A Record Half Built Below the Operating Line: Initiating Coverage of MUFG at Outperform

Results vs. Consensus

MUFG reports under Japanese GAAP in yen, publishes a full-year target rather than a forecast, and gives no half-year or quarterly guidance at all. The half is therefore scored against three yardsticks: the pre-print sell-side poll, progress against the company's own full-year target, and the prior-year half on both of the two bases the company publishes.

Interim scorecard

Year-on-year, on both bases the company publishes

Every line below carries two comparatives. The as-reported column ties to the income statement. The adjusted column removes the effect of the change in Krungsri's consolidation closing period from a January–December year to an April–March year, which loaded an extra quarter of the Thai subsidiary into the six months to September 2024. MUFG sizes that distortion at ¥79.6bn of operating profit and ¥22.1bn of net profit. Management speaks exclusively to the adjusted column.

Second quarter, derived

MUFG does not publish a standalone second quarter. The July–September figures below are the interim result less the three months to June 30, taken from the two consolidated summary reports. No standalone quarterly EPS is derived here, because the average share count behind the quarterly and half-year denominators differs.

Revenue assessment

The revenue story is better than the reported 0.8% suggests and worse than the adjusted ¥189.3bn implies, because the adjusted figure still carries the acquisitions. Strip the roughly ¥48bn acquisition effect out of the fee line and organic revenue growth on the adjusted basis is closer to ¥141bn, or about 5%. That is a respectable number for a bank of this size, and its source is the right one: the domestic loan book repricing upward as the Bank of Japan normalises. The domestic loan rate excluding government lending rose 27bp to 1.16% while the deposit rate rose only 14bp to 0.17%, so the spread on the part of the book that actually matters widened 12bp to 0.99%. On the quarterly series the second quarter alone printed a 1.18% lending rate against a 0.19% deposit rate.

The cleanest single evidence of transmission is at MUFG Bank. Its net interest income grew from ¥807.3bn to ¥842.7bn, an unremarkable 4.4%. But last year's figure contained ¥84.6bn of gains on the cancellation of investment trusts and this year's contains ¥0.2bn. On a like-for-like basis the Bank's net interest income grew 16.6%. The company discloses this in a footnote and did not mention it on the call, which is a pity, because it is the most encouraging number in the release.

Margin assessment

The call described the expense ratio as flat. The company's own slide shows 55.1% moving to 56.1%, up 0.9ppt as reported and 0.7ppt adjusted. Both statements can be reconciled only by rounding generously, and the disclosure is the one to trust. What matters more than the ratio is where the spend went: retail and digital absorbed ¥51.2bn of additional expense against ¥58.7bn of additional revenue, a 12.8% flow-through to operating profit, while asset management and investor services absorbed ¥36.8bn against ¥47.0bn. Those two groups account for ¥88.0bn of the ¥127.7bn of customer-segment cost growth and produced ¥17.7bn of the ¥85.0bn of customer-segment operating-profit growth. The rest of the franchise converted at a far better rate.

Earnings assessment

Reported return on equity of 12.5% is flattered. Strip the ¥100bn of one-time gains MUFG itself identifies, and tax the ¥130.2bn of equity-securities gains at the half's 23.0% effective rate, and the underlying half is closer to ¥1,092.6bn, an annualised return on equity near 10.6%. That is the number to carry into a valuation, and it is the number management is implicitly acknowledging when it defines its 12% target as a figure achieved with a 1.0% policy rate and no equity-disposal gains at all. The gap between 10.6% today and 12% later is the investment case, and the print gave real evidence that the bridge is being built.

Segment Performance

MUFG publishes business-group results on a managerial-accounting, local-currency basis, with the prior-year comparative for global commercial banking adjusted to remove the Krungsri effect. That basis does not tie to the consolidated income statement, and the gap is currency: the deck's own operating-profit waterfall puts the customer-segment increase at ¥93.6bn including roughly ¥10bn of foreign-exchange benefit, while the tables below sum to ¥85.0bn because they are currency-neutral. Both are the company's figures. The table is labelled accordingly and the two numbers are not interchangeable.

The headline fact in this table is that five of the six customer groups grew operating profit and the sixth is a Thai and Indonesian problem, while Global Markets went backwards. That is the shape a rate-normalisation thesis wants. The second fact is that risk-weighted assets for the six customer groups barely moved: Japanese corporate banking held at ¥31.2tn, global commercial banking at ¥7.5tn, and global corporate banking rose ¥0.8tn to ¥24.2tn. Group return on equity improved everywhere, in several cases by 2.5 to 4.0 percentage points, and it improved on stable denominators.

Entity view: where the profit actually comes from

Retail & Digital

Revenue growth of 13.0%, third-fastest of the six customer groups, and the worst conversion. Loan and deposit interest income rose ¥31.6bn and card settlement ¥16.6bn, both clean rate-and-volume outcomes, but expenses rose ¥51.2bn and the expense ratio moved to 74% from 72%. Net income nonetheless rose ¥29.5bn to ¥66.7bn and group return on equity went from 7.0% to 11.0%, which the disclosure attributes to one-time gains associated with equity investments rather than to operating flow-through.

Assessment: the retail build is the largest deliberate cost programme in the group and it has now run two halves without publishing any unit economics. An 11.0% return on equity that rests on investment gains is not the same as an 11.0% return on equity that rests on the build working. This is the group to watch and the group where disclosure is thinnest.

Commercial Banking & Wealth Management

The standout. Gross profits rose ¥61.8bn, of which ¥57.2bn came from loan and deposit interest income alone, while expenses rose only ¥12.7bn. The expense ratio collapsed seven points to 56% and operating profit rose 39.9%. Return on equity went from 11.5% to 15.0% on a risk-weighted asset base that grew ¥0.5tn. Investment product sales fell ¥4.3bn and real estate and inheritance fees fell ¥2.7bn, so none of the growth came from market-sensitive lines.

