Bank of Montreal (NYSE: BMO) — The Recovery Worked; the Currency Math Didn’t

Independent equity research · Report date: 2026-09-01 Primary listing used for valuation: TSX:BMO · Reporting currency: CAD · Fiscal year end: October 31 Prices at 2026-08-31 close: C$236.24 (TSX) / US$170.31 (NYSE) · Implied CAD/USD cross: 1.387 · Equity value: approximately C$165B / US$119B

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice, and not a recommendation to buy or sell any security. The analysis in the numbered sections below is deliberately position-free and carries no price target; this block is the single exception.

Verdict: HOLD / wait for a better entry / not a short. Conviction: HIGH on the valuation correction, MEDIUM on the earnings path. My current fair-value zone is C$205–230 on the TSX, or approximately US$148–166 at the 2026-08-31 exchange relationship. I would become interested below C$200 and would treat prices above C$250 as requiring unusually clean execution. At C$236.24, the stock is close enough to fair value that the dividend can carry a patient holder, but the risk/reward does not justify chasing a recovery whose main destination is already in the price.

The decisive change from the July 2 report is not BMO’s operations. It is that the earlier valuation was wrong. The prior memo labeled the NYSE price of US$174.13 as a TSX price of C$174.13, then divided that U.S.-dollar quote by Canadian-dollar book value. The actual July 2 TSX close was C$246.93. BMO therefore traded near 2.2 times book, not approximately 1.5–1.6 times. The earlier claim that the market paid only for a 12% return on equity and left the 15% plan unpriced cannot survive currency-consistent arithmetic. This report corrects the record explicitly rather than allowing a favorable operating quarter to conceal the error.

The corrected thesis is less exciting but more useful. BMO’s recovery is real: fiscal Q3 adjusted EPS rose 22%, adjusted ROE reached 14.0%, adjusted ROTCE reached 18.0%, total credit provisions fell to 41 basis points, and Canadian P&C earned a formidable 22.9% ROE. Yet the quote is 2.09 times reported book, 2.86 times a deliberately conservative tangible-book proxy, and 16.4 times trailing adjusted EPS. A justified price/book identity says 2.09 times book embeds sustainable ROE anywhere from 14.4% to 16.5% under plausible cost-of-equity and growth combinations. That is no longer a cheap option on management’s 15% objective; it is a security that needs much of the objective to endure.

The remaining debate is the U.S. bank. Canadian commercial banking has the financial fingerprint of a moat, but U.S. Banking earned only 9.8% adjusted ROE, with average deposits down 3% year over year and personal/business deposits down 7%. Two quarters of sequential commercial-loan growth are encouraging, not conclusive. The sale of acquired specialty-finance assets produced a C$962 million after-tax charge, primarily goodwill. Operational integration can be complete while acquisition economics remain inadequate. That distinction is the central analytical discipline here.

Capital return is similarly two-sided. BMO bought stock in Q3 at C$239.37, about 2.12 times book, and proposes authority for another 25 million shares. The trailing adjusted earnings yield at that price is only about 6%, below a reasonable 9%–10% equity hurdle. A full current-price authorization could lift constant earnings per share roughly 3.7% while mechanically reducing standalone book value per share about 4.1%. The buyback can work if U.S. returns and continuing earnings rise durably; EPS accretion by itself is not value creation.

What changes my mind: I turn materially more constructive if U.S. Banking sustains ROE above 11%, deposits stabilize, commercial loans compound for several quarters, and whole-bank ROE remains at least 14.5% after normalizing market-sensitive fees and divested earnings. I turn negative if total PCL returns above 55–60 basis points, U.S. ROE remains below 10% after the portfolio cleanup, or book value keeps shrinking through premium buybacks without a matching rise in sustainable earnings power. BMO is a good Canadian bank attached to an improving but still sub-hurdle U.S. franchise. At today’s price, that is a show-me proposition, not hidden value.

Tag: “The turnaround is visible; the margin of safety is not.”

Changes Since the 2026-07-02 Report

The prior report’s operating framework remains useful, but its valuation conclusion and absolute price history require correction.

1. Currency correction: the July 2 NYSE close of US$174.13 was mislabeled as C$174.13. The actual TSX close was C$246.93. Dividing the U.S. quote by Canadian book value understated P/B by roughly 30%. The prior “approximately 1.5–1.6 times book” thesis is withdrawn.
2. Price-map correction: the same listing/currency error affected historical TSX levels. The correct five-year TSX low was C$102.94 on October 27, 2023, not approximately C$74; the August 14, 2026 high close was C$257.88.
3. Q3 advanced the recovery: adjusted EPS rose to C$3.96, adjusted ROE to 14.0%, adjusted ROTCE to 18.0%, and PCL improved to 41 basis points. U.S. Banking ROE rose to 9.8% from 9.3% sequentially.
4. Acquisition economics suffered a real mark: BMO recorded a C$1.092 billion pre-tax/C$962 million after-tax charge, primarily goodwill, on the Transportation and Vendor Finance sale. Adjusted earnings exclude it; a capital-allocation assessment cannot.
5. Portfolio pruning accelerated: the branch sale transfers US$5.3 billion of deposits and US$0.7 billion of loans; Transportation/Vendor Finance removes US$9.2 billion of U.S. loans and C$1.7 billion of Canadian loans; BMO and RBC agreed to sell Moneris for about C$2.0 billion. The combined transactions are expected to release approximately 50 basis points of CET1.
6. Capital regulation was clarified: OSFI’s June reduction of the Domestic Stability Buffer to 3.0% makes the published D-SIB CET1 stack 11.0%, not the 11.5% stated previously. BMO’s 13.0% ratio therefore has 200 basis points of headline headroom.
7. The prior test score is mixed: the favorable PCL and whole-bank ROE thresholds now pass, but U.S. ROE remains below the prior 11% bear-falsification threshold and sustained loan compounding is not established. The bear case is weakened, not falsified; the end-FY2027 bull deadline has not arrived.

Sources: BMO Q3 2026 report, AZI TSX history, OSFI June 2026 DSB decision.

