NextFin News - Foreign investors sold Japanese medium-term government bonds at the fastest pace in two decades, official data showed this week, a reversal that lands on a market already pricing the steepest yen-denominated borrowing costs in thirty years and asking whether Japan's era of cheap debt is over for good.
The outflow, the largest from the medium-term segment since 2006, is more than a positioning footnote. For most of the past twenty years, overseas accounts were the marginal buyer of Japanese government debt, absorbing issuance that domestic banks and insurers would not touch. When that buyer steps back at a speed not seen since the mid-2000s, the market loses its shock absorber. And it is doing so as the benchmark 10-year yield touched 2.945% earlier this week, the highest level since September 1996, before easing to 2.90% on August 19.
The question this move forces is not whether yields have risen - they have, by 1.29 percentage points in a year - but whether the rise is a cyclical overreaction to fiscal noise or the first leg of a permanent repricing of Japanese credit. The answer determines who gets hurt, who gets paid, and how much room the Bank of Japan has left.
The last time foreign accounts exited the medium-term segment at this pace, in 2006, Japan was emerging from a different world. The Bank of Japan had just ended its first experiment with zero rates, the 10-year yield was climbing away from sub-1% territory, and the global search for yield was still in its infancy. Two decades later, the setting has inverted: the world is drowning in debt, Japan's central bank is unwinding the largest balance-sheet expansion in the Group of Seven, and the investors who once treated JGBs as a funding-leg trade are now being asked to hold them as a return-generating asset. The 2006 comparison is not a perfect parallel, but it marks the boundary of the data - and it tells you how rare a sustained foreign exit from this part of the curve has been.
The Medium-Term Segment Is Where Fiscal Credibility Gets Priced
The medium-term part of the curve - roughly the two- to ten-year maturities - is where monetary policy expectations and fiscal credibility collide. Short-term yields are anchored by the Bank of Japan's policy rate, lifted to 1.0% in June from 0.75%, the highest since 1995. Ultra-long yields are set by structural balance-sheet buyers: life insurers matching decades-long liabilities, and foreign accounts hunting duration at yields that finally compensate for risk. The belly of the curve has no natural owner. It is the most sensitive to where the policy rate goes next and whether the government can fund itself without a disorderly jump in borrowing costs.
That is why foreign selling concentrated in medium-term bonds is a sharper signal than selling at either end. It says investors are not merely taking profits on long duration - they are repricing the entire expected path of Japanese rates. By August 19, the two-year yield stood at 1.68%, the five-year at 2.13% after reaching a record 2.19% this month, the 20-year at 3.78%, the 30-year at 4.09% and the 40-year at 4.16%. The 30-year printed an all-time high of 4.20% in May. The yen traded near 158.4 against the dollar, a level that imports inflation through energy and food prices and keeps the central bank under pressure to move sooner rather than later.
The trigger sits in the fiscal 2027 budget, which Prime Minister Sanae Takaichi's administration will compile before seeking cabinet approval at the end of the year. The government has approved cutting the consumption tax on food to 1% from 8% for two years starting April 2027, a measure estimated to reduce annual revenue by roughly 4.4 trillion yen. Officials say the shortfall will be covered by a zero-based review of spending, tax breaks, subsidies and public funds rather than deficit-covering bonds, but the details remain scarce. Markets are pricing the uncertainty directly into the belly of the curve, where the extra supply from a widening deficit would land first.
The supply math is unforgiving. The Ministry of Finance must refinance hundreds of trillions of yen of maturing debt every year on top of any new issuance, and the interest cost on that stock is resetting higher with every auction. A 10-year bond issued five years ago at zero or negative yield rolls off into a market demanding nearly 3%. That is not a marginal change in the government's budget; it is a structural increase in the fixed cost of running the state. The Takaichi administration has pledged not to use deficit-covering bonds to fund the food-tax cut, but that pledge only holds if spending reviews and subsidy clawbacks deliver - and markets discount promises that have not yet been legislated.
Meanwhile the buyer that used to neutralize this supply is stepping back. The Bank of Japan's balance sheet swelled to more than 130% of GDP during the yield-curve-control years, and its outright purchase schedule has been cut quarter by quarter on a pre-announced path. Reducing purchases does not mean selling - the BOJ is not quantitatively tightening in the strict sense - but it means the marginal tonnage of new and refinanced debt must find a home elsewhere. In a market where foreign ownership of JGBs had already fallen to roughly 12%, near the lowest level in years, the medium-term outflow removes precisely the buyer the market needed most.
A Cyclical Wave Riding a Structural Shift
The right read separates two forces that are operating at once, because they point in opposite directions across time horizons.
