Sixteen Percent Arrives Years Early, and for the First Time the Fundamentals Gain on the Price

All figures are Canadian dollars unless marked US$. TD's fiscal Q3 2026 covers the three months ended July 31, 2026. Share prices are shown for both the NYSE (USD) and TSX (CAD) listings.

Results vs. Consensus

Q3 FY2026 Scorecard

The reported line is comparable this quarter, which is a change worth naming. Reported diluted EPS of $2.74 against $1.89 is a clean 45% increase, because the prior-year third quarter carried only ordinary items of note. This quarter's own items are two in number and small: $34M of intangible amortization and $41M from the terminated First Horizon capital hedge, together $56M after tax. Adjusted earnings therefore exceed reported earnings by 3 cents rather than the multi-dollar distortions that made the last four quarters' reported figures unusable.

Year-Over-Year Comparison (C$M unless noted, adjusted basis)

Operating leverage on the basis that matters for a bank carrying an IFRS 17 insurance book, revenue net of insurance service expenses against expenses, is positive 485bp. That is the fifth consecutive quarter of positive operating leverage and by a wide margin the largest: fiscal Q2 produced 130bp and fiscal Q1 roughly 330bp. Pre-tax pre-provision earnings on the same basis grew 15.4%, and management's own version of that figure, which additionally strips foreign exchange and the U.S. strategic cards portfolio, was 17%.

Quarter-Over-Quarter Comparison (C$M unless noted, adjusted basis)

Three of the 12.1% sequential earnings increase is the calendar. TD's fiscal third quarter runs 92 days against the second quarter's 89, and management invoked the day count in three separate segment disclosures. Normalized for it, adjusted earnings per calendar day rose 8.4%, which is still the largest sequential move in the four quarters we have covered. The point cuts the other way from last quarter, when the same arithmetic rescued a headline that looked like deceleration.

Assessment: revenue. Revenue net of insurance service expenses grew 8.7%, faster than the 6.7% of last quarter, and the composition improved again. Net interest income grew 8.8% and now leads non-interest income at 7.9%, on a balance sheet that expanded 5.9% year over year in loans and 2.8% sequentially. Total loans net of allowance reached $991.0B, up from $936.1B a year ago, and total deposits rose sequentially for the second consecutive quarter to $1,260.7B. At initiation the central question we posed was whether TD could generate double-digit revenue growth on something other than a shrinking balance sheet. Two quarters later the answer is settled: loans, deposits and margins are all contributing, and net interest income is the fastest-growing component of the three.

Assessment: margins and costs. Both banking margins expanded and both had been guided to. U.S. Banking reached a record 3.47%, up 6bp sequentially and 28bp year over year, against a guide of a modest increase; Canadian P&C reached 2.88%, up 3bp sequentially and 5bp year over year, against a guide of relative stability. The cost side is the larger story. Adjusted expenses grew 3.9% in absolute terms, but on the basis management guides to, which strips variable compensation, foreign exchange and the U.S. strategic cards portfolio, growth was 1%. That is the lowest figure in the fiscal year and it arrives in the same quarter that variable compensation had to rise to pay for a record Wholesale result. A cost base that absorbs a record bonus accrual and still prints 1% underlying growth is doing something structural.

Assessment: EPS. Adjusted earnings available to common shareholders grew 21.0% and adjusted EPS grew 25.9%. The 4.9-point gap is the share count, down 3.9% on an average diluted basis. We have flagged this contribution every quarter because it has a known expiry, and this quarter the flag needs a correction: the contribution fell from 5.9 points last quarter to 4.9 points here, while the expiry moved further away rather than closer. The effective tax rate also fell 77bp, so roughly a point of the earnings growth is below the line, which is a modest and disclosed offset to an otherwise operational result.

Segment Performance

Adjusted Net Income by Segment (C$M)

The U.S. Banking column no longer needs the Schwab basis note that qualified it for the last four quarters. TD sold its remaining Schwab stake in February 2025, so the prior-year third quarter is the first comparative that contains no Schwab equity income. The $1,074M against $956M is a clean like-for-like comparison, and every U.S. growth rate below is on the same basis without adjustment.

Where the Earnings Growth Actually Came From

This is the one table in the report that got worse. At initiation Wholesale contributed 37% of the year-over-year increase from 12.8% of segment earnings; last quarter it was 30.8% from 14.1%, which we described as improving; this quarter it is 40.0% from 15.6%, which is the highest concentration in the three quarters we have measured. The offsetting fact is that the base improved underneath it: excluding Wholesale entirely, group adjusted earnings still grew 13.9%, against 11.8% last quarter. Both statements are true, and the honest reading is that the rest of the bank accelerated while the most cyclical segment accelerated faster.

Canadian Personal and Commercial Banking

Net income of $2,095M was a record, up 7.3% year over year on revenue of $5,517M growing 5.3% against expenses of $2,131M growing 3.1%. That is 212bp of segment operating leverage, an efficiency ratio of 38.6% against 39.4%, and a 32.3% return on equity. Average loans grew 5% with personal at 4% and business at 8%, and average deposits grew 2%. Net interest margin reached 2.88%, up 5bp year over year and 3bp sequentially. Pre-tax pre-provision earnings of $3,386M compare with $3,175M a year ago.

Underneath the aggregate, the acquisition metrics kept improving: a record third quarter in digital sales, up 17%; small business banking acquisition up 13%; commercial client acquisition up 10% year to date; record penetration rates for both consumer and small business credit cards; record proprietary originations in real estate secured lending with loans up 4%. The business bank grew loans and non-term deposits 8% each.

"What we've been able to do is maintain disciplined pricing in spite of what's been a competitive market. And so you see this fueling both our NIM expansion and to the broader segment strategy, what you see is leading sequential NIM expansion, NII and PTPP growth." — Sona Mehta, Group Head, Canadian Personal Banking

The one adverse line is credit. Impaired provisions of $446M rose 19% year over year, which the filing attributes to "credit migration in the consumer lending portfolios," offset by a performing provision of $32M against $87M a year ago. Total segment provisions as a percentage of credit volume were 0.30%, down 1bp year over year and 3bp sequentially.

Assessment: The return on equity is 32.3% against 32.5% a year ago, and that flat line is the most informative number in the segment. This is the pillar we described at initiation as a high-return annuity already sitting at its own medium-term targets, and two quarters later it is still generating records with no improvement in the return. It funds everything else and it will not lift the group return. The Q4 margin guide was upgraded from "relatively stable" to a modest increase, which is a genuine positive in a competitive Canadian mortgage market, and it is worth a few basis points at the segment level rather than anything visible at the group.

U.S. Banking

Adjusted net income of US$771M grew 10.9% year over year and 9.8% sequentially, on adjusted revenue of US$2,987M up 5% and adjusted expenses of US$1,830M up 5.7%. The segment's adjusted return on equity was 10.2%, up from 9.6% in each of the prior two quarters and 8.9% a year ago, and return on tangible common equity expanded more than 210bp to 15.6%. The adjusted efficiency ratio improved to 61.3% from 62.2% sequentially and 61.0% a year ago. Net interest margin of 3.47% was a record for the segment, up 28bp year over year.

The volume picture confirms what last quarter's spot-balance claim implied. Total average loans fell 4% year over year, an improvement from the 7% decline last quarter, and core average loans excluding the portfolios identified for sale or run-off grew 3% year over year and 1% sequentially. The run-off book has shrunk from US$19B of average balances a year ago to US$8B, so the drag is close to exhausted. Card balances grew 20%, mid-market lending balances 15% with commitments up 9%, and home equity lending 6%.

"So I think that from a volume perspective, I feel quite comfortable. And that's before we get to the things that we're doing to deliberately strengthen our distribution and our coverage models." — Leo Salom, Group Head, U.S. Banking

Fee income was the sequential surprise, up 36% reported and 2% adjusted, and management noted the headline understates it because the Nordstrom partnership agreement shifted some revenue recognition out of fees and into net interest income. The more interesting disclosure was where the growth is coming from.

