Annaly Capital Management, Inc. (NYSE: NLY) — The Best-Diversified House on a Structurally Bad Street, Bought at Its Richest-Ever Price

Independent equity research. Report date: July 11, 2026. Reference price: $22.86 (July 10, 2026 close).

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice. The analysis that follows takes no position and carries no price target.

Verdict: HOLD. A genuinely better-run, better-diversified mortgage REIT than the pure-play peers — but you are being asked to pay the richest price in Annaly’s history for it. Own it for the ~13% income if you already do; do not chase it here. Accumulation zone is ~1.0x book or below (roughly ≤ $20), where the premium disappears and the diversification optionality is free. Not a short — the book is stabilizing, the dividend was just raised, and the policy tailwind is real.

Annaly is the highest-quality expression of a low-quality business model. Unlike AGNC’s pure agency book, Annaly runs three legs — Agency MBS (56% of capital), an Onslow Bay/OBX residential-credit securitization engine (23%), and a deep out-of-the-money MSR book (21%) — and the legs are genuinely complementary: MSR gains when rates rise and prepayments slow, partially self-hedging the agency book, while OBX is a real, hard-to-replicate fee-and-securitization franchise growing issuance 60%+ year-on-year. That is why 2025’s +20.2% economic return included a real +5.5% book-value gain rather than pure mark-to-market noise, why the dividend was just raised ($0.70 → $0.75, the first increase after the 2023 cut), and why EAD of $0.76 actually covers the payout. This is not AGNC’s “return-of-capital treadmill.” It is a better machine.

But the tape has already figured this out. At ~$22.86 against a Q1’26 book value of $19.82, Annaly trades near 1.15x book — its richest-ever multiple (P/B in the 99.8th percentile of its own decade-long history) — after a +32% total return over the trailing year that carried it to a five-year high. The empirical driver of that move is not durable momentum (the stock carries essentially no momentum-factor load); it is a −0.76 InterestRate loading — a levered bond proxy that re-rated as the market priced a benign-rate, GSE-supported, Basel-friendly regime. That regime can persist, and mid-teens new-money returns support the dividend. But at 1.15x book you have pre-paid for the good news, your carry is thinner than the headline yield (part of every “13%” is return of your own capital), and the same −0.76 rate sensitivity that lifted the stock will take book value down hard if the term premium backs up or the GSE-conservatorship-exit tail redefines what an “agency” security is. Great operator, wrong entry point. Framing: quality-compounder-of-a-bad-business at a full price — a HOLD, not a chase.

Conviction: Medium. Flips bullish if the stock de-rates back toward/below book (~$20) while EAD coverage and book stability hold — then you get a best-in-class operator at no premium. Flips bearish if leverage creeps up, EAD slips back below the dividend, or a disorderly GSE conservatorship-exit headline hits the agency basis. Tag: “You’re paying a premium for the one mortgage REIT that arguably deserves one — which is exactly why it isn’t cheap.”

📈 Stock Price Action — Five-Year Event Map

Over five years Annaly round-tripped a full rate cycle: split-adjusted, the stock fell from ~$16.8 (late 2021) to a $9.07 trough in October 2022 amid the fastest Fed hiking cycle in four decades, then ground back up as rates peaked and spreads normalized, reaching a five-year high of $22.99 in July 2026. At $22.86 it sits essentially at its five-year high (~0.6% off), inside a 52-week range of $17.07–$22.99. (All prices split-adjusted for the 1-for-4 reverse split effective September 23, 2022; unadjusted the stock trades in the low-$20s versus a mid-single-digit pre-split handle.)

Cycle narrative. (1–2) The 2022 drawdown was the defining event: as the Fed took the funds rate from ~0% to ~5%, agency MBS spreads blew out and every levered agency REIT’s book value cratered — Annaly bottomed near $9 (adjusted) in October 2022, a −46% peak-to-trough, before a sharp rate-vol-peak bounce. (3) 2023 was a grind: the March regional-bank failures and “higher-for-longer” kept book values pinned near their trough. (4–5) The re-rating ran through 2024–2025 as the Fed pivoted and spreads stabilized; crucially, 2025’s +20.2% economic return contained a real +5.5% book-value gain, and the diversified MSR/credit legs let Annaly rotate capital opportunistically — the market rewarded this with a ~40% total-return year. (6) The January 2026 Trump GSE directive to purchase up to $200B of agency MBS tightened the basis and spiked the stock to ~$23. (7) The June-quarter Middle-East energy shock lifted the term premium and cost 1.9% of book in Q1’26, but the Basel bank-capital reproposal (residential-mortgage RWA −30%, MSR capital-deduction relief) provided a technical offset. (8) The June 2026 dividend raise confirmed management’s confidence and the stock has held its five-year highs. (Price moves are Fact; attributed drivers are Interpretation, cross-referenced to earnings prints, 8-K events, and the news feed.)

1. Executive Summary

Annaly Capital Management is the largest US mortgage REIT (~$16.7B market capitalization, ~$136B balance sheet, ~707M shares), internally managed, and — unlike the pure-play agency peers — deliberately diversified across three housing-finance strategies: Agency MBS (56% of dedicated capital), Residential Credit via the Onslow Bay/OBX securitization platform (23%), and Mortgage Servicing Rights (21%). The central investment tension is not whether Annaly is a good operator — it demonstrably is one of the best in a bad neighborhood — but whether that quality is already more than reflected in a price that sits at the richest multiple of book value in the company’s history.

The business is structurally challenged but unusually well-run. At its core Annaly is a levered spread vehicle: it borrows short (repo at ~3.9%) against agency MBS and earns a net interest spread (~1.42%), amplified ~5.7x. This is a no-moat, commodity-asset model whose book value erodes across rate cycles and whose ~13% dividend is, in economic substance, partly a return of capital. What differentiates Annaly is (i) a genuine internal hedge — the MSR book, whose value rises when rates rise and prepayments slow, partially offsetting agency-MBS mark-to-market losses; (ii) a real, scaled fee franchise — OBX is the largest non-bank securitizer of residential credit, with issuance growing 60%+ year-on-year; and (iii) sector-leading cost efficiency (1.29% opex ratio). These are why FY2025’s +20.2% economic return included a +5.5% book-value gain rather than pure marks, and why the board raised the dividend ($0.70 → $0.75) for the first time since the 2023 cut — with EAD of $0.76/quarter covering the payout.