The cyclical leg is clear and mean-reverting. Foreign capital in Japanese government bonds is not ideologically committed in either direction; it is yield-sensitive and tactical. Overseas investors net sold a record ¥10.79 trillion of JGBs for the full year 2022, then returned. In June 2026 they offloaded ¥3.12 trillion of Japanese bonds, the largest outflow in three years, only to have bought aggressively into the 20- to 30-year segment once yields broke above 3.5%. A record ¥9.3 trillion flowed into longer-dated Japanese debt in 2025 alone. On this evidence, the current medium-term selling looks like a reaction to a specific set of conditions - fiscal headlines, a weak yen, elevated hedging costs - that can unwind as quickly as they assembled.
The structural leg will not revert on its own. Japan spent two decades as the silent subsidizer of cheap global borrowing. That era is over. The Bank of Japan's policy rate sits at 1.0%, up from -0.1% in 2024, and inflation has exceeded the central bank's 2% target for much of the past four years, with tight labour markets, solid wage growth and rising inflation expectations pointing to entrenched reflation. As domestic yields rise, Japanese institutions are repatriating capital - selling $29.6 billion of US debt in the first quarter of 2026 alone - and removing a historically reliable buyer from markets already navigating large fiscal deficits. This is a regime change in the cost of Japanese money, not a temporary dislocation.
The transmission runs through three channels, and all three are open at the same time. First, the supply channel: a larger deficit means more bond issuance, and the central bank is no longer absorbing it at the pace it once did. The Bank of Japan has been reducing its monthly purchases on a pre-announced schedule, handing price discovery back to the market. Second, the inflation channel: a weak yen and elevated energy costs keep price pressures above the comfort zone of policymakers, which argues for faster rate hikes and pushes the whole curve up. Third, the currency channel: foreign holders of yen bonds hedge their currency exposure, and when the yen weakens or hedging costs rise, the all-in return deteriorates even if the local yield does not move. That is the mechanism behind a medium-term outflow that is about Japan's rates path, not just Japan's bond prices.
The hedging channel deserves its own emphasis, because it is the piece most investors outside Japan miss. A US or European asset manager buying a five-year JGB at 2.13% does not earn 2.13%. It earns the yen yield minus the cost of swapping the coupon and principal back into dollars for the life of the bond. When the yen was stronger and Japanese rates were negative, that swap cost was low or even favourable. With the yen near 158 and the rate gap between the Federal Reserve and the Bank of Japan still wide, the hedged return on a medium-term JGB can be negligible or negative even as the local yield climbs. That arithmetic explains why foreigners can be sellers of the belly while still buying the long end: the ultra-long segment offers enough local yield - above 4% at 30 years - to overcome the hedge, while the medium-term segment does not. The outflow is rational portfolio construction, not a vote of no confidence in Japan's solvency.
Put against peers, the repricing looks less like a Japan-specific crisis and more like a global re-rating of sovereign duration. Long-term yields in the US, UK and Europe have also climbed to multi-year highs as fiscal deficits and sticky inflation reset the term premium worldwide. Japan's 10-year at 2.90% still sits well below the US 10-year near 4.7% and the UK 10-year above 5%. On that measure, Japanese bonds are not expensive; they are simply no longer free. The shift from "free money" to "fairly priced money" is what feels violent to a market that spent twenty years being paid not to hold risk.
The Counter-Thesis: A Fair-Value Reset, Not an Abandonment
The strongest case against a darker read is straightforward: foreign investors are not fleeing Japan; they are refusing to buy cheap. Net selling of medium-term bonds at 2% yields is entirely consistent with net buying of 30-year bonds at 4%. The data show foreigners piling back into the ultra-long segment once compensation arrived. On this view, the outflow is a valuation event, not a confidence event, and the "since 2006" framing is a statistical artifact of a segment that saw little foreign turnover during the yield-curve-control years.
There is also a buyer of last resort in the system. The Bank of Japan remains the largest holder of JGBs and retains the tools to cap yields if the move becomes disorderly. And on the fiscal side, some analysts argue that even after the proposed tax cuts, Japan's net government debt could fall to around 100% of GDP by the end of next year, which would make the bond-vigilante narrative premature. Gross government debt stood at 248.7% of GDP at the end of 2025, nearly 250%, but net figures strip out the government's financial assets, and Japan's external creditor position remains the largest in the world.
There is a second pillar to the counter-thesis: domestic demand. Japanese banks, flush with deposits and facing a steeper curve, have been expanding their JGB holdings as net interest margins improve. Regional banks in particular have room to add duration at yields their loan books cannot match. Life insurers, despite their unrealized losses, still need yen assets to match yen liabilities, and a 30-year at 4% is closer to the credited-rate assumptions baked into their legacy policies than anything available since the 1990s. If domestic accounts step in as foreigners step out, the medium-term selling is absorbed without a disorderly move - and the "since 2006" outflow becomes a footnote rather than a regime marker. The risk is timing: domestic buyers tend to accumulate gradually, while foreign selling can be fast, and the gap between the two is where volatility lives.
