End of The Off-Shored Empire (Part 1)
American’s most successful project was one it never admitted running. Now it’s being dismantled by people who didn’t know it existed.
Grüezi!
In April 1949, Joseph Dodge fixed the yen at 360 to the dollar. That rate held for more than twenty years.
Dodge, a Detroit banker on temporary assignment to occupied Tokyo, had embedded what became a durable export subsidy into the currency of the country that had been America’s principal Pacific enemy only four years earlier.
He’d also helped create a yen-denominated capital-allocation mechanism that no agency in Washington possessed, because none could have got such a mechanism through Congress.
Japan’s trade ministry, MITI, created a month later, would spend the next three decades putting it to use.
Between 1950 and 1978, Japanese firms signed 32,000 technology-licensing contracts, mostly with US companies, at a cumulative cost of about $9 billion. Developing the same technology in-house would have cost an order of magnitude more.
Every contract required prior government approval. That allowed MITI to force royalty reductions, demand cross-licensing, refuse foreign-equity participation, and steer technology to its preferred Japanese firms.
Washington watched this happen for decades and did nothing. It did nothing because that was the policy.
The standard story of the Asian miracle is grit, ingenuity and the Cold War. It is not wrong. It is incomplete.
The United States bought strategic alignment by tolerating, financing and protecting economic systems abroad that it could not have built at home.
That bargain was the price of undeclared empire in the face of the Soviet Union.
The empire granted market access, tariff tolerance and technology transfer. It got capital recycling. The bargain held because nobody in Washington had to admit it existed.
Now it is being taken apart by people with little interest in, or understanding of, what they are dismantling.
1 The Asian Miracle Maker
East Asian governments were unusually capable, unusually meritocratic, and unusually willing to override price signals in the service of long-term industrial planning.
That is the story of the Asian miracle.
And it is right about what these states did.
MITI ran a foreign-exchange budget that pushed capital to strategic industries. Korea’s Economic Planning Board went even further. It targeted sectors, set export quotas for each chaebol, and stopped lending to firms that missed them. Daewoo had loans pulled twice in the 1970s for falling short on shipbuilding orders.
Taiwan’s Hsinchu Park, opened in December 1980, combining five-year tax holidays and duty-free imports with very deliberate geographical placement: it sat next to National Tsing Hua University and National Chiao Tung University. Both of those had been rebuilt with US aid in the 1950s. Both supplied ready-made engineering talent.
None of this is up for argument.
In the conventional telling, the Cold War is the reason the United States was willing to tolerate Asian protectionism. It was distracted and indulgent. The Soviet threat made this indulgence look like grand strategy.
But the Cold War was not an excuse for indulgence.
Indulgence was the policy.
The Cold War made the Asian model possible. The Korean War built the centralised states and bankrolled Japan. Vietnam did the same for Korea. An open US market gave them buyers. America’s security guarantees meant they didn’t have to defend themselves.
Push it further and Washington was tolerating, financing and protecting economic systems abroad that it could not have built at home.
In the United States, MITI’s directed credit would have meant tearing down the wall between commercial and investment banking.
The Economic Planning Board’s sectoral targeting would have required a national planning office Truman’s Congress had refused to create.
Keiretsu cartels would have triggered the Justice Department’s antitrust enforcers.
Currency management like the yen peg was not what the Fed was for.
None of these tools were available to American policymakers at home. All of them were being used in Tokyo, Seoul and Taipei.
This isn’t to say East Asian leaders were Washington’s puppets. They negotiated hard. Park Chung-hee was nobody’s fool. Lee Kuan Yew kept his own counsel. And Deng Xiaoping was operating at a strategic level Washington repeatedly underestimated.
Iran, Pakistan, the Philippines and several Latin American states had the same Cold War sponsorship and did not industrialise. But none of them had the Korean War or the Vietnam War next door — neither the military buying boom that funded Japanese and Korean industry in their formative decades, nor the strategic urgency that made Washington tolerate hard bargaining by the client state.
Cold War sponsorship without a hot war on your doorstep produced very different results.
American sponsorship was necessary but not sufficient. The East Asian states still had to build the bureaucracies, discipline the firms, and suppress the consumption that made the model run.
