Two ideas explain how the dollar rules the world. Both were inverted in the retelling — and the errors now shape policy from Washington to Beijing.

The dollar’s dominance is usually explained by two propositions: that America enjoys a free lunch, and that it must run trade deficits to feed the world with money. Neither proposition survives contact with the evidence. What survives is stranger, and more fragile.

The dollar was enthroned by accident, in a committee room

The Mount Washington Hotel had been shut since 1942, and the renovations were not finished when the delegates began arriving on the last day of June 1944. Seven hundred and thirty people came to a 234-room resort in the White Mountains of New Hampshire — roughly three times the number expected. The overflow was billeted in hotels up to five miles away. A single bridge crossed the river separating the grounds from the road, which made the proceedings easy to seal off from the world.

But the specific sentence that put the dollar at the centre of the postwar world was not written by an American. White had been careful to seed the draft articles with a deliberately vague formula — payments and subscriptions were to be reckoned in “gold and gold-convertible exchange” — a phrase that named no currency at all. It was A. D. Shroff, of the Indian delegation, worried about how rupees would convert against sterling inside the imperial preference system, who asked the drafters to say plainly what the phrase meant. The answer came from the British side: Dennis Robertson, an economist of considerable standing, proposed that holdings be expressed as gold and United States dollars. He checked the following day that the wording had survived into the text.

It was not the Americans who proposed to put the dollar at the centre of the world economy. It was a British economist, answering an Indian delegate’s question about rupees.

That is the first inversion in this story, and it sets the pattern for the rest. The architecture of dollar dominance is routinely narrated as something Washington designed and imposed. Much of it was improvised, conceded, or built by other people for their own reasons. The two concepts that have come to define the system — the “exorbitant privilege” and the “Triffin dilemma” — are both borrowed phrases whose modern usage has drifted a long way from what their authors meant. The drift is not a pedantic matter. Both misreadings are currently load-bearing in political argument, and both point policy in the wrong direction.

A Belgian economist read the accounts and found a trap

Robert Triffin was born in Flobecq in 1911, took American citizenship, and spent the 1950s as one of the architects of the European Payments Union before settling at Yale. He was, by temperament, an accountant of systems: a man who looked at the postwar monetary order and asked what would happen if you simply extended its arithmetic forward.

The arithmetic was uncomfortable. Under Bretton Woods, other countries held dollars as reserves because dollars were convertible into gold at thirty-five dollars an ounce. World trade was growing. If reserves had to grow with trade, and reserves were dollars, then foreign dollar claims on the United States would compound year after year. The American gold stock would not. At some point the claims would exceed the metal, and everyone holding dollars would know it. Triffin published the argument in two 1959 articles in the Banca Nazionale del Lavoro Quarterly Review, testified before the Joint Economic Committee of Congress in October 1959, and collected the material in Gold and the Dollar Crisis: The Future of Convertibility (Yale University Press, 1960).

What Triffin actually predicted: a confidence crisis in a gold-convertible reserve currency, followed by a run, followed by global deflation as the United States tightened to defend the peg.

The prescription mattered as much as the diagnosis. Triffin did not want America to stop supplying liquidity. He wanted the supply of international reserves taken away from any single national currency and vested in a reformed IMF — an institution that could create reserve assets deliberately rather than as a by-product of one country’s balance of payments. The Special Drawing Right, created in 1969, is the surviving fragment of that idea.

It is worth being precise about what he feared, because the modern usage has replaced it with something else entirely. Triffin’s dilemma was a claim about a gold-convertible system. Its whole force came from the fixed, finite backing. Remove the gold and the specific mechanism he described — claims exceeding metal, therefore a run — has nothing to bite on.

The French complaint was about seigniorage, not trade

On 4 February 1965, Charles de Gaulle told a press conference that the acceptance of dollars as the equivalent of gold allowed the United States to be indebted to foreign countries free of charge, paying its debts in money it could itself issue at will. He called the dollar a means of credit appropriated to one state and dressed up as an impartial international instrument.

