Executive Summary

Investment Thesis (3 Sentences)

Ally Financial is the largest all-digital direct bank in the United States ($146B retail deposits, 92% FDIC-insured, 90% of funding) bolted onto the #1 prime-focused franchised auto lender, and the market is pricing it at roughly tangible book value (1.0x P/TBV, 10x trailing EPS) because reported GAAP returns sit at a cyclical trough — depressed by a CECL provisioning cycle that peaked in 2023-2024 and lease-residual headwinds, not by a broken franchise. The bridge from a trough Core ROTCE of 10.4% (FY2025) toward management's reaffirmed mid-teens target is the entire investment case: it rests on falling net charge-offs (5th consecutive quarter of YoY improvement, retail auto NCOs 1.97% trending to a 1.6-1.8% through-cycle range), a re-pricing deposit book (cumulative beta 63% with ~$18B of ~4% CDs maturing in 2026), and a cleaner, more focused company after the sale of credit card and exit from mortgage. The risk that makes it cheap is equally clear and unhedgeable — Ally is a leveraged (11x assets/equity), prime-but-not-pristine consumer auto lender whose earnings are hostage to the U.S. auto credit cycle, used-car prices, and unemployment, so this is a "buy the cycle low, demand a margin of safety" situation rather than a buy-and-forget compounder.

Key Metrics Dashboard

Decision

WAIT — Accumulate below ~$38 (≈0.92x TBV), Strong Buy below ~$33 (≈0.80x TBV). Fair value range $44-52 if Core ROTCE normalizes toward 12-13%; $56-62 only if the mid-teens target is fully achieved. At $42.77 the stock is fairly priced for today's trough returns and offers no margin of safety for the central risk — an auto credit downturn — so the correct action is to wait for either a cheaper entry or harder evidence that ROTCE is durably inflecting.

1. Business Model — How Ally Actually Makes Money

Ally is a bank holding company (Category IV) built around four core franchises after a deliberate 2024-2025 simplification ("Focus Forward"):

1. Dealer Financial Services — Auto Finance (the engine, ~$5.57B of FY2025 net revenue). Ally is the #1 prime-oriented franchised auto lender in the US. It originates consumer auto loans and leases sourced through ~22,000 dealer relationships, funded primarily by deposits. Q1'26: 4.4M applications (record), $11.5B consumer originations (+13% YoY), 9.6% originated yield, S-tier (super-prime) concentration ~41%. It also runs a commercial floorplan book financing dealer inventory.
2. Insurance (~$1.73B net revenue FY2025). Dealer-channel F&I products (vehicle service contracts, GAP, dealer inventory insurance). Capital-efficient, synergistic with auto, and a record $389M written premium in Q1'26.
3. Corporate Finance (~$0.54B net revenue, but 26% ROE). Senior-secured, lead-agent lending to PE-sponsored middle-market companies; ~$13.7B portfolio, ~1,200 obligors, 60% average advance rate. Management states it has never recorded a loss since entering the business in 2019 and no loan has ever been criticized or placed on nonaccrual.
4. Ally Bank — the deposit franchise (the funding moat). Largest all-digital direct bank in the US: $146B retail deposits, 3.5M customers (+6% YoY), 92% FDIC-insured, ~90% of total funding. No branches → structurally low operating cost; a national consumer brand (sports sponsorships, "Do It Right" culture) drives best-in-class retention.

The economic model in one line: gather sticky, low-cost online deposits (cost of funds 3.5% and falling) and deploy them into higher-yielding auto loans (9.6% originated yield) and corporate-finance loans, earning a ~3.5% net interest margin and a fee/insurance overlay, with the whole thing levered ~11x. Earnings = (NIM × earning assets) + fees + insurance − provisions − opex − tax. Two variables dominate the outcome: NIM (currently expanding) and provisions (currently falling). When both move the right way, ROTCE rises fast; when credit turns, the leverage cuts the other way.

This is not a free-cash-flow business — FCF is a meaningless construct for a bank (the process_financials.py "FCF -$0.65B" output is an artifact of loan-book growth and should be ignored). Value here is driven by ROTCE, tangible book value growth, NIM, net charge-offs, the deposit franchise, and CET1 — exactly as the special-handling note requires.

2. Phase 1 — Risk Analysis (Inversion)

I invert: what would have to happen for ALLY to be a permanent loss of capital, and how likely is each?

Risk Register (P × Impact)

Crude additive expected drag ≈ -25%, dominated by Risk #1. These are not independent — a recession (Risk #1) would simultaneously hit NIM (#2), pressure capital (#3), worsen residuals (#4) and could trigger deposit nerves (#5). The realistic tail scenario (severe 2008-style auto recession) is a 50%+ drawdown with a dividend cut, which is exactly why the entry price must be below tangible book, not at it.

