sectors — were no longer profitable. These companies owed a lot of money to banks. If they stopped being able to pay back their loans, the banks would be forced to recognize bad debt on their books. This would get them in trouble with regulators (because of capital requirements), and it would also get them in trouble with the Japanese public.
So what the banks did was to lend even more money to the failing companies that already owed them a lot of money, at very cheap interest rates. The new loans were used to pay back the old loans, and the new loans would be classified on the bank’s books as “good” debt. This process — known as “evergreening” — kept banks from ever having to acknowledge their losses:
Peek and Rosengren (2005) document this empirically as well.
Evergreening kept a bunch of companies afloat — like Daiei — that had utterly broken business models. Theoretically, the companies could have eventually pivoted their business models and recovered, or Japan’s economy could have started booming again, etc. In practice, this never happened.
Caballero, Hoshi, and Kashyap argue that evergreening was very bad for the Japanese economy, because it hoovered up scarce resources that better companies could have used to grow. With all of those crappy loans clogging up their books, Japanese banks couldn’t lend to healthier companies. With big zombies like Daiei still able to employ large amounts of Japan’s best managers, young scrappy upstarts were deprived of talent. The authors argue that keeping all of this labor and capital locked up inside doomed companies contributed significantly to Japan’s long productivity stagnation.
Why did the Japanese government allow this to happen? Preserving employment at the zombie companies was probably a big part of it. Japan had a strong tradition of job security at that point in time, and to throw so many people out of work — even if they could have gotten new jobs eventually — would have been seen as cruel and unfair. Social unrest was a possibility. Bank bailouts may also have been deeply politically unpopular. In any case, whatever the reason, throughout the 1990s the government supported banks with various capital injections and regulatory forbearance, without forcing banks to cut off the zombies.
Anyway, that’s Japan. The question is whether something like this will happen in China.
China’s experience with its real estate bubble and bust doesn’t exactly parallel Japan’s, but there are some broad similarities. Since 2021, there has been a broad economic slowdown (probably more severe than the official numbers suggest), and a long-lasting chill in real-estate-related industries. This has predictably led to a rise in the number of loss-making companies:
You’ll notice on this chart that the share of non-performing loans has actually gone down since 2021, even as fewer companies are turning a profit. That suggests that lots of Chinese companies are being kept on life support by cheap bank loans. Here’s the Rhodium Group:
Some concrete data points suggest that China’s evergreening of debt is more widespread than is commonly the case in most market economies. The ratio of banks’ reported non-performing loans has decreased over the past years, while the share of loss-making enterprises increased…This would indicate Chinese banks have been sitting on large volumes of NPLs that have not yet been fully recognized. This is an open secret: The National Audit Office recently claimed in an annual audit report to the NPC that 16 of 43 audited banks last year had NPL levels that were double the officially reported figure…
Loan rollovers are a pervasive phenomenon in China…[T]he financial system…served as a shock absorber, channeling resources to enterprises facing losses to maintain output and prevent the defaults and bankruptcies that occurred in market economies.
Another Rhodium report finds that the proportion of loans made below benchmark rates has risen significantly since 2021, even though benchmark rates are lower than they were back then:
And the Dallas Fed has documented how more and more Chinese companies, especially in the real estate sector, aren’t making enough money to pay the interest on their loans:
All this — falling official NPLs, much more below-market lending, companies unable to pay their interest expenses, widespread suspicion that many of the companies whose loans are “performing” will never be able to repay those loans — matches the general pattern that Hoshi and Kashyap (2000) documented in post-bubble Japan. Banks have taken a bunch of losses, but have refused to recognize those losses, using a flood of cheap debt to keep their borrowers afloat.
A bunch of people have warned about this. Here’s Rhodium:
Because of the political incentives shaping China’s financial system, banks in China tend to extend or roll over debt to poorly performing or loss-making companies. This can have some of the same effects as a subsidy, by removing incentives for companies to stay profitable and isolating them from market forces that would otherwise lead to their restructuring or bankruptcy….Evergreening of credit, therefore, allows firms to…[reduce] domestic and global prices to unprofitable levels[.]
And here’s the Dallas Fed:
There is mounting evidence of “zombie lending” in China, banks rolling over bad loans to unprofitable firms and allowing the status quo to continue rather than recognize losses.
And here’s a Business Times story about how China’s government has allowed and even encouraged zombification, much as Japan did in the 1990s:
It’s impossible to quantify the true extent of the [bad debt] problem, though most economists say the ratio of bad loans is significantly higher than the 1.5 per cent official rate…One analyst at Absolute Strategy Research in London pegs it at about 10 per cent…Others say it could be double that amount…
While the [banks’] leniency [toward borrowers], largely condoned by regulators in Beijing, has helped maintain financial stability over the past few years, it also means the banking system is recycling capital into unproductive companies rather than spurring real growth in healthy firms…
[Government] officials have moved to bolster the nation’s six biggest banks with more than US$100 billion in fresh capital…[R]ather than cracking down on deadbeat borrowers, China’s banks are encouraged to cut them some slack. Regulators have for years urged the big banks to keep their reported bad loan ratio under 2 per cent, according to sources familiar with the guidance…As a result, banks routinely roll over maturing loans, extend repayment periods, or allow interest to be capitalised to avoid triggering NPL recognition.
Now you might be tempted to think — and I’ve seen a few people argue — that this only matters in a market economy. In a market economy, undercapitalized banks matter because banks have to succeed or fail on their own. In a state-directed economy like China’s, the theory goes, debt on the banks’ books might as well be on the government’s books.1 Banks can keep lending no matter how much bad debt they have, because the only entity that could punish them — the Chinese government — wants them to do so.
But while government control might avert a financial crisis, it doesn’t automatically solve the zombie problem, or make the comparison with Japan inappropriate.
First of all, it would be a mistake to see Japan’s government in the 1990s as operating at arm’s length from Japanese banks. It most certainly did not; in fact, it acted to support the banks that were supporting the zombies. The government bailed out the banks, deliberately turned a blind eye to the zombie problem, and encouraged banks to keep on lending to healthier companies despite the unrecognized bad loans on their books. That’s not too different from what China’s government seems to have done in response to the real estate bust, at least initially.
But simply having the government urge (or order) banks to keep lending didn’t solve the zombie problem in Japan, and it won’t solve it in China either. Even if the zombie companies don’t end up competing with healthier companies for capital, they compete with them for other resources. They compete for labor — workers who could be working at young, growing, healthy companies are instead being paid to continue to work for unproductive companies that are just spinning their wheels. They also compete for raw materials, for land, for energy, and so on.
These resources are not in infinite supply, even in China. As long as unproductive zombie companies are hiring workers, hoovering up metals and chemicals and watts of electricity, and taking up prime real estate, they’re holding back the rest of the economy. This doesn’t just manifest as higher costs for healthy companies — it also shows up as increased competition. In 1990s Japan, if a new retailer wanted to enter the scene, it had to compete with Daiei, the unproductive behemoth that was essentially being paid by banks to produce below cost. The same will be true in China.
In fact, this may be a reason for the “involution” that Chinese companies are experiencing. In the wake of the real estate bust, China’s government directed banks to lend to manufacturing companies instead of to real estate-related companies. They did this (though some of the loans ended up sneaking back into the real estate sector). In fact, a large percent of the “subsidies” that China dishes out to its manufacturing companies is through below-market-rate loans.
Some of these manufacturing companies will be successful and efficient — indeed, many already have been. But others are unproductive and inefficient. Instead of letting these die, China’s banks may keep them on as zombies as well, paying them to compete with China’s healthier companies. Here’s Alicia Garcia-Herrero from back in March:
In many sectors, including…electric vehicles, solar panels, batteries, and other green technologies…Chinese firms…keep selling at rock-bottom levels, sometimes below what it costs to produce, just to hold onto market share. A growing number of these companies cannot earn enough revenue to even service their debt…These “zombie” companies survive only because banks roll over loans and local governments provide subsidies to avoid job losses and keep tax revenues flowing…In newer, high-priority sectors like green tech, the share of zombie companies has hit 30 percent of total listed companies…
Without real productivity advances, [zombies] still join the price-slashing frenzy to stay in the game thanks to external support from banks or local governments. They cut prices aggressively…The outcome is predictable: collapsing profit margins across the board, even for the better companies, whose productivity is increasing.
When we Westerners think about the effect of Chinese zombification, we often think about the flood of cheap exports threatening to deindustrialize Europe and other regions. But while that export dominance might seem like a victory to China’s mercantilist leaders, it’s a double-edged sword, because zombification reduces productivity at home. In the long run, lower productivity hurts growth, despite the temporary bump from exports.