📈 Stock Price Action — Five-Year Event Map

BMO’s clean five-year TSX record begins at C$127.01 on September 1, 2021. It peaked at C$153.72 in March 2022, then the unadjusted price fell 33% to C$102.94 in October 2023 while the dividend-adjusted return was -26.8%. It more than doubled from the April 2025 tariff shock to an all-time closing high of C$257.88 on August 14, 2026. The August 31 close of C$236.24 was 8.4% below that high, 3.0% below the 50-day exponential moving average, and 11.3% above the 200-day average. The long trend remains positive; the near-term posture is a pullback.

Price moves are measured facts; causal labels are interpretations except where tied directly to result-day disclosures. Adjusted closes are used for return arithmetic and unadjusted TSX closes for displayed C$ levels. Sources: AZI BMO.TO data, BMO presentations and events, and contemporary coverage of the May 2024 and August 2024 results.

Factor evidence supports a cohort rather than purely company-specific reading. A July 31 all-factor model for the NYSE line loaded most heavily on the market, Canada, dividend yield, banks/financials, and value, with slightly negative quality and no selected momentum exposure; the model explained about 60% of return variation. Over the year to August 31, Canadian-country, dividend, and value regimes were favorable, but Canada reversed sharply over the latest ten days. The interpretation is an extended Canada/yield/value exposure undergoing a pullback—not a classic momentum crash in a uniquely high-quality compounder. This evidence is secondary and model-dependent. Because the vendor models BMO’s NYSE U.S.-dollar return, CAD/USD and Canadian-country exposures are embedded; its beta must not be merged with TSX-series statistics. The closest modeled peers—RBC, TD, CIBC, and Scotiabank—independently validate the fundamental comparison set. FactorsToday methodology and BMO loadings endpoint.

1. Executive Summary

Bank of Montreal is a C$1.5 trillion diversified North American bank with four earnings engines: Canadian Personal & Commercial Banking, U.S. Banking, Wealth Management, and Capital Markets. Its Canadian business participates in a protected, scale-intensive oligopoly; its U.S. business is a contested regional/commercial franchise enlarged by the US$16.3 billion Bank of the West acquisition in 2023. The portfolio is more geographically and commercially diverse than a domestic-only Canadian bank, but the U.S. leg also lowers consolidated returns and increases credit/funding complexity.

Q3-FY2026 demonstrates meaningful operating progress. Adjusted net income was C$2.859 billion, adjusted EPS C$3.96, adjusted ROE 14.0%, and adjusted ROTCE 18.0%. Total PCL declined to C$722 million/41 basis points from 47 basis points a year earlier. Revenue grew faster than expense, although adjusted operating leverage was only 1.6%. The share count fell 2.7% year over year. These are not cosmetic improvements.

Quality is uneven by segment. Canadian P&C generated C$983 million adjusted income, 22.9% ROE, a 42.8% efficiency ratio, and 7% commercial-deposit growth. U.S. Banking generated C$925 million adjusted income, but only 9.8% ROE; average loans rose 1% year over year while deposits fell 3%. Wealth and Capital Markets delivered strong earnings, yet their growth was assisted by markets, an acquisition, and equities trading. Approximately half of positive segment income came from lower-return U.S. Banking plus cyclical Capital Markets.

Credit has improved without becoming benign. Total PCL is within management’s near-term range and gross impaired loans declined sequentially, but new impaired formations edged higher. Canadian P&C now bears more pressure than the U.S. segment: 58 basis points of PCL versus 30 basis points in U.S. Banking, where a performing-loan reserve release helped. Total allowance for credit losses of C$6.030 billion equals 88.6% of gross impaired loans, but only C$1.268 billion is impaired allowance; the rest is performing allowance. Scope matters when interpreting coverage.

Capital appears ample on regulatory measures: CET1 is 13.0% versus an 11.0% supervisory stack, LCR 125%, and NSFR 114%. A C$6.064 billion fair-value shortfall on amortized-cost securities is outside book value and CET1. It is not an imminent realized loss; a deliberately harsh full-deduction sensitivity would reduce CET1 to about 11.7%. That narrows, but does not eliminate, the capital cushion.

Valuation is the pivotal correction. At C$236.24, BMO trades at 2.09 times C$113.06 BVPS, 2.86 times a conservative C$82.59 tangible-book proxy, and 16.4 times C$14.39 trailing adjusted EPS. Its 2.09 times book multiple is close to or modestly above a simple Big Six P/B-versus-ROE curve. Depending on cost of equity and long-run growth, the market embeds approximately 14.4%–16.5% sustainable ROE. BMO’s recovery is therefore visible in both earnings and price.

The thesis reduces to one question: can BMO convert portfolio pruning, commercial-loan momentum, fee growth, and efficiency into U.S. ROE of at least 11%–12% without weakening deposits or relying on unusually favorable market revenue? If yes, the current premium can be sustained. If not, the stock has little valuation protection from a return or credit normalization.

2. Business Overview

BMO’s business model is diversified by geography and revenue source, but each engine has a different economic character.

Canadian P&C is the moat-bearing anchor. It combines consumer deposits, mortgages, cards, small-business products, and a strong commercial-banking franchise. In Q3 it represented about 32% of positive segment adjusted income. Commercial relationships are especially valuable because operating accounts, treasury services, payments, lending, and advice create multiple switching costs. Personal banking benefits from the same distribution and technology scale, though falling personal deposits show customer captivity is not absolute.

U.S. Banking combines legacy BMO Harris with Bank of the West. It operates across the Midwest, California, and other western markets, with a meaningful commercial orientation. It contributed roughly 30% of positive segment income in Q3. The attraction is a much larger addressable market and the ability to cross-sell commercial, wealth, treasury, and capital-markets services. The cost is weaker industry structure: BMO competes with money-center banks, super-regionals, and local banks without Canada’s protected concentration. Its current return confirms the difference.

Wealth Management includes private wealth, asset management, direct investing, and insurance. It contributed roughly 16% of positive segment income. Client assets and referrals provide attractive fee economics, while markets create unavoidable cyclicality. Burgundy Asset Management expanded the franchise and boosted reported asset growth; acquisition effects should be separated from organic net flows.

Capital Markets includes investment and corporate banking and Global Markets. It contributed roughly 21% of positive segment income. BMO has genuine niches in Canadian equity capital markets and global metals and mining. Relationship capital and sector expertise matter, but talent is mobile and trading/underwriting revenue is cyclical. Euroz Hartleys adds Australian metals-and-mining capability rather than transforming the segment.

