Mizuho Raised the Guide Its Peers Would Not Touch and Doubled the Buyback, on a Quarter the Markets Book Half Explains

A note on sourcing. Mizuho publishes a full record of an earnings call, prepared remarks and analyst Q&A, only for its interim and full-year results in November and May. For the first and third quarters it holds a shorter net conference and publishes, in place of a transcript, a document titled "FY26 Q1 Results Call (Speaker: Makoto Samejima, Group CFO)" that reproduces the CFO's own remarks keyed to the pages of the deck he presented. That document, the Tanshin and its Selected Financial Information supplement, the July 30 buyback amendment, the August 3 repurchase-progress notice, the August 14 capital release and the fixed-income investor update are the basis for this analysis. No analyst Q&A was published for the July 30 call and no webcast was posted, so the absence of an Analyst Q&A section below reflects the absence of a public record, not an omission on our part. The next opportunity to hear management questioned is the interim results conference in November.

Results vs. Consensus

Mizuho reports under Japanese GAAP on a cumulative basis and gives no quarterly guidance, so a June quarter has to be measured against three yardsticks: the prior-year quarter, progress toward the company's own full-year outlook, and whatever quarterly poll exists. The Street models this name at the full-year level, and no Bloomberg-compiled or QUICK quarterly figure for the June quarter appeared in the post-print coverage; the three-contributor poll below is recorded because it brackets the quarter, not because it carries weight. On all three yardsticks the quarter cleared comfortably.

Q1 scorecard (three months ended June 30, 2026)

The prior-year expense ratio of 59.9% above is derived from the deck's own gross-profit and expense lines and their stated year-on-year changes; the company reports the improvement as 12.0 points from its unrounded internals.

Two conventions that matter before any of these numbers are used. First, Mizuho truncates rather than rounds its own percentage changes, so a figure recomputed from the yen amounts can differ from the filed percentage by a tenth of a point. The guide revision is the live example: it is +7.7% on the yen amounts and +7.6% as filed. Where the two differ below, both are given. Second, and more consequentially, two bases run through the disclosure and must never be mixed. The Tanshin's "consolidated gross profits" of ¥1,026.7bn and "consolidated net business profits" of ¥532.4bn exclude net gains on exchange-traded funds; the investor deck's identically named lines, ¥1,070.1bn and ¥575.8bn, include them. The difference is ¥43.3bn this quarter against ¥4.4bn a year ago. Wherever the deck basis is used below it is labelled.

Year-over-year comparison (¥bn, consolidated, excludes ETF gains)

Sequential context (¥M, as reported quarterly)

Japanese banks report cumulatively, so a standalone quarter is a subtraction. The series below is the reported quarterly progression, and it is worth having in view because the March quarter that immediately preceded this one was the weakest of the last five.

The sequential jump is arresting: ordinary profits nearly doubled from the March quarter and net profit rose 84.9%. Some of that is the March quarter's own weakness, which carried the bond-portfolio rebalancing and the year-end true-ups that Japanese banks habitually push into the fourth quarter. The rest is genuine, and the split between the two is the subject of the next section.

Where the ¥259.3bn increase in net business profits came from

Quality of the beat

Revenue. Consolidated gross profits rose 34.3% to ¥1,026.7bn, and every line contributed. Net interest income at ¥364.1bn was the largest single increase in yen terms, up ¥62.6bn, and it is the most repeatable of them: the domestic loan-and-deposit rate margin at the two banks widened to 1.26% from 1.10% for the whole of last year, and the loan book grew ¥3.3tn in three months. Net trading income of ¥276.6bn and net other operating income of ¥102.7bn together added ¥151.3bn, and those are the two lines a bank cannot promise to repeat. The distinction is not academic: management's own outlook implicitly agrees with it, which is why the raised full-year guide asks for so little from the remaining three quarters.

Costs and credit. The expense ratio of 47.7% is the lowest Mizuho has printed in the six years shown in its own historical chart, where the range runs from 62.5% to 64.6% and last year's full-year figure was 59.4%. It is also, on the CFO's own account, not a number to plan around. General and administrative expenses grew 9.8% year on year while gross profits grew 34.3%, and the gap is almost entirely revenue timing rather than cost discipline: a weaker yen and inflation pushed the cost line up, and the revenue line simply outran it. Credit-related costs were a ¥6.1bn charge against an ¥11.4bn reversal a year ago, a ¥17.6bn deterioration in the comparison that is entirely the prior-year reversal disappearing. At 5.5% of the ¥110.0bn full-year plan after one quarter of four, credit is not a story yet, and the forward-looking reserve balance barely moved, from ¥162.1bn to ¥165.2bn.

EPS and tax. Basic EPS of ¥173.53 rose 49.7%, faster than the 45.5% growth in net profit, because the average share count fell 2.8% after the group cancelled 47,016,600 shares during the quarter. Working the other way, the effective tax rate normalised from 23.0% to 28.4%, and that single line is why a 62.5% increase in ordinary profits became a 45.5% increase in net profit. Held at last year's 23.0% rate, profit attributable to owners of parent would have been roughly ¥455bn rather than ¥422.9bn. Last year's full-year effective rate of 22.7% was itself depressed by a ¥65.3bn deferred-tax benefit, so 28.4% is the normalisation arriving, not an aberration, and the full-year guide embeds it: ¥1,400.0bn of net profit on ¥1,980.0bn of ordinary profits is a 70.7% conversion.

Segment Performance

Mizuho runs five in-house companies. The figures below are the company's management-accounting presentation on a group-aggregate basis, with one property that materially changes how they should be read: segment income and expenses of foreign branches and foreign subsidiaries are translated at budgeted exchange rates, and the prior-year comparatives were re-translated at this year's budgeted rate. The segment table is therefore currency-neutral. The consolidated income statement above is not, and the yen was 12.2% weaker against the dollar at June 2026 than at June 2025. The prior year has also been restated for a change in the general-and-administrative expense allocation method made in April 2026.