Why the risks are survivable (the steelman of the bull)

• Funding quality is genuinely better than 2008. In the GMAC era Ally relied on wholesale/secured funding that froze. Today 90% of funding is retail deposits, 92% FDIC-insured, granular and low-balance — the franchise that nearly killed it has been replaced by the franchise that protects it.
• Prime mix and reserves. S-tier ~41% of originations; consolidated allowance coverage 2.53%, retail-auto coverage 3.75% — already reserved for a meaningfully worse environment than today's 1.97% NCO.
• Capital build. CET1 10.23% and rising ~60-120bp/yr; fully-phased-in-AOCI CET1 ~9%+ under the new standardized approach.
• CECL optionality. Because reserves were built ahead of losses in 2023-2024, a benign outcome releases reserves into earnings — a tailwind the bear case ignores.

3. Phase 2 — Financial Analysis

3a. The earnings trough is the whole story

GAAP net income to common: $847M (2023) → $558M (2024) → $742M (2025). The 2024 dip and the gap between GAAP EPS (~$2.37) and adjusted EPS ($3.81) is almost entirely elevated CECL provisioning during the post-pandemic auto credit normalization. Provision for credit losses fell from $2.17B (2024) to $1.48B (2025), a $689M tailwind, and is still rolling over (5 consecutive quarters of YoY NCO improvement). This is a textbook "earnings depressed by the denominator of the cycle" setup.

3b. ROTCE bridge (the valuation crux)

• FY2025 Core ROTCE 10.4% → Q1'26 11.1% (+440bp YoY).
• Management's mid-teens (≈14-15%) target rests on three levers, two already in motion:
  1. NIM from 3.52% → upper-3% (deposit beta 63% and rising, $18B of ~4% CDs repricing lower, asset mix shifting to higher-yield auto + corporate finance). Each ~10bp of NIM on ~$180B earning assets ≈ ~$180M pre-tax ≈ ~$0.45 EPS.
  2. Credit normalization — NCOs migrating from ~2.0% toward the 1.6-1.8% through-cycle range releases provision pressure.
  3. Expense discipline — noninterest expense guided +~1% for 2026; positive operating leverage.

3c. DuPont (bank version, normalized)

On ~$13.0B average tangible common equity:

3d. My own valuation

Method 1 — Justified P/TBV = (ROTCE − g) / (CoE − g), CoE = 11%, g = 3%, TBVPS ≈ $41:

Method 2 — Normalized P/E. A leveraged consumer-credit lender deserves a single-digit-to-low-double-digit multiple. On a normalized ~$5.00 EPS (≈13% ROTCE), 9-11x → $45-55. On the bull ~$5.60 EPS, 10x → ~$56.

Both methods converge on a fair-value range of roughly $44-52 in the base case (12-13% ROTCE), with $36 if the cycle stalls and $56-62 only if mid-teens is truly achieved. At $42.77 the stock is below base-case fair value but above the bear case — i.e., fairly-to-modestly-undervalued, with the discount entirely contingent on the credit cycle cooperating.

3e. Relative valuation sanity check (no analyst inputs)

At ~1.0x P/TBV and ~10x trailing earnings, ALLY trades cheaper than universal-bank peers (typically 1.3-2.0x TBV) and most card/consumer lenders. The discount is deserved in part — Ally is more cyclical, less diversified, and lower-return than a JPMorgan or an American Express — but the magnitude (≈1.0x TBV for a franchise generating an improving 11% ROTCE with a 2.8% dividend) is the source of the opportunity Tweedy Browne is pressing.

3f. Capital return

Dividend $1.20/yr (2.8% yield) at a conservative ~29% payout on TTM EPS — well-covered and likely to grow. $2.0B open-ended buyback (Dec'25); $147M repurchased in Q1'26 (≈1.1% of cap/qtr). At ~1.0x TBV, every dollar of buyback is accretive to TBVPS — a meaningful, underappreciated compounding lever if returns hold.

4. Phase 3 — Moat Analysis

ALLY has a narrow moat, concentrated in funding, not lending.

The honest assessment: Ally's durable advantage is the funding side — a genuinely differentiated, low-cost, sticky national digital-deposit base that few competitors can replicate. The asset side (auto lending) is fundamentally a commodity exposed to credit losses and price competition; Ally has scale and discipline but no pricing power over a loan. A moat on the right side of the balance sheet and a commodity on the left is why this is a "good bank at a cheap price," not a "great compounder."

5. Phase 4 — Synthesis, Position Sizing, Monitoring

5a. Expected-return tree (3-year horizon)

Probability-weighted ≈ +17% over 3 years (~5-6%/yr) at $42.77 — positive but not compelling for the cyclical risk taken. The math improves sharply at a lower entry: at $36 the same tree returns ~+13%/yr; at $33 (Strong Buy) the bear case is largely de-risked because you are buying below tangible book.

