MUFG Fixed Its Capital Ratio a Quarter Early and Beat by 28%, Then Lagged Every Peer for Raising Nothing

Results vs. Consensus

MUFG reports under Japanese GAAP, publishes a full-year target rather than a forecast, and gives no quarterly guidance at all. A quarter is therefore measured against three yardsticks: the sell-side poll, progress toward the company's own target, and the prior-year quarter. This one cleared all three by a wide margin.

Q1 scorecard (three months ended June 30, 2026)

A note on the consensus figure. The ¥633.4bn average is the number carried through post-print coverage and is the benchmark we use. A second yen-denominated estimate we hold puts the quarter at ¥689.3bn with a high end of ¥740.8bn, so the actual cleared even the top of that range by 9.3%. On any available reading the beat was large, and the ambiguity is about how large rather than whether.

Year-over-year comparison (¥bn)

Two lines in that table deserve to be read together before anything else. Gross profits grew ¥377.7bn and G&A grew ¥111.7bn, so 70% of the incremental revenue fell through to net operating profit. That is the arithmetic of a bank with a fixed cost base meeting a repricing cycle, and it is why the expense ratio fell 6.6 points to 53.4% without any cost programme being announced.

Progress against the company's own full-year plan

Every line is ahead of pace and credit costs are running well behind the budget, which is the favourable direction. Trailing twelve-month net profit is now ¥2,690.6bn, which is to say the company has already earned its full-year target over the last four quarters and has told the market nothing about that.

Revenue assessment

This is the first quarter in which the rate cycle, rather than a base effect, is the visible driver. Net interest income grew 27.7%, interest on loans and bills discounted grew 17.6% to ¥1,139.0bn, and interest and dividends on securities grew 30.9% to ¥539.2bn, against interest on deposits up only 12.0% to ¥554.3bn. That gap is the whole thesis in one line: a ¥236.9tn deposit book repricing more slowly than a ¥134.5tn loan book. The Bank of Japan's move to 1.00% on June 16 landed with two weeks of the quarter remaining, so almost none of its roughly ¥100bn first-year value is in these numbers.

The fee line is the second engine and it is not currency. Trust fees and net fees and commissions grew ¥104.3bn, or 20.9%, with the company naming solutions, lending-related business and asset management across both domestic and overseas franchises. On the deck's local-currency view, investment product sales alone added ¥16.3bn in the wealth-management group and ¥6.8bn in retail, which is what a rising domestic equity market does to a bank that also distributes funds.

Margin assessment

A 53.4% expense ratio is a remarkable number for this company. The medium-term plan targeted approximately 60%, which was hit a year early in FY2025, and the quarter has now come in six and a half points inside it. The seasonal objection does not do much work here: the June quarter of FY2025 also printed exactly 60.0%, identical to that year's full-year ratio, so this is not a quarter that habitually flatters. What it is instead is a ratio driven entirely by its denominator. Costs grew 13.7% and revenue grew 27.8%, and since a meaningful share of the revenue came from treasury positioning, equity disposals and currency, the full-year ratio will almost certainly land above 53.4%. Management has not said where.

Underneath the ratio, the cost base is still building for structural reasons that have not changed since the initiation: artificial intelligence deployment, cybersecurity, the retail digital programme, and Japanese wage inflation. Depreciation grew 10.1% to ¥97.8bn and goodwill amortization 19.7% to ¥11.2bn, both consistent with a group still investing and still acquiring.

EPS assessment

Basic EPS of ¥71.77 grew 50.9% against 48.2% net-profit growth. The 2.7-point wedge is entirely the share count: average outstanding shares fell from 11,484m to 11,279m as the ¥100bn repurchase was executed. It is worth being precise about what that buyback is worth, because it is the smallest annual authorization in five years. At an average price of ¥3,128 the programme retired 31,967,100 shares, or 0.28% of the shares outstanding excluding treasury. Repeated at that rate it adds a little over one point a year to per-share growth, against the two-plus points the ¥400bn to ¥500bn programmes of FY2022 through FY2025 delivered.

Diluted EPS of ¥71.59 sits ¥0.18 below basic, a 0.25% dilution gap that is immaterial and unchanged in character from the prior year's ¥0.10.

Segment Performance

MUFG discloses its businesses twice, on two different bases, and the two tell slightly different stories. The Summary Report's segment note is the Japanese GAAP view: it reconciles to consolidated ordinary profit and it is the basis on which the company's own bridge of the quarter is drawn. The Financial Highlights deck restates the same groups on a managerial-accounting, local-currency basis, which strips currency translation out of the year-over-year comparison and adds the revenue-line detail. We use the first for the shape of the quarter and the second for what happened inside each business.

One disclosure change matters before reading either. In the June quarter MUFG changed how it allocates net revenue and operating expenses among reporting segments, and restated the prior-year quarter on the new method. The comparatives below are therefore internally consistent, but they are not comparable to segment figures published before August 2026, and no bridge between the old and new methods was provided.

Segment note, Japanese GAAP basis (¥bn)

The number that reframes the quarter. Segment operating profit rose ¥278.3bn. The six customer groups contributed ¥205.8bn of that, or 73.9%; Global Markets contributed ¥67.7bn, or 24.3%. Compare the year we wrote up in May, when Global Markets alone supplied 90% of the consolidated net-operating-profit increase because the prior-year base carried a deliberate ¥991.4bn bond-rebalancing loss. That base effect is gone, and the customer franchise is now carrying the quarter on its own. This is the single most important thing in the print and it is the direct confirmation of the bull pillar we were least able to prove in May.

