National Bank of Canada (TSX: NA) — The Best Bank in Canada Isn’t a Canadian Bank

Independent equity research. Published 2026-07-18. Fresh coverage. All figures in Canadian dollars (C$) unless noted. National Bank reports in CAD under IFRS; fiscal year ends October 31. The bank is listed on the TSX only (there is no NYSE line and no ADR), is not an SEC registrant, and files with SEDAR+ rather than EDGAR — so there is no Form 4 insider feed and no 10-K. Primary sources are the Report to Shareholders, the quarterly Supplementary Financial Information package, and the Management Proxy Circular.

⚡ Claude’s Take

This block is the author’s own subjective opinion. It is general information, not investment advice, and not a recommendation to buy or sell any security. The analysis that follows takes no position and carries no price target — it examines National Bank only as embedded expectations and scenarios.

Verdict: AVOID here / HOLD if owned — a good company, a misunderstood one, and a genuinely bad price. This is the most expensive bank in Canada relative to what it actually earns on the capital it actually deploys. Don’t initiate at ~C$234. The zone where the math starts working is ~2.2–2.4× book, roughly C$175–C$195. Not a short — the franchise is sound, the balance sheet is a fortress, and shorting a compounding oligopoly member into a momentum tape is a good way to be right and still lose. Conviction: medium-high — unusually high for a valuation call, because the central evidence is a disclosure the company itself began publishing five months ago.

Here is the fact that reframes everything, and it comes straight from National Bank’s own Q2-FY2026 supplementary package rather than from any model of mine: the bank’s Canadian Personal & Commercial segment — the domestic retail and commercial bank, the thing everyone means when they call NA “the best-run bank in Canada” — earned an adjusted ROE of 11.8% in the first half of FY2026, and 10.9% in the second quarter alone. Its peers’ Canadian retail banks earn 20–30%. Meanwhile Wealth Management earned 58.2%, Financial Markets 27.4%, and — this is not a typo — ABA Bank in Cambodia earned 29.4%. National Bank earns nearly three times as much on its frontier-market bank in Phnom Penh as it does on its bank in Canada. The celebrated ROE premium that justified this stock’s decade of outperformance was never a superior retail franchise. It was a mix artifact: a capital-light wealth arm carrying almost no allocated equity, plus an outsized dealer, sitting on top of a subscale domestic bank that absorbs roughly 55% of the equity to produce roughly 27% of the earnings. Investors have been paying a quality premium for the right company for the wrong reason.

Now put the price against it. At C$234 the stock trades at 2.8–3.0× book and 3.3× tangible book — above its entire eleven-year range on P/E, P/B and P/TBV simultaneously — and on the standard justified-P/B bridge, (ROE−g)/(COE−g), it capitalizes a permanent ~18.3% ROE. National Bank has printed that twice in eleven years, both times in capital-markets boom years. If every announced CWB synergy lands in full and on time — including the C$200–250M of revenue synergies of which C$33M has actually been realized — the adjusted ROE ceiling is about 17.4%. The market is paying for a return above the fully-synergized ceiling, and above management’s own “17%+” FY2027 aspiration. The framing is not falling knife and not contrarian value: it is a crowded, low-volatility momentum trade inside a sector-wide melt-up — all six Canadian banks sit 24–27% above their 200-day EMAs, and NA’s +68.6% twelve-month return is the second-weakest of the six. You are not being paid for stock-picking here; you are paying a record multiple for cohort beta. And the capital-allocation direction of travel is wrong: management spent ~C$5.0B of stock on CWB and is buying Laurentian’s retail book next — deploying capital into the 11.8%-ROE segment. What would flip me bullish: two consecutive quarters of reported (not adjusted) ROE ≥16.5% with PCLs ≥35bps, plus P&C segment ROE breaking above 14% on a rising trend — that would prove the retail bank is being fixed rather than merely diluted. What would flip me bearish: Financial Markets dropping below 30% of segment earnings while PCLs pass 40bps and reported ROE prints below 14% — the mix unwinding and the cycle turning together. Tag: a capital-markets house with a subscale Quebec retail bank attached, priced like Royal Bank.

📈 Stock Price Action — Five-Year Event Map

National Bank did not round-trip; it compounded, then melted up. Over five years the stock ran from a June-2022 low of C$83.12 to an all-time high of C$235.32 on 2026-07-15, closing 2026-07-17 at C$234.06 — 0.5% off the record — against a 52-week range of roughly C$142.54–C$235.32, which is +64% off the 52-week low. The advance has been unusually smooth: annualized volatility of 17.6% and a worst five-year drawdown of only −22.6%. Over ten years NA has compounded at 22.6% annually including dividends, the best of the Big Six (RBC 18.6%, BNS 12.3%) — the outperformance is real and earned. But over the last twelve months its +68.6% is the second-weakest of the six, in a cohort that all returned +68% to +76%. (Price moves below are Fact; attributed drivers are Interpretation.)

Cycle narrative. (1) NA re-rated out of COVID with the group once OSFI permitted capital return. (2) The 2022 rate shock compressed every Canadian bank multiple regardless of quality — NA’s trough was June–October 2022. (3) The March-2023 global banking scare was pure contagion; NA has no US regional-bank exposure and the drawdown was shallow and brief. (4) The October-2023 low, on peak-rate and Canadian-housing fear, was the last genuinely cheap print. (5) The single most information-rich day in the five-year record is 2024-06-11: NA announced it would buy Canadian Western Bank for ~C$5.0B in stock at a 110% premium to CWB’s market price, and its own shares fell 8.5% in three sessions — the market immediately priced the deal as dilutive to returns, which the subsequent FY2025 ROE decline confirmed. (6) The stock nonetheless compounded through 2024 on rate cuts and a strong dealer. (7) A genuine air-pocket followed: a Q1-FY25 credit miss, the CWB close, and the April-2025 tariff shock. (8) From April 2025 the stock nearly doubled — and this leg is overwhelmingly a re-rating shared with the entire Canadian bank complex, not NA-specific alpha. (9) The most telling recent day is 2026-05-27: NA beat, raised the dividend, printed a 16.8% adjusted ROE — and fell 4%, as the market registered that the beat came from trading while the domestic bank weakened. That is the whole thesis compressed into one session.

1. Executive Summary

National Bank of Canada (founded 1859, headquartered in Montreal) is the smallest of Canada’s “Big Six” banks, with roughly C$618 billion of assets at 2026-04-30, more than 35,000 employees, and a market capitalization of approximately C$90 billion. It operates four segments: Personal & Commercial, Wealth Management, Financial Markets, and U.S. Specialty Finance & International (USSF&I, which houses Credigy and ABA Bank in Cambodia). Laurent Ferreira — who rose through the capital-markets side — has been CEO since 2021.

For a decade National Bank was the best-performing Canadian bank by almost any measure: the highest ROE of the Big Six (20.4% in FY2021, 19.1% in FY2022), the best-regarded efficiency, and a 22.6% ten-year annualized total return, the best of the group. That record is real and it is the reason the stock commands the second-highest multiple in Canadian banking today.

This report’s central finding is that the record was substantially misattributed. In February 2026, following its acquisition of Canadian Western Bank, National Bank began disclosing return on equity by business segment. Those disclosures show that the consolidated ROE has never been a story about a superior Canadian retail bank. For the first half of FY2026, on an adjusted basis: Wealth Management earned 58.2%, ABA Bank (Cambodia) 29.4%, Financial Markets 27.4%, Credigy 21.4% — and the Canadian Personal & Commercial bank earned 11.8%, falling to 10.9% in the second quarter. Peer Canadian retail banks run 20–30%. Roughly 55% of the group’s allocated equity sits in the domestic bank producing roughly 27% of the earnings, while Financial Markets absorbs roughly 27% of the equity to produce 37.5%. National Bank is best understood not as the best-run bank in Canada but as a capital-markets and specialty-finance house with a subscale Quebec retail bank attached. Management has effectively conceded the point: CEO Ferreira has called P&C returns “subpar versus our peers” and opened a strategic review.

The second finding is that capital is flowing the wrong way. The ~C$5.0 billion all-stock CWB acquisition (announced June 2024, closed February 2025) deployed shareholder capital directly into the 11.8%-ROE segment, diluting the diluted share count 11.6% and driving FY2025 reported EPS down 5.7% to C$10.07 even as net income rose 5.2%. Reported ROE fell to 13.7% (15.3% adjusted) from 17.2%. A further acquisition — Laurentian Bank’s retail and SME business — is scheduled to close in late 2026, again in P&C. The cost synergies are genuinely running ahead of plan (C$215M realized against a C$270M FY2026 target, since raised to ~C$300M); the revenue synergies are not (C$33M realized against a C$200–250M FY2028 target), and the acquired bank has barely grown since purchase.

The third finding is the price. At C$234.06 the stock trades at 2.8–3.0× book, 3.3× tangible book, and ~19× annualized first-half earnings — above its entire eleven-year range on all three measures simultaneously, and second only to Royal Bank across the Big Six. On the justified-P/B bridge, today’s price capitalizes a permanent ROE of roughly 18.3%. Full realization of every announced synergy implies a ceiling near 17.4% adjusted. The market is underwriting a return the company cannot reach even on management’s own plan — unless the Canadian bank cost of equity has permanently fallen to ~9.0%, which is the honest crux of the bull case and is examined directly in the valuation section.

The verdict the body supports: a sound, well-capitalized, genuinely well-managed institution whose principal earnings engines are its least durable ones, whose domestic franchise earns close to its cost of capital, whose capital allocation is moving toward the weakest segment, and whose shares carry a record multiple in a sector-wide momentum re-rating. The business is not in danger. The price assumes a version of the business that the company’s own segment disclosures say does not exist.

2. Business Overview

National Bank is a diversified financial institution with an unusually barbelled earnings structure. Four reportable segments, and the differences between them are the analytical heart of this report.

Segment detail — Q2 FY2026 (quarter ended 2026-04-30), from the Supplementary Financial Information (FACT):

How it makes money. Q2-FY2026 revenue of ~C$3.90 billion splits roughly 60% net interest income / 40% non-interest income, a more spread-dependent headline mix than it first appears — but the earnings mix is far more markets-weighted than the revenue mix, because Financial Markets and Wealth carry much lower capital and cost intensity. Personal & Commercial average assets were C$224.7 billion, of which mortgages are C$102.8 billion (average, Q2-FY26) — up from C$81.9B two years earlier, largely via CWB. Personal lending grew 11% and commercial lending 5% year-over-year in the quarter; P&C net interest margin was 2.26%, compressing from 2.36% in Q1-FY2025.

