The Rate Hike Came With Twelve Days Left, the Balance Sheet Grew ¥7tn, and the Target Did Not Move

Results vs. Consensus

MUFG reports under Japanese GAAP in yen, publishes a full-year target rather than a forecast, and gives no quarterly guidance at all. Japanese issuers are polled on cumulative year-to-date figures, so the market benchmark for this print is the nine-month consensus for April to December, not a standalone quarter. The print is therefore scored against three yardsticks: the pre-print consensus poll, progress against the company's own full-year target, and the prior-year nine months on both of the bases the company publishes.

Nine-month scorecard (April 1 to December 31, 2025)

Two things about that table matter more than the rest. The first is that every operating line improved and the two lines the market actually polls came in fractionally short. The second is that the consensus miss is trivially small in absolute terms and not small at all in signalling terms: a nine-month cumulative poll already contains the interim actual, so a ¥9.4bn shortfall against a nine-month number is a ¥9.4bn shortfall against the one unknown quarter, or roughly 1.8% of it.

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

Every comparative carries two columns. 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 to December year to an April to March year, which loaded an extra quarter of the Thai subsidiary into the prior-year period. MUFG sizes the nine-month distortion at ¥83.8bn of net operating profit and ¥23.2bn of net profit. Management speaks exclusively to the adjusted column.

Figures are reproduced as the company prints them. Component rounding means one or two change cells differ from the difference of the rounded cells above them by ¥0.1bn. The right-hand column is the Krungsri-adjusted change and by construction does not equal the difference of the two as-reported columns.

The reported net operating profit increase of ¥191.2bn is the headline, and ¥152.4bn of it is the absence of last year's losses on sales of debt securities. Strip that single line and the nine-month operating result grew ¥38.8bn on a reported basis. The foreign-exchange contribution stated by the company is roughly ¥55.0bn of gross profits against roughly ¥30.0bn of expenses, so about ¥25bn of the operating increase is currency translation rather than business.

The derived third quarter, October to December

MUFG publishes no standalone quarter. The figures below are the nine months less the half year, taken from the Selected Financial Information supplements filed with each. The prior-year base carries roughly ¥4.2bn of residual Krungsri effect, being the nine-month adjustment of ¥83.8bn less the half-year adjustment of ¥79.6bn.

The derived EPS is the nine-month basic EPS less the half-year basic EPS. The average share count differs between the two windows, so it approximates the quarter's per-share result rather than reporting it.

Revenue assessment

Nine-month gross profits grew ¥347.7bn as reported and ¥521.8bn adjusted, and the adjusted figure still carries acquisitions and currency. Removing the roughly ¥55bn of stated foreign-exchange benefit leaves roughly ¥467bn, and the fee line still contains the annualisation of the WealthNavi, MPMS and Zenhoren purchases that management sized at about ¥48bn at the interim. Call the organic number roughly ¥420bn, or about 10%.

The source of it is the right one. MUFG Bank's domestic net interest income grew from ¥590.0bn to ¥728.3bn, an increase of ¥138.3bn or 23.4%, while its non-domestic net interest income fell from ¥582.0bn to ¥538.9bn as the Federal Reserve cut. On a like-for-like basis, once the ¥78.5bn of prior-year investment-trust cancellation gains and this year's ¥9.2bn of cancellation losses are removed from both sides, the Bank's total net interest income grew 16.7% to ¥1,276.4bn. That is the cleanest single number in the release and, as at the interim, it appears only in a footnote.

The group-level net interest income line looks nothing like that, growing 0.9% as reported, because Krungsri's contribution fell from ¥528.3bn to ¥444.9bn on the closing-period change and the Thai slowdown. Interest income on loans and bills discounted actually fell 4.9% to ¥3,122.3bn while the loan book grew ¥8.9tn. Anyone reading only the consolidated income statement will reach the opposite conclusion about MUFG's rate sensitivity from the one the disclosure supports.

Margin assessment

The nine-month expense ratio improved a full point to 57.3%, and improved 1.3 points on the adjusted basis. That is a genuine reversal of the interim, where the ratio went the wrong way by 0.9 points. Expenses still grew ¥156.5bn, of which roughly ¥30bn is currency, so the underlying cost growth of roughly ¥127bn against roughly ¥293bn of currency-adjusted revenue growth is a flow-through of well over half.

The derived quarter's ratio of 59.6% is worse than the nine-month average, which is normal seasonality rather than a signal: the December quarter carries the bonus accrual and the year-end push on strategic spend. The comparison that matters is 59.6% against 66.2% a year earlier, and most of that improvement is the denominator recovering from the bond losses rather than the numerator behaving.

EPS and returns assessment

Reported return on equity of 11.55% is below the 11.67% of the prior-year nine months, which extends the small year-on-year decline already visible at the interim (12.5% against 12.6%), and it happened in a period when operating profit grew 11.2%. The arithmetic is straightforward: shareholders' equity grew from ¥20,520.4bn to ¥21,363.3bn while the earnings that sit on it are growing more slowly, because the prior year contained ¥488.1bn of equity-securities gains against ¥183.3bn this year.

Strip the roughly ¥100bn of one-time gains MUFG itself identified at the interim, and tax the ¥183.3bn of equity-securities gains at the period's 23.2% effective rate, and the underlying nine months is closer to ¥1,572.8bn. Annualised against average shareholders' equity of ¥20,941.8bn that is a return of roughly 10.0%, against the 10.6% we calculated on the same method at the interim. The gap between 10.0% today and the 12% management targets on a 1.0% policy rate is the investment case, and the policy rate moved to 0.75% inside this quarter.

Segment Performance

Two disclosures describe the same business and do not tie. The Consolidated Summary Report's segment note is Japanese GAAP and as reported. The Financial Highlights deck's business-group tables are managerial accounting on a local-currency basis. The gap is currency, and both are the company's own figures. Both are shown below, labelled.

