Retail Repairs the Margin, but Credit and Costs Leave the Premium Exposed

Results vs. Consensus

CIBC reported before the open on August 27 for the quarter ended July 31, 2026, and held its call at 7:30 a.m. ET. Financial figures and valuation are in Canadian dollars unless labelled US$. Returns use the Toronto listing; adjusted measures exclude the bank’s items of note.

Year-over-Year Comparison

Quarter-over-Quarter Comparison

Revenue. The C$1,114M year-over-year increase was spread across businesses, with Capital Markets contributing C$328M, or 29%. Non-trading net interest income rose 14% and average loans grew 6%, while market-related fees and trading remained favorable. The C$303M sequential increase in non-trading interest income matters more to our forecast than the headline beat: it raises the starting point for Q4 even after allowing for three additional days in Q3. Our May interest-income view was too cautious.

Margins. Adjusted pre-provision earnings rose 20.5% year over year, but only 3.9% sequentially as expenses increased 5.2%. The efficiency ratio deteriorated 40 basis points despite higher revenue. CIBC still has positive operating leverage, now for twelve consecutive quarters, but the marginal expense dollar is becoming less flexible as investment and base compensation replace performance pay as the main sources of growth.

EPS. Adjusted net income grew 25.9%, and adjusted earnings attributable to common shareholders grew 24.3%; the 2.0% reduction in diluted shares lifted per-share growth to 26.4%. The adjusted tax rate fell to 22.1% from 22.9% a year ago and 23.0% in Q2. Applying the Q2 rate to Q3 adjusted pretax income would reduce EPS by about C$0.03. Holding performing provisions at Q2’s C$57M, instead of this quarter’s C$48M release, would reduce EPS by roughly another C$0.09 after tax. These are sensitivities, not adjustments to the bank’s reported result, and consensus already allowed for some credit expense.

Reported versus adjusted. Reported EPS was C$2.47 and net income C$2,409M. The C$239M after-tax difference comprises a C$232M Caribbean disposal charge and C$7M of acquisition-intangible amortization. Adjusted revenue is C$8,370M, C$2M above reported revenue; adjusted expenses are C$4,408M versus reported C$4,685M. The disposal charge belongs outside ongoing operating earnings, but it is a real reduction in shareholders’ capital and remains inside book value.

Segment Performance

The U.S. segment’s reported revenue is C$863M and adjusted revenue C$865M. Its local-currency results below separate operating growth from currency translation. The existing segment structure remains in place through Q3, with restated reporting due in Q4.

Canadian Personal & Business Banking: the recovery we needed to see

Retail adjusted earnings rose 17% year over year to C$953M and 12% sequentially. Margin increased four basis points to 3.16%, adjusted efficiency improved to 49.0%, and ROE recovered to 28.2% from 26.2%. These directly challenge May’s concern that the margin advantage was losing its operating foundation. Demand-deposit spot balances grew 5% and card balances 6%, even though total average deposits fell 1% and loans grew only 2%. Lower-cost deposits and more remunerative lending can improve earnings without much balance-sheet expansion.

The credit benefit needs qualification. Retail impaired provisions barely changed, at C$394M versus C$397M, while the performing charge fell to C$33M from C$77M. The C$47M improvement in total provisions was therefore predominantly a smaller reserve build. Adjusted operating leverage also narrowed to 0.8% from 4.8%.

Assessment: Retail has earned a higher interest-income forecast, and its sub-50% efficiency test remains satisfied. It has not yet restored the broader structural-return pillar: expenses are catching up with revenue, and most of the quarter’s credit improvement did not come from lower impaired losses.

Canadian Commercial Banking & Wealth Management: growth is being consumed

Commercial revenue rose 11% to C$752M, supported by 7% loan growth and 8% deposit growth, meeting the mid-to-high-single-digit volume expectation carried from May. Wealth revenue rose 23% to C$1,285M. Canadian wealth assets under administration reached C$670B, with C$84B of the C$118B annual increase from markets and C$34B from net sales. The franchise is attracting money as well as benefiting from asset prices.

Yet the combined segment produced only 4% earnings growth to C$619M. Expenses rose 18%, essentially matching revenue growth, and impaired provisions increased to C$132M from C$97M in Q2 and C$25M a year ago. ROE slipped to 23.1% from 23.7%, and the efficiency ratio reached 50.9%.

Assessment: Volumes fulfilled management’s commitment; credit did not begin the promised moderation. This is the clearest reason that strong fee growth alone cannot support a higher group valuation. More of the revenue must reach shareholders before the structural-return pillar returns to ON TRACK.

