Globalization’s Overlooked Economic Benefits

Antiglobalist ideas have motivated many Trump voters, but free trade benefits the average American

Letter to The WSJ.

"In his column “What Happened to the Pragmatic Trump of the First Term?” (Editor At Large, May 5), Gerard Baker wisely decries President Trump’s second-term pursuit of a misguided and extreme ideological agenda. Mr. Baker points out that antiglobalist ideas motivated voters in 2016—views that globalism “facilitated mass migration and the elevation of international capital that ravaged communities at home.”

A better name for “the elevation of international capital” is “free trade.” This term reveals the increased freedom of ordinary people to spend their incomes as they choose, while avoiding the mistaken suggestion that lowering trade barriers benefits only Davos-vacationing capitalists at the expense of the masses.

And where are these “ravaged communities at home” that voters were so worried about? Politicians and pundits still talk incessantly about these communities, but scholars who make serious attempts to locate them encounter difficulties. Economist Jeremy Horpedahl studied the 10 metropolitan statistical areas in the U.S. that suffered the largest negative hits during the infamous “China Shock” of the early 2000s. According to Mr. Horpedahl, all of the metropolitan statistical areas “hit hard by the China Shock still managed to have significant and positive real wage growth across the distribution since 2001 . . . Wage gains in several of these places, in fact, are better than the national trends.”

Whenever economic change occurs, some particular workers lose jobs, and some particular locations lose business and population. Economic growth requires economic change and adjustment. This has always been and will continue to be the case. But the story of America is that ordinary people recover over time and become wealthier. It’s an error to single out the freer trade of the past few decades as a unique source of economic change that justifies greater skepticism of globalization.

Prof. Donald J. Boudreaux

George Mason University

Fairfax, Va." 

Airlines and Overzealous Antitrust Enforcers

It’s a mistake to blame deregulation for Spirit Airlines’ demise

Letter to The WSJ. 

"Regarding your editorial “Spirit Airlines and the Antitrust Left” (May 4): Many people think that because carriers like Spirit Airlines have lower costs, they should be able to compete with the major airlines by offering lower fares. That’s a fallacy.

On any route involving a hub city of a major airline, the major airline’s network will support more flights (and therefore more possible connection options) than the low-cost carrier, which relies on point-to-point traffic. This product advantage, among others, generally allows the major airlines to charge and receive higher ticket prices than the low-cost carriers.

The revenue from these premium tickets will normally cover the cost of a network carrier’s flight before all the seats are sold. This means the remaining seats can be sold profitably at any price necessary to fill them. Unless travel demand is so high, or industry capacity so low, that major airlines can fill their planes at premium prices, it will generally make economic sense for them to match any price that a low-cost carrier offers if doing so is necessary to fill a seat.

The antitrust left is now blaming deregulation for Spirit’s demise. But there are far more airline flights, with more destinations served, at lower prices in real terms, than before deregulation. This is because deregulation allowed airlines to develop networks, that efficiently aggregate and distribute traffic through mergers, international alliances and organic growth.

During the era of deregulation, I was on the staff of the Civil Aeronautics Board, which regulated airline routes and prices until 1978. The architects of deregulation, Michael Levine and Alfred Kahn, didn’t know what form airline competition would take. They were confident, however, that business executives, freed from regulatory constraints, would find the most effective ways to increase output and reduce price. That is what has happened, despite resistance from regulators and occasional missteps by misguided judges. Unfortunately, the Biden administration’s antitrust enforcers and Judge William Young prevented Spirit Airlines and JetBlue from helping airline competition continue to evolve.

Ben Hirst

Wayzata, Minn.

Mr. Hirst is former executive vice president of Delta Air Lines." 

The U.S. Indicts a Mexican Governor

Charges made in New York mean President Sheinbaum will have to choose a side

By Mary Anastasia O’Grady. Excerpts:

"Mexican civil-society groups have long accused the political class of aiding and abetting cartels. Activists, family members whose loved ones are among the 130,000 gone missing since 2006, and journalists are some of the brave Mexicans who have tried to raise the consciousness of their nation about what they allege is a link between the gangsters who terrorize them and the state. It’s dangerous work."

