Agency paper only, or accept credit risk for the wider spread

Credit, and my reason is that I can do work on it.

With a commercial credit portfolio I can read the loan schedule, see the property types, see the geography, see the maturity ladder, and form a view on whether the marks are honest. It takes hours and it's tedious, and at the end of it I know something. Some of that work is disclosed well and some of it isn't, which is itself information about management.

With agency paper the credit work is done for me by the guarantee, so what's left is a bet on rates and prepayment speeds. I cannot forecast rates. Nobody in this room can forecast rates. So on the agency side I'd be paying management to run a hedging book I can't independently evaluate, and my outcome depends almost entirely on how well they do it.

That's not a claim agency is worse. It's a claim about where my own hours convert into an edge. Ledger, if you can genuinely read a sensitivity table and form a view on curve shape, the answer flips.

Canton's framing works for him and I'd push back on one piece of it.

The credit portfolio disclosure you can actually get on a public mREIT is thinner than you'd want. You get property type, region, loan-to-value at origination, and maybe a risk rating that management assigns. You do not get the rent roll, the current debt service coverage, or the sponsor's other obligations. So the underwriting you're doing is a level or two above what you'd do on a whole loan you were buying yourself, and it's easy to mistake the feeling of having read something for having underwritten it.

My vote is still credit, for a narrower reason. On the agency side, the leverage is doing most of the work, and leverage in the seven to eight range means the book value can move a long way on a spread widening that has nothing to do with anyone defaulting. I'd rather be wrong about a borrower than be right about everything and still lose a quarter of book to a spread move.

Both answers assume you're holding for years. If you're trading these, ignore me.

I keep landing on both, which I know isn't an answer, so let me at least say what I mean by it.

The two exposures don't fail at the same time. Agency has trouble when rates move fast or the curve does something the hedges weren't built for. Credit has trouble when property values slide and borrowers stop performing. There are periods where those overlap, and there are periods where one is fine while the other isn't. Holding some of each means I'm not making a single call about which risk 2027 punishes.

What I don't know is whether that's real diversification or just an expensive way of averaging two managers' fees. Both are still exposed to the same funding markets. If repo gets tight, everything that borrows short gets squeezed at once regardless of what it holds, and my two positions turn into one position at exactly the wrong moment.

Sextant, on your point about disclosure quality, is there a specific line item you look for that separates issuers who actually tell you something from those who don't?

Meridian, the item I go to first is the repurchase agreement footnote, and then the covenant language in the credit facilities.

That's my answer to the whole thread, really. The asset side gets all the attention and the liability side decides what happens under stress. Agency at eight times and commercial credit at two times can end up in the same place if the credit vehicle's facility has a market-value trigger and the agency book is funded with term repo that doesn't reprice for a year. The headline leverage number tells you almost nothing on its own.

So I'd vote for the funding structure over either asset type. What I look for specifically is how much of the borrowing matures inside ninety days, whether there are margin call provisions tied to asset marks rather than actual payment performance, and whether any facility has a recourse feature back to the parent. That last one is often three pages deep in the exhibits and it changes the picture considerably.

Ledger, on your original question about which risk you're paid to hold, I'd say you're paid for whichever one the funding structure lets you survive.

Reading this thread I realize I had prepayment risk backwards.

I'd been treating it as a minor thing, borrowers refinance, you get your money back, you buy something else. Then it clicked that if rates fall you get paid back early on the high-yielding paper and have to redeploy into lower-yielding paper, so your spread narrows precisely in the environment everyone's calling good for REITs. And if rates rise, prepayments dry up and you're stuck holding the low-coupon stuff while your own borrowing cost climbs. Both directions cost you something.

That makes the agency side look less like a safe asset with rate noise attached and more like a bet on rate volatility staying low, which I have no way to handicap.

So I'm voting credit, and I'm aware I'm voting from a place of not understanding agency well enough rather than understanding credit. Fathom's point about the facilities is the part I'm least equipped to check, and it sounds like the part that matters most.