The customer groups' expense ratio fell from 57.4% to 52.2%, and every one of the six grew operating profit. Growth was not evenly distributed: the corporate and investment banking businesses and wealth management supplied ¥171.6bn of the ¥205.8bn, while retail, the overseas commercial banks and asset management supplied ¥34.0bn between them.

Business groups, managerial-accounting and local-currency basis (¥bn)

The table below is the company's currency-neutral view, which is the fairer read of what each business did rather than what the yen did to it. Net income by group is disclosed only on this basis.

Global Corporate & Investment Banking: the biggest single contributor, and the one to watch

GCIB produced the largest operating-profit increase of any customer group on the GAAP segment basis, ¥70.8bn, though on the currency-neutral view its ¥40.3bn is narrowly behind the corporate bank's ¥43.2bn. Either way the expense ratio collapsed nine points to 45%. Commission income drove it, up ¥21.5bn to ¥95.0bn, alongside loan and deposit interest income up ¥17.0bn to ¥104.1bn. Net income grew ¥17.1bn to ¥81.1bn. Coverage of the quarter singled out the US project-finance franchise and its data-centre pipeline as a demand source.

Assessment: This is the group we flagged in May as the sharpest illustration of the capital problem, having added ¥3.7tn of risk-weighted assets in FY2025 for a four-point ROE decline. The revenue answer this quarter is emphatic, and a 45% expense ratio on a growing fee base is a genuinely good business. What we cannot check is whether the return on the capital consumed has recovered, because MUFG does not disclose group-level risk-weighted assets or ROE at the first quarter. Given that consolidated risk-weighted assets grew ¥4.7tn in the same three months, the answer matters, and we have to wait until November for it.

Japanese Corporate & Investment Banking: fee mix continuing to do the work

Gross profits grew ¥49.4bn currency-neutral, with loan and deposit interest income up ¥24.5bn to ¥152.3bn and derivatives and solutions up ¥16.4bn to ¥31.4bn. Merger, debt and equity capital markets fees added ¥3.6bn and real estate and corporate agency another ¥3.6bn. Expenses grew only ¥6.3bn, taking the expense ratio to 36% from 41%, the lowest of any group. Net income reached ¥139.4bn.

Assessment: Roughly half of the growth is fee income rather than spread, which is the mix we said we wanted to see. The 36% expense ratio makes this the most profitable large business in the group on a revenue-to-cost basis. The unresolved question from May was that JCIB added ¥2.9tn of risk-weighted assets in FY2025 to produce ¥15.5bn of incremental net income. This quarter alone it produced ¥22.5bn of incremental net income, which is a far better trade if the balance-sheet growth behind it was proportionate. Again, undisclosed at Q1.

Commercial Banking & Wealth Management: still the cleanest expression of the rate cycle

Loan and deposit interest income grew ¥33.3bn to ¥122.2bn, up 37.5%, and investment product sales grew ¥16.3bn to ¥56.0bn as domestic equities ran. Derivatives and solutions added ¥10.3bn. Gross profits grew 29.4% currency-neutral against expenses up 10.9%, taking the expense ratio down eight points to 48% and net income up 40.2% to ¥87.9bn.

Assessment: This remains the purest read on what a normalizing deposit franchise is worth, and it is compounding faster than the group. Two engines are firing at once, spread income from the rate cycle and fee income from a record domestic equity market, and only one of them is structural. If Japanese equities stall, roughly a quarter of this group's revenue growth stalls with them. That is not a criticism of the quarter; it is a reason not to annualize it.

Retail & Digital: the cost curve finally bends

Gross profits grew ¥37.4bn currency-neutral, of which ¥29.0bn was loan and deposit interest income. Expenses grew ¥19.3bn, so for the first time in the period we have tracked, revenue growth meaningfully outran cost growth in this group: operating profit grew ¥18.0bn and the expense ratio fell three points to 72%. Net income, however, grew only ¥0.4bn to ¥23.4bn.

Assessment: In FY2025 this group turned ¥119.5bn of revenue growth into ¥5.4bn of operating-profit growth, and we called that either a defensible investment or a warning. One quarter of 48% flow-through is the first real evidence for the defensible reading. The flat net income against ¥18.0bn of operating-profit growth is the discordant note and is unexplained. Still no unit economics for the digital build: no cost per account, no revenue per active user, no payback period. The burden of proof has moved, but it has not been discharged.

Global Commercial Banking: the acquisitions are annualizing slowly

Krungsri gross profits grew ¥8.3bn to ¥126.4bn and Bank Danamon ¥7.5bn to ¥45.4bn, but expenses grew ¥7.1bn and the group's expense ratio worsened a point to 57%. Operating profit grew only ¥2.8bn, the weakest of the six customer groups in absolute terms. Net income grew ¥9.5bn to ¥34.0bn, and the split is instructive: Bank Danamon contributed ¥4.6bn of that increase while Krungsri's net income contribution actually fell ¥2.0bn to ¥18.5bn.

Assessment: The FY2026 plan expects ¥17bn to ¥18bn of incremental profit as the Southeast Asian acquisitions annualize. On this quarter's run-rate the group is roughly on track, but the composition is wrong: the contribution is coming from Danamon and from consolidation effects rather than from Krungsri, which is the larger platform and the one whose expense ratio is deteriorating, from 49% to 51%. This is the group most exposed to the Middle East supply-chain risk management sized last quarter, and it is earning the second-lowest return in the portfolio.

Asset Management & Investor Services: the smallest contribution against the second-largest expectation

Investor Services grew ¥5.0bn to ¥73.8bn, pension ¥2.8bn to ¥22.8bn and asset management ¥1.4bn to ¥37.8bn. Expenses grew ¥5.6bn against ¥9.2bn of revenue growth, so operating profit grew ¥3.6bn and the expense ratio improved a single point to 72%. Net income grew ¥2.0bn to ¥24.9bn.