5b. Position sizing

A cyclical, leveraged consumer-credit lender warrants a smaller-than-average position (1.5-3% target) and a staged entry tied to price, not a full-size, buy-and-forget allocation. The right behavior is to let the auto cycle and the tape hand you a below-book entry.

5c. Entry prices

• Strong Buy: ≤ $33 (~0.80x adjusted TBV, ~6.5x normalized EPS) — the price you want if a recession scare hits; bear-case downside largely pre-paid.
• Accumulate: ≤ $38 (~0.92x TBV) — start building with a real margin of safety to tangible book.
• Current $42.77 is ~12% above the accumulate line → WAIT.

5d. Monitoring triggers (act, don't watch)

• Retail auto NCO crossing back above ~2.3% and rising two consecutive quarters → credit cycle turning; pause/avoid.
• 30+ delinquency reversing its YoY-improvement streak → leading indicator of #1.
• NIM (ex-OID) failing to progress toward the upper-3% guide → core thesis lever breaking.
• CET1 falling below ~9.5% without a clear growth-driven reason → capital-return risk.
• Deposit balances declining ex-seasonality or deposit beta rising → funding-moat erosion / cost pressure.
• Used-car price index (Manheim) rolling over hard → residual + LGD risk.
• Corporate Finance first criticized loan / nonaccrual → the "zero losses since 2019" narrative breaking.

6. The Superinvestor Signal — Read Honestly

Tweedy, Browne — a deep-value, Graham-lineage shop — opened a new ALLY position in Q1 2026. The logic is transparent and aligns with this analysis: a franchise bank trading at ~tangible book and ~10x trough earnings, with an improving credit trend and a credible path to higher normalized returns. That is precisely the kind of statistically-cheap, asset-backed, mean-reversion idea Tweedy specializes in. But Tweedy buys at a discount to a conservatively-estimated value and sizes for the possibility of being early or wrong — it is not a high-conviction quality-compounder endorsement. I weight the signal as confirmation that the price is interesting, not as a reason to override the demand for a below-book margin of safety. My verdict (WAIT, accumulate below book) is consistent with, not contradictory to, the Tweedy thesis — they are getting paid to be patient at a discount; at today's ~1.0x TBV the discount has largely closed.

7. Conclusion

Ally is a genuinely improved, well-funded, cheaply-valued bank at a cyclical earnings trough, with a real funding moat and a credible (not guaranteed) path from ~10-11% to mid-teens ROTCE. It is also a leveraged consumer-auto lender whose fate is tied to a credit cycle no one can forecast. At $42.77 ≈ 1.0x tangible book you are paying fair value for the current trough and underwriting the recovery for free-ish — but with zero margin of safety against the one risk that matters. The disciplined value answer is WAIT: accumulate below ~$38 and back up the truck below ~$33, where tangible book itself becomes the floor. Buy the franchise, but make the cycle pay you to take the credit risk.

Primary-Source Citations

• Ally Financial FY2025 Form 10-K (SEC EDGAR CIK 0000040729, filed 2026-02-25): net income to common $742/558/847M (2023-25); total net revenue $7,914M; net financing revenue+OII $6,176M; provision $1,477M; total deposits $151,649M; total equity $15,498M; preferred $2,324M; goodwill $190M; CET1 10.23% ($15,629M); avg earning assets $180,071M; avg deposit cost 3.56%; segment net revenue Auto $5,572M / Insurance $1,725M / Corp Finance $538M.
• FY2024 and FY2023 Form 10-K (SEC EDGAR), for multi-year trend.
• Q1 2026 earnings call transcript (AlphaVantage): adj EPS $1.11 (+90%), Core ROTCE 11.1%, NIM ex-OID 3.52%, CET1 10.1%, adjusted TBV/sh $41, retail auto NCO 1.97%, 30+ DQ 4.6%, $146B retail deposits, deposit beta 63%, $147M buyback.
• Q4 2025 / FY2025 earnings call transcript (AlphaVantage): adj EPS $3.81 (+62%), Core ROTCE 10.4%, adjusted TBV/sh $40, $2.0B buyback authorization, retail NCOs ended year below 2%.
• AlphaVantage COMPANY_OVERVIEW: book value $43.22, TTM diluted EPS $4.12, dividend $1.20, shares outstanding 306.5M, insiders 10.23%.
• Historical prices: AlphaVantage TIME_SERIES_DAILY_ADJUSTED, 1,260 daily records 2021-06-01..2026-06-05 (5-yr return -22.3%).
• (Sell-side analyst price targets/ratings in the overview feed were deliberately excluded per project policy.)