Assessment: this is what domestic rate normalisation looks like with no offsetting cost build. If the thesis is that MUFG converts policy rate into profit, this group is the proof, and it is the group with the highest operating leverage left if the Bank of Japan continues.

Japanese Corporate & Investment Banking

Gross profits rose ¥25.4bn on broad contribution: derivatives and solutions +¥7.5bn, real estate and corporate agency +¥5.7bn, advisory and capital markets +¥4.3bn, loan and deposit interest income +¥9.1bn. Expenses rose ¥9.7bn and the expense ratio held at 39%, the best in the group. Net income rose ¥43.2bn, well ahead of the ¥15.8bn of operating-profit growth, on contained credit costs. Return on equity reached 15.5% on essentially flat risk-weighted assets of ¥31.2tn.

Assessment: the highest-quality result in the table, because the revenue is diversified and the balance sheet did not grow to produce it. The one caveat is that the net-income growth is credit-cost-driven and credit costs are at a cyclical trough.

Global Commercial Banking

The only customer group to shrink. Gross profits fell ¥8.9bn, entirely at Krungsri, whose gross profits fell from ¥253.4bn to ¥244.5bn on weaker Thai lending income. Expenses rose ¥2.9bn and the expense ratio moved to 57% from 54%. Operating profit fell ¥11.9bn. Net income nonetheless rose ¥11.0bn to ¥60.0bn on tighter credit screening at the partner banks.

"Credit costs also decreased at our overseas subsidiaries due to the effect of stricter screening criteria for new credit transactions in Asian partner banks." — Jun Togawa, Group CFO

Assessment: a revenue problem masked by a provisioning benefit. Krungsri's loan balance grew ¥0.6tn to ¥7.0tn while its revenue fell, which is the wrong combination. Asia is the part of the story that is not working, and management named the regional slowdown as the cause rather than anything company-specific.

Asset Management & Investor Services

Gross profits rose ¥47.0bn, of which ¥40.3bn came from investor services alone, reflecting the bundled-services expansion and the Link acquisition. Expenses rose ¥36.8bn, so operating profit gained only ¥10.2bn and the expense ratio rose to 72%. Net income of ¥49.5bn rose ¥5.2bn. Return on equity improved to 14.0% on economic capital that shrank slightly.

Assessment: the growth is real but almost all of it is being spent. A 78% cost-to-revenue-growth ratio in a business MUFG has named as one of three inorganic priorities is a warning that the acquisitions have not yet been integrated. The disclosure that outsourcing operations reached the mid-term plan target early is encouraging in isolation.

Global Corporate & Investment Banking

Commission income rose ¥27.2bn and was almost all of the group's ¥28.7bn revenue growth, since loan and deposit interest income actually fell ¥2.1bn. That is precisely the originate-and-distribute shift management has been describing. Expenses rose ¥14.4bn with the ratio flat at 52%, and net income rose 55.9% to ¥147.0bn on the rebound from last year's large overseas credit provisions. Return on equity went from 8.0% to 11.5%.

Assessment: the best evidence in the release that the capital-efficiency programme is working, because fee income grew while lending income fell and risk-weighted assets grew only ¥0.8tn. The 55.9% net-income jump is a provisioning artefact and should not be extrapolated. The fee shift should be.

Global Markets

Gross profits fell ¥4.8bn and operating profit fell ¥8.4bn. The split matters: sales and trading operating profit fell ¥15.9bn on reduced liquidity in bond markets, while treasury operating profit rose ¥2.9bn as last year's bond-portfolio rebalancing began paying. Net income still rose ¥6.4bn to ¥149.5bn, and return on equity edged to 10.5% on shrinking economic capital.

"In Q1, reducing the balance of super long-term JGBs, partially offsetting with redemption gains on bear fund and gains on sale of foreign bonds, that's for the first half of the year." — Jun Togawa, Group CFO

Assessment: the treasury book has turned from the group's biggest liability into a modest contributor, which is the mechanical consequence of taking the losses last year. Sales and trading is the weak spot and it is a market-condition problem rather than a franchise problem. The bigger news inside this group is balance-sheet shape, covered below.

Key Topics & Management Commentary

Overall Management Tone: Measured and unusually precise, with the call run almost entirely by the group chief financial officer working line by line through the slides. There was no attempt to lead with the record headline, and the one moment of visible defensiveness came on inorganic investment, where a press report was declined without comment. Management volunteered its second-half caution before being asked and conceded internal debate about whether the guidance revision was worth making at all, which reads as candour rather than hedging.

1. The two year-on-year columns, and why the operating line looks worse than the business

Every comparative in this release exists twice. The change in Krungsri's consolidation closing period loaded an extra quarter of the Thai subsidiary into the six months to September 2024, worth ¥165.4bn of gross profits, ¥85.7bn of expenses, ¥79.6bn of operating profit and ¥22.1bn of net profit. Remove it and operating profit rose 5.0% rather than falling 1.4%, and net profit rose 4.6% rather than 2.8%. The as-reported column is the one that reconciles to the income statement; the adjusted column is the one that describes the business.

Assessment: an investor who reads only the income statement concludes that MUFG's operating profit is shrinking. An investor who reads only the deck concludes it is growing 5%. Both are the company's own disclosure and the difference is a one-off accounting change that annualises out after this half. Carry the adjusted column for trend and the reported column for arithmetic.

2. Domestic rate transmission is finally showing up in the spread ladder

The domestic business segment at MUFG Bank shows the average loan rate rising 27bp to 1.09%, or 27bp to 1.16% excluding government lending, against a deposit and negotiable-certificate rate up 14bp to 0.17%. The loan-deposit spread excluding government loans widened 12bp to 0.99%. On the quarterly series the second quarter alone printed 1.18% against 0.19%. The pass-through ratio, roughly half of the loan-rate move retained, is the number that determines how much of the remaining policy path converts into profit.