"For more than a decade, the Cartel, under the rule of the Chapitos Leaders, and, before them El Chapo and El Mayo, has paid cash bribes to public officials at each level of the government, in exchange for protection of the Cartel’s drug trafficking operations. These corrupt government and law enforcement officials, including the defendants, are essential to the Cartel’s drug trafficking operations.”" 

Spirit Airlines and the Antitrust Left

A case study in how Lina Khan’s theories about competition failed in the real world

WSJ editorial. Excerpts:

"In 2022 JetBlue offered Spirit a $3.8 billion merger lifeline so the combined companies could offer more competition for the four U.S. airline giants. Mr. Kanter’s Antitrust Division sued to block the merger in 2023 and prevailed in court in January 2024"

"Federal Judge William Young admitted Spirit’s financial troubles. He also agreed that “an expansion of all aspects of JetBlue’s business—including network, fleet and loyalty program—would allow for more vigorous competition with the Big Four, which carry most passengers in the country.”"

"He still ruled the merger an antitrust violation because it would eliminate one low-fare option on some routes."

"Spirit declared bankruptcy in November 2024, long before the Iran war fuel spike. Now it’s shutting down for good."

"there will be less competition than if the merger had been allowed." 

Malta: A Free-Market Success Story

By Dan Mitchell. 

"I’m in Malta for a bit of research before speeches in Amsterdam and Reykjavik as part of the Free Market Road Show. 

So today is a good opportunity for a column on Malta’s rather-successful economy (something I’ve done for other countries, such as Poland, Chile, Botswana, Singapore, and Estonia).

Let’s start by looking at Malta’s score from the latest edition of Economic Freedom of the World.

Malta is ranked #18, putting it easily in the top quartile.

It gets very good scores in every category other than fiscal policy (somewhat similar to Nordic nations).

What are some of the best features of Maltese economic policy? Let’s look at some excerpts from a column in The Business Picture by Nima Sanandaji.

"…while the big economies of Europe are stagnating, several of the smaller ones are outpacing the US. Malta is the best example, since it led the European growth league in 2024 with a five per cent growth. …Malta is succeeding thanks to competitive taxes and regulations, combined with talent supply, which make it a growing brain business jobs hub. …The share of adults employed in these jobs has increased substantially over time. Currently 9.5 per cent of adults in Malta are employed in highly knowledge intensive jobs. After Switzerland, Ireland and the Netherlands, this is the highest rate in Europe. …Large economies like Greece, Spain, Italy and France have due to regulatory and tax burdens stagnation in share of adults in knowledge intensive jobs. The same countries also struggle with economic growth and job creation."

I’m not surprised that Malta is growing faster than the United States. It’s a classic example of convergence.

What’s more interesting is to look at examples of divergence.

Here’s a chart, based on the Maddison database, showing Malta’s long-run performance (in red) compared to regional competitors, as well as a sampling of other nations.

 

"A few years after World War II ended, Malta was very poor. It ranked lower than Madagascar and its level of per-capita GDP was less than half of Greece.

Now it is has shot way past those two nations, as well as other countries that used to be richer.

Amazingly, Malta has almost caught up with Italy, which had nearly four times as much per-capita GDP back in 1950.

Does this mean Malta has perfect economic policy? Of course not. But it does have better economic policy than most other nations, especially its Mediterranean neighbors.

The moral of the story is that there’s a recipe for growth and Malta is doing a decent job of following the recipe. Assuming they want prosperity, other nations should do the same thing."

AI and the future of labor demand

See You are not a horse by Brian Albrecht. Excerpts:

"For simplicity, suppose human labor demand goes to zero. Not low. Zero. What does that require? It means no dollar you spend, anywhere in the economy, passes through a human hand at any point in its supply chain. Not the person who made the thing. Not the person who shipped it. Not the person who designed it, sold it, maintained it, or cleaned the building where it was assembled. Zero human labor embodied in final expenditure. That’s the target. That’s what I’m going to take “humans become horses” to mean, stated precisely. [tractors replaced horses and horses did not find employment elsewhere-the number of horses in the USA fell greatly after tractors came in]

This is the input-output idea Leontief built his career on. You can trace any final purchase back through its supply chain and add up all the human labor that went into it, direct and indirect. A cup of coffee has the barista, but also the roaster, the trucker, the farmer, the person who made the truck. “Embodied labor” means all of it. For labor demand to collapse, every one of those links has to go to zero, in every product anyone buys

The economy is not one production function. It is many activities. When AI makes some of them cheaper, people don’t just buy more of the same thing. They buy something else.