That part could not be outsourced from Washington.
2 The Bargain
The bargain began with a National Security Council memo in October 1948. To contain the USSR, Harry Truman ended the punitive phase of Japan’s occupation and committed to Japanese industrial recovery.
The 1947 Deconcentration Law was meant to break up the wartime zaibatsu — the family conglomerates that had powered Japanese militarism. By 1949, only 18 of the 325 firms designated for break-up had been dissolved.
The keiretsu of the 1950s — Mitsui, Mitsubishi, Sumitomo — were the surviving cores of the same zaibatsu, reorganised around main banks rather than family holdings. Truman shelved the deconcentration plan. Reparations were put on hold.
Joseph Dodge arrived in Tokyo in February 1949 with the rank of minister. The plan he announced a few weeks later balanced the budget, suspended Reconstruction Finance Bank lending and ended producer subsidies.
But the fixed yen rate was what mattered. It was intended to track purchasing power — what 360 yen would buy in Japan against what a dollar would buy in America. Except the Japanese price index that was used to set it didn’t take account of the black market, where much of the effective economy was being priced. The rate ended up roughly a third too cheap.
And that gap became a long-lived export subsidy.
America gave Japan aid in the form of food, fuel and raw materials. The Japanese government sold those goods at home for yen. Normally, a recipient country would have kept that yen and spent it as it pleased.
But the occupation set up a special account, jointly controlled, that held onto the money. The resulting counterpart funds became part of the machinery through which the Japan financed its industrial reconstruction — steel mills, shipyards and the capital base of post-war industry, paid for out of the yen the government had earned selling American wheat and oil to its own people.
Two National Security Council papers in December 1949 made Japanese industrial recovery part of containment in Asia. A few months later, NSC-68 named Japan, along with Germany, as one of two industrial bases the Soviet Union could not be allowed to acquire.
MITI was created on 25 May 1949. Two laws — one in December 1949, one in May 1950 — gave it the powers that shaped the next thirty years. Every Japanese company that wanted dollars to buy a foreign machine or pay a foreign royalty had to ask MITI. Every contract with a foreign company that ran longer than a year had to be approved by MITI as “desirable” for the Japanese economy.
Then the Korean War turned Japanese post-war austerity into a Pentagon-financed procurement boom.
North Korea invaded the South on 25 June 1950. Toyota was producing 300 trucks a month for the domestic market and was close to bankruptcy under the Dodge squeeze. Within weeks, the US Eighth Army had placed an order for 1,000.
Between 1950 and 1953, America’s military bought $1.3 billion of supplies from Japanese factories — and another $1 billion or so on top for base construction, repair contracts and in pay spent locally by soldiers. Through the 1950s, it added up to about $6 billion.
Japan’s Prime Minister Shigeru Yoshida called the war tenyū, a gift from the gods.
Meanwhile, Japan exported goods to America and re-imported them as Treasury bonds. From 1965 onwards, Japan ran a surplus with America that grew every year, peaking at $59 billion in 1987 and reaching roughly half a trillion dollars cumulatively by the time the Cold War ended. By 1989, Japan was the largest foreign creditor of the United States.
The model — undervalued currency, low American long rates, suppressed domestic consumption — didn’t repeat precisely elsewhere.
But the core bargain did.
3 Dictators With Deliverables
South Korea was a democracy in May 1961. It had been one for nine months. The April Revolution of 1960 had overthrown an autocrat, and a new constitution adopted that June established a parliamentary system. Prime Minister Chang Myon’s cabinet was running the country.
In the early hours of 16 May, Major General Park Chung-hee led 3,500 troops across the Han River into Seoul. They seized the broadcasting station and army headquarters. At 3 a.m., the ROK Army Chief of Staff cabled US General Carter Magruder, the UN Commander, to say a coup was in progress.
Magruder had operational control over most of the Korean army. He could have ordered loyalist units to put the coup down.
But he did not.
By morning, Park had taken the cabinet. By evening, Chang Myon’s government had collapsed. Korea’s Second Republic was over.