U.S. Commercial Banking & Wealth Management: volume offsets spread pressure

In U.S. dollars, adjusted revenue rose 7% to US$619M and pre-provision earnings rose 10% to US$277M. Adjusted net income rose 22% to US$229M, helped by a US$23M credit-provision reversal. In Canadian dollars the reversal was C$32M, comprising a C$69M performing release against C$37M impaired provisions. The release was substantially related to a CRE loan disposal, which also reduced exposure; it is not a quarterly revenue stream.

Margin fell another 14 basis points to 3.76%, following Q2’s eleven-basis-point decline. Strong loan growth nevertheless lifted net interest income 9% year over year, and ROE improved to 10.8% from 9.3%. Management expects seasonal deposit inflows to help, while some loan repricing reflects borrowers moving to better credit grades.

Assessment: The improved credit and volume picture prevents us from treating every lost basis point as franchise deterioration. But the earnings increase materially outpaces pre-provision growth. We expect some seasonal margin recovery, while removing the CRE reserve-release benefit from the forward run rate.

Capital Markets: normalization has begun

Revenue of C$1,834M was 22% above last year and 2% below Q2. Global Markets contributed C$1,118M, down 5% sequentially; Corporate and Investment Banking contributed C$716M, up 3%. Expenses increased 19% year over year, while impaired provisions rose to C$53M from C$10M in Q2. Net income fell 9% sequentially to C$722M, inside the C$680M–C$780M range in our May forecast, and ROE fell to 23.0%.

Capital Markets supplied C$182M, or 33.5%, of the C$544M year-over-year increase in group adjusted earnings. That is materially below the approximately one-half contribution in each of the first two quarters. Retail continues to earn more and again generates a higher return. The bank is less dependent on the exceptional wholesale contribution than it was in May, even though lending growth of 17% remains rapid.

Assessment: This fulfills part of the normalization case rather than constituting a new surprise. We reduce the concentration risk’s status from MATERIALIZING to EMERGING. The remaining concern is that loans originated during the expansion have yet to season fully, while compensation and technology spending already sit in the cost base.

Corporate & Other: improved disclosure, volatile contribution

Adjusted net income was C$32M versus a C$47M loss in Q2 and a C$107M loss a year ago. Revenue rose C$65M sequentially to C$290M, including higher treasury revenue; adjusted pretax income remained a C$107M loss, offset by a C$139M tax recovery. The reported net loss of C$200M includes the Caribbean charge.

Assessment: The quarter’s C$79M sequential earnings improvement accounts for 45% of the group’s C$177M adjusted net-income increase. The filing explains more of the movement than the May call did, but positive net income resting on tax allocation is a weak basis for extrapolation. Our forward tax and earnings assumptions return toward normal rather than capitalizing C$32M as a new quarterly floor.

Key KPIs: Credit and Capital

Consumer stress is no longer worsening uniformly: cards and personal-loan delinquencies improved, but mortgage arrears and total consumer delinquencies increased. Mortgage write-offs remain below 0.01%, supported by collateral, so an arrears increase is not equivalent to an unsecured-loss increase. Meanwhile, gross impaired dollars rose even though the ratio fell one basis point as the loan book expanded. The benign ratio does not negate the commercial loss experience.

Key Topics & Management Commentary

Overall Management Tone: Management remained confident in the strategy, but was more explicit about how retail deposit mix supports earnings and less willing to promise an immediate credit turn. Compared with May, the central uncertainty shifted from whether Canadian pricing could improve to whether the resulting revenue can outgrow a more persistent cost and credit burden.

1. Pricing returns; the margin repair is incomplete

At the total bank level, we reiterate our expectation of a stable to gradual positive bias on our net interest margin over time.
— Robert Sedran, Chief Financial Officer

The Canadian P&C bridge combines personal and commercial banking and excludes the wealth business; its 3.04% margin is distinct from the retail segment’s 3.16%. Pricing/other is now positive in Canada and across the bank, reversing May’s unfavorable contribution. Product mix, however, is still negative at group level, and U.S. pricing remains a drag. The hedging contribution continues at one basis point; management did not provide a revised crossover date, so the 2027 expiry risk has not disappeared.

Assessment: Our margin range rises to 2.06%–2.10% from 2.04%–2.08%. This credits the demonstrated Canadian repair while retaining the gradual outlook. Restoring one contributor is enough to lift earnings estimates, but not enough to assume the full franchise can sustain the ROE embedded in the share price.

