Prices are rising while the economy turns down. A central bank in that position cannot raise on account of prices, nor cut on account of growth. Under a central bank that cannot move, real rates eventually fall, and the currency falls with them.
Note the tense. Real rates are not falling yet. Japan's 10-year real rate has risen from 0.27% at the start of the year to 0.84%, because expectations of rate hikes are priced into it. The fall begins when those expectations come off. This issue is about verifying, in numbers, the conditions for that to happen. The turn in the semiconductor cycle described in our companion piece is a sign that those conditions are falling into place.
One qualification: this is not the 1970s variety. Japan's price rises come from energy, not wages. It is not a wage-price spiral but a deterioration in the terms of trade, an outflow of income abroad. That distinction matters throughout.
What is actually moving in bond markets
Japan's 10-year yield has reached its highest level since 1996, Germany's since 2011, and the US 10-year has again traded above 4.6%. Reading all of this as an inflation scare would, in our view, be a mistake. Decomposing nominal yields into expected real short rates, expected inflation and the risk premium shows that the dominant mover in Japan and the United States is the first component: the repricing of the policy path itself. In Japan, roughly eight tenths of the year's rise comes from that channel. In the US, expected inflation has contributed almost nothing, sitting below its level at the start of the year even with oil up 50% and Hormuz traffic at one vessel a day: a broken thermometer declaring the patient's temperature normal.
Germany is the exception, and an instructive one. It is the only market where inflation expectations led the move, and also the one with the lowest risk premium of the three, because the ECB raised rates early and named the inflationary pressure for what it was. Acting credibly on prices protects the premium as a consequence. The case for the euro rests on this single point.
Common to all three: the risk premium is small and has narrowed everywhere this year. That is not reassuring. It means the spending made permanent by the resource blockage, the aid pledged to Ukraine and the enlarged defence budgets have not yet been priced into the long end at all.
Why the priced-in hikes are unlikely to arrive
Markets currently see Japan's policy rate some 180 basis points higher in five years, a repricing of a different order from the US or the euro area. That pricing rests on the premise that Japan's economy can withstand five years of rate rises. Three observations from the real economy undermine it: the global equity cycle appears to have peaked, Japan's production of electronic components has turned down with an unprecedented downward revision to forecasts, and inventory ratios have begun rising across Japan, the US and Taiwan. Production typically follows inventory ratios down within a month or two. If volumes fall, the demand-side case for hiking disappears; and since Japan's inflation is imported rather than wage-driven, responding to it with hikes would cut into demand that is already shrinking.
The fiscal architecture points the same way. Japan's government has replaced its primary-balance target with a falling debt-to-GDP ratio, to be achieved through nominal growth exceeding the interest rate, while the interest rate is decided by the central bank. The homework has been transferred; the person responsible was swapped, not the target loosened. But the strategy contains a flaw: growth driven by imported inflation does not lift nominal GDP, because the income leaves the country. Meanwhile the Bank of Japan's own balance sheet, earning a fraction of a percent on its bond holdings while paying policy rates on half a quadrillion yen of overnight deposits, has produced its first full-year negative carry. A central bank in that position cannot deliver the priced-in path. Our working conclusion is uncomfortable: the operation may succeed, and the patient may die, with the suppressed interest rate converted into an inflation tax through a weaker yen.
If the priced-in hikes come off, the ten-year yield need not fall: what comes off the policy component migrates into the risk premium. The same 2.9%, with the contents swapped. A worse 2.9%.
Erasing the premium, or making it unmeasurable
In the United States, the Treasury has expanded buybacks of long bonds, an operation too small to move a $31 trillion market, and attention is shifting to the November refunding announcement: whether long-bond issuance itself is cut, and whether operation sizes stop being announced in advance. That last detail matters more than it seems. If shorting long bonds becomes impractical, the risk premium can no longer be measured in the market. Whether the premium has fallen, or the means of measuring it has been removed, becomes indistinguishable. Regularity and predictability were themselves the collateral behind US Treasuries; removing them does not lower the premium, it makes it invisible.
One step further out sits the proposal to transfer the Federal Reserve's long-duration holdings to the Treasury and retire them. The probability is low but not zero, and the mechanics deserve attention, because two forces act on the premium at once. Thinner supply of long paper pushes yields down: insurers and pension funds must hold thirty-year assets against thirty-year liabilities. But the market also learns that the government erases debt when it becomes inconvenient, and there is precedent for what that lesson costs: when the US last exercised call provisions on its bonds, in 2009, investors were stripped of high-coupon paper and repriced the option into everything thereafter. A premium for interest-rate risk would be replaced by a premium against the government's intent. The first is benign when it narrows. The second produces no signal while it narrows, then opens all at once when confidence is lost.
Japan's lesson applies directly. Three buffers, a current account surplus, the world's largest net external assets and a captive domestic savings pool, and still, with rates held down, the currency halved. The pressure did not vanish; the place where it is priced moved from rates to the exchange rate. The United States has none of those buffers, and roughly 30% of its debt is held abroad.
The five-light test
Our test for the final stage is unchanged: widening breakevens, falling 10-year real rates, a steepening curve, a weaker currency and rising gold, all appearing together, signal a market concluding that the central bank has begun underwriting the deficit. Nothing is lit today, in either country, and that is the accurate reading of where we stand. The sequence to watch runs through Japan's September policy meeting, the US inflation prints, and above all the Treasury's November refunding. Dollar-yen, meanwhile, has moved barely 1% all year despite trillions of yen of intervention. That is not calm. It is two currencies sinking at the same speed, with the aggregate cancelling out the deterioration of both. The remaining yardsticks are the euro and gold, and we believe the period of large moves in currency markets has only just begun.