"Spreads for large corporates in red line is rising, thanks to the accumulation of large, highly profitable loans. Along with SMEs in orange, profit improvement measures have been successful, and the upward trend is continuing." — Jun Togawa, Group CFO

Assessment: this is the single most important disclosure in the release and it is buried in a supplement. A balance sheet 56% funded by deposits, keeping about half of each loan-rate increase as spread, with the policy rate at 0.5% and management's own return target assuming 1.0%, is a multi-year earnings tailwind that has barely started.

3. The ¥100bn target raise is the same size as the one-time gains already booked

The full-year target for profits attributable to owners of parent went from ¥2,000.0bn to ¥2,100.0bn, a 5.0% revision, against a half that delivered 64.6% of the original target. Management's own decomposition of the raise: roughly ¥30bn from the customer segment, roughly ¥40bn from one-off gains that were not in the plan, and roughly ¥30bn from revised financial-indicator assumptions, mainly a weaker yen. The one-time gains MUFG flags inside the half total approximately ¥100bn.

"There was internal discussion about whether a 5% revision was really necessary, but we decided to do so with the aim of disclosing our forecast appropriately at each point in time since the first half of last year. We may not have done this in the past, but that is our line of thinking." — Jun Togawa, Group CFO

Assessment: only about ¥30bn of the ¥100bn raise represents an improved view of the operating business, and even that is described as a full-year figure. This is a guidance philosophy that revises for what has already happened rather than for what is expected, which sets a low bar for the second half and a high bar for believing the target is a genuine ceiling.

4. Capital return at a record size, with the composition doing the work

Three separate actions landed together. The interim dividend was ¥35.00 against ¥25.00 last year, and the year-end forecast was raised from ¥35.00 to ¥39.00 for an annual ¥74.00, up ¥10 year on year and ¥4 against the initial forecast. A second-half buyback of up to ¥250bn and 130m shares, running from November 17 to February 27, takes the full year to ¥500bn, the largest in the company's history and above the ¥400bn run-rate of the prior two years. And 200m treasury shares, 1.66% of shares outstanding, are cancelled on November 28.

"Regarding share repurchase, a resolution was approved today to acquire an additional JPY 250 billion in the second half of the year, bringing the total amount for the full year to JPY 500 billion. As discussed in May, this is due to take into account total shareholder return over the past few years. We also announced today the cancellation of 200 million treasury shares." — Jun Togawa, Group CFO

Assessment: the total payout ratio reaches 64.0% against 61.3% last year and a dividend payout of 40.2%, and the cancellation is the piece that makes the buyback permanent rather than a warehouse. The one blemish is relative: the absolute number is a record, and it was still the smallest upgrade of the three megabanks on the same day.

5. Capital sits at the top of the range, and the second-half walk barely holds it

The ratio that matters for the mid-term plan is the common equity tier 1 ratio on the finalised and fully implemented Basel III basis, excluding unrealised gains on available-for-sale securities. It fell 30bp over the half to 10.5%, at the very top of the 9.5% to 10.5% target range. The regulatory ratio including those gains fell 10bp to 14.08%, a figure that should never be read against the target range. The walk: earnings +1.3ppt, dividends −0.4ppt, buyback −0.3ppt, growth investment and risk-weighted asset increase −0.5ppt, currency and other −0.4ppt.

"The CET1 ratio, excluding unrealized gains on the finalized and fully implemented Basel III basis fell 30 basis points from the end of March to 10.5% at the upper end of our target range due to growth investments and increase in loans, as well as yen appreciation versus end of March." — Jun Togawa, Group CFO

Assessment: management's stated second-half arithmetic is +80bp of earnings, −65bp of returns and −30bp of risk-weighted asset growth, a net −15bp that lands at 10.35% before Morgan Stanley's retained profit. Guiding to a range of 10% to 10.5% off that walk means the midpoint of the guide is below the midpoint of the arithmetic, and any overshoot in risk-weighted assets or any move in the yen takes the ratio below the top of the target band. The capital condition is comfortable today, and a large inorganic deal would have to fit inside the 50bp to 100bp between the guided 10% to 10.5% and the 9.5% floor.

6. Risk-weighted assets grew two and a half times faster than loans

Risk-weighted assets rose ¥3,877.5bn, or 3.6%, in six months, of which ¥3,546.1bn was credit risk. Loans on the consolidated balance sheet rose 1.5%. Overseas loans grew ¥2.0tn, or ¥1.3tn excluding currency, with the Americas up ¥0.7tn, Europe and the Middle East up ¥0.7tn and Krungsri up ¥0.6tn. Domestic loans were flat in total only because ¥2.5tn of government lending ran off while large-corporate lending grew ¥1.8tn and lending to small and medium companies grew ¥0.6tn.

Assessment: the balance sheet is being worked harder, not just larger. Common equity tier 1 capital grew 2.9% while risk-weighted assets grew 3.6%, which is why the regulatory ratio fell despite record earnings. Additional tier 1 capital grew ¥461.1bn, more in absolute terms than common equity did, so over half of the tier 1 capital increase came from issuance rather than retention. That is a legitimate tool and it is also a signal about how tight the organic capital generation is against the growth ambition.

7. Morgan Stanley is now 22% of the group's profit

Equity in earnings of equity-method investees rose ¥124.7bn to ¥381.9bn, which is 21.9% of ordinary profits. On the entity breakdown, Morgan Stanley alone contributed ¥287.9bn of the ¥1,292.9bn interim net profit, or 22.3%. The company attributes the increase to "the extremely strong performance of Morgan Stanley" and, when asked to size the contribution inside the guidance raise, declined to do so.