The phrase that stuck came from his finance minister. Valéry Giscard d’Estaing described the benefit accruing to the issuer of the world’s principal reserve currency as an exorbitant privilege. The attribution is not perfectly clean — some scholars trace the sentiment to Jacques Rueff, de Gaulle’s monetary adviser and a Banque de France eminence, and the phrase is frequently misattributed to de Gaulle himself — but Giscard is the standard source, and the contemporary record places it around 1964–65. Rueff supplied the mechanism, in a formulation of some elegance: when a key-currency country runs a payments deficit, it pays its creditors in dollars, and those dollars are of no use in Bonn or Tokyo or Paris. The same day, they are lent back into the New York money market. The debtor does not lose what the creditor gains. The key-currency country never feels the deficit at all. Deficits, as Rueff put it, without tears.

The proof arrived. The London Gold Pool, through which the major central banks had defended the thirty-five-dollar price, collapsed in March 1968 and the market split in two. Pressure built for three more years. Over a weekend at Camp David, Richard Nixon and a small circle — Treasury Secretary John Connally as principal architect — settled on a package, and on the evening of Sunday 15 August 1971 Nixon went on television to announce it: convertibility suspended, a ninety-day freeze on wages and prices, and a ten per cent surcharge on dutiable imports, the first general tariff increase since Smoot–Hawley. The Smithsonian Agreement that December devalued the dollar by roughly 8.5 per cent, moving the official gold price to thirty-eight dollars, and widened the fluctuation bands; Nixon called it the most significant monetary agreement in the history of the world. It lasted about fifteen months. By March 1973 the major currencies were floating.

Triffin appeared to have been vindicated on schedule. This is precisely where the misreading begins.

The version of Triffin everyone quotes is not the one he wrote

Open a newspaper, a policy paper or a political speech and you will find the Triffin dilemma stated roughly as follows: because the world needs dollars, the United States is obliged to run current account deficits to supply them; therefore American trade deficits are structurally unavoidable; therefore the deficits are somebody else’s fault, or alternatively the price of empire. This claim now appears in arguments for tariffs, for currency accords, and for reforming the trading system by force.

It has a serious problem. Michael Bordo of Rutgers and Robert McCauley, then of the Bank for International Settlements, examined the proposition directly in a study published as BIS Working Paper 684 (2017), an NBER working paper (2018) and, in final form, in the IMF Economic Review (2019). Their verdict on the trade-deficit version is blunt: it is popular, but anachronistic, and flawed in logic and in fact.

The logical problem is that reserves are not the same thing as a current account deficit. A country can supply the world with its currency by exporting capital — lending, investing, buying foreign assets — while running a current account surplus. Britain did exactly that for decades. Sterling was the world’s reserve currency during a long period in which Britain ran current account surpluses and exported capital on an enormous scale. The historical case that is supposed to prove the dilemma disproves the trade-deficit version of it.

The factual problem is that the sequence does not fit. The United States ran current account surpluses through most of the 1950s and 1960s — the very period in which the world was accumulating dollar reserves and in which Triffin was writing. American deficits on current account became a persistent feature only much later, from the 1980s onwards, long after the gold link had been cut. If foreign dollar reserves were built out of American trade deficits, they were built before those deficits existed.

Bordo and McCauley also found that the American gold position after the Second World War was no worse than the British position in 1900 — and sterling’s gold link survived until the First World War broke it. A weak reserve ratio is not, by itself, destiny.

Their broader argument is that Triffin’s enormous influence rested on reviving an interwar fear — that a shortage of gold would force deflation — and that what actually followed the 1960s was not American prudence and global deflation but American profligacy and global inflation. That is close to the opposite of the predicted failure mode. Better and entirely feasible US policies, they argue, could have kept the system going.

This is a contested reading, and it should be presented as one. Bordo and McCauley are pushing against a large body of work that treats the dilemma as broadly validated, and they concede the central point in its general form: a national currency doing the job of an international public good does generate real conflicts between domestic and global obligations. Charles Kindleberger and Ronald McKinnon made versions of the sceptical case decades earlier. What the sceptics deny is the specific mechanical claim — that reserve status compels trade deficits — which is the version that has escaped into politics.

The world’s dollars are manufactured offshore, by private banks

If American trade deficits do not supply the world’s dollars, what does? The answer is a banking system, and it sits mostly outside the United States.

A dollar deposit is a liability of a bank. It does not have to be an American bank. When a bank in London or Singapore or Paris lends dollars, it creates a dollar deposit in the act of lending — the loan and the deposit appear together, and no dollars need to travel anywhere. This is the eurodollar market, which grew from the late 1950s onwards for a tangle of reasons including American interest rate ceilings under Regulation Q, and which long outgrew the loopholes that started it. It is a dollar money supply created beyond the reach of the Federal Reserve, by institutions the Fed does not charter and cannot directly examine.