Ten weeks later, Washington had accepted the junta as the government it would deal with. Five months after that, the US ambassador was cabling Washington that Park was “a forceful, fair and intelligent leader who can be trusted with power.” The democracy America had welcomed in April 1960 was smoothly traded for the Park dictatorship within a few months.
Through much of the previous decade, Korea had been the world’s largest per-capita aid recipient. America sent about $13 billion in assistance from 1946 to 1976 — more than half in military aid.
Through the second half of the 1950s, America covered over a third of the Korean government’s budget, nearly 85% of its imports, and almost 75% of its fixed capital formation. America was not merely subsidising Korea. It was underwriting the fiscal and import base on which the Korean state ran.
A related mechanism operated in Seoul. The local currency proceeds from selling American aid goods on the Korean market sat in a special account that became the government’s investment budget. About 70% of all public investment from 1954 to 1960 went through it. A third went to defence. Korean military spending peaked at 7.2% of GDP in 1960.
Park’s First Five-Year Economic Development Plan, from 1962 to 1966, aimed for more than 7% annual growth. The economy averaged more than 9% for the next decade. Manufacturing rose from 14% of GNP in 1962 to 30% by 1987.
In January 1973, Park announced that Korea was going to build a heavy industrial base: steel, non-ferrous metals, shipbuilding, machinery, electronics and petrochemicals. A planning council was set up to make it happen.
For the next six years, an extraordinary share of bank credit was channelled into manufacturing, with heavy and chemical industries getting the lion’s share.
Most international economists said the strategy would fail. They were wrong.
In early 1969, the World Bank had refused to finance a Korean steel mill on the grounds that Korea was too poor to run one. Korea redirected a quarter of the reparations it received from Japan — $119 million — towards building one. The first blast furnace fired in June 1973, one month ahead of schedule. Today, Korea’s POSCO is one of the world’s largest steel producers.
The Vietnam war became the making of Korea’s economy. In March 1966, the Brown Memorandum set the terms for South Korea’s participation in another American war.
America would pay for equipment, give Korean firms first refusal on military contracts, and add $10 million to military aid that year. In exchange, Korea would send troops.
About 313,000 ROK soldiers served in Vietnam between 1964 and 1973, mostly in combat units. Korea earned some $1 billion in direct hard currency, with broader contracts totalling $5 billion. After the first two years, about 40% of Korea’s foreign-exchange earnings came from the war.
Korea wanted from Vietnam what Japan had got from the Korean War, and Korea got it. Hyundai Engineering & Construction’s earliest contracts had been US Army base construction in Korea in the 1950s. Its Vietnam contracts from 1966 onwards funded the move into civil engineering and shipbuilding that produced the chaebol of the 1980s.
Taiwan ran almost the same template through different instruments. American economic aid from 1951 to 1965 reached roughly $1.4 billion to $1.5 billion. Through the 1950s, aid covered about 40% of Taiwanese imports, 38% of gross domestic investment, and 6% of GNP. That stopped in 1965. The official reason was that Taiwan no longer needed it.
The deeper reason was Vietnam and the balance-of-payments pressure it intensified. The war was costing $25 billion a year and the United States was running into the pressure that would force Nixon to close the gold window in 1971. Aid budgets were being cut everywhere. Taiwan was a “success story” that made cutting easy.
But Taipei now needed a new growth model. Where Korea was building heavy industry on redirected credit, Taiwan took a different path. Premier Chiang Ching-kuo’s economic ministers — K. T. Li chief among them — decided the country’s future was in technology, and specifically semiconductors.
The Industrial Technology Research Institute, a state research body, was set up in 1973 to acquire the foreign know-how to make this work. Three years later, it struck the deal that brought it.
In 1976, ITRI signed a $3.5 million contract with RCA — then one of America’s leading semiconductor firms — for a licence to make integrated circuits. RCA agreed to train Taiwanese engineers in the process.
Nineteen ITRI engineers went to the United States for instruction in chip design, fabrication, testing and equipment maintenance. More followed. RCA’s American factory produced chips at a 50% yield rate. By October 1977, eighteen months after the deal had been signed, ITRI’s pilot factory in Hsinchu was producing chips at a 70% yield rate.