2. A lower provision total does not establish a credit turn

At the bank level, impaired losses are 37 basis points year-to-date as the losses remain consistent with the elevated stress in the macroeconomic cycle. We expect impaired losses in and around this range for the remainder of the year.
— Frank Guse, Chief Risk Officer

Total provisions fell C$41M sequentially because a C$105M improvement in the performing line exceeded the C$64M increase in impaired provisions. Canadian commercial losses of C$132M failed the first Q3 test of May’s promised second-half moderation. Q3 group impaired provisions of C$612M also exceeded the C$534M first-half quarterly average used in our prior plateau assumption. Q4 would have to fall to about C$456M for the second-half dollar average to equal the first half; the new 37-basis-point outlook implies roughly C$570M at the current loan base and a normal-length quarter.

There are counterweights: card write-offs improved, the U.S. impaired charge declined, and the most tariff-sensitive businesses represent less than 1% of group loans. But those limits on broad stress do not reverse the earnings consequence of losses already above the earlier base case. Management attributed the commercial increase to a small number of unrelated events, while acknowledging that the original base case had not anticipated the prolonged trade conflict and Middle East shock.

Assessment: We now expect C$600M of Q4 total provisions, roughly C$570M impaired plus a C$30M performing build. At the current loan base this is about 39 basis points, below May’s 42–46-basis-point total-credit framework. The smaller performing build is a positive forecast change that more than offsets the higher impaired-credit trajectory in the total provision estimate. Our credit risk remains MATERIALIZING because impaired losses are above the earlier path, while the more favorable reserve outlook helps absorb their earnings cost.

3. Wholesale growth is more than trading, and brings longer-lived risk

We look at high single growth of our earnings through the cycle. And when I look at Q4, I would tell you that year-over-year, we do expect some growth, but some moderation quarter-over-quarter given the uncertainties in the macro side.
— Christian Exshaw, Senior Executive Vice-President and Group Head, Capital Markets

The digital-infrastructure build-out connects financing to swaps, hedging, bond issuance and energy trading. That relationship makes a portion of wholesale revenue more durable than a favorable quarter of trading. Deposits rose 32% year over year to C$137B, faster than loans, and management described fee businesses in payments and depository services that can perform across environments. It has also hired 250 people year to date, about half in the U.S., creating a cost commitment that will remain if markets become less constructive.

The earlier H2 revenue bracket was above C$3,029M and below C$3,885M. Q3 revenue of C$1,834M leaves a wide C$1,195M–C$2,051M mathematical range for Q4 under that bracket. The more useful current signal is sequential moderation. We assume C$1,750M of Q4 Capital Markets revenue, which would put H2 at C$3,584M. Average loan growth has slowed from 20% to 17%, but still has distance to travel to the high-single-digit year-end ambition discussed in February.

Assessment: The franchise has improved, so a wholesale collapse is not our base case. Nevertheless, financing growth ties today’s fee opportunity to future credit exposure. We retain normalized wholesale earnings rather than applying a peak quarter’s return to the entire capital allocation.

4. AI is a capacity strategy, with a rising cost base

Second, we launched CIBC AdvisorAssist, an AI-enabled platform that helps advisers spend more time with clients. The system automates meeting notes, summaries and follow-up documentation while supporting regulatory compliance, and this reduces administrative time for advisers by up to 50%.
— Harry Culham, President and Chief Executive Officer

The deployment evidence is specific: more than 4,000 developers have AI capabilities with a stated 20% productivity benefit, and DocuMind processes 63,000 documents per month. Those are operational outcomes, although reduced task time is not the same as a reduction in total employee expense. Management explicitly expects headcount to grow over the next five years. The investment case is more business per employee and better advice, rather than a shrinking workforce.

The expense bridge now adds C$107M of investment and C$242M of operating cost, offset by only C$50M of efficiencies, before C$144M of performance compensation. Thus only about one-third of the C$443M year-over-year adjusted expense increase is performance-linked, against more than half in May. Core growth is 8%, up from 4%, and Q4 adjusted expenses are expected to rise sequentially.

Assessment: We move the technology pillar to AT RISK on financial conversion, while recognizing the useful deployments. The forecast increases expenses to C$4.55B in Q4 and assumes no separate AI savings windfall. The result that would restore confidence is revenue growth consistently outpacing this enlarged cost base, with disclosed productivity gains supporting it.