Recurring versus cyclical. The annuity-like layer is Personal & Commercial (sticky Quebec deposits, a conservatively underwritten mortgage book) and Wealth Management (fee income on AUM, the highest-quality earnings in the company and the smallest capital consumer). The cyclical layer is Financial Markets — trading, underwriting and advisory — which at 37.5% of segment earnings is the highest capital-markets weighting in the Big Six, and which in Q2-FY2026 declined year-over-year (C$488M versus C$501M) even as total earnings rose. USSF&I is a third category altogether: genuinely high-return, genuinely growing, and genuinely exposed to frontier-market and US consumer-credit risk that has little to do with Canadian banking.

Geography. Quebec remains the centre of gravity of the retail bank — roughly half the residential-mortgage book — with the CWB acquisition adding Alberta and British Columbia commercial banking, and ABA adding Cambodia. That is an unusual perimeter for a bank whose public narrative is “national Canadian champion.”

Verdict: a four-part company in which the two segments that define its public identity (the Canadian retail bank) and its public quality reputation (its ROE) are, respectively, its worst-returning business and a statistic produced by its other three businesses. This is not a normal Canadian bank; it is a specialty-finance and markets group with a domestic banking utility attached, and the distinction matters enormously for what multiple it deserves.

3. Industry Dynamics

The favourable fact first, because it is genuinely favourable. Canadian banking is one of the best banking markets in the developed world, and by deliberate design. Six banks — RY, TD, BNS, BMO, CM and NA — control roughly 90% of Canadian banking assets inside a federal architecture that actively protects them: the Bank Act and OSFI prudential supervision; “widely-held” ownership rules capping any single shareholder; a standing federal posture against large domestic bank mergers; and effective exclusion of meaningful foreign retail entry. The results are rational pricing, mid-teens sector ROEs well above US and European norms, sticky low-cost deposits, and a modern record without a single Big-Six failure — including through 2008.

In Greenwald’s taxonomy this is a textbook economies-of-scale-plus-customer-captivity industry, reinforced by a regulatory barrier to entry. The market-share-stability test — Greenwald’s single best moat diagnostic — passes emphatically at the industry level: Big Six shares have been broadly static for decades, with the only meaningful share shifts coming from acquisition rather than competition. In Marathon’s capital-cycle framework, the supply of new capital into Canadian banking is structurally constrained by regulation, which is precisely why sector returns stay high rather than mean-reverting to the cost of capital. Consolidation — NA absorbing CWB and Laurentian’s retail book — is supply reduction, which Marathon would read as constructive for the survivors’ pricing power.

But the industry verdict does not automatically transfer to every participant, and that is this report’s second structural point. Greenwald’s key insight about scale is that it is local, not absolute. A bank that dominates a region can out-earn a larger bank that is subscale everywhere. This is the standard explanation for National Bank’s historic ROE premium — Quebec density and linguistic/cultural incumbency producing genuine local scale — and it is the explanation this report tests and, in the competitive-position section, substantially rejects on the evidence of the segment disclosures. NA’s Quebec franchise faces a competitor most analysts underweight: Desjardins, the co-operative financial group that is the largest financial institution in Quebec and is not a Big-Six bank at all. NA’s Quebec position is strong but it is not monopolistic, and it is contested by an institution with a lower cost of capital and no shareholders to satisfy.

The mortgage market and the renewal wall. Canadian residential mortgages are full-recourse, reset every ~5 years (there is no 30-year fixed product), and the high-LTV tranche is CMHC- or privately-insured; OSFI’s B-20 stress test governs uninsured underwriting. This structure is a powerful shock absorber and it is why Canadian banks did not blow up in 2008. The 2025–27 concern was the repricing of pandemic-era sub-2% mortgages into 4–5% — the “renewal wall.” On the evidence, this risk has substantially passed: roughly 80% of the affected book has repriced, management describes payment shock as “significantly reduced,” and 90-day delinquencies sit below pre-pandemic levels. Bank of Canada rate cuts through 2024–25 did most of the work. The consensus fear of 2025 simply did not materialize, and its non-materialization is one of the four engines of the sector re-rating.

The macro is genuinely worse than the multiples suggest. Canada entered a technical recession with real GDP of −0.1% annualized in Q1-2026. Household debt reached 179.6% of disposable income, the highest in the G7. Loan growth over 2021–24 leaned heavily on record population growth, and the immigration-policy reversal removes a structural volume tailwind for the entire sector — this is underappreciated and it is not cyclical. The unresolved CUSMA renegotiation and US tariff friction threaten commercial and manufacturing borrowers concentrated in Ontario and Quebec; the April-2025 shock demonstrated how quickly that repricing happens. Against this, sector multiples sit at records — a divergence that prompted David Rosenberg to argue publicly that the Canadian banks have entered “bubble territory.”

Competitive intensity. Beyond Desjardins, the pressure points are EQ Bank and Tangerine in deposits (price competition at the margin, not share threats), the credit unions regionally, and — prospectively — open banking, which Canada has implemented slowly and which remains a medium-term rather than immediate risk. Interchange regulation continues to grind card economics lower. None of these threatens the oligopoly. All of them shave basis points.

Verdict: a structurally excellent industry — one of the best in global banking — facing a genuinely deteriorating domestic macro, at record valuations. The industry is good; the cycle is late; the price is high. Membership in this oligopoly is worth a premium multiple, and every Big-Six bank currently receives one. The analytical question for National Bank specifically is not whether the industry is good, but whether NA’s participation in the best part of it — dominant local retail scale — is as strong as its reputation and its multiple imply. the competitive-position section answers that.

4. Competitive Position

This is the section where National Bank’s reputation and National Bank’s disclosures part company.

4.1 The paradox, and its resolution

For a decade the puzzle was this: National Bank is the smallest of the Big Six — C$618B of assets against Royal Bank’s ~C$2.3 trillion, a market capitalization roughly one-fifth of RBC’s — in an industry whose advantages are overwhelmingly scale-driven. Technology spend, regulatory compliance, data and AI infrastructure, funding costs, brand: all scale. Yet NA earned the highest ROE of the six (20.4% in FY2021, 19.1% in FY2022) and delivered the best ten-year total return (22.6% annualized). Something had to explain it.

The standard explanation — repeated across sell-side research and the financial press — was Quebec local scale: NA dominates its home province, and Greenwald tells us local dominance beats national subscale. It is an elegant answer and it is largely wrong.

In February 2026, National Bank began publishing return on equity by segment. The data resolves the paradox definitively (FACT, Supplementary Financial Information, Q2-FY2026):

The Canadian bank earns 11.8%. The Cambodian bank earns 29.4%. National Bank of Canada generates nearly three times the return on equity in Phnom Penh that it generates in Canada. The consolidated ROE that built the company’s reputation is produced by Wealth Management — an asset-light fee annuity carrying so little allocated equity that it mathematically levers the group average upward — together with an oversized dealer and an emerging-market bank.

Back-solving allocated equity from the quarterly segment returns gives an indicative (ASSUMPTION — derived, not disclosed) split of roughly C$13.7B in P&C, C$6.8B in Financial Markets, C$2.8B in USSF&I and C$1.9B in Wealth. If approximately right, the Canadian retail bank absorbs roughly 55% of allocated equity to produce roughly 27% of the earnings, while Financial Markets absorbs roughly 27% of the equity to produce 37.5%. That single sentence is the company.

And this was true before CWB — which disposes of the “it’s just an acquisition artifact” defence. Reconstructing segment earnings shares back to FY2021, the year National Bank printed its celebrated 20.4% ROE (FACT):

Even in FY2021, at the peak of its reputation, Financial Markets was ~29% of earnings — already the highest capital-markets weighting in the Big Six. The premium was mix-assisted from the start; CWB did not create the distortion, it removed the camouflage. The FY2024→FY2025 movement is starker still: P&C earnings fell 17% (C$1,343M → C$1,111M) while Financial Markets earnings rose 34% (C$1,254M → C$1,684M). The domestic bank was already shrinking in profit terms before the acquisition closed.

The honest counter-argument, which the memo must carry. The CWB acquisition landed goodwill and intangibles in the P&C segment’s equity denominator, mechanically depressing its segment ROE. Backing out a plausible C$3.5–4B of CWB-related goodwill and intangibles lifts P&C’s return to roughly 14%. That adjustment is real and material. It does not rescue the thesis, for two reasons. First, 14% is still far below the 20–30% that peer Canadian retail banks earn. Second, and more fundamentally, it is an ex-goodwill figure — it flatters the segment by ignoring C$5 billion of real shareholder capital that was actually spent to acquire it. On the capital genuinely deployed, the domestic bank earns 10.9–11.8%, which is at or barely above a Canadian bank’s cost of equity. On that capital, the segment is creating little or no economic value.

The direction of travel is also wrong. P&C ROE moved from 12.7% (Q1-FY26) to 10.9% (Q2-FY26) on an adjusted basis, and the segment efficiency ratio deteriorated from 51.0% to 54.4% over the same two quarters. The strategic review was opened on the Q1 number; the Q2 number came in worse. Management has not disputed the framing — CEO Ferreira conceded on the call that P&C “is subpar versus our peers,” and when an analyst put the peer gap at 600–800 basis points and located it in the efficiency ratio, the bank’s own CRO answered, “You got it.” On the Q2 figures that 600–800bp framing is, if anything, charitable.

4.2 So what is the actual moat?

Naming the mechanism honestly, in Greenwald’s terms:

• Financial Markets (27.4% ROE, 37.5% of earnings) — a genuine but narrow advantage carrying an outsized load. NA is a legitimately strong Canadian mid-market dealer, ranked first or second in Quebec corporate and investment banking, protected by the same oligopoly. A 28.9% ROE at a 40.9% efficiency ratio sustained across several years is not a fluke, and the corporate/investment-banking half is the more annuity-like leg — it grew this quarter. But the advantage is a capability, not a captivity: it does not lock customers in, it must be re-won on price and execution every year, and the marginal outperformance came from global markets — the equity-derivatives and structured-products book, which is a long-volatility position rather than a franchise. Management effectively conceded this, naming equity structured products, equity finance and ETF/option market-making as the drivers, identifying April 4, 2025 as among the franchise’s most profitable days ever, and acknowledging that “trading will not always be that good.”