In other words, China’s fusion between the financial system and the state may have made zombification worse, not better. The Chinese state is not a ruthlessly efficient allocator of capital; it has sociopolitical goals just like any other state, and it fears the unrest that could result from widespread corporate failure and unemployment. Yes, it can tell banks to lend to manufacturers instead of property developers, but that just ends up adding more zombies to the horde.
And at some point, even state-owned and state-directed banks probably do care about profitability. Yes, the government can bail out any bank at will, but if you’re the bank executive or manager who dished out the bad loans and made a bailout necessary, your career might be over. This might be why corporate loans have started to fall slightly from the torrid pace of 2023-24:
Ultimately, when people write the story of China’s economy in the 2020s, zombification could end up being more fundamental to that story than exports. The parallels with Japan are not always real, but they’re real in this case — and so far, China’s government seems to be walking into a similar trap.
And since that debt is owed almost entirely domestically, the theory says that the debt doesn’t really matter in a macroeconomic sense; it’s just some Chinese people owing money to other Chinese people.
2026-06-13 16:25:32
About a month ago, I weighed in on an interesting debate over American vs. European living standards.
On one side, you have people who argue that Europe is much poorer than America, and is falling even further behind. Most of the people making this argument are Europeans themselves — especially economists like Mario Draghi, Philippe Aghion, Luis Garicano, and Antonin Bergeaud. They look with envy on the U.S. tech sector, which Europe has no real equivalent to, and they yearn for liberalizing reforms that would allow Europe to catch up.
On the other side, you have American liberals like Paul Krugman and Brad DeLong. They argue that Europe is not falling behind — that most of the gap in material living standards is due to Americans working more, and that America’s apparently faster productivity growth is an artifact of the way relative growth rates are measured.
This is an interesting debate, and in the end it comes down to some surprisingly technical measurement issues and data mysteries. I’ll have more to say on it in the future. But it’s really just the most recent exchange in a long-running debate over the “varieties of capitalism” — whether Europe’s stronger welfare and regulatory states deliver better outcomes for regular people, or whether America’s more free-market system is superior. I don’t think that debate is close to being settled — and may never be settled, since liberalizing reforms in Europe and expansions of welfare and regulation in America may shrink the differences between the two systems.
But there’s a second debate going on regarding Europe’s economy, with potentially far more devastating consequences. It’s the question of whether Europeans should be rich at all.
Instead of crowing about the superior performance of the European social model, leftist economists are increasingly arguing that Europeans are too rich, and ought to be poorer than they are. This was the upshot of a manifesto by Thomas Piketty and his World Inequality Lab, which I covered in my last roundup post. Piketty argued that “labour hour reductions, growth caps in rich countries, less material consumption, and changes in food habits” will be needed in order to beat climate change:
This is also the argument in a recent Guardian editorial by Piketty, Olivier De Schutter, Joseph Stiglitz, Jayati Ghosh, Kate Raworth, and Jason Hickel:
We live in an age of manufactured scarcity. In a world richer than ever before, roughly one 10th of the world’s population still lives in extreme destitution…
For decades, the recipe was simple: grow the economy, and poverty would gradually disappear. But the promise that economic growth would “lift all boats” has not been kept. While national incomes expanded, wages stagnated, work became more precarious and public services were cut. At the top, fortunes ballooned; at the bottom, families turned to food banks. Growth has become decoupled from shared prosperity…It has also become ecologically unsustainable…That is why we have come together to develop and support the “roadmap for eradicating poverty beyond growth”.
The essay — which was immediately identified by Pangram as being 100% AI-generated — was extremely vague on its plans for transforming the global economy. It often reads like a mishmash of buzzwords and slogans. Here’s a taste:
The real question today is not whether growth continues, but what kind of economies we are building, who they serve and whether they allow everyone to live in dignity within planetary boundaries…[W]e are united in the conviction that our economies must be redesigned around the fulfilment of rights and collective wellbeing within planetary boundaries, rather than maximising output at any cost. Human rights here are not an afterthought; they are the organising principle for how we measure progress, set priorities and resolve trade‑offs…All too often, policies affecting people in poverty are designed without them – and sometimes against them. When welfare systems are built around suspicion, sanctions and humiliating conditions, they deepen stigma and deter people from claiming their entitlements.
…And so on. In fact, this is pretty typical of degrowther writing. If you read their “research” papers, it’s also a bunch of stuff like this. Consider this excerpt from “Exploring degrowth policy proposals: A systematic mapping with thematic synthesis”, by Fitzpatrick et al. (2022), published in the Journal of Cleaner Production:
Degrowth is a multi-layered concept…It combines critiques of capitalism…colonialism…patriarchy…productivism…and utilitarianism…whilst envisioning more caring…just…convivial…happy…and democratic societies…Capturing the essence of degrowth is difficult because it carries at least three denotations…degrowth as decline of environmental pressures…degrowth as emancipation from certain ideologies deemed undesirable, like extractivism, neoliberalism, and consumerism; and...degrowth as a utopian destination, a society grounded in autonomy, sufficiency, and care.
Honestly, whatever AI wrote the Guardian op-ed did a better job than Fitzpatrick et al.
In fact, there’s a reason degrowthers write like this. They see degrowth as a grand project to unify and reinvigorate the political left — a Big Idea to fill the hole left by the collapse of communism in the 20th century. Each buzzword or stock phrase is a shout-out to a particular faction of the European left — “decolonial” leftists angry about colonialism, climate activists, old-line socialists still angry about “neoclassical economics”, unionists who want job guarantees, social democrats who care about housing and education and welfare, and so on. Leftism is famous for factional infighting over differences in doctrine and focus; the degrowthers want everyone on the left to know that they’re building the biggest of big tents.
But in practice, what unites the degrowthers is their conviction that Europeans ought to be poorer. The Guardian op-ed makes it clear that the postwar European social model is not enough:
Social protection and public services are essential, but they cannot indefinitely compensate for economies that by design generate poverty wages, insecure jobs and unaffordable housing.
Why do I say “Europeans ought to be poorer”, instead of “Westerners ought to be poorer”? Obviously, the degrowthers would love it if America also degrew. But in practice, no one in the U.S. is signing on to this agenda. American leftists dream of Green New Deals and government-funded megaprojects like high-speed rail. Progressives are focused on breaking up big companies in order to ensure free-market competition. Liberals like Krugman value living standards intrinsically — they trumpet the European model precisely because they think it makes average Europeans rich.
Degrowth has also notably failed to win much traction in Canada or Australia. Nor have rich countries in Asia shown any interest. Developing countries ignore the idea as well, trusting more in their ability to grow their own economies than in the promise of massive transfers from the Global North.
This is why in practice, the only people who are interested in degrowth are Europeans. The idea’s main proponents — Piketty, Jason Hickel, Kate Raworth, Timothée Parrique, and so on — are all Europeans.1 The degrowth conferences are mostly in Europe. Degrowth literature reviews show that it’s a “predominantly European movement”. The EU is the only government that has shown much interest in degrowth.
In other words, when degrowthers call for the “Global North” to make itself poorer for the sake of the planet — or for the sake of any number of other causes — they’re functionally addressing their message to Europe. Degrowth is a movement for European impoverishment.
It’s important to note how astonishing of a shift this is from the pitch leftists made a hundred years ago. Communism was supposed to be about abundance — the idea that central planning could out-produce the capitalist system, ensuring not just military might, but better material living standards for regular people. In the 1950s, Western leftists praised the growth “miracle” of the Soviet Union (and, occasionally, North Korea). Even back in 2006, Joseph Stiglitz — one of the few American authors of the Guardian article — praised Hugo Chavez’s Venezuela for supposedly promoting “higher growth”.
That famously didn’t work out, of course. The Soviet Union, North Korea, and post-Chavez Venezuela are not exactly places that we associate with widespread material abundance and successful economic development. But in the wake of that colossal failure, the European left is switching its sales pitch. Yes, the new line goes, leftism will make you poorer — but that’s a good thing, because you deserve it for the sins of colonialism, because it’ll be good for the climate, and because you really don’t need all those consumer goods anyway.
Astonishing.
