The Bank of Nova Scotia (NYSE/TSX: BNS) — The Laggard, Re-Rated on a Turnaround It Hasn’t Finished

Independent fundamental research — general information, not investment advice. Figures in Canadian dollars (C$) unless noted; fiscal year ends October 31. Scotiabank is a foreign private issuer (40-F/6-K); the body of this report carries no recommendation and no price target — the single exception is the clearly-labeled author’s opinion (“Claude’s Take”) immediately below.

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice. The body of this report that follows is deliberately position-free and carries no price target.

Verdict: HOLD / accumulate on weakness — not a short. The genuinely cheap Big-Six bank, but cheap for a real reason and now re-rated +149% off its low, so you are paying for the recovery rather than the trough. Conviction: MEDIUM. Fair-value zone ~C$84–95 (~US$61–69) on the base-case ~12–13% ROE — i.e., roughly today’s price. The attractive entry is a pullback into the C$70–78 (~US$51–57) band, ~1.0–1.15× book, where the 5.5% dividend pays you to wait and you get the 14%-ROE optionality nearly for free. Don’t chase above ~C$92–95.

Scotiabank is the cheapest seat in the best banking oligopoly on earth — ~1.2–1.35× book and a ~5.5% yield versus RBC at ~3.1× — and that discount is finally being attacked by a credible outsider CEO whose return-on-equity inflection is real: 13.2% in Q2-FY26 against a ~9% fiscal-2025 base, with 14%+ promised a year early. The framing is deep-value-turning-show-me, not momentum and not falling knife: the factor model files BNS as a low-volatility Value + Dividend-Yield + Canada-beta proxy (related to Canada ETFs and Manulife, not the other banks, with no momentum loading despite +64% on the year) — the market owns it for its yield, not yet for its quality story. That is the whole opportunity and the whole trap. The bull only needs the marginal buyer to re-file BNS from “Canadian dividend bond-substitute” to “quality bank converging on peers,” worth 30–60%. The bear note is that the easy money is made: the 13.2% print is flattered by capital-markets strength (+25%), a “high-watermark” Caribbean NIM, and one-off marks, while management raised its credit-loss guidance into a Canadian mortgage-renewal wall and a structurally noisier Latin-American book. I sit at HOLD because the base case is already in the price and the upside now requires execution, not just re-rating — but unlike RBC/CIBC at penthouse multiples, BNS still pays you 5.5% to be patient, which is why it’s the most ownable of the group on a pullback.

Conviction: MEDIUM. Flips bullish if BNS prints a clean, cyclically-adjusted ROE marching toward 14% with International PCLs moderating below ~140bp and Canadian loan growth catching the market. Flips bearish if ROE plateaus below ~12% with International PCLs stuck above ~150bp — confirming the cheapness is structural and the recovery was cyclical. Tag: “Paid 5.5% to wait on a turnaround that’s half-proven.”

📈 Stock Price Action — Five-Year Event Map

Scotiabank has round-tripped from a pandemic-recovery base in the low-C$50s, down through the 2022–23 rate shock to a five-year low of C$34.64 (TSX) on 2023-10-27 — the deepest trough of the Big Six, struck on the same day the sector bottomed — and then up a near-uninterrupted +149% to a five-year high of C$87.03 on 2026-06-18. It closed 2026-06-26 at C$86.16 (~US$62–63 on the NYSE at USDCAD ~1.37), roughly 1% off its five-year high, with a 52-week range of approximately C$60–87. The defining feature is that the laggard of the group has, over the past 20 months, been one of its strongest price performers — a re-rating off a uniquely depressed base, not an earnings breakout. (Price moves below are Fact; attributed drivers are Interpretation.)

Cycle narrative. (1) Scotiabank re-rated out of COVID as the Big Six resumed capital return once OSFI lifted its 2020 distribution freeze. (2) The 2022 rate shock compressed every Canadian-bank multiple, and BNS — the most emerging-markets-exposed of the group via its Pacific Alliance (Mexico/Peru/Chile/Colombia) franchise — fell further as global risk-off hit Latin-American currencies and credit. (3) The slide bottomed at C$34.64 on 2023-10-27, the lowest trough in the group, amid the U.S. regional-bank aftershock and peak-rate fear. (4) The turn coincided with new CEO Scott Thomson’s December-2023 strategy reset — a deliberate pivot toward the “North American corridor,” capital redeployment, and the August-2024 acquisition of a 14.9% stake in U.S. regional KeyCorp (~US$2.8B) — alongside falling-rate optimism, lifting the stock back above C$49 by end-2024. (5) The April-2025 US–Canada tariff shock produced the last real dip. (6) Through 2025 the whole group re-rated on an earnings-beat streak and easing credit, and BNS — starting from the cheapest base — participated fully, closing 2025 at ~C$72. (7) The 2026 leg has been powered by a genuine return-on-equity inflection: Q2-FY26 ROE of 13.2% (versus a ~9% fiscal-2025 base) and management’s claim that the 14%+ target will arrive in FY2027, a year early, drove the stock to an all-time high. The single most important fact for valuation is that the 2024–26 advance is overwhelmingly a re-rating off a distressed multiple, with the ROE recovery only beginning to validate it.

1. Executive Summary

The Bank of Nova Scotia is the structural laggard of the Canadian Big Six — a C$1.46-trillion-asset universal bank whose defining feature, a 130-year Latin-American and Caribbean franchise that no domestic peer replicates, has for a decade been the source of both its differentiation and its discount. BNS has earned the lowest return on equity in the group (~9.1–9.7% in FY25, versus RBC ~17%, CIBC ~14–15%, TD adjusted ~13%), the consequence of an International book that carries roughly triple the Canadian cost of risk, structural FX and political volatility, and — critically — a sub-scale follower position in most of its emerging markets rather than the local-oligopolist position that earns excess returns. Diluted EPS has gone nowhere for three years (C$7.99 in FY22 → C$5.84 in FY25). The market has been right to discount it: at ~1.2–1.35× book it trades at the cheapest multiple in the group, and the justified-P/B framework makes that discount internally coherent at a 9% ROE.

The investment question is whether that is changing. An outsider CEO, Scott Thomson (appointed February 2023), has executed a credible strategic reset — redeploying capital from low-return Latin America to the North American corridor (Canada–U.S.–Mexico), fixing the funding mix toward sticky deposits, growing capital-light wealth and fee income, and running returns rather than assets as the scorecard. The concrete moves are consistent with the words: a 14.9% optionality stake in KeyCorp, the exit of Colombia/Costa Rica/Panama to Davivienda (crystallizing a C$1.36B impairment that tallies a decade of value destruction), the Bank of Xi’an exit, and the Asia loan runoff. And the inflection is now showing in the numbers: Q2-FY26 ROE reached 13.2% (+270bps YoY), the productivity ratio hit ~52.5% (already at the medium-term target), Canadian Banking delivered its fourth straight quarter of NIM expansion and third of positive operating leverage, and Global Wealth posted a seventh consecutive quarter of positive net flows at a 17.9% segment ROE. Management now expects the 14%+ ROE target in FY2027, a full year ahead of plan.

The skeptic’s case is equally grounded. The ROE recovery is partly cyclically flattered — capital-markets revenue +25%, a “high-watermark” International NIM of 4.76%, a Caribbean spread propped up by the absence of U.S. rate cuts, and one-off marks/tax items — while credit is a headwind, not a tailwind: management raised its FY26 impaired-PCL guidance into a Canadian mortgage-renewal wall and a stressed consumer, and the International book ran a 166bp loss rate with recurring episodic corporate losses. After a +149% advance off the 2023 low to an all-time high, BNS trades at the 95.7th percentile of its own P/E history (on still-depressed earnings) — the easy, re-rating money is made. The factor model reinforces the tension: BNS is owned as a low-volatility Value/Dividend-Yield/Canada-beta proxy — clustered with Canada ETFs and Manulife, not the other banks — not yet as a quality re-rating story.

Netting it out, the body that follows takes no position. The body’s verdicts: a genuinely durable domestic oligopoly moat diluted by a structurally inferior international mix; the lowest-quality earnings base of the group, recovering toward respectability but not yet proven to have reached it; competent capital-allocation damage-control rather than a value-creating engine; and a valuation that prices partial credit for the turnaround — a sustained ~10.5–11% ROE — leaving balanced asymmetry around a base case near today’s price, with a 5%+ yield as paid-to-wait support and execution, not re-rating, as the swing factor from here.

2. Business Overview

The Bank of Nova Scotia (“Scotiabank”) is the third- or fourth-largest of Canada’s “Big Six” banks, with approximately C$1.46 trillion of total assets (FY25, fiscal year ended Oct-31), roughly 25 million clients, and about 85,800 employees (FACT; FY25 disclosure). Founded in Halifax in 1832 and headquartered in Toronto since 1900, it is dual-listed (TSX in C$, NYSE in US$) and files with the SEC as a foreign private issuer (40-F/6-K under the MJDS regime, not a 10-K filer, and with no Form 4 insider feed). Its defining characteristic — the thing that makes it different from every domestic peer — is that it is “Canada’s most international bank,” built on a 130-year-plus footprint in Latin America and the Caribbean that no other Big Six bank replicates.