Segment note, Japanese GAAP basis (¥m)

MUFG changed the method of allocating net revenue and operating expenses among reporting segments in this nine-month period and restated the prior-year comparative on the new basis. No bridge between the old and new methods was published. Separately, MUFG corrected this table on February 5, the day after the release: Global Commercial Banking's net revenue was restated upward by ¥34,446m, with an offsetting change in the unallocated reconciliation line, leaving consolidated ordinary profit unchanged. The corrected figures are used throughout.

On this basis Global Markets supplied ¥164.9bn of the ¥212.9bn increase in segment operating profit, or 77%. Global Commercial Banking went backwards by ¥78.8bn, which is the Krungsri closing-period effect running through an unadjusted comparative. Neither fact is available from the deck's tables, which is why both disclosures are reproduced here.

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

Currency-neutral, the six customer groups grew operating profit ¥133.8bn, or 9.7%, on a 41.5% flow-through of revenue growth to profit. Global Markets grew ¥169.7bn. So on this basis the markets book supplied 56% of the segment increase and the customer franchise 44%. At the interim, the customer groups supplied all of it and Global Markets went backwards. The difference is entirely the ¥223.5bn of debt-securities losses that sat in the December 2024 quarter and could only ever produce a one-time base effect on the way out.

Risk-weighted assets and return on equity are not disclosed by business group at the third quarter. They appear only at the interim and the full year. The central question we raised at initiation, whether the groups consuming the most capital are earning an adequate return on it, therefore cannot be tested against this print.

The derived December quarter by business group (¥bn)

Nine-month deck tables less interim deck tables, on the same managerial local-currency basis. The derivation assumes that basis is unchanged between the two decks. It would not be safe to do the same with the Summary Report's segment note, because that note was restated on a new allocation method this quarter and the interim note was not.

Two readings fall out of that table. The customer franchise grew operating profit 10.4% in the quarter, which is the honest run-rate once the base effect is stripped. And Global Markets earned ¥9.4bn of operating profit in the December quarter against an average of roughly ¥102bn in each of the two prior quarters. The markets book did not carry this quarter. It merely stopped losing money in the comparative.

Commercial Banking & Wealth Management

The clearest expression of the thesis in the release. Nine-month gross profits grew ¥91.3bn, of which ¥81.7bn came from loan and deposit interest income alone, while expenses grew ¥20.6bn. The expense ratio fell six points to 56% and operating profit grew 36.6%. Investment product sales were flat at ¥143.9bn, so none of the growth came from market-sensitive lines. In the derived quarter the group grew operating profit ¥21.6bn on ¥29.5bn of revenue growth, a 73% flow-through.

Assessment: this is what domestic rate normalisation looks like when no offsetting cost build is attached, and it is the group with the most operating leverage left if the Bank of Japan continues. It is also the group whose result is most nearly a pure function of the policy rate, which cuts both ways.

Global Corporate & Investment Banking

Commission income grew ¥35.2bn and supplied 72% of the group's ¥48.6bn revenue increase, while loan and deposit interest income contributed ¥5.0bn against a loan book that grew substantially. The expense ratio improved a point to 50%, operating profit grew ¥28.0bn and net income grew ¥47.4bn, again on the rebound from prior-year overseas provisions.

Assessment: the shape is exactly what the originate-and-distribute strategy promises, with fees growing far faster than lending income. The problem is that the balance sheet grew anyway. Overseas loans added ¥4.7tn in the quarter. A distribution business whose inventory rises ¥4.7tn in three months is either warehousing a great deal or distributing rather less than it originates, and the disclosure does not let an outsider tell which.

Retail & Digital

Nine-month gross profits grew ¥93.8bn, the largest absolute revenue increase of any customer group, with loan and deposit interest income up ¥49.2bn and card settlement up ¥24.2bn. Expenses grew ¥77.0bn, so operating profit grew ¥16.8bn on a 17.9% flow-through and the expense ratio rose a point to 73%. Net income grew only ¥6.4bn to ¥67.1bn, and on the derived quarter it was essentially nil at roughly ¥0.4bn against ¥23.6bn a year earlier, with the card subsidiary's ¥30bn system-integration charge and ¥83.8bn of nine-month consumer-finance credit costs both landing here.

Assessment: the flow-through improved from 12.8% at the interim to 17.9%, which is progress on the worst number in the interim release. It is still the worst number in this one. Three consecutive periods of elevated spend have now passed without a single unit-economics disclosure for the retail brand, the digital bank or the securities integration.

Japanese Corporate & Investment Banking

Gross profits grew ¥23.0bn on broad contribution, with real estate and corporate agency up ¥6.6bn, derivatives and solutions up ¥6.1bn, advisory and capital markets up ¥5.0bn and loan and deposit interest income up ¥6.1bn. Expenses grew ¥13.7bn and the ratio moved a point to 40%, still the best in the group. In the derived quarter, though, operating profit fell ¥6.4bn year on year on slightly lower revenue.

Assessment: the highest-quality revenue mix in the table and the weakest sequential quarter of the six. Large-corporate domestic lending grew ¥1.8tn in the quarter while this group's lending income barely moved, which says the incremental volume is being written at spreads that do not change the arithmetic.

Global Commercial Banking

Gross profits grew ¥9.5bn against expenses up ¥13.9bn, so operating profit fell ¥4.5bn and the expense ratio rose two points to 58%. Krungsri's operating profit fell ¥5.9bn and its expense ratio rose three points to 51%, while Bank Danamon added ¥1.3bn. Net income nonetheless grew ¥15.0bn to ¥85.3bn.

Assessment: the same revenue problem masked by the same provisioning benefit that we flagged at the interim, one quarter further on and no better. Asia is the part of this franchise that is not working, and the Krungsri cost ratio is now moving in the wrong direction alongside it.

