Comparative Advantage Still Moves World Trade in the Age of Containers and Tariffs

Two glasses of port and a bolt of English wool, arranged in a small table of labor hours, once did more to organize the planet than any armada. David Ricardo’s 1817 example still runs the world’s supply chains, though it now travels in a forty-foot steel box and clears customs under rules no classical economist imagined. Trade economics is the story of comparative advantage colliding with distance, dockworkers, tariffs, and politics, and the collision is far from over.

Portugal’s Wine, England’s Cloth: One Table That Rearranged the World

David Ricardo came to economics from the trading floor, not the lecture hall. He started as a clerk on the London Stock Exchange at fourteen and retired in his early forties with a fortune reported at the time to run into the hundreds of thousands of pounds, an enormous sum in Regency England. Legend attaches part of it to government bonds bought shortly before news of Waterloo reached London, a story retold with equal parts admiration and envy. Restless in retirement, he purchased a country estate, then a seat in Parliament through an Irish borough, and settled into a second career writing dense treatises between debates. In 1817 he published On the Principles of Political Economy and Taxation, and in its seventh chapter, on foreign trade, sits a small table of four numbers that economists have been arguing about ever since.

Ricardo built the example to provoke. He imagined Portugal producing both wine and cloth with less labor than England requires for either good: fewer worker-years per barrel, fewer worker-years per bolt. Portugal, in his telling, is flatly better at everything. Common sense, and most trade policy before and since, concludes that Portugal should make both goods at home and sell its surplus, while England should shelter whatever industry it can keep breathing. Ricardo’s table concludes the opposite. Let Portugal concentrate on wine, England on cloth, and let the two exchange, and both countries end up consuming more wine and more cloth than either could manage alone. Not a winner and a loser. More of both goods, in both places.

Gold on the Scoreboard

To feel the shock of that claim, recall the orthodoxy it detonated. Mercantilism had governed European statecraft for roughly two centuries: exports counted as victories, imports as defeats, and the score was kept in gold. A nation grew rich, on this view, by selling much, buying little, and hoarding the difference in its treasury. Trade was war by quieter means, and one country’s gain was necessarily another’s loss. Adam Smith opened the first crack in 1776 with a homely observation: Scotland could raise grapes in hothouses and make its own wine at roughly thirty times the cost, he noted, but only a fool would call that prosperity. Buy wine where wine is cheap, and sell what you yourself make cheaply. That is absolute advantage, and it persuades easily. It also leaves an uncomfortable question hanging. Suppose a country is the cheapest producer of nothing at all. Smith’s logic seems to leave it with nothing to sell and no reason to be traded with.

Opportunity Cost, the Only Honest Price

Ricardo’s answer is the part that still bends minds. Being worse at everything turns out not to matter, because goods are never paid for with superiority. They are paid for with forgone alternatives. The real cost of a bolt of English cloth is not the labor poured into it but the wine England could have made with the same hands and hours. Measured that way, the rankings flip. England, though clumsier at both trades, sacrifices less wine for each bolt of cloth than Portugal does, precisely because Portugal’s vineyards are so productive that every worker diverted to a loom costs the country dearly in barrels. Portuguese excellence at wine is exactly what makes Portuguese cloth expensive. England therefore holds a comparative advantage in cloth: not because it is good at weaving, but because weaving is the thing it is least bad at. Each country specializes where its opportunity cost is lowest, the two exchange at some ratio between their internal tradeoffs, and the arithmetic quietly delivers the gains. No ideology is involved; the result follows from subtraction.

Domestic life runs on the same theorem. Picture a senior lawyer who types faster than her assistant. She holds the absolute advantage in typing, and she should still never do it, because an hour spent typing is an hour not spent on work billed at many multiples of any typist’s wage. Her assistant types the brief not despite being slower but because his forgone alternatives are cheaper. Surgeons do not sterilize their own instruments; star chefs rarely wash dishes. Everyone specializes where the next-best use of their time is least valuable, and everyone is richer for the swap. Countries, Ricardo saw, are no different, only larger and angrier about it.

Samuelson’s Wager on Comparative Advantage

Durability made the proposition a running joke in the profession. Paul Samuelson liked to recount a challenge put to him by the mathematician Stanislaw Ulam: name one proposition in all of social science that is both true and non-trivial. Years passed before the right answer occurred to him, and it was comparative advantage. True, Samuelson argued, because it follows from arithmetic that two centuries of critics have failed to refute. Non-trivial, because generations of intelligent people had proven unable to grasp it unaided or to believe it after patient explanation. The doctrine stands as a permanent rebuke to the scoreboard instinct: trade balances are not match results, imports are not defeats, and a country can no more be priced out of world commerce entirely than the slow-typing assistant can be priced out of the office.