Assessment: The FY2026 plan requires ¥38bn to ¥40bn of incremental profit from this group, the second-largest identified contribution in the whole bridge. Q1 delivered ¥2.0bn of net-income growth. That is roughly a fifth of the pace the plan needs, and it is the one line in the segment table that is visibly behind. The impairment cycle at the overseas asset-management businesses that suppressed FY2025 has at least not recurred, but nothing has replaced it either.

Global Markets: the normalization we underwrote, arriving faster than expected

Global Markets gross profits grew ¥107.9bn currency-neutral to ¥271.5bn, and the split is stark. Sales and trading grew ¥20.1bn to ¥99.7bn, a respectable 25%. Treasury grew ¥97.3bn to ¥170.0bn, and at the operating line treasury produced ¥156.0bn against ¥59.8bn, a ¥96.2bn swing. The expense ratio fell fifteen points to 28%. Net income was ¥128.7bn against ¥81.7bn.

Assessment: In May we wrote that treasury was still ¥65.1bn in the red at the full-year operating line and that another leg of normalization was available without market help. It arrived in a single quarter and then some. The caution is symmetrical: a treasury book that can add ¥96bn in a quarter can subtract it, and FY2024's ¥991.4bn bond-rebalancing loss and FY2025's ¥200bn hedging-review charge are both recent memory. Sales and trading, the genuinely repeatable half, grew 25%, which is the number we would extrapolate. Treasury is not.

Entity view: where the ¥809.4bn was earned

Subsidiary figures other than the two banks are approximate and stated before consolidation adjustments, and MUFG's ownership percentage is not reflected in the net income column for the partly-owned entities. Read as a distribution rather than a sum.

Assessment: Morgan Stanley's ¥222.2bn is 27.5% of consolidated net profit from a single 24.2% stake in a company MUFG does not manage. That is the bear point we opened at Neutral in May, and this quarter it is larger, not smaller. Note also ACOM, where net income fell ¥14.9bn despite operating profit rising ¥4.3bn, on a deferred-tax charge arising from a change in the company's classification for deferred-tax-asset recoverability. That is a technical item and MUFG discloses it as such in a footnote, but it is a reminder that the consumer-finance subsidiaries can put several billion yen of noise between operating profit and net income in either direction.

Balance Sheet, Capital and Credit

For a bank in the middle of a capital rebuild, the balance sheet is the story rather than the appendix. Three things happened in the quarter: the capital ratio was repaired ahead of schedule, the loan book kept growing, and asset quality improved sharply for reasons the company did not explain.

Capital: the commitment was met a quarter early

MUFG quotes its capital ratio on four bases and the target range applies to only one of them. The published comparison, on the same disclosure page for both dates, is below.

"The Common Equity Tier 1 capital ratio was 12.95% and excluding impact of net unrealized gains on available-for-sales securities was 10.8%. Furthermore, the Common Equity Tier 1 ratio on Finalized and fully implemented Basel Ⅲ basis(note1) was 11.4% and excluding impact of net unrealized gains on available-for-sales securities was 9.5%."
— MUFG, published FAQ on the first-quarter results, August 3, 2026

In May, management said it expected the ratio to recover "to near the lower end of the target range during the first half of the year." It reached the lower end of the range in the first quarter of that half. The underlying capital arithmetic, from the August 14 Basel 3 release, is a Common Equity Tier 1 capital base up ¥1,199.9bn, or 8.0%, to ¥16,202.1bn, against risk-weighted assets up ¥4,738.9bn, or 3.9%, to ¥125,020.7bn. Capital grew twice as fast as the assets it supports, which is what a ¥809.4bn quarter does.

Assessment: This is a clean and early delivery on the most consequential commitment management made in May, and it is the reason a downgrade here is about price rather than performance. It is worth being precise about the mechanism, though. The ratio did not recover because the balance sheet was restrained. Risk-weighted assets grew 3.9% in three months, an annualized pace of about 17%, faster than the 12.5% that consumed the buffer in FY2025. What repaired the ratio was an exceptional earnings quarter, one in which the lines management does not control on a quarterly cadence were equivalent to 44.5% of net profit. If earnings normalize toward the plan while risk-weighted assets keep growing at this rate, the second-half buyback decision gets harder, not easier.

The buyback: authorized ¥100bn, spent ¥99,999,771,608

The balance sheet corroborates it precisely: treasury shares rose 31,841,171 to 611,946,162 over the quarter and treasury stock rose ¥99,846m to ¥1,033,983m. The programme was finished five days inside its window with ¥228,392 of the authorization unspent. The yen cap was always the binding one: 45,000,000 shares against ¥100bn implies an average price of ¥2,222, and the shares never traded below ¥2,982 during the repurchase window.

Assessment: Management executed the full authorization at speed rather than dribbling it out, which is the behaviour of a board that wanted the capital returned rather than one hedging its options. That is a mild positive signal for the second-half decision. The offsetting read is that at a ¥3,128 average MUFG bought its own stock at 1.59 times March book value against 1.81 times today, so the repurchase was executed at what turned out to be a good level, and there is now nothing running while the stock re-rates.

Balance sheet and loan book

On the banking and trust account basis the deck uses, total loans grew ¥0.5tn to ¥135.7tn, and the composition inside that small net number is the interesting part. Lending to governments and governmental institutions fell ¥2.7tn to ¥0.7tn, essentially completing the runoff that began last year. Domestic corporate lending grew ¥1.4tn to ¥61.5tn. Overseas lending grew ¥1.7tn to ¥57.3tn, or ¥1.0tn excluding currency. Housing loans were flat at ¥14.4tn.