Asset Management & Investor Services

Gross profits grew ¥56.5bn, of which ¥45.0bn came from investor services alone on the bundled-services expansion and the First Sentier and Link acquisitions. Expenses grew ¥43.0bn, so operating profit gained ¥13.5bn and the ratio rose a point to 72%. First Sentier's own gross profits fell ¥3.6bn and its net income fell ¥3.6bn to ¥10.8bn.

Assessment: 76% of the revenue growth in this group is being spent, in a business MUFG has named as one of three inorganic priorities. The acquisitions are adding revenue and not yet adding margin, and the largest of them is going backwards on its own numbers.

Global Markets

Nine-month gross profits grew ¥177.8bn and operating profit grew ¥169.7bn, entirely on treasury, which swung from a ¥51.8bn operating loss to a ¥122.7bn profit. Sales and trading operating profit fell ¥7.8bn. The derived December quarter is the one to read: treasury gross profits were negative ¥11.0bn and treasury operating profit was negative ¥25.8bn, against sales and trading operating profit of ¥30.5bn.

Assessment: the treasury book went back into the red in the quarter the Bank of Japan raised rates and the JGB curve moved, and the nine-month presentation of a ¥172.3bn treasury swing conceals it completely. Sales and trading carried the group in the quarter, which is the reverse of the nine-month story. For a book that produced a ¥991.4bn loss last fiscal year, a quarter of renewed negative carry deserved a sentence somewhere in the disclosure and did not get one.

Entity view: where the ¥1,813.5bn was earned

Two of those lines need their footnotes. MUFG Bank's net income fell ¥51.1bn despite operating profit growing ¥222.4bn, because it carries an approximately ¥56.0bn pre-tax write-off of overseas subsidiary shares that is eliminated in MUFG's consolidated accounts. The Trust Bank's ¥50.4bn net-income decline is largely a base effect: the prior-year comparative contains an approximately ¥46.0bn pre-tax dividend paid up by the asset-management subsidiary. The Bank's explanation is new this quarter. The Trust Bank's repeats a footnote the interim deck already carried against that bank's column.

Balance Sheet, Capital and Credit

Capital: the third consecutive quarterly decline

The ratio that matters for the medium-term plan is common equity tier 1 on the finalised and fully implemented Basel III basis, excluding unrealised gains on available-for-sale securities. It is published nowhere in the Summary Report and nowhere in the deck. It appears only in a two-line answer on the written FAQ.

The medium-term plan range is 9.5% to 10.5%. The ratio is inside it and is 50 basis points lower than it was at March. The regulatory 14.00% should never be read against that range; the gap between the two measures is 370 basis points.

The mechanics are in the February 13 Basel 3 release. Risk-weighted assets grew from ¥110,808.0bn at September to ¥117,496.9bn at December, an increase of ¥6,688.9bn or 6.0% in three months, and ¥10,566.4bn or 9.9% over nine months. Common equity tier 1 capital grew ¥847.5bn, or 5.4%, in the same quarter. Capital grew quickly and risk-weighted assets grew faster, which is the whole explanation for the ratio.

One detail cuts the other way and deserves credit. Tier 1 capital less common equity tier 1, which is the additional tier 1 layer, grew only ¥31.6bn in the December quarter against ¥461.2bn in the first half. At the interim we flagged that over half of the tier 1 increase had come from issuance rather than retention. In this quarter essentially all of it was common equity. The balance sheet shows where that came from: shareholders' equity before other comprehensive income rose only ¥6.0bn and retained earnings fell ¥155.7bn, while accumulated other comprehensive income rose ¥479.5bn, led by foreign-currency translation (+¥472.6bn) and available-for-sale gains (+¥246.3bn). The capital build no longer leans on issuance. It leans on the yen and on market values instead of retained profit.

Assessment: the ratio has fallen in each of the three quarters since the plan was set, from 10.8% to 10.3%, by 10, 20 and 20 basis points on the published one-decimal figures. On the current trajectory it approaches the middle of the range by March and reaches the bottom of it during the next fiscal year. Management guided at the interim to a March ratio "somewhere between 10% and 10.5%", and that guide now looks like the ceiling rather than the midpoint of the outcome.

The buyback: three quarters executed, one quarter to run

The ¥250bn second-half repurchase authorised on November 14 runs to February 27. The February 2 progress notice shows 70,044,500 shares bought for ¥178,892,561,898 to January 31, at an average of ¥2,554. January alone accounted for 22,985,400 shares and ¥63.6bn at an average of ¥2,767.

That is 71.6% of the value authorisation and 53.9% of the 130m-share cap, with roughly four weeks left at January 31. The share cap will not bind, because the average price paid is well above whatever price the 130m figure was set against. Combined with the first-half programme the full-year total reaches ¥500bn, the largest in the company's history, and the 200m treasury shares cancelled on November 28 have already reduced shares outstanding from 12,067.7m to 11,867.7m.

Assessment: the capital-return commitment made at the interim is being executed on schedule and without fuss. What is entirely absent is any signal about the fiscal-2026 policy. The ¥500bn is 25% above the two-year run-rate, and whether it is the new base or a one-year peak will not be answered until May.

Balance sheet and the loan book

The loan categories shown are the three the company breaks out; housing loans and a residual line make up the remainder of the total.

The loan book grew ¥7.0tn in three months, from ¥124.8tn to ¥131.8tn: domestic loans up ¥2.7tn, overseas loans up ¥4.7tn, and the residual category down ¥0.4tn. Within the domestic increase, large-corporate lending rose ¥1.8tn to ¥28.5tn and lending to small and medium companies rose ¥0.7tn to ¥28.5tn. Those two books are now the same size for the first time in the disclosed series, and the small-company book grew again, by ¥0.7tn after ¥0.6tn in the first half, which reverses one of the mix concerns we carried out of the interim.