  It is already rolling over. In Q2-FY2026 Financial Markets revenue fell 2%, net income fell 3% year-over-year (C$488M versus C$501M), and the segment efficiency ratio deteriorated 430 basis points, from 36.6% to 40.9%. The gap between a modest, genuine mid-market advantage and 37.5% of group earnings is filled by cyclical trading profit, not by moat.

• Wealth Management (58.2% ROE) — the best business in the company and a real switching-cost moat. Advisor relationships, custody and administrative friction, and household inertia produce genuine customer captivity. It is also the smallest capital consumer, which is why its ROE is spectacular. If NA has a durable, high-quality franchise, this is it — and it is roughly a fifth of earnings.

• Personal & Commercial (11.8% ROE) — the weakest link, and the one everyone assumes is the moat. The Quebec franchise does have real customer captivity: linguistic and cultural incumbency, roughly 369 branches concentrated in the province, ~51–53% of residential secured lending there, and genuinely better Quebec credit performance. But a moat must show up in financial outcomes, and this one does not. If a moat claim cannot be tied to a financial outcome that would deteriorate without it, it is not a moat. A 10.9–11.8% ROE with a deteriorating efficiency ratio is not the signature of a dominant local franchise; it is the signature of a subscale one.

  The structural explanation most analysts miss is Desjardins. In its own home market National Bank is not the incumbent — it is the challenger to a member-owned cooperative that is the largest financial institution in Quebec, requires no market return on equity, and can therefore compete on price indefinitely. That is close to the worst possible local competitor, and it is a persuasive explanation for why a Quebec P&C efficiency ratio runs near 55.7% while peers’ domestic segments run 40–45% in provinces where their competition is other shareholder-owned banks. The same problem recurs in the newly acquired territory: Alberta’s ATB Financial is Crown-owned and carries the identical cost-of-capital advantage. National Bank has bought its way from competing against one non-profit-maximizing incumbent into competing against two.

• USSF&I (26.3%) — high returns, but not a moat. Credigy buys US consumer-loan portfolios: a skill and pricing-discipline business, not a franchise. ABA Bank’s 29.4% return reflects Cambodian market structure and risk pricing, not a competitive advantage NA can defend or extend.

4.3 Versus the peer group

Comp set computed at 2026-07-17 TSX spot prices on a consistent per-share basis, in Canadian dollars throughout (note that Canadian bank quotes are frequently mixed between the TSX C$ and NYSE US$ lines — all figures here are the TSX line). NA’s P/B is shown on the aggregator book basis (C$79.13); on a filing-consistent common-equity basis (C$82.90) it is 2.82×. The conclusion is unchanged across the range.

National Bank is the second-most-expensive bank in Canada on both price/book and price/tangible-book, behind only Royal Bank — which earns a comparable consolidated ROE with the widest franchise, the largest scale position, and a domestic retail bank earning roughly 30%. NA is priced above CIBC, TD and BMO while being the smallest institution in a scale industry, carrying the highest cyclical-earnings weighting in the group, and running the weakest domestic franchise returns.

Verdict: a genuine but narrower and lower-quality moat than the multiple implies. National Bank has one excellent business (Wealth), one strong-but-cyclical capability franchise (Financial Markets), two high-return non-Canadian specialty businesses (Credigy, ABA), and a domestic retail bank that earns close to its cost of capital and is getting worse rather than better. It does not have what its reputation and its price assume it has: a dominant, high-returning Quebec retail franchise. The “best-run bank in Canada” label is a misreading of a mix effect — and the market is paying for the label.

5. Growth History and Forward Opportunities

The ten-year record is genuinely strong, and it is important not to let the valuation argument obscure it. Revenue compounded from C$6.02 billion in FY2016 to C$14.02 billion in FY2025 — a 9.8% annual rate — while diluted EPS ran from C$3.29 to C$10.07, a 13.2% rate. Net income roughly tripled. Over that period the share count was broadly flat until 2024, meaning the growth was overwhelmingly organic and per-share. That is a good record by any standard and it is why the ten-year total return leads the Big Six.

Historical revenue and earnings (FACT, ROIC.ai reconciled to the filings; C$ millions except per-share):

The FY2025 line is the most important row in this report. Revenue rose 22.9% and net income rose 5.2% — and diluted EPS fell 5.7%, from C$10.68 to C$10.07, because the diluted share count rose 11.6% to fund the CWB acquisition. Reported ROE fell from 17.2% to 13.7%. This is the arithmetic signature of an acquisition that bought revenue and assets without buying returns. Some of the FY2025 depression is genuinely one-time — most notably a C$230M Day-1 IFRS 9 provision on the acquired performing CWB loan book, a non-cash accounting charge that does not recur — and on an adjusted basis FY2025 ROE was 15.3% rather than 13.7%. But adjusting for the charge does not change the direction: adjusted ROE still fell from 16.7% to 15.3%.

FY2026 is inflecting, and the inflection is real. Q1-FY2026 net income rose 26% year-over-year to C$1,254M and Q2 rose 38% to C$1,234M; adjusted ROE printed 16.6% and 16.8%. Critically, the Q1 beat was not purely acquisition arithmetic — excluding CWB, revenue grew 11% and pre-provision pre-tax profit grew 12%. Management raised FY2026 ROE guidance from ~15% to ~16% and articulated an ambition to exceed 17% in FY2027.

Where the forward growth is supposed to come from — and the quality of each source:

1. CWB cost and funding synergies — high confidence, largely banked. C$215M realized by Q2-FY2026 against a C$270M FY2026 target, with the target raised to ~C$300M and pulled forward a full year. This is genuine execution and management deserves credit for it.
2. CWB revenue synergies — low confidence, and the load-bearing assumption. C$33M realized against a C$50M FY2026 target and a C$200–250M FY2028 target. That requires a four-to-five-fold step-up. Revenue synergies are the most frequently missed category in bank M&A because they depend on cross-selling to customers who did not choose you. More concerning: the acquired bank has not grown since it was bought. CWB’s stand-alone revenue was flat at ~C$298M year-over-year; loans were flat through integration; term deposits fell ~C$1B and commercial loans ~C$400M in Q2-FY2026, with management acknowledging deposit attrition from a “noncore monoproduct customer segment.” Management’s claim that CWB’s book inflects in the second half of FY2026 is precisely the falsifiable proposition on which the deal thesis rests.
3. The Laurentian Bank retail/SME acquisition — closing late 2026, guided to 1.5–2% EPS accretion at a cost of 25–30bps of CET1. More consolidation of Quebec retail, into the 11.8%-ROE segment.
4. Buybacks and capital release. The NCIB was expanded to 14.5M shares (~3.7% of shares outstanding), with 6.4M repurchased at an average of C$165.75 — good execution against today’s C$234. CET1 is being deliberately walked down from 13.5% toward ~13% by FY2027, releasing roughly C$2.5B. AIRB approval on the CWB book should release a further 50–75bps.
5. ABA Bank (Cambodia) — still the fastest-growing unit (loans +12%, deposits +15% year-over-year in Q2-FY2026), though net-income growth is decelerating: 16% → 9% → 10% across recent quarters.

A critical observation on the composition of the FY2027 target. Of the roughly 100 basis points of ROE improvement management projects between FY2026 and FY2027, approximately 70 basis points come from buybacks (~40bp) and the Laurentian acquisition (~30bp) — that is, from financial engineering and M&A rather than from operating improvement in the underlying businesses. Buybacks at 2.9× book are also a questionable use of capital on their own terms: repurchasing stock well above book value is accretive to EPS but dilutive to book value per share, and it flatters ROE by shrinking the denominator rather than growing the numerator.

Verdict: high-quality historical growth, materially lower-quality forward growth. The ten-year record was organic, per-share, and driven by segments with genuine returns. The forward plan depends on cross-sell synergies that have barely begun in an acquired bank that is not currently growing, a second acquisition into the weakest segment, and buybacks above three times book. The growth is not disappearing — but its character has changed from compounding to assembling, and the market is still paying the compounding multiple.

6. Financial Quality

Capital and balance sheet — genuinely strong (FACT, Q2-FY2026). Total assets C$617.7 billion. Common equity C$31.97 billion after deducting C$1.61 billion of preferred and hybrid capital. CET1 ratio 13.5%, down from 13.8% at FY2025 year-end — a decline reflecting risk-weighted-asset growth and buybacks, not stress, and well above the OSFI requirement. Tangible common equity to total assets of 4.40% (the bank’s own disclosed ratio is 4.435%, confirming the calculation). Liquidity is ample. There is no balance-sheet fragility anywhere in this analysis, and the bear case in this report is emphatically not a solvency or credit-blowup case.

Book value (FACT, with a documented discrepancy). Common equity of C$31,970M over 385,633,934 shares gives BVPS of C$82.90. The ROIC.ai aggregator reports C$79.13, deducting a further ~C$1,450M presumed to be limited-recourse capital notes (AT1). Goodwill of C$3,098M plus other intangibles of C$1,688M gives tangible common equity of C$27,184M and TBVPS of C$70.49 (aggregator: C$70.32 — agreement to 0.2%, so the tangible figure is robust). The CWB deal added roughly C$2.2 billion of goodwill and intangibles, widening the book-to-tangible-book gap by about C$5.70 per share. This report carries P/B as a range of 2.82×–2.96×; every conclusion holds across it.

Profitability — the multi-year picture (FACT):

The ROE series is the story: a peak of 20.4% in FY2021, erosion through FY2023, a partial FY2024 recovery, then the CWB-driven drop to 13.7% reported in FY2025, and a recovery to 16.7% adjusted in H1-FY2026 that remains well short of the FY2021–22 peak. Return on assets has drifted down across the whole period, from 0.91% to 0.77% — a cleaner signal than ROE because it is unaffected by leverage or by the equity issued for CWB, and it says the asset base is becoming less productive, not merely more capitalized.

Efficiency — the reputation does not survive the data. National Bank has long been credited with the best efficiency ratio in Canada. At the consolidated level Q2-FY2026 was 50.4% adjusted, which is competitive. But the domestic segment tells a different story: P&C adjusted efficiency ran 53.2% in FY2024 → 55.5% in FY2025 → 51.0% in Q1-FY2026 → 54.4% in Q2-FY2026. The CWB acquisition brought a less efficient cost base and the ratio deteriorated roughly 230 basis points in FY2025 without fully recovering. Peer Canadian retail segments run materially better. Consolidated expense growth was +9.5% year-over-year in Q2-FY2026 (+7.4% excluding litigation and property-tax items) — running ahead of most revenue lines.

