The 10-year Treasury started October at 5.28% and closed Friday at 4.94%. That happened in a month when two Fed members voted for another hike, so the short end and the long end are pulling in different directions right now. Mortgage News Daily's 30-year index finished the week at 7.27%, its lowest in more than a month, and Freddie Mac's survey has dropped two weeks running to 7.24%.

Most of why the long end fell sits outside the Fed. Core PCE came in at 3.1%, right on forecast, so there was no inflation scare. Money moved out of tech stocks and into bonds. And there's been talk that heavy AI-related corporate borrowing, one of the things pushing rates up this year, might slow down.

Here's how I'm reading it for the files I underwrite. On a $264,000 loan (20% down on a $330,000 house), a quarter-point move either way changes an exit buyer's budget by about $6,500. That's roughly what one strong or weak jobs report can do to mortgage rates in a week, and the October report lands Friday with a 55,000 consensus after September's 29,000.

So I'm not pricing flips off the last two weeks of relief. If the number runs hot and rates give back a quarter point, a resale priced to the dip has lost $6,500 of buyer budget before it even lists. On rentals, a $250,000 takeout runs about $1,709 a month at 7.27%, $1,751 at 7.52% and $1,667 at 7.02%. If a file only covers at the low end of that $42 band, it doesn't really cover.

Election Day is Tuesday too, which usually means some overnight bond volatility before the jobs number even shows up.

I put the full weekly recap, with the week-ahead calendar, on my website, which you can find through my profile. Investors can sign up there to get new reports by email.

Question for the group: are you underwriting exits at today's rate, or building some of October's drop into your numbers? And has anyone holding rentals actually locked a refinance in the last two weeks to grab it?