The overseas increase splits as the Americas up ¥1.8tn to ¥17.9tn, Asia and Oceania up ¥1.4tn to ¥15.8tn, Europe and the Middle East up ¥1.1tn to ¥10.3tn. At the interim the overseas book was ¥48.4tn, under a plan that targets a doubling of fee income from originate-and-distribute business. Two quarters later the balance is ¥53.1tn.

The funding side is the quiet good news. Deposits grew ¥5.3tn in the quarter, individual deposits grew ¥1.6tn, and the loan-to-deposit ratio finished at 56.7%. The Bank of Japan current account fell ¥24.8tn over nine months and borrowed money fell ¥11.6tn as the growth-supporting funding facility ran off, so the loan growth is being funded out of a shrinking pile of near-zero-yielding central-bank deposits. The economics of that substitution are unambiguously favourable. The capital cost of it is what shows up in the ratio.

Domestic spreads: the engine, and what is actually driving it

On the cumulative nine-month series, MUFG Bank's domestic loan rate excluding government lending reached 1.17% from 0.91%, and the spread reached 1.00% from 0.86%.

The composition is worth stating plainly, because it is not what the phrase "loan repricing" implies. Over the year the lending rate rose 25 basis points and the deposit rate rose 12, so the bank retained roughly half the move. Meanwhile the two corporate lending-spread series MUFG publishes have barely moved at all, sitting inside a 0.60% to 0.65% band across twelve quarters, with the small-company line at 0.65% the highest point on the chart. The spread widening is a deposit-repricing lag, not a change in loan pricing power. That is a perfectly good source of profit and it is a finite one: the deposit beta will eventually catch up.

Securities: a smaller domestic book, a bigger unrealised hole

The available-for-sale JGB book has shrunk ¥8.23tn since March and the foreign bond book has grown ¥4.41tn, moving from an unrealised loss to a small unrealised gain. Available-for-sale unrealised gains reached ¥3.07tn, of which ¥3.10tn is the domestic equity portfolio, whose carrying value grew again to ¥4.10tn because market appreciation is outrunning the disposal programme.

Against that, held-to-maturity unrealised losses deepened from ¥625.2bn at March and ¥632.3bn at September to ¥872.4bn at December, of which ¥814.9bn is domestic bonds and ¥573.1bn is Japanese government bonds. The held-to-maturity balance grew ¥1.84tn in the same period, so the bank is adding to a book whose mark is deteriorating. There is no commentary on this anywhere in the release, and there was none at the interim either.

One further line moved and was not discussed. Net deferred losses on hedging instruments in shareholders' equity deepened from ¥894.2bn at March to ¥1,307.5bn at December, a ¥413.3bn deterioration over nine months in the accounting that sits between the yen rate book and the income statement. Comprehensive income of ¥2,188.8bn still exceeded net profit of ¥1,813.5bn, so the overall mark-to-market is favourable. The hedging line is the exception, and it is moving against the bank as rates rise.

Credit: a record-low ratio and a deteriorating region

The ratio is below 1% for the first time in the disclosed series, from 1.51% at March 2024 and 1.01% at the interim. Note what is doing the work: the numerator fell ¥77.8bn and the denominator grew ¥10.6tn. Roughly half the improvement in the ratio is loan growth diluting the book rather than problem assets being resolved.

Inside that total, one region moved sharply the wrong way. The company's own regional breakdown shows Americas non-performing loans rising from ¥124.0bn at March to ¥205.5bn at December, an increase of 66%, while Europe and the Middle East fell from ¥127.8bn to ¥80.1bn, Asia was flat at ¥491.7bn and domestic fell from ¥786.7bn to ¥675.1bn. The Americas book is where the ¥1.8tn of quarterly loan growth also went.

Total credit costs of ¥219.8bn for nine months are 62.8% of the ¥350.0bn full-year target, and the disclosed composition is ¥160.5bn overseas, ¥83.8bn at the consumer-finance subsidiaries and a ¥27.0bn net reversal at MUFG Bank, with a small remainder elsewhere in the group. The derived quarterly charge was ¥143.4bn against ¥76.4bn for the whole first half. The deck calls this "within the initial outlook range, despite large builds and reversals of loan loss provisions", which is true, and understates a near-quadrupling of the quarterly run-rate.

Two further disclosures belong here. MUFG adopted a new internal credit rating system during the period and recalculated its allowance on it, describing the impact as immaterial. And the forward-looking qualitative adjustment to loss rates, the overlay held against the Russia-Ukraine situation and trade policy, fell from ¥33.6bn at March to ¥15.5bn at December. An ¥18.1bn release of management overlay in a year when the overseas credit charge is running at ¥160.5bn is a choice, and it is disclosed only as a number in a paragraph.

Key Topics from the Disclosure

Overall disclosure posture: confident and thin. The deck leads with a record nine-month net income and an 86.4% progress rate, gives the operating improvement no qualification beyond the standard foreign-exchange footnotes, and devotes no page at all to capital or shareholder return. What has improved since the interim is the footnoting: the MUFG Bank write-off and the card subsidiary's integration charge are both explained on the page where they land. What has not improved is that the quarter's two most consequential movements, the risk-weighted asset build and the return of negative treasury carry, appear in no narrative anywhere.

1. The Bank of Japan moved on December 19, with twelve days of the quarter left

The policy rate went to 0.75% from 0.5% on December 19, 2025, the highest level since 1995. The Bank held at 0.75% on January 23, 2026 ahead of a February general election, with one board member dissenting in favour of 1.00%. MUFG's interim deck sizes the pre-tax profit impact of taking the policy rate from 0.5% to 1.0% at approximately ¥360bn, on its path to the medium-to-long-term return target.