Within the domestic book, large-corporate lending grew ¥1.8tn to ¥30.9tn while lending to small and medium-sized companies fell ¥0.4tn to ¥30.6tn. That is a mix shift away from the higher-spread half of the domestic franchise, and it went entirely without comment.

The domestic spread continued to widen. On the combined bank and trust bank basis, excluding loans to the government, the average lending rate reached 1.47% against 1.13% a year earlier while the deposit and certificate rate rose to 0.30% from 0.17%, taking the spread to 1.16% from 0.95%. Overseas lending spreads were broadly flat at about 1.4%.

On the funding side, deposits fell ¥2.5tn. Retail deposits grew ¥0.6tn to ¥94.8tn while corporate and other domestic deposits fell ¥3.9tn to ¥87.1tn and overseas deposits grew ¥0.7tn to ¥54.9tn. Negotiable certificates of deposit, which are wholesale funding, grew ¥1.7tn. The balance at the Bank of Japan fell ¥6.0tn to ¥63.3tn.

Assessment: The deposit line is the one we would put a marker against. A ¥3.9tn decline in domestic corporate deposits replaced in part by ¥1.7tn more of negotiable certificates is a small step down the funding-cost curve, and it is precisely what a rising policy rate does to corporate treasurers who have had nowhere to put cash for twenty years. The retail book, which is the cheap and sticky part, grew. Nothing here is alarming at this scale against ¥236.9tn of deposits, but the deposit franchise is the entire bull case for this name, and this is the first quarter in which the aggregate has gone backwards.

Securities: shorter domestic duration, larger unrealized gains, a bigger held-to-maturity hole

Assessment: MUFG bought ¥1.37tn more Japanese government and domestic bonds while shortening the portfolio's average duration from 1.5 years to 1.2. That is the correct trade for a bank that expects the Bank of Japan to keep moving: more carry, less price risk, and it explains why available-for-sale unrealized losses on domestic bonds barely moved despite the June hike. The item that did move the wrong way is held to maturity, where unrealized losses deepened ¥0.14tn to ¥1.26tn. Those losses do not touch capital under the current treatment, but they are real and they grow with every further move in yields, and MUFG does not discuss them.

Credit: a sixteen basis point improvement nobody explained

By region, the improvement is overwhelmingly domestic: non-performing loans to Japanese borrowers fell from ¥674.2bn to ¥503.7bn, while EMEA fell from ¥103.6bn to ¥71.9bn, the Americas from ¥205.6bn to ¥199.5bn and Asia from ¥477.2bn to ¥462.7bn.

Total credit costs of ¥72.1bn break down as a ¥14.8bn net reversal at MUFG Bank, a ¥29.9bn charge at the consumer-finance subsidiaries and a ¥56.5bn charge at the overseas subsidiaries. Against a full-year plan of ¥350bn, the quarter used 20.6%. The forward-looking overlay held against Middle East geopolitical risk rose to ¥28,446m from ¥24,357m at March.

"For the three months ended June 30, 2026, such assumptions remained substantially unchanged because no significant changes were observed subsequent to the previous fiscal year end with respect to the events or circumstances underlying the outlook relating to the geopolitical environment, including the situation in the Middle East."
— Consolidated Summary Report for the three months ended June 30, 2026, Additional Information

Assessment: A ¥172.3bn reduction in restructured loans in a single quarter, driving 77% of the total improvement, is a large event at a bank whose entire non-performing book is ¥1.2tn. It could be repayment, an upgrade in internal credit ratings, a disposal, or a write-off, and each has a different meaning for the forward credit trajectory. MUFG disclosed the movement and explained none of it, and with no conference call there was no mechanism to ask. A 0.80% non-performing loan ratio is the lowest this group has reported and we take it at face value, but an improvement of this size arriving unexplained is a lower-quality data point than the same number arriving with a cause attached.

Key Topics from the Disclosure

Overall disclosure posture: Minimal and unusually silent for a quarter this good. MUFG published the numbers, the deck and a two-question FAQ, held no call, changed nothing in its targets and offered no interpretation of a result that came in 28% above the Street and 30% of the way to a full-year plan in one quarter. Where the company did speak, it was precise and volunteered the useful basis rather than the flattering one: the capital answer gives all four CET1 bases including the one the target range is set on. Where it did not speak, the omissions cluster around the two questions the market actually has, which are what the second-half buyback will be and why the loan book keeps growing faster than plan.

1. The capital commitment was delivered a quarter ahead of schedule

The one thing that could have derailed the investment case in May was the possibility that the CET1 shortfall was structural rather than a timing problem. It was a timing problem, and the timing was shorter than management's own guide. The ratio on the target basis moved from 9.2% to 9.5%, and it did so while the entire ¥100bn buyback was bought and while risk-weighted assets grew ¥4.7tn.

Assessment: Our May working assumption was 9.6% by the end of the first half and 9.9% by year end. The first-half figure was effectively hit in the first quarter, which puts the year-end number ahead of our estimate if the earnings run-rate holds and risk-weighted asset growth moderates. This resolves the single largest uncertainty in the thesis in the shareholder's favour.

2. The target was left unchanged, and that is the quarter's most consequential non-event

MUFG earned 30.0% of its full-year net-profit target in a quarter that is ordinarily its smallest, and reaffirmed the ¥2,700bn target and the ¥96 dividend without comment. In May, management told the market it had abandoned the practice of setting a target it expected to beat, and had instead set the most likely outcome with a stated willingness to revise downward if conditions worsened.