Honesty requires reading the theorem no more broadly than it is written. It speaks of nations and aggregates: specialization plus exchange enlarges what a country as a whole can consume. It says nothing about how the enlargement is divided, and it never promised that every worker inside a winning nation wins. English weavers prosper in Ricardo’s world; Portuguese weavers must find something else to do, and finding something else is slow, painful, sometimes generational. The gains are large enough in principle for winners to compensate losers, but the theorem does not perform the transfer; politics does, or fails to. Ricardo, dry even by the standards of his subject, spent little ink on the point. Two centuries later the unpaid remainder of that bill would come due in shuttered mill towns, and this story will have to return to collect it.

Distance Refuses to Die: Gravity in the Trade Data

Endowments on Paper, Anomalies in the Tables

Two Swedish economists, Eli Heckscher and Bertil Ohlin, gave Ricardo’s insight a source. Comparative advantage, they argued in the early twentieth century, springs from what a country holds in abundance. Capital-rich economies should export machinery and chemicals, goods that soak up capital. Labor-abundant economies should export apparel and footwear, goods stitched together by many hands. Land-rich countries should ship grain and beef. The logic felt almost geological: endowments lie beneath the trade statistics the way bedrock lies beneath a landscape, and the pattern of exports simply reveals what was always there.

Then Wassily Leontief checked. In 1953 he ran American trade through the input-output tables he had painstakingly built, expecting to find capital pouring out of the most capital-rich economy on earth. Instead the arithmetic showed American exports embodying more labor per dollar than American import-competing production did. The result, quickly christened the Leontief paradox, embarrassed the cleanest theory in the field. Partial rescues arrived over the following decades: American labor was not ordinary labor but skilled labor, a form of human capital the tables did not measure; American technology differed from everyone else’s, so the theory’s assumption of shared techniques failed; natural resources and trade barriers muddied the sample. Each repair helped, yet the episode left a lasting bruise. Elegant theory had met the data and blinked first.

Rich Countries Selling Each Other the Same Things

A second embarrassment sat in plain sight. Endowment theory predicts that trade should flow between countries that differ: cloth for wine, machines for shirts. By the postwar decades, though, the largest trade flows on the planet ran between countries that resembled each other almost perfectly. Germany and France, similar in income, capital stock, and skills, shipped each other enormous volumes of cars, and not obviously different cars. Economists built intra-industry trade indexes to measure the overlap, and for neighboring rich economies those indexes showed that a large share of commerce, in many industries reported at well over half, consisted of two-way exchange within the same product categories. Nothing in the factor-proportions playbook explained why that should happen at all.

Paul Krugman, then in his twenties, supplied the explanation in a few short papers around 1979 and 1980. Drop the assumption of constant returns. Let factories get cheaper per unit as they scale, and let consumers prize variety for its own sake. Under those conditions no country can profitably produce every variant of every good; each specializes in a slice of the product space and trades for the rest. Germans buy Peugeots and the French buy Volkswagens not because either country has the wrong endowments but because scale economies force specialization while shoppers insist on choices. Trade between similar economies stops being a puzzle and becomes a prediction. Nearly three decades later, in 2008, the Nobel committee cited precisely this body of work.

Newton’s Formula, Borrowed Without Apology

While theorists argued, empiricists stumbled onto something close to a law of nature. Jan Tinbergen noticed in the early 1960s that bilateral trade behaves like gravitational attraction: the flow between two countries rises with the product of their economic masses, their GDPs, and falls with the distance separating them. The borrowing from Newton was cheeky, and it worked absurdly well. Estimated on almost any dataset, from any decade, for any sample of countries, the gravity equation accounts for a large majority of the variation in bilateral trade, a reliability that almost nothing else in economics can match. Typical estimates put the distance elasticity near one, meaning that doubling the distance between two trading partners roughly halves their trade. That relationship has held from the age of steamships through the age of container megaships and fiber optic cable, which is exactly the period when distance was supposed to stop mattering.