The bank reports in four segments:

• Canadian Banking — the foundational anchor and the most profitable, highest-return business: retail, small-business, and commercial banking across Canada (chequing/savings, residential mortgages, personal lending, credit cards, GICs, commercial lending and cash management). FY24 adjusted earnings were roughly C$4.28B (+7% YoY) (FACT). This is the segment management is “doubling down” on.
• International Banking — the differentiator and the source of the valuation discount: retail and commercial banking across the Pacific Alliance (Mexico, Peru, Chile, Colombia) plus a 200-plus-branch Caribbean network. FY24 adjusted earnings ~C$2.86B (+11%) (FACT). It carries a structurally higher net interest margin (Q2-FY26 NIM 4.76% vs. Canadian Banking ~2.3–2.4%) but also structurally higher provisions for credit losses (Q2-FY26 PCL 166bp vs. Canadian Banking 50bp) and far higher FX and political volatility (FACT; Q2-FY26 transcript).
• Global Wealth Management — the capital-light, fee-oriented growth engine: full-service brokerage, the 1832 Asset Management fund complex (ScotiaFunds), private banking, and international wealth across ~13 markets. FY24 adjusted earnings ~C$1.61B (+10%), AUM ~C$373B; Q2-FY26 segment ROE was 17.9% — roughly double the consolidated bank’s (FACT; transcript). This is the highest-quality earnings stream in the company.
• Global Banking & Markets (GBM) — the wholesale/capital-markets arm (corporate lending, debt/equity origination, M&A advisory, sales and trading) serving corporate, institutional, and government clients, weighted to the Americas. FY24 earnings ~C$1.69B (FACT). The Asia loan book is being deliberately run off (transcript).

Key operating subsidiaries include Tangerine (the digital-first bank acquired as ING Direct Canada), Scotia Capital and 1832 Asset Management in the wholesale/wealth complex, and the international franchises Scotiabank Inverlat (Mexico) and Scotiabank Chile. In a notable recent capital deployment, the bank holds a 14.9% equity stake in KeyCorp, a ~US$187B-asset U.S. regional bank, acquired for ~US$2.8B (announced Aug-2024) — a minority “optionality” position in the U.S. leg of the North American corridor (FACT; company release, Aug-2024). It is simultaneously exiting parts of Latin America: in January 2025 it agreed to transfer its Colombia, Costa Rica and Panama operations to Davivienda in exchange for ~20% of the combined entity (regulatory approval received Nov-2025), converting three sub-scale country operations into a passive minority stake (FACT). Against that, in June 2026 it proposed to buy in the minority of Scotia Group Jamaica for ~C$0.5B — increasing a Caribbean holding even as it trims the LatAm mainland (FACT; AZI news, 2026-06-12).

How it makes money. Like any bank, roughly half of revenue is net interest income — the spread on a ~C$1.46T balance sheet — and roughly half is fee, commission, wealth, and trading income. The earnings mix skews more toward spread-and-credit (Canadian + International retail/commercial banking together generate the clear majority of earnings) and less toward fee/markets income than higher-multiple peer RBC. The structural feature that distinguishes BNS is geographic: a meaningful slug of earnings is generated in emerging-market currencies (Mexican peso, Peruvian sol, Chilean peso) and is therefore subject to both higher nominal growth and higher translation and credit volatility when converted to Canadian dollars.

Verdict. Scotiabank is a complete, four-engine universal bank built on a high-return Canadian core, a capital-light wealth franchise, a respectable wholesale arm, and a genuinely unique — but lower-return, higher-risk — Latin American/Caribbean retail bank. The franchise is real and diversified; the open question is whether the international leg is a diversifier that has earned its capital or a value-drag that has dragged consolidated returns to the bottom of the group for a decade.

3. Industry Dynamics

The starting point is the same structural advantage that benefits every Big Six bank, and it is one of the best in global banking. Canadian banking is a federally engineered oligopoly. Six banks — RBC, TD, Scotiabank, BMO, CIBC, and National Bank — control roughly 90%+ of Canadian banking assets (FACT; industry data, consistent with the RY/CM analyses). This concentration is not an accident of competition but the product of a deliberate regulatory architecture: the Bank Act, prudential supervision by OSFI, “widely held” ownership rules that cap any single shareholder, a long-standing federal posture against large bank mergers, and effective barriers to foreign-bank retail entry. The practical results are rational pricing, benign (not value-destroying) intra-group competition, very high barriers to entry, and a regulator whose conservatism — high CET1 and liquidity buffers, the mortgage stress test, full-recourse mortgage lending — kept the system out of every modern banking crisis, including 2008. On the Greenwald framework this is a textbook economies-of-scale-plus-customer-captivity industry with high barriers; on the Marathon capital-cycle lens, the supply of new banking capital is constrained by regulation, which is exactly why Big Six returns stay high and stable rather than mean-reverting. Scotiabank, as a Canadian D-SIB, sits inside this protected garden and runs a fortress 13.3% CET1 ratio against an ~11.5% requirement (FACT; Q2-FY26).

Where Scotiabank’s industry exposure diverges — and this is the crux — is that a disproportionate share of its assets and earnings sit outside that protected Canadian oligopoly, in the structurally different banking markets of the Pacific Alliance and the Caribbean.

The Latin American tilt: higher growth, lower risk-adjusted returns. Latin American banking offers genuinely higher nominal growth than mature Canada — driven by low banking penetration, favorable demographics, rapid digital/fintech adoption, and (often) higher structural interest rates that produce wide margins. Scotiabank’s International NIM of ~4.4–4.8% is roughly double its Canadian NIM (FACT; transcript). But that higher gross spread is not free money: it compensates for materially higher credit losses (International PCLs run ~150–170bp vs. ~30–50bp in Canada), higher operating costs, sovereign and political risk, and FX volatility that can erase peso/sol earnings on translation to C$. The honest read (INTERPRETATION) is that LatAm banking is a higher-nominal-return, lower-risk-adjusted-return business than Canadian banking — which is precisely why the market has historically applied a structural valuation discount to BNS’s international earnings and why those earnings have been chronically more volatile.

Mexico as the crown jewel. Within the international portfolio, Mexico is the asset management most wants to keep and grow: it is the largest, most strategically coherent international franchise (Scotiabank Inverlat), the natural beneficiary of North American nearshoring and the USMCA/CUSMA corridor, and it is performing — Q2-FY26 Mexican revenue rose 8% and earnings 25% year-over-year (FACT; transcript). Mexico is the one international market where BNS’s “North American corridor” logic and its LatAm footprint actually overlap.

The Canadian-side risks are real too. Two sector-specific headwinds bear on the Canadian book. First, the 2025–27 mortgage-renewal wall: Canadian mortgages reset roughly every five years, so a wave originated at pandemic-era ~2% rates is renewing at materially higher rates, raising household debt-service burdens (the same dynamic flagged in the RY and CM analyses). Second, the highly-levered Canadian consumer: management explicitly cites “stresses in the Canadian portfolio,” prolonged inflationary pressure on “vulnerable client segments,” and elevated energy costs, and raised its FY26 impaired-PCL guidance to the mid-50bp range (FACT; Q2-FY26). CEO Thomson counters with a constructive 2027 view (oil-exporting nation, fiscal stimulus, a business-friendly government, CUSMA endurance), but that is management hypothesis, not evidence.

Verdict: a mixed industry exposure — a great domestic oligopoly diluted by a structurally inferior international mix. The Canadian half of Scotiabank sits in one of the best banking markets in the world. The international half sits in markets that are higher-growth but lower-return on a risk-adjusted basis, more cyclical, FX-exposed, and — critically for the competitive section below — markets where BNS is generally not the local oligopolist. The net is an industry mix that is structurally good but materially worse than a Canada-pure peer’s. The entire BNS debate reduces to whether that international tilt is a diversification benefit worth owning at a discount, or a persistent value-drag the bank is right to be shrinking.

4. Competitive Position

Domestically, Scotiabank owns the same oligopoly moat as its peers — scale, customer captivity, and regulatory protection — but it is the junior partner of the top tier. In Greenwald’s taxonomy the Canadian moat is economies of scale + customer captivity + government/regulatory barrier: a national branch/ATM/digital network, ~25M sticky primary-banking relationships, high switching costs (primary chequing accounts, pre-authorized payments, mortgages, bundled wealth), and the Bank Act/OSFI/widely-held-ownership wall that keeps new entrants and foreign retail banks out. That moat is real and shows up in financial outcomes — stable deposits, a defensible Canadian NIM, and a Canadian Banking segment that earns a high allocated-equity return. Scotiabank reinforces it with two genuine assets: Tangerine, a credible digital-only brand, and the Scene+ loyalty program (the Empire/Sobeys grocery partnership), a real customer-acquisition and cross-sell flywheel. But within the oligopoly, BNS holds roughly 13–14% of the Canadian banking market — a clear second tier, behind RBC and TD (each ~22%) and on par with BMO (FACT). It is a participant in the best banking market in the world, not its leader.