5. The buyback continues at the higher share price

This quarter, we bought back 7.5 million shares.
— Robert Sedran, Chief Financial Officer

CIBC spent C$1.199B: C$307M completing the old bid and C$892M under the new one. Purchases under the new authorization averaged C$162.18, above the post-results share price. That contradicts the inference in our May note that a lower average repurchase price revealed a valuation ceiling. The bank’s stated framework, with repurchases an output of organic capital needs, is a better explanation of its behavior.

CET1 fell 19 basis points on the bank’s unrounded bridge to 13.4%. Earnings net of dividends added 43 basis points, while RWA growth consumed 21, repurchases 33 and other items eight. The balance sheet remains capable of funding distributions, but the capital ratio is not rising automatically while both lending and buybacks expand.

Assessment: The capital-return pillar remains ON TRACK. Repurchases help EPS, yet buying above book value reduces the book-value growth available for our valuation. We reflect that cost explicitly rather than adding the full earnings retention rate and buyback yield together.

6. The transaction charge is smaller; capital benefits remain ahead

For the third quarter of 2026, we reported earnings per share of $2.47, which included the previously announced, albeit smaller, charge related to our Caribbean operations of $232 million after tax that was treated as an item of note in our Corporate and Other business unit.
— Robert Sedran, Chief Financial Officer

The Caribbean charge was C$269M pretax, below the roughly C$350M charge anticipated in May. The sale of the 91.67% interest to Butterfield still awaits shareholder and regulatory approval, with closing expected in the first half of 2027. The C$232M after-tax cost is recognized now; the previously indicated capital benefit and retained Butterfield earnings interest arrive at closing. A smaller charge protects book value relative to the prior expectation, but is not recurring growth.

The &Partners transaction closed July 6. The filing now discloses a 19.15% economic interest, 4.99% voting interest, a board seat and contingent consideration payable in 2029 if performance targets are achieved. Equity-accounted income sits in U.S. Commercial Banking & Wealth Management, resolving the location that our May discussion had placed in Corporate & Other. The upfront purchase amount remains undisclosed.

Assessment: The Caribbean exit still simplifies the franchise and the wealth investment adds distribution. We keep the disposal’s previously indicated modest EPS dilution in the medium-term range and do not book the expected closing capital benefit early. Neither transaction by itself closes the gap between current returns and the valuation’s required returns.

7. Investor Day gives the return debate a date

Against that backdrop, I'm pleased to announce that we will host our next Investor Day on December 9.
— Harry Culham, President and Chief Executive Officer

This is progress on the most consequential unresolved commitment in May: a venue now exists to explain the next phase of profitability. It is not yet a refreshed medium-term ROE target. Management deferred more detailed retail growth guidance to Q4 and Investor Day, and the new four-business reporting structure will arrive with Q4 restated comparatives.

Assessment: December 9 becomes a dated catalyst rather than an indefinite request for disclosure. A credible return target, supported by expense and credit assumptions, could justify a premium and challenge our rating. Announcing the event alone does not meet the earlier 17%-plus return condition, much less the approximately 18.9% implied by today’s price.

Guidance & Outlook

Implied earnings path. Nine-month adjusted EPS of C$8.03 already exceeds our old C$7.75 trajectory through Q3. We now expect C$2.55 in Q4, taking FY26 to C$10.58 versus about C$10.20 previously. That is an estimate increase, not a guidance raise: CIBC provided no consolidated numerical revenue or EPS range. The Q4 estimate is below Q3’s C$2.73 despite stronger non-trading interest income, because trading moderates, expenses rise and the performing-reserve release does not recur.

Guidance style. The bank’s growth ambitions remain confident, but credit has run above the original base case and the expense outlook is directional. We give more weight to those observed misses and the specific cost bridge than to an assumption that another beat will automatically sustain the multiple.

Analyst Q&A Highlights

What lower deposits mean for the retail franchise

The question challenged whether flat-to-declining Canadian deposits undermine further margin growth. The response distinguished everyday balances from maturing term deposits and linked retained client assets to wealth fees.

Q: I guess, looking at Canada and the U.S., I see Canada, the loan growth is still positive and deposit growth has been kind of flat to down a bit for the last 4, 5 quarters. And just wondering if you expect a turnaround there? Or does deposit growth dynamic, is that going to remain unfavorable and maybe present more of a challenge going forward?
— Gabriel Dechaine, National Bank Financial

A: We've actually been leading the tables in terms of our growth in mutual funds. And a lot of those GIC funds have actually been going by client preference and because of the advice we give them into the investment side with managed money. And so overall, the vast majority of the GIC outflows, we do see us retaining those funds.
— Hratch Panossian, Group Head, Personal and Business Banking

Assessment: This is a substantive answer to May’s deposit concern: lower reported deposits can coexist with retained client economics and better funding mix. It supports our higher fee and interest-income base. The trade-off is greater exposure to asset prices in wealth, so the earnings benefit is not identical to stable deposit growth.