Net interest margin in P&C compressed from 2.36% (Q1-FY2025) to 2.26% (Q2-FY2026), consistent with competitive deposit pricing and the acquired CWB book’s funding profile.

Credit quality — currently a genuine strength, and the honest counterweight to the bear case. Impaired PCLs are running 26–33 basis points, within management’s unbroken 25–35bp guidance. The bank has built performing allowances for fifteen consecutive quarters and carries an allowance covering 5.9× charge-offs. Q2-FY2026 total PCL was C$233M, below expectations. The year-over-year optics are distorted and must be handled carefully: Q2-FY2025’s C$545M included the C$230M Day-1 CWB provision, so the apparent collapse in provisions is substantially an accounting artifact rather than a credit improvement. Residential secured lending is ~51–53% Quebec — a province whose credit performance has outrun the national average — with ~27% insured. The mortgage renewal wall is largely passed. The genuinely new exposure is Western Canadian commercial real estate inherited from CWB: a single insured file produced more than half of Q2-FY2026’s commercial impaired formations, which is a concentration worth watching rather than a loss.

One credit item deserves separate flagging. ABA Bank in Cambodia carries roughly C$1 billion of gross impaired loans with low write-offs, low resolutions, and loan-to-values migrating from the high-40s into the 50s. On the Q2 call an analyst identified C$47M of formations matching an identical C$47M increase in gross impaired loans, and management conceded “resolutions are remaining low.” INTERPRETATION: reported impaired balances in a frontier market with slow resolution may lag realized losses. This is not currently a material earnings risk — ABA earned a 29.4% ROE — but it is an opaque credit book in a jurisdiction without Canadian-standard supervision, and it warrants an open question rather than comfort.

Quality of earnings — the wedges to know. Reported and adjusted diverge by C$0.68 across H1-FY2026 (reported diluted C$6.14 versus adjusted C$6.48), the difference being CWB integration charges plus amortization of acquisition-related intangibles. This report carries both series throughout. The amortization of intangibles purchased with C$5 billion of real shareholder currency is an economic cost, not an accounting artifact — so reported is treated as the conservative anchor and adjusted as management’s framing, on the principle that management commentary is a hypothesis rather than evidence. The other distortions to normalize: the C$230M Day-1 CWB provision (FY2025, depresses), a Q1-FY2026 result that some coverage characterized as tax-assisted, and the CWB purchase-accounting credit mark, which accretes into net interest income over time and modestly flatters margin.

Verdict: economics do not clearly improve with scale here — which is the most damaging finding available about a bank that just bought scale. Return on assets fell as the balance sheet grew. The domestic efficiency ratio worsened after the acquisition. Net interest margin compressed. ROE fell and has not returned to its pre-deal level. The balance sheet is strong and credit is clean, so this is a returns problem rather than a risk problem — but a bank whose central strategic act was to acquire scale, and whose productivity ratios deteriorated as a result, has not yet demonstrated that the scale was worth buying.

7. Capital Allocation

Capital allocation is where a business’s value becomes a shareholder’s value, and it is the section where National Bank’s recent record is weakest — not because management is careless, but because the direction of deployment contradicts the returns data the company itself now publishes.

7.1 The Canadian Western Bank acquisition — the defining act

The facts (FACT). Announced after the close on 2024-06-11. All-stock: 0.450 National Bank shares per CWB share, valuing CWB at C$52.35 per share against a pre-announcement close of C$24.89 — a 110% premium — for total equity consideration of roughly C$5.3 billion (C$5.6 billion including shares already held). Funded partly by a concurrent ~C$1.04 billion subscription-receipt raise at C$112.30, below the pre-announcement price, with a private placement to a CDPQ affiliate. Final Minister of Finance approval 2024-12-20; closed 2025-02-03. National Bank’s shares fell 8.5% over the three sessions following the announcement.

The price, stated honestly in both directions — because one framing alone misleads. The “110% premium” headline is a premium to a depressed market price: CWB traded at roughly 0.57× book because it earned only ~10% on equity. On book value, National Bank paid approximately 1.19× book. And it paid in its own stock, which was then trading at roughly 1.9× book. Issuing currency at 1.9× book to acquire assets at 1.2× book is accretive to book value per share. The deal was therefore not a straightforward overpayment, and analysts who stopped at the 110% headline got it wrong.

But it was unambiguously dilutive to returns, and that is what matters. National Bank bought a ~10%-ROE bank into a ~19%-ROE company. The dilution was not a mistake in the price; it was inherent in the asset. FY2025 recorded exactly that: revenue +22.9%, net income +5.2%, EPS −5.7%, reported ROE from 17.2% to 13.7%. The market understood this immediately, which is why the stock fell 8.5% on announcement.

Execution since has been genuinely good on costs and genuinely poor on revenue. Cost and funding synergies reached C$215M by Q2-FY2026 against a C$270M FY2026 target, with the target raised to ~C$300M and pulled forward a year — real, creditable execution. Revenue synergies stand at C$33M against a C$200–250M FY2028 target. And the acquired business is not growing: flat loans, flat stand-alone revenue, C$1B of term-deposit outflow and C$400M of commercial-loan decline in Q2-FY2026. Cost synergies shrink an acquisition; revenue synergies justify one. So far National Bank has delivered the former and not the latter.

Quantifying whether the returns gap can even be closed (the decisive calculation). Taking every announced synergy at face value — C$300M of cost and funding plus C$250M of revenue synergies, C$525M pre-tax in total, taxed at 24%, on ~C$32.0B of common equity — the full run-rate contribution is roughly C$399M after tax, or about +1.25 percentage points of ROE. Added to the current 16.7% adjusted return, that gives a fully-synergized adjusted ROE ceiling near 17.4% (a reported ceiling nearer 15.5%). Hold that number; the valuation section shows the market is paying for more than it.

7.2 The rest of the capital-allocation record

Further M&A — and a distinction that matters. The Laurentian Bank retail and SME acquisition is scheduled to close in late 2026 (guided 1.5–2% EPS accretion, −25 to −30bps CET1). Management has additionally signalled appetite for further “tuck-in acquisitions in P&C and wealth” while still integrating CWB.

It would be lazy to lump CWB and Laurentian together as “more P&C M&A.” In Greenwald’s framework they are close to opposites. Laurentian is the correct deal: it buys retail and SME share inside National Bank’s region of dominance, spreads an existing Quebec cost base over more Quebec customers, and removes a competitor — Laurentian is exiting retail banking. It is small, cheap, low-integration-risk, and it is plausibly the single most effective available lever on the P&C efficiency ratio. CWB is the adverse deal: roughly C$5.0 billion of stock for a branch network in Alberta and British Columbia where RBC, BMO, CIBC and ATB Financial already hold local scale. National Bank arrives subscale in someone else’s home market — its own Quebec advantage, turned against it.

INTERPRETATION: the acquisition appetite is not uniformly misdirected; it is misdirected in its largest expression. The Wealth ambition is well-aimed (that segment earns 58%), the Laurentian deal is well-aimed, and the C$5 billion CWB deal — by far the biggest capital commitment of the three — is the one that runs against both the segment-return data and the local-scale logic.

A precise and damning decomposition of the FY2027 ROE bridge. Management’s own walk to the “17%+” target attributes approximately +110bps organic, +20bps CWB revenue synergies, +30bps Laurentian, +40bps from levering CET1 down to 13% via buybacks, and −100bps from RWA growth. Two observations follow. First, the C$5.0 billion CWB acquisition contributes just 20 basis points of ROE improvement, while buybacks contribute 40 — financial engineering delivers twice what the largest acquisition in the bank’s history does. Second, the bridge assumes no improvement whatsoever from P&C and none from credit. That is either unusually conservative guidance, or management declining to underwrite its own strategic review producing anything.

Buybacks. The posture reversed sharply and instructively. Management initially declined an NCIB even at a 13.9% CET1 ratio, with Ferreira arguing that “buybacks are a complement, not a growth strategy” — and an analyst pointedly asking why the bank would run “an embarrassingly low 2% NCIB.” The programme was subsequently launched (September 2025, 8.0M shares) and then upsized to 14.5M shares (~3.7% of shares outstanding) in February 2026, with 6.4M repurchased by Q2 at an average of C$165.75. Against today’s C$234 those repurchases have worked well. Looking forward, however, buying stock at 2.9× book is EPS-accretive but book-value-dilutive, and it improves ROE by shrinking equity rather than by earning more on it.

Dividends. A consistent, well-covered record: raised 6 cents to C$1.24 (December 2025) and 8 cents to C$1.32 (May 2026, +6%), for C$5.28 annualized — a 2.26% yield on roughly a 41% payout of adjusted earnings. This is the lowest yield in the Big Six, which is itself a valuation signal.

Share count. Broadly flat from FY2016 to FY2024 (339.9M to 342.8M diluted) — a genuinely good record of not diluting holders — then +11.6% in a single year to 382.5M for CWB. Nine years of discipline spent in one transaction.

Incentive alignment (OPEN QUESTION). The Management Proxy Circular could not be retrieved during this engagement, so the precise weighting of ROE within executive short- and long-term incentive metrics is unverified. This matters more than usual: if management is compensated primarily on EPS growth, total shareholder return, or absolute earnings rather than on return on equity, then an acquisition strategy that grows earnings while diluting returns is a rational response to the incentive scheme. This is the single most important unresolved question in the file and it is carried into the fact-vs-interpretation table and the diligence appendix.

Insider activity (OPEN QUESTION). National Bank is not an SEC registrant, so there is no Form 4 feed. Canadian insider filings sit in SEDI, which is not reliably retrievable programmatically. The insider read is therefore open, not clean — absence of evidence, not evidence of absence.

Verdict: a good operator making a defensible-but-returns-dilutive strategic bet, with the execution risk concentrated in the half of the thesis that has not yet delivered. The CWB price was reasonable on book; the cost synergies are ahead of plan; the dividend record is strong; the buyback has been well-timed. But the strategy directs shareholder capital into the segment earning 11.8% while the segments earning 27% and 58% are not the acquisition targets, and the FY2027 ROE bridge leans roughly 70% on buybacks and further M&A rather than operating improvement. On the evidence to date this is capital allocation that grows the company more reliably than it grows per-share value.