ITRI spun out United Microelectronics Corporation in 1980 to commercialise the process. In 1987, Taiwan Semiconductor Manufacturing Company was founded with Morris Chang — the Taiwanese-American engineer who K. T. Li had recruited from Texas Instruments — as chairman.
The Taiwanese state put up nearly half of the founding capital. Philips of the Netherlands put up a quarter, bringing the production technology and patent licences. Taiwanese industrial families, pressured by the government to participate, put in the rest.
The most strategically important company in the world today began as a state-led joint venture financed in part by a Washington-approved American technology transfer.
4 Ports, Spies and Balance Sheets
The bargain worked differently for Singapore and Hong Kong. The former British dependencies were too small for national-champion strategies. They were given different jobs.
Singapore at independence in 1965 was a poor port city with no natural resources, mass unemployment, and a British military presence that was about to disappear. The conventional wisdom for newly independent countries in the 1960s was to build national industries behind tariff walls. Lee Kuan Yew did the opposite. Singapore’s Economic Development Board, set up in 1961, four years before independence, was given one job: bring in foreign multinationals.
That strategy was Albert Winsemius’s. The Dutch economist had been Lee’s adviser since 1961 and would stay with him for the next twenty-three years. Forget national champions, Winsemius told him; let the Americans come in instead.
They did.
Texas Instruments opened a chip factory in Singapore in 1969 — going from contract signature to first production in seven weeks. National Semiconductor had arrived the previous year. Hewlett-Packard came in 1970. Fairchild, GE, Seiko and Siemens followed.
By 1997, Singapore hosted nearly 200 American manufacturing firms. It had no national champion and no significant local industry. Its GDP per capita had also risen from about $500 to $26,000, while unemployment had fallen from 14% to just under 2%.
The security side of the bargain was less visible. In January 1968, Britain announced it would withdraw all military forces “east of Suez” by 1971. For Singapore this meant the loss of perhaps 20% of GDP — British bases were the country’s biggest employer — and, more importantly, the loss of a security guarantor.
The Five Power Defence Arrangements signed in 1971, with Britain, Australia, New Zealand and Malaysia, replaced the formal British presence. In full recognition of British post-imperial pusillanimity, it obliged the parties simply to “consult” if Singapore came under attack.
Singapore’s actual guarantor was US Pacific Command. The 1990 Memorandum of Understanding and the 1998 Strategic Framework formalised US Navy access to Singaporean naval bases at Sembawang, Paya Lebar and Changi. Singapore was not formally an American protectorate, but its strategic position rested increasingly on the US Seventh Fleet.
Hong Kong did two jobs. The first was intelligence. The US Consulate-General became one of Washington’s largest overseas diplomatic and intelligence platforms, and the CIA’s principal listening post on China.
The second job was commercial. Throughout the years that America embargoed direct trade with the People’s Republic, Hong Kong was the back channel.
Citibank had been there since 1902. Bank of America arrived in 1947. Chase Manhattan and HSBC were part of the financial plumbing that handled legitimate trade, remittances and grey-zone flows.
The textile-quota regime of the 1970s, designed in Washington to protect American garment workers from Asian imports, became from Hong Kong’s side a system of tradeable export rights into the US market. Hong Kong textile and garment exports rose from less than $1 billion in 1965 to more than $25 billion at handover in 1997.
The 1992 United States-Hong Kong Policy Act gave the colony separate trade, export-control and visa treatment from the rest of China — a regime that survived the handover in 1997 and lasted until Trump suspended it in 2020.
Bargain Busters
In September 1985, America turned the screw on Japan for the first time. Treasury Secretary James Baker convened the finance ministers of Japan, West Germany, France and Britain at the Plaza Hotel in New York and made clear that the dollar was going to come down.
Japan, running a $46 billion trade surplus with the United States and absorbing the political fallout from American manufacturing job losses, was the real target.
The yen rose from about ¥250 to the dollar at the start of 1985 to about ¥120 by 1988. In dollar terms, the yen had almost doubled. Japanese exports were suddenly no longer competitive. Japanese manufacturers responded by moving production to lower-cost countries in Southeast Asia, the same countries where Japanese government aid had spent twenty years building ports, power stations and industrial estates.
