Japan & Korea Allocator Intelligence · Lead Signal
Japan's 3% Bond Yield Resets the Hurdle for Every GP Pitch
Japan's benchmark 10-year government bond (JGB — the bond Tokyo issues to fund itself, and the safest yen asset a Japanese institution can hold) touched 3.00% on 1 September 2026, its highest since 1996.
A Japanese insurer buys foreign credit hedged, for the spread over yen rates. So a 3% JGB does not ask "does America still pay more than Japan?" It asks a question about spreads, not rates. Does your credit spread — after the hedge, fees and a higher capital charge — beat the 1.4–2.4 points of risk-free term premium Tokyo now pays on its own bonds? For senior direct lending the answer is still yes, but the margin has gone from comfortable to case-by-case.
Japan's benchmark 10-year government bond (JGB — the bond Tokyo issues to fund itself, and the safest yen asset a Japanese institution can hold) touched 3.00% on 1 September 2026, its highest since 1996. The Ministry of Finance's closing reference yield printed 2.987% that day and 3.006% on 2 September, before easing to 2.935% by 7 September. The 30-year closed the week at 4.01%. Set those against yen cash — three-month TIBOR, the Tokyo interbank rate, at 1.56% — and the JGB now pays a term premium of about 1.4 points at ten years and about 2.4 at thirty. That premium comes with no credit risk, no currency risk and no manager to select.
Start from how Japanese institutions actually buy foreign credit. A life insurer does not take the dollar unhedged and pocket the higher American interest rate. It hedges: it sells the dollars forward for yen, rolling one- to three-month contracts. Under covered interest parity the forward price hands back roughly the gap between short-term dollar and yen rates, so the hedge costs about that gap. The dollar interest rate never reaches the insurer. What survives is the credit spread — the part of the coupon paid for taking credit and illiquidity risk. For floating-rate private credit the algebra is exact: a loan paying SOFR plus a spread, hedged into yen, yields the yen short rate plus that spread, less the basis.
That is why "Japanese rates were zero, so anything abroad beat zero" was never the reason the money went overseas. Japan's own bond market offers almost no credit spread; foreign markets offer a great deal. That spread is still there. What has changed is the other side of the comparison — the JGB — and the regulator's price for holding each asset.
The hedge cost itself, two ways. Three-month US Treasury bills yielded 3.77% on 4 September against three-month TIBOR at 1.56%: a gap of 2.2 points a year, the rate the forward market actually prices. Overnight rates give a wider gap: the 90-day average of SOFR, 3.65%, against the Bank of Japan's 1.0% policy rate is 2.65 points. The difference is that yen money-market rates already price the BOJ's next hike. Add the cross-currency basis, a premium yen holders pay above the rate gap for hedged dollars: 15–20 basis points at the start of 2026, 35 at the April 2025 peak. Call the all-in cost 2.4–2.8 points a year. It is what the insurer pays to make the dollar disappear.
Now do the arithmetic as spread against spread, with 1 September 2021 as the sub-1% comparison point.
| 1 September 2021 | 4 September 2026 | |
|---|---|---|
| Three-month yen rate (TIBOR) | 0.06% | 1.56% |
| Three-month dollar rate (SOFR / T-bill) | 0.05% | 3.65% / 3.77% |
| Hedge cost from the rate gap | ~0 | 2.2–2.65 points |
| 10-year JGB | 0.04% | 2.91% |
| 30-year JGB | 0.65% | 3.97% |
| JGB term premium over yen cash (10y / 30y) | 0.0 / 0.6 points | 1.4 / 2.4 points |
| Hedged gross yield on a loan paying 5.3–8.3 points over the dollar rate* | ~5.3–8.3% | ~6.7–9.7% |
| Excess over the 30-year JGB, before fees | ~4.7–7.7 points | ~2.7–5.7 points |
*Senior direct lending prices at roughly 9–12% gross in dollars today, which is 5.3–8.3 points over SOFR; the table holds that spread constant across both dates. The 2026 column deducts 0.2 points of basis; the 2021 column deducts none, which flatters it slightly. Industry pricing, not a quoted rate.
Read the last row. A hedged Japanese investor never earned the dollar rate, so the fall in JGB-relative margin is not "five points of vanished rate gap". It is the two points by which the JGB's own term premium has widened. Take off an illustrative 1.5–2.5 points of management fee and carry, and the net excess over a 30-year JGB is now roughly 0.2–4.2 points. Industry practice asks 1.5–3.0 points of extra return for locking capital up in a private fund. The top of the direct-lending range clears that comfortably. The bottom no longer does. That is the pitch problem: not "the JGB beats you", but "your low case no longer pays for its illiquidity".
