Initiated 2026-10-14 · Price $19.91 (close 2026-10-13, stockanalysis.com) · Mkt cap $21.25bn · Financials / super-regional bank · Model: verified

Rating: Hold — Conviction: Low

KeyCorp has largely finished its repair. The 2024 securities repositioning and Scotiabank's equity stake rebuilt capital, NIM has climbed from 2.41% to 2.89% in six quarters, and about $30bn of fixed-rate assets reprice through 2027. But at ~1.4x tangible book and ~10x forward earnings, the market already prices most of the path to the >15% ROTCE goal for 2027. Our base case gets there about a year late. The upside needs the 3.25%+ NIM target on time; the downside is a commercial-credit or deposit-cost turn while forwards price rate hikes. Hold.

Conviction tests (3a-v-c): T1 fail (Hold → Buy at −1pp: Base +23.0%; −5.4% at +1pp) · T2 pass · T3 pass · T4 pass

Model value range vs price
Bear $11.27Base $21.29Bull $30.29Price $19.91

Business overview

KeyCorp is a $184bn-asset Cleveland bank. It has ~940 branches across 15 states, a national middle-market commercial and investment banking franchise (KeyBanc Capital Markets), payments, and a wealth business with a record $74bn of AUM. Loans are $110.4bn, roughly 70% commercial. Average commercial loans are guided +8–10% in 2026, against +4–5% for total loans. Revenue was ~64% net interest income and ~36% fees in Q2 2026 ($1,258mm TE NII, $706mm noninterest income). Three variables drive earnings. First, NIM: reported 2.89% TE in Q2. It is lifted by ~$9bn of fixed-rate assets repricing in 2H26 and ~$21bn in 2027, and squeezed if deposit costs reprice up. Second, fee growth from capital markets, payments and wealth. Third, commercial credit costs, currently 40–45bp of loans.

Competition

Key competes for middle-market C&I loans and Midwest/Northeast deposits with Fifth Third (which absorbed Comerica), Huntington (which added Veritex and Cadence), PNC (which bought FirstBank), Citizens and Regions. Its rivals are mostly buying scale, while Key says it will grow organically rather than through bank M&A. Trepp's Q2 2026 review found super-regionals, Key named among them, winning C&I share back from private credit (current as of Q2 2026). Key's average loans grew 4.1% year on year. Regions' average loans grew about 2% quarter on quarter, with line utilization up 100bp to 33.5%. Pressure would show up first in deposit pricing: Key's deposit costs have risen only 1–2bp so far in Q3. Huntington is the closest listed comparable by size and mix, at 1.60x tangible book and 9.4x forward earnings (stockanalysis, 2026-10-13).

Bull case

  1. Repricing delivers NIM above target. Fixed-rate asset repricing lifts NIM to ~3.35–3.40% by 2028, and operating leverage pushes efficiency to ~55%. ROTCE then reaches the 16–19% long-term range. Plays out if deposit betas stay low (deposit costs up only 1–2bp so far this quarter). Model: nim, efficiency_ratio
  2. Commercial share and fee normalization. C&I share regained from private credit sustains ~5% earning-asset growth. Middle-market M&A and sponsor activity normalize against what management calls a record pipeline, lifting fees toward 1.9% of earning assets. Plays out if credit spreads stay tight and deal activity recovers. Model: ea_growth, fee_pct_ea

Bear case

  1. Deposit costs eat the repricing. Forwards price rate hikes (per the CFO in September), the 10Y sits at 5.21%, and deposit competition pushes NIM back toward 2.85–2.90%. Efficiency stays ~62–63% with investment spend unchanged. Plays out if Q3/Q4 deposit costs rise faster than asset yields and the 4Q26 NIM exit misses 3.00%. Model: nim, efficiency_ratio
  2. Commercial demand stalls. Higher rates cut loan demand to zero growth and capital-markets fees fall back. Plays out if sponsor activity stays weak and line utilization reverses. Model: ea_growth, fee_pct_ea
  3. Commercial credit turn. CRE refinancing and leveraged-C&I stress push NCOs to ~95bp at peak, and the allowance rebuilds by 30bp of loans. Plays out if the slow default cycle (S7b) arrives. Model: nco_rate, reserve_build

