The Bank of Japan's quantitative tightening is real, fast and moving long-dated JGB yields. It is a weak yen defense, though, because the currency's funding economics are set at the short end, where the BOJ has barely moved. QT's more consequential effect lands elsewhere. By pushing the 10-year JGB to 3.09% and the 30-year to 4.21%, the BOJ has already made domestic duration beat currency-hedged US Treasuries for Japanese institutions. That pulls Japanese capital toward home through the bond market, slowly, while carry traders keep selling yen at the front end. The policy is tightening two channels that point in opposite directions for the currency.
A Balance Sheet Shrinking Faster Than the Rate Path Suggests
The scale is not in dispute. Total assets fell another ¥15.5 trillion in the quarter ended September 30. That brings the year-over-year decline to ¥70.8 trillion and the balance sheet to ¥625 trillion, the lowest since March 2020, according to the BOJ's own balance sheet. From the Q1 2024 peak, the bank has shed ¥131 trillion, or 17.4% of total assets.
The bond line is moving fastest. JGB holdings dropped ¥14.3 trillion in the quarter, the largest quarterly decline since QT began, to ¥504 trillion. Since the 2023 peak, holdings of Japanese government securities are down ¥88.2 trillion, or 14.9%. The BOJ no longer holds any Treasury bills. Every yen of remaining government paper on its books sits further out the curve.
Policy rates tell a different story. The first hike came in March 2024, and the policy rate has reached only 1.25%. That is the highest in 31 years, and it is still trivially low against inflation that Wolf Richter, writing at Wolf Street, describes as flourishing. The Fed, by contrast, ended QT nearly a year ago. The two central banks have swapped tools: the Fed holds its balance sheet steady and leans on the policy rate, while the BOJ runs down its balance sheet and keeps the policy rate close to the floor.
This combination is not accidental. A central bank that wants higher long yields without raising the cost of short-term credit for a heavily indebted state and a fragile banking system would choose this mix. The question for the yen is whether long yields are the lever that moves it.
Where the Currency Actually Gets Priced
Currency carry is funded short. A trader who borrows yen to hold dollar assets pays the Japanese short rate and earns the US short rate. With the US 2-year at 4.84% (as of October 5) against a BOJ policy rate of 1.25%, the front-end gap is roughly 3.6 percentage points. The policy rate is an imperfect stand-in for the JGB 2-year, which is not cited here. The order of magnitude still holds. That gap is the carry, and QT does nothing to it directly. Running off long JGBs does not raise the overnight rate a yen borrower pays.
The long end looks better for the BOJ but still unfavorable. The US 10-year sits at 5.31% against the 10-year JGB at 3.09%, a 222 basis point spread. That spread was much wider when JGB yields were negative in 2020, so QT has done real work in compressing it. The yen has kept weakening anyway. The most plausible reading is that the front-end differential, not the long-end spread, is the binding variable for the exchange rate.
This is the central problem with describing QT as the BOJ's primary yen tool. It is a powerful instrument aimed at the part of the curve that matters least for short-horizon FX flows.
The Hedged-Return Squeeze on Japanese Duration Buyers
The second-order effect is where QT has teeth. Japanese life insurers, pension funds and banks are among the largest foreign holders of long-dated global bonds. When they buy Treasuries, they choose between hedging the currency and leaving it open.
Hedging costs approximately the short-rate differential between the two currencies. Using the 3.6-point front-end gap as a rough proxy, a currency-hedged 10-year Treasury yields something close to 1.7% to a yen-based investor: 5.31% minus about 3.6%. The domestic 10-year JGB now pays 3.09%. On a hedged basis, home wins by roughly 140 basis points, and the 30-year JGB at 4.21% widens that gap further for liability-matching buyers who need long duration.
The precision deserves a caveat. Actual hedge costs run off three-month forward points, not 2-year yields, and institutions hedge different shares of their books. The direction does not depend on the decimal point. Each ¥14 trillion quarter of BOJ runoff pushes JGB yields higher, which raises the hurdle every hedged foreign bond has to clear for a Japanese buyer.
