1. Executive summary

Three-sentence thesis. Sallie Mae is the #1 US private student lender — a deposit-funded specialty bank earning a 31% return on tangible common equity, trading at 6.2x trailing earnings and 2.0x tangible book while management retires ~6% of the share count a year. The market is pricing SLM as a melting-ice-cube consumer lender facing a tough graduate-employment cycle, but it is missing a once-in-a-generation tailwind: the 2025 federal "Big Beautiful Bill" PLUS-loan reform caps federal graduate/parent borrowing, handing SLM an estimated **$5B incremental annual origination opportunity (70% growth over 2025)** phasing in from July 2026 through 2028. At $22.39 — about 72% of a conservative $31 fair value — you are paid a ~14% total shareholder yield (2.3% dividend + ~12% buyback) to wait for a catalyst that is already legislated.

Metrics dashboard

2. Phase 0 — Why does this opportunity exist? (Klarman)

A specialty consumer lender with a 31% ROTCE should not trade at 6x earnings. The mispricing comes from a cluster of identifiable, non-permanent sources:

1. "Consumer-credit-in-a-weak-job-market" stigma. 2024–2025 headlines about elevated recent-graduate unemployment and AI job displacement caused the market to extrapolate rising charge-offs. The stock fell ~30% over the trailing year (52w high $34.40 -> $22.39). But management's own data shows recent-grad unemployment has normalized to Nov-2023/2024 levels, late-stage delinquencies are stable at ~1%, and FY2025 charge-offs actually fell 4bps YoY.

2. Optics of "down" 2026 EPS guidance. SLM guided initial 2026 EPS to $2.70–$2.80, below 2025's $3.46 — because it is deliberately spending ~$120M of one-time investment to capture PLUS. The market read a guide-down; the reality is invest-ahead-of-the-curve. (By Q1 2026 management already raised the guide to $3.10–$3.20 via buybacks and loan-sale gains.)

3. Model-change confusion. The shift to selling new originations through the KKR strategic partnership (vs only seasoned portfolios) distorts reported credit ratios via denominator effects. Management added an entire 10-K/deck appendix to bridge GAAP-vs-non-GAAP — a textbook "complexity/stigma" Klarman setup that scares off lazy screens.

4. Institutional neglect of a boring, single-product lender. No glamour; a regulated bank; policy-dependent. This is exactly where mispricing persists.

Einhorn's fresh 1.41% stake (Q1 2026) is corroborating evidence that a respected value mind sees the same gap — but the thesis below stands on its own first-principles math, not on coattails.

3. Phase 1 — Risk analysis (inversion first)

"All I want to know is where I'm going to die, so I'll never go there." — Munger

How could this lose 50%+ permanently?

The bear's only credible path to permanent capital loss is a credit blow-up: a deep recession (or AI-driven structural unemployment among new grads) drives charge-offs from ~2% toward the 5–7% peak seen in the 2008–2010 era, the allowance proves inadequate, and the ~$2.1B equity base is impaired before the portfolio reprices. SLM is leveraged ~10x on a tangible-equity basis (inherent to banking), so credit — not the income statement — is the kill switch.

Risk register (P x impact)

Tail scenario (lollapalooza). A 2008-style combination — recession + funding freeze + political push to discharge private student debt — could stack into a >50% drawdown. This is the genuine fat tail and the reason this is a 1.5–2.5% position, not 5%.

The 3-sentence bear case (stated better than the bears)

SLM is a 10x-levered monoline lending unsecured money to 18–22-year-olds with no income, at the exact moment AI is gutting the entry-level white-collar jobs those graduates need to repay; the reported 31% ROTCE is flattered by reserve releases and gain-on-sale financial engineering that will reverse when credit turns; and the "PLUS tailwind" invites a price war that compresses the very gain-on-sale premium the bull case depends on. If I cannot refute this, I should not own it — so I address each point in Phase 2/3.

Non-price sell triggers (pre-committed)

1. Net charge-off rate breaks above 3.0% of loans in repayment for two consecutive quarters with rising late-stage roll rates (not a denominator/loan-sale artifact).
2. CET1 falls below 10% or the buyback is suspended to preserve capital.
3. Originations decline in 2027 (PLUS thesis broken) absent a deliberate balance-sheet shrink.
4. Federal legislation makes private student loans dischargeable in bankruptcy.

4. Phase 2 — Financial analysis (lender lens: ROTCE, book value, NIM, credit)

For a deposit-funded bank, value is driven by return on tangible common equity, tangible book value growth, net interest margin, and credit quality — not free cash flow. (FCF is not meaningful for a balance-sheet lender; the project FCF fields are set to null in the summary.)