The business mix is a strength in stress diversification but a complication in quality assessment. Canadian P&C earns well above the cost of equity. Wealth is capital-light but market-sensitive. Capital Markets can clear its hurdle over a cycle but is volatile. U.S. Banking remains below a plausible equity hurdle. Consolidated ROE is therefore not simply a function of growth; it depends on mix, capital intensity, and where incremental balance-sheet dollars are deployed.

Q3’s positive-segment mix—Canadian P&C C$983 million, U.S. Banking C$925 million, Wealth C$480 million, and Capital Markets C$649 million—also explains why BMO can report strong group results before the U.S. bank fully works. Diversification buys time, but it can obscure the economics of the acquisition that was supposed to create the next return engine. BMO Q3 segment tables provide the primary data.

The revenue model deserves a bank-specific reading. Net interest income is the spread between income on loans/securities and the cost of deposits/wholesale funding. Fee income includes payments, cards, advice, investment management, underwriting, trading, and insurance. Loan growth is useful only when deposits, pricing, and credit quality make the incremental spread adequate. Deposit growth is useful only when customers do not demand away the economic value through high rates or promotional pricing. This is why BMO’s U.S. loan acceleration and deposit contraction must be analyzed together rather than celebrated separately.

Balance-sheet leverage also changes the meaning of accounting returns. A small deterioration in credit costs can consume a large portion of pre-provision earnings, while a modest release can make earnings growth appear spectacular. Conversely, a high-quality commercial relationship can produce spread, payments, treasury, and advisory revenue from the same client. BMO’s best growth opportunity is not raw assets; it is deepening multi-product relationships without increasing loss content or expensive funding.

Corporate Services is more than a residual. It houses technology/operations and enterprise items, and it absorbed much of the current divestiture charge. Q3’s C$178 million adjusted loss reduced the sum of segment profits. Investors should resist allocating every corporate cost away: a four-engine bank requires shared infrastructure, liquidity, cybersecurity, compliance, and central capital. A segment can look attractive before these costs while the consolidated return remains ordinary.

The U.S. dollar introduces a second accounting layer. BMO reports in Canadian dollars, but U.S. revenue, loans, deposits, and earnings translate with CAD/USD. Currency can lift reported Canadian-dollar earnings while leaving source-currency economics unchanged, and it creates large positive foreign-translation AOCI. For that reason, U.S. operating growth is assessed in U.S. dollars, consolidated valuation in Canadian dollars, and the NYSE quote only after translation. This convention prevents the exact mistake corrected in this report.

3. Industry Dynamics

Canadian banking is structurally attractive because entry, ownership, capital, technology, compliance, and distribution impose high fixed costs. A small group of federally regulated banks commands national scale and dense customer relationships. The result is rational competition, durable deposit franchises, and through-cycle returns above the cost of equity. This advantage belongs to the industry before it belongs to BMO.

The Greenwald test is local. A high national market share does not automatically create advantage in every product or geography. BMO’s Canadian commercial bank has the strongest company-specific evidence: management reports stable number-two commercial-loan and number-three commercial-deposit positions, deep treasury/payment relationships, and improving efficiency. Q3’s 22.9% segment ROE and 7% commercial-deposit growth are the observable financial fingerprint. Public data do not provide a consistent independent annual market-share series sufficient to prove that share moved less than two percentage points, so the stability claim remains directional rather than precisely scored.

The U.S. market has different economics. Scale is valuable, but regional banking is not supply-constrained in the Canadian sense. Deposits are contested, commercial customers can multi-bank, and larger competitors possess lower technology/compliance unit costs. BMO’s sub-10% U.S. return, falling deposits, and branch/asset sales are inconsistent with a franchise-wide moat today. Local relationship pockets may be attractive without making the consolidated U.S. platform advantaged.

The capital cycle supports this distinction. BMO expanded aggressively by paying US$16.3 billion for Bank of the West, issuing equity, and absorbing integration costs. It is now in harvest-and-prune mode: selling branches and specialty finance, recognizing lost goodwill, releasing capital, and selectively reinvesting in fee businesses. This is disciplined ex post if new uses clear the hurdle. It is also evidence that part of the expansion attracted capital at poor returns. Capital-cycle analysis asks whether management is harvesting after overinvestment, not merely whether reported growth has resumed.

Regulation remains a stabilizer. OSFI cut the Domestic Stability Buffer from 3.5% to 3.0% in June 2026. Adding the 4.5% minimum, 2.5% conservation buffer, 1.0% D-SIB surcharge, and 3.0% DSB produces an 11.0% CET1 supervisory stack. The relief applies to every domestic systemically important bank; it is sector capacity, not a BMO moat. BMO retained its 12.5%–13.0% operating range rather than immediately lowering its capital objective. OSFI’s decision also emphasizes household debt and housing vulnerabilities.

The macro backdrop is mixed. The Bank of Canada held its policy rate at 2.25% in July, described the economy as weak but improving, and highlighted trade and geopolitical risks. Its assumptions placed average U.S. tariffs on Canada at 5.0% and Canadian tariffs on the United States at 1.5%, with most North American trade still tariff-free. The direct tariff rate is less important for banks than second-order effects on employment, business investment, borrower cash flow, and collateral. BMO’s commercial mix makes those transmission channels material. See the July 2026 Monetary Policy Report.

Canadian housing risk is structurally different from U.S. subprime history. Mortgages commonly reset after shorter contractual terms, exposing borrowers to renewal shocks, while high-loan-to-value loans are generally insured and mortgages are typically full recourse. Those protections reduce ultimate loss severity but do not prevent payment stress, delinquencies, slower consumption, or unsecured-credit losses. The relevant BMO indicators are renewal payment changes, Canadian P&C impaired PCL, cards, insolvencies, unemployment, and performing allowance—not house prices in isolation.

Rate movements have offsetting effects. Falling policy rates can reduce asset yields and deposit costs with different lags, affecting NIM; they may also relieve borrowers and increase capital-markets activity. A steeper curve may improve reinvestment economics while marking securities. BMO’s Q3 U.S. NIM of 4.02% was 20 basis points higher year over year but one basis point lower sequentially, illustrating that balance mix and deposit competition can matter more than the direction of the policy rate alone.

