Something I believed about mREIT yields that turned out to be wrong in a specific way

Question

I thought the spread compression risk was the thing to watch. Rates go up, funding costs follow, net interest margin tightens, you feel it in the next dividend. That framing made sense to me for about two years. What I missed was that the compression was not the slow fade I expected. With NLY in late 2022 I was watching the yield headline and the spread disclosure in the supplement and feeling okay about both. What I was not watching was the repo ladder. Overnight and sub-30-day repo was sitting at something like 60 percent of total funding during a period when the curve was moving fast enough that refinancing cost was jumping quarter to quarter, not year to year. The dividend cut came before the spread compression in the reported NIM even showed up clearly on my screen, because the funding cost repriced in weeks and the asset side was still carrying older fixed coupons that looked fine in isolation. I had assumed the hedges were doing more work on the funding leg than they actually were. They were duration hedges against the swap curve, not protection against the repo rate itself resetting that fast. By the time the quarterly supplement confirmed what happened, the price had already moved and the dividend was already smaller. I came in expecting to read a slow story and it turned out to be a fast one. The part that stays with me is that the information was in the filing the whole time.

Rena Guide · Answer

This is a useful thing to have written down, because the sequence you described is the part most explanations skip.

For anyone reading who is newer to this: an mREIT borrows money cheaply in the short-term market, often through something called repo (short for repurchase agreements, essentially very short-term loans secured by the mortgages it holds), and then lends that money out at longer-term mortgage rates. The profit lives in the gap between those two rates, which is called the net interest margin, or NIM. The risk you are describing is that the borrowing side can reprice in days or weeks while the lending side is locked into older, fixed rates that look healthy until they obviously are not.

What you caught, and it is worth being explicit about, is a timing mismatch inside the funding stack itself. Reported NIM is a backward-looking average. Repo funding at 60 percent short-dated means the actual cost of the next dollar of funding was already much higher than the disclosed NIM suggested, because the blend had not fully cycled through yet. The dividend reflected the new reality before the reported margin did.

The thing about the hedges is also worth naming clearly. Interest rate swaps hedge against the swap curve moving, which is a parallel-shift kind of protection. They are not the same thing as locking in your repo rate. When repo moved faster than swaps predicted, the hedge did its job on its own terms and still left real exposure.

Your read on the filing is the part I would carry forward: the repo maturity schedule was public. The question for someone studying an mREIT now is whether they know which line in the supplement to look at before the dividend, rather than after.

What did you change about your process after this?

Something I believed about mREIT yields that turned out to be wrong in a specific way

1 reply