Borders turn out to matter even more than miles. John McCallum reported in the mid-1990s that Canadian provinces traded with one another on the order of twenty times more than with American states of comparable size and distance, despite a boundary so friendly it barely registers in daily life. Later work, notably by James Anderson and Eric van Wincoop, showed that the original estimate overstated the effect by ignoring how each region’s trade depends on its trading costs with every partner at once; their correction pulled the number down substantially, yet the surviving border effect stayed large, a multiple rather than a rounding error. Other frictions show up just as reliably. Sharing a language raises bilateral trade by a measurable margin. Colonial ties linger in the flows generations after the empires dissolved. Currency unions appear to lift trade too, and although early estimates of that boost were later trimmed, the direction of the effect survived scrutiny.

All of this lands on a conclusion that flatters no ideology. Container ships, cargo jets, and undersea cables collapsed the cost of moving goods and information to a fraction of what merchants once paid, and pundits declared the world flat more than once on the strength of it. Trade data keep issuing corrections. Distance still taxes commerce, borders still tax it more heavily, and shared history still subsidizes it. Technology shrank the price of distance without abolishing its pull. Anyone who wants to predict how much two countries trade should skip the manifestos and the ministerial communiques, take out a map, note the size of each economy, and measure the space between them. That humble procedure, closer in spirit to Newton than to any political program, still explains the world’s commerce better than the theories written to redesign it.

Steel Boxes and Six-Dollar Tons: Freight Costs Fall Off a Cliff

Malcom McLean did not love ships. He loved trucks, or more precisely the money trucks earned when they were moving, and he hated watching that money die on a pier. His drivers idled for days at East Coast docks while longshore gangs wrestled cargo piece by piece, crate by barrel by bale, into the holds of aging freighters. McLean’s answer was to stop handling the cargo at all. In April 1956 a converted wartime tanker called the Ideal X left Newark for Houston carrying fifty-eight aluminum boxes on a reinforced deck. No one broke a single crate open along the way. The boxes came off in Texas, settled onto truck chassis, and rolled inland with their contents untouched by human hands since New Jersey.

Marc Levinson, whose history of the container remains the standard account, reported the before-and-after arithmetic. Loading loose cargo by hand in the mid-1950s cost on the order of 5.86 dollars a ton. Loading the Ideal X cost roughly 16 cents a ton. Reductions of that magnitude do not improve an industry; they abolish it and put something else in its place. Ocean freight had long been a serious line item in the price of anything traded across water, often a tenth or more of a good’s landed cost. Within a generation it shrank toward a rounding error.

Piers Torn Down, Twistlocks Locked In

Dockworkers grasped immediately what the box meant, which is why they fought it for a decade. Longshore work had been brutal but plentiful; a container crane and a handful of operators could replace a sweating gang of twenty. Peace was eventually purchased rather than won. On the American West Coast the union traded its old work rules for guaranteed income and pensions, in settlements reported at the time as among the costliest in the industry’s history, and similar bargains were struck from New York to Rotterdam. Port geography changed just as sharply. Cranes needed deep water and acres of flat land for stacking, which the finger piers of Manhattan and the crowded docks of London could not supply. Cargo migrated to purpose-built terminals: Oakland across the bay from a fading San Francisco waterfront, Rotterdam at the mouth of the Rhine, Singapore at the hinge of Asia’s sea lanes.

Standardization did the quiet work. Through the late 1960s, committees hammered out ISO specifications for box dimensions, corner castings, and locking fittings, so that a container lifted off a ship in Rotterdam could ride a German train and then a Dutch truck without anyone touching what was inside. War forced the pace. American military logistics had collapsed into chaos at Vietnamese ports, and McLean’s Sea-Land won contracts to containerize the supply line; his ships, which would otherwise have steamed home empty, began calling in Japan and carried early waves of Japanese exports to America at close to marginal cost. Vessels grew to match the ambition. Where the Ideal X carried fifty-eight boxes, the largest carriers afloat today are reported to load on the order of 24,000.

Goods Made Everywhere, Assembled Somewhere

Once moving a ton of freight across an ocean cost less than trucking it across a county, the logic of where to make things came apart. Factories no longer had to sit near their suppliers or their customers. Production fragmented into global value chains: design in one country, components from half a dozen others, assembly wherever labor and logistics aligned, with intermediate parts sometimes crossing borders several times before any finished good crossed one at retail.