Industry concentration should not be confused with immunity from capital-cycle mistakes. Canadian banks have repeatedly sought growth outside their protected home market because domestic share is difficult to add. That pushes capital toward U.S. regional banking, wealth, and international assets where barriers and return structures differ. When every institution pursues the same external growth, acquisition prices and competitive intensity can rise. Bank of the West is therefore best evaluated as a move from a constrained, high-return market into a larger but less protected one—not as a simple extension of Canada’s moat.

The supply-side question now reverses. BMO is withdrawing capital from single-digit-return branches and specialty finance while regulators release system capacity. If peers all use that capacity for loans or buybacks, the benefit can be competed away through pricing or richer acquisition/repurchase multiples. The favorable outcome is disciplined scarcity: capital remains selective and margins/returns hold. The unfavorable one is synchronized deployment that raises asset prices faster than economic profit. BMO’s own repurchase multiple makes this more than a theoretical risk.

4. Competitive Position

BMO has a strong Canadian commercial moat, a shared Canadian retail moat, niche fee advantages, and an unproven U.S. moat. Treating the whole company as either advantaged or undifferentiated loses the relevant market boundaries.

Canadian commercial banking is the clearest strength. BMO can spread technology, compliance, risk, and product costs across a national platform while binding customers through operating deposits, credit, payments, treasury, and advice. Commercial loans grew 3% and deposits 7% in Q3. The combination of high returns, relationship depth, and efficiency supports customer captivity and scale economies. It would be difficult for a new entrant to reproduce this network at economic cost.

Canadian personal banking shares the Big Six structural advantages: trusted brand, branch/digital reach, bundled accounts, recurring payments, mortgage relationships, and regulated scale. Yet personal/business deposits fell 5% year over year. Digital tools and artificial intelligence may improve cost and service, but peers can adopt the same technologies. They reinforce scale; they are not independent barriers.

Wealth benefits from referrals, advisor relationships, and tax/estate complexity. Q3 revenue rose 16%, AUM 33%, and AUA 18%, but management attributed much of the increase to stronger markets, client assets, and Burgundy. Expense rose 23%, adjusted ROE fell from 52.2% to 42.4%, and efficiency worsened. The franchise is attractive, though the quarter is not a clean organic-growth sample.

Capital Markets has credible relationship/intangible advantages in Canada and metals and mining. Q3 revenue rose 20%, adjusted income 46%, and ROE reached 16.1%; Global Markets grew 27%, mainly on equities trading. That proves relevance, not permanence. Underwriting calendars, trading conditions, and employee mobility limit the durability of excess returns.

The U.S. bank fails a strict present-tense moat test. Revenue and adjusted income rose 5% and 9% in U.S. dollars, NIM improved 20 basis points year over year to 4.02%, and sequential loan growth was encouraging. But average deposits fell 3%, including 7% in personal/business and 9% in wealth, while ROE reached only 9.8%. The pending sale of 138 branches removes US$5.3 billion of deposits for only US$0.7 billion of loans. The roughly 5% deposit premium confirms that funding franchises have scarcity value; it also makes replacement funding an important follow-up as loans grow faster than deposits.

Peer returns show the consolidated consequence. BMO’s 14.0% adjusted ROE trails RBC’s 18.1%, CIBC’s and National Bank’s 16.8%, and TD’s 16.0%, and is near Scotiabank’s 14.2%. BMO deserves a discount to the highest-return franchises until its U.S. economics improve. The protected Canadian profit pool prevents a low-quality label; the U.S. denominator prevents a top-quality one.

A useful counterfactual is to ask what would deteriorate without the advantage. Without Canadian commercial scale, BMO would lose low-cost operating deposits, treasury fees, cross-sell, and cost efficiency; the economics would clearly worsen. Without the U.S. network at its present performance, consolidated size would fall but returns might rise. That does not mean the U.S. bank has no option value. It means the burden of proof is incremental economic profit, not footprint.

The same counterfactual disciplines the “eighth-largest North American bank” narrative. Size is an input. It creates value when it lowers unit costs, broadens products, and increases customer captivity more than it raises complexity, compliance, funding, and credit costs. BMO has demonstrated the integration synergies; it has not yet demonstrated the required post-integration U.S. return. The next two to four quarters should be judged on that output.

5. Growth History and Forward Opportunities

Reported history is distorted by Bank of the West. FY2022 included a large purchase-price hedge gain; FY2023 absorbed acquisition closing, day-one credit provisions, integration costs, and the Canada Recovery Dividend; FY2024 brought a U.S.-led credit shock; FY2025 began normalization. Reported diluted EPS moved from C$19.99 in FY2022 to C$5.76 in FY2023, C$9.51 in FY2024, and C$11.44 in FY2025. The series is less useful than normalized earnings, segment returns, credit, and per-share book progression.

The table shows why a single trailing P/E can mislead in both directions. FY2022 earnings were not a durable peak; FY2023 was not a durable trough; Q3-FY2026 reported earnings include a real but non-recurring capital-allocation loss. The proper normalized base is neither the highest reported year nor whichever adjusted number produces the most favorable multiple. It is a through-cycle return applied consistently to the equity required to generate it.

FY2025 provides a cleaner baseline. Revenue was C$36.274 billion, PCL C$3.617 billion, pre-tax income C$11.550 billion, net income C$8.725 billion, and diluted EPS C$11.44. Gross loans were essentially flat year over year and deposits declined modestly. This was earnings repair, not broad balance-sheet growth. BMO’s FY2025 annual report is the source.

Five forward levers matter:

1. U.S. balance growth: commercial loans grew 4% sequentially in both Q2 and Q3 and became positive year over year. Management attributes about one-third of the bridge from roughly 10% to 12% U.S. ROE to client balances. Several more quarters are needed to establish compounding.
2. Fees and cross-sell: another third of the U.S. bridge is expected from fees. Treasury, payments, wealth, and Capital Markets referrals can raise revenue without equivalent balance-sheet growth, but disclosure should show that the fees originate from continuing clients rather than markets alone.
3. Efficiency and credit: the final third of the U.S. bridge is management’s efficiency/PCL normalization. This is plausible after integration and portfolio pruning, but normalization creates less economic value than new franchise revenue.
4. Capital-light wealth and niches: Burgundy and Euroz Hartleys extend attractive capabilities. Markets and acquisitions inflate near-term reported growth, so net flows, fee margins, and expense discipline are better quality indicators than AUM alone.
5. Divestiture/redeployment: selling single-digit-return assets can raise consolidated ROE by releasing 50 basis points of CET1 and shrinking the denominator. Continuing earnings must replace the 3% of adjusted Q3 net income contributed by the sold businesses. A higher ratio created by disposals is not equivalent to organic profit growth.

