Japan and the Price of Duration Everywhere

On 20 January 2026 the yield on the 40-year Japanese government bond rose above 4% and peaked at 4.24%. That maturity was introduced in 2007 and had never traded with a four handle, and the 30-year moved 25 to 30 basis points in the same session, its largest single-day change since 1999.

Japanese long bonds are not usually interesting; that is rather the point of them, and it is why the move matters well beyond Japan. For thirty years Japan has been the world’s structural buyer of duration: a large pool of domestic savings, a central bank suppressing yields, and institutions that had to look abroad to find any. This piece argues that the January repricing is best read as the marginal buyer of global duration discovering it has something to do with its money at home, and that the transmission to other markets runs through capital flows rather than through any change in their own fundamentals.

What happened in January

Prime Minister Sanae Takaichi called a snap election and announced a stimulus package of ¥21.3 trillion, roughly $115 billion, including a two-year suspension of the 8% consumption tax on food; the bond market’s response was immediate and concentrated at the long end.

Figure 1 shows the move against the threshold it crossed: the 4.00% line had held since the maturity was created, and the market went through it in a single session.

Three structural conditions made the reaction that large. The Bank of Japan had already raised its policy rate to 0.75%, ending the era in which the long end faced no competition from cash; inflation had returned; and gross government debt stands near 240% of GDP, which means a fiscal expansion is priced against a balance sheet with no headroom.

Figure 2 is the part worth sitting with. The Bank of Japan sets 0.75%. The market sets 3.90%. A gap of more than three percentage points on the same currency is not a forecast of where policy is going, because no plausible tightening path reaches 3.90%. It is a price for something else, and the something else is the probability that the government keeps borrowing at this pace.

Figure 3 gives the fiscal backdrop. Gross debt near 240% of GDP is roughly double the United States on the same measure, and the stimulus that triggered the January move was added on top of it.

Why this is not a Japanese story

Japan’s savings had to go somewhere: for three decades, with domestic yields near zero, they went abroad, and Japanese institutions became one of the largest foreign holders of US, European and Australian government debt. The yen carry trade is the leveraged version of the same flow: borrow cheaply in yen, buy something yielding more anywhere else.

Both mechanisms depend on the same condition, which is that Japanese assets offer nothing. When a 40-year JGB yields 3.90%, a Japanese life insurer with yen liabilities no longer needs to take currency risk to meet them; it can buy its own government’s paper instead.

The reversal has begun, and Japanese investors sold $29.6 billion of US debt in the first quarter of 2026 alone.

Channel	Mechanism	Where it lands
Repatriation	Domestic yields become adequate for yen liabilities	Long-dated sovereign markets in the US, euro area, Australia
Carry unwind	The funding leg stops being free	Emerging market currencies and high-carry assets
Term premium	One large price-insensitive buyer withdraws	Steeper curves everywhere, independent of local policy
Hedging cost	Wider rate differentials change the cost of hedged buying	Reduces appetite even where nominal yields look attractive

Transmission channels as described in market commentary following the January repricing. Sources: CNBC, Wright Research, Trading Economics.

Note that none of these channels requires anything to go wrong in the receiving country. A country with stable inflation, a credible central bank and a sound budget can still see its long yields rise, because the buyer who used to absorb its issuance is buying at home instead. Therefore the correct reading of a term premium increase in 2026 is often not a judgement on the issuer at all.

The same mechanism, a different creditor

This is the second time in a year that a large pool of foreign capital has turned inward. Gulf sovereign wealth funds did it for fiscal reasons after the Iran war, with the Saudi Public Investment Fund cutting its international allocation from a peak of 30% to 20% of a portfolio near $925 billion. Japanese institutions are doing it for yield reasons.

The causes are unrelated and the effect is identical: in both cases the marginal buyer of dollar-denominated assets is smaller than it was. Furthermore, this is happening while the Federal Reserve reduces a balance sheet of roughly $6.7 trillion through quantitative tightening, which withdraws a third source of demand.

Hence three independent buyers of duration are stepping back at once, none of them coordinated with the others and none of them responding to the others. That is precisely why the combination is easy to miss: each withdrawal has a sensible local explanation, and no single one of them looks systemic.

What would falsify this

The repatriation thesis makes a testable prediction: Japanese holdings of foreign bonds should keep falling, and the fall should be largest at the long end. If Japanese institutions resume buying foreign duration while domestic yields stay near 4%, the mechanism described here is wrong and something else explains the January move.

Furthermore, a single quarter of $29.6 billion in net US debt sales is a thin basis for a structural claim. Japanese investors have sold before and returned, often within two quarters, and quarterly flow data is noisy enough that the direction matters more than the magnitude. The honest position is that one quarter is consistent with the thesis and does not establish it.

There is also a straightforward alternative reading of January. A snap election creates uncertainty about fiscal policy, and uncertainty widens term premia temporarily. On that account the move was political rather than structural, and the yield should compress once the election result is known and the package is scored. The retreat from 4.24% to 3.90% by July is weak evidence for this view, though it is equally consistent with a market that overshot and then settled at a permanently higher level.

The Indian case, briefly

The clearest illustration of the transmission is a market with no obvious domestic problem. Indian equities saw foreign portfolio outflows of roughly ₹2.29 lakh crore in the first seven months of 2026, more than the ₹1.66 lakh crore withdrawn across all of 2025, and the rupee reached a record low of 96.84 against the dollar on 20 May.

India did not change its policy rate framework, lose fiscal control or suffer a banking problem in that window; it was simply on the receiving end of a carry unwind and a higher oil import bill at the same time. That combination is what global duration repricing looks like from the outside.

Conclusion

The Bank of Japan has spent three decades making Japanese long bonds uninvestable, and in doing so it exported a large and reliable bid for everyone else’s. January 2026 was the moment that arrangement started to unwind, and it unwound at the maturity where the effect on global portfolios is largest.

The 40-year at 3.90% against a policy rate of 0.75% tells you the market is no longer pricing Japanese duration off the central bank; it is pricing it off the budget. Furthermore, this is the third large withdrawal of demand for long-dated assets in twelve months, alongside Gulf repatriation and Federal Reserve balance sheet reduction, and the three are unrelated in cause and additive in effect.

Investors accustomed to reading a rise in their own long yields as a verdict on their own fiscal position should consider that in 2026 it may simply be a Japanese insurer buying at home.