Twelve days of a ninety-two day quarter is roughly 13% of the period and about 3% of a year, so only a sliver of whatever the step is worth annually could have reached this print, before any consideration of repricing lags on a loan book that reprices over months rather than days. The December quarter's lending rate of 1.20% is up just 2 basis points on the September quarter's 1.18%, which is consistent with almost none of the December step having arrived yet.

Assessment: the single most important thing about this quarter is a number that is not in it. The results describe a bank earning a 1.00% domestic loan-deposit spread at a 0.5% policy rate. The policy rate is now 0.75%, and management's own 12% return-on-equity target assumes 1.0%. Whatever one thinks of the print, the earnings base under it moved up after the balance sheet date.

2. The target was left untouched at 86.4% complete, and the implied fourth quarter is very small

The Summary Report states it in one sentence: "MUFG has an earnings target of 2,100.0 billion yen of profits attributable to owners of parent for the fiscal year ending March 31, 2026. (There is no change to our earnings target released on November 14, 2025.)" The dividend forecast of ¥74.00 was likewise left alone.

An implied fourth quarter of ¥286.5bn is 45% below the ¥520.6bn just delivered. It is 2.5 times the ¥114.0bn MUFG earned in the January to March quarter of last year, but that quarter absorbed the bulk of the bond-portfolio rebalancing loss and is not a normal base.

Assessment: at the interim, management conceded it had debated internally whether a 5% revision was worth making at all, which told us the ¥2,100bn contained an undisclosed cushion. This print does not remove the cushion; it enlarges it. Leaving the number alone with 86.4% booked is either genuine caution about the fourth quarter or a deliberate reserve. Without a call, there is no way to ask which, and that is precisely the cost of the first-and-third-quarter disclosure regime.

3. Risk-weighted assets grew 6% in a quarter and the capital ratio fell for a third time

Risk-weighted assets went from ¥110,808.0bn to ¥117,496.9bn in three months. Loans grew ¥7.0tn, of which ¥4.7tn was overseas. The target-basis capital ratio went from 10.5% to 10.3%, having been 10.8% at March and 10.7% at June.

At the interim, risk-weighted assets grew 3.6% in six months against loan growth of 1.5%, and we rated that bear point contained because the ratio was at the top of its range. This quarter risk-weighted assets grew 6.0% in three months against loan growth of 5.6%, and the ratio is now 20 basis points below the top of the range rather than at it.

Assessment: this is the pillar that moved this quarter. The capital position is comfortable and the direction is consistent, quarter after quarter, in one direction. Nothing here is a problem in the next twelve months. Everything here is a problem for anyone assuming that the ¥500bn capital-return run-rate is repeatable alongside a balance sheet growing at this pace.

4. Overseas loans grew ¥4.7tn in three months against a distribution story

The overseas book went from ¥48.4tn at September to ¥53.1tn at December: Americas ¥16.1tn to ¥17.9tn, Asia and Oceania ¥14.4tn to ¥15.8tn, Europe and the Middle East ¥9.2tn to ¥10.3tn. The deck reports the nine-month increase as ¥6.7tn, or ¥3.3tn excluding currency, and offers no explanation of the quarter.

Global Corporate & Investment Banking's own revenue mix argues the opposite way: commission income grew ¥35.2bn while lending income grew ¥5.0bn, which is the signature of an originate-and-distribute franchise doing exactly what it says.

Assessment: both things cannot be true indefinitely. Either the distribution engine is working and this is a timing effect in warehoused inventory, in which case the balances fall back in the first half of the next fiscal year, or the balance sheet is quietly becoming the product again. The Americas non-performing loan build over the same period is not proof of the second reading, and it is not nothing either.

5. Morgan Stanley supplied a quarter of the nine months and a third of the quarter

Equity in earnings of equity-method investees grew ¥154.8bn to ¥582.9bn, which is 23.2% of ordinary profits. On the entity breakdown, Morgan Stanley alone contributed ¥465.6bn of the ¥1,813.5bn nine-month net profit, or 25.7%, against 22.3% at the interim. Subtracting the interim figure, the Morgan Stanley contribution in the December quarter was ¥177.7bn, or 34.1% of the quarter's ¥520.6bn.

The deck's only commentary is that equity-method earnings "increased YoY, led by solid performance of Morgan Stanley".

Assessment: this is the largest single unhedged exposure in the earnings model and it is a US capital-markets cycle rather than a Japanese rate cycle. It is a genuine asset and a structurally advantaged alliance. It is also the reason a sum-of-the-parts on this company should carry a discount rather than a premium, and the dependence rose again this quarter in both absolute and relative terms.

6. Below-the-line items are 42% of the profit

Net gains on equity securities of ¥183.3bn and equity-method earnings of ¥582.9bn together total ¥766.2bn of pre-tax income sitting below the operating line, against ¥1,813.5bn of after-tax profit. The comparison mixes bases and is still worth making, because both lines are outside the operating business an investor is underwriting.

The equity-securities line is also shrinking on purpose. Gross sale gains fell from ¥529.3bn to ¥234.1bn as the disposal programme slowed, while losses on sales grew from ¥21.6bn to ¥42.3bn.

Assessment: the mix is moving in the right direction and slowly. Equity-securities gains fell ¥304.8bn year on year and the group still grew net profit, which is the single best evidence in this release that the operating business is replacing the disposal gains. The offset is that equity-method earnings grew ¥154.8bn at the same time, so roughly half the replacement came from another non-operating line.

7. The equity-disposal programme has almost stopped

Cumulative sales during the current medium-term plan reached ¥362bn on an acquisition-cost basis at December, from ¥339bn at September. That is ¥23bn of disposal in three months. Including agreed but unsold positions, the expected total rose from ¥539bn to ¥557bn against a ¥700bn target. For comparison, the prior fiscal year sold ¥276bn in full and ¥225bn in its first three quarters; this year's first three quarters have sold ¥86bn.