Valuation & balance sheet

Metric (definition) Current Own history (range or 5y avg) Peers Source, as-of
P/TBV (price ÷ period-end TBV/share) 1.46x on $13.62 (stockanalysis 1.46x) TBV/share $12.83 Q2'25 → $13.62 Q2'26; P/TBV history n/a (unverified) FITB 2.16x · HBAN 1.60x · CFG 1.69x · RF 1.98x (stockanalysis basis) er2q26 6/30/26; stockanalysis 2026-10-13
Forward P/E (stockanalysis consensus) 10.10x n/a (unverified) FITB 10.80x · HBAN 9.42x · CFG 10.74x · RF 9.94x stockanalysis 2026-10-13
ROTCE (company non-GAAP, continuing ops) 12.89% (Q2 2026) −3.03% FY24 · 11.85% FY25 n/a (unverified) er2q26; er4q25
NIM (TE, reported) 2.89% (Q2 2026) 2.16% FY24 · 2.69% FY25 n/a (unverified) er2q26; er4q25

Key trades at a ~20% P/TBV discount to the 1.84x peer median, and in line with them on forward P/E (peer median 10.3x). The discount reflects lower current ROTCE, not cheap earnings. FITB's 17.6x trailing P/E is inflated by Comerica merger costs.

Model-implied value range (from model-summary.json; bank, residual income and justified P/TBV, midpoints): Bear $11.27 · Base $21.29 · Bull $30.29 per share, i.e. implied returns of −43.4% / +7.0% / +52.2% vs $19.91. These ranges show how the bull and bear drivers translate into value; they are not price targets. The price sits just below Base and roughly midway between Bear and Bull, so the market already credits most of the NIM recovery and leaves little margin for a late or partial one. Re-rating: on the model's own basis, today's multiple is 1.42x FY2025 TBV/share ($14.05). Base justified P/TBV ($20.32) embeds ~1.45x, a ~2% re-rating. Residual income ($22.27, ~1.58x) is 9.6% higher and embeds ~12%, because it credits ROTCE rising to ~15% by Year 3 and terminal growth. No tail sensitivity was run: no single product exceeds half of revenue, and Bear is a break, not a plateau.

Balance sheet: net leverage not meaningful (deposit-funded bank). CET1 11.2% (Q2 2026, estimated). Management's binding constraint is marked CET1 including AOCI, targeted at 9.5–10%, which the CFO said could dip below 9.5% in Q3. AOCI was −$2.35bn and TCE/tangible assets 7.74% at 6/30. Deposits are $153.1bn; long-term debt $10.97bn (aggregate; maturity ladder not retrieved, so the model's single 2030 bucket is a placeholder). Model TCE/earning assets is 8.5% in Y1, and PPNR/provision coverage is 6.8x in Y1. ACL is 1.56% of loans. Ratings: n/a (unverified).

Model note: Verified (LibreOffice recalculated all 3,394 formula cells and they match). Unverified inputs: none. Assumptions without basis: none. Scenario consistency checks: all OK. Conventions: NIM is the reported TE margin, and the TE adjustment and ~$140mm/yr of preferred dividends are folded into a 28% all-in tax_rate. Base Y1 EPS of $1.83 is consistent with H1 diluted EPS of $0.88; the company gives no EPS guidance. AOCI is not modeled, so the rate-shock rows understate the capital channel: buybacks slow when AOCI widens. The debt line's coupon and maturity are placeholders and do not affect value.

Relative value vs peers
MetricKEYFITBHBANCFGRFPeer medianvs median
P/E (TTM)11.7x17.6x11.9x13.8x11.0x12.8x−9%
P/E (forward)10.1x10.8x9.4x10.7x9.9x10.3x−2%
PEG0.6x0.7x0.8x0.5x1.1x0.8x−23%
P/B1.2x1.4x1.0x1.1x1.3x1.2x+1%
P/TBV1.5x2.2x1.6x1.7x2.0x1.8x−20%
Dividend yield4.1%3.4%4.1%2.9%4.4%3.7%+0.4pp
P/E (TTM)FITB17.6xCFG13.8xHBAN11.9xKEY11.7xRF11.0x
P/E (forward)FITB10.8xCFG10.7xKEY10.1xRF9.9xHBAN9.4x
PEGRF1.1xHBAN0.8xFITB0.7xKEY0.6xCFG0.5x
P/BFITB1.4xRF1.3xKEY1.2xCFG1.1xHBAN1.0x
P/TBVFITB2.2xRF2.0xCFG1.7xHBAN1.6xKEY1.5x
Dividend yieldRF4.4%KEY4.1%HBAN4.1%FITB3.4%CFG2.9%