Whether wholesale growth has an earnings floor

The question tested how much of Capital Markets growth is structural and whether its earnings can expand without further capital. The response described diversification and reinvestment but did not set a quantified floor.

Q: On the capital market side, really nice growth there. Just how much of that improvement that you've been seeing is kind of structural versus cyclical? And is there another leg of growth that you can achieve without materially increasing the capital allocated to that business? And then kind of on a broader question, like what do you see as the floor of that business, just given the cyclicality that we know exists there?
— Matthew Lee, Canaccord Genuity

A: We've had, I would say, 3 quarters of exceptional constructive markets. And the business that we've built at CIBC is very well diversified. It reads, I would say, the results of that client-led strategy that we've had for nearly, I would say, 20 years. So in terms of answering the question, everything really when you look at our financials is centered around our clients and very difficult for us, obviously, to speculate on what the environment will be. What I can tell you, though, is that whenever the markets are constructive, we do capture a lot of that upside. Look at the PPPT growth over the last 7 quarters. We've invested heavily in people. So year-to-date, for instance, we've hired 250 people. Roughly half of that is in the U.S., 40% of that is in Canada. We continue building product suites that fit our clients, for instance, in the AI space, whether it's around project finance, infrastructure and energy. We continue to invest in the overall infrastructure of our business and technology. That's one of our big differentiators, if you think of our ASG business, if you think about our Canadian depository business, both businesses are fee-based businesses and very much, I would say, foundational to what it is we do and kind of like neutral to what the environment would be.
— Christian Exshaw, Group Head, Capital Markets

Assessment: A diversified client business plausibly earns more through the cycle than the older franchise, but the answer does not establish a minimum earnings level or a ceiling on capital. We credit the franchise by avoiding a collapse scenario in the base forecast while retaining a downside case for market normalization.

How much U.S. margin compression signals worse economics

The challenge focused on expensive deposits and a loan-to-deposit ratio near 100%. Management split the move between seasonal balance mix and pricing, then explained that some pricing declines reflect improving borrowers.

Q: And then in the U.S., the double-digit compression we saw this quarter, your flow chart, which is very helpful. It highlights pricing as the issue. Is that just the deposits are getting more expensive because you're running more of a 100% loans-to-deposits ratio there? And I guess if you can give some outlook commentary for the U.S. as well, that would be great.
— Gabriel Dechaine, National Bank Financial

A: But just to go into a little bit more detail about the 14 basis point decline, kind of building on what Rob said, about half of that is mix driven. About half of it is pricing compression. On the mix, it's very strong loan growth. It's $1.5 billion or about 3.5% sequentially, loans outpacing deposit growth, which was about flat. And we do expect to see a seasonal reversion there in the coming quarters as we saw last year. On pricing, I guess, I'd say two things. As you've seen from a lot of our peers in the U.S. and in Canada, yes, it's competitive for high-quality borrowers. That's for sure. But I think it's really important to note, about half of the difference in pricing actually relates to client situations where credit profile is improving. So loans are getting repriced downward based on moving down the price grid, which implies improving credit quality.
— Kevin Li, Chief Executive Officer, U.S. Region

Assessment: Better-rated borrowers can justify lower spreads and reduce expected losses, so the NIM decline overstates the deterioration in risk-adjusted economics. That is a useful counterargument. It still leaves competitive pricing pressure and seasonal funding to recover; our forecast assumes stabilization, not an immediate return to Q1’s 4.01% margin.

AI productivity does not imply fewer employees

The question asked management to choose between headcount reductions and revenue growth with a stable workforce. The answer set a different direction: more people, supported by more productive technology.

Q: Do you view AI as an opportunity to take CIBC's headcount lower over the next, say, 5 years? Or is it more a scenario where the revenue growth can remain strong while you keep headcount relatively stable? Which of those two is most likely?
— Mario Mendonca, TD Cowen

A: But what I would say, to answer your question, we actually see headcount growth over the next 5 years in the organization and when we look forward, but we see significant productivity increase to take our results to the next level and drive outperformance. So it's a tale of 2 things. We are going to invest in people and we're going to invest heavily in technology.
— Harry Culham, President and Chief Executive Officer

Assessment: The strategic answer is clear. Investors should expect earnings leverage through capacity and revenue, with a growing expense base. That can create value, but it also makes a market-driven slowdown harder to absorb than the May performance-pay-heavy cost bridge suggested.