8. Changes and Headwinds — Last Two Years

Strategic and corporate (FACT).

• June 2024: CWB acquisition announced; shares −8.5% in three sessions; ~C$1.04B subscription-receipt raise at C$112.30, followed by a further equity raise in July 2024 once the shares recovered.
• December 2024: final Minister of Finance approval, following OSFI and Competition Bureau review and a public consultation.
• February 2025: CWB acquisition closed. C$230M Day-1 IFRS 9 provision on acquired performing loans. Share count +11.6%.
• FY2025: African assets divested. ABA Bank (Cambodia) retained, with no strategic-review commentary on any of the five most recent calls.
• Late February 2026: systems conversion for CWB completed; segment-level ROE disclosure began — the disclosure that underpins this report.
• 2026: Laurentian Bank retail/SME acquisition announced, closing late 2026. AIRB approval on the CWB book still pending, expected mostly in Q4-FY2026 (50–75bps of capital release).
• September 2025 / February 2026: NCIB launched (8.0M shares), then upsized to 14.5M.

Operating and financial.

• Reported ROE fell from 17.2% (FY2024) to 13.7% (FY2025), recovering to 15.8% in H1-FY2026 (16.7% adjusted).
• FY2026 ROE guidance raised from ~15% to ~16%; FY2027 ambition articulated at “17%+.”
• CWB cost/funding synergies ahead of plan (C$215M at Q2-FY2026, target raised to ~C$300M); revenue synergies behind (C$33M).
• Dividend raised twice, to C$1.32 quarterly.
• Canadian P&C segment ROE deteriorated from 12.7% to 10.9% quarter-over-quarter (adjusted), efficiency ratio from 51.0% to 54.4%; CEO opened a strategic review and conceded returns are “subpar versus our peers.”

Macro and regulatory headwinds.

• Canada entered a technical recession (real GDP −0.1% annualized, Q1-2026); household debt reached 179.6% of disposable income.
• Immigration-policy reversal removes the population-growth tailwind that supported 2021–24 loan volumes — a structural, not cyclical, headwind for the whole sector.
• US tariffs and the CUSMA renegotiation remain unresolved; the April-2025 “Liberation Day” shock demonstrated the speed of repricing (NA −4.7% on 2025-04-04, +6.5% on the 2025-04-09 pause).
• FRTB implementation makes market-risk RWA more volatile — material for a bank with NA’s trading weighting; management has raised the issue with OSFI.
• Canadian bank sector valuations at records, prompting public warnings of “bubble territory.”

Tailwinds, stated fairly.

• The mortgage-renewal wall largely passed without the predicted default wave — ~80% of the book repriced, delinquencies below pre-pandemic levels.
• Sector-wide capital-markets boom: Big Six capital-markets profits +27% in Q2-2026, and National Bank is the most levered to it.
• PCL releases flowing to earnings across the group, with capital ratios strong enough to permit further buffer release.

Verdict: the last two years strengthened the balance sheet and the footprint while weakening the returns profile — and the market has priced only the first half. National Bank is larger, more geographically diversified, and better capitalized than two years ago; it is also lower-returning, less efficient in its domestic bank, more acquisitive, more dependent on trading, and operating into a materially worse Canadian macro. The strategic logic of reducing Quebec concentration is sound. The execution has so far delivered scale without returns.

9. Risk Analysis

Catastrophic-loss assessment. The risk of permanent capital impairment is low. National Bank is a systemically important, federally supervised member of a protected oligopoly, with 13.5% CET1, 5.9× allowance coverage, a full-recourse and partly insured mortgage book, and no US regional-bank or commercial-real-estate concentration of the kind that has damaged peers. The realistic downside is multiple compression against merely adequate earnings, not solvency. The two tail risks that could produce something worse are a severe Canadian housing and employment shock striking a Quebec-concentrated book simultaneously with a trading-revenue collapse, and — smaller but genuinely non-zero — an AML or sovereign event at ABA Bank producing a headline and capital charge out of proportion to the segment’s size.

10. Valuation Discussion — Embedded Expectations

No price target and no recommendation appears in this section. The purpose is to establish what the current price requires to be true.

10.1 Where the stock trades

At C$234.06 (2026-07-17 TSX close), against a market capitalization of approximately C$90.3 billion (FACT — note that aggregator data reports a market capitalization near C$79 billion, which is a stale 2025-10-31 fiscal-period-end snapshot; every aggregator-published multiple for the latest quarter is understated by roughly 14% and has been recomputed at spot throughout this report):

EV-based multiples are meaningless for a bank and are not reported. Aggregator services publish an EV/EBITDA near 200× and “net debt” of C$72.8 billion for National Bank; both are artifacts of treating customer deposits and wholesale funding as debt. They have been discarded rather than adjusted.

10.2 Own-history context — the single highest-signal datum

The own-history percentile series normally used for this comparison returns empty for TSX-only tickers, so the range was rebuilt manually from eleven years of fiscal-year-end multiples (FACT):

National Bank trades above its entire eleven-year fiscal-year-end range on price/earnings, price/book and price/tangible-book simultaneously — effectively a 100th-percentile own-history reading. It is not at the top of its historical range; it is outside it, by 24–38% depending on the measure. For context, the COVID trough saw an intra-year price/book near 0.96×.

10.3 The justified-P/B bridge — the core result

For a bank, the cleanest embedded-expectations framework is the standard justified multiple:

Justified P/B = (ROE − g) / (COE − g)

Using a cost of equity of 9.5% and sustainable growth of 5% — and a book value per share of C$79.13:

Now invert it — what does C$234.06 actually underwrite?

On these assumptions, today’s price capitalizes a permanent ROE of roughly 18.3%. National Bank has printed 18.3% or better in two of the last eleven years — FY2021 (20.4%) and FY2022 (19.1%) — both capital-markets boom years, and both before the CWB acquisition structurally lowered the return profile. Actual reported ROE was 13.7% in FY2025 and 15.8% in H1-FY2026; adjusted was 15.3% and 16.7%; management guides to ~16% for FY2026 and aspires to “17%+” in FY2027.

The decisive comparison. From the capital-allocation section, full realization of every announced CWB synergy — C$300M of cost and funding plus C$250M of revenue synergies, C$525M pre-tax, C$399M after tax on ~C$32.0B of equity — adds approximately 1.25 percentage points of ROE, implying a fully-synergized adjusted ceiling near 17.4% and a reported ceiling near 15.5%.

The market’s implied 18.3% sits above the fully-synergized ceiling, and above management’s own FY2027 aspiration. What is priced is not the plan working. It is the plan working, plus roughly a further point of return from an unidentified source — or a permanent re-rating of the discount rate.

The inversion the bear case rests on. If reported ROE simply stays where FY2025 left it, justified value is 1.94×–2.16× book, implying C$154–C$171 — the stock sits 37–52% above that. If the adjusted 16.7% is the true sustainable figure, justified value is 2.60× book, or C$206 — still roughly 14% below spot.

10.4 The honest bull rebuttal: this is a cost-of-equity argument

Intellectual honesty requires stating the strongest form of the counter-case, and it is not about ROE at all. Solving for the discount rate that makes today’s 2.96× fair:

The entire bull case reduces to a single question: has the Canadian bank cost of equity permanently fallen to roughly 9.0%? At a 9.0% cost of equity, 5% growth and the current 16.7% adjusted return, today’s price is approximately fair, and the FY2028 synergy run-rate is upside. At a 9.5–10.0% cost of equity, it is 14–35% expensive.

There is a real argument for the lower number: National Bank’s realized beta is 0.49, its CET1 ratio is 13.5%, and it operates inside a federally protected oligopoly with no modern failure record. Those are genuine grounds for a lower discount rate than a generic bank. Against that: 37.5% of segment earnings come from capital markets, the least durable and most volatile business in banking, and a low realized beta measured during a decade-long bull market in Canadian bank equities is weak evidence about the discount rate that will apply through a full cycle. A 50-basis-point difference in the assumption moves justified value 15–20%. This is the crux of the disagreement, and it is presented as the crux rather than resolved by assertion.

10.5 Scenario analysis

Three-year scenarios; book value per share compounds at ROE × (1 − 42% payout). Assumptions are explicit and are the point of the exercise.

The asymmetry is the conclusion. The bull case requires a permanent structural re-rating of both the return on equity and the discount rate, simultaneously, and pays +12.8% annually. The bear case requires only that credit normalizes and capital markets mean-revert — the two most reliably cyclical phenomena in banking — and costs −3.1% annually. Note also that the bear case’s 2.00× exit multiple is still above every fiscal-year-end price/book National Bank printed between 2015 and 2024; a genuine reversion to the 1.81× eleven-year mean would imply roughly C$177 on unchanged book value.

Verdict: the price embeds a return the company has reached twice in eleven years, cannot reach on its own fully-synergized plan, and is currently deploying capital away from. The valuation is not absurd — a 9.0% cost of equity makes it defensible — but it requires the most optimistic available assumption on every one of the three variables that matter simultaneously.

11. Variant Perception

The consensus view. “National Bank bought its way out of the Quebec-concentration problem. CWB delivers a genuine Western Canadian franchise, the synergies are running ahead of plan with the target raised from C$270M to C$300M, management raised FY2026 ROE guidance to 16% and is targeting 17%+ for FY2027, and the ROE dip is a transitory acquisition artifact that reverses as integration completes. Best-in-class operator, now with a national footprint.” The +68.6% twelve-month return and the price sitting 0.5% off an all-time high say this view is fully held and fully paid for.

The strongest bull case, stated at its best. The ROE decline is genuinely mechanical rather than operational — it is the arithmetic of adding C$5.0 billion of equity and C$2.2 billion of intangibles mid-year — and it reverses as synergies annualize. Adjusted ROE is already back to 16.8%. Cost synergies are ahead of schedule and the target has been raised, which is evidence of an integration going well, not badly. Credit is clean, the mortgage renewal wall has passed without incident, CET1 is 13.5% with C$2.5 billion of release ahead plus 50–75bps from AIRB approval, and the bank now has a Western Canadian growth runway it never had. If the Canadian bank cost of equity really has fallen to ~9% — defensible at a 0.49 beta inside a federally protected oligopoly — then today’s multiple is approximately fair on the current adjusted return, and the FY2028 revenue-synergy ramp is free upside.