Of course, as anyone following the saga of the degrowth movement knows, it’s pure snake oil — a whirlwind of factual distortions and poor scholarship that would make any serious researcher blush. When economists have read through the degrowth literature, they have found it to mostly be repetitions of the same old buzzwords, empty rhetoric, and unsupported assertions. Savin and van den Bergh (2024) write:
In the last decade many publications have appeared on degrowth as a strategy to confront environmental and social problems. We undertake a systematic review of their content, data and methods…Based on a sample of 561 studies we conclude that: (1) content covers 11 main topics; (2) the large majority (almost 90%) of studies are opinions rather than analysis; (3) few studies use quantitative or qualitative data, and even fewer ones use formal modelling; (4) the first and second type tend to include small samples or focus on non-representative cases; (5) most studies offer ad hoc and subjective policy advice, lacking policy evaluation and integration with insights from the literature on environmental/climate policies; (6) of the few studies on public support, a majority concludes that degrowth strategies and policies are socially-politically infeasible; (7) various studies represent a “reverse causality” confusion, i.e. use the term degrowth not for a deliberate strategy but to denote economic decline (in GDP terms) resulting from exogenous factors or public policies; (8) few studies adopt a system-wide perspective – instead most focus on small, local cases without a clear implication for the economy as a whole. We illustrate each of these findings for concrete studies.
A few economists, like Vincent Geloso, have done yeoman’s work studying and rebutting the claims of Jason Hickel, Kate Raworth, and other degrowthers:
I myself, of course, have written about Hickel’s ideas and reviewed Raworth’s book, and I’ve come to the same conclusions that Geloso has.
This pattern of intellectual sloppiness and utter disregard for the data is evident in the Guardian op-ed. “We’ve done the maths”, the headline proclaims, but neither the article nor the “roadmap” it promotes contain any math. The statement that “growth has become decoupled from shared prosperity” is simply false; the evidence is unequivocal that growth is good for the poor. No matter how you measure poverty, there is no country that has escaped poverty without growth, and there is no country that has substantial amounts of poverty after growth has occurred:
This relationship works in reverse, too. If degrowth gets serious purchase in European policymaking, regular Europeans will suffer. In keeping with degrowth’s mission to unite the entire political left, the Guardian op-ed promises all kinds of middle-class goodies: “investing in children, housing, health, education and transport through universal public provisioning”. As the authors surely know — and as any cursory attempt at actual “maths” would have easily shown — this will be utterly impossible if European countries are forced to degrow.
In other words, even as American conservatives look down their noses at “Europoors”, and liberals try to debunk the “Europoor” notion, European leftists are hard at work trying to make “Europoors” a reality.
But that’s not the only reason the degrowth program is pernicious. Right now, with America having turned inward in a spasm of rightist isolationism, Europe is one of the few remaining champions of the ideals of universal human rights set forth in the post-WW2 global order — a lonely, beleagured bulwark against the likes of Russia and China.
Right now, Europe is facing a dire military threat. Although economically and demographically it overmatches Russia, its militaries do not yet know how to use drones — and Russia’s does. Furthermore, Europe’s supply chains are utterly dependent on Chinese parts and materials; if China decides to cut Europe off during a Russian attack, while continuing to furnish Russia with everything it needs, Russia would have a good chance of triumphing.
In addition, European industry is facing a dire economic threat. Its manufacturers are bearing the brunt of the Second China Shock — the wave of massively subsidized exports that China is using to try to prop up its slowing domestic economy. This comes on top of the cutoff of Russian energy and Trump’s tariffs.
To embrace the poisonous nonsense of degrowth now — to shut down nuclear power plants, to regulate the AI industry out of existence, to forcibly shorten working hours, to bar the construction of houses and factories, etc. — would be to cripple one of the last few remaining economic engines of the free world, at precisely the time when it’s under its greatest external challenge.2
In other words, there has never been a better time to ignore the pronouncements of Thomas Piketty, Joseph Stiglitz, and the other rogue economists who want to turn “Europoor” from a slur into a grim reality.
I’m counting the UK as Europe. Sorry, Brexiteers.
One suspects that degrowthers know this, and that the anti-Westernism that has always animated far-left movements is an unspoken but important motivating force behind the movement.
2026-06-11 15:38:27
“No. I am Hugh.” — Hugh
I remember the first time I heard about the invention of the iPhone, back in 2007. My friend, who followed Apple products with an almost religious zeal — there were many such people in those days — entered the room and announced “This is the convergence!” We spent the next few minutes gaping in awe at the idea that every single piece of portable consumer electronics was about to be combined into a single device.
For many, it felt like a messianic moment. The iPhone was probably the last big innovation that we Americans embraced as a whole society. Everybody had an iPhone, or wanted one. Engineers loved how the thing was engineered. Humanities PhD students showed off the latest model at parties and admired the sleek design.1 Kids in working class neighborhoods were glued to their iPhones in math class. It was a supercomputer in your pocket, a voice for the voiceless, the tricorder and the communicator from Star Trek, all that and more.
It was also big money. Years before Benedict Evans wrote “The smartphone is the new sun”, every ambitious tech entrepreneur and content creator in America was in on the game. Social media — that infinitely scrolling vertical feed — was the killer app of the smartphone, what the spreadsheet had been for the PC or e-commerce had been for the internet. In 2012, Facebook’s monster IPO kicked off a gold rush, and everyone moved to San Francisco to strike it rich.
But you didn’t need to be a tech entrepreneur in order to get in on the action. The smartphone meant far more eyeballs glued to far more screens for far more minutes of the day, and that meant dollar signs for content creators. YouTubers, Instagram fashion influencers, and Twitter activists became whole new economic classes. Old-style content businesses like newspapers and TV networks saw their doom, but also a potential lifeline. (Eventually even econ bloggers got our piece of the pie; plenty of you signed up for this blog through a scrollable app.)
Beautiful design coupled with brilliant engineering. Technology anyone could use. Economic opportunity for the masses and for the elite. A way to have your ideas and opinions heard by millions of people thousands of miles away, at any hour of any day. No wonder Steve Jobs was the last technologist that everyone agreed was an American hero.
But when I recall that fateful day in 2007, I remember not joy, but a sudden surge of foreboding. That was strange, and out of character for me. I’ve always been a technophile at heart — I grew up as a hardcore Star Trek fan, and until that moment in 2007, each new marvel — broadband, the internet, the laptop computer, etc. — had felt like it was moving us toward that bold utopian future. The iPhone felt different. Some small voice in the back of my head told me: “This is a mistake.” And though I tried to ignore that voice for many years, it remained.
What was I worried about? I think some part of me knew that someday, I would end up saying something like this:
The internet was wonderful because it was a place you could go — a complement to real life, not a substitute. The iPhone promised to put that fantasy universe in our pockets 24/7, and we would never escape. It would be physically possible to turn off our phones, of course, but it wouldn’t be socially possible — you could always touch grass, but after everyone had an iPhone, that would be the only conduit through which you could touch another human mind.
The idea of perpetually tying every human into a global hive mind tripped alarm bells. It reminded me too much of the hive minds I had seen depicted in science fiction nightmares — the Borg from Star Trek, the Blight from A Fire Upon the Deep, the Human Instrumentality Project from Neon Genesis Evangelion. Humans were meant to be individuals — unique, independent incubators of ideas and desires, not terminals or the fingers of a world-mind.
We had spent centuries trying to escape the small, localized versions of the hive mind. The printing press, the car and the telephone had offered freedom from the crushing conformity of small-town life. When broadcast television threatened to smother us with a centrally dictated monoculture, it sparked a decades-long resistance. When the internet arrived, we spent two decades using it to revel in our individuality — we made our personal websites, started blogs, joined small online communities centered around our interests.
Sometime around 2014 or 2015 we woke up to the fact that the world of the Old Internet no longer existed. “The internet” no longer meant the Web — it meant a tiny handful of big platforms. Twitter and Reddit for screaming about politics, Facebook and Instagram for being jealous of your friends’ vacation pics. Gone were the days of painting our individuality on the canvas of the Web. The platforms were the hive minds, we were the neurons, and the smartphone was the axon that kept each of us wired tight into the collective.
“‘Social media is bad,’ he typed on social media!!” This is the perpetual and instantaneous response of many of the neurons…er, people…in my timeline. Indeed, if social media is so bad, why don’t you just put down the phone? But this idea displays a fundamental misunderstanding of the nature of network effects. Suppose I decide to get off Instagram and go play pickup basketball instead. If everyone else is on Instagram instead of playing pickup basketball, who am I going to play with?