"So if you look at our mid-market business, which has had a really strong run, total balances are up 15%, commitments this quarter were up 9%, our transactional fee revenues in that subsegment were up 28%. And I'd say a portion of that, a very important portion is the synergy that we're creating between ourselves and being able to afford our corporate clients the ability to avail themselves of broader debt, equity, capital market capability and M&A advisory capabilities in the wholesale bank." — Leo Salom, Group Head, U.S. Banking

Deposits remain the unresolved side. Average balances fell 5% year over year, with sweeps down 13%, business down 3% and personal down 2%. The CFO's characterization was narrower and more favourable than the segment table: "Excluding sweeps in our government banking business, deposits were flat year-over-year." Those two statements are on different bases and the bridge between them was not provided.

Assessment: The number we said had to move, moved. We wrote last quarter that the 9.6% adjusted return on equity was the figure to watch and that if it were still 9.6% at fiscal year end the U.S. leg of the return bridge was not working. It reached 10.2%, on a 60bp sequential gain built from margin, volume and a 90bp improvement in the efficiency ratio rather than from credit. Expense growth also fell into the guided mid-single-digit range for the first time this year, from 9.9% last quarter to 5.7%. This is the pillar that made the most progress in the quarter. It is also still the segment diluting a bank that earns 16.0%, and the 61.3% efficiency ratio has to travel to the mid-50s by fiscal 2029 with no interim milestone published.

Wealth Management and Insurance

Net income of $841M grew 19.6% year over year and was flat sequentially. The split is the story: Wealth Management earned $653M, up 25% year over year and 17% sequentially, while Insurance earned $188M, up 3% year over year and down 33% sequentially. Assets under administration of $831B grew 17% and assets under management of $644B grew 13%. Revenue net of insurance service expenses grew 15.6% to $2,439M and the efficiency ratio on that basis improved 160bp to 53.1%. The Wealth return on equity reached 72.5% against 62.4% a year ago.

The sequential swing is entirely Insurance, and its cause is disclosed precisely: catastrophe claims of $117M this quarter against nil in the prior quarter and $36M in the prior-year quarter. Insurance service expenses rose $248M or 18% sequentially, which the filing attributes to catastrophe losses and higher claims frequency. Insurance return on equity of 23.1% against 35.9% last quarter is the mirror image of last quarter's benign draw.

"In Q3, Direct Investing referred $1.4 billion to advice, up 34% year-over-year, deepening relationships by serving more of our clients' needs." — Kelvin Vi Tran, Chief Financial Officer

Assessment: Two quarters ago we said to model this segment off the Wealth trajectory and treat the Insurance line as a claims draw rather than a run rate. That has now been tested in both directions inside two quarters: 35.9% on nil catastrophes, then 23.1% on $117M of them. The average of the two Insurance quarters, roughly $233M, is a more defensible modelling anchor than either. Wealth is doing the compounding, with new accounts up 26%, trades per day up 20%, a record $24B of year-to-date referrals and 17% asset growth, and it did so in a quarter management itself flagged as helped by market conditions. The market help is real; the account and referral growth is not market-dependent.

Wholesale Banking

Adjusted net income of $743M was a record, up 75.7% year over year and 21.4% sequentially, on record revenue of $2,581M up 25.1%. The return on equity of 16.7% is 680bp better than a year ago and well above the segment's 13% Investor Day target. The adjusted efficiency ratio improved to 61.7% from 70.8%, and the average gross lending portfolio grew 15.5% year over year to $111.8B.

The composition is better than the headline implies, and it is worth separating carefully because the obvious reading is wrong. Trading-related revenue of $975M grew 11.7% year over year and 12.3% sequentially, and it is 15% below the $1,146M the segment produced in fiscal Q1. Everything else, which is lending, underwriting, advisory and equity commissions, grew 35.0% year over year to $1,606M and has risen in each of the three quarters of the fiscal year. This was not a trading quarter carrying a segment; it was a larger lending and advisory platform with a normal trading contribution.

What the headline does overstate is credit. Impaired provisions were $6M, against $80M last quarter and $63M a year ago, a $57M year-over-year tailwind in a segment whose provisions swung by $221M in a single quarter as recently as fiscal Q1. A $35M performing build was taken against credit migration and volume growth, so the segment is at least provisioning for the balance sheet it is adding.

"I would maybe just point to the overall growth in RWA in the quarter was 10%, but I always like to compare that to revenue growth. So the revenue growth was 2.5x the RWA growth in the quarter." — Tim Wiggan, Group Head, Wholesale Banking

Assessment: Last quarter we raised our normalized Wholesale earnings assumption because a record was earned with trading revenue down 24% sequentially. That logic applies more strongly here, not less: the record is 21% larger and trading revenue is still below the fiscal-year peak, with the growth coming from a lending book 15.5% bigger and a transaction-banking deposit base 18% bigger. Strip a normalized impaired provision back into the segment and the quarter is closer to $700M, which is where we now set the normalized run rate. The caveat management supplied is the right one and applies mostly to advisory and underwriting: those revenues came in conditions the CEO described as favorable, and favorable market conditions are a cyclical input.

Credit Quality

Every line in this table moved favourably, which has not been true in any quarter we have covered. The total provision ratio of 37bp sits below the bottom of the 40–50bp full-year guide, and management responded by narrowing the guide rather than by claiming a one-off. Formations of 20bp fell $149M sequentially, gross impaired loans of 51bp fell $138M, and the impaired-only provision ratio moved from roughly 42bp to roughly 35bp. Two quarters ago that same ratio was roughly 48bp.

"Looking forward, while results may vary by quarter and are subject to changes to economic conditions, I now expect total PCLs in 2026 to come in near the lower end of our previously guided range of 40 to 50 basis points." — Ajai Bambawale, Chief Risk Officer

The allowance fell 1bp to 96bp and still contains approximately $500M of reserves set aside for policy and trade risks, which remain largely unused. The one line that did not improve is the one we flagged last quarter: Canadian consumer delinquencies, which the Chief Risk Officer described as stable overall and "slightly higher on RESL," with impaired provisions in the consumer asset classes slightly up while every non-retail portfolio fell.

Assessment: This is now the least worrying of the four risk points we carry, and it earned that position over three quarters rather than one. The wholesale single-name problem that dominated fiscal Q1 has fully resolved, with impaired provisions in that segment down from $216M to $6M across two quarters. The Canadian consumer question that replaced it is still present and still slow-moving, and it is being absorbed by a group provision ratio that keeps falling. What remains open is the forecast, which the Chief Risk Officer explicitly deferred to the fourth-quarter call. A bank guiding below its own range with 96bp of coverage and $500M of unused tariff reserve is carrying a credit position that could deteriorate materially before it becomes a thesis problem.

Key Topics & Management Commentary

Overall Management Tone: Management was more expansive than at any point in our coverage and the shift was directional rather than merely confident: the language moved from defending a repair plan to describing where capital gets deployed next, with an investment "super cycle," a hundred new branches and 450 additional bankers taking the airtime that remediation held a year ago. The one place the answers stayed thin was price, both the price being paid for the buyback and the price of the remaining anti-money-laundering work, and neither thinning was tested. Analyst pushback was almost absent, with six questioners and a queue the operator closed early on a record quarter.

1. The Return Target Was Reached, and the Bridge Stopped Being Published

For three consecutive quarters the central artifact of this story has been management's own arithmetic connecting a delivered return on equity to the 16% medium-term target. At initiation it was 100bp of capital plus 150bp of cost onto 14.2%, summing to 16.7%. Last quarter it was 90bp plus 110bp onto 14.4%, summing to 16.4%, a downward revision nobody asked about. This quarter the delivered return is 16.0% and the bridge was not run at all.