Management targets 15% or better ROE exiting FY2027. The current 14.0% quarter places the bank near the destination, but composition matters. Canadian P&C and Capital Markets already clear the hurdle; U.S. Banking does not. Wealth is high-return but market-assisted. The best confirmation would be simultaneous U.S. loan/deposit growth, an 11%–12% U.S. return, positive operating leverage, and credit remaining in the low-to-mid-40 basis-point range.

Growth quality is therefore mixed-positive. BMO has credible self-help and cross-sell opportunities, yet the fastest Q3 growth came from businesses with market sensitivity, while the most strategically important engine remains below its hurdle. The stock’s re-rating makes the distinction more important than the headline growth rate.

The Investor Day plan can be translated into falsifiable operating evidence. Balance growth should appear in average source-currency loans and deposits, not only end-period figures. Fee growth should appear in continuing-business revenue without a matching rise in acquisition or market beta. Efficiency should generate cumulative positive operating leverage after restructuring charges. Credit normalization should hold even as risk appetite and loan growth return. Capital deployment should raise total economic profit and per-share book over time, not only ROE through a smaller denominator.

This translation also prevents double counting. Selling low-return assets can improve U.S. and group ROE; using the released capital for premium buybacks can improve EPS; lower performing PCL can improve net income. All three may occur without equivalent underlying revenue growth. A valuation that already embeds mid-teens returns should pay once for the outcome, not separately for each accounting route to it.

6. Financial Quality

Earnings and normalization

Q3 reported net income/EPS/ROE were C$1.750 billion/C$2.38/8.4%; adjusted figures were C$2.859 billion/C$3.96/14.0%. The C$1.109 billion after-tax gap is unusually large. It includes the C$962 million goodwill-related divestiture charge, Burgundy contingent-consideration remeasurement, acquisition-intangible amortization, and smaller integration/divestiture items. Adjusted results better describe continuing operating momentum; reported results preserve the acquisition’s capital cost. Both are necessary.

Year to date, reported net income was C$6.869 billion and adjusted net income C$8.143 billion, a C$1.274 billion gap versus C$304 million a year earlier. Adjusted revenue grew 9%, expense 8%, and operating leverage was 1.0%. Q3 adjusted operating leverage was 1.6%. The recovery has positive leverage, though not yet enough to call efficiency the dominant earnings driver.

Trailing adjusted EPS is approximately C$14.39, combining Q4-FY2025 C$3.28 with Q1–Q3-FY2026 C$11.11. Annualizing year-to-date EPS produces C$14.81; annualizing Q3 produces C$15.84. The three figures frame current earnings power rather than offering a forecast. The gap between trailing and quarterly run rate is precisely what the valuation already capitalizes.

Segment quality

Credit

Total PCL of C$722 million/41 basis points improved from C$739 million/45 basis points sequentially and C$797 million/47 basis points a year earlier. Impaired PCL was C$708 million; performing PCL only C$14 million. The year-to-date improvement is largely reserve-build normalization: performing PCL fell to C$26 million from C$465 million, while impaired PCL also declined.

Gross impaired loans were C$6.803 billion/0.97%, down from C$6.939 billion/1.01% sequentially. New formations, however, rose slightly to C$1.452 billion. Falling balances with rising formations can occur when resolutions and write-offs outpace additions; it is progress, not an all-clear.

Credit has rotated geographically. Canadian P&C recorded C$507 million/58 basis points of PCL, including C$60 million performing PCL. U.S. Banking recorded C$173 million/30 basis points, helped by a C$50 million performing-loan release. Canadian cards and consumer insolvencies remain a watch item even as U.S. wholesale credit improves.

Allowance totaled C$6.030 billion: C$4.762 billion performing and C$1.268 billion impaired, with C$5.247 billion recorded against loans and C$783 million in other liabilities. Total ACL/GIL is 88.6%; loan ACL/GIL 77.1%; impaired ACL/GIL 18.6%; loan ACL/gross loans 0.746%. Those ratios answer different questions and should not be substituted for one another.

Reserve interpretation requires three distinctions. First, performing allowance estimates losses not yet individually impaired; a decline in new performing PCL can lift earnings even when the stock of allowance remains. Second, impaired allowance is not designed to cover gross impaired loans dollar for dollar because collateral, guarantees, recoveries, and expected cash flows matter. Third, GIL is a balance while PCL is a period expense. A bank can report lower GIL and still have high formations if resolutions exceed additions. For BMO, the favorable facts are lower total PCL and sequential GIL; the cautions are slightly higher formations and a Canadian shift in pressure.

The U.S. performing-loan release makes geographic comparison especially important. A 30 basis-point U.S. PCL ratio looks much better than Canada’s 58 basis points, but US$ source-currency deposit contraction and commercial cyclicality remain. Conversely, Canadian consumer stress is more visible now, but full-recourse lending and insurance reduce loss severity. Neither segment should be extrapolated from one quarter’s total ratio.

Capital, liquidity, and book quality

CET1 capital was C$59.273 billion on C$454.757 billion RWA, producing 13.0%. Tier 1 was 14.7%, total capital 16.6%, leverage 4.2%, TLAC 29.2%, LCR 125%, and NSFR 114%. Internal capital generation in Q3 was offset by buybacks and higher source-currency RWA, leaving CET1 flat.

Common equity was C$78.817 billion, or C$113.06 per share. Goodwill was C$16.086 billion and intangibles C$5.158 billion. BMO does not report end-period TBVPS in the filed table. Deducting all goodwill and gross intangibles produces a deliberately conservative C$82.59 proxy; BMO’s own TCE definition deducts acquisition-related intangibles net of deferred tax liabilities, so the official economic figure should be modestly higher. The proxy is useful only if labeled.

AOCI was positive C$7.207 billion, but C$6.759 billion came from foreign-operation translation and C$1.288 billion from pensions; cash-flow hedges were negative C$674 million and FVOCI only positive C$69 million. Positive AOCI is therefore not a clean securities cushion.