Japanese foreign investment in East Asia had run at around $1 billion a year in the first half of the 1980s. By 1989, it had reached $8 billion. By 1991, investment flows from OECD economies to the ASEAN-4 — Thailand, Malaysia, Indonesia and the Philippines — had overtaken flows to the original Tigers for the first time.
Thailand’s Eastern Seaboard programme — Laem Chabang container port, Map Ta Phut petrochemicals — was financed by Japanese government loans through JBIC. Mitsubishi Thailand exported its first Thai-built vehicles in 1987.
Malaysia’s Penang cluster had begun fifteen years earlier with Intel’s first overseas factory: 5 acres on a former paddy field, $1.6 million, 100 employees. By the 2020s, Intel’s cumulative investment in Penang was more than $5 billion. Mahathir’s Look East Policy of 1982 sent more than 15,000 Malaysian students and trainees to Japan and South Korea to learn how the miracle worked.
Indonesia under Suharto combined oil-funded protectionism in the 1970s with foreign-investment liberalisation after 1986.
The bargain held together until 1997. Then it broke.
Through the early 1990s, Western and Japanese banks had been lending heavily and short-term to Thai, Indonesian and Korean banks, which had been lending the money on to local firms at longer maturities.
When things went sour in mid-1997, Western lenders refused to roll over their loans. The Bank of Thailand abandoned its peg to the dollar on 2 July 1997. The baht collapsed. The same wave hit Indonesia in October and Korea in December.
Then the IMF arrived, bringing rescue packages and something else: reforms.
Many of the reforms had little operational connection to the short-term debt-rollover problem that the rescue was officially designed to address. Treasury and the Fund used the stabilisation crisis to swap one model of capitalism for another, on terms that paid off for the United States in three ways.
First, American capital came in. Within five years of the crisis, Goldman Sachs had bought Kookmin Bank, Newbridge Capital had bought Korea First Bank for a fraction of book value, and Carlyle had bought KorAm. American investors ended up owning meaningful stakes in firms the Korean state had spent thirty years building up.
Second, Wall Street advisory firms got a new market. Asian companies had to start behaving the American way: quarterly reports, independent boards, share-price discipline, IFRS accounts. Korean chaebol that had been run as long-term industrial dynasties were now required to generate shareholder returns.
Third, Asia was wired directly into American capital markets. Asian savings — household, corporate, and eventually sovereign — flowed into US Treasuries, US corporate bonds and US equities at scale.
The “Asian savings glut” that Bernanke would later credit for keeping American long rates low throughout the 2000s was the same money.
From 1997 onwards, the Tigers operated inside a financial regime less protected and more open to American discipline than the one on which their rise had been built.
Now they were paying America rent for the privilege.
5 The Tiger Who Came To Tea
In March 1969, Soviet and Chinese troops fought a series of pitched battles on the Ussuri River. The Sino-Soviet split, building since the late 1950s, was erupting, and Nixon saw an opening. If America could bring China onto its side of the Cold War, the Soviet Union would be isolated.
Kissinger flew secretly to Beijing in July 1971 and came back with an agreement that Nixon would visit within seven months. The result was the Shanghai Communiqué, signed in February 1972.
It had two clauses that mattered. An anti-Soviet one committed both countries to opposing “hegemony” in Asia — code for countering Moscow. Then there was a Taiwan clause that was deliberately ambiguous.
America acknowledged that “all Chinese on either side of the Taiwan Strait maintain there is but one China and that Taiwan is a part of China”. Acknowledged, but did not endorse. This let America keep its security relationship with Taipei while opening its commercial relationship with Beijing.
Carter normalised diplomatic relations on 1 January 1979. Deng Xiaoping came to America the following month and got full White House military honours and a Stetson at the Simonton rodeo in Texas. The wedge had worked. China was now publicly aligned with America against the Soviet Union.
After 1979, two things brought China further in. China got most-favoured-nation tariff treatment — the same low rates America extended to its allies — though subject to annual review by Congress. And the 1979 US-China Science and Technology Agreement opened American laboratories to Chinese researchers and allowed for joint government and industrial cooperation.
For more than four decades, the agreement was renewed with little public debate. It ran continuously below the level of US political argument. It was arguably the most important technology-transfer instrument America ever signed.