"Total bank ROE reached 16% this quarter, in line with the medium-term target we shared at Investor Day. This reflects strong momentum and disciplined expense and capital management across our businesses as TD continues to deliver on the commitments we've shared with you." — Raymond Chun, Group President and CEO

The absence matters more than any restatement would have. Neither lever was withdrawn: the capital ratio is still 14.3% against a 13% target, and the structural cost programme is still running with management now describing upside to it. If both legs still apply on last quarter's sizing, the destination is 18.0%. If neither does, the target has been reached and there is nothing further published. Management chose a third option, deferring the whole question.

"And so come next quarter, not only will we lay out for you the Q4 numbers. But if there are any adjustments to our medium-term outlook, that's when we'll communicate some of that." — Raymond Chun, Group President and CEO

Assessment: A management team that has beaten its own plan for three consecutive quarters and has just reached its headline destination is telling you it intends to publish a new one in ninety days. That is a dated, high-probability catalyst, and it is the single event most likely to move this stock in either direction over the next two quarters. It is also why we are not extrapolating the 18.0% figure as though management endorsed it. They did not restate the bridge, and we should not restate it for them.

2. The Capital Ratio Did Not Fall, Which Is the Most Consequential Number in the Print

TD repurchased approximately 14.5 million shares in the quarter, consuming 37bp of CET1, and the ratio ended at 14.3%, unchanged. Organic capital accretion plus a 10bp benefit from model updates across the Canadian real estate secured lending, Canadian Business Banking and Wholesale portfolios absorbed the entire buyback. Surplus over the 13% managed target actually rose, from roughly $8.3B to roughly $8.5B, because risk-weighted assets grew 1.9% to $653.4B.

We modelled this lever as depleting. In fiscal Q1 the surplus was roughly $9.5B, in Q2 roughly $8.3B, and the natural extrapolation was a shrinking prize. That extrapolation was wrong, and the CEO said why.

"TD is different than peers. We don't just start from a position of strength in capital. We also produce capital at a very strong rate. As a result, it will take time for the bank to reduce its CET1 ratio." — Raymond Chun, Group President and CEO

He then quantified it, which is the disclosure of the quarter.

"To illustrate the potential magnitude of capital return, assuming continued strong organic capital accretion and RWA growth in line with fiscal 2026 year-to-date, TD could return over $13 billion in capital in fiscal 2027 to reach a 13% CET1 ratio by the end of that year." — Raymond Chun, Group President and CEO

Assessment: Over $13 billion of repurchases at the print-day price is roughly 4.7% of the shares outstanding, in one fiscal year, on top of a 2.67% dividend. That is a shareholder yield near 7.4% before a dollar of earnings growth, and it reframes the pillar we have been carrying. The capital lever is not depleting on schedule, which slightly delays the return benefit; it is much larger than we modelled, which materially raises the total-return arithmetic. On balance that is the strongest single argument the bull case gained this quarter, and it is why our share-count assumption moves up rather than down.

3. The Regulator Cut the Buffer and TD Did Not Cut Its Target

In June, inside TD's fiscal third quarter, OSFI lowered the Domestic Stability Buffer to 3.0% of risk-weighted assets from 3.5% and lowered the top of its range to 3.0% from 4.0%. The consequence for a Canadian systemically important bank is that the expected CET1 floor moved to 11.0% from 11.5%. The CEO referenced the change in a single sentence.

"TD has significant capital flexibility. The lower DSB requirement and range create even more capital flexibility than we had before." — Raymond Chun, Group President and CEO

What did not change is TD's own 13.0% managed target, reaffirmed for the second half of fiscal 2027. Holding that constant while the regulatory expectation fell 50bp means TD's self-imposed cushion widened from 150bp to 200bp above the floor.

Assessment: There is an unexercised option sitting in that gap. A bank whose managed target is 200bp above the regulatory expectation, while it carries an open consent order and an open FINTRAC review, is holding a deliberate reserve against regulatory outcome rather than against credit. That is defensible, and the arithmetic is worth stating: the 50bp the regulator released is roughly $3.3B of capital TD chose not to claim, and the full 200bp cushion it now carries above the floor is roughly $13.1B. Nobody on the call asked whether the 13% target survives the resolution of the consent order, and it is the highest-value unasked question of the quarter.

4. Expenses Grew One Percent on the Basis Management Guides To

Adjusted non-interest expenses grew 3.9% year over year in absolute terms. On the guided basis, excluding variable compensation, foreign exchange and the U.S. strategic cards portfolio, growth was 1%, the lowest reading of the fiscal year against a 3–4% full-year target that fiscal Q1 overshot and fiscal Q2 met at the bottom.

"Expenses increased 4% year-over-year, with approximately 3% driven by variable compensation, foreign exchange and the impact of the U.S. strategic cards portfolio. This disciplined expense management reflects significant structural cost reductions, coupled with continued investments in business growth." — Kelvin Vi Tran, Chief Financial Officer

The structural programme behind it is now described in different language than a quarter ago. The $900M fiscal 2026 goal was banked at the halfway mark; this quarter the CEO added that the medium-term $2.0–2.5B target will be reached sooner and that he sees upside to the range itself.

"The $2 billion to $2.5 billion that we had planned for over the MTO, we're significantly on track to get there sooner and certainly see upside on the $2 billion to $2.5 billion." — Raymond Chun, Group President and CEO

Assessment: A 1% expense growth rate in a quarter that had to accrue a record Wholesale bonus pool is the strongest cost datapoint of our coverage. The caveat we have carried since initiation is unchanged: there is still no phased schedule for the remaining $1.1–1.6B of the medium-term programme, and "upside" without a number is not a commitment. But the pattern has now repeated four times, and at some point a repeated pattern is evidence rather than a claim.

5. The Efficiency Ratio Reached Its Medium-Term Target in Year One

The group efficiency ratio net of insurance service expenses was 55.2%, improved 180bp sequentially and 260bp year over year. The CFO placed it against the plan directly.

"Our efficiency ratio, net of ISE, this quarter was 55.2%, in line with the medium-term target that we shared at Investor Day." — Kelvin Vi Tran, Chief Financial Officer

This is the second published medium-term destination reached inside twelve months of the Investor Day, alongside the 16% return on equity. Nine-month efficiency of 56.4% is not yet there, so the quarter is running ahead of the year, and the quarter carried the benefit of a revenue mix skewed toward Wholesale and Wealth.

Assessment: Reaching two of the plan's headline destinations in year one of a multi-year plan is either extraordinary execution or a plan that was set conservatively. Both readings support the same conclusion for the shares, which is that the published targets have stopped carrying information about the earnings trajectory. What matters now is what replaces them at the fourth-quarter call.

6. One Hundred New U.S. Branches, Announced With the Consent Order Still in Place

The filing states that TD expects to open 100 new stores by the end of calendar 2028, subject to regulatory approval, against a network that has shrunk from 1,100 stores a year ago to 1,048. That is roughly a 10% expansion of the footprint, and it is the first expansionary commitment the U.S. bank has made since the October 2024 global resolution.

"Obviously, we're -- at the same time, I'll say that we're focused very much on our AML program. It still remains our #1 priority. We have made significant progress on that program. I'd say today, the degree of maturity and effectiveness of that program is much stronger. And as a result, in discussions with our regulators, et cetera, I feel quite comfortable to be able to announce today that we'll be opening up those 100 stores." — Leo Salom, Group Head, U.S. Banking

Management paired the announcement with roughly 450 additional bankers across retail, advice, mortgage and vertically specialized commercial coverage, and framed the whole build as self-funded by the productivity agenda and by moderating governance and control spend through calendar 2027.

Assessment: The financial content of this is small and the informational content is large. A hundred branches over roughly two and a half years is a modest capital commitment for a bank with $8.5B of surplus, and management says it will not raise the expense growth rate. What it signals is that the segment head is comfortable enough with the regulatory relationship to announce a footprint expansion in public. That is a different posture than a year ago, and it is the first thing in three quarters that has genuinely shortened our sense of the remediation timeline. It is also, as he immediately clarified when pressed, not a statement about the asset cap.