Assessment: this is the largest single unhedged exposure in the earnings model and it is a US capital-markets cycle, not a Japanese rate cycle. It is a genuine asset, the alliance dates from 2008 and is structurally advantaged, but an investor buying MUFG for domestic rate normalisation is getting roughly a fifth of the earnings from a business with completely different drivers. It deserves a discount, not a premium, in any sum-of-the-parts.

8. Credit costs are a reversal, and the full-year target was left alone

Total credit costs were a ¥76.3bn charge against ¥185.7bn last year. The composition is unusual: MUFG Bank on a non-consolidated basis recorded an ¥82.3bn net reversal, offset by ¥56.9bn at the consumer finance subsidiaries and ¥100.9bn overseas. The non-performing loan ratio fell to 1.01% from 1.11%, the lowest in the disclosed series, with the improvement concentrated in doubtful loans, down ¥91.6bn, and restructured loans, down ¥30.6bn. Coverage rose to 80.6%. The full-year target of ¥350.0bn was explicitly left unchanged.

"Taking the current situation into account, we kept our full year outlook for credit costs unchanged." — Jun Togawa, Group CFO

Assessment: holding a ¥350bn full-year target after a ¥76.3bn first half implies ¥273.7bn in the second, 3.6 times the run-rate. Either management sees something it has not described, or this is a deliberate reserve against a target it intends to beat. Given the quality of the asset-quality disclosure, the second reading is more likely, and it is the largest identifiable cushion inside the full-year number.

9. The balance sheet shrank ¥8.7tn, and almost nobody asked why

Total assets fell from ¥413.1tn to ¥404.3tn. Three moves explain it. Borrowed money fell ¥10.5tn as the Bank of Japan's growth-supporting funding facility ran off. Deposits with the Bank of Japan fell ¥16.7tn. And available-for-sale Japanese government bond holdings fell ¥4.96tn, with domestic bonds maturing within one year down sharply as the collateral requirement disappeared. Foreign bond holdings rose ¥4.21tn in the same window, and the foreign bond book's average duration shortened from 4.5 years to 4.1 years while the domestic book's lengthened slightly to 2.3 years.

"Short-term JGBs decreased as the BOJ's growth-oriented lending support operation is gradually coming to an end and need for short-term JGBs as collateral has decreased. The balance of short-term government bonds has fallen significantly." — Jun Togawa, Group CFO

Assessment: a large, low-yielding, policy-driven balance sheet is unwinding and being partly redeployed into foreign bonds at a shorter duration. The economics are favourable, because the assets leaving earn almost nothing. The disclosure is thin, and the securities book's held-to-maturity unrealised loss deepened to ¥632.3bn with no commentary at all.

10. The equity-disposal programme is halfway done and the profit base after it is undefined

Cumulative sales during the current mid-term plan reached ¥339bn on an acquisition-cost basis, of which ¥63bn came in this half against ¥276bn in the prior year. Including agreed but unsold positions, the expected total rose to ¥539bn against a ¥700bn target. Meanwhile the domestic equity book's carrying value grew to ¥3.83tn and its unrealised gain grew ¥0.36tn to ¥2.82tn, because market appreciation outran disposal.

"the cumulative sales during the current MTBP were JPY 339 billion on an acquisition cost basis, which is about half of the JPY 700 billion target. The agreed amount has reached nearly 80% of the target, and we are making steady progress toward achieving this target." — Jun Togawa, Group CFO

Assessment: the disposal pace fell from ¥170bn in last year's comparable half to ¥63bn, while ¥134.3bn of sale gains ran through the income statement. When the programme completes, that line stops, and management has been explicit that the 12% return target assumes it has already stopped. The number investors need, and have not been given, is what the profit base looks like in the first year without it.

11. The 12% return target, put in writing for the first time

The deck's return-on-equity page states the two assumptions explicitly: a Bank of Japan policy rate of approximately 1.0%, and no gains from the sale of equity holdings. The path is built from mid-term-plan growth strategies, the rate uplift, organic growth in Japan and Asia, and inorganic growth in three named areas. Cumulative operating-profit growth across the plan is put at ¥500bn, with Asia business profit rising 1.6 times and originate-and-distribute fee income doubling.

"we originally began the discussions to set the 12% target by trying to see how much we can increase our profit under the assumptions that Japan's policy interest rate will be around 1% and that we have no gain on sale of equity holdings, which I strongly insisted." — Jun Togawa, Group CFO

Assessment: putting the assumptions in writing is a governance improvement and it also sets a trap. Against an underlying return on equity of roughly 10.6% today, 12% requires both the rate assumption and the growth plan to land. The far more useful disclosure was the follow-on: that potential investments are now screened on whether they contribute to the 12% target. That is a capital-allocation discipline statement, and it is new.

12. Retail brand, digital bank and an OpenAI collaboration

The retail brand launched in June 2025 produced credit-card issuance up roughly twofold year on year for the June to September window and roughly threefold against two years ago, new account openings up roughly 1.2 times, and brokerage account openings under the wholly owned eSmart Securities up roughly 11.5 times. New services including a digital bank are scheduled for FY2026. Separately, implemented artificial-intelligence use cases reached 116 with a target above 250 by FY2026, and expected cumulative benefit across the three-year plan is put at ¥30bn.

"The number of AI use cases has reached 116, and the aim is to increase to over 250 cases by FY '26. Current estimates suggest that the cumulative benefits over the 3 years of the current MTBP is approximately JPY 30 billion. The launch of a new strategic partnership with OpenAI is expected to accelerate use of AI across the company and to collaborate on various services, primarily in the retail sector such as digital banking." — Jun Togawa, Group CFO

Assessment: ¥30bn of cumulative benefit across three years is roughly 1.3% of this year's ¥2,250bn operating-profit target, which is a modest claim honestly made rather than an artificial-intelligence narrative. The retail multiples are impressive as ratios and meaningless as economics until a revenue or cost number is attached to them. The digital bank is not scheduled to launch until FY2026, which means the retail cost line carries another year before it earns anything.