China’s reform process unfolded under this umbrella. Foreign firms entering China were required to take a Chinese partner and accept government-imposed equity caps — 50% in car manufacturing, telecoms and financial services, lower in some sectors.
The four original Special Economic Zones, authorised in August 1980, gave foreign investors tax breaks and infrastructure in exchange for export production.
Shenzhen, a fishing town across the border from Hong Kong, had GDP of RMB 270 million in 1980. By 2018, it had RMB 2.42 trillion. It overtook Hong Kong along the way.
In June 1989, the People’s Liberation Army cleared Tiananmen Square. The death toll was estimated at somewhere between a few hundred people and several thousand, but it unfolded on US network news and American public opinion swung overwhelmingly against the Chinese government.
The Bush administration imposed limited sanctions but declined to touch most-favoured-nation status. Within four weeks, Bush sent a secret mission to Beijing to reassure Deng that the relationship would continue. Declassified records show the US envoy telling Deng in the Great Hall of the People: “We have benefited — both sides — strategically with respect to the Soviet Union.”
Bush kept his word. He vetoed bills in 1991 and 1992 that would have linked most-favoured-nation status with human-rights compliance.
Clinton came in promising to do just that, then reversed his position by 1994.
The campaign to keep most-favoured-nation treatment was run by the Business Coalition for US-China Trade: Boeing, GM, Motorola, AT&T and GE. Its political manager was Robert Rubin, who became Treasury Secretary the following year.
American multinationals had built their China strategies on the assumption that the bargain would not break. Washington confirmed that assumption.
Ten years after Tiananmen, in November 1999, the US-China WTO agreement was signed in Beijing. China’s most-favoured-nation status became permanent rather than annually reviewable.
Charlene Barshefsky, the US Trade Representative who had negotiated the deal, told an audience in April 2000 what it actually meant: “broad-ranging, comprehensive, one-way trade concessions on China’s part. The United States doesn’t lower any tariffs. We don’t change any trade laws. We do nothing.”
Clinton, speaking a month earlier, had articulated the democratic utopianism of the time:
“By joining the WTO, China is not simply agreeing to import more of our products. It is agreeing to import one of democracy’s most cherished values, economic freedom.”
China cut industrial tariffs from roughly 25% in 1997 to 9% by 2005. It opened up nine of twelve services sectors to foreign firms. It agreed to give foreign companies full trading and distribution rights by 2004 and accepted fifteen years of non-market-economy treatment in anti-dumping cases.
That non-market-economy clause expired in December 2016. Donald Trump had been elected the previous month.
The model Japan had run with America from 1965 onwards now ran with China at vastly larger scale. China sold America vastly more goods than it bought. The bilateral US deficit widened from $10 billion in 1990 to a peak of $419 billion in 2018. The cumulative US merchandise deficit with China between 1985 and 2019 was some $5.5 trillion.
And the dollars did come back. The People’s Bank of China’s reserves rose from $11 billion in 1990 to $1 trillion in 2006, when they overtook Japan’s, to nearly $4 trillion in June 2014. Chinese holdings of US Treasuries rose from $60 billion in 2000 to a peak of $1.3 trillion in November 2013.
China was now both the world’s largest manufacturer — overtaking America in 2010–11 — and the largest foreign creditor of the American government.
The concentrated cost fell on American manufacturing workers and the regions most exposed to import competition. Three US economists working with Census Bureau data published “The China Syndrome” in October 2013. They argued that Chinese import competition explained about a quarter of the decline in American manufacturing jobs between 1990 and 2007.
Their follow-up estimated that 2 million American jobs had been lost between 1999 and 2011, of which nearly half had been in manufacturing. American manufacturing employment had been 17 million in 2000. By 2010, it was 11.5 million.
What did America want? In September 2005, Robert Zoellick, Deputy Secretary of State and the man who as US Trade Representative had finalised China’s WTO accession four years earlier, gave a speech listing what Washington expected from the new arrangement.
China would protect intellectual property. China would liberalise its market. China would be transparent about its military build-up. China would cooperate on energy. China would commit to no use of force against Taiwan. China would improve relations with Japan.
