The Forced Bond Seller Has a Name

Japan sends its pension capital home and the 30-year yields more than it did at the 2023 peak with a Fed funds rate a full point lower.

Settlements as of Friday, July 10, 2026

The Forced Seller Has a Name, and It Is Not a Bond Vigilante

The 10-year Japanese government bond touched 2.865% on Wednesday and the 20-year reached 3.85%, the highest levels since September 1996 and July 1996 respectively. Over twelve months the 10-year JGB has repriced by roughly 137 basis points. Then on Friday, Finance Minister Satsuki Katayama said the government would encourage domestic pension funds to increase their holdings of Japanese financial assets, and the 10-year fell roughly 10 basis points on the remark.

Read that sequence again, because the sovereign has just told you the mechanism out loud. Japan has published a draft roadmap calling for more than ¥370 trillion in public and private investment through fiscal 2040. That has to be funded. A government that needs domestic buyers for an expanding issuance calendar reaches for the largest pool of domestic capital in the world and points it homeward. The Government Pension Investment Fund does not conjure new yen to do this. It rotates. A repatriation into JGBs is, mechanically, a supply of foreign bonds.

For three decades the Japanese institution was the reliable marginal buyer at the long end of the U.S. curve. That was never charity. It was a carry trade underwritten by a domestic yield of zero. With the 10-year JGB near 2.9% and the 20-year near 3.9%, and with the hedging cost of dollar exposure where it now sits, the arithmetic that made the Treasury bid rational for a Tokyo lifer or a pension board has simply stopped working. The buyer is not staging a protest. He is being pulled home by a rate structure his own government created.

This is an Asian bond market adjustment, and we want to be precise that it is not, at this moment, a U.S. confidence event. The distinction matters, because the lazy version of this story (”yields are up, therefore America has lost the market”) is contradicted by our own settlement data, and we would rather make the argument that survives the table than the one that flatters the thesis.

Look at what the U.S. curve actually did in 2026. The 2-year is up 73.9 basis points year to date. The 30-year is up 24.0 basis points. The long end has sold off the least of any tenor. The 2s30s spread has compressed from roughly 136 basis points at year end to 86.2 today. Friday itself was the same shape in miniature: the 2-year gave up 4.6 basis points while the 30-year gave up 1.7. That is a bear flattener. A bear flattener is the front end repricing a central bank, not the long end repricing a country.

So where is the fiscal premium. It is in the level, and it is hiding in plain sight. The 30-year yield settled Friday at 5.070. On October 18, 2023, at the height of the bond vigilante panic, the 30-year traded 4.994 with a 2-year at 5.22. Today the front end sits roughly 100 basis points below where it stood on that day, and the long bond still yields more. Strip out the Fed and the long end has not merely failed to rally. It has ratcheted higher against a materially easier policy rate.

That is the fingerprint. Foreign selling into a curve that is simultaneously pricing domestic hikes does not print as a steepener. It prints as a long end that refuses to participate in the rally the front end is entitled to. The 30-year made a new cycle high of 5.181 earlier this year, above the 5.103 of October 2023. The market is absorbing supply, and time will tell whether it absorbs it cleanly. Many readers know that we believe the FOMC knows that higher rates are contributing to inflation now, whether they hint at the opposite to the masses doesn’t change our data and our facts that we too believe this is the case.

Our view on the front end is a forecast, not a description. The 2-year has already done its work, roughly 83 basis points off the 3.379 low set in February. From here we expect the front end to remain anchored by a patient FOMC. Nine of eighteen participants penciled in at least one hike before year end and the median 2026 funds rate moved to 3.8% from 3.4%, but Kevin Warsh submitted no dot, ran a 130-word statement, and has spent his short tenure arguing that supply shock inflation should be looked through. A chair who declines to publish a rate path is not a chair preparing to surprise you with one. We do not think the next move is a hike.

Which returns us to the only question that matters for the long end: who is left to buy it. Japan is going home. And as Section X details, the fastest growing pool of dollars in the world right now is buying bills, not bonds. The curve is not telling you a story about growth. It is telling you a story about who is buying which maturity.