Every system
exports its contradictions.

The story usually told is about goods. Money expanded, cheap imports arrived, the consumer price index stayed quiet, and policy stayed loose because the index said it could. That is the first loop and it is real.

The second loop is the one that goes missing. The dollars paid for those goods did not stay abroad. They came back and bought government bonds, which held long term interest rates down, which made deficits cheap to run, which financed more buying. One country was absorbing the monetary expansion on the goods side and financing it on the bond side at the same time.

That matters because the two loops fail for different reasons and people therefore treat the failures as unrelated news. They are not unrelated. They are the same absorber withdrawing from two circuits.

The surpluses are still being run. In July 2026 the monthly trade surplus was 112 billion dollars. Those dollars have to be held as something, and the something changed.

Chinese holdings of United States government debt were 633 billion dollars in June 2026, roughly half the 1.3 trillion peak of 2013 and down 14 percent since the start of 2025. Over the same stretch the People's Bank of China bought gold for twenty consecutive months, and accelerated its buying through a quarter in which the gold price fell 16 percent. A buyer who buys more when the price falls is not trading. That is a reserve being rebuilt.

The European Central Bank recorded the crossing point. At the end of 2025 gold was 27 percent of global official reserve assets against 22 percent for United States government debt. Gold is now the largest reserve asset in the world for the first time since 1996, and central banks hold more than 36,000 tonnes, close to the levels of the Bretton Woods era.

When both stabilising loops are gone, one balancing move remains: let the currency carry the adjustment. Inflation lifts nominal output, tax receipts and wages, while old fixed rate debt stays denominated in yesterday's money. The debtor gains, the creditor is repaid in full in a weaker unit, and nothing has to be declared.

The catch is arithmetic. That trick works on debt that is long and fixed. The size of the real gain scales with how many years of it are locked in. Shorten the average maturity and the stock reprices at the new higher interest rate before inflation can erode it, so the same inflation that used to reduce the burden simply raises the coupon instead.

Now look at what is being done about the bond market. Buy back the long end, lean on shorter issuance, keep the thirty year quiet. Each of those steps shortens the average maturity. The tactic that relieves the symptom is dismantling the cure. Stage four is that trade off, and you can move the slider yourself.

The West's absorber was a foreign labour force. That labour force has an absorber of its own. Between 280 and 320 million people in China are counted as employed while holding no permanent contract, no employer obligation and no insured position, and that number has climbed through exactly the years the industrial buildout matured and investment turned negative. The gig economy is doing there what the China trade did here: holding a surplus that the formal system has no room for.

This is the shape worth taking away from the lesson. A system under strain does not resolve the strain, it relocates it, and it stays legible as long as there is somewhere further out to put it. The currency is the last absorber in the chain for one reason only. It is the only one with no counterparty to push to.

The numbers on the gauges are real and sourced, but the model is a teaching instrument and not a forecast. Three limits worth stating plainly.

The erosion figure in stage four is a first order approximation. It treats a sustained inflation surprise as eroding the fixed portion of the debt stock in proportion to its remaining average life. Real debt dynamics involve the primary deficit, the growth rate and the reaction of the interest rate, none of which are in this dial.

The loops are drawn as clean circuits. Real ones have lags, and the lags matter. Tariffs break the price loop within months while the reshoring they are meant to cause takes years, so there is a stretch with the inflation and none of the capacity. Nothing in this diagram shows that delay.

And the last stage is a point about measurement, not a prediction. If every currency loses purchasing power at a similar rate, the exchange rates between them barely move and the instruments most people watch will report that nothing has happened. That is not evidence of stability. It is the wrong ruler.