As of 2026-10-14. One source and one definition for every company: aggregator TTM and consensus-forward multiples; P/TBV on the source's tangible book. Not the model's own TBV basis (see the report's valuation table). Quotes are the 2026-10-13 4:00 PM ET close, fetched 2026-10-14. Dashed line = peer median. Source: stockanalysis.com /stocks/<ticker>/statistics/, fetched 2026-10-14.

Notes. The source's PEG figures, based on its analyst growth estimates; not computed here. FITB's 17.6x trailing P/E is inflated by Comerica merger costs after the deal closed; HBAN's trailing figures also include the Veritex and Cadence deals. Omitted: EV-based multiples are not meaningful for deposit-funded banks. Omitted: bank operating cash flow mixes loan-sale and trading flows, so FCF yield is not meaningful.

Scenario stress test

Reasoned from the bull/bear drivers above. The model column comes from the scenario overlays (Base case + shock).

Scenario Effect Mechanism Magnitude Model Δ value vs Base ($/sh)
S1 Fast equity crash − Beta-1.09 bank de-rates; capital-markets fees pause; no credit channel without recession Med −$2.46
S2 Slow bear / recession − Commercial losses, lower loan demand and fees, operating deleverage. About half of Base−Bear because it's transitory, while Bear permanently lowers NIM and raises credit costs High −$4.47
S3 Rapid rate shock − Higher CoE; deposit costs outpace remaining fixed-rate repricing; AOCI widens and slows buybacks (not modeled) Med −$2.01
S4 Slow rate grind − Loan demand slows and CRE refinancing strains; NIM roughly neutral as fixed-rate assets reprice up Med −$2.79
S5 Soft-landing cuts + Lower CoE, AOCI recovery frees capital, loan and IB demand improve; NII guided as relatively rate-insensitive Med +$2.76
S6 Recession-driven cuts − Labor-market break drives commercial and consumer losses; asset yields fall faster than deposit costs High −$3.67
S7a Credit liquidity shock − 2023-style regional-bank funding scare: deposit flight risk and spread widening de-rate the group High −$3.68
S7b Slow default cycle − Leveraged-loan and CRE defaults flow into C&I and CRE charge-offs Med −$2.87
S8 Stagflation − Higher CoE, borrower strain, cost inflation; NIM roughly neutral High −$3.25
S9a Dollar spike 0 no material effect, not modeled Low $0.00
S9b Dollar slide 0 no material effect, not modeled Low $0.00
S10 Melt-up + Regional-bank risk premium compresses; capital-markets fees benefit High +$3.20
S11 Energy supply shock − Rates spillover and modest credit strain Low −$1.05
S12 Mega-cap/AI derating 0 no material effect, not modeled (direct data-center exposure ~$700–800mm) Low $0.00

Currently active/on watch per the playbook: S3 active; S8, S10 and S11 on watch.

Model value change vs Base, by scenario
S2 Slow bear / recession−$4.47S7a Credit liquidity shock−$3.68S6 Recession-driven cuts−$3.67S8 Stagflation−$3.25S7b Slow default cycle−$2.87S4 Slow rate grind−$2.79S1 Fast equity crash−$2.46S3 Rapid rate shock−$2.01S11 Energy supply shock−$1.05S9a Dollar spike$0.00S9b Dollar slide$0.00S12 Mega-cap/AI derating$0.00S5 Soft-landing cuts+$2.76S10 Melt-up+$3.20

What would change the call

Upgrades if: 4Q26 NIM exits at or above 3.05% with 2027 guidance confirming 3.25%+, while NCOs stay at or below 45bp, or the stock falls toward ~1.2x tangible book with that path intact. Downgrades if: the 4Q26 NIM exit misses 3.00% on deposit costs, NCOs move above ~60bp, or marked CET1 pressure forces buybacks below the $1.3bn 2026 plan.

Watch items

Sources