That leaves Japanese institutions with three options for their overseas duration:
- Repatriate into JGBs at 3% to 4%, which supports the yen at the margin as foreign assets are sold and proceeds converted.
- Stay abroad unhedged, which is a bet that the yen keeps weakening, so the position adds to yen selling rather than relieving it.
- Stay abroad hedged at a yield now below what home pays, which is the option QT makes hardest to justify.
QT therefore helps the yen through capital flows, not through carry. The channel is real but slow. Insurers rebalance over quarters, not days, while carry positions turn over on the front-end differential every session.
The Long End Japan Is Exporting
The cross-asset consequence follows from the same mechanism. If Japanese duration buyers step back from hedged Treasuries, US term premium loses a structural bid. With the US 10-year at 5.31% and the 2s10s curve positively sloped at 0.47 percentage points, the US long end is already trading as if marginal buyers are demanding more compensation. This piece does not attribute that move to Japan. US fiscal supply and the Fed's own policy path are larger and better documented drivers. The direction of Japan's contribution is unambiguous all the same. A creditor nation whose central bank is draining ¥14 trillion of bonds a quarter, and whose domestic yields are climbing toward foreign ones, is withdrawing from global duration rather than adding to it.
That positions BOJ QT as a quiet input to global long rates, not just a domestic housekeeping exercise. The Fed stopping QT eased one source of Treasury supply to private hands. The BOJ continuing QT is removing one source of Treasury demand from Japanese hands.
The Countercase: Maybe QT Is Doing Exactly What Tokyo Wants
The strongest objection is that this misreads the BOJ's intent. The bank may not be aiming at the spot exchange rate at all. It may be aiming at the long-run current-account mix: higher domestic yields that keep Japanese savings at home over years, so the yen stops leaking on structural outflows even if carry dominates week to week. Read that way, the repatriation channel is the target, and judging QT by the near-term yen path is the wrong test.
There is support for this view in the balance-sheet composition itself. The BOJ has dumped every T-bill, let pandemic loan programs run down 55% from their Q1 2022 peak to ¥68 trillion, and taken corporate bond and commercial paper holdings down 91% from peak. That is a bank normalizing the long-term structure of its footprint, not one fine-tuning an exchange rate.
The equity book points the other way, and it is where the patience shows most. The BOJ began selling ETFs and J-REITs this year and disposed of ¥90 billion in Q3, measured at acquisition cost. Because those holdings were bought from 2011 onward and are not marked to market, while the Nikkei 225 has risen about 500% over that span, the market value of what was sold is several times the booked figure. Even so, combined equity-type holdings are down only 1.3% from peak. It took nearly a decade to exit the bank-stock portfolio, with the last of it sold in Q3 2025. Equity sales at this pace carry no meaningful FX or rates signal, and the bond runoff is the only part of QT doing macro work.
The countercase is coherent, but it concedes the thesis's main point. If QT is a multi-year capital-account tool, it is not propping up the yen on any horizon a currency trader uses. A collapsing yen in the meantime forces the BOJ back to the front end it has been avoiding.
The Spread That Settles It
The decision variable is the gap between the two parts of the differential. If the BOJ continues to run off ¥14 trillion a quarter and the 10-year spread to Treasuries narrows meaningfully below 200 basis points while the yen keeps falling, that confirms the front end is the binding constraint and QT alone cannot defend the currency. The next policy move would then have to be a faster hiking pace, which the BOJ has resisted since March 2024. If, instead, Japanese institutions show sustained net selling of foreign bonds over the coming quarters and the yen stabilizes without a front-end hike, the slow repatriation channel is working and the countercase wins.
Until one of those prints, the cleaner reading is that Japan has built a tightening program that pressures global long yields more reliably than it supports its own currency.