When rapid wholesale loan growth becomes a credit test

The question linked two years of wholesale expansion to later loan seasoning. The risk response accepted the approximate timetable while distinguishing larger dollar losses from worse loss rates.

Q: it's the exceptional growth in the wholesale lending platform, $23 billion in new wholesale loans in the last 2 years. And as I look through the supplement, I can see where it's coming from. It looks like it's leveraged to the U.S. It's in financial institutions, business services, commercial real estate. And again, it seems like it's U.S. So the first question is, to what extent is this wholesale lending growth really in support of the broader capital markets business? By that, I mean growing underwriting and advisory -- growing the underwriting and advisory business, that's the first part. Is this really necessary to drive the other source of revenue growth? And then for Frank, the rule of thumb I've used in the past for when I see growth at this level is to sort of look forward 2 years and say, in 2 years, they're going to pay some kind of price. Now it doesn't have to be a bad price, but they're going to pay some kind of price on the credit front. Is it your view -- is that rule of thumb, 2 years, is an appropriate rule of thumb to use?
— Mario Mendonca, TD Cowen

A: You asked a little bit about the rule of thumb. I think the rule of thumb of about 2 years as it relates to maturing of some of those originations is right. I would say it's probably a tiny bit longer on the business and government side, maybe a little bit shorter on the cards portfolios, but in and around that 2-year range. But we are working very closely with the businesses. We do understand those businesses that we are originating in very, very well. And we are not compromising from a risk quality perspective. So what that doesn't give me is any concerns that we will see a deterioration in the rates, yes. But you're right. On a larger portfolio, we may see larger loan losses coming up. But I expect the rates essentially to be very consistent with past experiences. That having said, those portfolios can be a little bit lumpy. But while you will see it in 2 or 3 years from now, we watch very, very closely ongoing migrations and so on, and we have no concerns with those underlying books as to what we are seeing so far.
— Frank Guse, Chief Risk Officer

Assessment: This is an acknowledgment of seasoning, not a forecast of deteriorating underwriting. It means the larger loan book can consume more earnings through absolute provisions even at unchanged rates. The commercial and Capital Markets loss events therefore warrant monitoring alongside migration data as the new vintages mature.

Why the reserve release is not an economic all-clear

The follow-up asked directly whether the quarter’s release marked the pivot to sustained releases. Management explained where it arose and emphasized that Canadian portfolios were still building reserves.

Q: And then second part is, are we in a period where maybe impaired PCLs continue to go up or deteriorate a little bit, but you are in that range or you're in that pivot point where you start to release PCL as the economic outlook maybe slightly improves. You talked about unemployment improving. I'm just wondering if we're in that period where the market and we should start to expect steady releases on the performing loan side?
— Doug Young, Desjardins Capital Markets

A: Actually, if you distill that a little bit, our release this quarter is actually a build in our Canadian businesses. It is a build driven by a lot of that uncertainty that we still continue to see. And the releases that we saw in our U.S. business and in our capital markets business is actually largely driven by the sale of a CRE portfolio that I talked about and some portfolio migration that we saw in those businesses. So again, not driven by the economic outlook, much more driven by those 2 specific events.
— Frank Guse, Chief Risk Officer

Assessment: The direct answer rejects extrapolating the release. The financial statements also identify favorable economic assumptions as one component, but the call makes clear that disposal and migration dominated management’s explanation. We therefore retain a small performing build in Q4 and treat sustained releases as upside contingent on clearer forward indicators.

What They’re NOT Saying

1. A refreshed sustainable ROE target. December 9 supplies a venue, while the number, time horizon and capital assumptions remain open. Those details will determine whether a valuation requiring almost 19% sustainable returns is credible.
2. A measured financial return from AI. Adoption, task-time savings and development productivity are tangible; enterprise-level net savings after implementation cost are not disclosed. The 8% core expense increase makes conversion into profit the relevant missing link.
3. The size and economics of the U.S. CRE disposal. The reserve release is visible, but the call did not quantify the sold balances, sale price or ongoing income forgone. We cannot treat the release as a recurring benefit without that bridge.
4. Enough detail to underwrite wholesale credit seasoning. The diversified private-credit exposure is about C$17B and under 3% of loans, but concentration protections do not reveal the future loss profile of the wider digital-infrastructure lending expansion. Vintage and migration evidence will matter more than another quarter of origination growth.
5. The restated segment economics and &Partners purchase dollars. The reporting change remains due in Q4. The wealth stake and accounting location are now known, but the cash price and contingent consideration amount remain absent, leaving its return on invested capital unresolved.