Internationally, the moat largely evaporates — and this is the decisive competitive fact about BNS. In most of its Latin American markets, Scotiabank is a sub-scale follower, not the local oligopolist. It is the #5-ish player in markets where the top two or three domestic incumbents own the deposit franchise, the payment rails, and the pricing power. The contrast with a genuine local champion is stark: Credicorp/BCP in Peru is a dominant, captive-deposit, ~30%-share national franchise earning a high-teens-to-20%+ ROE precisely because it sits at the top of a concentrated local oligopoly. Scotiabank’s Peru, Chile, and Colombia operations have never occupied that position — which is exactly why the bank is now swapping its Colombia/Costa Rica/Panama operations for a passive ~20% stake in Davivienda (a better-positioned local player) rather than trying to win those markets head-on (FACT). A follower in a higher-risk market earns the risk without the franchise advantage that would justify it. That is the textbook signature of a weak moat: international scale that has not converted into international returns.

The proof is in the consolidated return. Scotiabank has earned the lowest ROE of the Big Six for years — roughly 9.1–9.7% in FY25 — against RBC at ~17%, CIBC at ~14–15%, National Bank ~16–17%, and TD (adjusted) ~13% (FACT). It also carries the group’s highest dividend payout ratio (~74% in FY25) and historically the highest reliance on more expensive, more volatile wholesale funding (a function of a high loan-to-deposit ratio the new strategy is explicitly trying to fix). The internal segment gap is stark: International Banking earns an allocated-equity ROE materially below the Canadian Banking segment’s — i.e., the international segment dilutes consolidated returns, and the more capital it consumes the worse the blended number looks. A moat that does not translate into returns is, by definition, a weak moat. For a decade, Scotiabank’s international franchise was a moat-shaped object that destroyed relative value: the bank grew its balance sheet across Latin America but consistently earned below its higher-quality, Canada-concentrated peers and was rewarded with the group’s lowest price-to-book.

Direct comparison. Versus RY (the prime franchise: highest ROE, most diversified, best capital), BNS is structurally lower-return and lower-multiple. Versus TD (large U.S. retail, currently constrained by its U.S. AML/asset-cap penalty), BNS trades the U.S. regulatory overhang for LatAm credit/FX volatility — a different, arguably comparable, source of discount. Versus CM (more domestically concentrated, more uninsured-mortgage-levered, but now earning ~14–15% ROE), BNS is more diversified geographically but worse on returns — diversification that has not paid. On the metric that matters — risk-adjusted return on equity — BNS has been the laggard of the group, full stop.

Verdict: structurally disadvantaged within a good oligopoly. Domestically, Scotiabank shares a genuinely durable moat but ranks in the second tier. Internationally, it is a sub-scale follower whose footprint has produced higher risk without commensurate return, dragging consolidated ROE to the bottom of the Big Six. The competitive question is no longer whether the international franchise is a moat — the decade-long return gap answers that — but whether new management can fix the return on the capital that remains and redeploy the rest into higher-return North American businesses. The moat the market will eventually pay for is the Canadian core plus wealth; the international book has to earn its multiple, and so far it has not.

5. Growth History and Forward Opportunities

The history is the laggard’s history: flat-to-down earnings since the FY22 peak. Diluted EPS ran C$7.67 (FY21) → C$7.99 (FY22, the cyclical peak) → C$5.75 (FY23) → C$5.91 (FY24) → C$5.84 (FY25) (FACT). On a continuing-operations basis the FY25 figure is ~C$6.22, with ~C$0.38 of divestiture/impairment charges (notably a ~C$1.36B Q1-FY25 charge tied to the Colombia/Davivienda wind-down and KeyCorp-related items). Either way, earnings are roughly flat-to-down over three years while peers re-based higher — the financial fingerprint of a bank whose international engine stalled (FX, credit, sub-scale returns) and whose Canadian engine, though healthy, could not carry the group alone. Revenue grew over the period (C$30.9B in FY21 to C$37.1B in FY25), but the gap between top-line growth and stagnant EPS is itself the indictment: growth that did not drop to the bottom line, the classic low-quality “growth without economics.” Historically, much of the international expansion was acquired (Inverlat in Mexico, Colpatria in Colombia, various Caribbean/Pacific-Alliance deals) rather than organic, and the integration/return record of that M&A is precisely what dragged consolidated ROE.

The strategy: a North American-corridor pivot under new leadership. CEO Scott Thomson (an outsider, appointed Feb-2023) laid out a medium-term plan at the December 2023 Investor Day with explicit targets: an EPS CAGR of ≥7%, a consolidated ROE of ~14%, and a productivity (efficiency) ratio of ~53% over the medium term (FACT; Scotiabank 2023 Investor Day). The strategic content is a deliberate reallocation of capital — directing the bulk of incremental capital to the higher-return, lower-risk Canada–U.S.–Mexico corridor, and away from the lower-return international markets. The concrete actions are consistent with the words: the 14.9% KeyCorp stake (a low-integration-risk U.S. entry with “optionality”), the Davivienda swap (converting three sub-scale LatAm country operations into a passive minority stake), deposit-led Canadian growth (the new relationship-tiered Scotia High Interest Savings Account, “>90% retention of GIC maturities,” primary-client deepening, Scene+/Tangerine), fee growth in wealth and GBM, and a stated capital-priority stack of organic growth > buybacks > small tuck-in M&A. Importantly, management is buying back stock because of the valuation gap to peers — rational, given the lowest P/B in the group.

The inflection is now showing in the numbers — and it looks more than purely cyclical. Q2-FY26 (quarter ended Apr-30) delivered adjusted EPS of C$2.02, pre-tax pre-provision profit +16% YoY, consolidated ROE of 13.2% (+270bps YoY), and a productivity ratio of 52.5% — at or near the Investor Day target (FACT; transcript). Management now expects to hit the 14%+ ROE target in FY2027, a full year ahead of plan. The composition matters for the quality judgment:

• Canadian Banking: PPPT +13%, a fourth consecutive quarter of NIM expansion, third consecutive quarter of positive operating leverage, accelerating commercial loan growth (+2% q/q) and high-single-digit small-business growth, with a deliberate shift toward higher-quality fee income (45% of new card acquisition premium; record branch mutual-fund sales, fees +21% YoY) and sticky deposits — structural mix improvement, not just a rate tailwind.
• Global Wealth Management: net sales of C$4.7B (4x the prior-year quarter), a seventh consecutive quarter of positive flows, ROE 17.9%, and C$9B YTD of two-way referrals from the banking franchise — the highest-quality, most capital-light growth in the company, and genuinely organic.
• International Banking: PPPT +12%, revenue +7%, YTD positive operating leverage of 3.2%, with Mexico the standout (+8% revenue / +25% earnings). The strategy here is a balance-sheet mix shift — non-mortgage retail growing twice the pace of mortgage, deposits +5–6%, expensive deposits optimized out — i.e., improving returns on a flat-to-shrinking book rather than chasing volume.
• GBM: revenue +9%, capital markets +25%, with the Asia book in deliberate runoff.

The honest caveats (INTERPRETATION): part of the YoY ROE jump is helped by easing comparisons, divestiture clean-up, and a favorable International NIM that management itself flagged as partly seasonal (~7–10bp Q2 benefit) and partly a windfall from the absence of U.S. rate cuts propping up Caribbean spreads. PCL guidance was raised for FY26 (mid-50bp), the Canadian consumer remains stressed, and a Brazilian corporate impairment showed the international book’s idiosyncratic-loss tail. So the cyclical tailwind is real. But the operating-leverage story — four straight quarters of Canadian NIM expansion, three of positive operating leverage, a productivity ratio already at target, seven quarters of wealth inflows, and an improving deposit/fee mix — is structural and management-driven. The improvement is being manufactured, not just received.

Verdict: historically low-quality growth, now inflecting toward higher quality — and the inflection looks partly real, not purely cyclical. For a decade BNS grew its balance sheet without growing per-share economics; the laggard’s discount was earned. The Thomson strategy is the right one — fix the funding mix, deepen primary clients, grow capital-light wealth and fee income, redeploy capital from low-return international markets, and let returns (not assets) be the scorecard. Two-plus years in, the evidence is encouraging. The thesis-defining question is durability — whether the 14%+ ROE survives the Canadian mortgage-renewal wall, a stressed consumer, and the next LatAm credit/FX shock, or whether it proves to be a cyclically-flattered peak. The turnaround is real enough to take seriously; it is not yet proven enough to call won.

6. Financial Quality

Scotiabank is a C$1.46 trillion-asset bank whose financial signature within the Big Six is unambiguous: the lowest profitability, the highest dividend payout, the weakest coverage, and the widest gap between “reported” and “adjusted” earnings. The investment debate is entirely about whether the FY26 inflection — adjusted ROE rising from a FY25 trough toward a 14%+ FY27 target — is a durable repair of a structurally sub-par franchise or a cyclical bounce dressed up by adjusting items. The evidence says: partly real, materially flattered.