But the valuation has caught up and then some. At $22.86 against Q1’26 book value of $19.82, Annaly trades near 1.15x book — the 99.8th percentile of its own ten-year P/B history — after a +32% total return over the trailing twelve months to a five-year high. Empirically the move is a rate/spread re-rate (InterestRate factor loading −0.76; essentially no momentum load), not a fundamentally new earnings stream. The market is underwriting a benign, GSE-supported, Basel-friendly regime as a durable state. If it is, mid-teens new-money returns sustain the dividend and the premium is defensible; if the term premium backs up or the GSE conservatorship-exit process turns disorderly, the same −0.76 rate sensitivity that lifted the stock takes book value down.

Bottom line for the committee: Annaly is the highest-quality vehicle in a structurally poor category, but it is priced as such. The embedded expectation is continuity of a favorable macro/policy regime and continued double-digit economic returns — reasonable but not conservative at 1.15x book. This memo takes no position and sets no price target (see Claude’s Take above for the single labeled exception).

2. Business Overview

Annaly is an internally-managed Maryland REIT (NYSE-listed since 1997) that describes itself as “a leading diversified capital manager with investment strategies across residential mortgage finance.” The core mechanism is that of every mortgage REIT: “We use our capital coupled with borrowed funds to invest primarily in real estate related investments, earning the spread between the yield on our assets and the cost of our borrowings and hedging activities” (FY2025 10-K). It is the largest US mortgage REIT by equity (~$16B common equity; $138.5B total assets at 3/31/26), run by just 212 employees — an extraordinary asset-per-employee ratio that underpins its low-cost claim. Leadership is deep and tenured: CEO/Co-CIO David Finkelstein (ex-NY Fed MBS strategist), CFO Serena Wolfe, President/COO Steven Campbell, plus dedicated heads of Agency (V.S. Srinivasan), Residential Credit (Mike Fania), and MSR (Ken Adler).

What separates Annaly from the pure-play peers is that it runs three distinct investment groups, deliberately diversified across the housing-finance stack (capital allocation at 3/31/26):

The long-term target is 50% Agency / 30% Resi Credit / 20% MSR — a deliberate tilt toward credit and MSR from today’s agency-heavy mix. Note that Agency is ~89% of assets but only 56% of capital, because the credit and MSR legs run at far lower leverage.

How each leg makes money. The Agency book is the commodity carry trade: interest income on government-guaranteed bonds (FY25 avg asset yield 5.36%), funded with ~35-day repo (FY25 avg cost of funds 4.75%), levered ~5.7x — NLY is paid to take rate, spread, prepayment and extension risk, not credit risk. FY25 Agency net interest income was $754.9M. This leg is undifferentiated from AGNC’s. The Residential Credit leg is the genuine operating business: OBX is the largest non-bank securitizer of residential credit ($50B+ issued since inception; 29 deals / $15.2B in FY25). NLY sources whole loans through a correspondent channel (~80% of Q1’26 acquisitions), warehouses them, then securitizes — retaining subordinate/residual tranches and issuing non-recourse senior debt, capturing both net spread and the whole-loan-to-securitization “asset-creation” arbitrage at 12–15% returns. Credit quality is high (locked-pipeline WA FICO 762, CLTV 67%). The MSR leg (WA note rate 3.28%, ~250bps out-of-the-money, so near-zero prepay risk) throws off a servicing-fee annuity — FY25 net servicing income $519.3M — and, critically, acts as a partial rate hedge: MSR value rises when rates rise and prepayments slow, offsetting the mark-to-market loss on the agency book in the same scenario. This is the mechanical core of the “diversified housing finance” thesis.

Recurring vs. mark-to-market. GAAP results are dominated by non-cash marks (FY25 net interest-rate-swap losses of −$716.8M swung from +$2.1B in FY24). The recurring stream is EAD ($0.76/share in Q1’26, covering the dividend), and the truth-telling metric is economic return on book (+20.2% FY25, +1.5% Q1’26 — the 10th consecutive positive quarter). Verdict: Annaly is a three-legged levered fixed-income vehicle, but meaningfully less of a pure commodity carry machine than AGNC — two of its three legs (OBX securitization, MSR servicing) are genuine, scaled operating platforms, while the largest leg (56% agency) remains the same no-moat commodity spread trade as its peers.

3. Industry Dynamics

Analyzed through Greenwald (barriers to entry) and Marathon/Chancellor (capital-cycle) lenses, cross-reading the AGNC 2026-07-10 report, which established the agency-mREIT framing.

At its core, an agency mortgage REIT is a leveraged, publicly-traded spread account: it borrows short (repo, ~35-day maturity) against long-dated, government-guaranteed mortgage bonds and levers the net margin ~5–10x. It is a price-taker in a ~$9 trillion commodity market — one price per coupon, no product to differentiate, no customer to retain, no way to earn an abnormal margin on the asset. The entire result is a function of five exogenous variables the industry does not control: the level and shape of rates, MBS-to-Treasury spreads, prepayment speeds, funding cost/availability, and Fed/GSE policy.

This is structurally a bad business on three counts. (1) No moat, commodity asset — the MBS NLY owns is identical to the one AGNC, a bank, or a foreign central bank owns; every Greenwald test fails on the agency leg. (2) Book-value erosion over cycles — the industry periodically destroys large amounts of book value (2013 taper tantrum, 2020 COVID, 2022 rate shock); NLY’s own split-adjusted BVPS fell from ~$30 (2021) to a $19.12 trough (2023) and had only recovered to $19.82 by 3/31/26, still ~35% below its 2021 level after a banner year. (3) Reflexive external-capital dependence — these vehicles issue equity at a premium to book and stall at a discount; the premium→accretive-ATM→larger-base→lower-opex-ratio flywheel is genuine but self-reversing the instant the stock falls below book (NLY raised $2.6B via ATM in FY25 — a warm-cycle signal).