Market Reaction

• Pre-print setup: C$163.82 at the August 26 Toronto close, up 31.7% year to date and 57.8% over twelve months, but down 3.1% over the prior 30 days. The pre-print 52-week closing range was C$100.68–C$171.76.
• August 27 reaction session: opened at C$163.82, traded C$155.80–C$163.85, and closed at C$159.16, down C$4.66 or 2.8%.
• Participation and benchmark: 3.6M shares traded versus a 2.2M 30-day average, or 1.7 times normal volume. The S&P 500 rose 0.7% that session and was up 12.1% year to date entering the print.

The shares recovered part of their intraday decline, but a larger earnings beat than May’s still failed to produce a positive close. The setup was less stretched than immediately before Q2: CIBC had already retreated from its July highs. Nonetheless, its substantial twelve-month advance left the stock exposed to doubts about the durability of the higher earnings base.

Our reading is that improved retail results were insufficient to settle the credit and expense questions at a premium valuation. Contemporaneous sector discussion likewise focused on whether strong markets-facing profits justified rich bank multiples. Price action cannot isolate each disclosure’s contribution, and the sell-off is not proof of our valuation; the case rests on the earnings and book-value bridge below.

Street Perspective

Debate: Does retail mix justify a structural premium?

Bull view: The optimistic case is that everyday deposits, credit cards and advised wealth can generate superior earnings without rapid mortgage growth. The Canadian pricing recovery and 28.2% retail ROE give this argument more substance than it had in May.

Bear view: The skeptical case is that asset-price-sensitive wealth fees and a hedging tailwind are still helping the result, while expenses and commercial losses prevent the same return improvement across the bank.

Our take: Retail improvement deserves credit in earnings. It is insufficient evidence for a near-19% group return, because the commercial/wealth return fell and the U.S. improvement includes reserve releases.

Debate: Is credit reaching a turning point?

Bull view: Falling card write-offs, better U.S. impaired losses and improving unemployment assumptions offer a plausible path to lower provisions. Strong allowance coverage gives management room to absorb individual events.

Bear view: Canadian commercial losses rose again, group impaired provisions exceeded the earlier plateau, and the performing release was heavily transaction-related. The growth in wholesale loans adds future seasoning exposure.

Our take: Consumer stabilization is real but partial. We favor a cautious dollar provision forecast over a sustained-release thesis until Canadian commercial losses moderate and forward conditions justify releasing reserves.

Debate: Can AI and Investor Day unlock another valuation increase?

Bull view: The optimistic view treats agentic tools and adviser productivity as operating leverage still to come. A new return framework in December could reveal earnings capacity that the current group ROE understates.

Bear view: The cautious view points to 8% core expense growth, expected hiring and a stock already priced for a much higher sustainable return. A catalyst can reveal a shortfall as readily as close one.

Our take: The event has option value, reflected in our bull case. The base forecast should earn the return through revenue and costs already supportable today; it should not assume the target before management publishes it.

Our Estimates & Valuation

We raise near-term earnings because Canadian interest income is stronger than we assumed in May and our total-credit forecast is lower. The increase is smaller than annualizing C$2.73 would imply. Our Q4 estimate absorbs higher expenses, normalizes trading and the tax rate, and assumes a small performing build after Q3’s net release. These are our estimates; CIBC has not supplied consolidated numerical guidance.

The earnings bridge

FY27’s C$20.60B non-trading interest income assumes roughly 5% earning-asset growth from the Q3 base and margin near 2.09%; C$11.10B of non-trading fees assumes continued wealth flows and client activity. Trading revenue of C$2.80B is below four times Q3’s C$794M. Expenses grow with hiring and technology, with modest efficiency gains but no separately credited AI windfall. The provision assumption keeps dollar losses elevated as the loan book grows; the Caribbean transaction’s small dilution can be absorbed within this rounded earnings framework.