"MUFG has an earnings target of 2,700.0 billion yen of profits attributable to owners of parent for the fiscal year ending March 31, 2027. (There is no change to our earnings target released on May 15, 2026.)"
— Consolidated Summary Report for the three months ended June 30, 2026

Assessment: Two readings are available and the company did nothing to help choose between them. The generous reading is that a Japanese bank does not revise a full-year target on one quarter, particularly when a large part of the beat came from treasury positioning and equity disposals that it has no intention of repeating at that rate. The unfavourable reading is that the "most likely outcome" framing management adopted in May implies a target that moves when the most likely outcome moves, and the most likely outcome plainly moved. Both of MUFG's closest peers reported within days of this print; one of them raised. The market appears to have taken the second reading.

3. The customer franchise is now carrying the growth, which is the pillar we could not prove in May

Six customer groups produced ¥205.8bn of the ¥278.3bn increase in segment operating profit, or 73.9%, with the customer expense ratio down from 57.4% to 52.2%. In the year we initiated on, Global Markets supplied 90% of the consolidated increase and the customer groups grew 13.1%, which was a good number obscured by a spectacular one. This quarter the relationship is the right way round.

Assessment: This is the most important structural datapoint in the print and it moves Bull 2 from confirmed-on-a-thirteen-percent-basis to confirmed-emphatically. The caveat is that the currency-neutral view of the same groups shows ¥153.4bn of operating-profit growth rather than ¥205.8bn, so roughly a quarter of the customer-segment increase is the yen rather than the franchise. Both figures are the company's own and both are correct on their own basis.

4. The Bank of Japan moved inside the quarter, and almost none of it is in these numbers

The FY2026 plan assumed a policy move by July. The Bank of Japan raised the policy rate 25 basis points to 1.00% on June 16, the highest level since 1995, then held at the late-July meeting. Management has sized the sensitivity at approximately ¥100bn of first-year profit for each 25 basis points.

Assessment: The move landed with two weeks of the quarter remaining, so the ¥809.4bn was earned almost entirely on the pre-existing rate structure. Roughly ¥95bn of that ¥100bn first-year benefit is still ahead of the company, spread across the September, December and March quarters. Set against a full-year target of ¥2,700bn, of which ¥809.4bn is banked, that is a meaningful tailwind the plan already assumed but the first quarter did not receive. The asymmetry we flagged in May remains: the plan contains no benefit from any second move, and a policy rate of 1% is still well below any plausible neutral level for Japan.

5. Domestic spreads widened again, and the mix within the domestic book got worse

The domestic lending rate excluding government borrowers reached 1.47% from 1.13%, against deposits at 0.30% from 0.17%, so the spread widened 21 basis points year on year to 1.16%. Underneath that, domestic corporate lending grew ¥1.4tn, but the whole of it and more came from large corporates, up ¥1.8tn to ¥30.9tn, while lending to small and medium-sized enterprises fell ¥0.4tn to ¥30.6tn.

Assessment: Large-corporate lending is the lower-spread half of the domestic book and small and medium-sized enterprise lending is the higher-spread half. Growing the first while shrinking the second is a mix shift in the wrong direction for spread income, and it is happening at the same moment that the group is consuming capital rapidly. One quarter is not a trend and Japanese corporate lending is lumpy. It is, though, exactly the sort of thing an analyst would have asked about on a call.

6. Overseas loans grew when they were supposed to shrink

In May, management attributed a ¥9.2tn increase in overseas lending to bridge loans for Japanese acquirers and to a timing difference between warehousing and sell-down in the originate-to-distribute business, describing roughly ¥5tn of it excluding currency as containing ad hoc factors. We wrote at the time that if those positions cleared in the first half the capital ratio would recover on its own, and that if distribution markets tightened they would not. Overseas loans grew a further ¥1.7tn in the quarter, or ¥1.0tn excluding currency, to ¥57.3tn. The Americas added ¥0.6tn, EMEA ¥0.5tn and Asia and Oceania ¥0.4tn.

Assessment: The warehoused positions did not run off. Either the originate-to-distribute pipeline is still congested, or new origination more than replaced what distributed, or the ad hoc characterization was too generous in the first place. The capital ratio recovered anyway, on earnings. This is the single clearest negative in the quarter and it is the one that keeps Bear 1 alive: the group is still buying growth with balance sheet, and the only reason that has stopped showing up in the capital ratio is that the profit line got very large.

7. Global Markets treasury swung by ¥96bn and nobody had to explain it

Treasury operating profit was ¥156.0bn against ¥59.8bn. On the full year just ended, this same operation lost ¥65.1bn at the operating line after absorbing a ¥200bn charge from the review of yen interest-rate hedging. The swing between those two states is larger than the entire increase in the six customer groups' operating profit.

Assessment: We flagged in May that treasury being loss-making offered another leg of normalization without market help. That leg was taken in a single quarter, which is a good outcome and also the reason to discount it. Sales and trading, the part that reflects client activity, grew 25% to ¥99.7bn of gross profits, and that is the number we carry forward. A treasury result of this magnitude is a positioning outcome in a quarter when Japanese rates rose and the yen fell 12%, and positioning outcomes reverse.

8. Equity disposals: the gains are large, the programme is barely moving

Net gains on sales of equity securities were ¥103.5bn against ¥31.5bn, and net gains on equity securities overall were ¥98.9bn against ¥30.3bn. Meanwhile the cumulative reduction of equity holdings under the current medium-term plan moved from ¥441bn to ¥460bn, roughly ¥19bn of acquisition cost in the quarter, against a ¥700bn target. Including positions agreed but not yet sold, total expected disposals rose ¥7bn to ¥603bn. Because Japanese equities rose, the domestic equity securities carried on the balance sheet grew from ¥3.74tn to ¥4.09tn and the unrealized gain on them grew from ¥2.82tn to ¥3.19tn.