The scale is not marginal. The BIS global liquidity indicators track credit to non-bank borrowers outside the currency’s home area. At the end of March 2026, the stock of dollar-denominated foreign currency credit to non-banks outside the United States stood at $14.7 trillion, growing at 7.3 per cent year on year, with roughly thirty per cent owed by borrowers in emerging market and developing economies. Dollar credit expanded 8.5 per cent in 2025, the fastest annual growth since the third quarter of 2014.

And that is only the debt that appears on balance sheets. Claudio Borio, Robert McCauley and Patrick McGuire have argued in a series of BIS studies that foreign exchange swaps, forwards and currency swaps create forward dollar payment obligations that are functionally borrowing but are recorded as derivatives, and therefore vanish from standard debt statistics. Analysing the 2022 Triennial Central Bank Survey, they put these obligations at over $80 trillion worldwide — with non-banks outside the United States owing some $25–26 trillion off balance sheet, roughly double their on-balance-sheet dollar debt, and non-US banks owing upwards of $35 trillion. Much of it is very short-term, which means it must be rolled over constantly.

The world does not receive dollars from America. It manufactures them, in London and Tokyo and Singapore, and then discovers that it cannot manufacture them in a crisis.

This changes the shape of the problem entirely. The vulnerability in the dollar system is not that America might stop running trade deficits. It is that an enormous, largely offshore, largely short-term dollar funding structure exists which no single authority supervises, and whose obligations are invisible in the statistics that policymakers actually consult. When it seizes, the demand is not for American goods. It is for dollars, immediately, from an institution able to create them.

America’s external accounts are a bank’s, not a household’s

The second concept requires the same treatment. What exactly is the exorbitant privilege, measured?

Pierre-Olivier Gourinchas and Hélène Rey gave the term a precise modern meaning: the excess return the United States earns on its external assets over what it pays on its external liabilities. Working with Nicolas Govillot, they assembled a dataset of American external positions at market value running back to 1952. They found a persistent gap. Estimates vary with the period and the vintage of the data — figures around 2.7 percentage points since 1952 appear in their work, with roughly 2.6 points for the Bretton Woods era and about 2.4 points for the years after — but the finding itself is robust and has been replicated widely.

The mechanism is a balance sheet. The United States borrows short, safe and liquid — Treasury bills, bank deposits, the world’s reserve assets — and invests long, risky and illiquid — foreign equity, direct investment in factories and firms. Gourinchas and Rey called this being the world’s banker, and later, as the risk profile of American holdings shifted further out, the world’s venture capitalist. A bank earns the spread between what it pays depositors and what it earns on its loan book. So does America.

The current data make the scale of the balance sheet visible. At the end of the first quarter of 2026, the Bureau of Economic Analysis put American external assets at $43.37 trillion and external liabilities at $64.64 trillion, giving a net international investment position of minus $21.27 trillion. The net figure is the one that gets quoted. The gross figures are the ones that explain the system: over a hundred trillion dollars of cross-border claims running in both directions through a single balance sheet.

The net position moved from −$27.61tn at end-September 2025 to −$21.87tn at end-December and −$21.27tn at end-March 2026 — swings of trillions in single quarters. The current account deficit over the same period ran at roughly $200bn a quarter. Valuation effects, not trade, dominate the arithmetic.

This is why the household analogy — America as a spendthrift borrowing to consume — misleads so badly. Households do not have leveraged trading books whose mark-to-market swings dwarf their income. Banks do. And banks, unlike households, are exposed to the specific risk that everyone wants their money back at once.

The privilege is not a free lunch. It is an insurance premium

Here the argument turns. Gourinchas, Rey and Govillot asked the obvious follow-up question: if the excess return is a payment, what is it a payment for?

The answer they document is insurance. A portfolio that is long global risk and short safe dollar claims makes money in normal times and loses catastrophically in bad ones — and bad times are precisely when the dollar appreciates, inflating the value of what America owes while the value of what it holds collapses. During the 2007–09 crisis, they estimate, wealth transfers from the United States to the rest of the world amounted to around twenty per cent of American GDP, while the dollar appreciated some eight per cent in real terms. Related work by Gourinchas, Rey and Kai Truempler puts the deterioration in the net international investment position between the fourth quarter of 2007 and the first quarter of 2009 at twenty-one per cent of GDP, of which roughly sixteen points were pure valuation loss — on the order of $2.2 trillion.