Revenue composition and NIM (FACT). FY25 total revenue was ~C$37.1–37.7B, up ~12% YoY. The growth is higher-quality than the headline suggests: all-bank net interest margin expanded to 2.33% in FY25 from 2.16% in FY24 (+17bps), and to 2.40% in Q4-FY25 (Scotiabank Q4-FY25 release, 2025-12-02). Two engines drove this. (1) Canadian Banking posted a fourth consecutive quarter of NIM expansion through Q2-FY26, as the mortgage book reprices upward on renewal and the deposit mix shifts toward sticky savings/chequing balances; management retains “>90% of retail GIC maturities” even amid deposit competition. (2) International NIM reached 4.76% in Q2-FY26, a “high watermark,” helped by Latin-American funding-cost relief (Mexico/Chile/Peru rate cuts) and — a rate-environment gift — no U.S. rate cuts propping up Caribbean deposit spreads (CFO Viswanathan, Q2-FY26 call). Fee income is also improving in mix: Canadian non-interest revenue grew ~10% YoY in Q2-FY26 on premium cards (45% of new acquisition), record branch mutual-fund sales (+21% YoY), and insurance — and Global Wealth posted a seventh consecutive quarter of positive net flows (C$4.7B in Q2-FY26, 4x prior year) at a 17.9% segment ROE. (INTERPRETATION) The revenue quality is genuinely improving — more spread from mix and repricing, more capital-light fee income — and this is the most credible structural leg of the recovery.

The ROE story (FACT). Consolidated reported ROE was 9.1% in FY25 (~9.7% on the bank’s own basis), the lowest of the Big Six and below the FY22 cyclical peak of 13.6%. EPS has gone nowhere for three years. The recovery is in the quarters: Q2-FY26 ROE was 13.2% (+270bps YoY) on ~13% revenue growth and 16% PPPT growth, and management asserts ROE is “on track to hit 14%+ in fiscal 2027, one year ahead of the Investor Day target.” On tangible common equity (stripping ~C$16.2B of goodwill + intangibles from ~C$76.9B common equity, leaving ~C$60.8B tangible), reported ROTCE was ~12% in FY25 and adjusted ROTCE ~15% (author’s calculation from balance-sheet data) — the metric bulls point to for a re-rating, though note it is itself an adjusted figure.

Efficiency (FACT, with a wrinkle). Reported FY25 productivity worsened to ~59.7% — but this is distorted by the Colombia held-for-sale accounting and restructuring noise. On an adjusted basis the trend is the opposite: Canadian Banking productivity improved 230bps YoY in Q2-FY26, all-bank productivity hit ~52.5%, year-to-date operating leverage was positive, and the bank delivered three consecutive quarters of positive operating leverage. The medium-term target is ~50%. (INTERPRETATION) The cost-out is real and management-driven, not cyclical — the highest-conviction part of the turnaround.

Credit quality — the bear case (FACT). This is where the durability of the inflection is most exposed. Full-year FY25 PCLs were C$4.71B (up from C$4.05B in FY24), a ratio of roughly ~60bps. In Q2-FY26 the all-bank PCL ratio rose to 66bps (impaired 61bps), and management raised its impaired-PCL guidance to the mid-50s bps range for the remainder of FY26 — explicitly higher than the December guide, because “the macro environment has evolved meaningfully” (CRO McGinnis, Q2-FY26). The split is stark and structural:

• Canadian Banking PCL ~50bps, with management flagging that “prolonged inflationary pressures could further strain already vulnerable client segments” — i.e., the levered Canadian consumer and the mortgage-renewal wall are a live, not abated, risk. The Q2 improvement came from collections efforts, not underlying healing.
• International Banking PCL ~166bps — roughly 3x the Canadian rate — structurally embedding the higher loss content of the LatAm book. Q2-FY26 also carried a single Brazilian investment-grade corporate (“fallen angel,” ~7bps of all-bank impaired), which management calls episodic.
• Total allowance for credit losses reached C$7.3B, ~96bps of loans (Q2-FY26) — a meaningful buffer build.

(INTERPRETATION) Credit is currently a headwind, not a tailwind, to the ROE recovery. The 14% FY27 target implicitly assumes PCLs moderate from H1-FY26 levels; if Canadian retail deteriorates on the renewal wall or LatAm credit normalizes higher, the inflection stalls. This is the single biggest swing factor and the reason the market still applies a discount.

CET1 and balance-sheet strength (FACT). CET1 was 13.2% at FY25 year-end (13.3% in Q2-FY26), comfortably above OSFI’s ~11.5% D-SIB minimum, and held even after completing the 2025 NCIB and starting the 2026 program. RWAs were ~C$474B (Q2-FY26). Common equity/assets is ~5.5%. Capital is not a constraint; it is a strength. (Aggregator EV/EBITDA ~196x, net-debt/EBITDA, and interest-coverage figures are meaningless for a bank and are disregarded.)

Quality of Earnings — the adjusted-EPS bridge

The defining QoE issue at Scotiabank is the size and nature of the adjusting items, which are not random noise but the realized cost of fifteen years of value-destructive Latin-American capital allocation (FACT/INTERPRETATION):

(Sources: Scotiabank Q4-FY24, Q1-FY25, Q4-FY25 releases; Q2-FY26 call. FY23 adjusted approximate.)

The C$1.25/share (21%) FY25 gap between reported (C$5.84) and adjusted (~C$7.09) EPS is the crux. Management’s “adjusted” figure conveniently excludes the cost of unwinding its own prior bad M&A — the C$1.36B Colombia impairment crystallized a decade-plus of sub-cost-of-capital deployment. A skeptical reader should treat the adjusted numbers as the right forward run-rate (the divestitures are genuinely non-recurring) while remembering that the reported numbers are the honest tally of capital destroyed. Encouragingly, by H1-FY26 the adjusting items have shrunk to near-zero — the cleanup is largely done.

Decomposing the ROE recovery (INTERPRETATION):

• Structural / durable: Canadian NIM expansion and mortgage repricing; deposit-mix shift to sticky balances; cost-out (productivity improvement, positive operating leverage); capital-light wealth and fee mix; removal of the low-ROE Colombia/Xi’an drag. Perhaps half to two-thirds of the improvement.
• Cyclical / non-repeatable: GBM capital-markets revenue +25% YoY (a trading/IB rebound, not a moat); International NIM at a 4.76% “high watermark” including ~7–10bps of seasonality and a rate-dependent Caribbean tailwind; ~C$35–46M of one-off corporate-segment mark-to-market gains (which management guided to reverse); a Peru tax refund and LatAm inflation adjustments lowering the effective tax rate. And critically, credit is not yet helping — it is a drag with rising guidance.

Verdict. The economics are improving and the inflection is real but only partly durable. The cost-out and Canadian NIM/mix repair are structural and management-controlled; the divestitures remove a genuine low-return anchor; and the adjusted-EPS quality is improving as adjusting items fade. But the climb from 9% to a 14% FY27 ROE leans on cyclical capital-markets strength, rate-dependent Caribbean spreads, one-off marks/tax items, and an assumption that credit moderates — even as PCL guidance was just raised and the Canadian-consumer/mortgage-renewal and LatAm credit risks remain unresolved. This is the lowest-quality earnings base of the Big Six, recovering toward respectability, not yet proven to have reached it.

7. Capital Allocation

Scotiabank’s capital-allocation record is the historical root of its valuation discount, and the current regime is best understood through Marathon’s capital-cycle lens: a fifteen-year, top-of-the-cycle build-out into high-“growth” Latin-American banking that earned sub-cost-of-capital returns, now being unwound in a classic late-cycle retreat — with the freed capital redeployed defensively into U.S. optionality, Canadian organic growth, and pro-cyclical buybacks.

The M&A history is a study in value destruction now being reversed (FACT). The legacy strategy assembled Banco Colpatria (Colombia), Cencosud card businesses, and a sprawling Pacific-Alliance/Caribbean footprint at prices and a moment (the 2010s EM-credit upcycle) that never produced peer-level returns — International Banking carried a structurally higher cost and loss content (FY25 PCL ~166bps vs. Canadian ~50bps) and a lower ROE than the domestic bank, dragging consolidated ROE to last in the group. Thomson’s team is now dismantling it:

• Colombia / Costa Rica / Panama → Davivienda (Jan-2025 announced, closed Dec-1-2025): transferred in exchange for an ~20% stake in the combined Davivienda Group, booking an after-tax impairment of ~C$1.36B and a ~10–15bps CET1 hit. (INTERPRETATION) The right move executed at a real loss — converting a controlled, capital-hungry, sub-scale franchise into a passive minority equity stake. It improves the quality of the remaining book but is an admission that the original investment failed.
• Bank of Xi’an (China): impaired ~C$309M after-tax in Q4-FY24 and being exited; the GBM Asia loan portfolio is in runoff.
• Scotia Group Jamaica buy-in (2026-06-12): a proposal to acquire the minority shares it does not own for ~C$0.5B cash (~5bps CET1), taking the Jamaican bank private. (INTERPRETATION) Notably this increases Caribbean ownership — slightly at odds with the “North American corridor” narrative — but it is a high-ROE, well-understood franchise being bought in cheaply; defensible as capital optimization rather than expansion.

