Holding an agency mREIT through two dividend cuts in 14 months and how to decide that in advance

The thing worth tracking is whether leverage was moving in the same direction as book during that window. If management was actively cutting the repo book and letting the portfolio shrink rather than defending yield, that usually tells more than spread data alone, because it means they were buying time rather than gambling on a rate reversal. What was the CPR doing on the underlying pool through Q3 2020? That is the number worth pulling before the second cut, and it rarely gets discussed clearly in the supplement.

Repo haircut trends told me more than dividend coverage ever did.

Tangible book per share reported quarterly is almost always stale by the time you read it. The more useful read is the implied book from secondary offerings, because management will not sell new equity below book if they believe book is real. A small offering in June 2020 priced at a level that said the board thought the floor was holding, and that is the sort of signal worth sitting still on. The failure case is just as instructive. Apply the same logic to a different name in 2018 and the offering price turns out to be optimistic by about 19 percent, which costs roughly eight months of recovery time. So the signal stays useful, just never on its own.

Honestly the framing of "is book intact" is probably the wrong question to be asking in real time. What I'd want to know is whether management has been through a dislocated rate environment before and actually survived it, not just managed through a benign one. The guys who were running books in 2008 and 2009 and still had the same shop in 2012 are a different animal. I'd rather own a smaller yield from a team with that kind of scar tissue than chase 11 or 12 percent from a newer operation that's never seen repo markets freeze. The second cut in October is the part that would've pushed me out, not because of the income but because two cuts that close together usually means they were slow to right-size the leverage in April rather than responding to something genuinely new. That lag tells you something about the decision-making speed inside the operation.

The part that gets missed is the hedging book, specifically whether the duration gap was being managed with swaps or swaptions and how far out of the money those positions were sitting when vol spiked. An agency book can look clean on paper, but if a two or three year duration gap was carried into March 2020 with swaption strikes nowhere near where rates actually moved, realized losses hit equity fast and reported book value was already stale before the 10-Q landed. Reviewing disclosures across several names from that period, the gap between what the hedge book actually offset versus what management described in investor presentations was often uncomfortable. A second cut in October 2020 is interesting because by then the acute spread blowout had already reversed, which suggests management was dealing with the liability side or possibly CPR acceleration on the MBS pool more than the market stress itself. That prepay speed issue is the one that catches people, because when rates dropped as hard as they did in 2020, agency premiums got destroyed faster than the dividend math could adjust.

Thirty-one percent total return through two cuts is actually a decent outcome for an agency name in that window, so the stubbornness might have been the right call even if the reasoning was murky. What I've found, sitting mostly on the passive side with long holds, is that the second cut is less informative than the gap between the first cut and management's commentary about CPR assumptions. If they were still anchoring to prepayment speeds that the market had already priced out of existence, that told me more about whether they understood their own book than any leverage ratio did. Annaly and AGNC both had that problem in different degrees during 2020 and watching how long it took each of them to revise CPR guidance publicly was basically a credibility test. What I'd want to know from you: when the first cut came in April 2020, what were they saying about their hedging duration relative to asset duration, and did that number move before or after book value started sliding?