For once, the United States would not get what it wanted.
6 The Unnamed Empire
The bargain had four operating components and one political one.
Market access. America took the goods. Asian exports flowed into America’s consumer market with low tariffs, preferential treatment under the Generalised System of Preferences, and most-favoured-nation status renewed annually for communist countries, permanent for everyone else. The US government tolerated trade deficits that would have been politically unacceptable in any other context.
Tariff tolerance. America accepted asymmetric protection. Asian governments were allowed to keep their tariffs and quotas, joint-venture requirements, foreign-equity caps, local-content rules, and distribution arrangements like the Japanese keiretsu that effectively blocked American firms from selling into Asian markets even when formal tariffs were low.
Technology transfer. America let its technology go. Licensing contracts, joint ventures with mandatory IP sharing, the 1979 Science and Technology Agreement with China, military co-development like the F-16 partnership with Korea — every channel that could move American industrial knowledge into Asian hands was kept open.
Capital recycling. America borrowed back what it had spent. Asian central banks took the dollars their economies earned from selling to America and bought US Treasury bonds. The result worked very nicely. America imported Asian goods, ran fiscal deficits to absorb them, and was lent the money back at low interest rates by the Asian governments doing the exporting.
Deniability. None of this was called an industrial policy. None of it was called an empire. None of it was called economic statecraft. And none of it was admitted to be a single coherent arrangement. This is what kept the first four hidden.
But the problem with empires, as the British had found out, is that not everyone benefits back home. And the problem with democracies is that every few years those people get to vote.
The Japanese miracle, the Korean catch-up, Taiwan’s semiconductor sector, Singapore’s electronics cluster, the post-Plaza redistribution to Thailand, Malaysia and Indonesia, and finally China’s accession to the world trading system — these were not parallel national achievements that happened to take place inside an American security order.
They were working parts of an American order that allowed client bureaucracies to do what Washington could not do openly at home.
The bargain held up for seven decades because nobody had to admit to it.
The Americans who lost manufacturing jobs to it through the 1990s and 2000s did not know they had lost them to a deliberate policy. They understood themselves to have lost them to globalisation.
Because globalisation was what imperial policy was called when it had to be given a name.
7 The Bargain Gets Hunted
In May 2015, the Chinese State Council published Made in China 2025. MIC2025 was the first time Beijing had put in writing what its industrial strategy actually was.
The plan was a three-stage vault up the technological ladder: manufacturing power by 2025, mid-rank by 2035, global leader by 2049 — the centenary of the People’s Republic. Ten sectors, from semiconductors and aerospace to biotech and electric vehicles. By 2025, the plan said, 70% of core components and key materials in these sectors would be Chinese-made.
China had been climbing the value chain for twenty years. But MIC2025 added a timetable.
The bargain could tolerate upgrading by allies. What it could not tolerate was a rival power declaring that it intended to breach the frontier. Beijing was publicly refusing the old division of labour. Washington noticed.
Washington’s response came in three stages.
First was a trade war. In March 2018, the US Trade Representative published a 215-page investigation finding that China had used joint-venture requirements, licensing restrictions, state-directed acquisitions and cyber theft to extract American technology, at an estimated cost of at least $50 billion a year.
It was the clearest official admission that the technology-transfer part of the bargain — tolerated, managed, or failed to stop by every administration from Carter to Obama — was a one-way intellectual credit line Washington could no longer afford to keep open.
Tariffs followed. By February 2020, the average American tariff on Chinese goods had risen from 3% to 19%, covering two-thirds of imports.
The second stage was investment screening. An August 2018 act expanded the review of foreign acquisitions of American firms — the Committee on Foreign Investment in the United States — to cover minority stakes in tech companies, critical infrastructure and firms holding sensitive personal data. CFIUS had been around since 1975 as a narrow national-security review. After 2018, it became an industrial-policy enforcer.
Broadcom had been the turning point. In March 2018, Trump blocked Singapore-based Broadcom from acquiring the American chipmaker Qualcomm — a $117 billion deal — on the grounds that a Singaporean acquirer would weaken Qualcomm’s position against Huawei in 5G. The decision had nothing to do with Chinese ownership. It was about preventing the consolidation of America’s semiconductor capability under any foreign owner, even an allied one.