Guidance & Outlook

The implied second-half column is derived by subtraction and is not a company disclosure. It is also the most informative part of the guide. Second-half net profit of ¥807.1bn would be 37.6% below the first half and 33.4% above the second half of last year. Second-half operating profit of ¥963.0bn would be 25.2% below the first half. And second-half credit costs of ¥273.7bn would be 3.6 times the first-half charge.

Implied half-over-half ramp: management walked through its own version of this before being asked. Foreign exchange contributes roughly ¥25bn of operating profit, on an assumption that the yen strengthens by about ¥5. Treasury trading gains were front-loaded, and the first-half-to-second-half difference is put at roughly ¥130bn. Expenses rise by roughly ¥100bn on the retail brand, information technology, artificial intelligence, cyber security, inflation-related costs and base wage increases. Netting those gives the roughly ¥50bn of full-year operating-profit upside management described.

"Within NOP, JPY 25 billion is from ForEx, assuming the yen to be about JPY 5 stronger. The rebound from treasury trading gains was concentrated in the first half, as I said, and the difference between first half and second half is about JPY 130 billion. Then there is increase in expenses, expense incurred in EMUTO, IT costs, AI, cyber-related impact from certain inflation-related costs, base wage increase, among others. All in all, about JPY 100 billion in expense increase. We are also considering a certain level of structural improvements for next fiscal year as profits are also strong. Averaging them all out, we expected an upside of about JPY 50 billion in NOP." — Jun Togawa, Group CFO

Where the assumptions sit: the revised target assumes a Bank of Japan policy rate of approximately 0.5%, a Fed funds rate of approximately 4%, a Nikkei 225 in the high ¥40,000s and a dollar-yen rate in the mid-¥140s. The last of those is the one to watch. The yen closed the half at 148.88 and the release landed with spot well weaker than the assumption, so the currency line in the second half is set up to be a source of upside rather than the drag the plan carries.

Guidance style: conservative, and newly so. Management said explicitly that it debated whether a 5% revision was worth making, and that the practice of revising at each disclosure point began only last year. The line "We are also considering a certain level of structural improvements for next fiscal year as profits are also strong" is the tell: costs are being pulled forward into a strong year. An investor should treat ¥2,100bn as a floor with a cushion rather than a forecast, while noting that the same conservatism removes any chance of a positive surprise being credited to the operating business before the fourth quarter.

Analyst Q&A Highlights

Whether a 5% revision was worth making at all

The opening question of the call went straight at the size of the raise, on two fronts: that the underlying market assumptions looked stale against where the Nikkei and the currency were actually trading, and that ¥100bn was a small revision after a half that had already delivered 64.6% of the original target. Management's answer conceded the internal debate rather than defending the number, and then separated the two objections: the currency assumption is genuinely conservative, the equity-market assumption is nearly irrelevant to earnings.

Q: "I would like to hear your thoughts on the upward revision from two perspectives. First, I wonder if the assumptions are too conservative considering the current levels of the Nikkei stock average and the dollar-yen exchange rate. Second, the revision of JPY 100 billion from JPY 2 trillion to JPY 2.1 trillion is not small, but it is a somewhat small revision to your bottom line profit. What was the aim and your thoughts on this small revision?"
— Ken Takamiya, Nomura Securities

A: "There was internal discussion about whether a 5% revision was really necessary, but we decided to do so with the aim of disclosing our forecast appropriately at each point in time since the first half of last year. We may not have done this in the past, but that is our line of thinking. Regarding the assumptions, the yen assumption against the dollar is quite strong given the current level. But depending on interest rate trends, it is not unreasonable for the yen to be in the mid-JPY 140s by the end of the fiscal year. The share price of around JPY 43,000 may also seem conservative, but the impact of share prices on our earnings is not significant. So this was not the reason for the conservative profit target."
— Jun Togawa, Group CFO

Assessment: the admission that the revision was debated internally is more useful than the revision itself. It confirms that ¥2,100bn is a disclosure-hygiene number rather than a best estimate, and that the company would have been comfortable saying nothing. For a modeller, the practical consequence is that the target carries an undisclosed cushion whose size management has now implicitly acknowledged.

The return target's assumptions, stated in writing for the first time

A follow-up pressed on why the 12% return-on-equity ambition had suddenly appeared on a slide with explicit assumptions attached, and whether the environment had changed management's thinking. The answer revealed that the assumptions had always been the internal basis, and that the decision to publish them was a response to investors modelling the target on more generous foundations. The more consequential part of the reply came at the end and was not asked for.

Q: "On Page 6, you explained verbally the general direction you are heading, including assumptions like interest rate of around 1% and no gain on sale from reducing your equity holdings. But I think this is the first time you have clarified this in writing. Regarding the mid- to long-term ROE target of 12%, I want to know if there were any changes in your thinking and the management's perspective, reflecting the changes in the environment or tailwinds."
— Ken Takamiya, Nomura Securities

A: "we originally began the discussions to set the 12% target by trying to see how much we can increase our profit under the assumptions that Japan's policy interest rate will be around 1% and that we have no gain on sale of equity holdings, which I strongly insisted. Since investors asked questions based on different assumptions such as including gain on sale of equity holdings, we made that clear. … One change in our thinking, both in terms of inorganic investment and the use of capital, as I may have mentioned before, is that we are now discussing potential investments internally based on whether or not they contribute to achieving 12% ROE."
— Jun Togawa, Group CFO

Assessment: the investment-screening comment is the new information and it should reduce the market's discount for deal risk. A management team that screens acquisitions against a return target is less likely to make the sort of purchase that destroyed value at the US subsidiary a decade ago. It is a statement, not yet a track record.

The capital path to March and whether the ratio really reaches the middle of the range

A direct challenge to the capital slide: the stated intention is for the ratio to settle around the midpoint of the target range by the fiscal year end, and the disclosed trajectory did not look capable of getting there. Management responded with a component-by-component walk rather than a restatement of the intention, and the answer landed on a range that sits materially above the midpoint it had described.