Meanwhile the domestic equity book's carrying value grew to ¥4.10tn and its unrealised gain to ¥3.10tn, because appreciation is outrunning disposal by a wide margin.

Assessment: the programme is running at roughly a third of the prior year's pace with ¥143bn of the target not yet even agreed. Management has been explicit that the 12% return target assumes no equity-sale gains at all. The longer the disposals take, the longer the profit base contains a line management has promised to remove, and the less informative the reported return on equity becomes.

8. Global Markets treasury went back into the red in the quarter

The nine-month presentation shows treasury operating profit swinging ¥174.5bn, from a ¥51.8bn loss to a ¥122.7bn profit. Subtract the interim figures and the December quarter alone shows treasury gross profits of negative ¥11.0bn and treasury operating profit of negative ¥25.8bn, against sales and trading operating profit of ¥30.5bn. Including the unallocated remainder the whole Global Markets group earned ¥9.4bn.

Net losses on sales of debt securities for the quarter were ¥71.6bn, against ¥2.9bn of gains for the entire first half.

Assessment: the quarter in which the Bank of Japan raised rates was a quarter in which MUFG's own bond book lost money again. That is not a scandal; a treasury book that is being rebuilt into a rising-rate environment will take realised losses on the way. It is, however, the exact pattern that produced a ¥991.4bn loss last fiscal year, and it is presented here inside a cumulative number that shows the opposite sign.

9. The domestic mix improved, and the spread source did not

Lending to small and medium companies grew ¥0.7tn in the quarter to ¥28.5tn, drawing level with large-corporate lending for the first time in the disclosed series and reversing the mix deterioration we flagged at the interim, when ¥2.5tn of government lending ran off and large-corporate lending took the space. Government lending was flat at ¥5.5tn this quarter, so the growth is genuine commercial volume.

The spread series tell a flatter story. Both published corporate spread series sit in a 0.60% to 0.65% band and have moved a handful of basis points in three years. The widening in the headline loan-deposit spread is deposit-side lag.

Assessment: mix improved, pricing did not. The bank is growing its highest-spread domestic book again, which is the right direction, and it is doing so at spreads that have not changed. The rate cycle is doing the work and the franchise is holding its share of it.

10. Disclosure quality: two explanations added, one correction issued, one page missing

This release footnotes three entity-level distortions. MUFG Bank's net-income decline is attributed to a ¥56.0bn write-off of overseas subsidiary shares, an explanation that is new this quarter. The card subsidiary's widening loss is attributed to ¥30.0bn of system-integration expense booked in the third quarter itself. The Trust Bank's decline is attributed to a ¥46.0bn prior-year dividend from the asset-management subsidiary, which repeats a footnote the interim deck already carried. All three appear as footnotes on a single deck page.

Against that, MUFG had to correct the segment note the day after publishing it, restating Global Commercial Banking's net revenue by ¥34.4bn, and the third-quarter deck contains no capital page at all. The only place the ratio that governs the medium-term plan appears is a two-line FAQ answer.

Assessment: the footnoting is a real improvement and should be credited. The structural problem is unchanged: for two quarters of every four, the most important single number this company reports is published in the least prominent place it could be published.

11. A chief executive was named in December, and this is his predecessor's last third quarter

MUFG announced in December 2025 that Junichi Hanzawa, president of MUFG Bank, becomes President and Group Chief Executive on April 1, 2026, with Hironori Kamezawa moving to chairman. Masakazu Osawa succeeds Hanzawa at the Bank. The Summary Report still lists Kamezawa as representative.

Assessment: the succession is internal, from the main banking subsidiary, and telegraphed. It matters for one reason: the fiscal-2026 capital-return policy and any revision to the medium-term plan will be set in May by a chief executive in his second month. New chief executives have an incentive to reset expectations downward, and the ¥500bn buyback run-rate is the most obvious thing to reset.

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 ¥6.7tn quarterly growth in risk-weighted assets, the ¥4.7tn quarterly increase in overseas lending, the return of negative treasury carry, the 66% rise in Americas non-performing loans, or the ¥18.1bn release of forward-looking credit overlay, because there was no forum in which to do so. The next such forum is the full-year results call in May. We flag the absence rather than pass over it, because for a quarter in which the balance sheet grew ¥13.8tn it is a material limitation on what any analysis of it can claim.

What They're NOT Saying

1. Why risk-weighted assets grew 6% in a quarter. The largest single movement in the release, disclosed nine days after the results in a separate regulatory filing, with no narrative attached in either document. Loan growth of 5.6% does not by itself explain a 6.0% risk-weighted asset increase, and the composition of the difference is not given.
2. Why overseas loans grew ¥4.7tn. The originate-and-distribute framing given in November implied these balances would normalise. They grew by roughly a tenth in three months instead, and the characterisation was not revisited.
3. The negative treasury quarter. A ¥25.8bn operating loss in the treasury book, visible only by subtraction, presented inside a cumulative number showing a ¥174.5bn improvement. The bond book is the line item that destroyed the most value in this company's recent history, and it gets one clause in a bullet.
4. The held-to-maturity mark. Unrealised losses of ¥872.4bn, up ¥247.3bn since March, on a book the bank is still adding to. 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, for the third consecutive period.
5. The hedging line. Net deferred losses on hedging instruments deepened ¥413.3bn over nine months to ¥1,307.5bn, inside shareholders' equity. It is disclosed as a balance-sheet line and nowhere else.
6. Why Americas non-performing loans rose 66%. From ¥124.0bn to ¥205.5bn since March, in the region where the loan book also grew fastest. The deck prints the regional table and says nothing about it, and the overall ratio improvement provides convenient cover.
7. The ¥18.1bn release of forward-looking credit overlay. Reduced from ¥33.6bn to ¥15.5bn during a period when the overseas credit charge ran at ¥160.5bn. Disclosed as a parenthetical in an accounting note.
8. Any explanation of the unchanged target. 86.4% of a full-year number booked in nine months, with an implied fourth quarter 45% below the third, and not a sentence about why.
9. Segment risk-weighted assets and returns on equity. Not disclosed at the third quarter, so the question of whether the groups consuming the most capital are earning an adequate return on it cannot be tested until May.
10. The segment allocation change. MUFG changed the method by which revenue and expenses are allocated across reporting segments this period and restated the prior-year 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.
11. The equity-disposal timeline. ¥23bn of cost-basis reduction in the quarter against ¥338bn still required for the ¥700bn target, with no restatement of when it will be reached or what the profit base looks like afterwards.
12. Anything about capital return beyond February 27. The current programme ends that day. The fiscal-2026 policy will be set in May by an incoming chief executive, and nothing at all was said about the framework.
13. Any unit economics for the retail build. Three consecutive periods of elevated spend, a flow-through of 17.9% for the nine months, and still no revenue or contribution figure for the retail brand, the digital bank or the securities integration.