7. The Anti-Money-Laundering Cost Guide Moved Up for the First Time

TD now expects fiscal 2026 U.S. BSA/AML remediation and related governance and control investments of approximately US$550M pre-tax, raised from approximately US$500M. The filing gives the reason as "increased costs associated with lookback activities."

"From a financial perspective, we expect our overall U.S. AML remediation expenses for the year to be approximately $550 million, and we continue to expect overall fiscal 2026 expense growth for U.S. Banking to be in line with our mid-single-digit expense guidance." — Leo Salom, Group Head, U.S. Banking

The arithmetic behind the change is unflattering to the prior guide. First-half spend was US$321M and the third quarter added US$125M, so nine-month spend is approximately US$446M. Against the old US$500M guide, the fourth quarter would have had to come in near US$54M, a level the run rate never supported. The new guide implies roughly US$104M in the fourth quarter, which is a genuine step down from US$125M rather than a fictional one. Meanwhile the filing repeats, for a fourth consecutive quarter, that the Suspicious Activity Report lookback runs into calendar 2027, that the scope of the lookback reviews may be further expanded, and that the overall timeline may extend beyond calendar 2027.

Assessment: This is the specific watch item we set last quarter, and it broke. We wrote that holding the US$500M guide required a 44% second-half step down in a line that had risen in each of the two prior quarters; instead the guide moved. The dollar amount is immaterial to earnings, roughly US$50M pre-tax against a bank earning $4.7B a quarter. The signal is not immaterial: the cost line moved on the precise mechanism, lookback scope, that the filing names as the way the timeline extends past 2027. The branch announcement is the counterweight and it is a real one, but a guide that moves the wrong way for the reason the risk factor predicted is an escalation, not a rounding error.

8. Credit: the Chief Risk Officer Says the Peak Is Behind

Every credit indicator improved and the full-year guide was narrowed downward rather than merely reaffirmed. The more consequential disclosure was the Chief Risk Officer's willingness, for the first time in our coverage, to characterize the cycle.

"And then just on peak, I mean, it's very difficult to call a peak. In many ways, I think we're past the peak. If you go back to 2025, you'll find Q2 '25 was a peak for total PCL at 58 bps. If you look at Q1 '25, our impaireds were 50 basis points. So in some ways, we are past the peak, okay?" — Ajai Bambawale, Chief Risk Officer

He attached three conditions immediately: the fiscal 2027 forecast is not yet built, trade policy and the Middle East conflict remain live inputs, and the answer would be revisited at the fourth-quarter call. He also reiterated that the 96bp allowance includes approximately $500M for tariffs, which remains largely unused.

Assessment: "Past the peak" from a Chief Risk Officer who spent fiscal Q1 explaining a single wholesale borrower is a meaningful change of register, and it is supported by the disclosed data rather than resting on it. The residual risk is now almost entirely Canadian consumer, concentrated in the cohort management named last quarter, and it is being absorbed inside a falling group ratio. Our provision assumption for the full year moves down again.

9. Wholesale Supplied Forty Percent of the Growth From Fifteen Percent of the Earnings

Wholesale Banking earned $743M against $423M a year ago, a $320M increase that is 40.0% of the group's entire $800M year-over-year earnings growth, generated by a segment that produced 15.6% of segment earnings. That is the highest concentration in the three quarters we have measured, against 30.8% from 14.1% last quarter and 37% from 12.8% at initiation.

The CEO named the dependency himself rather than leaving it to be inferred.

"This quarter's record results were also supported by favorable market conditions in our Wealth and Wholesale Banking businesses." — Raymond Chun, Group President and CEO

The offsetting evidence is that the platform is materially larger than it was and that trading is not what grew. The average gross lending portfolio grew 15.5% year over year to $111.8B, transaction-banking deposits grew 18%, non-trading revenue grew 35.0%, and the CEO noted that quarterly revenue has almost doubled since the TD Cowen acquisition closed. Trading revenue meanwhile sits 15% below the fiscal Q1 level. Wholesale is a bigger business, not merely a luckier one.

Assessment: The concentration is what we grade and the composition is what tempers it. A group return of 16.0% that leans 40% of its growth on the segment with the highest revenue volatility should be normalized before it is extended, but the normalization is smaller than the headline implies because the growth is lending and advisory rather than trading. We take roughly 30bp off the group return, to 15.7%, mostly for the $6M wholesale impaired provision and part of the Corporate tax benefit, and build the valuation from there. That is the single largest judgment in this report and it is the one most likely to be wrong in the bulls' favour if capital markets stay open.

10. The U.S. Return on Equity Finally Broke Out of 9.6%

The U.S. Banking adjusted return on equity reached 10.2%, after two consecutive quarters at 9.6%. Return on tangible common equity expanded more than 210bp to 15.6%, the margin set a record at 3.47%, and adjusted expense growth fell into the guided range at 5.7% from 9.9% last quarter. Total loans were positive sequentially for a second quarter, now with the average-balance measure confirming what the spot measure claimed early.

"In U.S. Banking, earnings were up 11% year-over-year, and ROTCE expanded by over 210 basis points to 15.6%." — Kelvin Vi Tran, Chief Financial Officer

Management extended the guidance a year further out than usual, which it has not done before in our coverage.

"As we look forward, Kelvin referenced it in his comments, I do think we will see NIM expansion in 2027, albeit slightly more modest than what we've seen in the last 2 quarters as we do begin to try to continue to take share in our respective markets, but long-winded way of saying, I think we've got great growth prospects going into 2027." — Leo Salom, Group Head, U.S. Banking

Assessment: This is the upgrade trigger we published last quarter, at least in its first half. We said the sustainable group return could only move above 17% if the U.S. segment broke out of 9.6%, and it did, by 60bp, on operating drivers rather than credit. What it has not yet done is close the gap to the group: a 10.2% segment inside a 16.0% bank is still dilutive, and it needs several more quarters like this one before the fiscal 2029 mid-50s efficiency target looks like a schedule rather than an aspiration. The direction is right and the distance is long.

11. A Canadian Investment "Super Cycle" Entered the Capital Story

For the first time, the capital-deployment discussion led with organic growth rather than with returning surplus to shareholders, and the reason given was a macroeconomic thesis rather than a balance-sheet one. The CEO opened the call by citing TD Economics' estimate of more than $1 trillion of Canadian spending over the coming decade, and returned to it when asked how capital gets used.

"But our primary use of the capital will be for organic growth. And you're seeing some of that where we've deployed, organic growth in our Wholesale Banking business, and we'll continue to find some of those opportunities as you see some of the investment opportunities I do think that potentially will arise in the sort of super cycle of investment potential in Canada over the next decade." — Raymond Chun, Group President and CEO

He placed acquisitions second, restricted to fee-income areas already named at the Investor Day, and explicitly not a priority. Shareholder return came third.

Assessment: This ordering is new and it cuts both ways. A bank that can deploy surplus capital into 10% risk-weighted-asset growth at 2.5 times the revenue growth rate, as Wholesale did this quarter, creates more value than one that buys back stock at 2.4 times book. That is the strongest available argument for a higher long-run growth rate, and our valuation framework has held that assumption fixed at 3% since initiation while the growth outlook improved. It is also the argument that most easily becomes an excuse: "primary use is organic growth" is the sentence that precedes most disappointing capital allocation, and the $13B illustration is a ceiling on returns rather than a commitment to them.

12. Artificial Intelligence Hit Its Annual Target With a Quarter to Spare

The $200M fiscal 2026 value target from artificial intelligence was essentially reached at three quarters, against $145M at the half. Management reframed the programme from a list of use cases into three named transformation areas: retail end-to-end credit, the software development life cycle, and contact centres.

"Three quarters into the year, we have essentially hit our fiscal 2026 target of $200 million in value from AI. We expect to extract further value through the remainder of the year across predictive, generative and agentic AI use cases." — Raymond Chun, Group President and CEO

The operational detail is more persuasive than the dollar figure: over 20,000 client-facing colleagues in Canada now supported by generative knowledge-management tools, roughly a third of manual funding processes automated in Canadian auto finance, the first agentic capability fully scaled in real estate secured lending, and an industry-first deployment of AI vehicle damage estimation in the insurance business.