The policy regime is the supply side. Because the asset is government-guaranteed, federal policy — not corporate capex — sets the supply dynamic, and three live vectors matter: (i) the Trump Jan-2026 GSE $200B MBS-purchase directive, a materially spread-supportive technical NLY credits with boosting agency demand (and which it exploited by rotating out of richened agency into credit/MSR — optionality AGNC lacks); (ii) the conservatorship-exit tail, where a GSE exit “without sufficiently robust U.S. government support” could redefine what constitutes an agency security, an asymmetric multi-year tail over NLY’s 56%-of-capital agency book; and (iii) Fed policy (75bp of 2025 cuts lowered funding cost; QT ended Dec-2025).

The Basel Endgame reproposal is double-edged for NLY specifically. Lowering residential-mortgage risk-weights (~−30%) and eliminating the MSR capital deduction (a) pulls banks back into the MBS bid (spread-supportive; helps the agency leg) but (b) re-arms banks as competitors in MSR and resi credit — the two businesses where NLY’s edge is precisely its non-bank status (banks retreated from MSR post-2014 under punitive capital treatment; that retreat is why NLY’s platforms scaled). This nuance does not apply to pure-agency AGNC.

The two better neighborhoods. Private-label securitization (non-QM, prime jumbo, RTL) is a genuinely growing, supply-favorable pocket — gross issuance ran $79B in Q1’26, +63% YoY — and OBX is the #1 non-bank issuer; incumbency and relationships matter here in a way they never do in agency. The MSR market likewise offers ample bulk supply and rewards scale. Verdict: the agency leg (56% of capital) sits in a structurally bad, no-moat, cyclical, policy-dependent industry where book value mean-reverts destructively; but NLY straddles two better markets (44% of capital — private-label securitization and MSR) that are growing and reward non-bank incumbency. On a blended basis NLY operates in a less-bad industry mix than AGNC, and its diversification genuinely lowers through-cycle book-value volatility — but the majority of its capital still earns roughly its cost of capital over a full cycle.

4. Competitive Position

The honest null hypothesis: on the agency leg NLY has no moat — identical logic to AGNC. Agency MBS is a commodity; no product differentiation, no customer, no switching cost, no network effect. The only question is whether NLY’s scale, structure, and the OBX/MSR platforms buy a durable, quantifiable edge.

(1) Scale / cost-of-capital advantage — real but shared and modest. NLY is the largest mREIT (~$16B equity, $138.5B assets), which confers better repo counterparty terms (funded through captive broker-dealer Arcola Securities) and the lowest operating-cost ratio in the sector (FY25 G&A 1.42% of equity; a ~1.29% recent-quarter “efficiency ratio”) versus ~3%+ for externally-managed peers. Real — a 200bp opex disadvantage compounds — but shared with AGNC and DX and fully replicable (internalize the manager). Notably, on the pure agency leg, AGNC’s ~1.19% opex ratio is lower than NLY’s blended 1.42%, because NLY’s credit/MSR/servicing operations carry more headcount: diversification is a cost on the agency leg even as it is a differentiator overall.

(2) The OBX / Onslow Bay securitization platform — the strongest moat candidate, a genuine (narrow) advantage. This is what AGNC does not have. OBX is the #1 non-bank issuer of Prime Jumbo and Expanded Credit MBS ($50B+ since inception; programmatic cadence of 29 deals/$15.2B in FY25). Management calls it “not easily replicated,” and the evidence partially supports a real barrier: (a) a correspondent network of money-center banks and originators sourcing proprietary flow (~80% of Q1’26 whole-loan acquisitions); (b) operational/tech infrastructure — OBX “introduced a number of innovative deal structures, which have since been adopted by numerous other market participants” (a standard-setter, evidence of first-mover intangible value); © a repeat-issuer reputation with bond buyers that lowers execution cost and enables bespoke private deals; (d) a $1B+ GIC joint venture extending reach. In Greenwald terms this is modest economies of scale (in issuance volume) plus counterparty captivity (correspondents and bond buyers who value NLY’s programmatic issuance). Its absence would visibly deteriorate the 12–15% asset-creation returns. But it is narrow — it protects ~23% of capital, the private-label market is growing (which invites entrants), and RITM (which owns the Newrez servicer/originator) is more vertically integrated. Real, but not wide.

(3) MSR servicing scale — modest, contestable. NLY’s top-2 MSR-buyer / top-5 non-bank-servicer position gives bulk-purchasing scale and a genuine portfolio synergy (pairing MSR against the agency book as a hedge, which AGNC cannot do at scale). But the servicing itself is subserviced (outsourced), so the edge is in capital access and portfolio construction, not servicing operations — more contestable than OBX.

Verdict: NLY has no wide moat, but it is not a pure commodity book either — it sits above AGNC on the franchise spectrum. Name the types plainly: cost leadership (real, shared, replicable); a narrow economies-of-scale + counterparty-captivity advantage in the OBX platform (the one genuine, if narrow, franchise); and modest, contestable MSR scale. Roughly 44% of capital sits behind legs with some durable advantage; 56% (agency) has none. The honest verdict: a mostly-commodity book with one genuine, narrow securitization franchise bolted on — better than AGNC’s no-moat purity, but not a wide-moat compounder.

5. Growth History and Forward Opportunities

Share-count growth is dilution dressed as growth. Post the 2022 1-for-4 reverse split, NLY’s share count went from ~355M (2021) to 707M (12/31/25) to ~730M (3/31/26) — roughly doubling in under five years, almost entirely via the ATM program (127.9M shares/$2.6B in FY25; 77.9M/$1.6B in FY24). This is AUM/balance-sheet growth, not per-share intrinsic-value growth, and it is accretive only because the stock trades above book — the same reflexive flywheel as AGNC, equally reversible.