The KeyCorp stake — optionality, not control (FACT). Scotiabank deployed ~US$2.8B for a 14.9% equity interest in KeyCorp (Cleveland-based U.S. regional, ~US$187B assets), completed in two tranches (4.9% Aug-2024; to 14.92%, ~163M shares, Dec-2024) at ~US$17.17/share. It is accounted for as an associate, contributing ~C$60M/quarter equity-pickup. In Mar-2026 Scotiabank received Fed approval to potentially increase the stake. (INTERPRETATION) A deliberately minimal-risk way to plant a U.S. flag — immediate earnings accretion and “future optionality” on a deeper deal, with none of the integration risk of a full acquisition. But it is precisely not value-creating growth: a passive minority in a rate-exposed regional buys de-risked exposure and an option, while ceding control and capping upside. It shrinks the discount by reducing volatility; it does not, by itself, lift through-cycle ROE.

Dividends and buybacks — highest payout, weakest coverage, pro-cyclical repurchase (FACT). Scotiabank paid DPS of ~C$4.72 in FY25 (raised by C$0.04/quarter in Q2-FY26), a yield of ~5.0–5.5% at C$86 — the highest of the Big Six. But the payout ratio is also the highest and the coverage the thinnest: ~75% on reported FY25 EPS, ~67% on adjusted EPS (vs. peers’ ~45–55% target band). The dividend is not at risk — CET1 is 13.2% and adjusted coverage is adequate — but the cushion is the smallest in the group, and dividend growth will stay modest. Buybacks tell a pro-cyclical story: share count actually rose from ~1,215M (FY21) to ~1,244M (FY24) as the bank issued through DRIP/scrip while conserving capital during the lean years, then fell only ~8M to ~1,236M in FY25 as repurchases resumed — 6.4M shares in Q2-FY26 — near the 5-year high (~C$87) rather than at the C$35 trough of late-2023. Total capital returned was C$7.5B over the LTM. (INTERPRETATION) Buying back stock at ~1.3x book is accretive given the persistent discount, and management explicitly cites the valuation gap — but the firm conspicuously did not buy when the stock was cheapest, because capital was being rebuilt. Rational sequencing, but the opposite of contrarian capital allocation.

Compensation and incentives (INTERPRETATION / ASSUMPTION — live proxy not retrieved). The medium-term scorecard is now anchored on an explicit ROE target (14%+ by FY27), an improvement over a pure EPS-growth orientation and better aligned with through-cycle value creation. The flag: reliance on adjusted metrics that exclude restructuring and divestiture losses means the cost of fixing past capital-misallocation does not fully hit incentive comp — the very mistakes being unwound were partly insulated from the scorecard.

Insider behaviour (FACT — disclosure limitation). Scotiabank is a foreign private issuer filing 40-F/6-K; Canadian insiders report via SEDI, not the SEC Form 4 feed, so insider open-market buy/sell activity is not retrievable in this workflow. No insider-conviction signal can be asserted either way, and none is implied.

Verdict. Thomson’s team is allocating capital more intelligently than its predecessors — exiting capital-hungry, sub-cost-of-capital LatAm franchises, planting a low-risk U.S. flag, and returning a rising share of capital — and the direction is right. But the honest characterization is de-risking, not value-creating redeployment. The Colombia exit crystallizes (rather than creates) value; the KeyCorp stake buys optionality while capping upside; buybacks are pro-cyclical; and the highest-payout/weakest-coverage dividend constrains reinvestment flexibility. This is competent damage-control that should narrow the discount, not yet a capital-allocation engine that earns a premium.

8. Changes and Headwinds — Last Two Years

The last two years are the entire investment story: a new, outsider CEO and a deliberate strategic reset that is finally showing in the numbers, set against a worsening macro and credit backdrop that threatens the timeline.

New leadership and the strategy reset (FACT). Scott Thomson — recruited from outside traditional banking (former CEO of Finning International), a Scotiabank director before being named President in Dec-2022 and CEO in February 2023 — brought a refreshed executive bench (CFO Raj Viswanathan, CRO Shannon McGinnis, and new heads across Canadian Banking, International, Wealth and GBM). The December 2023 Investor Day crystallized the pivot: redirect the bulk of incremental capital to the North American “corridor” (Canada–U.S.–Mexico), prioritize primary-client relationships and sticky, low-cost deposits to fix the historically high loan-to-deposit ratio, target ~50% productivity and 14%+ ROE, and step back from broad international expansion. (INTERPRETATION) Two-plus years in, the operational evidence (NIM expansion, positive operating leverage, wealth flows, deposit-mix shift) validates the execution on controllable levers; the strategy is credible and being delivered, which is why the stock has re-rated from ~C$45 to ~C$86.

Portfolio reshaping (FACT). The KeyCorp 14.9% stake (US$2.8B, 2024); the Colombia/Costa Rica/Panama exit to Davivienda (C$1.36B impairment, closed Dec-2025); the Bank of Xi’an exit and GBM-Asia runoff; the pending Scotia Group Jamaica buy-in (C$0.5B); restructuring/severance charges across FY24–FY25; and a deliberate Mexico build-out (Q2-FY26 Mexico revenue +8% / earnings +25% YoY). These collectively shrink the low-ROE, high-volatility tail and concentrate capital on Canada, Mexico and the U.S. (INTERPRETATION) Net positive for earnings quality and the multiple, at the cost of growth optionality — Scotiabank is becoming a more boring, more North-American, more buy-backable bank, which is the point.

Technology / AI (FACT). In Q2-FY26 the bank launched Scotia Intelligence (an enterprise AI governance/platform layer) and Scotia Navigator (employee-facing AI assistant), with technology spend +9% YoY and a “model-agnostic, security-first” posture. (INTERPRETATION) Table-stakes catch-up — historically Scotiabank’s digital UX ranked mid-pack — and a cost lever as much as a growth lever; not yet a differentiator.

The headwinds (FACT).

• Macro / trade war: Management repeatedly cited “unexpected geopolitical developments,” CUSMA/tariff uncertainty, elevated energy costs and higher Canadian unemployment as having “evolved meaningfully” since the December guide — directly driving the raised PCL guidance (mid-50s bps for H2-FY26). Thomson is publicly optimistic on Canada into 2027 but conceded the recovery is “more gradual than we had originally anticipated.”
• The mortgage-renewal wall: the Canadian consumer is renewing pandemic-era mortgages at higher rates; Q2 retail credit improved on collections efforts, not healing.
• LatAm political/FX/credit risk: International PCL ~166bps, episodic corporate “fallen angels” (the Q2 Brazil name), and earnings exposed to Mexican/Chilean/Peruvian rate, FX-translation and political cycles — structurally the more volatile book even after the Colombia exit.
• Cross-currency reporting: results are IFRS/CAD; the NYSE-listed common is exposed to USD/CAD translation for U.S. holders.

Verdict. The structural changes strengthen the thesis: a credible CEO, a coherent and partly-delivered reset, a cleaner portfolio, improving NIM/efficiency/wealth flows, and a closing reported-vs-adjusted EPS gap. The cyclical changes weaken the near-term setup: a deteriorating macro, a rising PCL trajectory, an unresolved Canadian-consumer/mortgage-renewal risk, and persistent LatAm volatility — all on the path to the 14% FY27 ROE target. On balance the trajectory is genuinely improving and management is doing the right things, but the re-rating from C$45 to C$86 has already priced much of the self-help, leaving the stock increasingly hostage to a credit-and-macro backdrop management does not control. The changes strengthen the franchise; the headwinds threaten the timeline — and at ~1.3x book the timeline is what matters now.

9. Risk Analysis

Scotiabank’s risk profile is the most complex of the Big Six because its earnings sit on two distinct fault lines — a Canadian consumer rolling into the mortgage-renewal wall, and a Latin-American franchise that adds political, currency, and episodic-corporate-credit risk no domestic peer carries. Layered on top is the central thesis risk: a turnaround whose 14% ROE target is only partly delivered and whose recent earnings are flattered by several items that may not repeat.

The turnaround-execution risk is the thesis. Scotiabank has earned the lowest ROE in the Big Six for three consecutive years (~9.3% FY23, ~9.4% FY24, ~9.1% FY25), and the stock’s +149% advance has been the market paying forward Thomson’s promise to close that gap to 14%+. Q2-FY26’s 13.2% is real and encouraging — Canadian Banking delivered its third consecutive quarter of positive operating leverage and fourth of NIM expansion, and International generated 3.2% YTD operating leverage. But the gap from a ~9% fiscal-year base to a sustained 14% is wide, and management is now reaching the easier part (cyclical NIM tailwinds, expense discipline) before the harder part (durable loan growth — still only “low single digits” and not expected to catch the market until year-end). If the ROE path stalls in the 11–12% zone, the stock has already priced a 14% outcome.

