The Bank of Japan has put a number on how much its own retreat from the bond market pushed long-term yields up, and the answer is about 25 basis points

The rise in Japanese long-term yields is usually laid at the door of the central bank that stopped buying. On Tuesday the Bank of Japan published its own estimate of how much of that rise its withdrawal actually caused, and the number is smaller than the framing implies.

About 25 basis points.

That is the combined contribution of the Bank's reductions in Japanese government bond purchases to the rise in long-term interest rates since those reductions began in summer 2024, on estimates published by the Bank's Monetary Affairs Department and Financial Markets Department. Ten year yields are currently in the upper 2 percent range. The paper is careful about its own arithmetic, describing the estimates as attempts based on a simple regression model whose results should be interpreted with some latitude.

Two effects, and only one of them works in both directions

The Bank separates the mechanism in the conventional way. A stock effect is the downward pressure on rates that comes from the size of the holdings. A flow effect is the pressure that comes from the act of buying.

Estimated on its own, the waning stock effect accounts for around 10 basis points of the rise since reductions began. Adding the flow effect takes the total to about 25.

The asymmetry underneath that second number is the more interesting result. In the phase when the Bank's holdings were increasing, the paper finds no statistically significant flow effect at all, which is to say that buying more did not measurably compress term premiums. In the phase when holdings are falling, the effect is significant, and the reductions push term premiums up. The exit, on these estimates, is not the entry run backwards.

The Bank also splits the yield rise itself into an expected short rate component and a term premium, and finds the two have contributed to roughly the same degree. It attributes the first to market participants revising up their view of where the policy rate peaks, against a gradual rise in underlying inflation.

Forty nine trillion yen out, ninety five trillion in

Between the end of June 2024 and the end of March 2026 the Bank's holdings fell by 49 trillion yen, because redemptions have been running ahead of purchases. Over the same period the outstanding stock of government bonds grew by 46 trillion yen. Holdings by everybody else therefore rose by 95 trillion yen.

The paper says where it went. Domestic depository corporations added 36 trillion yen, overseas investors 27 trillion, households 6 trillion. Life insurers went the other way and cut holdings by 5 trillion yen, having already lengthened their books to meet the economic value based solvency rules.

Overseas investors are now the largest single presence in the trading day. At the end of December 2025 they accounted for just under 50 percent of cash bond transactions and more than 70 percent of futures transactions.

Households are the smallest of the four and the one the Bank spends most words on. Their bond holdings are about 1 percent of roughly 2,400 trillion yen of household financial assets, and less than 2 percent of the bonds outstanding.

The market is answering questions again

The purchase plan has been running against a monthly amount of nearly 6 trillion yen before reductions started. By March 2027 it will be around 2 trillion, about a third of that, after which the Bank stops cutting.

What that has bought, on the paper's account, is price discovery. The dip in the curve at the seven to ten year point, where the Bank owned most of what was on offer, has largely gone. Futures prices now fall when an economic release beats expectations, which they did not reliably do while the Bank was buying at scale.

Fiscal 2026 issuance is planned at 112.2 trillion yen, down from 113.8 trillion the year before.