This is an extreme and simplified example, obviously, but the intuition here comes from real research. Bursztyn et al. (2024) have a paper called “When Product Markets Become Collective Traps: The Case of Social Media”. Here’s a quick summary:
While one would typically assume that a popular product benefits its users…Bursztyn et al. find evidence of a “product market trap:” At least among the college students in this experiment, large numbers of people are choosing to use social media platforms they also wish didn’t exist at all…[T]he authors found that [subjects] would…be willing to pay an average of $24 to deactivate the platform on their campus for four weeks. This amount rose to $43 when they also included the minority of students who weren’t TikTok users. (Results for Instagram were similar and are included in the full policy brief.) [emphasis mine]
The students said that they only stayed on TikTok because other people were on it too, and they were afraid of missing out. FOMO is not utility; it’s a bad equilibrium.
There’s also an impulse to do a sort of “Good Tsar, bad boyars” maneuver, where people say “it’s not the phones, it’s social media”. But this is an argument over whether guns kill people or bullets kill people. Yes, we could access social media from laptops or other stationary devices, but we could only do so for part of the day; that would force us to develop offline interactions and relationships during the other hours, like we did back in 2007. It’s the ever-present umbilical that enables — and perhaps even mandates — the replacement of in-person interaction with an online hive mind.2
In 2007 I suppressed my deep-seated doubts about smartphone technology. I am a techno-optimist, and this was just another miraculous new tool for human empowerment. And yet the two decades since 2007 seem to have only validated my misgivings, across a number of dimensions.
Plenty of evidence has linked smartphones — and the social media apps that take up the single biggest chunk of the time we spend on those phones — to rising unhappiness among the world’s young people:
Since I wrote about this in 2023, the evidence has only grown stronger. Here’s an experiment by Castelo et al. (2025):
We used a mobile phone application to block all mobile internet access from participants' smartphones for 2 weeks and objectively track compliance. This intervention specifically targeted the feature that makes smartphones "smart" (mobile internet) while allowing participants to maintain mobile connection (through texts and calls) and nonmobile access to the internet (e.g. through desktop computers). The intervention improved mental health, subjective well-being, and objectively measured ability to sustain attention; 91% of participants improved on at least one of these outcomes. Mediation analyses suggest that these improvements can be partially explained by the intervention's impact on how people spent their time; when people did not have access to mobile internet, they spent more time socializing in person, exercising, and being in nature. [emphasis mine]
Other recent experiments have yielded similar results.
The likeliest explanation is that there is simply something thin and insufficient about online interaction. It’s lacking in some essential emotional nutrient that human beings evolved to harvest from the physical proximity of other human beings. Perhaps it’s something cognitive — the richness of context that tells you that no, your friend’s life isn’t perfect just because they posted a cool video of their trip to Europe, and thus you don’t need to feel constantly envious and inadequate and left-out. Or perhaps it’s something physical — the tiny touch of a high-five or a hug, the simple feeling of the proximity of other human bodies.
Whatever this emotional nutrient is, our young people are starving for it, while they binge on the cheap sugar-alcohol of emoji reactions and story views. In other parts of the world, young people are just starting to break free of this collective trap, but not yet in the United States.
But making teenagers sad is one thing; putting an end to the Human Age on planet Earth is quite another.
The global fertility decline is a long-standing trend. Every country that escapes poverty, urbanizes, and teaches its people to read is going to transition from a high fertility rate (5-7 children per woman) to a much lower rate. Long before the smartphone burst on the scene, most of Europe and the richer parts of East Asia had fallen below replacement-level fertility. Everyone would crack jokes about Japan not having enough kids.
2026-06-10 15:10:26
In a post last week, I wrote about the progressive anti-monopoly movement’s increasing disconnect from reality. I wrote:
[C]onsider the movement’s choice of targets. These include some industries with high profit margins, but also some with very low margins. These include grocery stores, airlines, and health insurers. Grocery stores and health insurers both consistently have much lower profit margins than American corporations in general, often hovering near the zero mark.
Commenter Matthew argued that the low profit margins of insurers are not a reason not to worry about their market power:
The idea that health insurers have “low margins” so they are OK is nuts…Private health insurers in the US do not lower costs and do not improve patient care…In the flow of money between patients and providers, private insurers just sit in that flow like a tapeworm and take money out to sustain themselves…
There is a lot of evidence…[W]ith the current status quo, 10 -15$ out of every 100$ of healthcare premiums a person spends is just going to the private insurer….That would be fine if the insurance companies secured lower costs for their members; it would be the useful service they provide…But there is no evidence that they do.
Matthew’s argument doesn’t really address the point of my post. Private insurers might be inefficient, or even unnecessary, but this is very different from them being extractive monopolies. It is absolutely incredibly relevant that health insurers have very low profit margins. If $10 of every $100 spent on health care premiums goes to the insurer, but the insurer isn’t profitable, this just means that the $10 is going to cover the insurer’s operating costs. It is not money being funneled into the pockets of the people who own the insurance companies.
In fact, the more general fact here is that private insurers are not the main reason why American health care costs so much more than health care in other developed nations. Almost all of the excess cost goes to providers rather than to insurers. Private insurers may be an unnecessary middleman, but the amount they extract from the system is not large compared to the amount that gets either appropriated or wasted by the people providing the care.
So why do Americans — especially American progressives — focus so obsessively on health insurers instead of health providers? In a post two years ago, I hypothesized that it’s because insurers are the part of the system we have direct contact with — the people who have to tell us “no” when we can’t afford some treatment.
Insurers have thus become what Jeremiah Johnson calls “sin-eaters” — the hapless fall guys who bear the brunt of all Americans’ rage, despair, and frustration at a broken system in which the insurers play only a very minor role.
The more progressives focus on venting rage and making accusations at insurance companies, the less effective they will be in actually delivering Americans cheaper health care.
Anyway, here’s the post I wrote back in 2024, which fleshes this all out in greater detail.
“I’d rather die than owe the hospital til I get old” — Courtney Barnett
When UnitedHealthcare CEO Brian Thompson was gunned down in the street in cold blood the other day, a bunch of people on the internet gloated and cheered:
The jokes came streaming in on every social-media platform, in the comments underneath every news article. “I’m sorry, prior authorization is required for thoughts and prayers,” someone commented on TikTok, a response that got more than fifteen thousand likes. “Does he have a history of shootings? Denied coverage,” another person wrote, under an Instagram post from CNN. On X, someone posted, with the caption “My official response to the UHC CEO’s murder,” an infographic comparing wealth distribution in late eighteenth-century France to wealth distribution in present-day America…On LinkedIn, where users post with their real names and employment histories, UnitedHealth Group had to turn off comments on its post about Thompson’s death—thousands of people were liking and hearting it, with a few even giving it the “clapping” reaction. The company also turned off comments on Facebook, where, as of midday Thursday, a post about Thompson had received more than thirty-six thousand “laugh” reactions.
In general, I think it’s a very bad look to endorse murder. And I think this kind of thing is a sign of how stressed-out and mentally unbalanced our country is after an era of unrest. (The chief suspect, who was just apprehended, looks like a random crazy guy rather than a leftist ideologue.)
But more fundamentally, I think the outpouring of schadenfreude1 at Thompson’s killing reflects some deep-seated popular misconceptions about the U.S. health care industry. A whole lot of people — maybe even most people — seem to regard health insurance companies as the main villains in the system, when in fact they’re only a very minor source of the problems.
All my life, Americans have been raging at health insurers. Who could forget this clip from the 1997 movie As Good as It Gets?
It’s not hard to understand why people hate health insurers. When you interact with the U.S. health care system, the providers — the hospital staff, the doctor, the nurses, the technicians — all just take care of you. The only time they ask you for money during your doctor visit is when you pay your copay at the front desk, and that’s usually not that big — if the bill is big, they’ll send it to you later. So for the most part, your interaction with the providers is just you walking up and asking to be taken care of, and them taking care of you.
Your interaction with the health insurer, on the other hand, feels like a struggle against an enemy who wants to destroy you. If you get a big hospital bill days after your visit, it’s because the insurer wouldn’t cover the whole cost. If the bill is a surprise because the provider didn’t tell you they were out of network, that also feels like the insurance company’s fault — why wasn’t that provider in their network?
Even more terrifying is when insurers deny coverage completely, which happens to about 10-20% of claims. It feels like you’ve been robbed. You paid this company a hefty premium every month, and in exchange you expected them to pay for your health care if you needed it. And now you needed it, and they won’t even uphold their end of the bargain? Why were you even paying them the premium in the first place?
Everyone knows that denying claims is in the insurance company’s financial interest. The more they can get away with taking your monthly premium and then weaseling out of their end of the bargain, the more their shareholders and executives can walk away with giant bags of money. They’re the ones buying huge houses and yachts and whatever on the money they made from finding some technical reason to send you and thousands upon thousands of people like you into medical bankruptcy after your chemotherapy. Who wouldn’t be mad?