Book value per share is the truth-teller, and it has gone backwards. BVPS fell from ~$30 (2021) to a $19.12 trough (2023) to $19.82 (3/31/26) — after five years and a +20.2% FY25 economic-return year, per-share book is still ~35% below its 2021 level. The ~13% dividend has delivered cash, but per-share book compounding over the five-year window has been negative — the classic mREIT pattern where headline yield masks principal erosion. FY25’s +5.5% book gain and 10 straight positive economic-return quarters are a genuinely strong recent run, but they follow a deep hole.

The higher-quality growth is in the operating legs. Resi-credit lock volume was $7.4B in Q1’26 (+41% YoY), whole-loan acquisitions $6.7B/quarter, and the credit portfolio grew 30% QoQ to $10.3B — real volume growth in a growing market at 12–15% new-money returns; OBX securitization ran $15.2B in FY25 (+38% YoY). The MSR flow channel is scaling ($24B UPB committed in Q1’26), and the dynamic capital rotation (shifting +6pts to credit/MSR when agency richened in Q1’26) is optionality AGNC structurally lacks. Forward opportunities — the march to the 50/30/20 target, private-label market growth, ample MSR supply, Fed easing — are real but all regime/policy-dependent, and Basel relief could reverse the bank retreat that created the MSR/credit opportunity.

Verdict: mixed — low-quality headline growth (share count, AUM) wrapped around a genuinely higher-quality kernel (OBX/credit + MSR volume). The balance-sheet doubling is empire-adjacent AUM growth, accretive only while the premium-to-book persists, and it has coincided with a ~35% decline in per-share book. But unlike AGNC’s purely commodity growth, a real slice of NLY’s expansion (44% credit/MSR capital, growing toward 50%) is deployment into higher-return, franchise-protected businesses. Judge NLY as a high-yield, cyclically-managed income vehicle with one genuine growth engine (OBX) — not a per-share compounder.

6. Financial Quality

Read the right numbers. For a mortgage REIT, GAAP net income is close to noise — it swings with mark-to-market on both assets and hedges (2023 GAAP: −$3.61/share; 2025 GAAP: +$2.92) and reconciles to nothing useful. The metrics that matter are: (i) book value per share (the balance-sheet reality); (ii) economic return = (change in book + dividends) / beginning book; and (iii) earnings available for distribution (EAD), the recurring net spread that funds the dividend. On all three, Annaly’s recent record is genuinely better than the pure-play peers.

Book value has stabilized and is recovering — the key contrast with AGNC. Split-adjusted book value per common share fell from ~$31 (2021) to a ~$19 trough (2023), then recovered: ~$19.15 (YE2024) → ~$20.20 (YE2025) → $19.82 (Q1’26, down 1.9% on the Q1 rate shock, then reported up ~4% quarter-to-date by late April). Critically, FY2025’s +20.2% economic return included a +5.5% book-value gain (FY2025 10-K) — real value creation, not just dividends paid out of eroding capital. This is the sharpest distinction from AGNC, whose comparable 2025 economic return was almost entirely mark-to-market recovery against a still-declining net-spread stream. (Fact: 10-K, Q1’26 transcript. Interpretation: the diversified book is why marks and carry both held.)

Earnings cover the dividend — and the dividend was just raised. EAD (ex-PAA) was $0.76/share in Q1’26, up $0.02 sequentially and above the $0.70 dividend; the EAD return on average equity was 13.71% in FY2025 (vs 13.28% FY2024, 14.47% FY2023). On the strength of that coverage the board raised the quarterly dividend to $0.75 for Q2’26 (announced June 10, 2026) — the first increase since the 2023 cut to $0.65. Annualized $3.00 is a ~13.1% yield on price and ~15% of book, with EAD of ~$3.04 annualized providing thin but real cover (~99–101% payout on EAD). (Fact.) The honest caveat: even a covered 13% distribution on a book that compounds in the low-to-mid-single digits means a meaningful slice of the “yield” is the return of the investor’s own capital — the total economic return, not the headline yield, is the number that matters.

Spread and funding. Net interest margin (ex-PAA) was 1.71% in Q1’26 and net interest spread 1.42%, aided by a 30bp decline in average repo cost to ~3.9% and higher TBA dollar-roll specialness, partly offset by lower swap income as SOFR fell. New-money returns are mid-teens in agency and 12–15% in residential credit — above the ~13% dividend, which is what makes accretive ATM issuance and the dividend raise defensible. (Fact: Q1’26 transcript.)

Leverage and liquidity are conservative for the model. Economic leverage was 5.7x at Q1’26 (vs peers often 6–8x), the economic capital ratio ~14%, and the firm held $7.4B of unencumbered assets (incl. ~$5B cash and unencumbered agency MBS) plus ~$1.6B of MSR fair value pledged but undrawn — ~$9B, ~55% of total capital, in financeable liquidity. Operating efficiency (1.29% of equity) is among the lowest in the sector despite running three businesses — a real, if modest, cost-leadership edge. (Fact.)

Quality-of-earnings flags. (1) GAAP EPS is unusable; do not anchor on the 6.98x “P/E” — it is an artifact. (2) The ~13% yield includes return of capital; judge the payout against EAD and economic return, not price. (3) MSR marks are model-derived (5.94x multiple, level-3 fair value) and can move with rate/prepay assumptions. (4) The 2020 and 2023 GAAP losses were mark-driven, not operating. Verdict: for its category, financial quality is above-average and improving — book is growing, earnings cover the (rising) dividend, leverage is conservative, and the diversified legs smooth the marks. But the underlying economics remain those of a levered spread book, where a covered double-digit yield still delivers only low-double-digit total economic returns in good years and negative ones in bad.

7. Capital Allocation

The ATM issuance machine — accretive, but only in the warm cycle. Shares outstanding roughly doubled 349.6M (FY2020) → 707M (FY2025), and in FY2025 alone NLY raised $2.6B of common equity through its ATM program “at accretive levels” (plus $510M in Q1’26). Selling stock at/above tangible book (~$22.2 TBV vs. issue prices ~$19–23) is mildly book-accretive — the premium-to-book is the fuel — but it is dilutive to legacy per-share economics the instant the stock slips below book, and its purpose is to grow the absolute asset/fee base and preserve REIT scale, not per-share intrinsic value. The buyer is underwriting management’s ability to redeploy fresh equity above the dividend — true in 2025’s wide-spread regime, false in a spread shock. (Fact + interpretation.)