Latin-American credit is structurally noisier and currently elevated. Q2-FY26 International provisions ran at 166bp — more than triple Canadian Banking’s 50bp — driven partly by a single investment-grade Brazilian corporate the CRO characterized as an episodic “fallen angel,” but also by genuinely elevated retail impairments. Management guided all-bank impaired PCLs up to the mid-50s-bp range for the rest of FY26 (from a lower December outlook), explicitly because Canadian-consumer inflation pressure and LatAm energy costs are worse than expected. The Pacific Alliance is the source of both BNS’s higher structural growth and its higher loss volatility and FX risk — the discount and the differentiator are the same variable.

Recent earnings are flattered, which matters at a 95th-percentile P/E. Three Q2-FY26 tailwinds are explicitly cyclical or one-time: capital-markets revenue +25% YoY (a high, not a run-rate); an International NIM of 4.76% the CFO called a “high watermark,” partly because the Caribbean book benefited from the absence of U.S. rate cuts; and ~C$35–46M of mark-to-market and equity-pickup gains in the “other” segment that management said will revert to a modest loss. Strip these and the underlying ROE is lower than the 13.2% headline — a critical caveat when the market is capitalizing trailing earnings at the richest multiple in the stock’s own history. The dominant risks are execution and cyclical-earnings risks, not franchise or solvency risks; the realistic adverse case is a 20–30% drawdown if the ROE recovery disappoints, not a wipeout.

10. Valuation

Scotiabank must be valued on bank-appropriate metrics — P/E, P/B, P/TBV, P/PPOP, ROE, and dividend yield. Corporate-style measures are meaningless for a balance-sheet business: an aggregator’s EV/EBITDA of ~173x and any net-debt/EBITDA figure for BNS are mechanical artifacts and are disregarded entirely. The whole valuation question is a single tension — BNS is simultaneously the cheapest Big-Six bank on absolute price-to-book and the most expensive version of itself on its own multi-year history — with the reconciliation found in whether the ROE turnaround is real and durable.

(a) Where BNS trades — own-history and cross-sectionally. At C$86.16, on TTM EPS of C$6.03 and AZI’s book-value basis of ~C$70.78, BNS trades at:

• ~14.3x trailing earnings — the 95.7th percentile of its own multi-year P/E range (richest on record), but on depressed trailing earnings still carrying ~C$0.38/share of FY25 charges. On a normalized/forward basis the multiple is far lower (~11–12x the ~C$7.5 annualized H1-FY26 run-rate).
• ~1.22x book — the 91.1st percentile of its own history, yet the LOWEST absolute price-to-book of the Big Six (RY ~3.1x, TD ~2.6x, CM ~2.4x, BMO ~2.1x). (Book-value definitions differ: on filing-based common equity of ~C$76.9B / 1,236M shares ≈ C$62, the multiple is closer to ~1.35–1.4x; ROIC’s BVPS of C$65.10 implies ~1.32x. On any basis BNS is the cheapest book multiple in the group; the percentile uses AZI’s own series.)
• ~1.9x sales — the 26.5th percentile (genuinely cheap), composite 71.1st. Tangible book ~C$50, so P/TBV ~1.7x.
• A dividend yield of ~5.0–5.5% — the highest of the Big Six — the “wait and get paid” feature, but resting on the group’s thinnest coverage (~74% reported payout).

The cross-sectional discount to peers is real and long-standing, and it exists for a reason: BNS has earned a structurally lower ROE (~9% vs. RY’s ~17%, CM’s ~14–15%) because of the lower-return, higher-cost-of-risk International book and weaker domestic operating leverage. The justified-P/B framework makes the discount internally coherent rather than an anomaly.

(b) Justified P/B = (ROE − g)/(COE − g), with a Canadian D-SIB cost of equity ~9.5–10% and long-run growth ~4–5%:

At a fiscal-2025 ROE of ~9%, the model justifies only ~0.9–1.0x book — i.e., at its trough earnings BNS deserved the ~1.0x book it traded at in 2023–24. At a sustained 14% ROE it would justify ~1.8–2.0x book, a CM/BMO-like multiple. The current ~1.2–1.35x book therefore prices a sustained ROE of roughly 10.5–11% — meaningfully above the depressed ~9% trough, but well short of the 14% target. The multiple gives the turnaround partial credit: it has paid for the move from “broken laggard” to “improving mid-tier,” but not for full convergence to the quality peers.

© Embedded-expectations read — what the price is underwriting. The ~1.2–1.35x book / ~14x trailing (~11–12x forward) multiple embeds a market view that BNS reaches a durable ~11% ROE — better than the trough, short of management’s goal. This is the crux: the market is pricing skepticism, not success. What it appears to price correctly: that the easy, cyclical part of the recovery (NIM expansion, expense discipline, capital-markets strength) is real and underway; that the dividend is safe enough to anchor a 5%+ yield; and that LatAm structurally caps the through-cycle ROE below RY’s. What it may be pricing incorrectly in either direction: it gives little credit for the 14% target actually landing on schedule (upside if Thomson delivers), but it also may be under-discounting how cyclically flattered the 13.2% print is (downside if capital markets, Caribbean NIM and the “other”-segment marks revert at once into a Canadian/LatAm credit normalization). The sell-side is more constructive than the tape — consensus is Neutral/Hold-heavy with targets clustered modestly above spot, reflecting the same “show-me” stance.

(d) Bear / Base / Bull scenarios (illustrative embedded-expectations framing — not a price target; USD at ~1.37):

The bull zone coincides with the published analyst-target range — i.e., the sell-side’s bullish targets require the full 14% outcome the market is not yet paying for. Verdict: BNS is the genuinely cheap Big-Six bank, but cheap for a structurally good reason (the lowest ROE in the group), and the +149% re-rating has already monetized the move from distressed to merely-discounted. From here the multiple prices partial credit for the turnaround: the asymmetry is more balanced than RY’s or CM’s penthouse-priced setups — a real (if cyclically-flattered) ROE inflection backed by a 5%+ yield gives genuine paid-to-wait support — but the upside now requires execution, not just re-rating, and the base case sits at roughly today’s price. This is no longer the deep-value layup it was at C$35; it is a show-me turnaround at a fair-to-full price on its own history.

11. Variant Perception

Consensus belief. The Street files Scotiabank as the Big-Six laggard in slow rehabilitation — the cheapest bank on book and the highest-yielding, with a credible new CEO finally addressing a decade of underperformance, but not yet worth a premium. Hence the Neutral/Hold-heavy consensus: nobody disputes the discount or the early ROE progress; the debate is whether the 14% target is achievable and durable, or whether LatAm and a Canadian credit turn keep BNS the perpetual also-ran. The prevailing view is “cheap, improving, but show me.”

The strongest bull case. Scotiabank is the cheapest Big-Six bank (~1.2–1.35x book) at the moment its ROE is genuinely inflecting — Q2-FY26 ROE of 13.2%, up 270bp YoY, with management guiding to 14%+ in FY2027, a year early. Canadian Banking is compounding positive operating leverage and NIM expansion; Wealth (ROE 17.9%, seven straight quarters of positive flows) is a capital-light bright spot; Mexico is delivering revenue +8% / earnings +25%. If Thomson delivers a sustained 14% ROE, the justified multiple moves from ~1.2x toward ~1.8–2.0x book — a re-rating toward the quality peers worth 30–60% — and you collect a 5%+ dividend while you wait. The cheapest valuation, the highest yield, the biggest self-help lever, the most room to converge: on a return-to-multiple basis it screens as the best risk/reward in the group if the turnaround is real.

The strongest bear case. BNS is a value trap that has finally, temporarily, worked. It has earned the lowest ROE in the Big Six for years because of structural reasons — a lower-return, higher-cost-of-risk Latin-American footprint and weaker domestic scale — that a strategy reset does not erase. The recent ROE recovery is cyclically flattered (capital markets +25%, a “high-watermark” International NIM, Caribbean NIM propped by no U.S. cuts, one-time “other”-segment marks), so the true run-rate ROE is below 13.2%. Meanwhile the easy money is made: +149% off the trough, now at a 95.7th-percentile own-history P/E and the highest payout/thinnest dividend coverage of the group, just as the Canadian consumer rolls into the renewal wall (management raised its FY26 impaired-PCL guide) and LatAm credit stays elevated (166bp Intl PCL, recurring episodic corporate losses). Buy the laggard at C$35 on 1.0x book; you are now paying up for the recovery at the high of its range, underwriting flawless execution against a deteriorating macro.

The 3–5 assumptions that matter most: (1) Does the ROE reach and hold 14%, or stall in the 11–12% zone once cyclical tailwinds fade? (2) Is the 13.2% print a clean run-rate or flattered by capital markets, Caribbean NIM, and one-off marks? (3) Does LatAm credit (and FX) normalize benignly, or do episodic corporate losses and Pacific-Alliance retail stress keep International PCLs elevated? (4) Does the Canadian consumer/renewal wall stay manageable? (5) Does the cross-sectional discount to peers close, or is it a permanent feature of a lower-quality franchise?

What would falsify each side. The bull breaks if FY26–27 ROE plateaus below ~12% with International PCLs staying north of ~150bp, confirming the cheapness is structural and the recovery cyclical — the stock then re-rates back toward ~1.0x book. The bear breaks if BNS posts a clean (cyclically-adjusted) ROE marching to 14% with LatAm credit moderating as guided and Canadian loan growth catching the market — in which case ~1.2x book is a real discount and the re-rating toward peers is deserved.