Assessment: The $200M is roughly 0.6% of the annual expense base, so the near-term financial content is small. What is worth tracking is the $1B medium-term figure, because it is the mechanism by which the structural cost programme gets from the $900M already banked to the $2.0–2.5B target without a headcount action. Hitting the annual number a quarter early, twice in a row, is the closest thing to a schedule that this programme has produced.

Guidance & Outlook

Two changes in the guidance posture matter more than any individual line. The first is that the qualifier moved again: fiscal 2026 targets went from "potential upside" at fiscal Q1, to "on track to outperform" at Q2, to "on track to significantly outperform" here. The second is the only line that moved against the bank, which is the anti-money-laundering budget. Every other revision this quarter was favourable, and management still declined to raise a single financial target, deferring the whole exercise to the fourth-quarter call.

Implied run-rate against the full-year guide: nine-month adjusted EPS of $7.59 against $6.19 is 22.6% growth, and the full-year guide remains 6–8%. For the guide to be merely met rather than beaten, fourth-quarter adjusted EPS would have to fall by roughly half against the prior-year fourth quarter, which nothing in the print or the call suggests. On the return on equity, 14.8% across nine months against a 13% full-year target says the same thing. The published fiscal 2026 targets no longer carry information about the fiscal 2026 outcome.

Street at: consensus entering this print was $2.47 of adjusted EPS, which the quarter cleared by 12.1%, after 5.8% in fiscal Q2 and 8.4% in fiscal Q1. The beats are getting larger, not smaller, three quarters into the pattern. On revenue the Street was exactly right, at $15.28B against $15,280M delivered, which locates the entire forecast error in costs and credit rather than in the franchise's top line.

Guidance style: conservative to the point of being uninformative on the published fiscal-year targets, and newly forthcoming a year out on the operating lines. The novelties this quarter are all forward: fiscal 2027 U.S. margin expansion, fiscal 2027 U.S. expense growth below fiscal 2026, a $13B fiscal 2027 buyback illustration, a hundred stores by calendar 2028, and an explicit promise to revisit the medium-term outlook in ninety days. A management team that has beaten every target it published is choosing to extend the horizon rather than raise the numbers, and the fourth-quarter call is where that choice gets resolved.

Analyst Q&A Highlights

The question period was the shortest of our coverage. Six analysts asked questions and the operator closed the queue, on a record quarter in which the bank reached its headline medium-term target. Two other Canadian majors held calls the same morning, which accounts for some of it. What it does not account for is that the three most consequential disclosures of the quarter, the unpublished return bridge, the widened capital cushion after the regulatory buffer cut, and the price being paid for the buyback, drew either no question or an answer that went unchallenged.

Whether the Branch Build Signals Movement on the Consent Order

The first question of the call went directly to the tension in the quarter's biggest announcement: a bank operating under a consent order and an asset cap declaring a 10% expansion of its store network. The initial answer described the strategic rationale and the regulator discussions behind it. The follow-up removed the ambiguity, and the response was the flattest statement management made all morning.

Q: "Okay. But you wouldn't necessarily make any inferences in the asset cap that you're under in the U.S. and the branch openings?"
— Matthew Lee, Canaccord Genuity

A: "No. No, Matt. Let me be clear. The consent order is still in place and we are working hard to satisfy every aspect of that consent order, and there is no greater priority for me than resolving that comprehensively and urgently."
— Leo Salom, Group Head, U.S. Banking

Assessment: The exchange did exactly what a good first question should, which is to stop a favourable announcement from being read as more than it is. The branch build is evidence that the regulatory relationship has improved enough to permit a footprint decision. It is not evidence that the asset cap is moving, and management refused to let it be taken as such. That refusal is more credible than the announcement, and it is why we treat the remediation timeline as shortened at the margin rather than resolved.

What Changes to Take Common Equity Tier 1 From 14.3% to 13%

A pointed question observed that the capital ratio has fallen only 50bp in a year while the plan calls for more than 100bp of reduction in the next sixteen months, and asked what changes. The answer reordered the capital priorities in public, placing organic growth first, opportunistic fee-income acquisitions second, and shareholder return third, then handed off to the wholesale bank to demonstrate what organic deployment looks like.

Q: "First question is on the capital plan. Ray, targeting 13% by the end of next year, what's going to change in the plan? The CET1 ratio is down 50 basis points over the past year, and we're talking about a more than 100 basis point haircut in the next 16 months."
— Gabriel Dechaine, National Bank Financial

A: "As I said at Investor Day and today, the way I think you should all think about the way we will be managing capital is that we are managing it differently. But our primary use of the capital will be for organic growth."
— Raymond Chun, Group President and CEO

Assessment: The question identified the arithmetic problem precisely and the answer did not solve it. Fifty basis points in twelve months against more than a hundred in sixteen requires either a step change in the buyback or a step change in risk-weighted-asset growth, and the answer chose the second without quantifying it. That is coherent with the $13B illustration, which is explicitly conditioned on risk-weighted assets growing in line with the year to date, but it converts a capital-return commitment into a growth forecast. Investors buying this stock for the buyback should notice that management put it third on the list.

Whether U.S. Margin Expansion Extends Past the Fourth Quarter

A long and well-constructed question laid out the case that the U.S. margin should keep expanding beyond the guided quarter: excess deposits allowing selective funding, a reinvestment-rate tailwind with several years to run, and loan growth that need not compress spreads. The response accepted the framing, quantified the loan-to-deposit ratio behind it, and then extended the guidance horizon a full year with an explicit deceleration attached.

Q: "So you're giving the positive NIM commentary for Q4, I would assume that extends beyond that as well."
— Gabriel Dechaine, National Bank Financial

A: "As we look forward, Kelvin referenced it in his comments, I do think we will see NIM expansion in 2027, albeit slightly more modest than what we've seen in the last 2 quarters as we do begin to try to continue to take share in our respective markets, but long-winded way of saying, I think we've got great growth prospects going into 2027."
— Leo Salom, Group Head, U.S. Banking

Assessment: This is the most useful forward disclosure on the call and it was volunteered rather than extracted. A 76% loan-to-deposit ratio is the structural reason the margin can expand while loans grow, because the funding is already in the building. The honest half of the answer is the qualifier: expansion will be more modest, because taking share costs spread. A segment guiding to slower margin expansion while accelerating volume is describing a mix shift toward growth, which is the right trade at a 10.2% return on equity and would be the wrong one at 16%.

What the Thirteen-Billion-Dollar Buyback Illustration Assumes About Price

The only question of the quarter that touched valuation asked two things: whether the capital-return figure is buybacks alone, and how management weighs the size of a repurchase against its effect on book value per share at the current multiple. The first half was answered immediately and precisely. The second half was not answered at all, and the response moved to the capital-priority ordering before closing on the promise of a medium-term outlook update.

Q: "No, I just wanted to confirm the $13 billion, Ray, that you talked about in terms of returning capital to shareholders, is that all -- is that buybacks, that's not including dividends? And the question also, just how do you measure that size of buybacks and the potential implications on just book value per share growth? Just trying to understand the mechanics there as well."
— Doug Young, Desjardins

A: "Just on the $13 billion, it is on specifically buybacks. And that's what we could do from a buyback. First, I just want to clarify that, Doug, right?"
— Raymond Chun, Group President and CEO

Assessment: This is the second consecutive quarter in which a direct question about buying back stock at a rising multiple has produced an answer about the amount rather than the price. Six months ago the shares were characterized as undervalued at 1.98 times book and a $7B programme was authorized on that basis. The stock has since risen 24% against book value per share that rose $1.49, and the repurchase pace has increased. The dodge is not evidence of a bad decision, and buying back stock at 2.4 times book is still accretive to earnings per share while it is dilutive to book value per share. But an answer would have been better than an ordering, and the question deserves to be asked a third time.