4.1 Returns — the core of the thesis

• ROTCE ~ 31% (FY2025 net income to common $729.1M / average common equity ~$2.35B). SLM carries negligible goodwill/intangibles, so ROE ~ ROTCE. This is a spectacular return for a lender — well above large-bank ROEs of 12–15% and the company's ~10–11% cost of equity.
• ROE 5-yr average ~24% (process_financials.py); every year comfortably clears Buffett's 15% bar.
• Why so high? A high-yield asset (private student loans yield well above prime) funded by cheap retail deposits ($21B FDIC-insured) plus securitization, run at a 30.6% efficiency ratio. The 5.29% NIM is roughly double a typical commercial bank's, and rising as funding costs fall.

4.2 DuPont (lender form)

ROE = (NIM-driven net margin) x (asset turnover) x (equity multiplier).

• Net profit margin (FY2025): net income $729M / total revenue $3.11B ~ 23%.
• ROA (TTM): 2.57% — strong for a bank (vs ~1.0–1.3% for diversified banks).
• Equity multiplier: ~10x (assets ~$29.7B / equity ~$2.5B at YE2025). High leverage is normal and intended for a deposit-funded bank; the relevant guardrail is CET1 12.4%, comfortably above the ~7% regulatory minimum + buffer.

4.3 Credit — the part that actually matters

• FY2025 net charge-offs $346M = 2.15% of average loans in repayment, down 4bps YoY.
• Q1 2026 NCOs $89M; 30+ day delinquency 3.98% (modestly improving sequentially); late-stage steady at ~1%.
• Reserve rate 6.05% of loan exposure — roughly 3x the running charge-off rate, a thick cushion. Allowance ~ $1.25B+ against a $20.3B book.
• Underwriting has tightened for years: cosigner rate 95% (up from 86% five years ago), average approval FICO 754. Management's long-term destination is a "high-1s to low-2s" NCO rate.
• Loss-mitigation works: >80% of modified borrowers complete their first six payments; ~75% of the 2023 modification cohort current after 24+ months.
• Refuting the bear: charge-offs are falling, not rising, despite the weak recent-grad job market — because of the 95% cosigner rate (parents backstop the school-to-work transition) and disciplined underwriting. The reserve build, not gain-on-sale, is doing the heavy lifting.

4.4 Tangible book value and the buyback flywheel

• Tangible BVPS $11.04; the stock at $22.39 is 2.0x TBV — but a 31%-ROTCE bank should trade at a large premium to book (see justified-multiple math below).
• Capital return is the secret weapon: ~58% of shares retired since 2020 at an average $17.15, far below today's price — genuinely accretive, not value-destructive. The new $500M (2-yr) authorization ~ 12% of the float; combined with the 2.3% dividend that is a ~14% total shareholder yield, with a payout ratio of only ~15% leaving ample room.
• Strategic loan sales (KKR partnership + seasoned portfolio sales) generate gain-on-sale and program-management fees, building recurring, capital-light fee income and freeing capital for buybacks — Q1 2026 alone: $3.3B sold for $146M gains.

4.5 Valuation (lender-appropriate)

A) Justified price-to-tangible-book (Gordon: P/TBV = (ROTCE - g) / (r - g)). Even haircutting ROTCE from 31% to a conservative through-cycle 22–25%, with cost of equity r = 11–12% and sustainable growth g = 4–5%:

B) Normalized through-cycle EPS x multiple. Normalized EPS ~$3.20–$3.50 (2026 trough $3.15 depressed by one-time PLUS spend; 2025 was $3.46). A specialty lender of this quality merits 9–11x:

C) 5-year forward DCF (PLUS catalyst). Off the 2026 $3.15 base, PLUS-driven origination growth (+50–70% over several years) plus continued buybacks supports a 10–14% EPS CAGR -> 2031E EPS $5.07–$6.07; at a still-cheap 9x and discounted 10%/yr, present value $28–$34.

Triangulated fair value: $28 – $36, base case ~$31. Current $22.39 = 72% of fair value ~ 28% margin of safety.

5. Phase 3 — Moat analysis

SLM's moat is narrow-but-real, built on three reinforcing sources:

1. Brand + school distribution (the core). "Sallie Mae" is the default name in private student lending; the company sits inside university financial-aid offices and the FAFSA/enrollment funnel. New entrants cannot replicate decades of school relationships and the organic marketing channels SLM "pioneered over the last five years" overnight. This is the asset that lets SLM capture the PLUS windfall — schools' #1 post-reform worry is filling classrooms, and SLM is the partner with product, underwriting, and terms already designed.
2. Proprietary underwriting data / credit models. ~20 years of private-student-loan performance data across cohorts and the school-to-work transition. The measurable result: tightening the buy box (cosigner 86%->95%, FICO 750->754) while growing originations and share — evidence of pricing and selection power competitors lack.
3. Cost-of-funds + scale advantage. $21B of stable, FDIC-insured retail deposits plus a deep, repeat ABS investor base ("flight to quality" in private credit favors SLM's high-quality paper). A 5.29% NIM and 30.6% efficiency ratio reflect a cost structure rivals cannot match at scale.

Moat durability test (wider or narrower in 10 years?)