Teardown studies of the iPhone became the canonical ledger of this system. Assembled in China, as the fine print concedes, but with chips from Taiwan, displays and memory from South Korea, and the design, the software, and the fattest margins residing in California. Researchers who traced the value chain reported that China’s captured value per unit was a modest slice of the retail price, a few percent in the early studies, even though customs data booked the entire factory-gate value as a Chinese export. Statistics built for a world of finished goods were now counting the same components again and again. Countries had begun trading tasks rather than products: discrete stages of production, each priced and sourced on its own.

Aggregates followed the boxes. Trade’s share of world output, stuck for decades at modest levels, climbed steadily from the 1970s onward; by the 2008 peak, total trade in goods and services was reported at roughly half of world GDP. No earlier era of globalization, not even the gold-standard decades before 1914, had knitted production itself so tightly across borders. Earlier globalizations moved wine and wheat between economies that remained self-contained. This one distributed the assembly line itself across a dozen jurisdictions and bet that the ships would always sail on schedule.

Six Days in the Suez, Ten Times the Rate

Tightly optimized systems fail expressively. In March 2021 the Ever Given, among the largest container ships afloat, wedged itself diagonally across the Suez Canal and stayed there for six days while hundreds of vessels, carrying cargo valued in the billions of dollars, queued at either end of the waterway. Salvage crews dug at her bow; the internet made her a folk hero; economists tallied the daily cost of a single blocked ditch in Egypt. That same year, pandemic-scrambled demand produced a container crunch in which spot rates on major routes were reported at around ten times their normal level, while empty boxes piled up in the wrong ports on the wrong continents. Shortages of everything from bicycles to garage doors taught households a phrase that had lived for fifty years in trade journals and logistics conferences. Supply chain entered dinner conversation, and it carried a question the six-dollar ton had long deferred: whether a trading system built to make distance free had, in the same stroke, made itself fragile.

Triangles at the Border: Anatomy of a Tariff

Strip away the politics and a tariff is a sales tax collected at the dock. Impose a duty of 25 percent on imported steel and the domestic price, for foreign and homegrown metal alike, floats up toward the world price plus the tariff. Everything else follows from that one price change. Buyers of steel, which in practice means carmakers, appliance plants, and construction firms, pay more for every ton. Domestic mills, sheltered behind the wall, sell at the higher price and expand output that could not survive open competition. The treasury, meanwhile, collects the duty on whatever imports still arrive.

Tally the ledger and it refuses to balance. Producers gain, the government gains, yet consumers lose more than both together, and the shortfall takes a precise geometric form on the economist’s diagram: two triangles of pure loss. One measures wasted production, resources pulled into making at home, expensively, what the world would supply for less. The other measures destroyed consumption, purchases that would have happened at the world price and now never occur. Nobody collects either triangle. That value simply evaporates at the border, and a standard result holds that the evaporation grows roughly with the square of the tariff rate, so doubling a duty quadruples the waste.

Theory does leave one loophole. A country large enough to move world markets can, in principle, force foreign sellers to absorb part of the duty by depressing the world price itself, an argument the textbooks call the optimal tariff. Whether that escape hatch opens in practice is an empirical question rather than a logical one, and the freshest evidence, taken up below, suggests that for the United States in 2018 it mostly stayed shut.

Rents That Sailed Back to Tokyo

Quotas inflict the same damage with one twist: the revenue disappears from the treasury’s books and surfaces somewhere less convenient. When Washington pressed Japan into a “voluntary” export restraint on automobiles in 1981, capping shipments at a reported 1.68 million cars a year, scarcity lifted prices across the showroom floor, Japanese and American models alike. Had that price gap been a tariff, the US government would have collected it. Under the restraint, Japanese firms held the scarce right to export, so the premium landed in their pockets; estimates reported at the time put the annual transfer in the billions of dollars, a subsidy to the competition delivered, in effect, by the country doing the protecting.

Constraint bred strategy. Any firm allowed to ship only a fixed number of cars will ship expensive ones, and the restraint duly pushed Toyota, Honda, and Nissan upmarket. Lexus, Acura, and Infiniti exist as brands in large part because a quota made cheap subcompacts a wasted allocation. The same firms then jumped the wall altogether, opening transplant factories in Ohio, Tennessee, and Kentucky. Within two decades those plants were exporting cars from America, an outcome nobody drafting the restraint had sketched. Protection meant to buy Detroit time ended up financing its rivals’ move upscale and onshore.