The FY26 revision is C$0.38: C$0.28 from Q3 versus the prior C$2.45 midpoint and C$0.10 from Q4. The Q4 increase includes a benefit from lower total provisions. On the same C$608.6B loan base and a 92-day quarter, May’s 42–46-basis-point framework corresponds to roughly C$644M–C$706M, versus C$600M now. The reduction from that band’s midpoint is worth about C$0.06 of EPS at our current tax and share assumptions. This is a comparison of credit-rate assumptions, since the prior recap did not publish a complete Q4 line-item bridge. Stronger interest income contributes alongside that credit benefit, with higher expenses offsetting part of the improvement. FY27 EPS of C$11.11 is 5.0% above the revised FY26 estimate. Each additional C$100M of annual provisions or expenses reduces EPS by about C$0.086 at a 23% tax rate and 900M shares. This sensitivity is why a modest revenue upgrade does not eliminate the risks in the cost and credit assumptions.

Book value, sustainable returns and the 12-month price

We retain the prior Gordon-growth price-to-book method: justified P/B = (sustainable ROE − terminal growth) / (cost of equity − terminal growth). Terminal growth remains 3%, with a 9%–10% cost-of-equity sensitivity. The base case uses 16.5% sustainable ROE and a 9.5% cost of equity, both inside the prior published ranges. We use the middle discount rate rather than leaning on the most favorable 9% case after another credit disappointment and a less flexible expense increase.

At C$159.16 and book value of C$65.14, the shares trade at 2.44 times book and 15.5 times trailing adjusted EPS of C$10.24. Inverting the central framework gives 18.9% sustainable ROE, compared with the 16.8% delivered this quarter. A medium-term target of at least 17% would improve confidence, but even 17% at a 9.5% cost of equity supports only 2.15 times book. The price needs more than merely fulfilling May’s target-refresh condition.

For an explicit twelve-month horizon we forecast book value near C$69. Starting from C$59.146B of common equity, we assume C$10.0B of common earnings, C$3.84B of dividends, and repurchases of 29M shares at an assumed average C$159, costing C$4.61B. With no OCI contribution, equity reaches about C$60.70B and shares fall to about 879M. This produces approximately C$69 per share, including the book-value cost of repurchases above book.

What changed from May. The prior C$140 fair-value reference was a midpoint of four cells using current book value, ROE of 16.4%–17% and cost of equity of 9%–9.5%; it was not a single-point target multiple. Our central 16.5%/9.5% assumptions value today’s C$65.14 book at C$135.29. Rolling that same multiple onto C$69 forward book raises the twelve-month value to C$143.31. Thus the modest increase over the old reference includes future retained earnings, while the current price has moved further above what we can support.

Why Underperform. In May, Hold reflected a wide bracket between multiple normalization and a continuing premium. We now give normalization more weight: impaired credit has again missed the earlier trajectory, underlying costs have accelerated, and the price requires a higher sustainable return despite those constraints. The lower total-provision forecast is a meaningful counterweight and supports our EPS increase; the downgrade depends on recurring impaired-loss pressure and costs preventing the much higher sustainable return embedded in the price, rather than on assuming every credit measure worsens. The base return is negative, below our explicit 6%–8% twelve-month S&P 500 total-return assumption. The relative judgment assumes unchanged CAD/USD. Maintaining today’s 2.44-times-book multiple on C$69 forward book would instead produce roughly 8.6% total return, explaining why a strong Investor Day could invalidate our call. The bull case requires both improved returns and a favorable discount rate; the bear case captures weaker earnings retention and multiple compression.

Thesis Scorecard Post-Earnings

The seven pillars below carry forward the standing May thesis. The quarter improves the Canadian revenue case and reduces the immediate concentration in Capital Markets, while worsening the financial conversion of technology investment and the credit trajectory.

Overall: the operating thesis is mixed rather than uniformly weaker. Retail delivered a meaningful repair and wholesale concentration receded. Those positives and a lower total-provision forecast raise our earnings estimate, but commercial impaired credit and core costs reduce confidence that the current valuation can be sustained. The rating changes because the expected return has become less attractive, not because a profitable bank suddenly lacks a viable strategy.

Action: Underperform, conviction 7/10. C$143.31 of base twelve-month value plus C$4.28 in dividends offers insufficient return from C$159.16. The next tests are commercial impaired losses moderating, the 37-basis-point Q4 credit outlook holding, expenses remaining supportable by recurring revenue, and December 9 providing a credible return framework. A price near the prior 2.0–2.1-times-book zone, now roughly C$130–C$137 on current book, would improve the return assessment; a higher target ROE would need a convincing earnings and capital bridge. We will also track the Q4 restatement, continuing buybacks and the 2027 Caribbean close against the commitments already made.