The strongest bear case, stated at its best. The market is paying an all-time-record multiple — above the entire eleven-year range on P/E, price/book and price/tangible-book — for a bank whose reported ROE just fell 350 basis points, which is the smallest of six in an industry where scale is the only durable cost advantage, whose domestic franchise earns 11.8% and is deteriorating, and which draws 37.5% of segment earnings from the least durable business in banking at a point when that segment is already declining year-over-year. The justified-P/B bridge says a permanent 18.3% return is priced; full realization of every announced synergy reaches roughly 17.4% adjusted. The gap is not a forecast — it is a subsidy from a re-rating that is sector-wide, momentum-driven, and reverses when either credit or capital markets normalize. Management is meanwhile deploying capital into the 11.8% segment and sourcing 70% of its FY2027 ROE improvement from buybacks and M&A rather than operations.

Where consensus is most likely offsides. Not on business quality — National Bank is a well-run institution and the bulls are right about that. Consensus is offsides on two specific things: (1) the composition of the ROE, which the company’s own new segment disclosures reveal to be a mix artifact rather than domestic-franchise excellence, and which most published analysis has not yet incorporated; and (2) the durability of the capital-markets earnings stream that produces the largest single share of profit. The market is applying a retail-bank multiple to a trading-weighted earnings mix.

The 3–5 assumptions that actually matter:

1. Is sustainable ROE 14.7% (reported), 16.7% (adjusted), or 18%+ (what the price implies)?
2. Is the Canadian bank cost of equity 9.0% or 9.5–10.0%? A 50bp difference moves justified value 15–20%.
3. Do the C$200–250M of CWB revenue synergies land? Only C$33M realized; the cost synergies are already banked.
4. Can the Personal & Commercial segment’s 11.8% ROE be structurally improved, or is a subscale domestic bank simply what National Bank is?
5. Does Financial Markets — 37.5% of segment earnings and already down year-over-year — hold, or mean-revert?

Falsification test, bull side. Two consecutive quarters of reported (not adjusted) ROE at or above 16.5% with PCLs at or above 35 basis points — that is, ROE recovering on operations rather than on a benign credit quarter or a favourable earnings definition — combined with P&C segment ROE breaking above 14% on a rising trend and cumulative CWB revenue synergies crossing C$120M by the end of FY2027. That combination would validate a sustainable return near 17.5% and make the current multiple defensible.

Falsification test, bear side. Financial Markets falling below ~30% of segment earnings for two consecutive quarters while total PCLs exceed 40 basis points and reported ROE prints below 14% — confirming that the FY2025 decline was operational and cyclical rather than an integration artifact. Alternatively, cumulative revenue synergies still below C$75M at the end of FY2027 would falsify the acquisition thesis outright.

Positioning read. The stock sits 0.5% off an all-time high, +68.6% over twelve months, 26% above its 200-day exponential moving average, with a realized beta of 0.49 and a five-year maximum drawdown of only 22.6%. Empirical factor-model coverage is unavailable for TSX-only tickers, so this read is built from the price series alone and is stated with that limitation. Critically, all six Canadian banks sit 24–27% above their own 200-day EMAs — this is a beta trade on the Canadian bank complex, not a National Bank-specific re-rating. The framing is a crowded, low-volatility momentum trade in a sector-wide melt-up: not a falling knife, not a contrarian value setup. The risk is not that the business breaks. It is that a bank compounding book value at roughly 9–10% annually and yielding 2.26% has nothing left to give from the multiple.

12. Fact vs. Interpretation

13. Open Questions

1. What are the actual executive incentive metrics, and how heavily is ROE weighted? The most important unresolved question in this file. If compensation rewards EPS growth, absolute earnings or total shareholder return rather than return on equity, then an acquisition programme that grows the company while diluting returns is a rational response to the scheme rather than a strategic error — and would be expected to continue.
2. Can Personal & Commercial’s ROE be structurally improved, or is 11.8% simply what a subscale domestic bank earns? Management has opened a strategic review. The specific question is whether the gap is a cost problem (fixable), a mix problem within the segment (partly fixable), or a scale problem (not fixable without more acquisitions, which is the problem).
3. Why does National Bank still own ABA Bank in Cambodia? It earns a 29.4% ROE, so the financial answer is obvious — but it is a frontier-market bank inside an institution whose narrative is “national Canadian champion,” it carries roughly C$1 billion of slowly resolving impaired loans, and no strategic-review commentary appeared on any of the five most recent calls. What is the long-term intention?
4. What is the true book value per share? Filing-consistent common equity gives C$82.90; the aggregator basis gives C$79.13, a ~C$1.45 billion difference presumed to be limited-recourse capital notes. This moves price/book between 2.82× and 2.96×.
5. Will the CWB loan book actually inflect in H2-FY2026 as management claims? This is the load-bearing, near-term, falsifiable proposition of the entire acquisition thesis.
6. How much of the recent PCL improvement is genuine credit performance versus the roll-off of the C$230M Day-1 CWB provision comparative?
7. What are peer Canadian P&C segment ROEs on a directly comparable basis? This report cites 20–30% for peers from secondary sources; sourcing them directly from RY, TD and CM supplementary packages would convert the central comparison from approximation to hard fact.
8. How much of Financial Markets’ returns are structural franchise versus long-volatility positioning? Management named equity structured products, equity finance and ETF/option market-making, and identified April 4, 2025 as among the franchise’s most profitable days — which suggests more of the latter than a franchise narrative implies.

14. What Must Be True

For the bull case

1. Sustainable ROE must be ≥17.5%, not the 16.7% currently delivered. Falsification: reported ROE fails to exceed 16.5% in any two consecutive quarters through FY2027 while PCLs sit at or above 35bps.
2. The Canadian bank cost of equity must genuinely be ~9.0%, and must stay there. Falsification: a macro or credit shock in which Canadian bank multiples compress more than 20% — demonstrating that the low realized beta was a bull-market artifact rather than a structural property.
3. CWB revenue synergies must materially land. Falsification: cumulative revenue synergies below C$120M at end-FY2027 against the C$200–250M FY2028 target.
4. The P&C segment must improve from 11.8% toward the high teens. Falsification: P&C segment ROE remains below 14% through FY2027, or continues the 12.7% → 10.9% deterioration.
5. Financial Markets must hold near current contribution levels rather than mean-reverting. Falsification: Financial Markets net income falls below 30% of segment earnings for two consecutive quarters.

For the bear case

1. The multiple must compress from its record. Falsification: the stock sustains 2.9×+ book for eight or more consecutive quarters while ROE stays below 17% — demonstrating a genuine permanent re-rating rather than a momentum overshoot.
2. Capital markets must mean-revert. Falsification: Financial Markets sustains ≥27% ROE and ≥35% of segment earnings through a full volatility cycle including a low-volatility year.
3. The domestic franchise must remain structurally subscale. Falsification: P&C ROE reaches the high teens on operating improvement — expense discipline and revenue synergies — rather than on goodwill amortization running off.
4. Credit must normalize toward through-cycle levels. Falsification: impaired PCLs remain below 35bps through FY2028 despite a technical recession and 179.6% household debt-to-income.
5. Capital allocation must continue flowing to the weakest segment. Falsification: management redirects capital toward Wealth (58% ROE) and away from P&C acquisitions, or returns the CET1 release to shareholders instead of deploying it into further P&C M&A.

15. Source Appendix

See NA.TO_source_appendix.md for the complete annotated source list. Principal sources:

Primary — company filings and disclosures

• National Bank of Canada, Supplementary Financial Information, Q2 Fiscal 2026 (quarter ended 2026-04-30) — consolidated and segment return on common shareholders’ equity, segment efficiency ratios, net interest margin, average loans and assets, PCL detail. The single most important source in this engagement.
• National Bank of Canada, Report to Shareholders, Q2 Fiscal 2026 and Q1 Fiscal 2026; Q4 Fiscal 2025 annual results.
• Supplementary Financial Information packages, Q4 FY2021 through Q4 FY2025 (multi-year trend series).
• Press release, “National Bank to accelerate domestic growth with the acquisition of Canadian Western Bank,” 2024-06-11, and accompanying investor presentation.
• Press releases: final Minister of Finance approval (2024-12-20); acquisition completion (2025-02-03); NCIB launch (2025-08-27) and amendment (2026-03-10); Q2-FY2026 results and dividend increase (2026-05-27).

Primary — management commentary (treated as hypothesis, not evidence)

• Earnings-call transcripts, FY2025 Q2/Q3 and FY2026 Q1/Q2 — CWB synergy progress, segment ROE commentary, CEO Laurent Ferreira on P&C returns being “subpar versus our peers,” Financial Markets regime-dependence, ABA Bank credit exchange, capital and buyback posture.

Market and quantitative data

• Daily price and total-return series for NA.TO and all five Big-Six peers (RY, TD, BMO, BNS, CM), accessed 2026-07-18 — spot prices, 21/50/200-day EMAs, beta and alpha, five-year and 52-week ranges, dividend history, cohort relative-strength comparison.
• ROIC.ai aggregated financial data — multi-year income statement, balance sheet, profitability ratios, valuation multiples. Third-party aggregated data, not primary; reconciled to the filings throughout, with two documented discrepancies (FY2025 ROE and book value per share) resolved in favour of the filing.
• Empirical factor-model coverage is unavailable for TSX-only tickers; the positioning read is built from the price series alone and is stated with that limitation.

Secondary — trade and financial press

• Reuters, Bloomberg, BNN Bloomberg, The Globe and Mail, Canadian Mortgage Trends, The Deep Dive — for event attribution, sector context, analyst commentary and the Canadian macro backdrop.

This article takes no position and contains no price target outside the clearly-labelled opinion block at the top, which is the author’s own subjective view. General information only — not investment advice.

APPENDIX A — Standard Diligence Questionnaire

National Bank of Canada (TSX: NA) — 2026-07-18

Supplemental to the main analysis. Figures in C$ unless noted. Where a question does not map to a bank’s business model, the correct sector analog is given.

General

What thoughtful questions have other investors asked about this company?