Dividend policy — history, and return OF vs. ON capital. Split-adjusted annual dividends fell from ~$4.17 (2020) to a ~$2.86 trough (2024) — the quarterly was cut to $0.65 in 2023 — before the $0.75/quarter declared June 2026 (run-rate $3.00, still ~28% below the 2020 level). The “raise” is a step off a deep trough, not a return to prior payout. Over 2020→2025, TBV/share fell ~36% (~$34.7 → ~$22.2) while the company paid out roughly $18–19/share cumulatively — a large slice of the historical “dividend” was return of capital, book-value erosion consuming a material fraction of the distributions (the classic mREIT critique). However, the current dividend is genuinely covered: FY2025 EAD was $2,024.9M (≈$3.1/avg common share) and Q1’26 EAD $0.76 covers the $0.75 payout. So the return-of-capital problem is a multi-year book-trend statement, not a current-coverage statement.

Preferred stock has quietly become expensive capital. Four series (~$1.8B total): Series F 6.95%, G 6.50%, I 6.75% — all now past their fixed-to-floating reset and floating at SOFR + ~4.2–5.0% (≈8.5–9.3% all-in) — plus a new Series J 8.875% fixed (2025, NLY’s first sizeable non-rated preferred in years, signalling the market demanded a fat coupon). A small slice of the stack, but the floaters are now a high-cost, rate-sensitive liability.

Buybacks are optics, not policy. A $1.5B repurchase program runs through 12/31/2029, replacing a prior $1.5B program that expired 12/31/2024 with essentially nothing bought — and zero shares were repurchased in FY2025. NLY is a structural net issuer that does not repurchase even when a below-book buyback would be the textbook accretive move; treat the authorization as a defensive option, not shareholder-return intent.

M&A — the 2021 simplification. NLY sold its Commercial Real Estate business to Slate Asset Management (~$2.33B) and its Middle Market Lending portfolio to Ares, exiting non-core credit to concentrate on the three housing-finance pillars. Value-neutral-to-positive: it removed businesses where NLY had no edge and freed capital for the higher-ROE resi-credit/MSR rotation — a sensible simplification, not a value-creating roll-up.

Compensation — well-designed metrics, thin skin in the game. CEO Finkelstein’s total comp was $18.6M (2025) on a $1.0M base flat since 2020 (~95% variable). The scorecard and PSUs run on Absolute and Relative Tangible Economic Return, EAD Return on Equity, and relative TSR — i.e., comp is tied to the right mREIT value driver (per-share economic return), not gross AUM or leverage, materially better-aligned than a size-driven or externally-managed fee structure. The weakness is ownership: all executives and directors together own <1% (~0.15%) of shares, dwarfed by BlackRock (7.8%) and Vanguard (6.7%) — large variable pay, little personal equity at risk.

Verdict: adequate, disciplined-within-a-flawed-model — not value-creating. Management allocates capital competently for what NLY is (economic-return-tied comp, sensible 2021 simplification, currently-covered dividend, book-accretive ATM while the premium holds). But the model is structurally a net-issuer treadmill — shares doubled, TBV/share fell ~36%, the buyback sits unused, and the preferred has turned expensive. This is intelligent stewardship of a no-moat levered-spread vehicle — capital preservation and scale, not per-share compounding. The negative verdict belongs to the model, not the allocators.

8. Changes and Headwinds — Last Two Years

Strategic rotation into higher-ROE housing credit. NLY has steadily shifted capital toward Residential Credit (OBX, ~23%) and MSR (~21%), with agency still the core (~56%), scaling the OBX shelf into a top non-bank correspondent/aggregator and repeat prime/non-QM securitizer. This diversifies the return stream away from pure rate/spread leverage and leans into the one area where NLY has a nascent defensible edge; the MSR book adds a structural rate hedge. The agency book itself grew +$22.3B YoY to $92.9B in 2025, funded by the ATM — a benign-regime scale bet. Modestly thesis-strengthening.

The dividend raise ($0.75/quarter, June 2026, EAD-covered) is the first increase in years and signals management confidence — mildly positive, though off a trough. Leadership: CLO Anthony Green retired 1/1/2026; Finkelstein, Wolfe and Campbell remain — continuity at the top; comp scorecard simplified in response to investor feedback. Neutral.

The regulatory backdrop is the dominant swing factor — supportive today, fragile. The GSE-privatization/FHFA-reform agenda and the Basel Endgame reproposal both bear directly on agency-MBS demand and spread stability. The administration’s pro-housing-finance posture (the $200B GSE MBS-purchase directive, housing-affordability agenda, and a bank-capital reproposal that pulls banks back into the MBS bid) currently tightens and stabilizes agency spreads — the exact regime in which NLY’s levered book and ATM flywheel work. But it is benign now and fragile: a GSE-reform misstep, a re-widening of spreads, or a rate/vol shock would compress book value and the price-to-book premium simultaneously (the classic mREIT double-hit), and the outcome is an external variable NLY cannot control. No material litigation overhang was identified.

Headwinds to weigh against the improvement: floating preferred now costing ~9%; multi-year book-value erosion (−36% TBV/share since 2020); structural dependence on a wide-and-stable-spread regime and on a ≥book share price to keep the ATM accretive; and no per-share compounding track record. Verdict: net neutral-to-mildly-strengthening on the margin, within an unchanged fragile model. The rotation, OBX scaling, covered dividend raise, and supportive regime genuinely improve the near-term earnings picture and diversify the return stream — but none of it changes what NLY is. The changes strengthen the cyclical case; they do not create a structural one.

9. Risk Analysis

Annaly’s risks are dominated by rates, spreads, and policy — the exposures embedded in any levered agency book — modulated by the diversification that is its differentiator.

