The Price Says Up. The Negotiating Table Says Otherwise.

In 1947, Levitt & Sons started building houses on seven square miles of Nassau County potato fields, about thirty miles east of Manhattan. They put up as many as a hundred and fifty a week, 17,447 in the end, and sold the first Cape Cods to returning GIs for around seven thousand dollars. No money down, about sixty dollars a month.

The entire point of Levittown was that an ordinary person could buy a house without a fight.

A year ago, homes in Levittown were selling for 102.8% of their asking price. Nearly six in ten closed above list. If you wanted one, you paid what the seller asked and then some, and you were grateful for the chance.

That is over. Levittown homes now sell for 98.8% of asking, under list. The share closing above list fell from 59% to 40%. Buyers there are saying no for the first time in years.

Here is the part that makes it interesting: the median price went up 4.7% anyway, to $728,000. If you only watched the price, you would think nothing had changed at all.

Four measures, all moving the same way, in twelve months. The one that is easiest to miss is the last: a home that took three weeks to sell now takes six.

None of that comes from our Alpha Score, which is built from a completely different set of inputs. So it is worth asking whether the score sees the same thing.

It does. Read those two middle bars together. Price Momentum sits at 61, so the price line really is still healthy. Inventory Health sits at 37, and that bar is inverted: a low score means supply has loosened.

A market with rising prices and loosening supply is one where sellers are about to lose leverage. Levittown has already lost it.

If Levittown were the whole story it would be a curiosity. So start with the number everyone quotes. Across American ZIP codes the typical home still sells for about 98.5% of asking, down roughly a tenth of a point from a year ago. On paper, sellers barely lost a step, and every national headline this year is built on that.

A national average is a blender. Here is what it is blending.

The peak is real. Most neighborhoods did barely move. But that tail is close to two thousand American neighborhoods where the negotiation changed hands, averaged into invisibility by the places where it did not. Levittown is one of them.

Before going further, it is worth saying why I trust one number over the other.

Median price moves when a different mix of homes sells. A few large houses trade, the median climbs, and it looks like appreciation that never happened to any individual owner.

Sale-to-list compares what a seller got to what they asked, house by house. Mix cannot touch it.

That is why every market here had to clear a second test. Price per square foot had to rise too. If the median rose but per-square-foot fell, the "prices are up" story is a mix artifact and the market is out. That one filter removed a lot of tempting examples.

Run that filter across every ZIP code in the country and 561 markets survive it. Here they are against everywhere else.

The grey cloud is the normal state of things: prices and negotiating power drifting together, in both directions. The coral cluster is where they came apart. Those are the markets worth naming.

These are the ten steepest losses of seller pricing power in the country among markets where prices still rose. Grey is where each one sat a year ago; coral is today. Every green figure on the right is a market whose median price went up over the same period.

The sharpest reversal in the country is in Plainview, Texas, out on the Panhandle plains between Lubbock and Amarillo. Sellers there went from 108.2% of asking, eight full points over, to 93.0%. That is a fifteen-point swing in twelve months, and the median price rose almost 20% while it happened.

Plainview knows what a hard year looks like. In February 2013, Cargill closed the beef plant that was the town's largest employer, after drought gutted the regional cattle supply. It took about 2,200 jobs with it, close to 14% of the entire county's workforce, in a town of roughly 22,500 people.

A place that has absorbed that does not panic over a housing statistic. It is still worth noticing when the people buying homes there stop paying over ask.

Topeka, Kansas lost ten points of pricing power on 140 sales, one of the largest samples in the set, in a state capital with a steady government payroll underneath it. Hardin County, Texas, in the piney woods outside Beaumont, went from 98.5% to 88.9%: more than a dime on the dollar in negotiating room, while prices rose 4.7%.

Which raises the obvious question: is every one of these 561 markets cooling? It would be tidier to say yes. That is not what the data says, and the gap is worth showing.

Two of the markets above make the point themselves. In Hardin County, inventory actually fell over the year even as sellers gave up nearly ten points. In Baltimore's 21218, the Charles Village and Waverly rowhouse blocks around Johns Hopkins' Homewood campus, price per square foot jumped 12% and the share selling over list dropped from 38% to 22%. Yet new listings declined.

Neither is a textbook cooling market. Both are markets where sellers are getting less of what they ask. If I only showed you the 24% where every arrow points the same way, I would be showing you a cleaner story than the one that exists.

So here is the practical version. If you are buying in one of these markets, the price chart says it is hot and the negotiating data says it is not. That gap is where your room lives.

Ask your agent one specific question: what did the last ten comparable homes sell for relative to their asking price? In Levittown a year ago the answer was "over." Today it is "under," and that is the whole difference.

If you are selling, the frenzy is already over where you are — the list price just has not heard yet. Price to the market that exists now, not the one your neighbor sold into last spring.

You can run this read on your own market in about a minute. Open its market page, find the sale-to-list ratio, and compare it to where it sat a year ago. If the ratio is falling while the price line rises, you are in one of these.

P.S. If something in here nags at you, my guess is it is the 24%. Three quarters of these markets do not show the full cooling picture, and that is a reasonable thing to sit uneasily with. My read is that sale-to-list tends to move first and the rest catches up unevenly. Worth checking again in a couple of quarters.

Metric: average sale-to-list ratio, per Redfin data, ZIP level, 90 days ending May 2026 versus the same 90-day window a year earlier. Both sides use the same window length, and the ratio is immune to changes in the mix of homes sold. Charts are generated directly from the database and stamped with their as-of date. See the methodology.