The most penetrating questions came from sell-side analysts on the FY2026 Q1 and Q2 calls, and they cluster on exactly the issues this report identifies as central:

• TD’s Mario Mendonca put the Personal & Commercial ROE gap at 600–800 basis points below peers and located it specifically in the efficiency ratio. National Bank’s own Chief Risk Officer answered, “You got it.” That is an unusually direct concession and it is the analytical starting point for this report. (Note: Mendonca’s framing was based on the Q1 figure of 12.7% adjusted; the Q2 figure deteriorated to 10.9%, so the gap is now wider than he stated.)
• RBC’s Darko Mihelic pressed on CWB’s stand-alone growth, observing that acquired revenue was flat at roughly C$298M year-over-year. The observation went unrebutted. If a C$5.0 billion acquisition is not growing eighteen months after announcement, the revenue-synergy target is in question.
• Mendonca separately caught an ABA Bank disclosure inconsistency — C$47M of impaired formations matching an identical C$47M increase in gross impaired loans — and management conceded “resolutions are remaining low.”
• Scotiabank’s Mike Rizvanovic characterized Q2-FY2026 as “decent” but “mixed in composition,” noting the EPS beat came from an outsized Financial Markets quarter offsetting weaker Canadian P&C and Wealth. The stock fell 4% that day despite a beat and a dividend increase.
• Nigel D’Souza / Ebrahim Poonawala-style capital questions: why run “an embarrassingly low 2% NCIB” at a 13.9% CET1 ratio? Management initially resisted buybacks (“buybacks are a complement, not a growth strategy”) before reversing and upsizing the programme to 14.5M shares.

The question this report believes investors are not yet asking: if the Canadian retail bank earns 11.8% and the consolidated bank earns 16.7%, what multiple should be applied to earnings that are 37.5% capital markets and 14.3% Cambodian and US specialty finance? The relative multiples across the Big Six are arguably the wrong way round — CIBC, the most retail-concentrated bank with the least mix assistance, earns a similar headline ROE and trades at the group’s lowest P/E, while National Bank trades at the highest.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? A cyclical high, on several dimensions at once (INTERPRETATION, well-evidenced). Canadian bank sector capital-markets profits rose 27% in Q2-2026; provisions across the Big Six are running below through-cycle averages with reserve releases flowing to earnings; and National Bank is the most capital-markets-levered of the six. Impaired PCLs at 26–33bps sit below a normalized 45–50bps. Management has itself described the trading outperformance as volatility-regime-dependent. The one genuine offset: the FY2025 base was depressed by the C$230M Day-1 CWB provision, so year-over-year growth optics overstate the underlying trend in both directions.

Driven by the external environment or internal actions? Both, and the split matters. Internal and creditable: CWB cost and funding synergies (C$215M realized, target raised to ~C$300M and pulled forward a year) are genuine execution. External: the credit normalization, the capital-markets boom, the rate-cut cycle and the sector-wide re-rating are environmental and shared with all five peers. The 2025–26 share-price advance is overwhelmingly external — all six Canadian banks returned +68% to +76% over twelve months, and National Bank’s +68.6% was the second weakest.

How stable are revenues? Bifurcated. Personal & Commercial (27.2% of segment earnings) and Wealth Management (21.0%) are genuinely stable annuity streams — sticky Quebec deposits, fee income on AUM. Financial Markets at 37.5% is not stable, and is already declining: Q2-FY2026 revenue −2%, net income −3% year-over-year, segment efficiency deteriorating 430bps. USSF&I at 14.3% is stable in growth terms but exposed to Cambodian and US consumer-credit risk.

Outlook for products/services? Canadian banking products are mature and commoditized; competition is on price, distribution and service rather than product. The growth vectors are share (Laurentian, CWB), wealth AUM, and — in the bank’s own framing — cross-sell revenue synergies that have so far delivered C$33M against a C$200–250M FY2028 target.

How big will this market be — growing, shrinking, domestic or international? Overwhelmingly domestic (Canada), with Cambodia and US specialty finance as the exceptions. The Canadian market’s structural growth rate is deteriorating: loan growth over 2021–24 leaned heavily on record population growth, and the immigration-policy reversal removes that tailwind permanently rather than cyclically. Canada is in a technical recession (Q1-2026 real GDP −0.1% annualized) with household debt at 179.6% of disposable income, the highest in the G7. This is a low-growth, high-leverage, mature market.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Less, at the structural level — and this is the strongest point in the bulls’ favour. Supply is contracting: Laurentian is exiting retail, CWB has been absorbed, HSBC Canada was acquired by RBC, CIBC is exiting the Caribbean, Scotiabank is retreating from Latin America. In Marathon’s capital-cycle framework, fewer independent competitors is the canonical positive inflection. The offsetting pressures are at the margin — EQ Bank and Tangerine on deposit pricing, credit unions regionally, prospective open banking, and continued interchange compression. The critical exception specific to National Bank: in its home market it competes against Desjardins, a member-owned cooperative requiring no market return on equity, and in its newly acquired market against ATB Financial, Crown-owned with the same advantage. Competing against non-profit-maximizing institutions in both core geographies is a structural handicap that does not appear in industry-level analysis.

How profitable is the business (ROIC, ROE)? ROIC is not a meaningful metric for a bank (interest expense is a cost of goods sold, not financing); ROE and ROTCE are the correct analogs. Consolidated ROE: 15.9% reported / 16.8% adjusted in Q2-FY2026; 13.7% / 15.3% in FY2025; a peak of 20.4% in FY2021. The segment detail is the whole story — adjusted H1-FY2026: Wealth 58.2%, ABA Bank (Cambodia) 29.4%, Financial Markets 27.4%, Credigy 21.4%, Personal & Commercial 11.8%. Return on assets has drifted from 0.91% (FY2021) to 0.77% (FY2025) — a cleaner deterioration signal than ROE because it is unaffected by the equity issued for CWB.

How profitable is the industry — how many competitors, what barriers to entry? Highly profitable and deliberately protected. Six banks hold ~90% of Canadian banking assets. Barriers are regulatory and government-granted: the Bank Act, OSFI licensing and prudential supervision, “widely-held” ownership rules capping single shareholders, a standing federal posture against large domestic bank mergers, and effective exclusion of foreign retail entry. In Greenwald’s taxonomy this is economies of scale plus customer captivity resting on a regulatory barrier — with the regulatory barrier primary, because it is what makes the other two durable. Greenwald rates government protection the strongest barrier type precisely because it does not expire on a schedule. The market-share-stability test — his best moat diagnostic — scores near maximum: Big Six shares have moved only low-single-digit percentage points over multi-decade spans, with no member entering or exiting and no modern failure or bailout.

Can the business be easily understood? Yes at the segment level, but National Bank is materially harder to understand than a plain Canadian retail bank, and that opacity is itself relevant. A reader must separately assess a Quebec retail bank, an institutional dealer with a large equity-derivatives book, a wealth manager, a US distressed-consumer-loan purchaser (Credigy), and a Cambodian commercial bank. The consolidated ROE conceals rather than reveals — which is why the February 2026 segment disclosure was so informative.

Can it be undermined by foreign low-cost labour? No. Banking is a domestically regulated, relationship- and license-based business. Back-office offshoring is a marginal cost lever available equally to all six.

Do brands matter? Moderately, and asymmetrically. In Quebec, National Bank’s brand carries genuine linguistic and cultural incumbency — real customer captivity. Outside Quebec, and especially in the newly acquired Alberta/BC territory, the brand carries little weight against RBC, BMO, CIBC and ATB. This is precisely why the CWB acquisition is strategically questionable in Greenwald’s terms: brand and local scale do not travel.

What is the nature of competition? Rational and oligopolistic in pricing, intense in service and distribution. Canadian banks do not compete destructively on price — mid-teens sector ROEs across decades demonstrate that. Competition is for share of wallet, advisor talent, and deposit relationships.

Customers’ switching costs? Real but moderate in retail — direct-deposit and pre-authorized-payment rewiring, mortgage discharge friction, and inertia produce genuine deposit stickiness. High in Wealth Management — advisor relationships, custody transfer friction and household inertia are the reason that segment earns 58%. Low in Financial Markets — institutional trading business is re-won on price and execution continuously, which is exactly why that segment’s returns are a capability rather than a moat.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The Quebec deposit franchise and the National Bank Financial advisor network are genuine economic assets carried at nil. Wealth Management’s 58.2% ROE is the financial fingerprint of an intangible asset — a distribution and relationship franchise — that appears nowhere on the balance sheet. Conversely the CWB acquisition converted an unrecognized asset problem into a recognized one: ~C$2.2 billion of goodwill and intangibles now sit in the P&C denominator.

Off-balance-sheet liabilities? Standard banking items: securitization vehicles, undrawn credit commitments, letters of credit, derivative notional exposures, and operating lease obligations. Nothing anomalous was identified. The genuine off-balance-sheet-like risk is the derivatives and structured-products book in Financial Markets — economically on balance sheet, but with risk characteristics (long volatility, regime dependence) that the reported figures do not convey. FRTB implementation will make market-risk RWA more volatile, and management has raised this with OSFI.

How conservative is the accounting? Broadly conservative, with two caveats. Conservative: fifteen consecutive quarters of performing-allowance build, an allowance covering 5.9× charge-offs, and the C$230M Day-1 IFRS 9 provision on acquired CWB performing loans taken up front. Caveats: (1) the adjusted-versus-reported wedge was C$0.68 across H1-FY2026, and adjusted figures exclude amortization of intangibles bought with C$5 billion of real shareholder currency — an economic cost, so reported is treated as the conservative anchor throughout this report; (2) purchase-accounting credit-mark accretion on the acquired CWB loan book flows into net interest income over time and modestly flatters margin. A third item warrants monitoring rather than criticism: ABA Bank’s ~C$1 billion of gross impaired loans with low write-offs and low resolutions may mean reported impaired balances lag realized losses.

How CapEx-hungry is the business? Low in the manufacturing sense; the correct analogs are technology spend and regulatory capital. Technology and compliance investment is the real capital intensity of modern banking, and it is scale-driven — which structurally disadvantages the smallest of six. Regulatory capital is the binding constraint: CET1 at 13.5% converging to ~13% by FY2027, releasing roughly C$2.5 billion, with a further 50–75bps expected from AIRB approval on the CWB book.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy? Free cash flow is not a meaningful bank metric; the correct analog is capital generation above the regulatory minimum. National Bank generated roughly C$4.0 billion of net income in FY2025 and is running at a ~C$5.0 billion annualized rate in FY2026. Uses, in order of magnitude over the past two years: the CWB acquisition (~C$5.3 billion in stock), dividends (~41% of adjusted earnings), buybacks (14.5M share authorization, 6.4M repurchased at an average C$165.75), and the pending Laurentian acquisition. The stated philosophy — “buybacks are a complement, not a growth strategy” — reveals a preference for M&A over return of capital.