Ghosts of 1930, Grudges over Chickens

History keeps two warnings on file. The Smoot-Hawley Act of June 1930 raised American duties on thousands of goods just as the Depression gathered force. More than a thousand economists reportedly signed a petition urging President Hoover to veto the bill; he signed it anyway. Dozens of trading partners retaliated or followed suit, and by 1933 the value of world trade had collapsed by a reported two thirds. Economists still debate attribution, since deflation and imploding incomes did much of the damage on their own, but the sequence of tariff, retaliation, and spiral became the founding trauma of the postwar economic order. The GATT, signed in 1947, is best read as an institutional vow never to run that experiment again.

Retaliation can also outlive its cause by generations. In the early 1960s the European Economic Community raised duties on American poultry, which was then flooding European markets. Washington answered in 1964 with a 25 percent tariff on light trucks, among other items. The poultry quarrel is long forgotten; the truck tariff, known ever since as the chicken tax, still stands six decades later. It helps explain why American roads belong to domestically built pickups, why foreign automakers assemble their trucks in North America, and why at least one manufacturer reportedly shipped vans with rear seats installed purely to clear customs as passenger vehicles, then stripped the seats out at the port.

Field Notes from the 2018 Experiment

Tariff waves in 2018 and 2019 handed economists something rare: a large protectionist experiment observed almost in real time with modern price data. Research teams led by Amiti, Redding, and Weinstein, and separately by Fajgelbaum and coauthors, converged on the same headline: pass-through into US prices was close to complete. Foreign exporters, on the whole, did not cut their prices to absorb the duties. American firms and households paid them, in roughly the proportion the border tax implied.

Washing machines supplied the cleanest case. After safeguard tariffs took effect in early 2018, retail prices rose by roughly 12 percent, and dryer prices, though never tariffed, rose in near lockstep because the two sell as pairs. Setting the total consumer cost against the jobs created at new domestic plants, researchers reported a bill on the order of 800 thousand dollars per protected job per year, several times what any of those jobs actually paid.

Steel told the arithmetic story. Industries that use steel employ many times more workers than industries that make it, by common estimates dozens of times more, so raising the input’s price endangered more jobs downstream than it sheltered upstream. Retaliation arrived on schedule as well: China targeted soybeans, American exports to their largest customer fell steeply, and Washington answered with farm support programs reported at roughly 28 billion dollars over two years, returning a large share of the tariff revenue to the countryside as compensation. Tax the consumer, pay the farmer, and the two triangles of pure loss sit exactly where the theory drew them.

Gains on Average, Grief in Particular: The China Shock Ledger

Every gains-from-trade theorem ships with fine print, and the fine print spent two centuries in small type. Ricardo’s arithmetic, and everything built on top of it since, establishes that trade enlarges the pie for the country as a whole. It says nothing about who eats. The formal result is a compensation principle: winners gain enough that they could reimburse the losers and still come out ahead. Could carries a great deal of weight in that sentence. Nothing in the theorem writes the check, and for most of the era of liberalization, nobody else wrote it either.

Invisible Dividends at the Register

Aggregate gains are real, and their invisibility is built into how they arrive. They come as a shirt that costs a few dollars less, a toaster cheaper than lunch, shelves stocked with varieties that simply did not exist in a closed economy. Studies of the import surge from China reported meaningful declines in the prices of traded goods, with the largest proportional benefit flowing to lower-income households, which spend more of their budgets on the tradable stuff of daily life. Import competition also pushed domestic producers toward higher productivity, and cheap imported inputs made downstream industries more competitive abroad. All of it is genuine wealth. None of it is legible. No shopper thanks trade policy at the checkout, because the dividend dissolves into millions of transactions, each too small to notice and impossible to attribute.

Furniture Towns and the Arithmetic of Concentration

Losses obey the opposite arithmetic. When David Autor, David Dorn, and Gordon Hanson began publishing what became known as the China shock research, the numbers landed hard because they were tied to places. Their estimates, widely reported, attributed on the order of one million lost American manufacturing jobs directly to Chinese import competition between 1999 and 2011, and roughly 2.4 million once supplier networks and local spending spillovers were counted. That damage did not spread thinly across the country the way the gains did. It clustered. Furniture making collapsed in the North Carolina Piedmont, in towns like Hickory and High Point that had built a century of identity around it. Textiles and apparel buckled across the Carolinas and Georgia. Toys, luggage, footwear, and consumer electronics assembly hollowed out their own particular counties elsewhere. A commuting zone specialized in exactly what China had begun to export absorbed a blow that national statistics barely registered.