They called this the exorbitant duty. It is the other half of the ledger, and it is systematically omitted from the political version of the story.

The hegemon earns a spread in calm years and pays it back, at scale, in the year the world catches fire. That is not a subsidy. It is an underwriting business.

The duty is not only financial. It is operational, and it falls on the Federal Reserve. When the offshore dollar system seized in 2008, the Fed extended currency swap lines to foreign central banks so that they could lend dollars to their own banks; outstanding swaps peaked at around $583 billion in December 2008, roughly a quarter of the Fed’s balance sheet at the time, with the European Central Bank and the Bank of Japan the largest counterparties. In March 2020 the Fed did it again, reactivating and extending the lines to nine additional central banks and adding a repo facility for foreign and international monetary authorities against Treasury collateral. Outstanding swaps peaked near $450 billion in late May 2020, with the Bank of Japan holding $223 billion and the ECB around $145 billion.

Consider what those facilities are. An American institution, accountable to the American Congress, creating dollars to rescue banks in other jurisdictions, because the alternative — a fire sale of dollar assets and a global funding collapse — would be worse for everyone including Americans. The BIS authors note the awkwardness bluntly: the swap lines were set in a fog, because nobody knew where the off-balance-sheet obligations actually sat. The hegemon was underwriting risks it could not measure.

The heaviest costs fall on countries that never chose the system

None of this makes the arrangement fair. Reframing the privilege as an insurance premium answers the question of whether America gets something for nothing; it does not answer the question of who bears the system’s costs.

Hélène Rey’s work on the global financial cycle supplies the sharpest version of the complaint. The textbook trilemma says a country may choose two of three: free capital movement, a fixed exchange rate, and independent monetary policy. Float the currency and you buy back monetary autonomy. Rey argued, in a 2013 paper delivered at the Jackson Hole symposium, that this is too generous. Capital flows, leverage and asset prices across the world move together in a cycle driven substantially by conditions at the centre — by the Federal Reserve, and by global risk appetite. For a country with an open capital account, floating does not restore independence. The choice is a dilemma, not a trilemma: you can have monetary autonomy only if you manage the capital account.

The consequences are concrete. A tightening cycle in Washington raises the cost of the $14.7 trillion of offshore dollar credit, of which around thirty per cent is owed by emerging and developing economies, regardless of conditions in those economies. Governments respond by accumulating reserves as self-insurance — an expensive habit that intensified after the Asian crisis of 1997–98, when the terms attached to emergency assistance persuaded a generation of finance ministers never to need it again. Reserve accumulation means buying low-yielding claims on the centre while paying more on your own liabilities: a negative carry that transfers real resources outward.

Estimates of the magnitude exist but should be handled with care. Work published in the Review of World Economics in 2021 estimated that developing economies as a group recorded a return differential of about minus three percentage points relative to developed economies over 2010–19, implying an average annual outward transfer on the order of $800 billion, or 3.3 per cent of their GDP. Figures of this kind depend heavily on valuation methodology and country coverage, and reasonable economists dispute them. The direction, however, is not seriously contested: the insurance premium the hegemon collects is paid disproportionately by countries that had no vote in designing the arrangement.

The petrodollar agreement everyone cites does not exist

A third misunderstanding deserves separate treatment, because it is not a subtle drift of meaning but a straightforward factual error, and it is astonishingly widespread.

The story runs that in 1974 the United States and Saudi Arabia signed a fifty-year agreement under which the kingdom would price oil exclusively in dollars in exchange for American security guarantees; that this agreement created the petrodollar system; and — in the version that went viral in June 2024 — that it expired on 9 June that year, unrenewed, with catastrophic implications for the dollar.

What actually happened in 1974 is documented. The United States–Saudi Arabian Joint Commission on Economic Cooperation was established on 8 June 1974, in the aftermath of the 1973 embargo. A separate and genuinely secret arrangement later that year committed Saudi Arabia to invest heavily in US Treasury securities in return for military support; it was brought to light in 2016 through a Freedom of Information Act request by Bloomberg News. Neither instrument contained a clause requiring oil to be priced exclusively in dollars, and Saudi Arabia continued to accept other currencies, including sterling, during 1974. The claim that a fifty-year exclusivity deal expired in June 2024 has been traced back through a chain of aggregator sites to no primary source at all, and was rebutted at the time by, among others, the chief economist of UBS Global Wealth Management.