Factor-positioning read (where consensus may be offsides). The factor model is unusually informative for BNS and tilts toward the bear’s framing of who owns it and why. BNS loads as a Value (+0.32) and DividendYield (+0.18), Canada-beta (Country:Canada 0.48) name with a beta of only ~0.54, idiosyncratic volatility of just 11.9%, a negligible Banks-industry loading (0.084) — and, strikingly, NO Momentum loading despite a +64% trailing-year return. Its factor-similar peers are Canada ETFs (BBCA/FLCA/EWC) and Manulife (MFC), NOT the other Big-Six banks. The interpretation: the market holds Scotiabank as a generic Canadian yield/value proxy — a low-vol dividend bond-substitute — not as a quality-bank re-rating story. That is precisely the disconnect the bull thesis needs the market to resolve in its favor: if BNS is owned for its yield and Canada-beta rather than for an idiosyncratic ROE turnaround, the buyer base is not the cohort that re-rates a stock toward higher-quality peers — it is the cohort that rotates out when value/dividend factors fall out of favor or rates back up. The risk-adjusted record reinforces the “laggard” identity beneath the recent run: a trailing-year Sharpe of 3.68 (maxDD −13.4%) sits on top of a five-year +12.1%/yr (maxDD −40.5%) and a lifetime +8.8%/yr at a Sharpe of just 0.27 with a −63.8% max drawdown — the worst long-run risk-adjusted profile of the cohort. Net: the tape says BNS is currently a crowded, beloved Canadian dividend/value trade that has been a wonderful low-vol ride — but that is a different thing from a market that has come to believe the quality story, and it is exactly the positioning that unwinds first if either the ROE recovery disappoints or the value/yield regime turns. Treated as input, not a price call, the factor read sharpens the bear’s timing concern and the bull’s “the re-rating hasn’t even started in the eyes of the marginal buyer” optionality in equal measure.

12. Fact vs. Interpretation Table

13. Open Questions

1. What is the “clean,” cyclically-adjusted ROE? Stripping the capital-markets high, the Caribbean-NIM windfall, and the one-off marks/tax items from the 13.2% Q2-FY26 print — is the underlying run-rate 11%? 12%? Management has not disclosed a normalized figure, and it is the single most important number for the thesis.
2. Does Canadian loan growth actually “catch the market” by year-end as guided, or does a stressed consumer keep it stuck in the low single digits — the missing organic-growth leg of the 14% ROE bridge?
3. How elevated do International PCLs stay? Is the 166bp Q2 rate a peak (episodic Brazil corporate) or a new structural level for the post-Colombia book?
4. Is the KeyCorp stake a precursor to a full U.S. acquisition (Fed approval to increase already obtained), and if so at what price/dilution — or does it remain a passive minority indefinitely?
5. What does the live proxy show on incentive metrics — does executive comp use per-share or ROIC-style measures, and are restructuring/divestiture losses excluded from the scorecard?
6. Does the Jamaica buy-in signal renewed Caribbean appetite or is it a one-off capital-optimization, and how does it square with the “North American corridor” narrative?

14. What Must Be True

Bull case — what must be true:

1. The ROE reaches and holds ~14% (not just touches it cyclically) — requiring durable Canadian loan growth, sustained positive operating leverage, and the wealth/fee mix-shift to continue.
2. International credit normalizes toward ~120–140bp and FX is benign — the post-Colombia book proves to be genuinely de-risked, not merely smaller.
3. The Canadian mortgage-renewal wall passes without a material retail-credit spike.
4. The marginal buyer re-files BNS from “Canadian yield/value proxy” to “quality bank converging on peers,” re-rating the multiple toward ~1.5–1.8x book.

Falsification test: Two consecutive quarters of cyclically-adjusted ROE below ~12%, or International PCLs sustained above ~150bp, or Canadian retail PCLs spiking on the renewal wall — any one breaks the bull case and argues the cheapness is structural.

Bear case — what must be true:

1. The recent ROE recovery is mostly cyclical (capital markets, Caribbean NIM, marks) and fades back toward 10–11% as those tailwinds reverse.
2. LatAm remains a perpetual drag — episodic corporate losses and Pacific-Alliance retail stress keep International returns sub-par.
3. Canadian consumer credit deteriorates as the renewal wall bites, pressuring the thinly-covered dividend.
4. The +149% re-rating de-rates back toward ~1.0x book as the market re-prices a stalled turnaround.

Falsification test: A clean (cyclically-adjusted) ROE marching to 14% with International PCLs moderating below ~140bp and Canadian loan growth catching the market — confirming the discount is unwarranted and the franchise re-rate is deserved.

15. Source Appendix

See Appendix B — Source Appendix in the combined report for the full source list with URLs and access dates. Primary sources: Scotiabank FY25 Annual Report and Q2-FY26 (quarter ended 2026-04-30) report and earnings call (2026-05-27); Scotiabank December 2023 Investor Day; Scotiabank/Davivienda transaction releases (Jan-2025 / Nov-2025 close); KeyCorp stake disclosures (Aug/Dec-2024); Scotia Group Jamaica proposal (2026-06-12). Quantitative/market data: ROIC.ai (statements, ratios, per-share), price history and valuation percentiles, and factor-model data (loadings, leaderboard). Public peer reference: Credicorp/BCP (Peru) for the local-oligopolist contrast.

APPENDIX A — Standard Diligence Questionnaire

The Bank of Nova Scotia (NYSE/TSX: BNS) — supplemental to the research memo. Figures in C$ unless noted; FY ends Oct 31.

General

What thoughtful questions have other investors asked about this company? The dominant question is whether the Thomson turnaround is a genuine, durable repair of a structurally sub-par franchise or a cyclically-flattered bounce in the lowest-ROE Big-Six bank. Sub-questions: (a) what is the “clean” run-rate ROE under the 13.2% Q2-FY26 headline; (b) is the Latin-American franchise a diversifier or a perpetual value-drag now best shrunk; © is the KeyCorp 14.9% stake a precursor to a full U.S. deal; (d) is the ~5.5% dividend (highest payout in the group) safe through a Canadian-consumer credit turn; and (e) does the cross-sectional discount to RBC/CIBC close, or is it permanent.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Mixed and transitional. Consolidated ROE (~9% FY25) was near a multi-year low; the Q2-FY26 13.2% is a recovery but is partly flattered by cyclical-high items (capital-markets revenue +25%, a “high-watermark” International NIM, one-off marks). Net: off the trough, not yet at a sustainable peak — and credit is a headwind, not the usual late-cycle tailwind.

Driven by external environment or internal actions? Both. Internal (durable): cost-out/productivity, Canadian NIM repricing, deposit-mix shift, wealth-flow momentum, divestiture of low-ROE units. External (cyclical): trading/IB rebound, LatAm rate cuts, a Caribbean-NIM windfall from no U.S. cuts, FX. Roughly half-to-two-thirds of the ROE improvement looks structural.

How stable are revenues? Canadian banking and wealth are stable/recurring; International retail/commercial is structurally more volatile (FX translation, EM credit cycles); GBM capital-markets is the most cyclical line. BNS has the least stable revenue mix of the Big Six because of the EM tilt.

Outlook for products/services / how big will the market be? Canadian banking is mature/low-growth but high-return; wealth is a structural grower; Mexico (nearshoring/CUSMA corridor) is the highest-conviction international growth market; the rest of LatAm is being de-emphasized. Net growth ambition: ≥7% EPS CAGR, ~14% ROE medium-term.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? The Canadian oligopoly is stable (deposit competition has intensified at the margin but structure is intact). LatAm markets are more competitive for BNS as a sub-scale follower facing dominant local incumbents (e.g., Credicorp in Peru).

How profitable is the business (ROIC/ROE)? ROE ~9% FY25 (lowest of group), recovering to 13.2% Q2-FY26; ROTCE ~12% reported / ~15% adjusted. For a bank, ROE/ROTCE is the relevant return metric (ROIC/EV-based measures are not meaningful).

How profitable is the industry / barriers to entry? Canadian banking is one of the most profitable, highest-barrier banking markets globally (Bank Act, OSFI, widely-held ownership rules, no large-bank mergers, foreign-retail-entry barriers). BNS shares those barriers domestically but ranks second-tier (~13–14% share).

Can the business be easily understood? Mostly — a universal bank — but the four-segment, multi-currency, EM-exposed structure makes it the hardest of the Big Six to model cleanly (FX translation, country-level credit, adjusting items).

Can it be undermined by foreign low-cost labour? Not materially; banking is regulated and relationship/deposit-franchise-based. (BNS does run global business services hubs offshore as a cost lever.)

Do brands matter? Switching costs? Yes — primary-banking relationships, mortgages, pre-authorized payments, bundled wealth, and the Scene+ loyalty/Tangerine assets create real switching costs domestically. Internationally the brand is weaker (follower position).