And yet when we take a hard look at the question of why Americans pay so much more for their health care than people elsewhere in the developed world, insurance companies and their profits just aren’t that big of a piece of the story.
First of all, insurance companies just don’t make that much profit. UnitedHealth Group, the company of which Brian Thompson’s UnitedHealthcare is a subsidiary, is the most valuable private health insurer in the country in terms of market capitalization, and the one with the largest market share. Its net profit margin is just 6.11%:
That’s only about half of the average profit margin of companies in the S&P 500. And other big insurers are even less profitable. Elevance Health, the second-biggest, has a margin of between 2% and 4%. Centene’s margin is usually around 1% to 2%. Cigna Group’s margin is usually around 2% to 3%. And so on. These companies are just making very little profit at all.
Here’s another way of visualizing that:
You can see that the company’s net income — i.e., its total profit — was $23.1 billion in 2023. That’s a lot of money, but it pales in comparison to the $241.9 billion that the company spent on medical costs. Even the company’s $54.6 billion in operating costs — of which Brian Thompson’s own $10 million salary represented 0.018% — are dwarfed by actual medical costs.
In fact, the actual health insurance business — taking premiums and paying out claims — is even less profitable than these numbers might suggest. As Axios recently reported, insurers’ profits are increasingly coming from other lines of business.
What does this mean? It means that if UnitedHealth Group decided to donate every single dollar of its profit to buying Americans more health care, it would only be able to pay for about 9.3% more health care than it’s already paying for. If it donated all of its executives’ salaries to the effort, it would not be much more than that.
What about those denials of coverage, copays, deductibles, and so on? In fact, Americans are paying a smaller percentage of their health costs out of pocket than people in most other rich countries!
Note that the song lyric at the top of this post, about a woman in anaphylactic shock worrying that she won’t be able to afford her hospital bills, is from a band in Australia, not the U.S. This isn’t a coincidence — although Australian medical costs are fairly low, the proportion they pay out of pocket is unusually high.
In other words, Americans’ much-hated private health insurers are paying a higher percent of the cost of Americans’ health care than the government insurance systems of Sweden and Denmark and the UK are paying. The only reason Americans’ bills are higher is that U.S. health care provision costs so much more in the first place.
On top of all that, health insurance companies don’t actually look very inefficient, in terms of their administrative costs. Yes, we all know that the fragmented U.S. health system is a paperwork nightmare, with different providers and insurers drowning each other in forms and approvals. And Elizabeth Warren has claimed that switching to national health insurance would save huge amounts of money by reducing administrative costs. But when we look at United Health Group’s operating costs in the diagram above, they’re only 22.6% of the actual cost of medical care.
In fact, the Kaiser Family Foundation does detailed comparisons between U.S. health care spending and spending in other developed countries. And it has concluded that most of this excess spending comes from providers — from hospitals, pharma companies, doctors, nurses, tech suppliers, and so on:
This means that eliminating all administrative waste and inefficiency in the entire U.S. health care system — not just at insurance companies, but administration of government insurance programs — could save Americans at most about $680 per person every year. And the true savings would probably not anywhere close to that amount — part of America’s greater costs are certainly an income effect, due to the fact that Americans have higher incomes than people in other rich countries.A few hundred bucks a year is not nothing, but it’s only a small fraction of the $5683 more that we pay relative to other countries.
So the fundamental reason your health care costs so much is not that the health insurance companies are lining their pockets. And it’s not that insurers are an inefficient mess. It’s that the actual provision of America’s health care itself just costs way too much in the first place.
The actual people charging you an arm and a leg for your care, and putting you at risk of medical bankruptcy, are the providers themselves. The smiling doctor who writes you prescriptions and sends you to the MRI and refers you to a specialist without ever asking you for money knows full well that you’re going to end up having to wrangle with the insurance company for the cost of all those services.
The gentle nurse who sets up your IV doesn’t tell you whether each dose of drugs through the IV could set you back hundreds of dollars, but they know. When the polite administrative assistants at the front desk send you back to treatment without telling you that their services are out of your network, it’s because they didn’t bother to check. The executives making millions at “nonprofit” hospitals, and the shareholders making billions on the profits of companies that supply and contract with those hospitals, are people you never see and probably don’t even think about.
Here’s a good Bloomberg story on predatory pricing by hospitals. Hopsitals bilk insurers, who are forced to pass on the higher costs to patients, who then blame the insurers instead of the hospitals.
Excessive prices charged by health care providers are overwhelmingly the reason why Americans’ health care costs so cripplingly much. But they’ve outsourced the actual collection of those fees to insurance companies, so that your experience in the medical system feels smooth and friendly and comfortable. The insurance companies are simply hired to play the bad guy — and they’re paid a relatively modest fee for that service. So you get to hate UnitedHealthcare and Cigna, while the real people taking away your life’s savings and putting you at risk of bankruptcy get to play Mother Theresa.
So the way to make our health care system affordable is not to browbeat insurers, in the hope that they will be able to reduce their profits and pay for us to have cheap health care. Insurance companies simply do not have the power to do that, even if you threaten to shoot them. What we need is to reduce costs within the actual medical system itself. One idea is to have the government insurance system play hardball with providers, negotiating lower prices. is what the Biden administration had Medicare do with some drug companies. There are some risks to this approach — if it’s executed clumsily it can suppress innovation — but it’s basically what every other rich country does, so the track record is decent. There are probably other ways to foster competition and increase efficiency in the medical care system.
But focusing all our anger on the middlemen of the U.S.’ bloated health care system is just a way of shooting the messenger.
Update: Matt Bruenig argues that insurers really are to blame. He claims that inefficiency in the system is a bigger driver of costs than I realize, because providers have to spend money dealing with insurers — something the KFF numbers don’t include. He also alleges that some of this “inefficiency” is actually intentional on the part of the insurers — a disguised way to pay themselves out.
But I don’t think this passes the smell test. If insurers were so good at extracting money from the system, why are they so unprofitable? The average company in the S&P 500 has a profit margin of 12%, but these insurers have margins of 1% to 3%. If they’re so good at extracting money from providers, why are shareholders — and executives who own stock — not getting a piece of that money? It doesn’t make sense.
Also, provider-side administrative costs don’t just include the money providers spend wrangling with insurers — it includes the money they spend on the executives, managers, and billing departments who figure out how to charge patients $700 for every injection of IV drugs, or hundreds of dollars for a hospital pillow, or $10,000 for an MRI.
So no, I don’t buy Bruenig’s argument here. Though I still do agree with him that national health insurance would be a good idea — mostly for the negotiating power, not for the administrative cost savings.
Update 2: Over at Tyler Cowen’s blog, a commenter argues that profit margins are not a good guide to the financial success of a business, and that instead one should look at return on equity (ROE). That’s fine, I wasn’t talking about the financial success of health insurance companies; I was simply showing that they don’t have the ability to pay for much more health care than they’re currently paying for. But for what it’s worth, the ROE of health insurers is pretty low. The S&P 500’s weighted average ROE is usually around 15%, and has been a bit higher lately. Here’s what health insurers were earning:
They had a couple of good years in there, but in general it’s underwhelming. UnitedHealth Group, which does include businesses other than the low-margin health insurance business (e.g. providing health care and pharmaceuticals), is doing pretty well with 26%.
But if you look at the list of companies with the highest ROE, you see health care providers or suppliers like HCA Healthcare (272%), Cencora (234%), Abbvie (84%), Mckesson (84%), Novo Nordisk (72%), Eli Lilly (59%), Amgen (56%), IDEXX Laboratories (53%), Zoetis (46%), Novartis (44%), Edwards Lifesciences (43%), and so on. If you want to know which shareholders are making the real money in the health care industry…well, it’s the shareholders of those providers and suppliers.
2026-06-08 15:42:08
Howdy, folks! Today’s roundup is mostly a bunch of follow-ups to posts I wrote before. It’s very hard to decide when to post about a particular topic, and it often happens that some relevant news story or piece of data comes out a little bit later. These roundups are a good way of cleaning up those loose ends.
Today we start with a truly wacky policy proposal by the esteemed Thomas Piketty…
Unlike many people, I never pretended to have read Thomas Piketty’s book, Capital in the Twenty-First Century. I simply didn’t read it. I did read a number of the papers that the book was based on, which is often a better and quicker way of getting the key points of a book like that. I thought those papers were a good and important addition to the economics literature, even if the messy reality of inequality didn’t always fit the simple story Piketty told, and the data he relied on was less reliable than we might want.