Guidance & Outlook

MUFG made no change to any published figure. The full-year target of ¥2,100.0bn, the ¥2,250.0bn net operating profit target, the ¥350.0bn credit-cost assumption, the ¥74.00 dividend forecast and the medium-term plan all stand as set on November 14. The Summary Report's dividend section carries the line "Revision of forecasts for dividends on the presentation date of this Consolidated Summary Report: None."

Implied quarter-over-quarter ramp. The fourth quarter needs ¥286.5bn of net profit, ¥344.1bn of net operating profit and ¥490.8bn of ordinary profit, against a third quarter that delivered ¥520.6bn, ¥618.9bn and ¥762.6bn. The credit-cost assumption leaves ¥130.2bn for the quarter against ¥143.4bn just charged, which is the only one of the four that looks tight rather than loose.

Street position. The pre-print consensus for the nine months was ¥1,822.9bn of net profit and ¥2,513.3bn of ordinary profit, both marginally above what printed. The published sell-side average target price was ¥2,771.67 across eleven analysts, described as moderately bullish, against a January 28 close of ¥2,750. The shares finished the reaction session at ¥2,879.5, which is 3.9% above that consensus target.

Guidance style. Unchanged from the interim, and now more conspicuous. Management adopted the practice of revising at each disclosure point only in the prior fiscal year, and at the interim it admitted debating whether a 5% revision was worth making. Holding a target at 86.4% complete is consistent with that philosophy, and it means the fourth-quarter number will almost certainly be a beat that the market has already assumed. The information content of MUFG's target has now fallen close to zero, which is a problem for the company rather than for the earnings.

Market Reaction

• Release timing. MUFG published at 16:00 Tokyo time on Wednesday February 4, after the 15:30 close and before the New York open. The New York listing therefore traded the news on February 4 and the ordinary shares on February 5.
• Pre-print setup, ordinary shares. 8306 closed at ¥2,950 on February 4, up 18.3% year to date against 7.9% for the Nikkei 225, up 51.9% over twelve months and up 15.9% over the trailing thirty days, inside a 52-week closing range of ¥1,496 to ¥2,990. The shares had entered the print having recovered 7.5% across the two preceding sessions from a February 2 close of ¥2,744, and sat 1.4% below the ¥2,990 January 16 high.
• Pre-print setup, New York listing. The ADR closed at $18.52 on February 3, up 16.8% year to date against 1.1% for the S&P 500, up 48.3% over twelve months and up 13.5% over thirty days, inside a 52-week closing range of $10.79 to $18.90.
• New York reaction session, February 4. The ADR opened at $18.65, traded $18.42 to $18.89 and closed at $18.55, up 0.2%, on 3.9m shares against a 2.9m thirty-day average, a 1.4 times volume multiple. The S&P 500 fell 0.5%.
• Tokyo reaction session, February 5. The ordinary shares opened at ¥2,948, traded ¥2,854 to ¥2,948 and closed at ¥2,880, down 2.4%, on 65.2m shares against a 43.5m thirty-session average, a 1.5 times volume multiple. The Nikkei 225 fell 0.9%, putting the relative move at roughly minus 1.5 points.
• Megabank peers, same session. Mizuho Financial Group fell 1.9%, Sumitomo Mitsui Financial Group rose 0.4% and Sumitomo Mitsui Trust rose 0.3%. MUFG was the weakest of the four.

The two listings did not disagree despite the opposite signs. At the February 4 dollar-yen rate of 156.81, the ADR's $18.55 close implies ¥2,909 per ordinary share, 1.4% below Tokyo's ¥2,950 pre-news close. Tokyo then settled at ¥2,880, a further 1.0% below the ADR-implied level. New York took roughly half the mark-down overnight and Tokyo completed it the following session.

The reaction was a verdict on positioning, not on the quarter. The shares entered the print having rallied 7.5% in two sessions, 15.9% in thirty days and 51.9% in twelve months, and sitting a fraction below an all-time closing high. A result that landed half a percent below the nine-month consensus, with no target revision and no capital-return news, gave that positioning nothing to work with. The Japanese financial press characterised the print as in line with expectations and attributed the fall to profit-taking, which is the correct reading.

The relative trade is the more interesting one. Sumitomo Mitsui reported on January 30 and raised its full-year forecast by ¥200bn to ¥1,500bn, with nine-month progress at 93%. Mizuho reported on February 2 with 90% progress and no revision. MUFG reported on February 4 with 86.4% progress and no revision. For the second reporting round in a row, the largest of the three megabanks produced the least news, and for the second reporting round in a row the market ranked it last on the day. At the reaction close MUFG traded at 1.53 times book, having been overtaken on that measure by Mizuho during the same session.