Significant acquisitions recently? Yes, the defining fact of the last two years. Canadian Western Bank: announced 2024-06-11 at 0.450 NA shares per CWB share (C$52.35, a 110% premium to CWB’s market price but roughly 1.19× its book value), ~C$5.3 billion equity consideration, closed 2025-02-03. Laurentian Bank’s retail/SME business is scheduled to close in late 2026 (1.5–2% accretion, −25 to −30bps CET1). Management has flagged appetite for further tuck-ins in P&C and wealth. The critical assessment: CWB and Laurentian are strategically opposite — Laurentian deepens the region of dominance and removes a competitor (Greenwald-correct); CWB extends into markets where others hold local scale (Greenwald-adverse) and is by far the larger commitment.

Buying back shares? Yes, after initial reluctance. NCIB launched September 2025 (8.0M shares), upsized February 2026 to 14.5M (~3.7% of shares outstanding), with 6.4M repurchased at an average of C$165.75 — well-timed against today’s C$234. Forward caution: repurchasing stock at 2.8–3.0× book is EPS-accretive but book-value-dilutive, and improves ROE by shrinking the denominator rather than earning more on it. Management’s own FY2027 bridge attributes ~40bps of ROE improvement to buybacks versus just ~20bps to CWB revenue synergies.

Issuing large amounts of new shares to insiders? No evidence of excessive insider issuance. The material dilution was external and transactional: diluted share count rose 11.6% in FY2025 (342.8M → 382.5M) to fund CWB, after eight years of near-flat share count (339.9M in FY2016 to 342.8M in FY2024). Nine years of dilution discipline were spent in a single transaction.

Compensation policy of directors/management? OPEN QUESTION — and the most important unresolved item in this file. The Management Proxy Circular could not be retrieved during this engagement, so the specific weighting of ROE within short- and long-term incentive metrics is unverified. This matters disproportionately here: if compensation rewards EPS growth, absolute earnings, or total shareholder return rather than return on equity, then an acquisition programme that grows the company while diluting returns is a rational response to the incentive scheme rather than a strategic error — and should be expected to continue. This remains an open question.

Motivations of management? INTERPRETATION, held with appropriate caution. CEO Laurent Ferreira rose through the capital-markets side, which is at minimum consistent with a firm whose dealer has grown to 37.5% of earnings. The observable behavioural pattern is a preference for building scale through acquisition over returning capital, combined with genuine candour about the problem — Ferreira publicly conceded P&C returns are “subpar versus our peers” and opened a strategic review rather than defending the number, and the bank voluntarily began disclosing segment ROE, which is the disclosure that makes this report’s critique possible. That candour is a genuine mark in management’s favour and should be weighed against the capital-allocation critique. Management is not concealing the problem; it is deploying capital as though the problem were smaller than the data suggests.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? None of these. National Bank is a Canadian federally chartered bank listed on the TSX only — there is no NYSE listing and no ADR programme (a thin OTC line exists but is not the primary market). It is not an SEC registrant: no 10-K, no 40-F, and no Form 4 insider feed. Canadian filings go to SEDAR+, and insider transactions to SEDI, which is not reliably retrievable programmatically — so the insider read for this engagement is open, not clean. US holders should note there is no US withholding-treaty simplification: Canadian withholding tax applies to dividends, generally 15% for US residents under the treaty in taxable accounts.

Dividend policy? A long, consistent record of growth. The stated policy targets a 40–50% payout of net income attributable to common shareholders, and the adjusted payout ratio has averaged 42% over ten years. Current dividend C$1.32 quarterly (raised 6 cents to C$1.24 in December 2025, then 8 cents to C$1.32 in May 2026, +6%), C$5.28 annualized, a 2.26% yield — the lowest of the Big Six, which is itself a valuation signal. Dividend increases were interrupted by OSFI between 2020-03-13 and 2021-11-04 across the entire sector.

How profitable is the business? See above. The one-line answer: consolidated 16.7% adjusted, which is good — assembled from a 58% wealth arm, a 27% dealer, a 29% Cambodian bank, and an 11.8% Canadian bank.

Is net income diverging from cash from operations? Cash-flow statements are of limited analytical use for banks — operating cash flow swings violently with trading-book, deposit and settlement-balance timing and is not a quality-of-earnings signal as it would be for an industrial. The correct analogs are the adjusted-versus-reported wedge (C$0.68 across H1-FY2026, being CWB integration charges and intangible amortization) and provision adequacy (allowance at 5.9× charge-offs, fifteen consecutive quarters of performing-allowance build — comfortable). No evidence of earnings quality deterioration in the accrual sense; the earnings-quality question here is about mix and durability, not about accruals.

Risks & Downside

What factors would cause the stock to decline? In descending order of likelihood: (1) multiple compression — the stock trades above its entire eleven-year range on P/E, P/B and P/TBV simultaneously and sits 26% above its 200-day EMA in a sector-wide melt-up where all six banks are 24–27% above theirs; (2) Financial Markets mean reversion — 37.5% of segment earnings, already down year-over-year with efficiency deteriorating 430bps; (3) failure of P&C to improve from 10.9–11.8%, confirming a structural rather than transitory returns problem; (4) CWB revenue synergies missing — C$33M realized against a C$200–250M FY2028 target, in an acquired bank that is not currently growing; (5) credit normalization toward 45–50bps; (6) Canadian macro deterioration — recession, tariffs, the immigration-driven loan-growth reversal.

Risk of a catastrophic loss? Low. National Bank is a systemically important, federally supervised member of a protected oligopoly with 13.5% CET1, 5.9× allowance coverage, a full-recourse and partly CMHC-insured mortgage book, no US regional-bank exposure, and no US commercial-real-estate concentration of the kind that damaged peers. The realistic downside case is multiple compression against merely adequate earnings — roughly −3% annually over three years — not impairment of capital.

Chance of a total loss? Negligible. No Canadian Big-Six bank has failed or required a bailout-style rescue in modern history, including through 2008. Total loss would require a systemic Canadian banking failure with regulatory forbearance exhausted — a scenario in which the entire cohort, and most Canadian financial assets, would be impaired simultaneously. The two identifiable tail risks that could produce something materially worse than the bear case are (a) a severe Canadian housing and employment shock striking a Quebec-concentrated book at the same time as a trading-revenue collapse, and (b) — smaller but genuinely non-zero — an AML enforcement or sovereign/currency event at ABA Bank in Cambodia producing a headline and capital charge out of proportion to that segment’s size.

Recent News & Events

Has the business environment changed recently? Yes, materially and in both directions. Improving: the mortgage-renewal wall largely passed without the predicted default wave (~80% of the affected book repriced, payment shock “significantly reduced,” 90-day delinquencies below pre-pandemic levels); provisions normalized; a sector-wide capital-markets boom (+27% Big Six capital-markets profits in Q2-2026); and the April-2025 tariff fear unwound. Deteriorating: Canada entered a technical recession (Q1-2026 GDP −0.1% annualized); household debt reached 179.6% of disposable income; the immigration-policy reversal permanently removed a loan-growth tailwind; and the CUSMA renegotiation remains unresolved.

Significant acquisitions? Covered above: CWB (closed 2025-02-03, ~C$5.3 billion) and Laurentian’s retail/SME business (closing late 2026). African assets were divested in FY2025; ABA Bank (Cambodia) was retained with no strategic-review commentary on any of the five most recent calls.

Change in accounting policies? No material policy change was identified in the period reviewed. The significant accounting events were acquisition-driven: the C$230M Day-1 IFRS 9 provision on acquired CWB performing loans (FY2025), purchase-accounting fair-value marks on the acquired loan book (accreting into net interest income), and ~C$2.2 billion of goodwill and intangibles added to the P&C segment. Note that National Bank restates its dividend payout ratio from time to time as EPS figures are revised; the restatements are real, though their causes were not verified in this engagement.

Recent changes — new markets, facilities, management?

• New markets: Alberta and British Columbia commercial banking via CWB — the first material geographic extension in the bank’s modern history.
• Systems: CWB conversion completed ~late February 2026; AIRB approval on the acquired book still pending, expected mostly in Q4-FY2026.
• Disclosure: from Q1-FY2026 National Bank began publishing segment-level ROE — the single most consequential change of the period for an outside analyst, and the basis of this report’s central finding.
• Capital: NCIB launched September 2025 and upsized February 2026; dividend raised twice; CET1 target set at ~13% by FY2027.
• Management: Laurent Ferreira remains CEO; Marie Chantal Gingras is CFO. No senior departures of note were identified. A strategic review of Personal & Commercial has been opened.

APPENDIX B — Source Appendix

National Bank of Canada (TSX: NA) — 2026-07-18

All sources accessed 2026-07-18 unless otherwise stated. Primary sources are listed first. Management commentary is labelled as hypothesis, not evidence.

1. Primary — Company filings and disclosures

National Bank of Canada is not an SEC registrant. There is no 10-K, no 40-F, and no Form 4 insider feed. Canadian continuous-disclosure filings go to SEDAR+; insider transactions to SEDI. The equivalent local-jurisdiction documents were used instead.

2. Primary — Management commentary

(treated as hypothesis requiring external validation, per this firm’s standing rule)

*Transcript coverage note: the aggregated transcript service carried FY2025 Q2/Q3 and FY2026 Q1 but had gaps at FY2025 Q4 and FY2026 Q2 — the two most decision-relevant calls — which were reconstructed from newswire and transcript-service coverage (Benzinga, Investing.com) and cross-checked against the primary Report to Shareholders and Supplementary Financial Information wherever a figure was load-bearing. *

3. Market and quantitative data

Data unavailable for this ticker, and disclosed as such rather than treated as a finding:

• The news and sentiment feed returns zero articles for TSX-only tickers. The the recent-changes section recent-events timeline was built from company press releases and trade press instead.
• The own-history valuation percentile series returns null. The eleven-year own-history range in the own-history valuation section was rebuilt manually from fiscal-year-end multiples — this is the substitute for the missing percentile ranks.
• The empirical factor model returns an empty loadings array (its universe is US common stock only). The positioning read in the variant-perception section is therefore built from the price series alone and is stated with that limitation — no factor loadings, no risk-adjusted leaderboard.

4. Secondary — trade and financial press

(used for event attribution and macro context; every load-bearing figure cross-checked to a primary source)

5. Analytical frameworks