Unlocking Value: 10 Fresh Picks and COF Insights

Value Radar: 10 Fresh Ideas + A Full Valuation Deep‑Dive on Capital One (COF)

Date: November 20, 2025 (United States)

This research is educational, not investment advice. Every key assumption is sourced so you can audit it.

10 new value candidates (quick scan)

  • Capital One (COF) — integrated Discover on May 18, 2025; scale, network economics, and cost/revenue synergies to 2027 now the core debate. Solid capital and improving card credit trends set up a classic “through‑the‑cycle” value test. Picked for the deep‑dive below. (reuters.com)
  • Verizon (VZ) — cash‑rich, debt‑heavy utility‑like profile; lead‑sheath remediation overhang seems manageable next to 5G FCF; dividend offers upside if capex glide‑slope holds.
  • AT&T (T) — de‑leveraging path continues; dividend covered by FCF after fiber/mobile mix shift; litigation/liability discount persists.
  • Publicis (PUB.PA) — data/AI push (e.g., Lotame) bolsters targeted ads; resilient organic growth through cycles. (ft.com)
  • Valero (VLO) — disciplined capex and strong export cracks; heavy cash returns with low net leverage; cyclical but mispriced in mid‑cycle scenarios.
  • Brookfield (BN) — fee‑bearing capital growth plus monetizations; macro/rates sensitive; long runway if fundraising cadence holds.
  • eBay (EBAY) — high free‑cash conversion, buyback‑heavy model; ad take‑rate tailwinds; needs execution on verticals/GMV.
  • Zimmer Biomet (ZBH) — procedure normalization and margin work; slower‑growth “quality at a price” screening interesting on setbacks.
  • Molson Coors (TAP) — cash generation, modest growth; de‑risked balance sheet; value hinges on brand mix and price discipline.
  • Canadian National Railway (CNI) — best‑in‑class OR discipline; secular moat; valuation re‑rates on network velocity and pricing power.

Capital One (COF): post‑Discover value case, from the ground up

Why this is timely

  • All major bank approvals to acquire Discover are in the rear‑view; close date was May 18, 2025. Conditions included remediation of Discover issues and fines, but the deal went through. (reuters.com)
  • Q3‑25 showed the first “clean” quarter with substantial Discover contribution: net revenue +23% q/q to ~$15.4B; pre‑provision earnings up 29%; provision fell sharply vs. Q2; CET1 a strong 14.4%. (nasdaq.com)
  • Management and contemporary reporting have referenced synergy potential (historically cited around $1.5B by 2027 and mid‑teens ROIC), which we pressure‑test below. (ft.com)

Bottom line to test: Is the market already pricing a sustainably higher mid‑to‑high‑teens return on tangible equity?


Business in one page

  • What they do now: a top‑tier U.S. card‑led bank plus the Discover payments network. Post‑close, Capital One became the largest U.S. card issuer by balances and controls a general‑purpose network, a strategic asset versus bank‑issuer peers without networks. (reuters.com)
  • Financial anchors around the close/integration:
    • Q3‑25: net income $3.2B; NIM 8.36% (8.43% adjusted); CET1 14.4%; allowance release $760M as credit metrics improved; integration and amortization items disclosed. (nasdaq.com)
    • Tangible book value per share (TBVPS) rose to about $105.18; period‑end TCE ratio ~10.8%. Shares outstanding ~636M. (barchart.com)
    • Q1–Q2 2025 show the step‑change in scale (loans +$100B+ in Q2 from consolidation) and a one‑off reserve build in Q2 followed by Q3 normalization. (investor.capitalone.com)
    • Card credit quality: Q3 domestic card net charge‑off rate directionally improved from Q2; monthly metrics earlier in 2025 showed elevated but manageable delinquency/charge‑off levels. (nasdaq.com)

Macro inputs that matter for cost of equity and earnings power:

  • 10‑year U.S. Treasury hovers near 4.1% in November 2025; this informs our discount rate build. (federalreserve.gov)
  • Forward‑looking equity risk premium estimates range broadly 4–5%+ in 2025 updates; we use a conservative span. (imaa-institute.org)

Cross‑checking the landscape (peers)