The regulator sets the second price. Japan's economic-value solvency regime for insurers (J-ICS, in force since 31 March 2026) replaced the 200%-floor book-value solvency margin with a 100%-floor ratio that marks assets and liabilities to market. Three of its rules drive allocation. First, currency risk is a factor applied to the net open position, 30% for dollars against yen. A hedged holding carries almost no currency charge; an unhedged one carries a heavy one. The regime itself pushes insurers to hedge.
Second, spread risk is a stress on the spread itself, capped at plus 1.5 points for any asset spread above 2 points. A floating-rate loan with three to four years of spread duration loses roughly 4.5–6% in that stress. But a fund the insurer cannot look through can fall into "other equities", where the shock is 49% (infrastructure equity in developed markets: 27%). How the vehicle is classified can be worth more than a point of yield.
Third, interest-rate risk. Japanese lifers' liabilities run longer than their assets, so long JGBs shrink the mismatch and the capital it consumes. Dai-ichi's own disclosure on the new basis (March 2023 figures) showed a 50-basis-point fall in yen rates cutting its group ESR by 19 points. A floating-rate private-credit fund has almost no duration and does nothing for that match. A 30-year JGB at 4% does two jobs at once; the fund does one.
The insurers describe the same trade-off themselves. Daiwa's April survey of the ten lifers' FY2026 plans found hedging costs had "improved significantly". Even so, hedged foreign bonds' "investment appeal wanes when compared to JGBs, whose yields have risen significantly". Lifers are shifting "from low-yield JGBs to high-yield JGBs". Where they do add hedged assets, they prefer "floating-rate assets that are resilient in the face of hedging cost fluctuations". For unhedged foreign bonds, "none plan to increase them". The BOJ's April Financial System Report records the same caution on unhedged bonds, with foreign-bond holdings "more or less unchanged".
Nippon Life is the whole argument in one balance sheet. It says it is open to becoming a net JGB buyer again from the fiscal year starting April 2027, after two years out nursing bond losses. In June it committed ¥1.5tn ($9.4bn) to a five-year Blackstone private-credit and real-estate mandate. The JGB book buys duration and term premium at zero credit risk; the Blackstone mandate buys spread. Both books get bigger at the same time.
GPIF is the exception that proves the rule. The world's largest pension fund, with ¥320.4tn (about $2.0tn) of assets at end-June 2026, holds its foreign bonds unhedged by policy. Its benchmark is the FTSE World Government Bond Index ex-Japan ex-China, "unhedged, yen basis". Its principles state that "currency-hedged foreign bonds should be positioned as domestic bonds". Its foreign-bond book is therefore a currency position, not a spread trade.
GPIF's 25/25/25/25 target has stood since April 2025. Domestic bonds were 25.59% of the portfolio at end-June 2026 by GPIF's own count, after 26.9% in March. Société Générale and Goldman Sachs estimated $76–80bn of further JGB room within the band without a policy change. Bloomberg reported seven straight quarterly losses on the domestic-bond book through Q2 2026, and GPIF then held its first August board meeting since 2019. Nothing is decided. If the domestic-bond target does rise, the cut falls on foreign bonds or equities. And any hedged alternative pitched to GPIF is, by its own rulebook, competing for the domestic-bond bucket.
The calendar moves both sides of the comparison. Ministries' FY2027 budget requests hit a record ¥143.07tn, and more issuance at higher assumed rates pushes the JGB term premium up. Governor Ueda says the BOJ will debate a hike "including in September", with a decision due 17–18 September. A hike lifts yen short rates, which makes the hedge cheaper, and lifts JGB yields, which raises the bar. For a floating-rate loan hedged into yen the two roughly offset. So the meeting matters less to the arithmetic than to the mood of the committee approving the commitment.
Korea, the same week, has a different problem: not a rising price, but a stalled decision process. NPS has gone more than a month past its expected date for naming a chief investment officer. KIC is still without a permanent one after its July search collapsed. Neither empty seat stops diligence, but both stop final sign-off. The concrete opening is elsewhere — Korean bank bad loans at an eight-year high and a shrinking SME loan book — and is covered in the Korea section below.
Which GPs Have Opportunities:
| GP profile | Why this fits | What they'd need |
|---|---|---|
| Senior secured and asset-based lenders earning 650bp or more over SOFR (the top half of today's 9–12% gross range) | Hedged into yen that spread, less 20bp of basis and about 2 points of fees, still leaves 2–4 points over a 30-year JGB — enough to pay the 1.5–3.0-point illiquidity premium | One page showing yen short rate + spread − basis − fees against the 10- and 30-year JGB, dated, with the rates used |
| GPs that can offer a yen-hedged share class or a Japanese unit-trust wrapper with asset-level look-through | Hedging removes J-ICS's 30% dollar currency factor and takes the roll and margin burden off the insurer; look-through keeps the fund out of the 49% "other equities" bucket | A hedged class that bears its own basis and roll costs, monthly NAV, and reporting an insurer's actuaries can map to the spread-risk stress |
| Multi-line platforms that can run a spread book beside a duration solution, on the Nippon Life–Blackstone pattern | The lifer is buying duration from JGBs and spread from private credit at the same time; a partner that helps with both is the one that gets the five-year framework | A framework proposal, not a single fund, stating each vehicle's currency, duration and spread duration, with a secondment offer |
Sources: note 1.
