Annaly hedged 100% and still lost 11% of its book. What does the 16% yield pay for?

Hedged 100%, down 11.4%: the risk Annaly cannot hedge away

Annaly Capital Management ($NLY) ran the textbook defence in the quarter to September 2026 and still lost $2.29 a share of book value. Book value per common share fell from $20.15 to $17.86, an 11.4% drop; with the $0.75 dividend added back, the economic return for the quarter was -7.6%. The GAAP result was a loss of $1.36 a share, while earnings available for distribution (EAD, the cash-style profit that funds the dividend) came in at $0.80, five cents above the payout.

Net losses on investments reached $4.72 billion; gains on derivatives recovered $3.07 billion, about two-thirds. The hedge ratio ended the quarter at 100%, up from 97%, and economic leverage was a modest 5.9x (5.6x in June). The Federal Reserve raised rates to 3.75%-4% on 16 September 2026, its first hike in more than three years, and the 10-year Treasury yield reached 5.23% on 25 September, its highest since 2007.

Annaly's own sensitivity table explains why 100% hedged still meant -11.4%. As of 30 September 2026, a quarter-point rise in interest rates would cut net asset value by just 1.8%; a quarter-point widening in mortgage spreads would cut it by 10.1%. Swaps and Treasury futures neutralise the first risk. Nothing on the balance sheet neutralises the second, and Agency MBS spreads widened 15-25 hundredths of a point in the quarter. It is a landlord insured against fire, not against the neighbourhood losing value.

Peers with more leverage lost more: AGNC's tangible book fell 16.9% to $7.13 and Dynex's book fell 15.4% to $10.91. Annaly's -11.4% is the best of the three, yet the stock dropped 19.2% in three months against a 2.7% gain for the S&P 500.

$757 million of new shares sold near $22, a book value of $17.86

Annaly raised $757 million through its at-the-market program during the quarter, lifting the share count to 784.7 million at 30 September 2026 from 750.6 million in June and 681.1 million in September 2025 - up 15.2% in twelve months. The new shares went out at roughly $22 apiece, above the June book of $20.15; 90 days later the book behind each one is $17.86.

Common dividends payable were $588.5 million for the quarter against EAD to common of $607.8 million - coverage of 1.03x, a cushion of $0.05 a share. On the other side of the ledger, all $442.5 million of the 6.75% Series I preferred stock was redeemed on 1 October 2026, trimming a preferred dividend line that ran at $41.3 million a quarter.

Insiders have been sellers, not buyers: 0 purchases and 9 sales worth $7.0 million in the twelve months to October 2026, including 50,000 shares sold by the CEO at $22.89 on 3 August 2026. At $18.30 the stock sits at 1.02x the new book value, 60 cents above its 52-week low of $17.70.

Why the model says $23.62 - and why that is 1.32 times a shrunken book

The FPI fair value stands at $23.62, which puts the stock 29.1% below fair value. The model blends a discounted cash flow (50%), mortgage-REIT peer multiples (30%) and analyst consensus (20%). Annaly's numbers flatter the first two legs: a P/E of 4.7x against an industry median of 10.9x, a return on equity of 18.4% against 8.2%, a dividend yield of 15.85% against 10.61% and a net margin of 40.5% against 19.7%. An annualised EAD of $3.20 a share on an $18.30 stock is a 17.5% earnings yield.

Other inputs argue the opposite way: a free cash flow yield of -2.28%, interest coverage of 1.57x, debt to equity of 7.38x and an EPS trend of -9.4%. A mortgage REIT's cash stream is simple arithmetic: spread times leverage times equity. Equity per share shrank 11% in 90 days, so the stream a DCF discounts shrank with it - and a daily-updated model digests that gradually. At $23.62 the fair value equals 1.32x the new book, against 1.17x the June book. The guide to valuing REIT stocks explains why book value matters more here than for an ordinary company.

Street vs model: a $23.50 median set when book was $20.15

The consensus rating is buy: 16 buy, 10 hold, 2 sell. Price targets as of 6 October 2026 cluster tightly - low $23, median and mean $23.50, high $24 - just $0.12 (0.5%) below the FPI fair value, with both around 29% above the price. On paper, model and Street agree.

The disagreement is between old targets and new ones. BTIG set $21 on 6 October 2026, the only target in the FPI database in the past month; secondary reports put Piper Sandler at $21.50 (from $25) on 9 October and JPMorgan at $19 (from $24) on 16 October. Those three equal 1.06x to 1.20x the new book; the $23.50 median equals 1.32x. Analysts price mortgage REITs as a multiple of book and re-mark after the print; the consensus leg of the model and the median still carry the $20.15 book. The real debate sits between $19 and $21.50, not around $23.50.

Three numbers to watch: a quarter-point of spread, $0.75 and $17.86

The upside is a spread story, not a rate story. A quarter-point tightening in mortgage spreads adds 10.4% to net asset value, about $1.86 a share; if EAD holds at or above $0.75 and the Fed pauses, the stock has a path toward the $21-21.50 of the newest targets. The downside is the mirror: another quarter-point of widening takes 10.1% off book, about $1.80 a share, and EAD slipping below $0.75 would remove the $0.05 cushion - with further share sales at or below book turning dilutive. The 16% yield is the payment for carrying that leveraged bet.

• Book value in the next print (earnings expected 10 February 2027, not confirmed) against $17.86.
• EAD against the $0.75 dividend; the net interest spread (asset yield minus funding cost) was 1.42% in the quarter to September 2026, down from 1.50%.
• The share price against book ($17.86), the 50-day average ($21.11) and the 200-day average ($22.14); the 10-year Treasury yield against 5%.