Amortized-cost debt securities carried at C$92.978 billion had fair value of C$86.914 billion, a C$6.064 billion unrealized shortfall outside book value/AOCI/CET1. A full, unhedged deduction would leave CET1 near 11.7%, about 70 basis points above the current requirement. That sensitivity assumes realization and ignores hold-to-collect economics, hedging, and funding. It is a tail-capacity lens, not a forecast or regulatory measure.

Book value itself requires care. Reported common equity includes positive translation AOCI, while tangible common equity removes acquisition-related intangibles. Using reported BVPS with reported ROE is internally consistent for justified P/B; using tangible book requires a matching ROTCE and denominator definition. Mixing reported price/book with adjusted ROTCE can overstate cheapness, just as mixing currencies did. This report therefore leads with reported book/reported or adjusted ROE bridges and labels the gross-intangible TBV proxy separately.

The bank’s own reported adjusted ROTCE of 18.0% is economically informative but not directly comparable to the 2.86 times conservative proxy because BMO’s TCE denominator adds related deferred tax liabilities and uses average balances. A correct tangible-book valuation would require the same official denominator across time and peers. The absence of end-period TBVPS is a reason for restraint, not an invitation to invent precision.

7. Capital Allocation

BMO’s capital-allocation record is the principal reason to distinguish adjusted operating performance from shareholder value creation.

Bank of the West: the US$16.3 billion acquisition expanded scale and geography, and management exceeded cost-synergy targets. Yet Q3/YTD U.S. adjusted ROE remains 9.8%/9.2%; deposits are shrinking; acquisition-intangible amortization persists; and selling acquired specialty finance produced a C$962 million after-tax goodwill-related loss. Integration execution may be good while the purchase price fails its hurdle. The return verdict remains open.

Portfolio sales: selling 138 branches at an approximately 5% deposit premium demonstrates scarcity value and frees resources, but removes low-cost funding. The Transportation/Vendor Finance sale removes a large loan portfolio and retains a 19.9% stake. Moneris should generate an estimated C$600 million after-tax gain for BMO. Combined, the announced sales should add about 50 basis points to CET1 and remove businesses that contributed 3% of Q3 adjusted income. Future comparisons must separate operating improvement, lost earnings, disposal gains/losses, and a lower capital denominator.

Buybacks: FY2025 repurchases were 22.2 million shares at C$152.97 for C$3.461 billion including tax—not C$3.24 billion as previously stated. Alongside C$4.630 billion of common dividends, common capital return totaled C$8.091 billion, 93% of net income attributable to shareholders before preferred distributions. This leaves limited retained organic capital.

Q3 repurchases at C$239.37 equaled 2.12 times BVPS and roughly 2.90 times the conservative tangible-book proxy. At that execution price, trailing adjusted earnings yield about 6.0% and annualized Q3 earnings yield about 6.6%, below a 9%–10% equity hurdle before growth. The proposed 25 million-share authorization equals about 3.6% of shares. If executed immediately at C$236.24 with earnings/book static, it would cost C$5.91 billion, lift constant EPS about 3.7%, and reduce standalone BVPS about 4.1% to C$108.48 before subsequent earnings, dividends, AOCI, RWA, or option issuance. Monitoring both EPS and total/per-share equity is essential.

Dividend: C$1.71 quarterly implies C$6.84 annually and a 2.90% yield. The payout is well covered by adjusted earnings. Its long record signals franchise durability, though the current yield offers less downside cushion than a typical high-yield bank setup.

Governance and incentives: CEO Darryl White received C$15.445 million of FY2025 direct compensation, including C$14.245 million variable, on a 117% bank multiplier; 75% of variable pay was deferred. He disclosed C$55.437 million of shares/DSUs/PSUs, 46 times salary. Thirteen of 14 director nominees are independent, chair/CEO roles are separate, and say-on-pay passed 96.34%. Formal alignment is strong.

The caveat is adjustment design. Compensation ROE/EPS exclude performing-loan PCL, and an earlier PSU result excluded Bank of the West goodwill/intangibles and unplanned regulatory-capital changes. These choices reduce incentive exposure to reserve building and acquisition accounting. A secondary SEDI-derived report also shows White sold 27,167 shares near C$225.74 in June; it is incomplete and partly award/tax-related, so it should not override his substantial remaining alignment. Sources: 2026 proxy, 2026 vote results, secondary insider report.

The preferred capital-allocation hierarchy is straightforward. First, retain enough capital and allowance to absorb stress without forced balance-sheet contraction. Second, fund organic relationships that earn above the cost of equity after credit and operating costs. Third, reinvest in technology and talent where measurable productivity or fee growth follows. Fourth, pay a sustainable dividend. Repurchase shares only when their earnings yield plus defensible growth exceeds the hurdle and when per-share equity damage is acceptable. Acquisitions come last because integration and purchase-price risk are highest.

BMO’s current actions straddle that hierarchy. Divesting sub-hurdle assets is rational, and the dividend is covered. Organic U.S. commercial growth could be attractive if funded by stable relationship deposits. The proposed repurchase authorization is less obviously economic at a 6%–7% earnings yield. Euroz Hartleys is strategically coherent but has no disclosed purchase price or return hurdle, making ex ante discipline impossible to verify publicly.

Management’s statement that released capital can be redeployed at 15% or better returns should be treated as a testable hypothesis. A buyback at 2.1 times book is economically equivalent to acquiring a claim on BMO’s own earnings at roughly a 6% current yield. It clears a 15% return hurdle only if earnings grow enough and the premium remains supported. That may happen, but the arithmetic should be shown rather than assumed.

8. Changes and Headwinds — Last Two Years

The net change is favorable operationally and less favorable financially than headline adjusted EPS suggests. Credit, efficiency, and loan momentum improved. At the same time, the divestiture charge crystallized a capital loss, funding weakened, and the share-price re-rating absorbed much of the recovery.

9. Risk Analysis

The most dangerous combination is not any single line. It is simultaneous U.S. deposit pressure, commercial credit deterioration, and premium buybacks: funding raises the cost base, losses reduce earnings/capital, and repurchases shrink book value. The most favorable combination is U.S. loan/fee growth with stable deposits, low-40s PCL, and capital redeployed into returns above 15%.