Assessment: The reservoir is refilling faster than it is draining. On this quarter's pace the ¥700bn target requires roughly ¥240bn more of cost-basis disposals against ¥19bn achieved, which is not a run-rate that reaches the target inside the plan period without a step change. The P&L consequence is the opposite of the disclosure headline: gains are running hot while the underlying reduction stalls, and when the programme does complete, a large annual profit item stops. Management gave no revised timeline and no indication of what the profit base looks like without it.

9. Morgan Stanley is a bigger share of the result, not a smaller one

Equity in earnings of equity-method investees was ¥261.6bn against ¥158.0bn, and the Morgan Stanley stake contributed ¥222.2bn of net income on a holding MUFG puts at 24.2%. That is 27.5% of consolidated net profit from one investment, up from a full-year FY2025 equity-method contribution of ¥845.5bn that was 34.8% of net profit.

Assessment: Directionally the ratio improved because MUFG's own earnings grew faster, but the absolute dependence increased and the concentration in a single holding rose. A US investment bank in a strong trading and dealmaking environment is a fine asset to own and it is also the most cyclical earnings stream in the group. An investor buying MUFG at 1.8 times book is paying a bank multiple for a block of earnings that is, in substance, a levered stake in Wall Street revenue.

10. Capital policy: the standing answer, unchanged

MUFG restated its capital-return framework in the FAQ without amendment.

"We will keep enhancing the content of shareholder returns with dividends as a primary vehicle. The dividend payout ratio is set at around 40%, aiming for a stable and sustainable increase in dividends per share through profit growth. We will flexibly repurchase stock as return profits to return profits to shareholders that will contribute to improving capital efficiency, taking into consideration our business performance and capital situation, opportunities for growth investment, and the market conditions, including the stock prices."
— MUFG, published FAQ, capital-policy section

Assessment: Two clauses in that answer now carry more weight than they did in May. "Opportunities for growth investment" is the clause under which loan growth outranks repurchases in the capital queue, and the loan book grew again this quarter. "The market conditions, including the stock prices" is a clause that argues for a smaller repurchase after a 22% three-month rally than before it. The framework is unchanged, but the facts it is applied to have moved in a direction that makes a large second-half buyback less rather than more likely, even though the capital ratio now permits one.

Analyst Q&A

There was none. MUFG holds a results conference call for the interim and the full year only, and this quarter's Summary Report records "Investor meeting presentation: None" on its cover page. No analyst put a question to management about the unchanged target, the ¥172.3bn fall in restructured loans, the continued growth in overseas lending, or the second-half buyback, because there was no forum in which to do so. The next such forum is the interim results call in November. We flag the absence rather than pass over it, because for a quarter with this many unexplained movements it is a material limitation on what any analysis of it can claim.

What They're NOT Saying

1. Why the target did not move. Thirty percent of a full-year target in the seasonally smallest quarter, with credit costs running at 20.6% of budget, and no commentary at all. The May framing was that the target reflects the most likely outcome rather than a conservative floor. If that is still true, the target should be under upward review, and if it is not still true, the guidance philosophy changed again without being announced.
2. The second-half buyback. The condition management set was the ratio returning to the target range, and it has. There is no interim signal, no framework for sizing, and nothing until November.
3. Why restructured loans fell ¥172.3bn. Seventy-seven percent of the improvement in the non-performing book came from one category in one quarter, with no explanation of whether it was repayment, upgrade, disposal or write-off.
4. Why overseas loans grew again. The bridge-loan and warehousing explanation given in May implied these balances would normalize. They grew ¥1.7tn instead, and the company did not revisit the characterization.
5. Any ex-equity-holdings ROE. Management volunteered a 10.4% ex-equity-holdings figure alongside the 11.3% headline at the full year, and it was the more useful number. The 14.4% quarterly headline has no such companion, in a quarter containing ¥103.5bn of equity-sale gains.
6. Segment risk-weighted assets or returns. Not disclosed at the first quarter, so the central question we raised in May, whether the businesses consuming the most capital are earning an adequate return on it, cannot be tested until November.
7. The revised equity-disposal timeline. Nineteen billion yen of cost-basis reduction against ¥240bn still required, with no restatement of when the ¥700bn target will be reached or what earnings look like afterwards.
8. The shrinking SME loan book. A ¥0.4tn decline in the highest-spread part of the domestic franchise, alongside a ¥1.8tn increase in the lowest-spread part, went unremarked.
9. The segment allocation change. MUFG changed the method by which revenue and expenses are allocated across business groups this quarter and restated the comparative, but published no bridge between the old and new bases. Anyone maintaining a segment model has to rebuild it and cannot reconcile the rebuild to what was published previously.
10. The held-to-maturity mark. Unrealized losses deepened to ¥1.26tn. The company shows the number in a table and discusses the available-for-sale book at length. The held-to-maturity hole gets no narrative at all.
11. A full-year expense-ratio target. The medium-term plan called for a ratio below FY2025's 60.0%. The quarter printed 53.4% and there is no statement of where the year is expected to land.

Guidance & Outlook

MUFG made no change to any published figure. The full-year target, the medium-term plan, the ROE goal and the dividend forecast are all as set on May 15.

Implied ramp. With ¥809.4bn banked, the remaining three quarters need ¥1,890.6bn to reach the target. The same three quarters last year produced ¥1,881.1bn. The target therefore implies net profit growth of 0.5% for the balance of the year, after a first quarter that grew 48.2%. On the operating line the bar is real rather than nominal: net operating profits need ¥2,091.0bn over the remaining three quarters against ¥1,834.2bn last year, or 14.0% growth. Ordinary profits need 4.8% growth on the same basis. The gap between a 14% operating requirement and a 0.5% net requirement is the plan's embedded assumption that equity-securities gains and equity-method earnings step down from last year's levels, which is consistent with the disposal programme completing.