Expense ratios by segment, both years on the same restated basis, are the second half of the picture and in two cases the more informative half.

Retail & Business Banking: the inflection we said the re-rating depended on

RBC is the segment our initiation put at the centre of the bear case. Last year it produced the largest absolute gross-profit increase of any customer company and still delivered lower net income than the year before, at a 5.9% return on the 19.3% of group risk capital it consumes, with an expense ratio of 76.8%. This quarter it converted. Gross profits rose ¥44.4bn to ¥258.5bn while expenses rose ¥7.4bn, so the entire revenue increase fell through: net business profits more than doubled to ¥82.2bn and net income rose 67% to ¥47.8bn. The expense ratio went from 81.4% to 70.3% in a single year.

The composition is credible. Non-interest income at RBC reached ¥119.0bn, up ¥14.3bn, with individual wealth management at ¥40.0bn (up ¥6.5bn) and the solutions business at ¥28.7bn (up ¥3.8bn); total assets in custody grew ¥5.5tn to ¥71.2tn. The rest is the rate cycle arriving in a mass-retail and small-business book that has been waiting for it for two decades. Domestic corporate loans inside RBC grew, individual loans shrank slightly, and the RBC corporate loan spread widened to 0.67% from 0.64% for the whole of last year.

"In the Customer Groups, RBC and CIBC led performance, with Net Business Profits increasing by JPY 44.4B and JPY 54.3B YoY, respectively. In addition to the effect of rate hikes, fee businesses such as investment banking and real estate also grew, driven by strong financing demand and active corporate actions."
— Makoto Samejima, Group CFO

Assessment: this is the single most important thing in the print. Our May note said the bridge to a higher multiple ran through RBC moving from 5.9% toward a cost-of-capital return, and one quarter of 67% net-income growth on an eleven-point improvement in the expense ratio is that bridge being built. The caveats are real: RBC carried ¥16.0bn of the group's ¥6.1bn net credit charge, so its credit line is still the segment's swing factor, and Mizuho does not publish segment risk capital or segment return on equity at a first quarter, so we cannot yet say what return this profit was earned on. But the direction is unambiguous and it is the direction that matters.

Corporate & Investment Banking: still the best business in the group

CIBC grew gross profits ¥57.1bn to ¥211.2bn against a ¥4.7bn increase in expenses, taking net business profits up 55% and net income up 37% to ¥127.3bn, more than half the entire Customer Groups total. Its expense ratio of 30.1% is the lowest in the group by a wide margin. The engine is the domestic corporate solutions business, which grew from ¥62.1bn to ¥94.3bn, and CIBC's own solutions line rose ¥28.2bn to ¥67.0bn.

One number inside that deserves attention. Of the ¥32.2bn increase in the domestic corporate solutions business, ¥28.5bn is credit-related fees, which rose to ¥64.3bn. Investment banking fees contributed ¥2.5bn of the increase and real estate ¥1.2bn. Credit-related fees are arrangement, agency and commitment income on lending, so they scale with the same loan book that is driving net interest income.

Assessment: CIBC is doing exactly what a corporate bank should do in a rising-rate, high-corporate-activity market, and it is why Mizuho is worth owning at all. The qualification belongs to our second bull pillar rather than to the segment: 88% of the growth in the group's flagship domestic fee business came from fees generated by lending. That is a good business. It is not the rate-independent engine the fee-income story is usually sold as, and if loan growth slows, this line slows with it.

Global Corporate & Investment Banking: the problem got worse

GCIBC grew gross profits ¥3.0bn, to ¥232.8bn, and grew expenses ¥16.4bn. Net business profits fell 12% and net income fell 34% to ¥63.7bn. The expense ratio deteriorated 6.4 points to 61.8%, the only segment in the group to move the wrong way on costs. Overseas loan balances did grow, with the Americas average balance up USD 3.2bn to USD 114.0bn and EMEA up USD 2.0bn, but the overseas loan-and-deposit rate margin narrowed to 3.17% from 3.44% and the overseas loan spread to 1.07% from 1.11%. Non-interest income at GCIBC rose only ¥5.7bn to ¥128.2bn, with investment-banking fees in the Americas up ¥6.4bn and credit-related fees down ¥2.1bn.

"GCIBC: Net Business Profits decreased YoY, mainly because the increase in Gross Profits was not enough to offset the increase in expenses. While investment banking business grew, net interest income from lending and deposit-taking businesses declined amid a lower interest rate environment, and certain products, such as leveraged finance, showed some softness due to market conditions."
— Makoto Samejima, Group CFO

Assessment: this is the half of our first bear point that did not improve, and it is the larger half. GCIBC consumed 26.5% of group risk capital last year against RBC's 19.3%, and it earned 8.0% on it. A quarter in which revenue is flat, costs rise ¥16.4bn and net income falls a third moves that number down, not up. Management's explanation is honest and mostly exogenous, which is the problem: a business whose economics depend on the dollar rate cycle running in Mizuho's favour is not a business the group controls. There was no cost action announced, no target and no date. Carry this one forward unresolved.

Global Markets: the quarter's largest single contributor, and its least repeatable

Global Markets produced ¥199.7bn of net business profits against ¥73.5bn a year ago, an increase of ¥126.2bn that is 48.7% of the group's entire improvement. The banking book, which runs the group's own interest-rate, credit and equity positions, went from ¥45.3bn to ¥144.4bn. Sales and trading, the client-facing half, rose 96% to ¥55.2bn on strength in equities and fixed income in and outside Japan; management attributed part of that to US equity trading capacity it has been building.

The number to hold on to is that ¥42.9bn of the ¥43.3bn of group-wide gains on exchange-traded funds sits inside Global Markets, against ¥3.2bn a year ago. Strip the ETF gains from both years and the segment's increase is roughly ¥86.5bn rather than ¥126.2bn. ETF gains are disposals from a legacy equity holding, marked and sold into a Nikkei that has risen more than a fifth this year. They are real cash and they are not a franchise.

