Tony Wang10 min readAmerica's Housing Market Didn't Crash — It Split in Two
173 months of Redfin data: the 2021 seller's market unwound into a frozen standoff, then split — Sun Belt to buyers, Rust Belt to sellers.
The headlines can't agree on whether US housing is crashing or booming, and the reason is that both are happening in different zip codes. We pulled the full monthly history — 173 months of Redfin Data Center statistics from 2012 to mid-2026, joined to Census income — and the picture that falls out isn't a crash or a boom. It's a market that froze at the top, then split down the middle.
The seller's market of 2021 has fully unwound
For a decade, US housing ran in a fairly stable band: roughly a fifth to a third of homes sold above asking, and homes took a month or two to sell. Then 2021 happened — and then it un-happened:
National, all residential. 2026 is the May reading; other years are each June. Two measures of who holds the leverage: homes closing over asking vs. active listings forced to cut.
At the 2021 peak, 56% of homes sold above asking and homes went under contract in a median of 17 days. Bidding wars were the default: the median home sold for 102.5% of its list price. That is as lopsided as the data goes.
Every one of those signals has since reversed. Homes sold above asking fell from 56% to 30%. The median sale-to-list ratio dropped back below 100% — homes sell at a discount to asking again. Days on market climbed from 17 back to 42. And the share of listings cutting their price rose to 20%, the highest in the entire 14-year series. On every measure of negotiating leverage, buyers have clawed back most of what they lost.
But prices never actually fell
Here is the twist that makes this a freeze rather than a crash: while buyers regained leverage, prices didn't give it back.
The national median sale price is $449,846 as of May 2026 — an all-time high, up 2.2% year over year and 16% above the 2021 peak. Inventory has recovered to 1.46 million active listings, months of supply is back to a normal-ish 3.1, and yet prices keep grinding upward. This is the standoff: sellers won't cut, so they wait; buyers won't chase, so homes sit. High prices and weak urgency held apart, with the gap absorbed by longer marketing times and more price drops rather than lower clearing prices. Nobody capitulates. The market doesn't clear — it freezes.
The national average hides a regional inversion
"Frozen" is a national average, and it's hiding the most interesting thing in the data. Sort the major metros by how fast homes sell and the country splits cleanly in two — and it's the opposite of the 2021 map. The Sun Belt boomtowns that led the frenzy are now the slowest; the affordable Northeast and Great Lakes metros that sat out the boom are now the hottest.
Rochester, New York is the single most competitive market in the country. Homes there sell in a median of 8 days, for 113.7% of list price, with 79% of homes closing above asking. Buffalo isn't far behind (12 days, 66% above asking), and the pattern extends across upstate New York, Pennsylvania and the Great Lakes. At the other end, the slow list is almost entirely Florida:
Show the regional split — hottest and coldest major metros
| Metro | Days on market | Sold above list | Cutting price | Sale-to-list | Price ÷ income |
|---|---|---|---|---|---|
| Rochester, NY | 8 | 78.8% | 17.6% | 113.7% | 3.37× |
| Buffalo, NY | 12 | 65.7% | 24.0% | 104.8% | 3.14× |
| Albany, NY | 10 | 55.9% | 22.2% | 102.1% | 3.95× |
| Grand Rapids, MI | 7 | 46.5% | 36.5% | 101.4% | 5.08× |
| Harrisburg, PA | 9 | 41.2% | 39.6% | 100.4% | 3.91× |
| — NATIONAL — | 42 | 30.0% | 20.0% | 99.3% | — |
| Savannah, GA | 77 | 15.9% | 28.8% | 98.3% | 4.72× |
| Naples, FL | 77 | 3.8% | 25.3% | 94.3% | 8.38× |
| West Palm Beach, FL | 81 | 8.0% | 24.0% | 95.3% | 7.24× |
| Miami, FL | 86 | 9.3% | 18.2% | 95.7% | 7.64× |
| Myrtle Beach, SC | 98 | 7.5% | 24.7% | 97.5% | 4.98× |
Redfin's own 2026 forecast reached the same conclusion independently — it named Great Lakes cities like Rochester and Buffalo the hottest markets of the year while flagging coastal Florida and Texas as the slowest, and Florida's statewide inventory has ballooned to roughly 84 days on market and 7.5 months of supply.
