Pay down a 7.1 percent loan with the 90k, or buy paper?

Your last paragraph is the real post. You've had flexibility be worth more than yield twice, which means you already know the paydown has a cost you feel and the spreadsheet doesn't show. Have you priced what a HELOC or a line against that fourplex would cost you as a middle path, so the money goes in but stays reachable. Terms and availability on lines against investment property vary a lot by lender, so get quotes before you assume it's an option.

The two options aren't measured on the same axis, so comparing 7.1 to 11 directly will mislead you.

Paying down the loan produces a certain reduction in interest expense at the note rate, with no mark to market and no manager. Its downside is that the capital is locked in the property until a sale or refinance, and current lender terms for pulling it back out have to be confirmed in writing before you count on them.

A mortgage REIT produces an uncertain total return made of two parts. The dividend is one part. The change in book value per share is the other, and it can be sharply negative. These vehicles borrow short term, often on repurchase agreements, and hold longer dated mortgage assets, so their earnings depend on the spread between funding cost and asset yield, and their book value depends on where those mortgage assets are marked. Widening spreads reduce book value even when every borrower is paying. Leverage of five to eight times equity on the agency side amplifies both directions. Prepayment adds another layer, since falling rates can retire the high yielding assets and replace them with lower yielding ones.

The structural backdrop people cite is real. Banks have pulled back from parts of real estate lending, demand for real estate credit is strong, and mortgage REITs sit on the lending side of that with public market liquidity. That describes the opportunity, not the outcome for any particular company in any particular year.

What the comparison depends on is your horizon and how much you'd pay for optionality. If you'd want that 90k back inside three years, the paydown is expensive to reverse and the mortgage REIT can be sold at a price you don't control. Neither is a cash equivalent.

Break the 90k into pieces instead of picking a winner, because your two problems have different sizes.

Write down what a full reserve looks like for the fourplex. Roof, one HVAC, one sewer surprise, and six months of debt service with the building half empty. For a fourplex that's often somewhere between 25k and 40k depending on the age of the systems. That amount doesn't go into anything with a price chart. It sits in cash or something maturing on a known date.

Whatever's left after the reserve is the money that actually gets to answer your question. If that's 55k, the choice between a 7.1 percent guaranteed saving and a levered mortgage vehicle is a much smaller decision than it felt like when the number was 90k, and the flexibility argument gets weaker too, because the reserve is now carrying the flexibility job.

On the paper side, rena-expert's point about book value is the one I'd hold onto. Look at any candidate over four years, add up dividends paid per share, add the change in book value per share, and see what's left. Do it for the same company across the last rate cycle. If the answer is that the dividend was mostly capital coming back, you know what the 11 percent is worth.

My own bias is that a 7.1 percent note at 24 years remaining is a large, certain, unglamorous drag and killing part of it is hard to regret. But I'd fund the reserve first regardless of which way you go.