Verdict: stable-to-modestly-widening. PLUS reform removes a government competitor (federal Grad PLUS/Parent PLUS) from SLM's highest-value segments — a regulatory tailwind that structurally widens the addressable moat, partially offset by new private competition.

6. Phase 4 — Management & incentives

• CEO Jonathan Witter (since 2020) and CFO Peter Graham have executed a coherent, disciplined strategy: tighten credit, raise NIM, build the strategic-partnership/loan-sale model, and return capital relentlessly.
• Capital allocation is the standout. Retiring ~58% of shares since 2020 at an average $17.15 — consistently below intrinsic value — is among the best buyback records in financials. The programmatic "buy more on down days" discipline and the willingness to monetize the loan/equity valuation gap (Q1 2026 $2B seasoned sale + $200M ASR) show owner-like thinking.
• Insider ownership ~1.27% — modest in absolute terms (typical for a large-cap bank run by professional managers, not founders), but compensation is equity-weighted and the buyback behaves like a giant per-share value transfer to continuing holders.
• Succession: deep finance bench (multiple Strategic-Finance MVPs, experienced CFO). No key-person cliff risk evident.

Munger's incentive test: "If I were management with these incentives, what would I do?" — exactly what they are doing: invest ahead of a legislated demand surge, fund it capital-efficiently via partnerships, and shrink the share count while the stock is cheap. No discrepancy between stated strategy and rational self-interest.

7. Phase 5 — Catalysts

This is the rare value setup with a legislated, dated catalyst — reducing time risk in Klarman's framework. With a catalyst present, the required MOS is ~20%; the actual ~28% MOS is more than adequate.

8. Phase 6 — Decision synthesis

Expected-return probability tree (5-yr, price-only; dividends/buyback on top)

Add the ~12% annual buyback accretion to per-share value and the through-cycle expected return is attractive relative to an ~11% cost of equity, with a legislated catalyst skewing the distribution to the upside.

Position sizing

A narrow-moat, 10x-levered monoline with genuine tail risk and a 31% ROTCE: 1.5–2.5% position, entered in tranches. Accumulate at/near current price; add aggressively below ~$20.

Recommendation

+---------------------------------------------------------------+
|            INVESTMENT RECOMMENDATION: SLM CORPORATION          |
+---------------------------------------------------------------+
| Current Price: $22.39    Date: 2026-06-06                     |
| Intrinsic value (weighted): ~$31      Margin of safety: ~28%  |
+---------------------------------------------------------------+
| RECOMMENDATION:  ACCUMULATE                                   |
| STRONG BUY  (>=30% MOS):   <= $20.50                          |
| ACCUMULATE  (~20% MOS):    <= $25.00  (already in zone)       |
| FAIR VALUE:                ~$31                               |
| TRIM:                      > $37                              |
| POSITION SIZE: 1.5–2.5% of portfolio                          |
| CATALYST: PLUS reform origination surge (Jul 2026 -> 2028)    |
| PRIMARY RISK: Severe credit cycle / new-grad unemployment     |
| SELL TRIGGER: NCO >3% two quarters w/ rising rolls; CET1 <10% |
+---------------------------------------------------------------+

Monitoring triggers

9. Post-analysis psychology check (Munger)

• Social proof (Einhorn): acknowledged but the thesis stands on independent ROTCE/valuation math.
• Deprival/price-drop bias: the -30% drop is why I looked, but value (28% MOS to $31) — not the drop — is the reason to act. Cross-checked on absolutes, not relatives.
• Authority bias: no analyst targets used; valuation is first-principles.
• Humility check / one thing that kills the thesis: credit. If new-grad unemployment becomes a structural AI phenomenon rather than a cyclical air-pocket, charge-offs and the 95% cosigner backstop are the variables to watch like a hawk.

Final Munger test — if SLM dropped 50% tomorrow on no credit deterioration, would I buy more? Yes — at ~$11 (1x tangible book for a 31%-ROTCE lender) it would be a table-pounding buy. That answer confirms conviction; hence ACCUMULATE now and add on weakness.

10. Sources

• SEC EDGAR 10-K FY2025 (filed 2026-02-19, CIK 0001032033), FY2024, FY2023 — data/10-K-*.htm
• Q1 2026 earnings call transcript (2026-04-23) — data/earnings-transcript-Q1-2026.md
• Q4/FY2025 earnings call transcript (2026-01-22) — data/earnings-transcript-Q4-2025.md
• AlphaVantage COMPANY_OVERVIEW, INCOME_STATEMENT, BALANCE_SHEET, CASH_FLOW — data/*.json
• AlphaVantage TIME_SERIES_DAILY_ADJUSTED (1,362 daily records 2021–2026) — data/historical-prices.json
• Processed: data/price-summary.md, data/financial-summary.md
• No analyst reports, price targets, or Yahoo Finance used. EODHD returned 401; AlphaVantage + SEC used.