Does an mREIT dividend make sense as a down payment holding account when the timeline is 18 to 30 months out

The $300 annual differential is probably understated, but the exit timing problem is worse than you've framed it. MREIT distributions are not steady coupons you can schedule around: they get cut precisely when rate moves compress the spread, which is the same environment that knocks 8 to 15 percent off book value. The two hits arrive together, so the scenario where you need the money and the position is down is not a tail risk, it is the correlated case you'd actually face.

The thing nobody has priced here is the bid-ask friction on a forced small-lot exit. If you need $4,000 out in month 19 because the punch list clears early, you are selling at whatever the market gives you that week, and on a thinly traded session that can be another half-percent or more below the quoted price.

Keep the full $14,000 in the HYSA and treat any dividend income from the mREIT idea as the speculative layer it actually is, meaning it only belongs here if you have a separate savings buffer that can absorb the position going to zero without touching your closing number. What's the punch list contingency that sets the short end of that 18-month window, and is 18 months actually realistic or is 30 the honest number?

The exit timing is exactly the right thing to worry about and nobody talks about it enough. I had a chunk parked in AGNC in early 2022 waiting on a lot purchase and the NAV dropped about 22 percent in six weeks, which on $4k would have been $880 gone right when I needed to move fast.

The part that gets people is the dividend reinvestment math only works if you can hold through the compression, and your window is 18 to 30 months, which means you could hit your target date right in the middle of a spread widening cycle with no exit that doesn't cost you principal.

What's driving the "maybe Zanesville or Lancaster" part, is that a price flexibility thing or a cash flow target, because if you can move the goalposts on the purchase price you might already have more optionality than you think?