"Markets: Banking was exceptionally strong, recording more than double the Net Business Profits YoY. Sales & Trading was also strong, with growth in both fixed income and equities, both in Japan and overseas."
— Makoto Samejima, Group CFO

Assessment: our May note observed that the entire year-on-year improvement in Global Markets in FY2025 was the banking book recovering from a loss. A year later the banking book has gone from recovery to windfall, and it is now supplying 38.2% of the group's profit growth on its own. Sales and trading nearly doubling is the part worth capitalising; the banking book tripling is the part that will mean-revert. Anyone extrapolating this quarter should discount roughly ¥100bn of it before starting.

Asset Management: small, and quietly going backwards on revenue

AMC produced ¥15.5bn of gross profits, down ¥2.0bn, and ¥2.2bn of net income. It held its net business profits up 15% only by cutting expenses ¥2.0bn. Non-interest income at AMC fell ¥1.8bn. Asset Management One, the group's asset-management joint venture, contributed ¥7.6bn of net business profits against ¥5.4bn a year ago.

Assessment: AMC is 1.4% of group gross profits and does not move the investment case either way, but a fee business shrinking its revenue in a quarter when domestic equity markets rose sharply and household assets in custody grew ¥5.5tn is a small warning. The retail asset-gathering growth is being captured in RBC, not here.

Group companies

Mizuho Bank on a non-consolidated basis reported ¥651.7bn of net business profits against ¥208.0bn, and ¥559.6bn of net profit against ¥187.2bn. Neither figure means what it appears to: the CFO disclosed that roughly ¥200bn of it is a one-off dividend from subsidiaries, and the consolidation removes ¥236.5bn of intercompany dividend income at the bank. The securities business, aggregating Mizuho Securities with the group's US and Europe-based entities, produced ¥90.0bn of net business profits against ¥51.0bn and ¥69.6bn of net income against ¥43.9bn. Equity in income from affiliates reached ¥21.0bn, of which ¥8.5bn is Vietcombank.

"Results by Group Company: BK (on a non-consolidated basis) shows a significant YoY increase, but this includes a one-off factor of approximately JPY 200.0B in dividends from subsidiaries. However, even after considering these factors, the performance trend is strong."
— Makoto Samejima, Group CFO

Assessment: flagging the ¥200bn intercompany dividend unprompted is the kind of disclosure discipline that earns a management team the benefit of the doubt elsewhere, and it is worth noting because it is not universal in this sector. The securities business reaching record earnings, helped by the Greenhill acquisition in the United States, is the clearest evidence that the overseas build is producing something even while GCIBC's lending economics deteriorate.

Key Topics & Management Commentary

Overall Management Tone: Confident, and unusually willing to argue against its own headline. The same remarks that announced a raised outlook and a doubled buyback also told investors that the 47.7% expense ratio is "somewhat too low," that the bank's own non-consolidated profit contains a ¥200bn one-off, and that the record earnings-per-share now in prospect is "merely a milestone." That is a marked change from May, when the posture was defensive: a guide that came in under the Street, a buyback a quarter the size of the prior year's, and a capital ratio that had gone backwards. Where management was less convincing was on the overseas corporate bank, where the explanation for a 34% fall in profit was framed entirely in terms of the external environment.

1. The guide went up in July, for the second year running

Raising a full-year forecast on one quarter of evidence is unusual for a Japanese bank, and both of Mizuho's closest peers declined to do it this season. It is not unusual for Mizuho. Last year the group opened with a ¥940.0bn forecast in May 2025, raised it to ¥1,020.0bn at the first-quarter results on July 31, raised it again to ¥1,130.0bn at the interim in November, and delivered ¥1,248.6bn. This is the same move at the same point in the calendar, and management framed it the same way.

"We have revised our FY2026 earnings outlook upward to Consolidated Net Business Profits of JPY 1,750.0B (+JPY 120.0B vs May) and Profit Attributable to Owners of the Parent of JPY 1,400.0B (+JPY 100.0B). This upward revision was decided based on solid earnings progress and the BOJ's policy rate hike last month in June. Barring a significant deterioration in the external environment, we are confident in achieving these targets."
— Makoto Samejima, Group CFO

The revision moved net profit up 7.7%, ordinary profits and net business profits up ¥120.0bn each, and left credit costs at ¥110.0bn and stock gains at ¥360.0bn untouched. Every yen of the increase therefore comes from the operating line. Both closest peers, reporting within four days on either side, changed nothing: Sumitomo Mitsui reiterated its outlook on July 31 and MUFG left its ¥2,700.0bn target in place on August 3.

Assessment: the third bear point in our May thesis was that Mizuho's guide came in 0.9% below the yen consensus, last of the three megabanks, and needed a normalised-basis defence four days later. Ninety days on, the revised guide sits 0.9% above the twelve-analyst average and the company is the only one of the three to have moved. The sign of that gap flipped because Mizuho moved, not because the Street did.

2. The buyback doubled, and the payout is still only forty percent

The board resolved on July 30 to expand the repurchase programme from ¥100.0bn to ¥200.0bn and from 25 million to 35 million shares, to extend the window from August 31 to September 30, and to move the cancellation date from September 24 to October 23. All repurchased shares will be cancelled. By July 31 the group had bought 9,147,800 shares for ¥71.4bn under the programme, including ¥31.8bn in July alone.

What the doubling does not do is close the gap to the stated policy. On the revised ¥1,400.0bn outlook, a ¥150 dividend across the current share count is roughly ¥365bn, which with the ¥200bn buyback gives a total payout near 40%. The policy floor is 50%, which would be ¥700bn. Roughly ¥135bn of capacity is therefore still uncommitted, and management said explicitly that it intends to revisit the question.