Despite the limitations of Piketty’s work, it sparked a long-overdue and generally healthy debate about inequality. And Piketty’s basic policy solution — tax rich people more — was pretty reasonable, even if his proposed numbers were too extreme. I did roll my eyes when Piketty stood on a stage at an academic convention and accused Greg Mankiw of being in the pocket of rich people.1 But overall his work seemed pretty serious and often reasonable.
However, after years of relative silence, Piketty has burst back onto the scene with some work that seems very unreasonable. He and his team at the World Inequality Lab — which includes his longtime co-authors Emmanuel Saez and Gabriel Zucman — have come out with a grand plan for fixing the world. And for the most part, it’s total nonsense.
Piketty described the new plan in a thread on X. Its main focus, perhaps surprisingly, is not inequality — it’s climate change!
First of all, Piketty’s baseline climate change scenarios appear based on a very outdated model — the RCP8.5 scenario, an extreme projection that essentially all serious climate scientists have now rejected. This choice of baseline suggests that Piketty et al. were trying to find ways to justify maximal policy intervention, instead of starting from the science.
Piketty’s preferred solution to climate change is degrowth. He envisions detailed central planning to achieve deliberate impoverishment of large portions of the world’s population — mandated reductions in the consumption of various specific goods, including food.
In addition to the dubious morality of deliberately impoverishing untold millions of human beings based on scientific models that have already been rejected, this kind of scheme is just utterly unworkable. Back in 2021, when I wrote about why degrowth is a political nonstarter, I declared that “implementing the kind of reallocation schemes that degrowthers throw around with abandon would require global economic planning that would put Gosplan to shame.” Piketty knows this, and thinks it’s a good thing.
Even more ridiculously, Piketty envisions a global fiscal authority to carry out this insane plan via global taxation:
Let’s set aside the obvious fact that countries are just not going to agree to give up their spending and taxation power — even the EU refuses to have a fiscal union, and it’s rather insane to imagine Indians and Chinese people agreeing to let themselves be taxed by Tanzania and Nigeria — and just point out how this proposal ignores the basic economics of climate change.
Climate change is a global negative externality — the reason countries don’t all just impose their own local carbon taxes and solve the problem is that there’s an incentive to free ride and let other countries handle it. That exact same free rider problem applies to the global fiscal authority that Piketty envisions. There’s a clear incentive for any country to simply drop out of the fund and let other countries fix climate change for them.
It’s obvious that Piketty et al. are just looking for a reason to levy high taxes on the global rich. This is the “World Inequality Lab” we’re talking about here. And it probably made sense to try to ally with other factions of the progressive movement — degrowthers, “decolonial” leftists, and so on — in order to get support for their desired policies.
But the result here is not going to be a good one for Piketty, Saez, Zucman, and their team. No country is actually going to embrace the idea of a global fiscal authority to fight climate change. In calling for this sort of thing, Piketty et al. simply make themselves look less like serious economists and more like opportunistic activists on the fringe of a “green” movement that’s already in steep decline.
In a post this week, I noted that “tokenmaxxing” — simply using as much AI coding output as you can and hoping that it pops out something valuable — is hitting its limits:
Well, here’s a follow-up. John Burn-Murdoch of the Financial Times recently made this nice chart, using data from the Demirer et al. (2026) paper that I discussed in my post:
The number of apps with significant usage is actually going down in the age of AI, even as people are releasing floods of new apps into the world. Meanwhile, Bob Elliott notes that since generative AI was created, there has been a rapid acceleration in many measures of text output, even though the economy hasn’t accelerated much:
And Sam Altman is now warning of a significant pullback on AI spending — the first such pullback since generative AI appeared.
This doesn’t look like a simple story of “bottlenecks” and “weak links” — if it were that, we wouldn’t see so many new apps and e-books hitting the market. The deeper story here may be that demand for many of the things that generative AI produces might be a lot more inelastic than we thought. The things we really want a lot more of may not actually be the things that generative AI is yet equipped to provide. As the AI industry advances, of course, that will probably change.
A couple of weeks ago, I wrote a post about how all militaries not based around large masses of cheap drones are now functionally obsolete:
This includes America’s military, which is based around a few big expensive “platforms” like fighter jets, aircraft carriers, and tanks. I’m not saying those weapons will all be useless in future wars, but if that’s all you have, and you don’t have masses of cheap drones, you will lose wars to countries that do have masses of cheap drones — such as China, if they ever get serious about turning their mighty industrial base toward making billions of weaponized drones.
The Lowy Institute has a good report explaining why Western militaries seem incapable of learning to use the essential weapons of modern warfare. They write:
Western military institutions…are failing to energetically learn from modern wars. Despite four years of unprecedented visibility into Ukrainian battlefield innovations, and the recent war in Iran, Western forces have not institutionalised key lessons into doctrine, force structure, or procurement priorities…The recent war in Iran has confirmed and amplified many of Ukraine’s lessons, particularly on the centrality of drone warfare, the inadequacy of Western counter-drone capabilities, [and] the effectiveness of cheaper long-range strike systems…And yet the response of Western institutions…has been characterised by rigidity, inertia, and what can be called a humility deficit: an unwillingness to genuinely confront the implications of what is being demonstrated in real time on real battlefields.
The U.S. military could, of course, learn from Ukraine — currently the #1 best country in the world in drone warfare. But the unrelenting hostility and disdain toward Ukraine from Donald Trump and the MAGA movement has prevented America from taking advantage of Ukraine’s expertise:
The Trump administration’s hesitancy in signing a major drone deal with Ukraine is slowing the U.S. military down in an area where it’s already trying to play catch-up…[T]he U.S. has so far refused to embrace Kyiv as a partner in its drone development…
[E]ven with senior Pentagon officials — including Defense Secretary Pete Hegseth and Army Secretary Dan Driscoll — lauding Kyiv’s drone abilities, the Trump administration is still biding its time on taking full advantage of the Ukrainian capabilities, a delay that experts say is potentially kneecapping the U.S. military…
“I don’t know what the hang-up would be in denying ourselves the ability to take advantage of that. I don’t think there’s any good reason,” Rebeccah Heinrichs, a senior fellow at the Hudson Institute think tank, said of Ukraine’s drone capabilities…One former official [called] the hold-up “lethargy” on the part of the Trump administration and “a certain amount of hostility towards Ukraine coming from the very top.”
MAGA basically created a fantasy world where Russia is a defender of Western values, Ukraine is somehow an arm of global wokeism, Ukraine is part of Russia’s legitimate “sphere of influence”, and Russia is a mighty superpower with a manly martial culture that would eventually be able to grind the Ukrainians down and inevitably triumph.
The problem with this fantasy was that it was fantasy, and if you believe in fantasy too long, reality tends to intercede. By allowing themselves to believe their own anti-Ukraine mythology, Trump and his followers are cutting themselves — and the U.S. Military — off from access to crucial modern military technology.
In my last post, I argued that Europe should put tariffs and other trade barriers on Chinese imports, in order to protect its own strategic defense-related industries. But this is actually a lot harder than it sounds. Even if Europe blocks final goods from China, China can still export intermediate goods to “third countries” that assemble those goods for final export to Europe. In fact, China has done this in response to American tariffs, reducing (though not eliminating) the decoupling effect.
But if that happens, it’ll be very good for the “third countries”! Assembly work isn’t the most valuable part of the supply chain, but it does create value, and it does create lots of jobs, and it does create local companies that then have the potential to climb up the value chain someday and start making their own components. In fact, this is exactly what China did! Back in the 2000s, China did a lot of the low-value assembly work for components made in Japan, Korea, and Taiwan; now, most of that has been onshored, but it was still important for China to go through that initial phase of learning to slap together iPhones and computers and cars.
So if putting tariffs and other trade barriers on Chinese-made goods just ends up shifting assembly to poor countries…well, that’s not the worst outcome in the world. It’ll help counteract Chinese companies’ home bias — their natural tendency to want to build factories in China instead of overseas.
In fact, as the WSJ reports, this is already happening:
“Made in China” is becoming “made by China”—all over the world…Faced with higher Western tariffs and weak demand at home, many Chinese factories are moving abroad, making everything from appliances to automobiles everywhere from North and South America to Eastern Europe…In Mexico, Chinese investment in industries such as the automobile sector generated more than 100,000 jobs from 2020 to 2023, according to one analysis…In 2024, Chery Automobile, China’s top car exporter, helped to rescue a small factory in Barcelona that struggling Japanese automaker Nissan no longer wanted…
Jeep maker Stellantis this month said it planned to build EVs with two separate Chinese companies in Spain and France. Ford and Geely are in discussions about a potentially similar deal in Spain, and have also discussed whether the collaboration might extend to the U.S…Midea, the home appliance maker…opened a roughly $100 million factory in Brazil making refrigerators and washing machines in 2024. Its subsidiary, Welling Auto Parts, opened its first overseas manufacturing facility in Mexico last year.