  • Large U.S. banks frame “mid‑teens” ROTCE as normal through‑the‑cycle, with JPMorgan often at 18–20% and Bank of America newly targeting 16–18% medium‑term. That context makes a sustained mid‑teens ROTCE for COF plausible but not trivial, given card cyclicality. (reuters.com)
  • Valuation yardstick: COF currently trades around ~1.9x price‑to‑tangible book (P/TBV) on recent TBVPS data—implying the market is underwriting high‑teens ROTCE if cost of equity and growth are typical. We’ll quantify that below. (gurufocus.com)

How we’ll value a bank like COF

For banks, residual‑income (a.k.a. excess‑return) is the cleanest framework: value equals today’s tangible book plus the present value of future excess returns (the spread between ROTCE and cost of equity) after growth. A steady‑state simplification is widely used:

  • P/TBV ≈ (ROTCE − g) / (COE − g)

Where:

  • ROTCE = normalized return on tangible common equity
  • g = long‑run growth in tangible book per share
  • COE = cost of equity

We will:

  1. Anchor TBVPS to audited/official figures,
  2. Infer a realistic normalized ROTCE after integration,
  3. Build COE from risk‑free rate and equity risk premium,
  4. Cross‑check with current multiples and credit metrics.

Audited anchors and near‑term facts (Q3‑25 unless noted)

  • TBVPS ≈ $105.18; TCE ratio ~10.8%; CET1 14.4%. (barchart.com)
  • Q3‑25 profitability optics (as integrated): net revenue ~$15.4B, pre‑provision earnings ~$7.1B, provision $2.7B, NIM 8.36% headline (8.43% adjusted). (nasdaq.com)
  • ROTCE run‑rate signal: the quarter annualizes to high‑teens—Q3 materials show ~18.8% ROTCE on average TCE—though that includes reserve release and early integration economics. We must normalize below that. (advfn.com)
  • Deal terms and synergy context: all‑stock Discover acquisition (~$35.3B headline) with synergy ambitions (widely cited ~$1.5B by 2027; 16% ROIC). This frames the upside case if delivered. (ft.com)

Core modeling assumptions (and why)

  • Starting TBVPS: $105.18 (Q3‑25, company tables/press). (barchart.com)
  • Normalized ROTCE (mid‑cycle, post‑integration): 16% base; 18% bull; 13% bear.
    • Evidence: Q3’s annualized ~18.8% plus large‑bank targets in the mid‑teens to high‑teens. We haircut for card cyclicality and early integration noise. (advfn.com)
  • Long‑run growth in TBVPS (g): 3.0% base; 3.5% bull; 2.5% bear.
    • Rationale: mature U.S. consumer credit, modest leverage growth, and share buybacks offset by through‑the‑cycle loss normalization. (nasdaq.com)
  • Cost of equity (COE): 11.0% base; 10.5% bull; 11.5% bear.
    • Build: risk‑free ~4.1% (10Y), plus 4.5–5.0% equity risk premium adjusted for U.S. downgrade nuances, times a beta near 1.2 for a card‑centric lender → low‑double‑digit COE. We round to a conservative 11%. (federalreserve.gov)

These inputs are cross‑referenced to public filings, third‑party reporting of Q3, and contemporaneous macro markers.


Residual‑income valuation (steady‑state form)

Formula: P/TBV = (ROTCE − g) / (COE − g)

  • Base case: ROTCE 16%, g 3.0%, COE 11.0% → P/TBV = (0.16−0.03)/(0.11−0.03) = 1.625.

    • Implied value = 1.625 × $105.18 ≈ $171 per share.
  • Bull case: ROTCE 18%, g 3.5%, COE 10.5% → P/TBV ≈ 2.07.

    • Implied value ≈ 2.07 × $105.18 ≈ $218 per share.
  • Bear case: ROTCE 13%, g 2.5%, COE 11.5% → P/TBV ≈ 1.17.

    • Implied value ≈ 1.17 × $105.18 ≈ $123 per share.

