Currencies move relatively: normally, if one falls, the other rises. What is unusual about the present phase is that the yen and the dollar are sinking in the same direction at the same time. Dollar-yen has moved little more than 1% year to date, but this is not calm; it is two currencies degrading at a similar speed, with the aggregate cancelling out the deterioration of both. The remaining yardsticks are the euro and gold, and gold's move towards record levels this summer is, in our reading, the answer key.
What is actually moving in bond markets
Ten-year yields have reached multi-decade highs this year in Japan, Germany and the United States. Reading this collectively as inflation concern would, we believe, be a mistake. Decomposing nominal yields into expected real short rates, expected inflation and term premium shows that the dominant driver in the United States and Japan is the first component: the market has reversed its pricing of the policy path, from cuts to hikes, while measured inflation expectations have barely moved and term premia have actually compressed.
The American case contains an irony: with crude up sharply and Hormuz effectively closed, breakeven inflation sits below its level at the start of the year. Since the crude price does not reflect physical reality, the inflation expectations derived from it do not move either: a broken thermometer reporting normal temperature. Germany is the exception that proves the rule: the only market where inflation expectations lead the move, and also the one where credible central bank action has kept the term premium the lowest of the three.
Japan's rise is of a different order. The market is pricing a policy rate several years out far above today's, a gap of roughly 180 basis points against the current rate, triple the equivalent repricing in the United States. This is not the removal of suppression on yields; it is the pre-pricing of hikes the market believes must come.
Japan: changing who does the homework
Japan's fiscal framework has quietly changed hands. The government has dropped its primary-balance objective in favour of lowering the debt-to-GDP ratio through nominal growth exceeding the interest rate, the Domar condition. The decisive detail is that while the government controls spending, the central bank controls the interest rate: the homework has been transferred from the ministry of finance to the Bank of Japan. Once that transfer is made, restraining rate hikes is not ignorance but consistency, the logic Sargent and Wallace called unpleasant monetarist arithmetic: with fiscal policy fixed, tightening today enlarges the debt that must eventually be absorbed, and therefore tomorrow's inflation.
The Bank of Japan's own balance sheet makes the bind concrete. Its bond portfolio yields a fraction of a percent and reprices only slowly, while the interest it pays on current-account deposits adjusts almost immediately, producing its first full-year negative carry. A central bank whose income improves by a third of a percent a year cannot indefinitely pay policy rates near 3% on half a quadrillion yen of deposits. Either the priced-in hikes are delivered, or they peel off, and if they peel off while purchases continue, what falls is not the nominal yield but its quality: the same 2.9%, with policy expectation swapped for term premium. A worse 2.9%.
The July intervention fits the same picture. Six to seven trillion yen of reserves, and euro sales by the US Treasury, bought a level that partially retraced within weeks, while the pressure re-emerged in yields. Intervention buys time, not levels, and there are fewer bullets the second time. It is also telling that Washington defended the yen without selling a single dollar.
The mechanism being prepared in Washington
The US Treasury has expanded buybacks of long bonds, an operation too small, at a few billion dollars a time, to move a $30 trillion market. The consequential idea sits one step further out: proposals to redefine the Treasury-Fed relationship so that the Fed's multi-trillion-dollar holdings of long-duration bonds could be retired rather than returned to the market. We assign this a low probability, but not zero, and the asymmetry matters. Removing duration from the market permanently converts long-term debt into an annual refinancing, the same structure as the Bank of Japan's negative carry, and if such an operation were ever executed for political convenience, the market would learn the lesson and charge a permanent premium on everything. A mechanism designed to eliminate the term premium would end by raising it.
Our discrimination method is a five-light panel: widening breakevens, falling real yields, curve steepening, currency weakness and rising gold, appearing together. That combination signals a market concluding that the central bank has begun to shoulder the fiscal deficit. The United States has not lit up. Japan is closer, and the Bank of Japan's September meeting is, in our view, the heaviest date on the calendar: what is at stake is not the level of the ten-year yield but its composition.
In a phase where both currencies sink, positions expressed purely in dollar-yen can only capture the difference in sinking speeds. The clearer expression, in our view, favours the yardsticks that do not degrade with either: gold, and the euro. Looking only at dollar-yen, this entire movement looks as if nothing is happening. We believe the phase of large moves in currency markets has only just begun.