Catastrophic-loss risk is low in the “total wipeout” sense (agency MBS carry an implicit government guarantee on principal), but a permanent impairment of 30–50% of book is a real tail in a rate/spread shock, as 2022 demonstrated. The asymmetric danger today is buying the premium: paying 1.15x book for an asset whose book can fall 20%+ in a bad year means price drawdowns compound the book drawdown. Verdict: rate and policy risk are structural and unavoidable; the incremental, self-inflicted risk at today’s price is the valuation itself.

10. Valuation

Mortgage REITs are valued on price-to-book, full stop. Multiples on GAAP earnings are meaningless (the “P/E” of 6.98x is an artifact of mark-driven net income); the market prices these vehicles as a multiple of the liquidation-relevant book value, with the multiple flexing on expected forward economic return relative to the cost of equity.

Annaly trades at ~1.15x book — the richest in its history. Against Q1’26 book of $19.82, the $22.86 price is ~1.15x (≈1.11x if book recovered ~4% into Q2 as management indicated). On an own-history percentile framework, that P/B sits in the 99.8th percentile of the last decade — Annaly has spent most of its history at 0.85–1.00x book, frequently at a discount. The composite valuation percentile is 83rd; the P/S 80th. However you anchor the absolute book figure (aggregators show ~1.0x on a differently-computed equity base; the company’s common book of $19.82 is authoritative and yields ~1.15x), the direction is unambiguous: this is the most expensive Annaly has ever been relative to its own assets.

Embedded-expectations analysis — what must be true at 1.15x book? A ~15% premium to book, for a vehicle whose assets are marked to market daily, is the market capitalizing an excess forward economic return above the cost of equity. Roughly: if the cost of equity is ~11–12% and Annaly can sustain a ~13–14% economic return on book (mid-teens new-money returns, dividend covered, book stable-to-growing), a ~1.1–1.2x multiple is internally consistent. So the price is not absurd — it is fair if the good times continue. The embedded assumptions are: (i) the benign-rate, GSE-supported, Basel-friendly regime persists; (ii) spreads stay range-bound (no 2022 repeat); (iii) the diversified book keeps generating book-value gains, not just covered dividends; and (iv) the GSE conservatorship exit, when it comes, is orderly. Each is plausible; none is conservative.

Scenario framing (economic return on book, illustrative, not a forecast):

• Bear (rate/spread shock à la 2022, or disorderly GSE exit): book falls 15–25%, the premium compresses to a discount, total return deeply negative — a repeat of the 2022 −46% drawdown is the template.
• Base (regime continuity): ~11–14% economic return, dividend covered at $3.00, book roughly flat-to-+MSD, price tracks book with the premium slowly bleeding toward ~1.0x — a mid-single-digit-to-low-double-digit total return dominated by the dividend.
• Bull (rates grind lower in orderly fashion, spreads tighten further, Basel demand materializes): book gains high-single-digits, EAD supports another dividend bump, the premium holds or expands — high-teens total return.

Relative value. Annaly’s ~1.15x is richer than most pure-play peers on absolute book multiple but is arguably more deserved given the diversified, book-growing model and covered/rising dividend — the market is paying up for quality within the category. Versus AGNC (1.32x book, pure agency, return-of-capital dynamics), Annaly is the more defensible premium; versus discounted names (DX, IVR, ARR near/below book), Annaly is the “pay-up-for-the-operator” choice. Verdict: not mispriced so much as fully priced. The valuation embeds regime continuity and continued double-digit economic returns; there is little margin of safety in the multiple, and the asymmetry (book downside in a shock vs. premium already paid) is unfavorable at 1.15x. No price target (see Claude’s Take).

11. Variant Perception

Consensus view. Annaly is the blue-chip mortgage REIT — the largest, most liquid, best-diversified, internally-managed name with a covered, just-raised ~13% dividend and a supportive policy backdrop (GSE MBS purchases, Basel relief). Sell-side is constructive (e.g., Piper Sandler Overweight, PT raised to $25 in July 2026), framing it as the highest-quality way to own the mortgage-rate recovery. The +32% trailing-year return and five-year-high price reflect this consensus embrace.

The strongest bull case. The three-legged model is structurally superior and the market is still under-appreciating the durability of the improvement. MSR provides a real internal rate hedge; OBX is a genuine, compounding fee franchise (largest non-bank resi-credit securitizer, 60%+ issuance growth) that deserves a franchise multiple, not a spread-book multiple; the Basel reproposal (residential-mortgage RWA −30%, MSR capital relief) structurally increases bank demand for exactly Annaly’s assets; and with new-money returns mid-teens against a ~13% cost of equity, book value can compound rather than erode. In this view, 1.15x book is cheap for a business transitioning from “levered bond fund” to “diversified housing-finance platform,” and the dividend has further upside.

The strongest bear case. Strip the narrative and Annaly is still a 5.7x-levered bond fund with an InterestRate loading of −0.76, trading at its richest-ever premium after a rate-driven re-rate. The “book-value gain” of 2025 was a spread-recovery cycle that mean-reverts; OBX and MSR are real but small (together ~44% of a book whose returns are still dominated by the agency leg and the level of rates); and the +32% return is a warning, not a green light — you are buying high after the easy re-rate, with the premium itself as downside. The 2022 −46% drawdown shows what the same book does when the regime turns, and paying 1.15x rather than 0.85x book means ~35% of incremental downside before you even touch the asset marks.

The 3–5 assumptions that decide it: (1) Does the benign-rate/spread regime persist, or does the term premium back up? (2) Is the 2025 book-value gain repeatable, or a one-cycle spread recovery? (3) Does OBX/MSR diversification justify a structural re-rating of the multiple, or is it a smoothing overlay on a spread book? (4) Is the GSE conservatorship exit orderly? (5) Does the premium hold, or mean-revert toward the historical ~0.9–1.0x?

Falsification. Bull is falsified if EAD slips below the dividend, book resumes eroding, or a rate/spread shock reprices the book (premium → discount). Bear is falsified if book continues to compound through the next rate cycle, OBX/MSR earnings visibly de-correlate returns from the level of rates, and the dividend rises again while the premium holds. The factor tape sides with caution: the +32% is a −0.76-rate-beta re-rate with no momentum load — a levered bond proxy that has already priced the good news, not a self-sustaining trend. Verdict: consensus is right about the quality and likely too sanguine about the price.