"In the course of future results announcements, we would like to consider additional returns."
— Kazuharu Sasaki, General Manager of Financial Planning (our rendering of the Japanese original)

Assessment: our second bear point was that capital-return intensity had stepped down from FY2025's 60% total payout to a ¥100bn opening tranche, with no decision date and no trigger. Doubling inside ten weeks answers the intent question. It does not answer the level question, and 40% against a 50% floor with a November checkpoint is the most identifiable catalyst in the name between now and March.

3. The capital ratio that carried the bear case repaired itself in one quarter

On the Basel III fully-effective basis excluding net unrealised gains on other securities, which is the measure management steers by and the one that fell from 10.3% to 9.9% over the course of last year, common equity tier 1 rose to 10.4% at June. On that basis capital grew ¥461.4bn, or 4.9%, while risk-weighted assets grew ¥428.1bn, or 0.5%. On the regulatory basis the ratio went from 13.16% to 13.51% and risk-weighted assets grew 1.6%, an annualised pace of about 6.5% against the 12.8% recorded for the whole of last year.

Management was careful not to declare the constraint over.

"Amid strong corporate financing demand, we will continue to monitor RWA trends and act flexibly regarding shareholder returns, balancing our business results and capital adequacy with growth investment opportunities."
— Makoto Samejima, Group CFO

Assessment: a fifty-basis-point recovery is a lot for one quarter, and it came from the numerator: capital rose 4.9% while risk-weighted asset growth slowed to a 0.5% drag. The honest qualification is that a quarter this profitable would repair almost any capital ratio, and that the loan book still grew ¥3.3tn while risk-weighted assets grew ¥1.3tn, which implies the growth is being taken in low-density assets. That is prudent, and it is also the kind of mix that can change quickly.

4. Net interest income, the domestic margin, and a rate rise that is barely in the numbers

Net interest income of ¥364.1bn rose ¥62.6bn, or 20.8%. The domestic loan-and-deposit rate margin at the two banks reached 1.26% against 1.10% for the whole of last year and 0.92% the year before; the return on loans went to 1.59% from 1.32% and the cost of deposits to 0.32% from 0.21%. At Mizuho Bank alone the quarterly margin was 1.27% against 1.04% a year ago. Interest on loans and bills discounted rose to ¥735.8bn from ¥647.2bn, and interest and dividends on securities to ¥262.9bn from ¥203.9bn, while interest on deposits rose to ¥438.9bn from ¥389.5bn.

The Bank of Japan raised its policy rate 25 basis points to 1.00% on June 16, the highest level since 1995, with fourteen days of the quarter remaining. Almost none of that hike is in these numbers, and the revised outlook assumes the policy rate stays at 1.00% for the rest of the year.

Assessment: this is the cleanest part of the print and the reason the first bull pillar stays on track. A margin that has widened 34 basis points against the FY2024 average, on a ¥103tn loan book funded by ¥175tn of deposits and negotiable certificates, is a large and mechanical earnings stream, and the June move is essentially unbanked. The offsetting fact is that the cost of deposits is rising faster in proportion than the return on loans: deposit costs are up 5.3 times from FY2024's 0.06% while loan yields are up 62%. Pass-through is running, and it is running earlier in this cycle than in the textbook.

5. Deposits fell while the loan book grew

Total deposits and negotiable certificates fell ¥2.1tn in the quarter to ¥175tn. Domestic yen deposits fell ¥2.5tn, of which corporate deposits fell ¥2.9tn and individual deposits rose ¥0.4tn. On the two-bank non-consolidated basis, domestic deposits fell ¥1.86tn while individual deposits rose ¥417bn.

"On the balance sheet (B/S), loan balance increased by JPY 3.3T compared to the end of March 2026, mainly driven by strong financing demand in Japan. While total deposits decreased by JPY 2.1T, this is mainly due to a seasonal decrease in corporate deposits and is not a particular concern. Retail deposits increased along with a rise in new account openings."
— Makoto Samejima, Group CFO

Assessment: the seasonality explanation for corporate deposits is plausible and standard, but the sequence is the point. A group that grew loans ¥3.3tn while its deposit base shrank ¥2.1tn is funding that gap somewhere other than core deposits. Retail deposit gathering was named as the top priority for the retail company at the May results with no target and no date attached; the deposit-rate move is the first concrete action taken in that direction, and it will show up in the cost of deposits from the second quarter.

6. Half the quarter is the markets book

The arithmetic bears repeating in one place. Of the ¥259.3bn increase in consolidated net business profits on the deck's presentation basis, ¥126.2bn came from Global Markets and ¥99.1bn of that from the banking book. Gains on exchange-traded funds contributed ¥38.9bn of the group increase, of which ¥39.7bn sits inside Global Markets. A further ¥46.4bn came from the "Others" line, which the company defines as items eliminated as internal transactions between segments, and which swung from a ¥6.1bn loss to a ¥40.3bn profit with no explanation offered.

Assessment: the four customer companies, which are what an investor is actually buying, grew net business profits ¥86.7bn on a currency-neutral basis, or 33.4% of the group increase. That is a very good number for a Japanese bank and it is not the number in the headline. Anyone building a forward model off ¥422.9bn of quarterly profit is capitalising a banking-book windfall, an ETF disposal programme and an unexplained elimination swing.

7. Credit costs at five percent of plan, and a forward-looking reserve that barely moved

Credit-related costs were a ¥6.1bn charge, 5.5% of the ¥110.0bn budgeted for the year, against an ¥11.4bn reversal in the prior-year quarter. Within it, the retail company took a ¥16.0bn charge while the domestic and overseas corporate banks recorded ¥4.0bn and ¥3.5bn of reversals. The balance of reserves recorded from a forward-looking perspective rose ¥3.1bn to ¥165.2bn. Non-performing loans fell ¥77.4bn to ¥842.6bn and the ratio improved to 0.70% from 0.80%.