Sources
[1] Japan's 3% Bond Yield Resets the Hurdle for Every GP Pitch.
- Japan's benchmark bond yield rises to 3% for first time in 30 years — Reuters via Yahoo Finance
- Japan benchmark bond yield hits 30-year high of 3% amid global debt selloff — Nikkei Asia
- GPIF would be justified in buying more Japan bonds, say analysts — Bloomberg
- Nippon Life to invest $9.4bn in Blackstone private-credit deal — Nikkei Asia
- NPS CIO selection drags past a month as cabinet reshuffle crowds the vetting queue — 파이낸셜뉴스
- Korea bank bad loans hit ₩18.9tn, an eight-year high on corporate credit — Businesskorea
- Interest rate (JGB reference yields), September 2026 — Ministry of Finance, CSV — 10-year 2.987% (1 Sep), 3.006% (2 Sep), 2.910% (4 Sep), 2.935% (7 Sep); 30-year 3.965% (4 Sep), 4.009% (7 Sep)
- Interest rate, historical series 1974– — Ministry of Finance, CSV — 1 Sep 2021: 10-year 0.035%, 30-year 0.654%
- JBA Japanese Yen TIBOR, 2026 daily file — JBA TIBOR Administration — 3-month 1.55991%, 31 Aug 2026
- JBA Japanese Yen TIBOR, 2021 daily file — JBA TIBOR Administration — 3-month 0.05727%, 1 Sep 2021
- SOFR Averages and Index — Federal Reserve Bank of New York — 90-day average 3.64571% (4 Sep 2026); 0.04378% (1 Sep 2021), via the Markets Data API
- 3-Month Treasury Bill Secondary Market Rate (DTB3) — FRED — 3.77%, 4 Sep 2026
- Statement on Monetary Policy, 31 July 2026 — Bank of Japan — overnight call rate "at around 1.0 percent", 8-1 vote
- FOMC statement, 29 July 2026 — Federal Reserve — target range 3.50–3.75%, 9-3 vote
- Are foreign investors more cautious over the US dollar? — Banque de France, 13 April 2026 — 3-month hedging premium for a Japanese investor: 35bp on 9 Apr 2025, 15–20bp at the start of 2026
- ESR for life insurance companies under new Japanese solvency regulation — Milliman, February 2026 (PDF) — 100% ESR minimum replacing 200% SMR; currency factor USD/JPY 30%; spread stress +max(0.4%, min(1.5%, 75%×spread)); equity level stress: developed listed 35%, developed infrastructure 27%, other equities 49%
- Prudential insurance regulation in Japan — Skadden, April 2025 (updated June 2026) — ESR effective 31 March 2026; 100% minimum; three intervention categories
- Response to new economic value-based regulation — Dai-ichi Life Holdings (PDF) — new-basis ESR 212% at March 2023; sensitivity −19 points to a 50bp fall in yen rates, +4 to a rise
- Life insurers' investment plans for FY26 — Daiwa Securities, 28 April 2026 (PDF) — hedged foreign bonds' "investment appeal wanes when compared to JGBs"; floating-rate preference; unhedged: "none plan to increase them"
- Financial System Report, April 2026 — Bank of Japan (PDF) — lifers "cautious ... unhedged foreign bonds"; foreign bondholdings "more or less unchanged"
- Investment results for the first quarter of fiscal 2026 — GPIF, 7 August 2026 (PDF) — total ¥320,373.2bn at end-June 2026; domestic bonds 25.59%, foreign bonds 24.60%, domestic equities 24.48%, foreign equities 25.33%
- Policy asset mix for the fifth medium-term objectives period — GPIF (PDF) — 25/25/25/25 from 1 April 2025; foreign-bond benchmark "unhedged, yen basis"; "currency-hedged foreign bonds should be positioned as domestic bonds"
Disclaimer: Japan & Korea Allocator Intelligence is published for general information and education only. It is not investment advice, and it is not an offer, solicitation or recommendation to buy, sell or hold any financial instrument, nor an invitation to engage in any investment activity. Nothing in it takes account of any reader's circumstances, objectives or needs. Figures and sources are believed accurate at the dates shown, but completeness and continuing accuracy are not guaranteed. Take your own professional advice before acting on anything here.