In order to get around EU tariffs, Chinese companies are fueling a Moroccan manufacturing boom:
This has helped accelerate Morocco’s growth to 5%. That’s in the range where growth starts meaningfully transforming a country.
So even in the worst-case scenario where trade barriers don’t reduce dependence on Chinese supply chains, they can help spread the blessings and bounty of industrialization to a bunch of poor countries who need the factories more than China does. The flying geese must fly!
I recently came across this chart, showing various aspects of India’s infrastructure boom:
This is all pretty incredible. India’s poor infrastructure has long been regarded as a bottleneck to urbanization, manufacturing, and economic growth in general. Whatever else you think of the government of Narendra Modi, it has built a lot of infrastructure.
But over that same period, overall growth has been slower than we’d like to see. Anand, Felman, and Subramanian have a recent paper in which they argue that India’s GDP growth rate from 2011 to 2023 was overstated by somewhere between a quarter and a third:
India’s annual economic growth during the boom years between 2005 and 2011 may have been underestimated by about 1–1½ percentage points on average, and subsequent growth between 2012 and 2023 may have been overestimated by about 1½-2 percentage points…The first methodological issue leading to the misestimation is that the economy’s formal sector has been used as a proxy for the vast informal sector, even though the latter was disproportionately hit after 2015 by demonetization, the introduction of the goods and services tax, and the COVID-19 pandemic…The second methodological issue…is that the deflators for many sectors have been based on commodity prices, which have moved sharply relative to others. [emphasis mine]
If Anand et al.’s estimates are right — and they marshal a huge amount of supporting evidence — then it suggests that Modi’s tenure in office has been mixed. A couple of big policy missteps — demonetization and a botched tax rollout — hurt the informal sector of the Indian economy, while massive infrastructure investments have helped.
The implication here is that Modi and his successors should lean into what works. They should focus more on marshaling national resources and applying those resources toward rapid growth — two things that China did very well in the 1990s and 2000s.
I’ve been writing over the years about how the right’s favorite immigration economist does shoddy, subpar work. Despite having a job at Harvard, George Borjas — whose analyses miraculously always seem to find that immigration is much worse than all the other economists think it is — consistently uses both poor data and flawed methodology. In another roundup back in February, I pointed out how Jianxin He and Adam Ozimek had found yet another example of Borjas doing subpar economics:
Borjas’s February 2026 working paper attempted to answer whether H-1B workers earn less than comparable native-born workers…[His] findings result from substantial data errors.…The most significant mistake is a…mismatch between his H-1B and native-born samples: the H-1B applications span 2020-2023, while the ACS data covers just 2023…[Accounting for this discrepancy cuts] the wage gap roughly in half…
The second error stems from controlling for geographic wage drivers using each worker’s PUMA (public use microdata area)…The problem is that Dr. Borjas uses the PUMA where visa holders work alongside the PUMA where native workers live. Consider a native-born software developer working at Google in Mountain View who resides in a cheaper area like Fremont. If residential areas have lower average wages than business districts, this mismatch systematically inflates the apparent native wage and negatively biases the H-1B wage gap.
Again and again and again, economists catch Borjas at it. It seems pretty obvious that Borjas simply wants to conclude that immigration is bad, and doesn’t much care about methodological errors as long as they reach his desired conclusion.
In order to fight back against this accusation, Borjas decided to accuse his critics of ideologically-driven research instead. In a paper with Nate Breznau, he wrote:
Our study exploits an opportunity to observe 158 researchers working…during an experiment. After being asked their position on immigration policy, they used the same data to answer the same empirical question: Does immigration affect public support for social welfare programs? The researchers estimated 1253 alternative regression models, and the estimated impacts ranged from strongly negative to strongly positive. We find that teams composed of pro-immigration researchers estimated more positive impacts of immigration on public support for social programs, while anti-immigration teams estimated more negative impacts. The differences arise because different teams adopted different model specifications. The underlying research design decisions are the mechanism through which ideology enters the process of producing parameter estimates.
The idea here seems to be to turn one researcher’s clear pattern of errors into a he-said/she-said sort of situation. If all researchers just engineer results based on their ideology, then why should we selectively get mad at Borjas for doing what everyone else does too?
But — surprise! — it turned out that this Borjas paper also contained critical errors that invalidated the whole result! Katrin Auspurg and Josef Brüderl pointed out in a comment paper that if you fix one simple coding error in Borjas’s analysis, his entire result about ideologically-driven research just vanishes into thin air:
Borjas and Breznau…recently reported that researchers’ ideology influences their empirical findings. Although we were able to reproduce B&B’s numerical results, our reanalysis shows that the reported association is not robust. Specifically, the association hinges on a coding error. Data from four teams that contradict the ideology hypothesis were excluded from the analysis due to idiosyncratic variable coding. Correcting this error renders the ideology effect no longer statistically significant. Also, B&B employed a different outcome variable and weighting scheme to that used in a previous paper based on the same data. These two analytical decisions further contribute to the observed ideology effect. Correcting the coding error or using the same specification as in the previous paper renders the ideology effect indistinguishable from zero. Therefore, we conclude that B&B do not provide robust evidence of ideological bias in this context. Instead, the reported association appears to be a statistical artefact resulting from questionable modelling decisions. [emphasis mine]
How does this just keep happening again and again, and why is it always Borjas?
In any case, I think the implication here is pretty clear: Friends don’t let friends cite George Borjas.
Greg Mankiw makes his money by selling textbooks.
2026-06-06 16:10:50
As regular readers of this blog know, I’m pretty ambivalent about trade barriers as an economic policy. On one hand I think targeted tariffs and other trade barriers can be used to protect strategic industries from surges in underpriced import competition, especially by geopolitical rivals. On the other hand, broad tariffs like the ones Trump has used are generally bad — they hurt domestic manufacturing by making intermediate goods more expensive, they limit scale for domestic companies, etc.
And yet I do think that Europe should erect much higher trade barriers — both tariffs and non-tariff barriers — against Chinese high-tech manufactured export goods. The basic reason is that it’s important to protect Europe’s nascent modern defense industry. But I also think that blocking Chinese exports might nudge China to change its economic model to one that benefits regular Chinese people more.
In other words, China-Europe trade has some unusual characteristics right now that make trade barriers a much smarter idea than usual.
First, let’s talk about what’s going on with the Chinese economy. For the past few years, China’s government has unleashed an unprecedented torrent of subsidies for high-tech manufacturing industries. This — along with structural factors about how the Chinese economy works — has resulted in China making big global market share gains in industries like autos, pharmaceuticals, and shipbuilding. No one knows just how much of China’s market share gains are a result of government support, but as Paul Hannon reports, the OECD estimates that it’s more than half:
Government subsidies have driven most of the increase in the global market share of Chinese businesses over the past two decades as they have received three to eight times more support than their competitors, the Organisation for Economic Cooperation and Development said Monday…The analysis is based on the OECD’s Manufacturing Groups and Industrial Corporations database, which includes subsidy estimates and financial information for 525 of the world’s largest manufacturing groups spread across 15 key industrial sectors…[T]he OECD database tracks the amounts that firms are actually given…
“Industrial firms based in China receive more subsidies than their competitors based everywhere else,” the OECD said…For Chinese businesses, however, the share of [market share] gains explained by subsidies was…60%.
The Rhodium Group has a deeper dive into China’s new industrial policy. Essentially, instead of selecting a few industries to specialize in, China’s leaders just want the country to dominate everything — not just manufacturing, but services as well:
China’s industrial strategy…is becoming more systemic and pervasive, extending across all layers of production, from upstream inputs and industrial equipment to downstream applications, services, and frontier technologies…China’s next-generation industrial policy represents a shift from targeted sectoral intervention to what can be described as an “industrial policy of everything.”…While [Made in China 2025] focused on a defined set of strategic emerging industries, current policy frameworks extend across mature sectors, foundational supply chain nodes, and frontier technologies alike…
Even in mature industries facing overcapacity and severe price pressures, Beijing is providing continued support and pushing firms to upgrade production technologies to gain market share and lower production costs, rather than cutting capacity…Services, relatively neglected in earlier rounds of industrial policy, are getting more attention[.]