Then the textbook adjustment failed to happen. Displaced workers were supposed to shift into expanding sectors or move toward stronger labor markets. Mostly they did neither. A decade and more after the initial hit, exposed counties still reported depressed wages, lower employment rates, and elevated enrollment in disability programs, which came to function as an unofficial early retirement scheme for men in their fifties. Follow-on research found the shadow in social statistics too: reported declines in marriage rates, rises in single parenthood, and elevated mortality from drugs and alcohol in the hardest-hit places. Labor proved far less mobile than the models assumed. People stayed where their houses, families, and churches were, and the loss stayed with them.

Stolper and Samuelson, Translated into Paychecks

None of this should have surprised the theorists. In 1941, Wolfgang Stolper and Paul Samuelson worked out who wins and who loses when a country opens to trade, and their logic fits in a sentence: trade raises the return to the factor a country holds in abundance and squeezes the factor it holds in scarcity. The United States is abundant in capital and educated labor and was scarce, relative to China in the 2000s, in less-educated manufacturing labor. So trade with China lifted returns to shareholders, college graduates, and firms plugged into global value chains, while it bid down the market value of exactly the skills concentrated in furniture towns and mill towns. In a poor country the same theorem runs in reverse and favors abundant low-wage labor. The theory never promised that everyone wins. It promised the opposite, with precision.

Compensation That Stayed Theoretical

The standard answer held that losers would be compensated out of the winners’ surplus. The actual instrument, Trade Adjustment Assistance, dates to 1962 and remained chronically miniature next to the shock it was meant to cushion. Annual spending reported in the hundreds of millions to low single-digit billions stood against labor-market losses plausibly measured in the tens of billions each year. Eligibility rules were baroque, take-up was low, and evaluations of the retraining on offer found modest results at best. European countries with thicker general safety nets buffered comparable shocks somewhat better, which suggests the failure was a policy choice rather than a law of nature. The compensation principle stayed a principle.

Backlash Runs on Concentration Too

Political economy had long explained protectionism with a simple asymmetry: a tariff’s benefits concentrate on a compact, organized industry, while its costs diffuse across inattentive consumers, so lobbies beat shoppers. Liberalization built the mirror image of that machine. Its gains diffuse across everyone and mobilize no one; its losses concentrate in communities that do not forget. Later work by Autor, Dorn, Hanson, and Kaveh Majlesi reported that congressional districts more exposed to import competition drifted toward ideological extremes, and parallel findings emerged across European electorates. Concentrated grief proved as politically potent as concentrated privilege ever was, and the tariff politics of the late 2010s grew directly out of soil the China shock had prepared.

Honesty requires holding the whole ledger open at once. The profession’s consensus on aggregate gains was right and remains right; its confidence that adjustment would be quick and diffuse was wrong, and the wrongness became measurable in disability rolls. The same ledger has a Chinese side: export-led growth coincided with what the World Bank reported as roughly 800 million people climbing out of extreme poverty, the largest such movement in recorded history. Cheaper goods, richer economies, hollowed counties, and the greatest escape from destitution ever measured are not competing claims. They are entries in a single account, and the era’s central failure was pretending the painful lines would reconcile themselves.

Rounds, Rules, and Rewired Maps: Trade Order Under Renovation

Twenty-three countries signed the General Agreement on Tariffs and Trade in Geneva in October 1947, a provisional pact drafted in the shadow of the interwar tariff wars it was designed to bury. Provisional, in this case, lasted nearly half a century. Round after negotiating round, from Annecy and Torquay through the Kennedy, Tokyo, and Uruguay Rounds, the contracting parties swapped concessions until average industrial tariffs among rich countries fell from figures commonly cited at roughly 40 percent after the war toward low single digits by the 1990s. Each round ran longer and drew more members than the last, as the easy cuts were exhausted and the agenda crept outward from tariffs toward subsidies, standards, and procurement. Few treaties have moved so much money with so little drama.

Two clauses did most of the quiet work. Most favored nation treatment meant that any concession granted to one member extended automatically to all, so a bargain struck between Washington and Bonn lowered barriers for Uruguay and Japan as well. Binding meant that a tariff, once cut, was capped: a government could not quietly raise it again without compensating its trading partners. Together the two rules converted thousands of bilateral haggles into a single ratchet, multilateralizing generosity while constraining backsliding. Lawyers rarely describe treaty language as elegant. These clauses earned the word.