"Credit-related costs were contained at a very low level of JPY -6.1B. Although there were concerns about the external environment, including the Middle East situation, in Q1, the actual impact was limited. We have built up approximately JPY 165.0B in forward-looking provisions and are well-prepared for risks."
— Makoto Samejima, Group CFO

Assessment: the ¥110.0bn credit guide was the least defensible line in the May plan, being ¥23.0bn below the prior year while management described the geopolitical overlay as unresolved. One quarter in, the guide looks conservative rather than optimistic, and there is real upside to the full-year outlook sitting in this line: even if the quarterly charge triples from here, the year lands near ¥61bn against a ¥110.0bn plan. The reason to hold judgement is that the retail book took ¥16.0bn in a benign quarter and rising rates in Japan work against exactly the leveraged small-business and real-estate borrowers concentrated there.

8. Cross-shareholdings: the plan is intact, the delivery is entirely back-loaded

Net gains related to stocks, on the deck basis that excludes exchange-traded funds, were ¥30.5bn, against a full-year outlook of ¥360.0bn that management left unchanged. That is 8.5% of the plan after one quarter of four, leaving ¥329.5bn to come in nine months. Gross sale proceeds tell a busier story than the net figure: on the Tanshin basis, gains on sales were ¥126.7bn and losses on sales ¥52.4bn, both up sharply. Cumulative reduction against the three-year target of over ¥350bn running from March 2025 to March 2028 stands at ¥125.2bn at acquisition value, or ¥175.8bn including sales already accepted, at a point 42% of the way through the window.

"The sale of cross-shareholdings in Q1 were limited. However, sales negotiations are progressing steadily, and we anticipate an increase in reductions and gains from sales from Q2 onward."
— Makoto Samejima, Group CFO

Assessment: ¥360.0bn of disposal gains is roughly 18% of the ordinary-profit outlook, which our May note flagged as making the programme a dependency rather than a cushion. Nothing this quarter changes that, and the concentration risk has increased: the year now needs an average of ¥110bn of gains in each of the next three quarters, against a Nikkei that has already risen 22.9% this year. If the equity market cooperates the plan is easy. If it does not, this is the line that breaks first, and it breaks in the same quarters the banking-book windfall reverses.

9. The expense ratio management says is too low

At 47.7%, the expense ratio is twelve points better than the same quarter last year and nearly twelve points better than last year's full-year figure of 59.4%. Management refused to bank it.

"The expense ratio landed at a favorable 47.7%, as revenue growth outpaced the increase in expenses. This level is somewhat too low, and for the full FY2026, we are aiming for a level slightly below the previous fiscal year's expense ratio. We will continue our disciplined expense management."
— Makoto Samejima, Group CFO

An aim of "slightly below" 59.4% for the full year, against 47.7% delivered in the first quarter, implies a ratio near 63% for the balance of the year. That holds whether the remaining nine months produce three times the first quarter's gross profits or merely repeat last year's final nine months: the two paths give 62.7% and 63.4% respectively at a 59.0% full-year outcome. Expenses themselves grew 9.8%, which management attributed to yen weakness, inflation, continued investment in growth areas and governance-related costs.

Assessment: the guidance is internally consistent with a company that expects markets revenue to normalise, and it is the clearest confirmation available that management does not regard this quarter's revenue base as the new run-rate. Treat 47.7% as the ratio of a quarter in which trading and disposal income arrived early, not as evidence of a structurally cheaper bank.

10. The securities arm hit a record, and the American build is producing

Non-interest income across the group reached ¥376.8bn, up ¥60.0bn or 18.9%, with increases in almost every division. The securities entities grew net business profits from ¥51.0bn to ¥90.0bn and net income from ¥43.9bn to ¥69.6bn.

"Non-interest business was strong, with YoY increases in almost all divisions. Investment banking was a key growth driver, performing well both in Japan and overseas. The effects of the Greenhill acquisition in the U.S. are being reflected in our results. In particular, the securities business segment achieved record-high earnings."
— Makoto Samejima, Group CFO

Assessment: this is the counterweight to the GCIBC disappointment and it matters more than its size suggests. The overseas strategy has two legs, a balance-sheet-heavy corporate bank and a fee-based advisory and securities build; the balance-sheet leg is losing its rate tailwind while the fee leg is compounding. A group that continues to shift the mix that way improves its return on risk capital without needing the dollar curve to cooperate.

11. The JGB book grew by five and a half trillion yen and got shorter

Japanese government bonds at the two banks rose from ¥14.9tn to ¥20.6tn at acquisition value in three months, with the securities line on the balance sheet up ¥7.2tn. Of the June balance, ¥14.0tn is treasury discount bills and ¥6.5tn medium-to-long-term bonds, and the average remaining period fell to 0.8 years from 0.9. Unrealised losses on Japanese government bonds actually narrowed slightly, to ¥52.5bn from ¥53.8bn. Foreign bonds grew ¥1.3tn at acquisition value with unrealised losses deepening ¥32.2bn to ¥387.0bn, and average remaining period extending to 1.4 years from 1.2.

"While our JGB balance has increased, this occurred within normal ALM management, with a focus on short-term instruments. Also, though the balance of the mid-to- long term bonds increased, the majority of bonds have short remaining duration. Thereby the overall average duration remained short at 0.8 years, and our prudent management policy is unchanged."
— Makoto Samejima, Group CFO

Assessment: this is the right way to be positioned into a tightening cycle and it is worth a moment's attention because it is the opposite of the mistake that has damaged bank balance sheets elsewhere. Two-thirds of a ¥20.6tn government-bond book in treasury bills is a cash proxy earning a rising policy rate, not a duration bet. The residual concern is the foreign-bond book, where unrealised losses of ¥387.0bn are still deepening and the duration went the other way.

12. Earnings per share, buyback mechanics and the milestone management refused to celebrate

The revised outlook implies full-year earnings per share of ¥575.08 against ¥502.92 delivered last year, growth of 14.4% against 12.1% on net profit; the difference is the share count. Issued shares fell 47,016,600 in the quarter as previously repurchased stock was cancelled, and treasury holdings fell from 51.3 million shares to 9.3 million.