The setup also carried a sector overhang. Japanese bank shares sold off sharply in the second half of January on Japanese government bond volatility and the losses it implies for bank securities portfolios, taking MUFG from ¥2,990 on January 16 to ¥2,730 on January 26. This release put a number on that concern rather than resolving it: held-to-maturity unrealised losses of ¥872.4bn and a ¥71.6bn realised loss on debt securities in the quarter.

Street Perspective

Debate: is a target left unmoved at 86.4% discipline, or an admission about the fourth quarter?

Bull view: the implied fourth quarter asks for ¥286.5bn after a ¥520.6bn third, with a credit-cost assumption that leaves ¥130.2bn against a run-rate of ¥143.4bn and a policy rate that has just moved up 25 basis points. On this reading the target carries ¥150bn or more of cushion and the full year lands well above ¥2,250bn.

Bear view: the third quarter contained ¥177.7bn from Morgan Stanley and ¥53.1bn of equity-securities gains, neither of which is controllable, and the fourth quarter carries the year-end cost push and the seasonal credit build. Management has more information than the market and chose not to move the number.

Our take: the bulls are right on the arithmetic and the bears are right that it no longer matters. A company that revises only for what has already happened has turned its target into a lagging indicator, and the market has learned to ignore it. The beat will come in May and will be discounted before it arrives.

Debate: is the balance-sheet growth a capital problem or a growth opportunity?

Bull view: loans grew ¥7.0tn in the quarter while near-zero-yielding central-bank deposits fell ¥8.1tn, at a loan-to-deposit ratio of 56.7%, in the first innings of a domestic rate cycle. Deploying idle liquidity into earning assets as rates rise is the correct use of this balance sheet, and the capital ratio is still inside its target range.

Bear view: risk-weighted assets have grown 9.9% in nine months, the target-basis ratio has fallen in each of the three quarters since the plan was set, and ¥4.7tn of the quarter's growth is overseas lending in a franchise that told investors it was moving to distribution. Americas non-performing loans rose 66% over the same nine months.

Our take: both, and the sequencing favours the bulls for now. Nothing in the ratio constrains the company over the next four quarters. What the trajectory does constrain is the capital-return policy set in May, and the market is pricing a ¥500bn run-rate that a balance sheet growing at this pace cannot sustain alongside a ratio at 10.3%.

Debate: how much of the earnings growth is the rate cycle and how much is the comparative?

Bull view: MUFG Bank's domestic net interest income grew 23.4% and its total net interest income grew 16.7% once investment-trust cancellation gains are removed from both years. The domestic loan-deposit spread reached 1.00% from 0.86%, at a policy rate that has since risen again.

Bear view: the reported nine-month operating profit increase of ¥191.2bn is ¥152.4bn the absence of prior-year bond losses and roughly ¥25bn currency, leaving very little that is business. Group net interest income grew 0.9% and interest income on loans actually fell.

Our take: the bears are describing the consolidated income statement and the bulls are describing the bank. Both are accurate and the bank is the one to underwrite, because the consolidated line is distorted by a Thai accounting change that annualises out and a bond loss that cannot repeat. The number that settles it is the six customer groups growing operating profit 9.7% currency-neutral with no help from either.

Debate: does 1.6 times book still work?

Bull view: the shares trade at 1.59 times a book value that grew 5.7% in nine months, on an underlying return on equity of 10.0% that management is targeting to take to 12% at a policy rate the Bank of Japan has now delivered three-quarters of. The dividend yield is 2.5% and the buyback is running.

Bear view: this is a bank that traded below book for most of the last decade, now at 1.59 times on one year of normalised earnings, with a published sell-side consensus target 7.7% below the current price and a capital ratio that has fallen three quarters running.

Our take: the multiple is defensible and is no longer cheap. At initiation in November the shares were at 1.32 times book and our value range implied roughly 16% of upside. At 1.59 times that gap has closed to about 1.5% even after raising the sustainable return assumption for the December rate move. The position still works. The margin of safety that justified building it has been substantially spent.

Model Update Needed

Valuation. Book value per share was ¥1,884.72 at December 31, from ¥1,783.37 at March and ¥1,834.3 at September, a gain of 5.7% in nine months. At the February 13 close of ¥3,003 the shares trade at 1.59 times book and 16.3 times the company's own full-year target, with a 2.5% forward dividend yield on the ¥74.00 forecast. On a sustainable return on equity of 11.5% to 13.0%, a cost of equity of 8.5% to 9.0% and 2% long-term growth, the justified multiple is 1.36 to 1.69 times book. Against a twelve-month book value of roughly ¥2,000, that is a ¥2,714 to ¥3,384 value range with a midpoint near ¥3,050, about 1.5% above the current price, or roughly 4% with the dividend. The equivalent ADR figure at the prevailing exchange rate is roughly $20.00 against a last close of $19.73, though any ADR target carries a currency view the yen-denominated one does not.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation on the FY2025 interim print and carried in our thesis of record. They are graded against what this quarter's disclosure showed, not re-derived.

Overall: thesis intact and less cheap. The rate-normalisation case got its strongest quarter of evidence yet and the policy rate moved up inside the period, which raises the sustainable return the franchise can earn. The capital pillar moved against us for the third consecutive quarter, and the concentration in one equity-method holding rose again. Nothing broke; the balance of the argument shifted from price to fundamentals, and the price moved further than the fundamentals.

Action: hold the position, stop adding. At 1.59 times book against an underlying return on equity of 10.0% and a policy rate three-quarters of the way to the level management's 12% target assumes, the case still clears the hurdle. It clears it by roughly 1.5% on price and roughly 4% on total return, against 16% when we initiated. Two things would change the view: a target-basis capital ratio that falls below 10% at March, which would put the ¥500bn capital-return run-rate in question before the new chief executive has to defend it; or a fiscal-2026 buyback set materially below ¥500bn in May.