12. Fact vs. Interpretation

13. Open Questions

1. How much of the 2025 +5.5% book-value gain is repeatable structural improvement vs. one-cycle spread recovery? (Needs another rate cycle to answer.)
2. What multiple does the market ultimately assign OBX/MSR earnings — spread-book or franchise? Does management ever break out segment ROEs cleanly enough to force the re-rate?
3. What is the concrete GSE conservatorship-exit path and timeline, and how would a privatized guarantee structure reprice the agency basis?
4. How far can ATM issuance scale before it saturates demand / stops being accretive, and how disciplined is management if the premium compresses toward book?
5. Where do non-QM 2024–2025 vintage delinquencies go if the consumer weakens under the energy-price shock — and is the resi-credit book’s ~1.4% D90+ a floor or a start?

14. What Must Be True

Bull case — what must be true: (i) the benign-rate, GSE-supported, Basel-friendly regime persists and spreads stay range-bound; (ii) EAD continues to cover the (rising) dividend with mid-teens new-money returns; (iii) OBX/MSR diversification keeps generating book-value gains, not just covered dividends, de-correlating returns from the level of rates; (iv) the market maintains/expands the ~1.15x premium. Falsification test: if, over the next 12–18 months, book value resumes a sustained decline OR EAD falls below the dividend OR the P/B mean-reverts toward ~0.95x while fundamentals are unchanged, the “structural re-rate” thesis is wrong.

Bear case — what must be true: (i) rates/spreads mean-revert wider (term premium backs up), taking book down as in 2022; (ii) the 2025 book gain proves to be a one-cycle recovery; (iii) the premium compresses toward the historical ~0.9–1.0x. Falsification test: if book value compounds positively through the next rate-vol episode, the MSR/OBX legs visibly cushion a spread shock (book down < the agency-only peer set), and the dividend is raised again while the premium holds, the “levered bond proxy at peak price” thesis is wrong.

Synthesis: The disagreement is entirely about price, not quality. Both sides concede Annaly is the best operator in the category. The bull pays 1.15x for a business he believes is structurally re-rating; the bear waits for the premium to compress, confident the underlying asset is still a levered spread book. The factor evidence (−0.76 rate beta, no momentum load, five-year-high price) favors patience.

APPENDIX A — Standard Diligence Questionnaire

Annaly Capital Management, Inc. (NYSE: NLY) — as of 2026-07-11

Supplemental to the analysis above. Fact / Interpretation / Assumption labels applied where material. Where a question does not map to a mortgage REIT, the correct sector analog is given.

General

What thoughtful questions have other investors asked? The recurring debates: (1) Is the diversified three-legged model (Agency + OBX credit + MSR) a structural re-rating story or just a book-value-volatility dampener on a levered spread book? (2) Does the ~13% dividend represent real income or return of capital? (3) At ~1.15x book (richest-ever), is the premium deserved and durable, or has the rate-driven rally priced in the good news? (4) How does NLY compare to AGNC (cleaner/cheaper agency) and RITM (deeper vertical integration)? (5) What does GSE conservatorship exit do to the agency basis?

Cyclicality & Earnings Nature

• Cyclical high or low? Interpretation: mid-to-late cycle. Economic returns are running hot (+20.2% FY25) off a benign wide-and-stable-spread regime; book value is recovering from the 2023 trough but still below 2021. Earnings power is above mid-cycle, not at a trough.
• Driven by external environment or internal action? Overwhelmingly external — rates, MBS spreads, prepay speeds, Fed/GSE policy. Internal action (capital rotation, OBX scaling, hedging discipline) modulates the outcome but does not drive it.
• How stable are revenues? GAAP revenue is highly unstable (mark-to-market). The recurring EAD/net-interest stream is more stable (NIM 1.71%, spread 1.42%) but compresses in funding-cost spikes.
• Market size / direction? Agency MBS is a ~$9T market (mature, policy-anchored); private-label securitization and MSR (NLY’s growth legs) are growing (+63% YoY gross private-label issuance).

Business Quality & Competitive Moat

• Industry more or less competitive? Agency: perpetually competitive commodity (no barriers). Resi credit/MSR: NLY has a nascent non-bank incumbency edge, but Basel relief could re-arm banks as competitors.
• How profitable (ROIC/ROE)? Standard ROIC/ROE are not meaningful for a levered spread book. Use economic return on book (FY25 +20.2%) and EAD ROE (13.71% FY25). Through-cycle, the agency leg earns roughly its cost of capital.
• Barriers to entry? None on agency; a narrow economies-of-scale + counterparty-captivity barrier around the OBX securitization platform (#1 non-bank issuer, “not easily replicated”).
• Easily understood? The structure is simple (levered spread + two operating legs); the risk (convexity, hedging, MSR modeling) is not.
• Undermined by foreign low-cost labor? No (capital-markets business). Do brands matter? Only OBX’s repeat-issuer reputation with bond buyers. Switching costs? None on agency; some correspondent stickiness on OBX.

Financial Condition & Balance Sheet

• Assets not fully recognized / off-balance-sheet? MSR is level-3 fair-value (model-derived, 5.94x multiple); OBX securitizations are largely non-recourse VIEs. Book value is broadly marked-to-market, so hidden value is limited.
• How conservative is the accounting? Marks are fair-value and audited; the honest issue is model risk (MSR, non-agency credit), not aggressiveness. GAAP EPS is noise; EAD/TBV are the run-rate.
• CapEx-hungry? N/A — the analog is asset acquisition funded by repo + equity issuance, not physical capex. Leverage 5.7x economic / 7.2x GAAP.