Where the Street sits. No quarterly consensus is published for this company; coverage is modelled at the full-year level. The pre-print full-year consensus was below the company's own target, and the company's target has not moved, so the arithmetic of the quarter has pushed the Street's number up rather than the company's.

Guidance style. In May, management stated plainly that the FY2026 target reflected the most likely scenario rather than a conservative floor, and that a deterioration would be met with a downward revision rather than absorbed by a buffer. A framework that revises down on bad news and stays put on good news is not the symmetric one that was described. One quarter is too early to conclude the philosophy reverted, but the November interim is now a test of it as much as of the numbers.

What a raise would require. If the remaining three quarters merely match last year's ¥1,881.1bn, the full year lands at ¥2,690.6bn, essentially on target. If they grow 8%, which is below the plan's own operating-profit requirement, the year lands near ¥2,840bn. The target is a floor on any reasonable path, and the company has chosen not to say so.

Market Reaction

• Pre-print setup. The ADR closed at $22.45 on July 31, up 41.6% year to date against 9.4% for the S&P 500, up 61.3% over the trailing twelve months and up 8.9% over the trailing thirty days. It sat inside a 52-week closing range of $13.78 to $23.10. In Tokyo the ordinary shares closed at ¥3,571 on July 31 and ¥3,567 on August 3, the session before the release.
• Release timing. The results were published on August 3 after the Tokyo close, so the ADR's August 3 US session was the first post-release session for the New York line and August 4 was the first for the primary listing.
• ADR reaction session, August 3. The ADR opened at $22.74, traded $22.35 to $22.87 and closed at $22.49, up 0.2% on 3.8m shares against a 3.4m thirty-day average, a 1.1 times volume multiple. The S&P 500 rose 1.5% that session, so the relative move was about minus 1.3 points.
• Tokyo reaction session, August 4. The ordinary shares closed at ¥3,470, down 2.72%, against a Nikkei 225 up 0.32%. Relative move: about minus 3.0 points. The ADR converged the same day, closing at $21.83, down 2.94%.
• Two-week aftermath. From the August 3 pre-release close through August 17, the ordinary shares rose 2.30% to ¥3,649 while the Nikkei 225 rose 8.57%. Over the same window Sumitomo Mitsui Trust rose 6.27%, Mizuho 5.39% and Sumitomo Mitsui Financial Group 4.60%. MUFG was the weakest of the four.

The single-session reaction was a mild negative and the two-week reaction was a clear one. A 28% beat, a capital ratio repaired ahead of schedule and a record first-quarter profit produced a stock that has lagged its home index by roughly six points and every one of its direct peers in the fortnight since. That is not a market that disbelieved the numbers. It is a market that had something else in mind.

Sell the news, and the news was the wrong kind. The ADR entered the print up 41.6% year to date and 8.9% in the preceding month, so a great deal was priced. More specifically, the composition of the beat is exactly what a sceptical holder discounts: a ¥96bn swing in treasury operating profit, ¥103.5bn of equity-sale gains into a record domestic equity market, ¥222.2bn from a stake in a US investment bank, and roughly ¥75bn of gross profits from a 12% weaker yen. Strip those and the result is very good rather than extraordinary, and very good was already in the price.

The peer comparison is the sharper explanation. Japan's megabanks reported within days of one another. Mizuho reported on July 30, raised its full-year forecast from ¥1.30tn to ¥1.40tn, and simultaneously doubled its buyback authorization from ¥100bn to ¥200bn, extending the window and moving the cancellation date. Sumitomo Mitsui Financial Group reported a 33% increase in quarterly profit, reiterated its outlook, and is running a ¥180bn repurchase authorized in May alongside a two-for-one stock split with a September record date. MUFG posted the largest profit growth of the three and was the only one of the three with nothing new to announce: no raise, no incremental buyback, and its own ¥100bn programme already finished and out of the market since June 25. In a sector being re-rated on capital return, having the best quarter and the emptiest announcement is a recipe for relative underperformance, and that is what happened.

What the price is now saying. At ¥3,649 the ordinary shares trade at 1.81 times the ¥2,016.68 of book value per share reported at June, against 1.52 times when we initiated in May. The market has done the re-rating that our May work said was the smaller half of the return. What it has not done is price a second-half buyback, which remains the one identifiable catalyst between here and the March year end.

Street Perspective

Debate: does leaving the target unchanged show discipline or a missed opportunity?

Bull view: A Japanese bank does not revise a full-year target on one quarter, and a first quarter this heavy in treasury and disposal gains is the worst possible basis for extrapolation. Keeping the target intact preserves the credibility of the "most likely outcome" framework and leaves room to raise at the interim with two quarters of evidence rather than one. Peers that raised early can be forced to raise again or, worse, to hold.

Bear view: Management explicitly abandoned conservatism in May and said the target reflects the most likely outcome. Thirty percent progress in the smallest quarter, with credit costs at a fifth of budget, moves the most likely outcome. A framework that revises down on deterioration but not up on outperformance is the old conservatism in new language, and the market has stopped paying for it.

Our take: The bear has the better of this. The company set the terms of its own framework and then did not apply them. Silence would be defensible from a management that had never made the May statement. The practical consequence is small, because a target is not an estimate, but the credibility consequence is the reason the stock underperformed peers who said something.

Debate: how much of the quarter is repeatable?

Bull view: The customer segments produced 74% of the increase in segment operating profit with the customer expense ratio down five points, every one of six groups grew, and the domestic spread widened another 21 basis points year on year. The Bank of Japan's June move contributed almost nothing to this quarter and is worth roughly ¥95bn still to come. This is a franchise result, not a markets result.