Every dollar you spend lands somewhere. Some dollars land in activities with lots of human labor inside them: a restaurant, a therapist, a roofer. Some land in activities with almost none: a streaming subscription, an automated checkout, cloud storage. So when we are tracing out what happens when AI gets cheaper, it’s not just “Can AI do my job?” It is “When everyone saves money because AI did my job cheaper, what do they buy next?”

Aggregate labor demand depends on three things: how much people spend in total, how much of that spending lands on activities with human labor inside them, and how much labor is embodied in each of those activities. For human labor demand to collapse, it’s not enough for AI to displace workers inside some activities. Every dollar of spending, wherever it lands, must lose all its embodied human labor. That’s three channels, and the horse argument needs all three to go wrong simultaneously.

The important starting point for thinking about labor is te idea that nobody wants labor. A restaurant doesn’t want waiters; it wants orders taken, customers reassured, mistakes corrected. So labor demand is derived demand. How does AI change how much firms demand?

When AI can do the things firms are actually buying, cheaper AI does two things at once. Firms substitute AI for workers, which reduces labor demand per unit of output. But cheaper AI also lowers output prices, output expands, and the expansion pulls labor demand back up. Whether labor demand rises or falls depends on which effect is larger. This is the Hicks-Marshall decomposition of derived demand into substitution and scale effects.

This is going to be the organizing principle for everything. When a dollar is saved, where is it redirected? To new tasks? To new jobs? To new sectors? It must go somewhere

This is obviously true for many things. Early models had this even. For example, the early GPT exposure paper by Eloundou, Manning, Mishkin, and Rock estimated that roughly 80% of the U.S. workforce could have at least 10% of tasks affected by LLMs. With complementary software, 86% of occupations cross the 10% exposure threshold

And lots has been done on this. The task-level evidence backs this up. In a large customer-support setting, access to generative AI raised issues resolved per hour by about 15%. In a professional writing experiment, ChatGPT reduced average task time by 40% and raised measured output quality by 18%. In a controlled GitHub Copilot experiment, developers completed a coding task 55.8% faster. These aren’t tiny effects.

But they’re effects on tasks. The saved dollar doesn’t vanish when a task gets automated. It creates new tasks within the same job, such as more review, more client management, more judgment calls. Just as there’s not some fixed amount of demand so the scale effects matter, there is not some fixed job.

There’s a ritual in AI discourse where someone posts a demo, the demo does a task associated with a job, and people conclude the job is doomed. Sometimes they’re right. But the inference skips about fifteen steps. What does it actually cost to deploy, errors included? Do customers trust it? Does management know how to reorganize around it? A chatbot demo can appear overnight. A hospital reorganizing clinical liability around AI cannot.

We need to think not just about jobs but organizations. Often the result is a team, not a replacement. A human-AI pair produces output. But complementarity is not free. A pair that produces only slightly more than the AI alone doesn’t justify the human wage. The human has to add something the AI can’t replicate cheaply.

Surgery, aviation, structural engineering, fiduciary advice, for legal reasons alone are areas where we can expect the damage from an error dwarfs the savings from cheaper production. Again, that can always change one day but not soon. When failure on one component destroys the value of all others, you don’t care about the sticker price. That’s the O-Ring logic. You care about cost per unit that actually works. When damage stakes are high enough, human-supervised production wins regardless of how cheap AI becomes.

Suppose substitution wins inside most jobs. The saved dollar escapes the workplace entirely. Where does it go? Most standard models aggregate into a single final good, so this question plays no role. The real economy has many sectors, and the dollar has to land somewhere.