Q: "First, let me confirm the full year CET1 ratio forecast on Page 20 again. It doesn't seem like it will approach the middle of the range. So if you could share with us your view on the level and the breakdown to the extent possible."
— Shinichiro Nakamura, BofA Securities

A: "First, regarding the outlook for CET1 ratio toward the end of FY '25, the end of March '26, approximately 80 basis points up in the second half from the accumulation of net income based on the revised performance targets, 65 basis points down due to shareholder returns, including dividends and share buybacks, as I explained earlier, around 30 basis points down from the planned increase in risk assets. And with Morgan Stanley's accumulated profit from its extremely strong performance, et cetera, we expect the ratio to be somewhere between 10% and 10.5%."
— Jun Togawa, Group CFO

Assessment: the question was right and the answer conceded it. The arithmetic given nets to minus 15 basis points from 10.5%, which is 10.35%, not the midpoint of a 9.5% to 10.5% band. Management is guiding to a ratio at the top of its own range, which means capital is not the constraint on returns this year, and also that the buffer to absorb a large risk-weighted asset build or an adverse currency move is thinner than the headline ratio suggests.

Private credit, data-centre lending and concentration risk

The most topical question of the call, and the one where the answer was shortest. The line of questioning connected widening credit spreads in US private credit to large syndicated financings for hyperscale data centres and asked how the bank thought about the resulting concentration. Management denied material exposure, pointed at the falling non-performing loan ratio as evidence, then answered a question that had not quite been asked about how project finance risk is underwritten.

Q: "So my question is on the current situation of private credit in the U.S. Although MUFG has not directly mentioned it, we are seeing large-scale loans to Oracle's data center investment, among others, which is widening credit spreads as a result. What are your thoughts on this increasing concentration of risk?"
— Shinichiro Nakamura, BofA Securities

A: "Regarding the private credit market, MUFG actually does not have a significant exposure. We have some exposure to companies that have been mentioned in the media. But as you saw earlier, our NPL ratio is declining. So I do not think we have a significant exposure. That said, the private credit market is extremely strong now. So we need to keep a close eye on the recent increase in volatility. I think the risk of lending to data centers depends on the project. We have extensive knowledge on project finance. So it is important to carefully select projects, taking into account factors like sources of cash flow and technical conditions, such as proper installation of high-voltage cables."
— Jun Togawa, Group CFO

Assessment: citing a falling non-performing loan ratio as evidence of low private-credit exposure is a non-sequitur, because the whole concern about that market is that stress does not show up in a backward-looking ratio. No exposure figure was given for either private credit or data-centre financing, and the global corporate group's fee income grew ¥27.2bn this half on expanded project finance in the Americas and Europe. The denial may well be correct; it is not evidenced.

Decomposing the guidance raise

A request for the components behind the revision waterfall, specifically whether the weaker currency sat in the financial-indicator bucket or somewhere else. The answer gave the fullest breakdown of the call and, in the process, disclosed that roughly ¥40bn of the ¥100bn raise is one-off gains the plan had never contained.

Q: "Upward revision of financial targets for FY '25. Can you give a more detailed breakdown? The graph on the bottom left shows a breakdown into customer segment, equity method investees and review on financial indicators. Can you give a breakdown of each of them? For example, weaker yen than the beginning of the year, would that be included in review on financial indicators or the equity market value?"
— Maoki Matsuno, Mizuho Securities

A: "Earlier, I said the customer segment is expected to continue making steady progress in the second half of the year and is expected to exceed the initial plan by around JPY 30 billion for the full year. Regarding equity and earnings of equity method investees, I must admit it is difficult to say how much is coming from Morgan Stanley, but a certain amount is factored in. There are also some one-offs. … Step-up gains from acquiring shares of JACCS, one-off gains from acquisition of Tidlor as a subsidiary and gains related to liquidation of local subsidiaries, a part of them were not factored in, accounting for approximately JPY 40 billion. The revision of financial indicators is expected to have an impact of approximately JPY 30 billion, mainly due to the weak yen."
— Jun Togawa, Group CFO

Assessment: ¥30bn of ¥100bn is operating improvement, ¥40bn is one-offs already banked, and ¥30bn is currency translation. The refusal to size the Morgan Stanley contribution inside the raise is the second time in the call that the single largest earnings variable was left unquantified. A raise that is 70% non-operating deserves to be read as the company marking to what has happened, not to what it now expects.

How the bond book gets run in the second half

The question observed that the first half had gone well through a combination of cutting super-long yen bonds and taking gains on foreign bonds, and asked what the second-half policy is. The answer covered both books and volunteered an explanation for a duration statistic that looks internally contradictory on the slide.

Q: "My second question is on the operational policy of Global Markets in the second half. In the first half of the year, it looks like you did well by drastically reducing yen bonds and super long-term bonds and making profits on foreign bonds. Is there anything you can speak about the operations of Global Markets in the second half of the year?"
— Maoki Matsuno, Mizuho Securities

A: "Regarding yen bond management from the second half onwards, our policy of gradually building up our yen bond positions, while monitoring the rise in Japan's policy rate remains unchanged. Short-term JGBs decreased as the BOJ's growth-oriented lending support operation is gradually coming to an end and need for short-term JGBs as collateral has decreased. … As for foreign bonds, the balance of long-term bonds appears to be increasing, while duration is decreasing and some might feel this doesn't sit well. This is due to categorizing mortgage bonds with long statutory maturities as long term. But overall duration shortened to 4 years."
— Jun Togawa, Group CFO

Assessment: a bank that intends to rebuild its yen bond position as the policy rate rises is telling you it expects to buy duration at better yields, which is the correct posture and the opposite of what it did in the year that produced the rebalancing loss. Pre-empting the mortgage-bond duration question without being asked is the kind of disclosure that builds credibility on a book that destroyed a lot of it two years ago.

