Everyone is watching the yield level. The yield level is not the mechanism that matters.

Japan's 10-year government bond yield has been trading near 2.9%, its highest since the late 1990s, and the 30-year has pushed past levels never recorded since that maturity existed. The Bank of Japan is widely expected to hike its policy rate to 1.25% at its September 18 meeting, even after Japan's revised second-quarter growth came in below consensus. That combination, a central bank tightening into a weak print, is the kind of thing that normally gets read as a policy mistake. It is being read instead as confirmation that Japan's decades-long export of capital is starting to reverse.

By the numbers
2.9%
10-year JGB yield, near 30-year highs as of Sept 8, 2026
$29.6B
US Treasuries sold by Japanese investors in Q1 2026
1.25%
BOJ policy rate target after the expected September 18 hike

The fact pattern is real. Japanese investors sold $29.6 billion of US Treasuries in the first quarter of 2026 alone, removing a buyer that has been one of the most dependable sources of demand for US government debt for a generation. At the same time, an unusual management meeting at Japan's Government Pension Investment Fund, a roughly two trillion dollar pool, has fueled speculation that it could raise its target allocation to domestic bonds. That is not a trading flow. That is a strategic reweighting of one of the largest asset pools on earth.

The mechanism nobody is naming correctly

The consensus explanation is simple: JGB yields are rising, so the yield gap that sent Japanese money abroad for years is closing, so capital comes home. That story is true but incomplete, and the missing piece is the one that actually drives the timing.

Japanese investors do not buy foreign bonds unhedged in scale. They buy them and hedge the currency exposure back to yen, and the return that matters to them is the hedged return, not the headline US yield. Hedging cost is a function of the short-term rate differential between the dollar and the yen. Every time the Bank of Japan raises its policy rate, it does not just make JGBs more attractive in isolation. It simultaneously raises the cost of hedging a dollar bond back into yen, which mechanically depresses the hedged return on that Treasury holding, regardless of where the US 10-year sits. Hedged Treasury returns for Japanese buyers have already spent stretches near zero or negative. A rate hike moves that number further into negative territory automatically, on the same day, independent of any US data.

That is the part the market is underpricing. This is not a story about a yield gap narrowing gradually. It is a story where every BOJ decision has a direct, mechanical, same-day effect on the profitability of holding US paper for the single largest foreign holder cohort. The repricing does not require a change in conviction. It requires nothing more than the BOJ doing what it has already signaled it will do.

The second-order effect being ignored

The genuine second-order risk is not a one-time capital flow. It is the reflexive loop between the yen carry trade unwind and US risk asset liquidity. Positions funded by borrowing cheap yen to buy higher-yielding assets globally, equities included, become less economic every time the BOJ tightens. When those positions get unwound, the selling is not confined to bonds. It shows up as dollar-yen volatility, and dollar-yen volatility has a documented habit of forcing deleveraging across unrelated risk assets that happen to share the same funding currency. The market treating this as a Treasury demand story misses that the transmission channel runs through funding markets first, and Treasuries are just where the pressure eventually surfaces.

A market-neutral book has to treat this as a volatility and correlation event before it treats it as a yield event. That is how I am framing it at Zentra Asset Management right now: less about direction in rates, more about the second-order convexity in dollar-yen and its spillover into cross-asset correlation.

The honest counter-case

The rush has not happened yet, and it may not happen the way the narrative assumes. Japanese life insurers have largely fulfilled long-standing regulatory purchasing obligations in domestic long-dated bonds, which changes the marginal buyer profile without necessarily forcing existing foreign holdings to be sold. Separate data over a trailing twelve-month window has still shown net Japanese purchases of foreign bonds, not net selling, meaning the shift so far looks selective rather than systemic. If Japanese growth data continues disappointing, the BOJ could slow its hiking path, which would remove the mechanical pressure on hedging costs entirely. I would be wrong if the BOJ blinks.

What I am watching next is simpler than the headline: the pace of GPIF's allocation review, and whether hedged Treasury returns for Japanese buyers stabilize or keep grinding lower after September 18. If they keep grinding lower, the flow follows the mechanism, not the narrative. Where is your own exposure actually funded, and in what currency?

Common questions

Why are Japanese government bond yields rising so much in 2026

The Bank of Japan has been stepping back from the bond purchases and yield curve control that kept rates suppressed for years, while inflation has stayed above target and the government has pursued more expansionary fiscal policy. The 10-year JGB yield has traded near its highest levels since the late 1990s, and the 30-year yield has breached levels never seen since that maturity was introduced.

What is Japanese capital repatriation risk

It refers to the possibility that Japanese institutions such as life insurers, banks and pension funds, which for decades bought foreign bonds because domestic yields were too low, start bringing that capital home now that JGB yields are more competitive. Because Japan holds a large share of foreign government debt including US Treasuries, a shift in that behavior can affect global bond demand.

Did Japan actually sell US Treasuries in 2026

Yes. Reporting has cited Japanese investors selling $29.6 billion of US Treasuries in the first quarter of 2026 alone, though separate data over the trailing 12 months has shown Japanese investors still net purchasing foreign bonds overall, suggesting the shift is uneven rather than a uniform exodus.

What is the yen carry trade and why does it matter here

The yen carry trade involves borrowing in low-yielding yen to fund purchases of higher-yielding assets elsewhere. As the Bank of Japan raises rates, the cost of that borrowing rises and the return advantage shrinks, which can force unwinding of those positions and create volatility in currency and risk asset markets.

Is the Bank of Japan expected to raise rates again soon

Markets have been pricing a Bank of Japan rate hike to around 1.25% at its September 18, 2026 meeting, with commentary describing the central bank as raising rates even as growth data has come in below consensus.

Is GPIF changing its asset allocation because of this

There has been speculation, following an unusual management meeting at Japan's Government Pension Investment Fund, that the roughly two trillion dollar fund could raise its target allocation to domestic bonds, though no formal revision to its medium-term objectives had been confirmed at the time of writing.

Article sourced from Bloomberg: Japan’s Rising Yields Stir Debate Over Growing Repatriation Risk. The commentary above is original analysis by Komey Tetteh.

Research that reaches you before the narrative does

One note each week connecting a global event to positioning in US markets. No paywall.

Subscribe free