Japan's three megabanks reported results for the quarter ended 30 June 2026 - and all posted sharp jumps.
But before the numbers, a correction to a common assumption: they did not report on the same day.
| Bank | Report date |
|---|---|
| Mizuho | 30 July |
| Sumitomo Mitsui | 31 July |
| MUFG | 3 August |
A three-day spread across two weeks.
What was reported
| Net profit | Change | Net interest income | Change | |
|---|---|---|---|---|
| MUFG | 809.4 | +48.2% | 882.4 | +27.7% |
| Sumitomo Mitsui | 501.4 | +33.0% | 791.9 | +26.4% |
| Mizuho | 422.9 | +45.5% | 364.2 | +20.8% |
In billions of yen
MUFG posted a first-quarter record, and Mizuho showed the sharpest percentage gain in ordinary profit - up 62.5%.
A note on Mizuho
Mizuho's net interest income looks materially lower than the other two, and that is structural rather than operational. Its large markets business books funding cost as interest expense while the offsetting revenue sits in trading income. A direct comparison of the three figures is not fully valid.
The mechanism: a widening spread
Here are MUFG's numbers, and they explain the entire phenomenon:
| Quarter | Year ago | Change | |
|---|---|---|---|
| Lending rate | 1.47% | 1.13% | +34bp |
| Deposit rate | 0.30% | 0.17% | +13bp |
| Spread | 1.16% | 0.95% | +21bp |
The bank charges borrowers 34 basis points more, and pays savers 13 basis points more. It passes on roughly 38% of the increase. The difference stays with the bank.
Exactly the same pattern at Sumitomo Mitsui: the lending rate rose to 1.61% from 1.26%, deposits to 0.30% from 0.18%, and the spread widened from 1.08% to 1.31%.
And at Mizuho the sharpest figure: interest income rose 5.1% while interest expenses were essentially flat - up just 0.8%.
The most important point: this has barely started
Here sits the figure that changes how all three reports should be read.
The Bank of Japan raised its rate to 1.0% on 16 June 2026 - the highest since September 1995. The quarter ended on 30 June, just two weeks later.
And Sumitomo Mitsui quantifies it explicitly
The bank discloses that its annual plan assumed a 0.75% policy rate while the actual is 1.0%. And it quantifies: every 25 basis points of policy rate is worth about ¥110 billion.
And of roughly ¥80 billion of annual benefit from the June rise, this quarter captured only about ¥30 billion.
Which means the profits seen here mostly reflect earlier hikes, not June's. The bulk of the effect is still ahead.
Behind that sensitivity sits a simple structure: Sumitomo Mitsui holds ¥51.0 trillion in its current account at the Bank of Japan. Those balances are remunerated at the policy rate, so each hike converts almost directly into revenue.
And who dared raise guidance
Only one of the three.
- MUFG: did not raise. Guidance held at ¥2,700 billion, 30% progress
- Sumitomo Mitsui: did not raise. Held at ¥1,700 billion, 29% progress
- Mizuho: raised to ¥1,400 billion from ¥1,300 billion - an addition of ¥100 billion
And Mizuho's CFO, Makoto Samejima, gave the reason explicitly: the revision was decided "based on solid earnings progress and the BOJ's policy rate hike last month in June".
Mizuho also doubled its buyback programme to ¥200 billion.
And the other side that is barely discussed
Rising long yields are not only good news for banks.
Sumitomo Mitsui booked a ¥19.8 billion loss on its bond portfolio - a ¥34.1 billion deterioration against a year ago. When yields rise, existing bond prices fall - which hurts the investment book even as it helps the lending book.
And Mizuho's response to this is worth noting: its portfolio duration stands at just 0.8 years. The CFO explained that the increase in bond holdings was made "within normal ALM management, with a focus on short-term instruments".
And what that means
We wrote in the weekly review that the Japanese 30-year bond yields 3.975%. Japanese banks' answer to that yield is not to buy it.
Mizuho holds a portfolio at 0.8 years duration - meaning it is not touching the long end. A bank that believes yields will keep rising does not lock them in today. That is a statement about expectations, not just about risk management.
Further figures
Credit quality is improving: Sumitomo Mitsui's non-performing loan ratio fell to 0.81% from 0.97%.
Capital returns: Sumitomo Mitsui completed a ¥180 billion buyback and will cancel 28 million shares on 20 August, alongside a two-for-one share split. Mizuho doubled its buyback to ¥200 billion and extended it to 30 September.
And the monetary context: the Bank of Japan held the rate at 1.0% at its 31 July meeting, on an 8-1 vote - where the sole dissenter, Hajime Takata, wanted to raise it to 1.25%.
The full map of Japan's earnings week: The Week Japan Reported
My Angle
A personal opinion of Ilan Abramov - not advice, not a recommendation
This is one of the easiest reports to misread, and precisely in the optimistic direction.
The headlines say: three banks, three jumps, plus 48%, plus 45%, plus 33%. It is easy to conclude the monetary reversal is already priced.
It is not. The Bank of Japan raised on 16 June. The quarter ended on 30 June. Two weeks. And Sumitomo Mitsui says it in numbers: of ¥80 billion of annual benefit, this quarter captured only ¥30 billion.
Which means what we saw here is mostly the 2024 and 2025 hikes maturing. June's hike shows up in the coming quarters.
And what caught me hardest is a technical detail: Mizuho holds a bond portfolio at 0.8 years duration. While the Japanese 30-year yields 3.975%, a major Japanese bank chooses not to touch it.
That is a statement. Someone who thinks yields have peaked locks them in. Someone staying short thinks they will keep rising. And if Japanese banks are right, that reaches far beyond Japan - they are among the world's largest bond holders, including of US Treasuries.