Whether the fee surge is durable or an acquisition artefact

Fees and commissions were the strongest revenue line of the half in both domestic and overseas markets, and the question was simply whether it repeats. The answer separated the acquisition effect from the organic pipeline and sized the former precisely, which is the single most useful number given on the call.

Q: "Net fees and commissions in the first half of the year was very strong for both domestic and nondomestic. Is this trend in the first half a temporary phenomenon? Or including the current pipeline, can we expect further growth going forward?"
— Ken Matsuda, Daiwa Securities

A: "Fee revenues, fee income partially include impact of acquisitions. Acquisition of WealthNavi, MPMS acquired by our Trust Bank and NICOS acquiring Zenhoren has resulted in a total acquisition effect of about JPY 48 billion. Apart from that, GCIB, in particular, is further promoting O&D initiatives, so fee income will grow. Domestically, fees related to loans such as MBOs and LBOs are growing. Solution-related fees are also growing. … In addition, AUM in asset management is growing steadily, and IS has also issued a press release stating that outsourcing operations have quickly achieved the MTBP target."
— Jun Togawa, Group CFO

Assessment: ¥48bn of the ¥99.2bn reported fee increase is acquisition annualisation, which will lap out. The remaining roughly ¥51bn is organic and the sources named (originate-and-distribute, leveraged and management buyout financing, solutions) are cyclical but not one-off. Fee growth in the mid-single digits rather than the reported 10% is the right assumption to carry.

What They're NOT Saying

1. The expense ratio: characterised on the call as flat year on year at 56.1%. The company's own slide shows 55.1% moving to 56.1%, up 0.9ppt as reported and 0.7ppt adjusted. The slide is the disclosure; the characterisation is not.
2. The trust bank's net income fell ¥76.3bn to ¥89.0bn, with gross profits down ¥22.2bn and operating profit down ¥26.5bn. It is the second-largest banking entity in the group and its results were not discussed in the prepared remarks or in Q&A.
3. Held-to-maturity unrealised losses of ¥632.3bn, deepened from ¥625.2bn, of which ¥557.0bn is domestic bonds. The commentary covered available-for-sale positions at length and this number not at all.
4. Inorganic investment: a press report was raised directly and declined. The answer named three target areas and gave nothing else, at a point when a large transaction would have to fit inside the 100bp between today's 10.5% capital ratio and the 9.5% bottom of the target range.
5. No unit economics for the retail build. Two halves of elevated spending, multiples disclosed as ratios (account openings up 1.2 times, card issuance up 2 times, brokerage accounts up 11.5 times) and not a single revenue or contribution figure. The digital bank is not due until FY2026.
6. The ¥161bn gap between the ¥539bn of equity disposals now agreed and the ¥700bn target was not addressed, nor was the shape of the profit base in the first year after equity-sale gains (¥134.3bn in this first half alone) disappear.
7. ¥2.5tn of government lending ran off in six months, holding total domestic loans flat while the higher-spread commercial book grew ¥2.4tn. That is a favourable mix shift and the call gave it one clause (loans excluding government lending up about ¥4tn), so the mechanism behind it was never explained.
8. Morgan Stanley's contribution was explicitly not sized inside the guidance raise, despite being the largest single driver of the increase in profits other than operating profit.
9. NICOS lost ¥6.9bn, a ¥3.8bn wider loss year on year, while its gross profits grew ¥16.9bn. A card business growing revenue 16% and widening its loss deserved a sentence.
10. Nothing on capital-return policy beyond this fiscal year. The ¥500bn is a record and the question of whether it is the new run-rate or a one-year peak was neither asked nor volunteered.

Market Reaction

MUFG published the interim results on Friday, November 14, after the Tokyo close, and held its results call the same evening Japan time. The New York listing therefore traded the news on Friday, a full session before the ordinary shares could. Both legs are set out separately below.

• Pre-print setup, ordinary shares: 8306 closed at ¥2,451 on release day, a 52-week closing high, up 32.8% year to date against 26.3% for the Nikkei 225, up 36.3% over the trailing twelve months and up 7.0% over the trailing thirty days. The 52-week closing range through the prior session was ¥1,495.5 to ¥2,450.
• Pre-print setup, New York listing: the ADR closed at $15.44 on November 13, up 31.7% year to date against 14.6% for the S&P 500, up 27.9% over twelve months and up 2.0% over thirty days, inside a 52-week closing range of $10.79 to $16.05. The ADR sat 3.8% below its own high while the ordinary shares made a new one, a gap created entirely by the yen weakening from 148.88 at end-September to 154.52 on release day.
• New York reaction session, November 14: the ADR opened at $15.57, traded $15.49 to $15.95 and closed at $15.79, up 2.3%, on 2.5m shares against a 3.2m thirty-day average, a 0.8 times volume multiple. The S&P 500 fell 0.05% that session, so the relative move was roughly +2.3 points.
• Tokyo reaction session, November 17: the ordinary shares opened at ¥2,423.5, traded ¥2,405.5 to ¥2,453, and closed at ¥2,426.5, down 1.00%, on 47.4m shares against a 41.4m thirty-session average, a 1.1 times volume multiple. The Nikkei 225 fell 0.10%, putting the relative move at roughly minus 0.9 points.
• Megabank peers, same session: Sumitomo Mitsui Financial Group rose 4.57%, Mizuho Financial Group fell 0.26% and Sumitomo Mitsui Trust fell 0.34%. All three megabanks reported the same day. MUFG was the weakest of the four.

The two listings did not actually disagree, despite the opposite headline signs. At the November 14 rate of 154.52 the ADR's $15.79 close implied ¥2,440 per ordinary share, about 0.4% below Tokyo's pre-news close; Tokyo then settled at ¥2,426.5, about 0.6% below the ADR-implied level. The ADR's 2.3% gain was largely catch-up from a Thursday US session in which the S&P 500 had fallen 1.66%. The real verdict on the print is the Tokyo session, and the Tokyo session said no.

