Updated for the July 2026 release
Methodology
We rank the 200 largest US metro areas by the 90-day change in the Zillow Home Value Index, and publish the full table, the filters, and the near-misses every month. Everything below is what you'd need to reproduce our numbers.
How it works, in 30 seconds. Every month I take Zillow's estimate of what the typical home is worth in each big US metro, and rank the metros by how much that value moved over the last 90 days. The rent and the gross yield sit right next to the price, because a falling price and a worsening investment are not the same thing — and I say where I'd buy, watch or avoid as someone who owns rentals. A 90-day move is a short-window signal, not a verdict on a market. The full method, filters and limitations are below; the table itself is free.
The ranking metric
Zillow Home Value Index (ZHVI) — all homes, mid-tier (35th–65th percentile), smoothed and seasonally adjusted.
ZHVI is not an average of what sold. It is a constant-quality index: it estimates the value of the typical home in a market, whether or not that home changed hands this month. That distinction is the whole reason we use it.
Why not median sale price? Because median sale price measures what happened to sell, not what homes are worth. If a quiet month is dominated by luxury closings, the median jumps and nothing about the market has changed. That's mix shift, and it makes median sale price a poor ranking metric — even though it's the number most headlines quote.
Window: 90 days
Latest month against three months prior. Long enough that one noisy month can't drive a ranking, short enough to still be news. A 12-month window would mostly re-rank last year's story.
Seasonally adjusted
Home values follow a strong seasonal pattern — spring up, winter flat. Without adjustment, a spring ranking is a list of nothing. We use Zillow's seasonally adjusted, smoothed series so that a decline means a real decline, not January.
The filters
| Filter | Value | Why |
|---|---|---|
| Geography | Metropolitan Statistical Areas | Housing markets are commuter sheds, not city limits |
| Size floor | Zillow SizeRank ≤ 200 | Keeps #1 from being a 40,000-person town where six sales move the index |
| Completeness | Current and 90-day-prior values required | No imputation, no gap-filling |
Applied to the July 2026 release, this leaves 195 ranked metros.
Tie-breaks and near-ties
Ranked ascending by 90-day % change. Ties broken by (1) the larger 12-month decline, then (2) the larger dollar decline.
We disclose near-ties. In the July 2026 list, #10 (Dallas, −0.95%) and #11 (Stockton, −0.94%) are separated by one one-hundredth of a percentage point. That is not a meaningful difference, and we say so on screen rather than pretending the cutline is precise. The rankings page shows every metro, so you can always see who just missed.
What the second source is for
We pull Redfin Data Center metro files alongside Zillow, but not as a second ranking. Redfin's median sale price is unadjusted and mix-driven, and its release lags Zillow's by about two months. Comparing the two headline numbers directly produces nonsense.
A worked example. In the July 2026 data, Zillow's index fell in all ten of our metros over 90 days — while Redfin's median sale price rose in all ten. Both are correct. Redfin's window ran February to May, which is winter into spring, and its figure isn't seasonally adjusted.
Redfin is used for market-condition context: days on market, months of supply, share of listings with price drops, share selling above list, sales volume. Those describe how a market is moving. Zillow's index describes where values are. When two credible sources disagree, we show both and explain the disagreement.
The landlord layer
Every ranking also carries a gross yield estimate: annualised Zillow Observed Rent Index (ZORI) divided by ZHVI.
gross yield = (ZORI × 12) ÷ ZHVI × 100
This is deliberately a gross figure — before taxes, insurance, maintenance, vacancy, management and financing. It is not a return. It is a comparison tool: it shows how a market's price moved relative to the rent it produces.
A market where prices fell and rents held is a market that got cheaper to own for income — and that fact is invisible in a price ranking alone. We report it because falling prices and worsening investments are not the same thing, and most housing coverage conflates them.
Sources & attribution
- Zillow Research — ZHVI, ZORI, for-sale inventory, share of listings with a price cut, median days to pending. zillow.com/research/data
- Redfin Data Center — metro market tracker. redfin.com/news/data-center
Both are used under their public-data terms with attribution. We publish our derived rankings, not their raw files. If you want the source data, get it from the source.
Update cadence
Monthly, following Zillow's release (typically mid-month, covering the prior month). Each month's list is archived. We do not silently revise past rankings — and when a source revises its own history, which Zillow does, the revision is noted rather than quietly absorbed.
Known limitations
- Index revisions. Zillow restates prior months. A metro can move in the rankings retroactively.
- Metro-level averaging. "Austin" is a five-county area. Suburbs and core can move in opposite directions and net to zero.
- 90 days is still short. A three-month move is a signal, not a trend. We show 12-month and peak-to-now figures alongside it precisely so you can tell the difference.
- ZORI covers fewer metros than ZHVI (751 vs. 895). Where a ranked metro has no rent figure we leave it blank rather than estimate.
What this is not
This is not investment advice, a forecast, or a valuation of any individual property. A metro-level index says nothing about a specific house on a specific street. Every number here describes the past.
Change log
- 2026-07 — First published methodology.
Found a problem with any of this? Tell me — corrections get published, not buried.