Whether Credit Losses Have Peaked

Credit drew the most questions of any topic, from two separate analysts in consecutive turns, and produced the most quotable answer of the call. The first asked what had gone better than expected across the year; the second asked directly whether the cycle has peaked. The response to the second was unusually willing to be pinned, with a specific historical peak cited and three conditions attached.

Q: "And I'm certainly not trying to [ pin ] you on timing, but it seems like potentially, we're at the point where we're close to reaching that sort of peak credit losses in this current cycle. And I'm just wondering, like you've had some really good progress in the last couple of quarters. And is there anything anomalous in terms of what TD might be doing?"
— Multiple analysts incl. Mehmed Rizvanovic, Scotiabank; Doug Young, Desjardins

A: "And then just on peak, I mean, it's very difficult to call a peak. In many ways, I think we're past the peak. If you go back to 2025, you'll find Q2 '25 was a peak for total PCL at 58 bps. If you look at Q1 '25, our impaireds were 50 basis points. So in some ways, we are past the peak, okay?"
— Ajai Bambawale, Chief Risk Officer

Assessment: Compare this with fiscal Q1, when the same seat was explaining a $221M sequential increase and a single wholesale borrower. The claim is anchored to disclosed peaks rather than to sentiment, and the conditions attached are the right ones: no fiscal 2027 forecast yet, trade and geopolitical inputs unresolved, revisit at the fourth-quarter call. What is missing is any framing of what a normalized through-cycle level looks like from here, which matters because a 37bp provision ratio is well below where any of this bank's medium-term planning assumes it will settle.

Flat Mortgage Balances While Margins Expanded

One question probed the softest volume line in the Canadian bank: mortgage balances roughly flat sequentially in a competitive summer market. The answer reframed it as a deliberate trade, claimed record proprietary originations alongside the flat balance, and attributed the outcome to distribution investment and a newly scaled automated decisioning capability rather than to price.

Q: "Just in terms of the mortgage growth in the quarter or I guess it was more flattish this quarter sequentially. Can you maybe just talk about the spread dynamic? Is this just the market getting a bit more competitive and you're looking to protect spreads? Or is there something else sort of driving that flat result?"
— Mehmed Rizvanovic, Scotiabank

A: "What we've been able to do is maintain disciplined pricing in spite of what's been a competitive market. And so you see this fueling both our NIM expansion and to the broader segment strategy, what you see is leading sequential NIM expansion, NII and PTPP growth."
— Sona Mehta, Group Head, Canadian Personal Banking

Assessment: Last quarter the same seat described competitive pricing as having "set in" and delivered 2bp of sequential margin; this quarter it delivered 3bp with real estate secured lending balances up 4% year over year. Choosing margin over volume in Canadian mortgages is the correct call for a bank whose scarce resource is return rather than balance sheet, and it is the specific reason the Q4 margin guide could be upgraded from stability to a modest increase. The cost is visible in the volume line and the benefit is visible in the 32.3% segment return.

What a Ten-Percent Larger Branch Network Does to Expense Growth

The final question of the call made the obvious objection to the store announcement: branches cost money on day one and earn revenue much later, so a 10% footprint expansion should show up in expense growth. The answer enumerated three investment pockets, sized the banker hiring, and then made a commitment about next year's expense growth that had not been made before.

Q: "So I don't think the opening of 100 new U.S. stores is necessarily new as you highlighted. But maybe you can give us some comfort around how that's going to impact, I guess, really PTPP or earnings growth over the next 2 years, right? Like it's roughly expanding the branch footprint by 10%. So it's significant. We all know you open a branch, day 1, it results in expenses with no revenue."
— Paul Holden, CIBC

A: "So in many ways, if I just -- we'll give you more detailed guidance in the fourth quarter, but I fully intend to be able to deliver an expense profile -- an expense growth profile in 2027, despite the increased investment in growth initiatives that is lower than what we have posted in 2026."
— Leo Salom, Group Head, U.S. Banking

Assessment: This is a real commitment and it is the one to carry into next quarter's scorecard. Fiscal 2026 U.S. segment expense growth is tracking mid-single-digit; the promise is that fiscal 2027 comes in below that while absorbing a hundred stores and 450 bankers. The stated funding sources are the productivity agenda and a moderation in governance and control spend, which is the same anti-money-laundering line that just had its annual guide raised. Two commitments that depend on the same declining cost line are one commitment, and it is the thing most likely to be renegotiated at the fourth-quarter call.

What They're NOT Saying

1. Why the return-on-equity bridge was not restated. Two quarters ago the levers were 100bp of capital plus 150bp of cost onto 14.2%; last quarter 90bp plus 110bp onto 14.4%; this quarter the delivered return is 16.0% and the arithmetic was not run. Neither lever was withdrawn. If both still apply the destination is 18.0%, which nobody said, and if neither applies the medium-term target has been reached, which nobody said either.
2. Why the 13% capital target did not move when the regulatory floor did. The Domestic Stability Buffer cut in June took the systemically-important-bank CET1 expectation to 11.0% from 11.5%. Management called it additional flexibility and reaffirmed 13.0%, widening its own cushion from 150bp to 200bp. No one asked whether the 13% survives the resolution of the consent order.
3. What price the $13 billion of repurchases assumes. Asked directly about the book-value implications of buying back stock at the current multiple, the answer covered the amount and the capital-priority ordering and never reached the price. This is the second consecutive quarter the same question has gone unanswered.
4. How U.S. deposits are flat. The segment table shows average deposits down 5% year over year, with personal down 2% and business down 3% before any sweep adjustment. The CFO said deposits were flat excluding sweeps in the government banking business. The two statements use different bases and the bridge was not given. Last quarter produced the same pattern on loans, where a spot-balance claim was made against average-balance disclosure.
5. What the $52 million performing build is composed of. Attributed to the Wholesale and Canadian Commercial portfolios, driven by volume growth and "some credit migration." Two drivers, no split, third consecutive quarter with the same gap.
6. What the fourth-quarter anti-money-laundering run rate is. The full-year guide moved to approximately US$550M and nine-month spend is approximately US$446M, implying roughly US$104M in the fourth quarter against US$125M in the third. Management raised the annual figure and left the quarterly path off the record, in the same quarter it promised fiscal 2027 U.S. expense growth below fiscal 2026 partly on the strength of that line declining.
7. How much of the 16.0% return was market conditions. The CEO stated that the record was supported by favorable market conditions in Wealth and Wholesale. He did not size it, no analyst asked him to, and no normalized figure was offered on a call whose headline was a medium-term target being reached.
8. What the FINTRAC review could produce. The filing sentence that the review "is ongoing, and subject to the outcome, may result in additional regulatory actions" has now appeared for four consecutive quarters. Not in the prepared remarks, not in the question period, four quarters running.

Market Reaction

• Pre-print setup: the shares closed at US$119.43 on the NYSE line and C$165.87 on the TSX line the session before the print. Entering the print the stock was +26.8% year to date in U.S. dollars, +56.9% over the trailing twelve months and -0.8% over the trailing thirty days, inside a 52-week closing range of US$72.85 to US$124.80. In Canadian dollars the comparable figures are +28.2% year to date, +58.0% over twelve months and -2.4% over thirty days, against a 52-week closing range of C$100.27 to C$175.27. The S&P 500 was +12.1% year to date.
• Print-day session (August 27, before the open): the NYSE line opened at US$120.17, a 0.6% gap, traded US$118.94 to US$121.42, and closed at US$121.09, up 1.4% or US$1.66. The TSX line opened at C$166.10, traded C$164.88 to C$168.27, and closed at C$167.84, up 1.2% or C$1.97. The 20bp gap between the two listings is currency, not disagreement.
• Volume: 2.9 million shares on the NYSE line against a 2.3 million thirty-day average, 1.3 times normal; 3.5 million on the TSX line against a 4.4 million average, 0.8 times normal.
• Peer context: two other Canadian majors reported the same morning and two had reported two days earlier. Across the six large Canadian banks on that session, TD was the only one to close higher. The moves ranged from -2.8% to TD's +1.2%, with the largest decline belonging to a bank reporting the same morning. The S&P/TSX Composite closed +0.1% and the S&P 500 +0.7%.
• Follow-through: across the four sessions after the print the TSX line closed at C$168.61, C$167.67, C$166.23 and C$168.20, finishing +1.4% against the pre-print close while the S&P/TSX Composite fell 2.0%. The gain was held rather than faded, which was not true after the fiscal Q1 print.