Why the map inverted: affordability, not weather
The tell is in the last column. The hot markets aren't hot because they're desirable in the glossy sense — they're hot because they're affordable and supply-starved. Rochester's median home costs 3.37× the median local income; Buffalo's 3.14×. In both, the affordability math is actually negative — a median household can afford more than the median home — so any local with a job can compete, and 500 of them show up for every listing. These metros never overbuilt and never bubbled, so there's nothing to unwind.
The frozen Sun Belt is the mirror image. Florida and the coastal Southeast overbuilt through the 2021–2023 boom, then got hit with the specific costs that make Florida ownership uniquely expensive right now — surging property insurance, special assessments on aging condos, and property taxes reset to boom-era valuations. Supply piled up (Naples sits at 8.38× income, Miami 7.64×), demand cooled, and the result is 80-to-98-day marketing times and one in four sellers cutting price. The 2021 winners became the 2026 laggards not because the weather changed but because price ran past what local incomes could carry, and the correction is coming out as time and discounts rather than lower prices.
This is the same divergence our Airbnb-versus-housing study found from the supply side — the Sun Belt vacation metros that carry more Airbnb listings than homes for sale are exactly the ones now cooling as owner-occupier markets — and it sits alongside our affordability map of the salary you need to buy in each state.
What this means if you're buying, selling, or investing
For buyers, the single most useful fact in this data is that "the housing market" is not one market. In Florida you have leverage — homes sit, sellers cut, and 95% of list is a realistic offer. In upstate New York you have none — you're competing against dozens of cash-flush locals and paying over asking. National mortgage-rate commentary tells you almost nothing about which of those two you're walking into.
For sellers, the price-drops column is the early-warning system. A market where 25% of listings are already cutting price (most of Florida) is one where the first list price matters enormously, because the discount is coming out of your number, not the market's. In a Rochester-type market, underpricing to start a bidding war is the dominant strategy; in a Naples-type market, overpricing means 90 days of silence.
For investors and researchers, the whole 173-month series is queryable — every metro, every month, every leverage metric — over one REST API. The regional inversion is reproducible for any market pair you care about, and the affordability join is already done.
Query the US housing market yourself
173 months of Redfin market statistics — price, days on market, price drops, sale-to-list and Census-joined affordability, by metro, county, city or zip — over one REST API. 2,000 free credits a month.
Frequently asked questions
Is the US housing market crashing in 2026?
No. Crawlora's tracking of 173 months of Redfin data shows the national median sale price at an all-time high of $449,846 in May 2026, up 2.2% year over year. What changed is negotiating leverage, not price: homes sold above asking fell from a 2021 peak of 56% to 30%, price drops rose to a series-high 20%, and days on market climbed from 17 to 42. The market froze into a high-price, low-urgency standoff rather than crashing.
What is the hottest US housing market in 2026?
Rochester, New York — homes sell in a median of 8 days at 113.7% of list price, with 79% closing above asking. The Northeast and Great Lakes lead the hot list (Buffalo, Albany, Harrisburg, Grand Rapids), which Redfin's own 2026 forecast independently named as the most competitive markets of the year.
Why is the Florida housing market slowing down?
Florida overbuilt through the 2021–2023 boom and then absorbed costs that make ownership uniquely expensive — surging property insurance, aging-condo special assessments, and taxes reset to boom-era valuations. Supply piled up while demand cooled, so 8 of the 12 slowest-selling major US metros are now in Florida. Miami sits at 86 days on market and 9% above asking; Naples sells for 94.3% of list.
Which US metros favor buyers vs. sellers right now?
The country split regionally. The Sun Belt, especially Florida and the coastal Southeast, favors buyers (long marketing times, frequent price cuts, sales below asking). The affordable Northeast and Great Lakes — Rochester, Buffalo, Albany, Grand Rapids, Harrisburg — strongly favor sellers, with bidding wars and sales well above asking. National mortgage-rate commentary doesn't tell you which one a given metro is.
Did US home prices fall in 2026?
No, nationally they rose to an all-time high ($449,846 median, +2.2% year over year) even as buyers regained negotiating leverage. The market's correction has come out as longer marketing times and more price drops rather than lower clearing prices — a freeze rather than a decline. Individual metros vary: some Florida markets show small year-over-year declines while Rust Belt metros post gains.
What data is this based on?
Crawlora's Housing Markets dataset — Redfin Data Center monthly market statistics per region and property type since 2012 (173 months), joined to Census ACS median household income for affordability. Figures are aggregate market statistics (price, days on market, sold-above-list share, price drops, sale-to-list, inventory) — no individual properties or owners. Latest complete period is May 2026.