The causation runs the other way. Oil is priced in dollars because the dollar is dominant. The dollar is not dominant because oil is priced in it.

The historian David Wight has made this point directly, and the logic is straightforward: a seller wants a currency that is deep, liquid, universally accepted and usable to buy anything else in the world. That was true of the dollar before 1974 and would remain true if every oil contract were redenominated tomorrow. Oil invoicing is a symptom of dominance, not its foundation.

The myth persists because it offers something the real explanation does not: a single, dated, revocable document. If dominance rests on a contract, it can be cancelled. If it rests on network effects across payments, invoicing, funding, collateral and reserves, it cannot be cancelled by anyone — only eroded, slowly, by the accumulated decisions of thousands of institutions. The second story is true and unsatisfying. The first is false and thrilling.

There is a Triffin problem. It is fiscal, not commercial

Having dismantled the popular version, honesty requires stating the version that may well be real.

Strip out the gold and the trade deficits and a structural tension remains. The world’s demand for safe dollar assets grows with the world economy. The capacity of the United States to supply genuinely safe assets grows with the American fiscal base — which is shrinking as a share of the world. Emmanuel Farhi and Matteo Maggiori formalised this in A Model of the International Monetary System, showing that the dilemma can resurface even under floating rates, because a reserve asset carries an implicit promise that it will not be devalued when the world suffers a disaster, and that promise becomes harder to keep as the issuer shrinks relative to the claims on it. Ricardo Caballero, Farhi and Gourinchas developed the parallel argument about a global shortage of safe assets and its macroeconomic consequences. Rey has framed the same tension as a new Triffin dilemma: demand for dollar liquidity rising with the world economy, the relative fiscal capacity backing it falling.

Bordo and McCauley are sceptical of this version too, arguing that it overstates both the demand for safe assets and the inflexibility of their supply. The dispute is live and unresolved, which is the honest thing to say about it.

Foreign holdings of US federal debt reached roughly $9.2 trillion in December 2025 — about 31 per cent of publicly held debt of $30.1 trillion, a share that has fallen in recent years. Of those holdings, 41.9 per cent was official and 58.1 per cent private. Japan held about $1.2tn, the United Kingdom $0.9tn, China $0.7tn.

That last set of numbers deserves attention, because it quietly refutes another common claim — that foreign governments finance America and could therefore destroy it by selling. Foreign official holders own a minority of a minority: roughly two-fifths of the 31 per cent of publicly held federal debt that foreigners hold at all. The marginal buyer of Treasuries is increasingly a private investor, foreign or domestic. A Chinese decision to sell would move prices and would hurt China’s remaining holdings; it would not determine American funding costs.

The real threat is to the asset’s legal quality, not its economics

In February 2022, following the invasion of Ukraine, Western governments immobilised roughly $300 billion of Russian central bank reserves held abroad. Whatever the merits of the decision, it taught every reserve manager on earth a lesson that no amount of economic analysis could have taught: a reserve asset held in another jurisdiction is contingent on a political relationship.

The response is measurable, and it has gone into metal rather than into rival currencies. World Gold Council data show central bank net purchases of 1,082 tonnes in 2022 — the highest on record in the modern series — followed by 1,037 tonnes in 2023 and about 1,045 tonnes in 2024, against an average of roughly 473 tonnes a year over 2010–21. Purchases moderated to about 863 tonnes in 2025, still far above the historical norm. Gold has no issuer, no counterparty and no jurisdiction. That is now the point of it.

Infrastructure has moved too. Project mBridge, a wholesale central bank digital currency platform built with the central banks of China, Hong Kong, Thailand and the United Arab Emirates, reached minimum viable product stage in mid-2024, with Saudi Arabia joining as a full participant. On 31 October 2024, BIS general manager Agustín Carstens announced that the BIS would hand the project to the participating central banks, framing the exit as a function of the project’s maturity rather than of politics, and distancing mBridge explicitly from Russian proposals for a BRICS payments bridge. The BIS has since concentrated its own efforts on Project Agorá, with seven advanced-economy central banks and a large group of private institutions. Whatever the intentions, the effect is two development tracks running along a geopolitical seam.