Then came the Entity List. Huawei was added in May 2019. American firms could no longer sell it components or software without a licence, and the presumption moved sharply against approval. A year later, the Foreign Direct Product Rule was extended: any chip made anywhere in the world using American semiconductor equipment now required an American licence before it could be sold to Huawei.
That forced TSMC to stop fabricating chips for Huawei’s design subsidiary HiSilicon. SMIC, China’s largest domestic foundry, was added to the list in December 2020 with a footnote that meant any application for sub-10-nanometre technology would be presumed denied.
China’s path to cutting-edge chip production was now blocked at the key external supply points.
The squeeze tightened in October 2022. The Bureau of Industry and Security at the Commerce Department published a rule restricting American export licences for chips at the leading edge in every major category — logic, memory, manufacturing equipment. Anything Chinese factories were trying to build at the frontier was now under American export control.
Under Section 744.6, American citizens, permanent residents and US-incorporated entities now needed a licence from the Commerce Department to support the development or production of advanced chips at fabrication facilities in China.
Export control had gone from policing the movement of America’s goods to restricting the movement of its people.
Even workarounds were shut down. An October 2023 rule closed a loophole Nvidia had engineered by stripping its top chips’ bandwidth to keep them under the original thresholds: the A800 and H800 had been sold into China before being locked out. A December 2024 package added 140 more Chinese companies to the Entity List, imposed the first countrywide controls on high-bandwidth memory, and added more than two dozen types of semiconductor manufacturing equipment.
By the end of 2024, the United States had built the most extensive technology-export-control regime since the Cold War, all aimed at a single country, with the announced aim of preventing China from reaching the leading edge of semiconductor production.
The third response was domestic. The CHIPS and Science Act and the Inflation Reduction Act were signed within a week of each other in August 2022. Together they constituted the most explicit American industrial-policy turn since the early Cold War.
CHIPS put $52 billion into direct subsidies for building chip factories on American soil. By late 2024, the Commerce Department had committed much of it. The largest single beneficiary was Taiwan Semiconductor Manufacturing Company, the firm that already produced most of the world’s leading-edge chips. America was paying TSMC $6.6 billion to build American fabs.
There were strings attached. Recipients of CHIPS money could not engage in any significant expansion of advanced-semiconductor manufacturing capacity in China, Russia, Iran or North Korea for ten years. Take the subsidy and foreign expansion stops.
The IRA did the same thing for electric vehicles and clean energy. Its consumer credit tied eligibility to rising North American and free-trade-partner content requirements for critical minerals and battery components, while excluding vehicles using battery components from Chinese-controlled entities.
Tax credits were reverse-engineering three decades of offshoring.
CHIPS and the IRA admitted what the bargain had been all along. Biden’s National Security Advisor, Jake Sullivan, said so at Brookings in April 2023. The previous American economic strategy had been built on the assumption that markets would allocate capital efficiently and that integration with authoritarian competitors would produce convergence.
Neither had happened.
Trade had hollowed out the American industrial base. That hollowed-out base helped produce the political conditions in which American democracy itself began to deteriorate.
The new approach that Sullivan advocated was “a modern American industrial strategy.” Direct government investment in strategic sectors. Supply-chain resilience. Technology denial to adversaries: “small yard, high fence.”
America would draw a narrow perimeter around a small number of frontier technologies — semiconductors, artificial intelligence, biotech, quantum — and defend it at any cost, while letting trade in everything else continue.
What is happening now is the collapse of the bargain’s deniability. Industrial policy that created an East Asian offshore empire, that ran in the shadows for seven decades because it could not be rolled out at home, is now being repatriated and rehabilitated.
The political coalition that built the offshore empire — American multinationals with Asian production bases, financial-services firms with Asian asset positions, the security establishment that traded market access for strategic alignment — never had to defend its bargain publicly because the bargain was not publicly visible.
But the coalition that has now won the public argument is dismantling it without a clear account of what held it together, or what the consequences of dismantling it will be.
What gets built in its place? What comes after the offshore empire?
That is the subject of Part 2!
Thanks for reading.
Bis bald,
Adrian