Start with software as a microcosm. This is a sector that has already been heavily automated by digital inputs for decades. If substitution were going to drive labor out of a sector, this is where you’d see it first."

"The most software-intensive industries don’t just retain human labor;they have a higher labor share (67%) than the least software-intensive ones (55%). Heavy digital inputs didn’t drive out human labor. If anything, the industries that automated the most are the ones that spend the most on workers. BLS projects U.S. employment to increase by 5.2 million from 2024 to 2034. Software-developer employment? Up 17.9%, despite direct AI exposure.

The scale effect won within the sector most exposed to digital automation. The BLS could be completely off but the evidence so far points strongly toward the scale effect dominating in software-intense industries.

Software is one extreme but we basically have the same pattern holds across the whole economy, over a much longer period.

For another angle on the problem, let’s go bigger and look across the biggest sectors in the economy: services vs. goods. In 1929, most consumer spending went to physical goods. Today, roughly two-thirds goes to services. As manufacturing got cheaper, people didn’t just buy more stuff. They shifted spending toward healthcare, education, restaurants, personal services. That’s the saved dollar in action at a more not-quite macro but close level — the savings from cheaper goods flowed toward services.

In terms of our guiding decomposition, coods got cheaper."

"Demand for physical stuff didn’t explode. Instead those freed-up dollars migrated to services, and the scale effect showed up there. The substitution effect won inside goods-producing industries. The scale effect won across sectors. Output overall expanded. So if you’re thinking as a macroeconomist, the scale effect dominated."

"But migration alone doesn’t help workers unless the destination still has human labor inside it."

"Services consistently pay a higher share to labor than goods-producing industries. Spending didn’t just migrate. It migrated toward sectors where more of each dollar ends up in someone’s paycheck."

"there is a margin of adjustment, there is an escape hatch when you are looking at an economy as diverse as the modern U.S. economy."

"comparative advantage always pops up fighting against this. When automation makes some things cheap, the things that remain expensive tend to be the things that are hard to automate. And the things that are hard to automate are, almost by definition, the things where humans still have comparative advantage. The saved dollar drifts toward where humans are still worth paying. That’s not optimism. That’s what comparative advantage means.

"In early textiles, power looms cut labor per yard of cloth. But cloth got so cheap that demand exploded, and total employment in textiles rose for decades. Same in early steel, early autos. Eventually demand saturated, prices stopped falling fast enough, and automation reduced employment in each sector. The question for AI isn’t “does automation destroy jobs?” It’s “which phase are we in, for which sectors?”

Where might the AI-saved dollar land today? Healthcare is already 18% of GDP and rising. Elder care will grow as populations age."

"this time isn’t different: new tasks appeared, comparative advantage held, products we couldn't imagine created new work." 

"If AI keeps inventing new varieties of goods that compete with human-produced ones, even a strong initial preference for human labor gets diluted by expanding choice.

I take this seriously. It’s a possible scenario.But notice what it requires. Not just that AI-produced variety expands (which it will) but that it expands fast enough and broadly enough to pull spending away from every human-intensive category at once. The question isn’t whether AI competes with some human goods. It’s whether any human-intensive island survives. Does anyone still spend money on something with a person inside it?

The numbers still have to be extreme. Suppose AI eats 85% of the economy. Software, accounting, law, medicine, logistics, most management, most media. All gone or nearly gone as human labor categories. Suppose the remaining 15% of spending goes to things with at least 30% human labor inside them. Elder care, in-person education, surgery, live performance, skilled trades, therapy, status goods. Then the aggregate human labor share is at least

S ≥ 0.15 × 0.30 = 0.045

That may not sound great but I’m literatlly just putting a bound. Knowing nothing else, we can sustain this. Not large. Not utopia. But not zero, and that’s the absolute lowest possible bound. And remember, labor share declining is not the same thing if the pie is growing much larger."