Basically, China does not want1 to exist in a trading system, where goods are traded for other goods. China wants to make all the goods, and have other countries pay for those goods with debt.
There are two basic reasons China is doing this. The first is pure mercantilism; China is trying to export its way out of the economic slump created by its housing bust. The second, as the Rhodium Group report explains, is power. If China controls key segments of other countries’ supply chains, it can use the threat of export controls to bring those countries to heel.
What should other countries do about this? The U.S. has chosen to respond with tariffs. These are of limited effectiveness, but they do appear to be doing something; even when you take into account the intermediate goods that China exports to America via third countries like Vietnam and Mexico, China’s share of America’s imports has fallen slightly from 2021 (or from 2017):
There are almost certainly much more effective tools that the U.S. could use to accelerate the decoupling of the two economies and reduce dependence on China…but since when has U.S. policy been driven by a desire for effectiveness?
The question is now what Europe and other developed countries — who have marginally more rational decision-making processes — are going to do about China’s attempt to dominate all tradable industries. One proposal — which Germany seems to be following so far — is to do nothing, and to simply let China make all the physical objects in the world, while focusing on services instead. This is essentially the proposal of Tej Parikh, who writes that China “has a comparative advantage in industrial policy itself”, and that trying to compete with China in any manufacturing industry is therefore doomed to fail.
This annoys me, because it represents a deep misunderstanding of the entire concept of comparative advantage! The theory of comparative advantage is about traded goods; it’s about which traded goods can be produced relatively more cheaply by which countries. If I’m better at making TVs than cars, and you’re better at making cars than TVs, then I’ll make TVs and you’ll make cars and then we’ll trade. That’s how comparative advantage works. This is why you cannot have a “comparative advantage in industrial policy”. Industrial policy is a production input, not a traded good. No one buys and sells industrial policy!
“OK, Noah,” you’re about to say. “Stop being a pedant. You know what he means. He means China is better at making anything and everything, because they use industrial policy for everything.”
Yes, I know that’s what he means. And yes, this reflects a deep misunderstanding of the concept of comparative advantage.2 Even if one country is better at making everything, it doesn’t have a comparative advantage in everything. That’s impossible. Every country has a comparative advantage at something!
That’s why in the theory of comparative advantage, trade is balanced. In the real world, China’s massive trade surplus means that trade is not balanced; much of the time, China isn’t trading goods for other goods, it’s trading goods for IOUs. That kind of unbalanced trade is something that just doesn’t happen in the theory of comparative advantage.
OK, so that was a bit of a rant. The real point here is that Parikh’s preferred solution — that every country except China should focus on innovation, and leave the making of everything to the Chinese — is simply ridiculous. First of all, it doesn’t deal at all with the issue of supply chain vulnerabilities. Second of all, China has an industrial policy for innovation, too — in fact, it’s China’s most important industrial policy. The idea of “We’ll do the innovation while China makes everything” sounds straight out of 2002 — and it was obviously wrong even back then.
The cold, hard fact is that Europe needs to do something, or risk losing its sovereignty to foreign conquerors. China — the very country that Europe’s free-traders are now suggesting should supply every single manufactured good — is waging a proxy war against Europe even as we speak. China trains Russian soldiers, provides Russia with battlefield intelligence in its war against Ukraine, helps out Russian defense manufacturers, and even does some defense manufacturing for Russia — in addition to buying Russian oil and keeping the Russian economy afloat.
And this is all while Russia is actively threatening to invade the EU. If Russia eventually does invade, Europe will need to make large amounts of drones to resist the invasion. All militaries that are not centered around large masses of drones are now obsolete — when NATO conducts war games against drone-equipped Ukrainian units, the Ukrainians easily triumph.
But both Europe and Ukraine cannot currently make drones from scratch without relying on Chinese industry. Many of the components and materials that go into making a drone are controlled by China — things like radio modules, lithium-ion batteries, electric motors, navigation cameras, and even carbon frames. Europe cannot currently make these — or can’t make many of them, at least.
If Russia were to invade Europe, China could simply decide not to sell Europe the components it needs to make drones. Why wouldn’t it? China has already proven itself perfectly willing to use export controls on rare earths and other upstream technologies to throttle other countries’ defense industries. And a Europe cowed and dominated by China’s most important ally would probably be more useful to Xi Jinping than a free and independent Europe that steers its own destiny.
If Russia invaded Europe and China simultaneously halted the export of drone components, Europe would be a lost cause. Unless Europe could assemble upstream industries for drone components from scratch before Russia’s drone-equipped armies marched across the Baltics and into Poland, the war would quickly be lost for lack of weapons.
Whether they realize it yet or not, Europe’s dependence on China for the manufacture of many key defense inputs puts it at China’s mercy. This is a downside to free trade that the folks who advocate a European retreat from manufacturing simply fail to engage with or acknowledge. It provides a strong rationale for putting up trade barriers against the import of certain intermediate goods — something that harms economic efficiency, but is necessary for defense.
When invading armies are burning your country to the ground, you should worry less about deadweight loss than about being dead.
But those who wring their hands about the economic losses should take heart. Blocking the import of Chinese goods might harm economic efficiency, but it could have some positive knock-on effects in terms of political economy.
For all China’s high-tech wizardry, its big industrial policy push doesn’t seem to be doing much to help the actual people of China.
The real estate industry, which previously created plenty of labor demand and broad-based wealth for regular Chinese people, is still in the dumps and may even be getting worse. The continued property bust is weighing on aggregate demand — Fixed-asset investment is shrinking, while retail sales have flatlined.
“Industrial policy for everything” was supposed to fill the hole left by real estate, but it isn’t doing a very good job of it. Because the rise in Chinese manufacturing output is being done mostly for export, regular Chinese people aren’t able to share in the bounty the policies are creating. For example, Chinese motor vehicle consumption is below where it was a decade ago, despite surging exports:
In fact, this shift dates back to the pandemic. Matt C. Klein has a good series of charts on China’s anemic consumption. Here’s an example:
This is often framed as China helping producers at the expense of consumers. But often it’s not even that. China’s industrial subsidies pay a bunch of different companies to produce the same goods, competing their profits to zero even as they also undercut the overseas competition. A prime example of this is the solar industry:
China’s solar exports have enjoyed a surge since the bombing [of Iran] began. But that will be small cheer to its companies…Domestic demand for their products is falling for the first time in decades because the country’s power grids—far and away the biggest market for solar panels—have become overloaded with the things. Solar-panel supply, meanwhile, is overabundant because of years of splashy investment in factories…Most companies have been running at a loss since 2024 because of brutal price wars; bankruptcies are mounting.
But it’s not just undifferentiated commodity products like solar that are suffering this fate; China’s vaunted auto industry, which came out of nowhere to leapfrog all other countries with its mastery of EVs, is locked in an endless brutal price war:
China’s efforts to cool its automotive price war are faltering as BYD Co. and rivals expand discounts to avoid ceding ground in the world’s largest car market…The average price reduction for BYD cars accelerated to 10% in March…Discounts by competitors…also edged higher…Regulators’ missives aimed at halting deflationary momentum have fallen on deaf ears so far, and industry observers say it won’t stop the discounting trend anytime soon.
China’s industrial policy is accomplishing its central goal of national greatness. China’s technology level is advancing, its companies are winning global market share, and it’s gaining control over key strategic technological choke points. But China’s workers, its savers, its investors, and even its entrepreneurs are on a treadmill, unable to enjoy the fruits of their country’s industrial dominance.
European trade barriers could potentially nudge China out of this toxic political economy. If Xi Jinping & co. see that they can’t forcibly deindustrialize the West by subsidizing infinite exports, their cost-benefit calculations may shift. Providing growing living standards for Chinese people might once again become the central goal of policy, as it was during the time of Deng Xiaoping, Jiang Zemin, and Hu Jintao.
So Europe should push back against the Chinese import flood, not just for their own security, but also for the sake of regular Chinese people. Fortunately, there are indications that European leaders have had enough of Xi’s little game, and are preparing to take real action. Hopefully this newfound resolve doesn’t get lost in the maze of European bureaucracy and inertia like so many other worthwhile initiatives.
This is a colloquial expression. Countries don’t want things; I’m taking about what the Chinese government, or at least Xi Jinping, wants for China.
Parikh is confusing comparative advantage with something called “competitive advantage”. In the theory of comparative advantage, competitive advantage — who makes which good more cheaply in the absolute sense — doesn’t end up mattering for the patterns of international trade. That’s why the theory is so brilliantly counterintuitive.