A 1.2% gain on a 12.1% earnings beat is a small absolute response and a large relative one. Every other large Canadian bank fell on the session, including the two that reported the same morning, and TD has now led the group on its own print day for two consecutive quarters after landing mid-pack at fiscal Q1. The relative move, not the absolute one, is the signal.

The setup explains the modest absolute response better than the print does. Unlike fiscal Q2, when the stock entered the print within 0.3% of its 52-week closing high, this print arrived after a pullback: the shares were down 2.4% over thirty days in Canadian dollars and roughly 5% below the 52-week closing high of C$175.27. A stock that had already given back some ground responded to a large beat by recovering part of it, and volume on the TSX line was below the thirty-day average, which is a striking datapoint for a print day. The engagement was moderate and the direction was clear.

The four sessions afterward are the more informative half. The market gave the gain back after fiscal Q1 and held it after fiscal Q2; this time it held the gain through a week in which the Canadian index fell 2%, an outperformance of roughly 3.4 points. That pattern is consistent with a stock being accumulated on fundamentals rather than traded on a headline, and it is consistent with published targets that were refreshed upward into the high C$170s and above after the print, against a traded price near C$168.

Street Perspective

Debate: Does Reaching the 16% Target Mean the Destination Moves Up, or That There Is Nothing Left?

Bull view: the bull case on the Street is that a target reached in year one of a multi-year plan was set too low, and that management has all but said so. The published levers were never withdrawn, the capital ratio has not moved, the structural cost programme is described as having upside to its own range, and the CEO promised to communicate any adjustments to the medium-term outlook at the fourth-quarter call. On that reading the 16% is a waypoint and a higher number is coming with a date attached.

Bear view: the bear camp observes that the bridge stopped being published in the same quarter the destination was reached, which is exactly what you would expect if there is nothing further to publish. The two prior versions of that bridge were revised downward, from 16.7% to 16.4%, and a management team that revises a destination down twice and then stops describing it is not signalling upside. The quarter that delivered the 16.0% also carried a $6M wholesale provision, a Corporate segment helped by tax benefits, and market conditions the CEO himself called favorable.

Our take: the bears have the better read of the disclosure and the bulls have the better read of the base rate. Three consecutive quarters of early delivery is real evidence about how this management team sets targets, and the honest expectation is that the medium-term outlook goes up at the fourth-quarter call. We are raising our exit-fiscal-2027 return assumption from 15.9% to 17.5%, anchored on a market-normalized 15.7% rather than the delivered 16.0%, plus the capital lever taken in full and the cost lever haircut by one sixth. That is a 160bp upgrade and the largest single-quarter revision we have made to this name. It is also, for the first time, a sustainable return that is no longer below what the current price requires, which moves the entire argument onto the cost of equity.

Debate: Is the Capital Story a Buyback Story or a Growth Story?

Bull view: the optimistic reading is that both, and that the $13B illustration understates it. The capital ratio did not fall despite 37bp of repurchases, so the bank out-earns its own buyback. The regulatory buffer cut released a further 50bp of headroom that has not been claimed. At the print-day price, $13B is roughly 4.7% of the shares outstanding in a single year, which with the dividend is a shareholder yield near 7.4% before any earnings growth.

Bear view: the skeptics point out that management explicitly ranked shareholder return third, behind organic growth and behind opportunistic fee-income acquisitions, and that the $13B figure is conditioned on risk-weighted assets growing only in line with the year to date. If the Canadian investment super-cycle thesis is right and the bank deploys into it, the same capital funds loan growth instead of repurchases. The illustration is a ceiling, and the ordering of the priorities tells you which way management will lean.

Our take: this is a better problem than it looks. Deploying surplus capital into risk-weighted-asset growth that generates revenue at 2.5 times the asset growth rate, as the wholesale bank did this quarter, creates more value than repurchasing stock at 2.4 times book. We would rather own the growth outcome than the buyback outcome at this multiple, and we would rather own the buyback outcome at 1.9 times. What we would not do is assume both, and the $13B figure has already been reported in some places as including dividends, which it does not.

Debate: Is the Anti-Money-Laundering Chapter Closing or Widening?

Bull view: the constructive reading is that the branch announcement is the tell. A bank that has cleared its footprint expansion in discussions with its regulators, describes its programme's maturity and effectiveness as much stronger, sees opportunities to accelerate remaining deliverables, and has advanced lookback reviews across multiple populations, is a bank approaching the end of a remediation rather than the middle of one. The cost increase is small and is being absorbed inside a mid-single-digit segment expense guide.

Bear view: the bear camp notes the guide moved up for the first time, that the stated cause is lookback scope, and that lookback scope is the exact mechanism the filing names for the timeline extending beyond calendar 2027. The Suspicious Activity Report lookback still runs into 2027, internal-audit validation runs alongside it, the FINTRAC review remains open and unmentioned for a fourth quarter, and the segment head restated that the consent order and asset cap are unchanged.

Our take: we are escalating this risk point from contained to emerging, and the escalation is about signal rather than dollars. US$50M is nothing to a bank earning $4.7B a quarter. A cost guide that moves the wrong way for precisely the reason the risk factor predicted is the first hard evidence in three quarters that the timeline could extend, and it arrived in the same quarter management promised fiscal 2027 U.S. expense growth below fiscal 2026 partly on this line declining. The branch announcement is a genuine counterweight and it does not neutralize the point.

Debate: Does 2.4 Times Book Have a Case?

Bull view: the bulls argue that the multiple is measuring the wrong thing for a bank compounding at this rate. Adjusted earnings per share grew 25.9%, the return on equity is 16.0% and rising, the share count is falling 4% a year with a $13B programme illustrated for next year, the dividend yields 2.67%, and both published medium-term destinations have been reached inside a year with a target revision promised. At an unchanged multiple that arithmetic delivers a low-to-mid teens total return without any re-rating, and the loop of beat-then-raise has three quarters of history behind it.

Bear view: the bears note that the shares now trade at 2.41 times reported book against 2.29 a quarter ago and 1.98 two quarters ago, and that on our framework the price requires a 17.4% sustainable return at a 9% cost of equity, against roughly 16.7% a quarter ago. The adjusted price-to-earnings ratio of 17.2 times is barely changed from last quarter's 17.0 only because earnings caught up with a price that had already moved. A Canadian bank at 2.4 times book is priced for a destination reached in a quarter that management itself said was helped by market conditions.

Our take: the gap between the two answers is smaller than it has been in either prior quarter, and that narrowing is the finding of this report. Delivered return rose 160bp while the return the price demands rose roughly 70bp. Our fair-value estimate rose about 14% while the share price rose 7.5%. That is the first quarter of our coverage in which the fundamentals gained on the price rather than the reverse. It is also still a gap, of roughly 4% before the dividend, and a stock whose expected twelve-month return sits near zero on our framework and in the low teens on the constant-multiple arithmetic is precisely as undecided as it was three months ago, just at a higher level of both.

Model Update & Valuation Framework

The table below records the changes to the assumptions we set last quarter. Where management publishes a target we anchor to it and state where we sit relative to it.

Valuation: Price-to-Book Against Sustainable Return on Equity

We use the same construction as at initiation. Justified price-to-book equals return on equity less growth, divided by cost of equity less growth, with long-run growth held at 3%, run across the plausible range of both inputs rather than resolved to a single point. Book value per share is $69.69 as reported at July 31, 2026.