"Following the upward revision to our earnings outlook, EPS is also expected to pass our historical high of JPY 551, but we see this as merely a milestone and will aim for further EPS growth."
— Makoto Samejima, Group CFO

Assessment: a management team that frames its own record as a milestone rather than an achievement is signalling that the medium-term target is what it wants to be judged on. That target, set in May, is a return on equity above 12% by FY2028 assuming no further rate rises. On a trailing twelve-month basis the group is already at 12.5%, with the policy rate now above the level the target assumed. The FY2028 target has effectively been reached two years early and is now stale; what should replace it is the open question management did not address.

13. What the raised guide actually asks for

Subtract the first quarter from the revised outlook and the remaining three quarters need ¥977.1bn of net profit. The same three quarters last year produced ¥958.1bn. The raised guide therefore embeds growth of 2.0% over the balance of the year, after a first quarter that grew 45.5%. On net business profits the arithmetic is the same shape: ¥1,174.2bn implied for the final three quarters against ¥1,144.6bn a year ago, or 2.6%.

Assessment: there are only two readings. Either management believes the banking book, the exchange-traded-fund gains and the tax and credit tailwinds genuinely reverse to the point where the rest of the year barely grows, or the guide remains conservative and will be raised again in November. The company's own record points hard at the second. Last year's July raise took the forecast to ¥1,020.0bn, the November raise took it to ¥1,130.0bn, and the outcome was ¥1,248.6bn: the year finished 32.8% above the number set in May and 10.5% above the last number management put its name to. We think ¥1,400.0bn is a floor rather than a target, and that is the central reason for the rating change.

Guidance & Outlook

The assumptions behind the plan are published and are worth reading against reality. The outlook assumes a Bank of Japan policy rate of 1.00%, the Nikkei 225 at ¥57,000 and USD/JPY at ¥150. The policy rate is at 1.00% today and the central bank held there on July 31 while warning that core inflation is running above its 2% target, so the assumption is current but is not a forecast of no further tightening — it is an assumption that none is banked. The Nikkei closed at 66,216.79 on August 20, 16% above the level in the plan. Spot USD/JPY at the June quarter end was ¥162.45, so the plan's ¥150 assumption is materially stronger than the rate that actually prevailed, which flatters nothing in the outlook and understates the translation benefit if the yen stays where it is.

Implied ramp for the balance of the year. The revised outlook needs ¥977.1bn of net profit across the September, December and March quarters, against ¥958.1bn produced by the same three quarters last year. That is growth of 2.0%, following a first quarter that grew 45.5%. On net business profits, ¥1,174.2bn is implied for the final three quarters against ¥1,144.6bn a year ago, or 2.6%. On the disposal line, ¥329.5bn of stock gains is implied against ¥30.5bn delivered, an average of ¥110bn per quarter.

Street position. The revised ¥1,400.0bn guide sits 0.9% above the ¥1,387.0bn twelve-analyst average that was in the market on the day. English-language coverage described the raise as having aligned with Bloomberg-compiled forecasts. The relevant comparison is with three months ago, when the ¥1,300.0bn opening guide sat 0.9% below a ¥1,311.2bn consensus and was the lowest growth rate of the three megabanks. As of August 6 the domestic poll of eleven analysts carried an average twelve-month target of ¥8,707, which is 7.7% above where the shares closed on August 20 and is a materially more comfortable position than the largest peer, whose shares have moved above their own consensus target.

Guidance style. Consistently conservative, and consistently revised upward on the same calendar. Last year's forecast went from ¥940.0bn in May to ¥1,020.0bn at the July results, to ¥1,130.0bn at the interim, and finished at ¥1,248.6bn: 32.8% above the opening number and 10.5% above the final one. This year the opening number was ¥1,300.0bn and the July revision took it to ¥1,400.0bn. A group whose first quarter delivered 30.2% of the revised full-year figure, whose credit costs are running at 5.5% of budget and whose disposal programme has 91.5% of its planned gains still ahead has not set a stretching target.

What They're NOT Saying

1. Still no first-half forecast. The Tanshin's fiscal 2026 first-half row for profit attributable to owners of parent is blank, exactly as it was in May. A company confident enough to raise the full year on one quarter's evidence is not confident enough to frame the half, which removes the only interim checkpoint before the November results.
2. The ¥46.4bn swing in "Others" is unexplained. The elimination and consolidation-adjustment line went from a ¥6.1bn loss to a ¥40.3bn profit and supplied 17.9% of the group's entire increase in net business profits. Gross profits in that line swung ¥58.6bn. Neither the deck nor the CFO's remarks address it.
3. No sizing of the currency benefit. The segment table is translated at budgeted rates and is therefore currency-neutral, but the consolidated income statement is not, and USD/JPY moved from ¥144.82 to ¥162.45 year on year. Management named yen depreciation as a driver of the expense increase and did not quantify what it added to revenue. The largest peer disclosed roughly ¥75bn of gross-profit benefit from the same move; Mizuho disclosed nothing.
4. The ¥120bn-per-25bp rate sensitivity was neither repeated nor revised. In May management said on the record that the disclosed figure was probably too high and that it would be revised. The first-quarter pack contains no interest-rate sensitivity disclosure at all, so the number an investor would use to size the June rate rise is one its author has publicly disowned and has now had two quarters to replace.
5. No quantified Rakuten Bank impact. The alliance takes effect on October 1, 2026 and gives Mizuho a 10.52% voting stake. The financial effect was described as under review at the May results and was not mentioned in the first-quarter materials. With ten weeks to go the order of magnitude is knowable.
6. No target or date for the overseas corporate bank. GCIBC's expense ratio rose 6.4 points and its net income fell 34%. The explanation given was market conditions and the dollar rate environment. There was no cost programme, no revenue target, no return threshold and no timeline.
7. Nothing about what replaces the FY2028 return-on-equity target. The target of above 12% assumes a flat 0.75% policy rate. The rate is 1.00% and trailing return on equity is already 12.5%. Management repeated its ambition to grow earnings per share past the record without addressing the fact that the medium-term target it set three months ago has been overtaken.
8. The consolidated subsidiary count rose from 239 to 270. Thirty-one additional consolidated subsidiaries in twelve months is disclosed in a single line of the Selected Financial Information with no commentary. For a group whose "Others" segment just swung ¥46.4bn, the composition is worth knowing.
9. No update on the retail deposit franchise as a franchise. Individual deposits grew ¥417bn and new account openings were cited, but the deposit-gathering priority named in May still carries no balance target, no cost-of-deposits ceiling and no date. The only retail-funding datapoint offered was the account-opening remark itself.
10. Foreign-bond unrealised losses deepened and were not discussed. Unrealised losses on the foreign-bond portfolio widened ¥32.2bn to ¥387.0bn while average remaining period extended from 1.2 to 1.4 years. The commentary on the securities portfolio addressed the Japanese government bond book only.