Bear view: Treasury swung ¥96bn, equity-sale gains tripled into a record equity market, the Morgan Stanley stake supplied 27.5% of net profit, and a 12% weaker yen added about ¥75bn of gross profits. On the company's own currency-neutral view the customer segments grew ¥153.4bn rather than ¥205.8bn. Sales and trading, the repeatable half of Global Markets, grew 25%, which is good and is a fraction of what the headline implies.

Our take: Both descriptions are accurate and the reconciliation is that the franchise result and the windfall are both real and roughly equal in size. The franchise number to underwrite is the customer segments' currency-neutral ¥153.4bn of operating-profit growth, or about 33% on the same basis, together with 25% growth in sales and trading. That is a very good bank growing quickly. It is not a bank that earns ¥3.2tn annualized, which is what a naive extrapolation of this quarter produces.

Debate: is the capital constraint over?

Bull view: The ratio is back at the bottom of the target range one quarter ahead of guidance, having absorbed the whole ¥100bn buyback on the way. Common equity tier 1 capital grew 8.0% against risk-weighted assets up 3.9%. Capital generation of that magnitude funds both growth and a materially larger second-half repurchase, and the condition management set for that decision has been satisfied.

Bear view: Risk-weighted assets grew ¥4.7tn in three months, an annualized pace faster than the 12.5% that broke the ratio last year, and overseas lending grew ¥1.7tn when it was meant to run off. The ratio recovered on an earnings quarter that will not repeat at this level. Being at 9.5% is being at the floor, not inside the range, and the capital-policy answer explicitly puts growth investment ahead of repurchases.

Our take: The constraint has loosened, not ended. Both facts are true at once: the commitment was delivered early, and it was delivered by the numerator rather than the denominator. A second-half authorization is now more likely than not, but sizing it toward the ¥400bn to ¥500bn era requires balance-sheet growth to moderate, and there is no evidence yet that it has.

Debate: does 1.8 times book still work?

Bull view: Trailing twelve-month earnings are ¥2,690.6bn, so the shares trade at roughly 15 times an earnings stream that has already reached the company's full-year target with the rate cycle barely started. Book value per share compounded 2.2% in a single quarter. A bank sustainably earning 12% or better on equity deserves a premium multiple, and a policy rate at 1% against a plausible neutral well above it is free optionality nobody is paying for.

Bear view: This is a stock that traded below book for most of a decade, has risen 22% in three months, and is now valued at 1.81 times book on a quarter whose quality is the subject of the debate above. The published sell-side consensus target sits below the current price. The re-rating from 1.5 to 1.8 times happened without a single new disclosure about the two things that matter, which are second-half capital return and the return on the capital being consumed.

Our take: We said in May that the re-rating was not the opportunity and the earnings growth was. The re-rating has now happened anyway, and it has happened faster than the earnings. On a sustainable return of 12% to 13% against a cost of equity of 8.5% to 9%, the justified multiple is roughly 1.5 to 1.75 times book, and the shares are at the top of that. The business is better than we underwrote and the price is ahead of it.

Model Update

The operating assumptions move up and the valuation conclusion moves down. Both are consequences of the same quarter.

Valuation. At the August 17 close of ¥3,649 the shares trade at 1.81 times the ¥2,016.68 of book value per share reported at June, 15.3 times trailing twelve-month earnings of ¥239.0 per share, and 15.2 times the ¥239.9 implied by the company's own full-year target. The dividend yield on the guided ¥96 is 2.63%. The ADR at $23.03 corresponds to ¥3,672 per ordinary share at a dollar-yen rate of 159.43, a premium of about six-tenths of a percent to the Tokyo close and within the normal basis for this line.

Valuation framework, unchanged in method. We anchor on price to book against sustainable return on equity. Our May work used approximately 12% sustainable ROE, a 9% cost of equity and 2% long-term growth for a justified range of 1.4 to 1.6 times book. This quarter argues for lifting the return assumption to 12.5% to 13%, given the customer franchise carrying the growth and the rate cycle only beginning to land, and for a cost of equity toward the lower end of the plausible Japanese range at 8.5%. That produces a justified multiple of roughly 1.5 to 1.7 times. Applied to an estimated March 2027 book value of about ¥2,100, the twelve-month value range is ¥3,150 to ¥3,570, with a midpoint near ¥3,360. Against the ¥3,649 close that midpoint is about 8% below the current price, and a 2.6% dividend yield takes the expected twelve-month total return to roughly minus 5%. The ADR equivalent of the midpoint is approximately $21.10 at the prevailing exchange rate, with currency a first-order risk in that translation rather than an afterthought.

What would change it. A second-half buyback of ¥300bn or more announced at the interim, a full-year target raise, or a de-rating back toward 1.5 times book, which is roughly ¥3,150, would each restore the asymmetry we underwrote in May. On the other side, a second-half authorization at or near ¥100bn again, or a further quarter of risk-weighted asset growth at this pace, would take the justified multiple back toward 1.5 times and make the current price harder to defend.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation and carried in the standing thesis. They are graded against what this quarter's disclosure showed, not re-derived.

Overall: the thesis is stronger than it was in May. Four of the seven pillars improved on the quarter, one bear point was pushed back for the first time, and the capital commitment that was the largest open question was delivered early. The two that did not improve are the two that concern how the balance sheet is grown and where the earnings come from, and neither is new.

Action: hold what you own; do not add here. The rating change is a valuation judgment and not a verdict on the business. The shares have moved from 1.52 to 1.81 times book in three months while our estimate of fair value moved much less, so the margin of safety that justified building a position in May has been spent. The two events that would restore it are a second-half buyback of ¥300bn or more at the November interim, and disclosure at the same interim that the businesses consuming the capital are earning an adequate return on it. Both arrive on the same day, and that day is the next real checkpoint in this name.