Capital Allocation & Management

• FCF generation / use / philosophy? The analog is EAD ($2.02B FY25), distributed as the dividend (REIT payout requirement). Philosophy: distribute EAD, grow the base via ATM when accretive, hold conservative leverage/liquidity.
• Significant acquisitions? The 2021 divestitures (sold CRE to Slate ~$2.33B; MML to Ares) — simplification, not roll-up.
• Buying back shares? No — $1.5B authorization sits unused; NLY is a structural net issuer.
• Issuing shares to insiders / compensation? ATM issuance is to public markets, not insiders. CEO comp $18.6M (~95% variable, tied to economic return/EAD ROE/relative TSR — well-designed). Insider ownership <1%.
• Motivations of management? Metrics aligned to the right driver (economic return), but thin personal equity at risk.

Valuation & Market Data

• ADR / MLP / K-1? No — a US REIT (1099-DIV; dividends largely ordinary income, not qualified). Dividend policy? $0.75/quarter (raised June 2026), ~13.1% yield, EAD-covered.
• How profitable? See economic return / EAD ROE above.
• Net income diverging from cash flow? Constantly — GAAP net income is mark-driven and unrelated to distributable cash (EAD). Anchor on EAD and economic return, never GAAP EPS/“P/E.”

Risks & Downside

• What would cause the stock to decline? A term-premium back-up / rate-vol shock (InterestRate loading −0.76), agency spread widening, a disorderly GSE conservatorship exit, EAD slipping below the dividend, or simple premium compression from 1.15x toward the historical ~0.9–1.0x book.
• Catastrophic loss risk? Low in the “total wipeout” sense (agency principal is government-guaranteed), but a 30–50% book impairment in a rate/spread shock is a real tail (2022 was −46% peak-to-trough).
• Chance of total loss? Very low — the agency book is money-good on principal; the risk is drawdown and dividend cuts, not zero.

Recent News & Events

• Environment changed recently? Yes — the Jan-2026 GSE $200B MBS-purchase directive tightened the agency basis; the June-quarter Middle-East energy shock lifted the term premium (book −1.9% Q1’26); the Basel reproposal is a supportive-but-double-edged technical.
• Significant acquisitions / accounting changes? None recently (LDTI/fair-value framework stable). Recent changes? Dividend raised to $0.75 (6/10/26); Series J 8.875% preferred issued (2025); CLO Anthony Green retired (1/1/26); capital rotating toward the 50/30/20 target.

APPENDIX B — Source Appendix

Annaly Capital Management, Inc. (NYSE: NLY) — as of 2026-07-11

Primary sources first. Facts in the memo trace to these; management commentary is treated as hypothesis and validated against filings/financials/external data.

Primary — SEC filings (EDGAR, CIK 0001043219)

• FY2025 Form 10-K (filed 2026-02-12; nly-20251231.htm). Business overview, three-segment detail, capital allocation, economic return (+20.2% incl. +5.5% book gain), EAD ($2,024.9M), G&A (1.42% of equity), leverage (5.6x YE), ATM issuance ($2.6B / 127.9M sh), $1.5B buyback authorization (zero used), preferred series F/G/I/J, portfolio ($92.9B agency). https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001043219
• Q1’26 Form 10-Q (filed ~2026-04-29; nly-20260331.htm). Book value/share $19.82, capital allocation 56/23/21, resi-credit portfolio $10.3B (+30% QoQ), loan locks $7.4B (+41% YoY), MSR $4.2B, leverage 5.7x.
• FY2022 Form 10-K (filed 2023-02-16; nly-20221231.htm). 1-for-4 reverse stock split effective 9/23/2022.
• FY2021–FY2024 Form 10-Ks (2022-02-18, 2023-02-16, 2024-02-15, 2025-02-13). Multi-year book value, dividend, segment history.
• DEF 14A proxy (2026). Executive compensation (CEO $18.6M, ~95% variable), incentive metrics (Absolute/Relative Tangible Economic Return, EAD ROE, relative TSR PSUs), insider ownership (<1%), board.
• 8-K, June 10–11, 2026 — quarterly common dividend raised $0.70 → $0.75.
• 8-K earnings releases (1/28/26, 4/21/26, and prior quarters) — EAD, book value, economic return, dividend declarations.
• Form 4 corpus (2025–2026) — insider transactions: routine grant/vest/sell only; CEO sales 10b5-1 (plan 5/16/2024); no open-market (code P) buys.

Primary — Company disclosures

• Q1 2026 earnings call transcript (2026-04-22). CEO Finkelstein / CFO Wolfe / Co-CIO Fania. Book value $19.82, economic return +1.5%, EAD $0.76, leverage 5.7x, repo 3.9%, NIM 1.71%, MSR 5.94x multiple / 3.3% WAC / 4.2 CPR, OBX 8 securitizations/$4.7B, capital rotation 38%→44% credit/MSR, “up 4% QTD” book (late April), new-money returns mid-teens agency / 12–15% credit, 50/30/20 long-term target. Source of record: ROIC.ai transcript tools.
• Q1 2026 Investor Presentation & Financial Supplement (company IR — Presentations). Referenced for portfolio, hedge, and liquidity detail.

Market & quantitative data

• Public market data — share price and 5-year split/dividend-adjusted price history (5yr low ~$9.07 Oct-2022, high ~$22.99 Jul-2026; 52-week range $17.07–$22.99); trailing valuation percentiles (P/B ~99.8th of own decade history — richest-ever), dividend yield ~13%.
• Aggregated fundamentals (income statement, balance sheet, per-share, enterprise value, valuation multiples, FY2020–FY2025) reconciled to the 10-K/10-Q; company-reported common book value ($19.82) used as authoritative.
• Factor/risk data — factor loadings (Mortgage-REIT, Value, DividendYield, LowVol positive; Quality negative; InterestRate −0.76; ~zero Momentum) and risk-adjusted track record (1-yr +31.5%, 3-yr +22.0% ann., 5-yr +4.8%, 10-yr +6.0%, max drawdown −60%).
• Recent news — dividend raise (6/10/26); Piper Sandler Overweight, PT $25 (7/2/26).

Method notes

• No BUY/SELL or price target in the memo body; the single labeled exception is Claude’s Take.
• GAAP EPS treated as noise; economic return, EAD, and tangible/common book value are the run-rate metrics.
• Management commentary validated against filings, financials, and external data. Third-party aggregator/factor data labeled and reconciled; the filing wins on any material discrepancy.