"As AI makes commodities cheap, real incomes rise, and richer people systematically shift spending toward what he [Alex Imas] calls “relational” goods 

 There's a huge literature in economics on structural change, the long-run pattern where spending shifts from agriculture to manufacturing to services as countries get richer. The big question is why. Is it because prices change and people buy more of whatever got cheaper? Or is it because incomes rise and people just want different stuff? Comin, Lashkari, and Mestieri, for example, decompose the two and find that income effects account for over 75% of the shift. That matters here. If spending migration were mostly about chasing cheap goods, AI making things cheaper would pull dollars toward AI-produced stuff. But it's mostly about what richer people want. And richer people have consistently wanted more services with humans in them."

"Human-created artwork gains 44% in value from exclusivity, versus 21% for AI-generated artwork. AI-made goods feel copyable. Human-made goods feel scarce even when they aren’t. People want what other people can’t have. That wanting doesn’t run out, and it sticks to things a person made."

"income effects dominate price effects by three to one. When basic needs get cheaper, humans don’t say “good, I’m done wanting.” They invent new ways to compare themselves with neighbors. Whether the new wants land on human-made goods or AI-made goods is the open question, and the experimental evidence so far favors humans.

A falling labor share is not falling labor demand. There is a range where labor’s share of income is declining but total labor demand is still rising, because the pie is growing faster than labor’s slice is shrinking. That range may be where we are right now. It would look like “AI is taking over” in share terms while employment keeps growing. The popular argument runs these together and they are not the same claim.

We already see that. Higher income people consume more services. Services tend to be high labor share. Again, that can always flip in the future but this is the evidence we have."

Affordable manufactured housing versus unaffordable climate regulations

By Ben Lieberman of CEI.

"The Biden administration had a field day piling on one costly climate-related regulation after another, not knowing – or caring – that affordability would emerge as a much more pressing concern for Americans than climate change ever was. But now, the Trump administration and Congress have the opportunity to undo these ill-advised rules that are driving up costs for everything from utility bills to cars and light bulbs. We have already seen some progress, but there is much more to do. Next on the list should be Department of Energy (DOE) regulations targeting manufactured housing.

The housing affordability challenges are real, and government is a big part of the problem.  According to the National Association of Home Builders, regulations at all levels of government account for almost 25 percent of the cost of a new single-family home. This includes a growing contribution from federal climate measures, such as those raising the price of major home appliances like air conditioners and furnaces. Worst of all are rules that make the most affordable homes less affordable, thus threatening the dream of homeownership for low-income and younger households. That is why the 2022 DOE energy efficiency rule for manufactured housing warrants a second look. 

The DOE sets energy efficiency standards for manufactured housing. And, as with appliance standards, the agency has a knack for rules that raise up-front costs beyond what is likely to be recouped through energy savings. In this case, the agency admitted that the 2022 rule raised home prices up to $4,500, though manufacturers fear higher costs will outweigh any energy savings.

For perspective, estimates suggest that every $1,000 increase in a median-priced home disqualifies about 156,000 prospective homebuyers. And the effects may be more severe at the lower end of the home spectrum, including manufactured homes, which are the choice of the most price-sensitive buyers. Indeed, it is quite possible that the DOE rule alone is enough to place the dream of homeownership out of reach for hundreds of thousands of lower-income Americans.

As was often the case for the Biden DOE, climate change was a finger on the scale favoring its draconian energy limits on manufactured housing. In fact, the final rule mentions the social cost of carbon dioxide and other greenhouse gases a whopping 50 times. By the agency’s own estimates, the rule’s climate benefits fell short of the claimed consumer savings. Even so, they undoubtedly played a role in the agency’s decision to adopt such stringent standards, despite their effect on prices.   

Fortunately, the president and Congress have not ignored the regulatory plight facing manufactured homes and their prospective purchasers. President Trump’s March executive order titled Removing Regulatory Barriers to Affordable Home Construction, specifically mentions manufactured housing in its section urging regulatory reforms.

Both the House and Senate have passed bills addressing housing affordability, and both contain provisions specific to manufactured housing. Importantly, both bills eliminate the costly and unnecessary requirement that manufactured homes have a steel chassis, however they also fell short of repealing the DOE rule.

A separate House-passed bill, H.R. 5184, the Affordable HOMES Act, would have completely repealed the DOE rule, but it has not been taken up by the Senate. Total repeal deserves consideration if Congress is serious about addressing housing affordability."