Financial Condition & Balance Sheet

Assets not fully recognized / off-balance-sheet? The ~20% Davivienda stake and the KeyCorp 14.9% stake are carried as associates (equity method) and may hold optionality not fully reflected; AUM (~C$373B) and AUA generate fee income off-balance-sheet. Customer relationships/deposit franchise are not capitalized.

Off-balance-sheet liabilities? Standard bank items (commitments, guarantees, derivatives); nothing flagged as unusual versus peers. Pension obligations are modest.

How conservative is the accounting? Mixed. IFRS/OSFI framework is conservative on capital. But management’s adjusted EPS conveniently excludes the recurring costs of unwinding prior bad M&A (restructuring, divestiture impairments) — read reported EPS as the honest tally. CET1 13.3% is genuine strength.

How CapEx-hungry is the business? Low physical capex (a bank); the real “capex” is technology spend (+9% YoY, the Scotia Intelligence/Navigator AI build) and credit-loss provisioning. Capital intensity is regulatory-capital, not PP&E.

Capital Allocation & Management

How much FCF, and how is it used? Bank “FCF” ≈ earnings less retained capital for RWA growth. Capital priority: organic growth > buybacks > small tuck-in M&A. ~C$7.5B returned LTM (dividends + buybacks).

Significant acquisitions recently? KeyCorp 14.9% stake (~US$2.8B, 2024); Scotia Group Jamaica minority buy-in proposed (~C$0.5B, 2026). Divestitures dominate: Colombia/Costa Rica/Panama → Davivienda (closed Dec-2025, ~C$1.36B impairment); Bank of Xi’an exit; GBM-Asia runoff.

Buying back shares? Yes, resumed — 6.4M shares Q2-FY26; but share count had risen 2021–24 (DRIP/scrip during lean years) and buybacks are pro-cyclical (near 5-yr highs, not the 2023 trough).

Issuing large amounts to insiders? No unusual insider issuance flagged; dilution historically via DRIP/scrip, now reversing.

Compensation policy / motivations of management? Medium-term scorecard now anchored on an explicit ROE target (14%+ by FY27) — an improvement. Caveat: reliance on adjusted metrics insulates the cost of fixing past misallocation. CEO Scott Thomson (outsider, ex-Finning) is credible; the entire re-rating is tethered to his strategy.

Valuation & Market Data

ADR/MLP/K-1? Neither an ADR (dual-listed common, NYSE + TSX) nor an MLP/K-1. Canadian withholding tax applies to dividends for U.S. holders (often reduced/recoverable in registered accounts per treaty). Foreign private issuer (40-F/6-K).

Dividend policy? ~C$4.72/yr, ~5.0–5.5% yield (highest of Big Six); payout ~74% reported / ~67% adjusted (thinnest coverage). Modest growth (C$0.04/qtr Q2-FY26 bump).

How profitable? Lowest-ROE of the Big Six (~9% FY25), recovering. Profitable but sub-peer on returns.

Net income vs cash from operations diverging? Bank CFO is volatile and not a clean quality signal (driven by balance-sheet/trading flows); use earnings, PPPT, and ROE instead. The relevant divergence is reported vs adjusted EPS (a ~21% gap in FY25), which reflects divestiture/restructuring charges.

Risks & Downside

What would cause the stock to decline? A stalled ROE recovery (plateau at 11–12%); a Canadian-consumer/mortgage-renewal credit spike; a LatAm credit/FX shock; reversal of the cyclical tailwinds (capital markets, Caribbean NIM); a value/dividend-factor regime turn that unwinds the crowded yield-proxy ownership; a general multiple de-rate after +149%.

Risk of catastrophic loss? Very low — a D-SIB at 13.3% CET1 in an OSFI-supervised system that survived 2008/COVID without bailouts. The realistic adverse case is a 20–30% drawdown, not impairment.

Chance of total loss? Negligible absent a systemic Canadian banking collapse — not a base-case scenario for a fortress-capitalized D-SIB.

Recent News & Events

Has the business environment changed recently? Yes — a deteriorating macro (CUSMA/tariff uncertainty, higher Canadian unemployment, elevated energy costs) drove management to raise its FY26 impaired-PCL guidance to mid-50s bp. Offsetting positives: the ROE inflection (13.2%), wealth-flow momentum, Mexico strength, and the dividend hike/NCIB.

Significant acquisitions / accounting changes / new markets? Davivienda close (Dec-2025); Jamaica buy-in proposal (Jun-2026); KeyCorp stake (with Fed approval to increase, Mar-2026); AI platform launches (Scotia Intelligence/Navigator); GBM-Asia runoff. No accounting-policy red flags beyond the adjusted-vs-reported framing.

APPENDIX B — Source Appendix

The Bank of Nova Scotia (NYSE/TSX: BNS). Report date 2026-06-27. Primary sources before secondary; access dates 2026-06-27/28. Fact / Interpretation / Assumption distinctions are carried in the memo body.

Primary — Company filings, disclosures & events

• Scotiabank FY2025 Annual Report / Q4-FY25 results (fiscal year ended 2025-10-31; released 2025-12-02) — segment earnings, NIM, PCL, CET1, BVPS, EPS, dividend. https://www.scotiabank.com/ca/en/about/investors-shareholders.html
• Scotiabank Q2-FY26 report & earnings call (quarter ended 2026-04-30; call 2026-05-27) — primary source for ROE 13.2%, PPPT +16%, segment detail, NIM 4.76% (Intl), PCL 66bp / mid-50s guide, CET1 13.3%, buyback, dividend +C$0.04, “14%+ ROE in FY2027” guidance. Speakers: Scott Thomson (CEO), Raj Viswanathan (CFO), Shannon McGinnis (CRO), segment heads. [Transcript via ROIC.ai]
• Scotiabank Q1-FY26, Q3/Q4-FY25, Q1-FY25 quarterly reports/releases — quarterly EPS/NI bridge, the ~C$1.36B Q1-FY25 Colombia/Davivienda charge.
• Scotiabank 2023 Investor Day (December 2023) — medium-term targets (≥7% EPS CAGR, ~14% ROE, ~53% productivity), North American corridor strategy.
• Scotiabank / Davivienda transaction releases — Colombia/Costa Rica/Panama transfer for ~20% of combined entity (announced Jan-2025; regulatory approval Nov-2025; closed Dec-1-2025).
• KeyCorp investment disclosures — 4.9% stake (Aug-2024), increase to 14.9% (~US$2.8B, ~163M shares at ~US$17.17, Dec-2024); Fed approval to potentially increase (Mar-2026).
• Scotia Group Jamaica minority buy-in proposal (~C$0.5B cash), 2026-06-12.
• Scotiabank corporate profile / segment descriptions / subsidiary list — https://www.scotiabank.com

Quantitative data services

• ROIC.ai (mcp) — income statement, balance sheet, profitability/credit/per-share ratios, valuation multiples, enterprise value, and the Q2-FY26 earnings-call transcript. Third-party aggregated; reconciled to filings. (Note: bank EV/EBITDA, net-debt/EBITDA, interest-coverage outputs are not meaningful and were disregarded; certain bank book-equity fields cross-checked against the filing.)
• AZI — daily price/OHLCV history (TSX, C$), the valuation-index own-history percentile ranks (P/E 95.7th, P/B 91.1th, P/S 26.5th, composite 71.1st as of 2026-06-26), and the news feed (5 items; the 2026-06-12 Jamaica proposal).
• FactorsToday — factor loadings (Value +0.32, DividendYield +0.18, Country:Canada 0.48, Market 0.54; Banks-industry 0.084; no Momentum), leaderboard (y1 +64.2%, Sharpe 3.68, maxDD −13.4%; lifetime +8.8%/yr, Sharpe 0.27, maxDD −63.8%), specific volatility (11.9% annual), related stocks (Canada ETFs + MFC).

Peer & sector reference (public)

• Credicorp Ltd. (BAP) and Intercorp Financial Services (IFS) — Peruvian banks; the dominant-local-oligopolist contrast for the competitive section (public filings/investor materials).
• Canadian Big-Six peers RY, TD, CM, BMO, National Bank — public filings and disclosures for oligopoly structure, OSFI/mortgage-renewal framing, and cross-sectional valuation comparison.

Secondary — market & consensus

• MarketScreener — BNS analyst consensus & target range. https://ca.marketscreener.com/quote/stock/THE-BANK-OF-NOVA-SCOTIA-1409234/
• Yahoo Finance — BNS analyst estimates / price-target distribution. https://finance.yahoo.com/quote/BNS/analysis/
• General financial press (Reuters / Bloomberg / Globe & Mail) for Big-Six sector context, the 2023 sector trough, and the 2025–26 re-rating.

Note on currency & figures: results are reported in C$ under IFRS; the NYSE-listed common trades in US$ (USDCAD ~1.37 at report date). Where book-value-per-share figures differ across sources (ROIC ~C$65, AZI ~C$71, filing-based common equity ~C$62), the spread is noted in the Valuation section; percentile rankings use AZI’s own multi-year series. Insider transaction data is unavailable (foreign private issuer; Canadian insiders file via SEDI, not SEC Form 4).