Market Reaction

Mizuho released after the close of the Tokyo Stock Exchange on Thursday July 30, which put the American depositary shares and the ordinary shares in different sessions: the New York line reacted the same afternoon, the Tokyo line the following morning. Both rose, and only one of the two moves carries information.

• Pre-print setup: the ordinary shares closed at ¥7,818 on July 30, up 37.2% year to date and 78.3% over twelve months, against a 52-week closing range of ¥4,064 to ¥8,766. The ADR closed at $9.81 on July 29, up 34.0% year to date and 66.6% over twelve months, against a 52-week closing range of $5.84 to $10.66. Both lines had gone almost nowhere in the month before the print: the ordinary shares were up 1.0% over thirty days and the ADR 2.4%. The S&P 500 was up 6.9% year to date entering the print.
• ADR reaction session (July 30): the ADR opened at $9.76, a 0.5% gap down, traded $9.70 to $10.35 and closed at $10.35, up 5.5% and at the high of the day, on 4.4M shares against a 4.2M thirty-day average. The S&P 500 rose 1.7% the same session.
• Tokyo reaction session (July 31): the shares gapped open at ¥8,264, up 5.7%, traded ¥8,003 to ¥8,310 and closed at ¥8,167, up 4.5% or ¥349, on 14.0M shares against a 9.5M thirty-day average. The Nikkei 225 rose 4.03% the same session.
• Three weeks on (to August 20): the ordinary shares reached ¥8,775 on August 13, 12.2% above the pre-print close, before giving most of it back to finish at ¥8,085, up 3.4% from the print. The ADR followed the same path, $10.35 to $10.99 on August 13 and $10.13 on August 20.
• Peer reaction: over the same July 30 to August 20 window the largest peer fell 1.7% and the second-largest fell 1.8%, while the trust bank rose 1.1% and the Nikkei rose 7.0%. Mizuho was the best performer of the four banking names and the only one up more than 1.2%.

The Tokyo session is close to uninterpretable and should be treated that way. The Nikkei rose 2,494 points, or 4.03%, that morning on a semiconductor rally and the Bank of Japan's widely expected decision to hold at 1.00%, one of its strongest single sessions of the year. A 4.5% move in a bank on a day its index rose 4.0% is not a verdict on the print. The New York session is the cleaner read only because the American tape was moving for entirely unrelated reasons, a technology-led rally that lifted the semiconductor index more than 8%; MFG's 5.5% cleared the S&P by 3.8 points on a day when banks were not the story anywhere.

What actually reads as a verdict is the three weeks that followed, because that window contained both peers' prints and a common macro shock. Mizuho outperformed the two larger megabanks by roughly 5.2 points from July 30 to August 20, and it did so through a session on August 19 that took 5.0% off the stock inside a Nikkei down 3.2% and hit all four banking names. The relative move survived the drawdown. Set against a Nikkei up 7.0% over the same window, though, the whole banking group lagged: this was a rotation quarter for Japanese equities and the banks were not where it went.

The more useful frame is where the shares sit against the sell-side rather than against the tape. The domestic poll of eleven analysts carried an average twelve-month target of ¥8,707 a week after the print, which is 7.7% above the August 20 close. Every target action reported after the results was a raise, and not one of them came with a rating change. That is a Street marking its earnings numbers to a raised guide without re-examining what the stock is worth, and it leaves the shares below the published average, which is the opposite of the position the largest peer occupied after its own print.

Street Perspective

Debate: Is the raised guide a genuine upgrade or a catch-up to a number the market already carried?

Bull view: Mizuho was the only one of the three megabanks to move its full-year number this season, and it moved it 7.7% on one quarter of evidence while both peers stood pat. The raise came entirely from the operating line, with credit costs and disposal gains left untouched, and it followed the Bank of Japan's June move, most of which is not yet in the numbers.

Bear view: the revised ¥1,400.0bn lands 0.9% above a twelve-analyst average that was already at ¥1,387.0bn, so the company has done little more than validate where the Street had modelled it. English-language coverage described the raise as having aligned with existing forecasts, and the shares reacted in line with their index at home.

Our take: the bears are right about the level and wrong about the significance. What changed is not the number but the sign of the gap to consensus, which went from 0.9% below in May to 0.9% above in July, and it flipped because the company moved rather than because the Street did. For a management team our May note criticised specifically for guiding last and lowest among its peers, moving first and above is the relevant fact.

Debate: How much of the June quarter can be capitalised?

Bull view: the repeatable half is the half that comes from the balance sheet. Net interest income rose 20.8% on a domestic loan-and-deposit margin that widened to 1.26% from 1.10%, the loan book grew ¥3.3tn in a single quarter, and the June rate rise sat in the numbers for only fourteen days. Non-interest income rose 18.9% with the securities arm at a record. None of that is a trading print.

Bear view: Global Markets supplied 48.7% of the increase in net business profits and its banking book alone supplied 38.2%. Gains on exchange-traded funds contributed ¥38.9bn of the group increase, and a further ¥46.4bn came from an elimination line nobody has explained. The four customer companies, which is what an investor is buying, grew ¥86.7bn, a third of the total, and one of them went backwards.