The rebuttal is that the medium-term segment does not care about the ultra-long bid. A government that must fund itself at 2.13% on five-year money and 2.90% on ten-year money faces a higher interest burden regardless of what happens at 30 years. Debt servicing costs rise on the stock of refinanced debt, not on the marginal 40-year issue. And the Bank of Japan's ability to cap yields is precisely what created the inflation and currency problem in the first place; using it again would trade bond-market stability for yen weakness and imported inflation. The counter-thesis holds only if yields stabilize here. If the 10-year pushes decisively through 3% and the 30-year retests its May high of 4.20%, the "fair-value reset" story gives way to a repricing of fiscal risk.
What Comes Next: Scenarios Across Three Horizons
The base case is continued volatility with an upward bias in yields. The trigger is the fiscal 2027 budget. If the package delivers tax cuts without a credible offset, the medium-term curve has room to reprice higher, and the 10-year spends more time above 3% than below it. The upside case for bonds - lower yields - requires either a more hawkish-than-expected Bank of Japan that convinces the market inflation is being tackled, or a fiscal pivot that restores confidence in the medium-term path of debt. Economists surveyed expect the policy rate to reach 1.25% in the fourth quarter and 1.50% by the second quarter of next year, a path that would cap the two-year yield's rise but leave the long end exposed to supply concerns. The downside case is a disorderly move through 3% on the 10-year, which would test the central bank's tolerance and force a choice between yield control and currency stability - the same trade-off that made the "widow-maker" trade famous.
The calendar of catalysts is crowded. The Bank of Japan's next policy meeting in September is the first inflection point: a 25-basis-point hike to 1.25% would validate the market's pricing and could paradoxically calm the curve by removing uncertainty, while a hold would be read as the central bank falling behind the curve and could push the 10-year through 3%. The summary of opinions from the previous meeting already showed a growing number of board members calling for a stronger response to inflation, which is why the hike is widely anticipated. The second inflection point is the autumn budget formulation, when the size of the fiscal 2027 spending envelope becomes concrete. The third is the US Federal Reserve's path: a weaker dollar would relieve pressure on the yen and give the Bank of Japan more room to move gradually; a stronger dollar would tighten the noose.
Short-term, the market is driven by sentiment and positioning: the weekly foreign-flow print, the yen, and headline risk from budget negotiations. A single week of foreign net buying would calm nerves; a second consecutive week of heavy selling would reignite the question of who is left to buy. Medium-term, fundamentals dominate: the size of the deficit, the pace of the Bank of Japan's balance-sheet reduction, and whether underlying inflation stays above target. Long-term, the structural question is whether Japan can normalize policy without a gilt-style disordering of the long end.
The asymmetry runs through the curve and through the investor base. Banks benefit from a steeper curve and wider net interest margins, which is one reason financials have outperformed. Life insurers sit on large unrealized losses and face a dangerous zone if the 30-year breaches 4.5%, where mark-to-market pressure could force selling into a falling market. Exporters face a weaker yen as a tailwind on overseas earnings but higher domestic financing costs on their yen debt. Foreign bondholders who stay hedged must accept that the cheap-funding subsidy is gone; those who stay unhedged are making a currency bet as much as a rates bet.
The signal that would prove the structural-shift judgment wrong is specific and observable: if the 10-year yield settles back below 2.5% within three months and foreigners return to net buying of medium-term bonds for three consecutive weeks, this is a cyclical dip, not a regime change. The signal that would confirm it is a sustained break above 3% on the 10-year with the Bank of Japan still holding at 1.0%.
"JGBs are increasingly moving from 'uninvestable' to 'investable' for global bond investors," said Masahiko Loo, senior fixed income strategist at State Street Investment Management. "Foreign investors have piled back into the 20- to 30-year segment as yields broke above 3.5%, with a record 9.3 trillion yen flowing into longer dated Japanese debt in 2025 alone."
That quote captures the fork in the road. The same investors who sold medium-term bonds this week are the ones who bought the long end last year. They are not ideologues; they are price-sensitive. The market is no longer asking whether Japanese yields will rise. It is asking how high they must go before someone stops selling - and whether the answer is a number the government can afford.
For the global bond market, the stakes extend beyond Japan. Japanese investors are the largest foreign holder of US Treasuries, and their repatriation - $29.6 billion of US debt sold in the first quarter of 2026 alone - removes a price-insensitive buyer from the deepest market in the world. If Japanese institutions continue to bring money home as domestic yields rise, the US Treasury market must find replacement demand at higher yields. That is the second-order effect of a Japanese medium-term outflow: it is not just a local rates story, it is a reduction in the global supply of patient capital. The world's cheapest funding currency is becoming a normal funding currency, and every asset priced off the assumption that it would stay cheap has to be re-underwritten.
Explore more exclusive insights at nextfin.ai.