Justified multiple = (ROE - g) / (COE - g) with g = 3.0%. Implied prices apply the multiple to reported book value per share of $69.69 and are rounded to the nearest dollar. The first row uses the adjusted return earned across the fiscal year to date, which is the market-condition-neutral read. The second uses the return actually earned this quarter, which is also the published medium-term target. The third is our own estimate: the delivered 16.0% normalized to 15.7%, plus the 90bp capital lever taken in full and the 110bp cost lever haircut by one sixth. Applying those same two levers to the unnormalized 16.0% gives 18.0%, which at a 9% cost of equity is worth C$174; management did not restate that arithmetic this quarter, so we show it as our extension rather than as their guidance.

Where the stock sits. TD closed the print at C$167.84, or 2.41 times reported book value, against 2.29 one quarter earlier and 1.98 two quarters earlier. The traded price now exceeds eight of the nine cells above, and for the first time since initiation one cell sits above it. The median of the nine implied prices is C$129, roughly 23% below the traded price. Taking the two most defensible cells on the same method we have used each quarter, the top return row at a 9% and a 10% cost of equity, gives a fair-value range of C$144 to C$168 with a midpoint near C$156, about 7% below the close. Adding the 2.67% forward dividend yield produces a base-case twelve-month total return of roughly negative 4%. Last quarter the same construction produced negative 6%, and the quarter before that positive 5%. Our fair-value estimate rose about 10% this quarter while the share price rose 7.5%, which is the first quarter of our coverage in which that comparison ran in the shareholder's favour.

The single most important cell is the top-left one. Our own 17.5% sustainable return at a 9% cost of equity justifies C$168, which is the traded price to within a fraction of a percent. That is not a coincidence to be dressed up as insight; it is a statement that our framework and the market agree exactly, provided the cost of equity is 9%. Move it to 10% and the same return justifies C$144, or 14% below. For two quarters this rating rested on the argument that TD could not earn what its price required. That argument is now spent. What is left is the risk premium.

What the multiple has to believe. Solving backwards, 2.41 times against $69.69 of book requires a 17.4% sustainable return at a 9% cost of equity, or a 19.9% return at 10%. One quarter ago the requirement was roughly 16.7% at 8.9%. So the bar rose about 70bp while the delivered return rose 160bp, and the distance between what the bank earns and what the price demands narrowed from 2.3 points to 1.4. Our modelled 17.5% now sits marginally above the 17.4% the price requires, which it has not done in any prior quarter of this coverage. That crossover is the substance of this quarter.

Which makes the cost of equity the whole argument. Nine percent is the aggressive end of a defensible range for this bank. It implies the market requires no incremental premium for an open U.S. consent order, an asset cap the segment head reaffirmed on this call, a Suspicious Activity Report lookback running into calendar 2027 with a filing warning it may extend beyond, an open FINTRAC review in its fourth quarter of silence, and a remediation cost guide that has just moved up for the first time. A bank two years past a criminal resolution, whose regulatory work is on the record as unfinished, is not obviously entitled to the same required return as a peer that never had one. That is the entire case against the multiple now, and it is a narrower and more honest case than the one we made in either prior quarter.

The assumption now doing the most work is the growth rate. We have held long-run growth at 3% since initiation, which was defensible for a bank shrinking its U.S. balance sheet under a consent order. It is a harder assumption to defend against a quarter in which the wholesale bank grew risk-weighted assets 10% at 2.5 times the revenue-to-asset ratio, the U.S. bank committed to a hundred new branches and 450 bankers, and the CEO framed the entire capital plan around a decade-long domestic investment cycle. At g of 4% and a 17.0% sustainable return, the justified multiple is 2.60 times, or C$181, which is 8% above the traded price rather than 4% below it. We are not making that change on one quarter of evidence. Readers should know that it is the single input that flips our conclusion, and that the evidence moved toward it this quarter.

The competing arithmetic, stated fairly. The framework assumes the multiple converges toward a justified level within a year, and multiples routinely do not. If 2.41 times simply holds, the twelve-month return is earnings growth plus the dividend. Fiscal 2027 adjusted EPS growth of 8–12%, which is what a 4.5–5.0% share-count reduction plus mid-single-digit pre-provision growth supports, plus a 2.67% yield, is a low-to-mid teens total return. Published price targets refreshed after the print cluster in the high C$170s to high C$190s against a traded price near C$168, implying a high-single-digit to mid-teens price return before the dividend. The three methods span roughly negative 1% to positive 17%, and the entire spread is multiple persistence.

Sensitivity that matters most. The cost of equity still dominates: one point of it moves fair value by between C$14 and C$25 per share on this book value, which exceeds the entire remaining gap between the delivered return and our exit-fiscal-2027 assumption. At 2.41 times the shares carry no cushion against a higher required return. The specific mechanism by which that required return would rise is a regulatory outcome, which is why the anti-money-laundering guide moving up matters more to the valuation than its dollar amount suggests.

Dividend. The quarterly dividend was $1.12, unchanged from the rate declared last quarter and up 6.7% from $1.05 a year ago. It annualizes to a 2.67% yield at the print-day close, down from 2.87% at the fiscal Q2 print purely because the share price rose faster than the dividend. The reported payout ratio of 40.7% is the lowest of the fiscal year and leaves ample room for the next increase.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation, carried forward unchanged and graded against what this quarter's print and call revealed. Status tags follow the standing thesis: bull pillars are on track, at risk or broken; risk points are contained, emerging or materializing.

Overall: the standing thesis is stronger on seven of eight points and one risk point escalated. This is the first quarter of our coverage in which the fundamentals gained on the price. Our fair-value estimate rose about 14% while the share price rose 7.5%; last quarter those figures were 3% and 15.6%, and at initiation the price was already ahead. Two published medium-term destinations, a 16% return on equity and a mid-50s efficiency ratio, were reached inside twelve months of the Investor Day that set them. The capital lever turned out to be larger and slower than we modelled, which raises the total-return arithmetic more than it delays it. The one thing that moved against the bank is the anti-money-laundering cost guide, and it moved for the reason the filing's own risk factor predicted.

Action: Maintaining Hold. It is worth stating plainly that this call has been wrong in market terms: the shares are up roughly 24% since we initiated at Hold in February, and we have now watched the operating story beat our expectations three quarters running. The error was underestimating the delivery speed, not the valuation framework, and the framework still says the same thing. What has changed is the shape of the argument. At 2.41 times book the price requires a 17.4% sustainable return at a 9% cost of equity; we now carry 17.5%, a 160bp upgrade, so the return we model finally clears the bar. The Hold no longer rests on the earnings and rests instead on the risk premium, and the honest version of that is narrow: a 9% cost of equity for a bank with an open consent order, a live asset cap, a lookback running into calendar 2027 and a remediation guide that just moved up is thin, but it is not indefensible. Full delivery of everything management has named, an 18.0% return at 9%, is worth C$174, or 4% above the traded price, which with the dividend is roughly 6.5%. A stock where the bull case fully realized returns 6.5% is not an Outperform, and a bank compounding at this rate with a $13B buyback illustration and a target revision promised in ninety days is not an Underperform. Upgrade triggers, revised: a medium-term outlook at the fourth-quarter call that raises the return destination above 18%; evidence supporting a long-run growth assumption above 3%, most plausibly sustained double-digit risk-weighted-asset growth at Wholesale's revenue-to-asset ratio; resolution or material narrowing of the U.S. consent order, which is the specific event that would justify a cost of equity below 9%; or a de-rating toward 2.1 times book, roughly C$146. Downgrade triggers, revised: a multiple above 2.65 times book without an upward revision to the return destination; the fourth-quarter medium-term outlook arriving unchanged at 16%, which would mean the bridge really is spent; fiscal 2027 U.S. anti-money-laundering spend guided above the fiscal 2026 level; or Wholesale supplying more than 40% of year-over-year earnings growth for a third consecutive quarter. For existing holders this is a well-run bank that has reached its own targets early, is returning 7% of its market capitalization a year, and has a dated catalyst in ninety days. For new capital, the price still assumes the destination and the fourth-quarter call is a cheap thing to wait for.