Our take: the bear arithmetic is correct and is also the company's own view, which is the part the debate keeps missing. A management team that raises the full year by ¥100.0bn after a quarter that beat its own straight-line run rate by ¥97.9bn is telling you it expects the remaining nine months to grow 2.0%. The expense-ratio guidance says the same thing from the other direction. Capitalise the customer groups and the margin, discount the banking book and the exchange-traded-fund gains to near zero, and the quarter is still good enough to carry a raised guide.

Debate: Does doubling the buyback settle the capital-return question?

Bull view: the authorisation went from ¥100.0bn to ¥200.0bn inside ten weeks, the share cap from 25 million to 35 million, the window was extended and every repurchased share is cancelled. Capital was the binding constraint in May and it is not now: common equity tier 1 on the fully-effective basis excluding unrealised gains recovered 50 basis points in a quarter, and management said on the record that it intends to consider further returns at future results.

Bear view: ¥200.0bn of buyback plus roughly ¥365bn of dividends against a ¥1,400.0bn guide is a total payout near 40%, against a stated floor of 50% or more. The company has now announced two tranches and is still ¥135bn short of its own minimum with two quarters to go, while simultaneously flagging strong corporate financing demand as a call on the same capital.

Our take: both are describing a company that has answered the intent question and deferred the level question, and the deferral is the opportunity. A 50% floor implies a third tranche of roughly ¥135bn on the company's own guide and ¥165bn to ¥190bn on our higher full-year estimate, and the natural venue is the November interim, the same event at which this company raised its forecast a second time last year. That is the most identifiable catalyst in the name between here and March, and it is not in the price at a 1.9% dividend yield.

Debate: Should Mizuho still trade at parity with the largest peer?

Bull view: the two facts that made parity look wrong in May have moved. The guide went from the slowest of the three to a 12.1% growth rate the peers did not match, and the retail company that anchored the return problem grew net income 67% with an eleven-point improvement in its expense ratio. On our numbers the shares trade at roughly 13x forward earnings with a return on equity already through the medium-term target.

Bear view: at 1.71x June book against the largest peer's 1.72x, the market is being asked to pay the same multiple for a bank whose overseas corporate arm just posted a 34% fall in net income, whose disposal programme has 91.5% of its planned gains still ahead, and whose best quarter in years was half a markets result. Parity was the bear case in May and the shares have since risen 15.8%.

Our take: the bear case has weakened enough to change the rating without being wrong. The relative-value objection assumed a company that would keep guiding low, returning less and earning single-digit returns in half its franchise; two of those three assumptions broke this quarter. What remains genuinely unresolved is the global corporate bank, and it is the smaller half of the problem now rather than the larger. We are no longer prepared to argue that a bank compounding book value at roughly 8% with a raised guide, a doubled buyback and ¥135bn of undeployed payout capacity should trade at a discount to a peer doing none of those things this season.

Model Update & Valuation Framework

Our May base case assumed a company that guides low and beats. One quarter in, that assumption has been confirmed harder than we modelled, and the revisions below are mostly upward. The rows carry our May numbers so the change is visible.

Valuation. At the August 20 close of ¥8,085 the ordinary shares trade at 1.71x the ¥4,730.1 book value per share struck at June 30, 14.4x trailing twelve-month earnings of ¥560.55 and 14.1x the company's own revised full-year figure of ¥575.08. The dividend yield on the ¥150 estimate is 1.86% and the announced buyback is worth a further 1.02% of the ¥19.68tn market capitalisation, so the committed cash return is roughly 2.9%. Book value per share compounded 1.94% in the June quarter, an annualised pace near 8%, and it did so while the group cancelled 47 million shares.

What the shares are worth on our numbers. A ¥1,460–1,510bn full year against an average share count near 2.42 billion is roughly ¥600 to ¥625 of earnings per share, against the ¥575.08 the company now guides. Holding the current trailing multiple of 14.4x and allowing a modest expansion to 15.0x on the capital-return step-up gives a twelve-month value range of ¥8,640 to ¥9,375, roughly 7% to 16% above the August 20 close, with the dividend on top. The midpoint of that range sits above the ¥8,707 published sell-side average, and the difference is almost entirely the view that ¥1,400.0bn is a floor.

The bridge that decides the rating, revisited. In May we wrote that for the shares to re-rate on anything other than the sector's rate trade, the retail company had to move from a 5.9% return toward its cost of capital and the global corporate bank had to stop losing ground at 8.0%. One quarter later the first has happened faster than we thought possible, with net income up 67% and the expense ratio down eleven points, and the second has gone the other way, with net income down 34% and the expense ratio up 6.4 points. Half of the arbitrage inside Mizuho has closed. The rating change reflects that half, and the residual risk is now concentrated in a single segment consuming roughly a quarter of group risk capital rather than spread across two consuming nearly half.

Thesis Scorecard Post-Earnings

These are the six pillars established at initiation in May, graded against this quarter. For a bull pillar, Confirmed means the quarter supported it; for a bear point, Confirmed means the risk showed up and Challenged means the quarter argued against it.

Overall: the thesis is materially stronger than it was in May. Four of the six pillars moved in Mizuho's favour, two bear points de-escalated a tag, and the two that did not, the disposal programme and the global corporate bank, are both back-loaded risks rather than present ones. The quarter's own quality is the qualification rather than the case: half of it is a markets result that will not repeat, which is why the argument rests on the guide, the payout and the capital ratio instead.

Action: buy. The rating change is not a call on the June quarter. It is a call on a company that raised a forecast its peers would not touch, doubled a buyback ten weeks after setting it, repaired the capital ratio that carried our bear case, and still has roughly ¥135bn of payout capacity uncommitted against its own floor. We would revisit on a November interim that fails to add a third buyback tranche, on evidence that the disposal programme cannot deliver ¥110bn a quarter, or on a further leg down in the global corporate bank.