The relative trade, not the absolute one. Sumitomo Mitsui gapped open 4.8% higher and held the gain. The difference between the two prints was not the quarter, it was the guide: Sumitomo Mitsui raised its full-year forecast 15.4% and Mizuho raised 10.8%, against MUFG's 5.0%. A market that had spent the preceding month positioning for Japanese bank earnings upgrades got the largest upgrade from the second-largest bank and the smallest from the largest. On a day when all three reported together and the index was flat, that ranking was the entire story.

Positioning was the other half. The ordinary shares entered the print at a 52-week closing high, up 7.0% in thirty days and 32.8% for the year, having already re-rated ahead of an earnings season everyone knew would be good. A beat of 22.6% against a consensus that had modelled a 16% decline is, on paper, a very large surprise. It moved the stock down one percent, which tells you the beat was in the price and the composition was not what the buyer wanted. The parts of the beat that were genuinely new (credit-cost reversals, equity-method earnings, ¥100bn of one-time gains) are the parts nobody capitalises.

Street Perspective

Debate: is the ¥2,100bn target conservative, or is the second half genuinely weaker?

Bull view: the implied second half asks for ¥807.1bn after a ¥1,292.9bn first half, with credit costs assumed at 3.6 times the first-half run-rate and a currency assumption in the mid-¥140s that spot has already passed. On the bull reading, the target contains at least ¥100bn of cushion and the company said as much when it admitted debating whether to revise at all.

Bear view: the first half contained approximately ¥100bn of one-time gains, ¥134.3bn of equity-sale gains and an ¥82.3bn credit reversal at the main bank, none of which repeat. Treasury trading was front-loaded by roughly ¥130bn and costs step up roughly ¥100bn. On the bear reading the second half is honestly guided and the first half was the anomaly.

Our take: the bears are right about the first half and wrong about the target. The one-offs are real and the run-rate is lower than ¥1,292.9bn implies, but a ¥350bn full-year credit-cost assumption against a ¥76.3bn first half and a falling non-performing loan ratio is a reserve, not a forecast. We model the year above the target and treat the second half as the more honest picture of the underlying business.

Debate: does the quality of this beat matter, or is the rate cycle all that matters?

Bull view: quality of one half is noise against a structural repricing of a ¥227tn deposit book. The domestic loan rate excluding government lending moved 27 basis points in twelve months, the deposit rate moved 14, and the policy rate is halfway to the level management assumes for its 12% return target. Everything else is timing.

Bear view: a bank that reports a record half with operating profit down 1.4% and a fifth of its profit coming from a US investment bank is not a clean rate play. Underlying return on equity near 10.6% is well short of the 12% headline, and the gap is filled with disposal gains that management has itself promised will end.

Our take: both, and the sequencing favours the bull. The rate cycle is the dominant variable and it is early, but the bear is right that the entry multiple should be set against the 10.6% underlying return rather than the 12.5% reported one. On that return, an 8.5% to 9.0% cost of equity and 2% growth justify 1.23 to 1.32 times book, so at 1.32 times the shares are fully valued today. The upside in our value range depends on the underlying return rising to 11.0% to 12.5% as the rate cycle runs.

Debate: is a record ¥500bn buyback generous, or is MUFG losing the capital-return race?

Bull view: ¥500bn is the largest programme in the company's history, 25% above the prior two years' run-rate, paired with a ¥10 dividend increase, a 64.0% total payout ratio and the permanent cancellation of 200m shares. Judged against itself, this is a step change.

Bear view: judged against the day it was announced, it is the least exciting of three. Both peers raised full-year profit forecasts by more than twice MUFG's percentage and both expanded repurchase programmes. Capital returns are a relative game and the market ranked MUFG last, which is precisely what the reaction session showed.

Our take: the bear wins the session and the bull wins the year. MUFG is returning more absolute capital than either peer and doing it from a ratio at the top of its target range, which is the definition of sustainable. But the relative framing will keep capping the multiple until the second-half programme is either extended or the fiscal-2026 policy is set out, and neither happens before May.

Model Update Needed

Valuation. Book value per share was ¥1,834.3 at September 30, up ¥50.9 from ¥1,783.4 in six months after the year-end dividend and the first-half repurchase. At the November 17 close of ¥2,426.5 the shares trade at 1.32 times book, 13.2 times the company's own revised full-year target and a 3.05% forward dividend yield. On a sustainable return on equity of 11.0% to 12.5%, a cost of equity of 8.5% to 9.0% and 2% long-term growth, the justified multiple is 1.29 to 1.62 times book. Against a twelve-month book value of roughly ¥1,934, that is a ¥2,495 to ¥3,133 value range with a midpoint near ¥2,814, about 16% above the current price. Adding the dividend gives a total return in the high teens. The equivalent ADR figure at the prevailing exchange rate is roughly $18.10, against a last close of $15.79, but any ADR target carries a currency view that the yen-denominated one does not.

Thesis Scorecard Post-Earnings

This is our first published work on MUFG, so this quarter establishes the pillars rather than grading a standing thesis. The statuses below are the opening positions that subsequent quarters will be scored against.

Overall: thesis established. The operating evidence for the rate-normalisation case is stronger than the headline suggests, and the quality of the reported half is weaker. Those two facts point in the same direction for a buyer, because the market is pricing the reported half.

Action: build the position. At 1.32 times book against an underlying return on equity of 10.6% that management is targeting to reach 12% on a policy-rate assumption only halfway realised, the re-rating is not the opportunity; the further earnings growth is. The two things that would change this view are a second-half risk-weighted asset build that takes the capital ratio below the target range, and any sign that the Morgan Stanley contribution is peaking.