Risk correlations make a simple matrix look safer than the business can be. A trade shock can reduce commercial loan demand, weaken borrower cash flow, increase provisions, lower capital-markets activity, and prompt rate cuts that pressure spreads—all while CAD/USD moves reported earnings and capital. Premium buybacks executed before that shock would reduce the starting book cushion. Stress analysis should therefore combine variables rather than add isolated downside cases.

Concentration disclosure matters at the portfolio level. Commercial and industrial credit is diversified by borrower but can share macro drivers; CRE is collateralized but refinancing-sensitive; Canadian mortgages have low historical losses but transmit renewal pressure into unsecured consumption. The most useful early-warning dashboard is formations, impaired PCL by segment/product, allowance migration, criticized/watchlist balances where disclosed, deposit flows, and source-currency NIM. Headline gross loans are too blunt.

Operational and cyber risk deserves equal seriousness despite lower reported frequency. Scale provides larger security budgets but also creates a broad attack surface and complex integration estate. Bank of the West systems, data, third parties, payments, and model-driven decisions expand the control perimeter. A severe outage or breach can create remediation cost, regulatory limits, deposit attrition, and reputational harm without first appearing in traditional credit metrics.

Regulatory risk is likewise endogenous. OSFI can raise buffers as vulnerabilities build; U.S. regulators can impose operational or capital constraints; Basel implementation can increase RWA without asset growth. The June DSB cut is reversible capacity, not permanent excess capital. A repurchase program should be evaluated against stressed and future requirements, not only today’s 200-basis-point headline surplus.

BMO’s 2026 U.S. company-run stress test provides useful but limited comfort. The U.S. subsidiary’s CET1 is projected to fall from 14.0% to a 12.0% minimum under approximately 10% unemployment, a 30% housing decline, and a 39% CRE decline; projected loss rates include 5.9% C&I and 7.2% CRE. It supports subsidiary resilience but does not replace consolidated stress analysis. BMO U.S. stress test.

10. Valuation Discussion (Embedded Expectations)

Banks should be valued primarily on earnings, book/tangible book, return on equity, growth, and capital—not enterprise value or EBITDA. Deposits and wholesale funding are operating inputs rather than ordinary financing. All per-share arithmetic below matches TSX CAD price to CAD financial statements.

Current snapshot

Justified book value

For a mature bank, P/B = (ROE - g) / (cost of equity - g). Solving for sustainable ROE at the observed 2.09 times book produces:

The algebra is exact; the assumptions are not. A 5% perpetual growth rate for a mature bank is generous, but a 9.5% equity cost may be reasonable for a protected Canadian franchise. A 4% growth rate and 10% cost are more conservative and require returns above management’s 15% goal. Across the grid, current valuation requires at least Q3-like returns and often more.

Peer curve

Prices are August 31 TSX closes; book and earnings use current Q3 filings. Adjustment policies differ. A simple six-bank P/B-on-Q3-ROE regression has approximately 0.92 R-squared and places BMO near 1.98 times versus 2.09 times observed, a +0.11 times residual. This is too small and model-dependent to call a mispricing, but it shows no obvious BMO discount. The stock is cheaper than higher-return peers and modestly dearer than Scotiabank despite a slightly lower quarterly ROE.

Operating scenarios

These are return/multiple regimes, not share-price forecasts.

Observed 2.09 times lies above the through-cycle regime and below the full high-return case. The market underwrites more than an ordinary 13%–14% bank but not the complete 15.5%-plus outcome. The biggest swing variable is not the Canadian franchise; it is whether U.S. Banking crosses its hurdle without expensive replacement funding.

Return decomposition and sensitivity

Long-run shareholder return for a bank can be approximated as starting dividend yield plus sustainable book-value growth plus change in the P/B multiple. At today’s 2.90% dividend yield, a base case needs several percentage points of per-share book/earnings growth merely to reach a conventional equity return; multiple expansion is not a free assumption when P/B already exceeds a through-cycle justified level.

Retained earnings do not automatically equal book growth because dividends, repurchases above book, AOCI, FX, option issuance, and acquisition/disposal accounting intervene. If BMO earns 14% on book but distributes most earnings and repurchases at a premium, BVPS growth can lag the headline ROE. If U.S. returns rise and capital is recycled from low-return assets, earnings can grow faster than book. The proper forecast identity reconciles beginning equity, comprehensive income, dividends, repurchase cost, issuance, and ending shares.

The current 6.1% adjusted earnings yield provides another lens. If normalized EPS grows 6%–8% and the multiple is unchanged, return can be respectable with the dividend. If growth falls to low single digits and the P/E normalizes toward the mid-teens or lower, the dividend absorbs only a small part of multiple compression. If earnings reach the Q3 annualized run rate and U.S. ROE crosses 12%, the present multiple becomes easier to defend. This is a duration bet on the return bridge rather than an obvious current-yield anomaly.

Scenario asymmetry comes from the bank valuation formula’s denominator. When cost of equity and growth are close, small assumption changes produce large P/B moves. That mathematical sensitivity is a warning against false precision. The 1.21/1.81/2.33 scenario multiples are regimes, not confidence intervals. What can be observed directly is that the current quote exceeds the central through-cycle regime and sits near peer value for its present ROE.

The peer table also embeds different risks. RBC earns a premium for scale and higher returns; TD’s multiple includes both higher current return and U.S. regulatory repair; CIBC’s higher ROE comes with domestic/wholesale concentration; Scotiabank carries transformation and international complexity. BMO’s lower absolute multiple is not automatically a discount when its current return is lower. Relative valuation must control for ROE, growth, risk, and adjustment policy.

Finally, translating the conclusion to the NYSE line introduces FX, not a second valuation. At August 31, C$236.24 and US$170.31 imply about 1.387 Canadian dollars per U.S. dollar. A U.S. holder earns BMO’s Canadian-dollar economics plus CAD/USD movement. The two listings cannot rationally support different intrinsic values after conversion and market frictions; using one listing’s price with the other’s currency-denominated book recreates the prior error.

11. Variant Perception

Consensus-like bull view: BMO is ahead of schedule on its 15% ROE bridge. Credit has normalized, U.S. commercial loans have turned, NIM is better, portfolio sales release 50 basis points of CET1, and buybacks accelerate EPS. The strong Canadian/fee engines provide support while U.S. returns converge.