Courtrooms Without Judges

Marrakesh, 1994: the GATT acquired a constitution. The World Trade Organization opened its doors the following year, extending the rulebook to services and intellectual property and, most consequentially, adding a dispute settlement system with a standing Appellate Body, something close to a world trade court. For two decades it functioned remarkably well. Hundreds of disputes were filed, panels convened, and rulings largely obeyed, even by the biggest economies. Bananas, cotton subsidies, aircraft: the docket read like an atlas of commercial grievance.

Then the court lost its judges. Washington blocked new appointments to the Appellate Body for years, citing judicial overreach, and by December 2019 the bench lacked a quorum to hear anything at all. Trade lawyers now speak of appeals filed into the void: a losing party can appeal a panel ruling to a tribunal that no longer exists, suspending the case indefinitely. Rules remain on the books; enforcement has become optional for anyone patient enough to appeal. A coalition of members, the European Union among them, improvised an interim arbitration arrangement to preserve appellate review among themselves, yet the world’s most consequential commercial rivalry sits outside it.

Blocs in the Vacuum

Regionalism accelerated where the multilateral track stalled. Europe’s single market remains the deepest integration ever attempted among sovereign states, harmonizing regulations, opening services, and treating national borders as administrative details for goods, capital, and people alike. NAFTA, in force from 1994, was renegotiated into the USMCA in 2020 with tighter automotive content thresholds and new labor provisions. Across the Pacific, the TPP survived American withdrawal in 2017 by rebranding as the CPTPP, eleven economies concluding that the deal was worth keeping without its largest member; the United Kingdom later joined. RCEP, signed in 2020, stitched fifteen Asia-Pacific economies, including China, Japan, and South Korea, into a bloc reported to be the world’s largest by population.

Every preferential deal carries a hidden tax called rules of origin. Lower tariffs apply only to goods genuinely made inside the bloc, so each agreement specifies, product by product, how much regional content qualifies. Firms selling into several blocs must therefore document compliance with several overlapping rulebooks, the tangle Jagdish Bhagwati memorably named the spaghetti bowl. Textile chapters can demand yarn-forward compliance, tracing a finished shirt back to the spinning of its thread before any preference applies. Some exporters reportedly pay the ordinary tariff rather than shoulder the paperwork of proving they deserve the preferential one.

Subsidies, Sanctions, and Connector Economies

Industrial policy, pronounced dead in polite company for a generation, has returned with a budget. America’s CHIPS and Science Act of 2022 committed on the order of 50 billion dollars to semiconductor manufacturing at home; the Inflation Reduction Act channeled tax credits estimated in the hundreds of billions toward clean energy, with content rules favoring North American production. Brussels answered with green industrial plans of its own and loosened state aid rules. Restrictions arrived alongside the subsidies: sweeping export controls, announced in October 2022 and later coordinated with Japan and the Netherlands, on advanced semiconductors and the machines that make them, aimed squarely at slowing China’s progress at the technological frontier.

New vocabulary followed the money. Friendshoring means sourcing from geopolitical allies; derisking means reducing dependence on a rival without attempting a full divorce. On the ground the pattern looks more like rerouting than retreat. Mexico and Vietnam have emerged as connector economies: China’s share of American imports has fallen since the tariffs of 2018, Mexican and Vietnamese shares have climbed, and Mexico reportedly overtook China as the largest source of American goods imports in 2023. Meanwhile both connectors buy sharply more from China than before. Components increasingly take a detour, collecting a stamp of assembly en route to American shelves. Supply chains, like rivers, answer a dam by cutting another channel.

Grand talk of deglobalization deserves a colder look at the numbers. World trade as a share of global output has been roughly flat since 2008, plateauing near its peak rather than collapsing, a pattern the business press nicknamed slowbalization. Integration stopped deepening; it did not go into reverse. The system is being renovated while occupied, walls moved and wiring redone, tenants complaining about the noise.

Comparative advantage has survived Napoleonic blockades, two world wars, Smoot-Hawley, and a pandemic that idled the factories of entire nations, and the present renovation reroutes it rather than repeals it. Trade’s center of gravity is shifting toward Asia’s internal corridors, toward services sold over fiber optic cable, toward maps drawn as much by security councils as by cost curves. The boxes keep sailing, millions of them on the water at any given moment, watched now by politicians as closely as by economists.