Interpretation: the current market P/TBV near ~1.9x implies the street is underwriting ~18% ROTCE at “mature‑market” growth and a low‑double‑digit COE. Algebraically, using base COE 11% and g 3%, a 1.88x P/TBV back‑solves to ROTCE ≈ 18.0%. That is very close to the Q3 annualized figure, i.e., little margin for downside normalization. (gurufocus.com)


Comparable‑multiples checks

  • P/TBV: COF ~1.9x; Citi typically ≈1.0–1.1x; JPMorgan commonly >2x on 20% ROTCE. COF’s ratio between Citi and JPM fits a mid‑to‑high‑teens ROTCE underwriting. (Peer targets see BofA’s 16–18% guide for context.) (gurufocus.com)
  • P/E: With Q3‑25 GAAP EPS ~$4.83 and adjusted items noted, annualized EPS would sit well above $15, but card losses are cyclical—P/E snapshots are less robust than P/TBV for banks. We prefer residual‑income as primary and use P/E only as a sanity check. (nasdaq.com)

What must go right (upside)

  • Deliver the synergy math: cost saves and network economics to reach or exceed the ~$1.5B synergy bogey by ~2027 without eroding customer experience or network acceptance. (ft.com)
  • Keep credit normalization orderly: Q3 reserve release and improving card NCO trajectory need to persist into 2026 as labor stays resilient. (nasdaq.com)
  • Sustain capital return while keeping CET1 robust (≥13% currently) so ROTCE stays mid‑teens through the cycle. (investor.capitalone.com)

What could go wrong (downside)

  • Deeper credit cycle: Card NCOs re‑accelerate, forcing renewed reserve builds (Q2‑25 showed how large this can swing), compressing ROTCE. (investor.capitalone.com)
  • Integration drag: delayed synergies, higher tech/integration spend, or network‑remediation costs tied to Discover’s historical issues and consent orders. (reuters.com)
  • Higher structural COE: if the risk‑free rate or ERP rises, the justified P/TBV falls even if ROTCE holds. (federalreserve.gov)

DCF cross‑check (why we don’t lead with it for banks)

Discounting free cash flow to equity is awkward for regulated banks because equity “investment” largely happens via retained earnings rather than explicit capex. Residual‑income embeds regulation via capital and earnings power directly; that’s why we anchor on P/TBV and ROTCE. We still reconcile the intuition: sustaining mid‑teens ROTCE with modest growth and buybacks can support per‑share value compounding in the low‑ to mid‑teens, consistent with our base case and peer targets. (reuters.com)


Valuation range and conclusion

  • Valuation range (residual‑income): $123 (bear) to $218 (bull), with a base case around $171 per share.
  • What the range means today: With the market pricing near ~1.9x P/TBV (implied ~18% ROTCE using standard g/COE), COF looks fairly to slightly richly valued for strict value investors unless you believe high‑teens ROTCE is durable through the next credit cycle. (gurufocus.com)

Recommendation for value‑focused buyers:

  • Favor entries at or below ~$175 (≈ our base intrinsic value), and accumulate more aggressively near ~$150–$160 if credit jitters push shares toward ~1.4–1.5x TBV without a deterioration in capital/credit.
  • To upgrade to “buy at market,” we’d need clearer evidence that:
    • sustainable ROTCE ≥18% post‑integration (not just one quarter),
    • synergy capture tracks toward ~$1.5B, and
    • credit metrics (delinquencies/NCOs) stay on an improving or stable path into 2026. (ft.com)

Step‑by‑step audit trail (what we did and why)

  1. Anchored TBVPS, TCE, CET1, Q3 profitability, and integration line‑items to company press/third‑party reposts for Q3‑25 and earlier 2025 quarters. (barchart.com)
  2. Confirmed regulatory approvals and close date; documented attached conditions/fines, to frame execution risk. (reuters.com)
  3. Mapped normalized ROTCE using Q3 annualized figures as an upper bound and peer return targets as sanity checks. (advfn.com)
  4. Built COE from observable risk‑free (10‑year ~4.1%) and 2025 ERP references; used a conservative 11% for a card‑heavy lender. (federalreserve.gov)
  5. Applied the steady‑state residual‑income identity to translate ROTCE/COE/g into P/TBV and per‑share values; back‑solved the market‑implied ROTCE for a cross‑check.

Monitoring checklist (next 12–18 months)

  • Credit trajectory: monthly card delinquencies/NCOs and quarterly allowance moves; look for stability rather than sharp reserve builds. (tipranks.com)
  • Synergy scorecard and Discover network milestones: cost saves, tech consolidation, acceptance/volume growth (evidence that network scale is creating value). (ft.com)
  • Capital return vs. CET1: buyback cadence relative to earnings and regulatory buffers, given CET1 ~14% in Q3‑25. (nasdaq.com)

If you want, I can share a compact residual‑income calculator so you can flex ROTCE, growth, and COE to see how sensitive P/TBV is to each lever.