Hey there —

A number stayed with me this week, and it is not one of the ones I expected to be writing about.

3,600. That is the median number of renter-occupied homes in the entire metropolitan area, for the 157 US metros that have a published home-value index and no rent index at all (🟢 Census ACS 5-Year Estimates). Not 3,600 listings. Every rented home in the whole metro.

Those markets came up for a boring reason. Run a rent-to-price screen across every metro we can measure — 737 of them — and not one clears today's bar. The median sat at 0.49% a month, less than half of the old plain 1% Rule. Which reads like a verdict on American rental real estate, and I do not think it is one.

What is actually going on is that the screen has a blind spot, and the blind spot has a shape. The metros missing from the data are systematically smaller and cheaper — exactly the profile most likely to clear. And here is the part I keep coming back to: the share of households that rent in those places is completely ordinary. Thirty percent, against 31.5% in the well-covered metros. People rent there at the same rate as everywhere else. There are just very few of them.

That changes what a ratio means. In a metro with a few thousand rentals, the rent figure is not really a market rate — it is the middle of a handful of listings that happened to be live when somebody measured. One of the markets that cleared our screen has two rent readings in the entire year. February, then July, with a 37% jump in between and nothing to check it against. Another has an unbroken monthly record going back more than two years. They sit at the same rank, in the same column, looking equally solid — I put the two side by side here.

A rank shows you the number. It never shows you the confidence behind it.

Here is the part worth doing something with. If a market cleared your screen recently, spend ten minutes finding out how many months of rent history sit behind that number. Steady readings mean you have a real signal and can move on to pulling actual comparable rents. Scattered readings mean you have an estimate with a wide error bar, and you should price the deal off three comps you find yourself, on a page that opens the calculator with a worked deal already in it. And if you cannot find three comparable rentals in that market at all — that is not an inconclusive result. That is the answer, and it just saved you a trip.

Run the same bar one level down — ZIP codes instead of metros — and eighteen clear it. Every single one is a $48,000–$109,000 urban submarket. The screen does not fail — it bottoms out into risk it cannot price, which is a different and much more useful finding.

So: when was the last time you checked how much data was behind a number you acted on? Not a trick question — I suspect the honest answer for most of us is "never, it was in a column and the column looked fine." Hit reply and tell me. I read every one.

Martin