Variant response: the operations are improving, but the valuation already capitalizes approximately 14.4%–16.5% sustainable ROE. U.S. ROE is still 9.8%, deposits are shrinking, and a C$962 million capital loss emerged from the acquired perimeter. Premium buybacks can manufacture EPS accretion while reducing book value. The debate is duration and composition, not whether Q3 was good.

Consensus-like bear view: Bank of the West was a failed acquisition, Canada is overleveraged, and credit will relapse. BMO should trade near the weakest peer.

Variant response: that case understates the protected Canadian franchise, 22.9% Canadian P&C ROE, improving aggregate credit, strong liquidity, diversified fees, and real capital release. U.S. economics are unproven, not necessarily doomed. A short thesis also fights positive operating momentum and a powerful long-term price trend.

Most differentiated observation: BMO’s risk is no longer that investors fail to see the recovery. It is that investors confuse an improving ROE numerator with durable value creation after purchase-price losses, divested earnings, and premium share repurchases. A bank can reach a higher ROE by earning more, holding less equity, or selling low-return assets. Only the first is unambiguously high quality; the second and third require careful per-share accounting.

12. Fact vs. Interpretation

The central factual correction is non-negotiable: U.S.-dollar and Canadian-dollar per-share inputs cannot be mixed. The central interpretation is debatable: the fair cost of equity and sustainable growth rate. That is where reasonable conclusions should differ.

13. Open Questions

1. Can U.S. Banking sustain commercial-loan growth for four quarters while total deposits stabilize after the branch sale?
2. How much of the final U.S. ROE bridge comes from continuing-business revenue versus divested capital, lower credit costs, and denominator reduction?
3. What are the exact continuing earnings and funding costs lost with the three divestitures after closing adjustments?
4. Will Canadian P&C PCL remain near 58 basis points, or do cards, insolvencies, and mortgage renewals create another leg?
5. Does Capital Markets retain Q3 revenue when equities trading and issuance normalize?
6. Can Wealth improve efficiency while separating Burgundy/market effects from organic net flows?
7. At what valuation will management execute the new NCIB, and will BVPS/TBVPS grow after distributions?
8. How will the C$6.064 billion amortized-cost securities shortfall evolve as securities mature and rates change?
9. Can independent market-share data verify the stability of BMO’s Canadian commercial position?
10. What return hurdle and purchase price apply to Euroz Hartleys, which were not publicly disclosed in the announcement?

14. What Must Be True

The prior July bear-falsification test required U.S. ROE above 11%, sustained commercial-loan compounding, PCL at or below the mid-40s, and whole-bank ROE at least 13.5%. Two operational conditions now pass; the U.S. conditions do not. The prior bull-break test—U.S. ROE below roughly 10% with whole-bank ROE only 11%–12% at end-FY2027—has not triggered and is not yet due. The corrected valuation adds a stricter requirement: success must not merely occur; it must exceed what 2.09 times book already discounts.

15. Public Source Appendix

Company filings and investor materials

• BMO Q3-FY2026 Report to Shareholders / MD&A, August 25, 2026 — earnings, segment, credit, capital, corporate events, book value.
• BMO Q3-FY2026 Earnings Release, August 25, 2026 — headline results, dividend, Q3 repurchases, proposed NCIB.
• BMO Q3 interim financial statements, August 25, 2026 — securities, AOCI, allowances, goodwill/intangibles, equity.
• BMO Q3-FY2026 conference-call transcript, August 25, 2026 — U.S. ROE bridge, credit commentary, capital redeployment.
• BMO Q2-FY2026 conference-call transcript, May 27, 2026 — sequential U.S. loan and Canadian consumer-credit context.
• BMO FY2025 annual report, December 4, 2025 — five-year financial baseline, FY2025 capital return, business/risk disclosures.
• BMO 2025 Form 40-F, December 4, 2025 — issuer/filing record.
• BMO Investor Day presentation, June 16, 2026 — FY2027 targets, segment strategy, commercial ranks.
• BMO Investor Day transcript, June 16, 2026 — management’s return bridge and operating commitments.
• BMO 2026 management proxy circular, March 11, 2026 — governance, compensation, ownership, incentive adjustments.
• BMO 2026 voting results, April 15, 2026 — director, auditor, and say-on-pay votes.
• BMO 2026 U.S. stress test, 2026 — U.S. subsidiary severe-scenario capital/loss rates.

Corporate events

• Transportation and Vendor Finance sale, May 11, 2026.
• Euroz Hartleys acquisition announcement, June 29, 2026.
• Moneris sale announcement, August 10, 2026.
• Proposed 25 million-share NCIB, August 25, 2026.

Regulation, macro, market, and methodology

• OSFI Domestic Stability Buffer decision, June 19, 2026 — 3.0% DSB and current CET1 stack.
• Bank of Canada July 2026 Monetary Policy Report, July 15, 2026 — policy rate, growth/inflation/trade backdrop.
• AZI BMO.TO daily history, retrieved September 1, 2026 — TSX price/event map and matched-currency valuation.
• FactorsToday, retrieved September 1, 2026 — factor and risk context applied to the NYSE line with Canadian-issuer/currency caveat: methodology, loadings, leaderboard, stock information, specific volatility, related stocks, and factor history.
• Canadian Insider SEDI-derived report, June 3, 2026 — secondary, incomplete insider-activity context.

Peer primary reports

• Royal Bank of Canada Q3 2026 report, August 26, 2026; accessed September 1, 2026.
• CIBC Q3 2026 full report, August 27, 2026; accessed September 1, 2026.
• TD Q3 2026 results, August 27, 2026; accessed September 1, 2026.
• Scotiabank Q3 2026 report, August 25, 2026; accessed September 1, 2026.
• National Bank Q3 2026 report, August 26, 2026, and Q4/FY2025 results, December 3, 2025; accessed September 1, 2026.

Method note: all per-share valuation arithmetic uses TSX prices in Canadian dollars against BMO’s Canadian-dollar financial statements. NYSE prices are shown only after currency translation. Adjusted measures are management-defined non-GAAP measures and are paired with reported outcomes where capital-allocation conclusions differ. No enterprise-value or EBITDA framework is used for this deposit-funded bank.