The repo book is the first thing I read in an mREIT filing, and most people skip straight to the dividend

My brother would call this whole conversation fake real estate and change the subject, but the repo maturity clustering point is the one I had to learn the hard way watching NLY in Q4 2022 when their swap roll and repo window landed inside six weeks of each other and book value dropped almost two dollars before the dividend moved at all.

The thing I don't see people track is the counterparty concentration inside that repo book. A lot of these 10-Ks will show you the maturity buckets clearly enough but bury the fact that three dealers represent 60-plus percent of the outstanding balance. When one of those dealers tightens haircuts, the effective funding cost jumps before the rate itself does, because you're suddenly posting more collateral against the same assets. That collateral call is a liquidity event, not just a margin event, and it shows up in cash and unencumbered assets, not in the repo rate line.

Do you know what percentage of the repo is with the top three counterparties in the filing you're looking at?

The manager's framing is clean and the math holds. The assumption doing the most work in your $0.24 figure is that the asset yield is static across the stress period, which is reasonable for a fixed-coupon agency book but starts to drift once you factor in prepayment. A 50bp funding cost spike that also triggers a refinancing slowdown extends duration on the asset side, so the mismatch window widens at exactly the wrong moment. You are not just absorbing a margin compression, you are absorbing it for longer than the original duration assumption implied.

The point about the swap-repo cluster is where I think most analysts underweight the execution dimension. The P&L hit from a compressed spread is modelable. The liquidity dimension of rolling repo in a quarter where counterparties are also managing their own balance sheets, and where the swap desk is simultaneously marking against you, is harder to model and faster to move. Haircut expansion on the collateral side is the variable I see skipped most often in that scenario. A counterparty that moves the haircut from 2 percent to 4 percent on agency MBS is effectively a margin call without calling it one, and it hits the leverage ratio before the income statement shows anything.

On the hedge coverage question: the disclosure is rarely granular enough to tell you the exact swap maturity distribution against the repo bucket breakdown, so the 10-K table usually needs to be read alongside the interest rate sensitivity table in the same filing and cross-referenced with the maturity schedule in the notes. If a manager is only disclosing notional swap coverage as a percentage of the repo book without bucketing it by maturity, that gap in disclosure is itself a signal worth noting.

What is the collateral haircut disclosure like in the filing you are working through, and has it shifted quarter over quarter?