And yet the currency data stubbornly refuse to show a collapse. The IMF’s COFER survey put the dollar at 57.13 per cent of official foreign exchange reserves in the first quarter of 2026, up from 56.42 per cent the previous quarter, against the euro at 20.03 per cent, the yen at 5.44 per cent and the renminbi at 1.99 per cent. Total reserves are around $13 trillion. The renminbi’s share has gone essentially nowhere for a decade.

The measurement trap: in the second quarter of 2025 the dollar’s reported share fell from 57.79 to 56.32 per cent, which read as dramatic diversification. At constant exchange rates it fell 0.12 percentage points. The dollar index had dropped more than ten per cent in the first half of 2025, its largest such move since 1973. Almost the entire “decline” was arithmetic.

The dollar’s long slide from around seventy per cent of reserves at the start of the century to the high fifties is real, but it is a slow diversification into a scatter of smaller currencies and gold, not a transfer to a rival. Meanwhile the dollar accounted for around 89 per cent of global foreign exchange turnover in the BIS 2025 Triennial Survey and roughly half of payment value on the SWIFT network — a figure that varies depending on whether intra-eurozone payments are counted, which is itself a useful reminder that most headline dominance statistics are measuring different things.

The genuinely new development is not in the reserve data. It is in the behaviour of the assets themselves. After the tariff announcements of early April 2025, the BIS observed that the historical correlations between US Treasuries and other safe assets — highly rated sovereign bonds, gold — and between Treasuries and gauges of risk such as the VIX had approached zero, which it suggested might indicate a weakening of Treasuries’ safe haven properties. The equivalent correlation for the German bund had, if anything, strengthened. The dollar fell while Treasury yields rose, inverting the pattern of decades in which stress drove money towards both.

For half a century the dollar strengthened when the world was frightened. In April 2025 the world was frightened and the dollar fell. That is the datum worth watching.

What the two ideas, corrected, actually tell us

Set the corrected versions side by side and a coherent picture emerges, quite different from the one in general circulation.

The exorbitant privilege is real but mispriced in the public imagination. It is the return on an intermediation business — borrowing safe, lending risky — and it comes with a contingent liability that is called in during global crises, when America transfers wealth outward on a scale measured in tens of percentage points of its own GDP and its central bank becomes the world’s lender of last resort. Whether the premium fairly compensates the risk is a genuine question. Whether it is a free lunch is not.

The Triffin dilemma is real in its general form — a national currency serving as an international public good does generate conflicts between domestic and global obligations — but the specific mechanism that made Triffin famous belonged to a gold-convertible world that ended in 1971, and the trade-deficit version now in political circulation is, on the evidence, an anachronism. Britain ran surpluses while sterling ruled. America ran surpluses while the world stocked up on dollars. The world’s dollars come from banks, not from the American trade balance, and the $14.7 trillion of offshore dollar credit is the proof.

The practical significance is that the two misreadings recommend the wrong policies. If dominance rests on trade deficits, then closing the trade deficit is a monetary strategy — which is roughly the argument now being deployed for tariffs and currency accords. If dominance rests on the depth of the Treasury market, the credibility of the institution that stands behind it, and the willingness of the issuer to act as lender of last resort in a crisis, then the things that would actually end it are fiscal deterioration, a compromised central bank, and a demonstrated readiness to make access to the asset conditional on political alignment.

Nobody knows what a genuine transition would look like, because the last one — sterling to dollar — took two world wars and about half a century, and even then the succession was settled in a committee room by a British economist answering a question from an Indian delegate about the convertibility of the rupee. Monetary orders do not usually end on a scheduled date, by the expiry of a document. They end the way Ernest Hemingway said men go bankrupt: gradually, then suddenly. The useful work, in the meantime, is to be clear about which of the load-bearing walls are actually load-bearing.

Robert Triffin’s deepest point was never the gold arithmetic. It was that entrusting an international public good to a national institution creates a structural conflict of interest that no amount of good faith fully resolves. That claim has survived every revision of his dilemma, and it is the one that matters now. The question facing the system is not whether America can keep importing. It is whether the country that issues the world’s money still wants the job on the terms the job requires — and whether the rest of the world can bear the answer either way.