The steepest single-family price declines are not where high mortgage rates hit hardest. They are where someone was still free to sell. The source reports drops of 11% to 26% in 16 bigger markets and 9% to 10% in another 13. Read the pattern alongside the San Francisco exception and the common explanation, that expensive mortgages are deflating housing across the board, gets the mechanism backwards. Expensive mortgages mostly freeze prices in place. Prices actually fall when a seller arrives who does not care what rate they are giving up.
Why Rate Shock Props Up Prices First
Start with what intuition expects. Higher mortgage rates cut how much buyers can borrow on a given income. Less purchasing power should mean lower bids and therefore lower prices everywhere. That half of the story is correct.
The other half works in the opposite direction. Most existing owners carry mortgages written when rates were far lower. Selling means giving up that loan and financing the next home at today's rate, which turns a move into a large rise in monthly cost. So owners stay put. Resale listings dry up, and when supply shrinks as fast as demand, prices hold. This lock-in is the reason national price indices stayed so stubborn after the rate shock. Demand fell, and the existing-home supply that would normally have forced price discovery fell with it.
This means price declines are a supply story before they are a demand story. To find where prices give way, look for supply that the lock-in does not restrain.
The Sellers Who Were Never Locked In
Three groups fit that description, and they gather in the same kinds of markets.
The first is homebuilders. A builder holding finished inventory has no low-rate mortgage to protect. It has carrying costs, land already paid for, and a quarterly earnings report to deliver. Builders clear inventory by cutting prices or, more often, by paying to buy down the buyer's mortgage rate. Either way, the new home competes directly with resale homes nearby. A resale seller who refuses to cut is now competing against a subsidized new house down the road.
The second is investors and second-home owners who bought during the pandemic rush. Their math is about rent coverage and carrying cost, not about holding on to a cheap primary-residence loan. When rents level off and insurance and property taxes rise, selling becomes the rational move.
The third is forced sellers: relocations, divorces, estates. They exist in every market, but they set prices only when the other two groups have already pushed listings up.
Builders put up the most homes in the markets where land was cheap and permitting was quick. Investors piled into the same metros during the migration boom of 2020 to 2022. The source excerpt does not list its 16 markets, so the specific names cannot be confirmed here. The size of the declines, though, fits the profile of the pandemic boomtowns, mostly in the Sun Belt and Mountain West, that saw the largest run-ups. Declines of 26% are not a gentle cooling. They are what happens when unlocked supply meets a buyer pool whose borrowing power has already been cut.
The builder channel can be checked against filings. D.R. Horton, the largest US homebuilder by volume and concentrated in Sun Belt markets, reports gross margin each quarter. Incentives and rate buydowns show up there as a cost. If builders are the marginal price-setters in the declining markets, that cost should appear as pressure on margins even while unit volume holds up. That is the trade a builder makes when it chooses to sell homes rather than hold them.
San Francisco Runs on a Different Input
San Francisco is the control case that proves the mechanism, not an exception to it. The source describes the market as going haywire as AI money spills into housing. Look at what changes in that sentence: the source of purchasing power.
In most markets the buyer's ceiling is set by income multiplied by the borrowing capacity a mortgage rate allows. In the Bay Area right now, a meaningful share of buyers get their down payments, or entire purchase prices, from equity compensation and liquidity events at AI companies and the firms around them. A buyer paying cash, or putting down half the price from vested stock, is barely sensitive to rates. The cost of a mortgage drops out of the bid.
San Francisco also lacks the unlocked supply that sank the boomtowns. Building there is slow and limited by geography and permitting. Its own owners face the same lock-in as everyone else. So the city gets a new source of demand that ignores rates, with no supply valve to absorb it. The result is a bidding war in a market that has spent several years as the punchline of post-pandemic real estate.
The broader lesson concerns what housing actually tracks. In San Francisco, home prices are behaving like a leveraged derivative of local equity wealth. The exact size of the AI-driven buying is not documented in the source material, and how widely equity proceeds reach housing across the buyer pool is an inference, not a measured figure. The direction is consistent with the mechanism even so: when the buyer's balance sheet comes from the stock market rather than the mortgage market, rates stop setting the price.
The Countercase: Declines From a Peak Are Not a Crash
The strongest objection deserves a direct hearing. Declines of 11% to 26% may be measured from each market's peak, and the source's word "already" points to a running total. If so, much of the drop reverses a run-up that was itself extraordinary. A market that rose 50% and gave back 20% still sits well above where it started. On that reading the declines are a normalization, and calling them distress overstates the case.
That objection is partly right, and it sharpens the thesis rather than breaking it. Normalization still needs a mechanism, and lock-in rules out the usual one. Prices are not drifting down as ordinary owners slowly accept lower bids. They are being marked down by sellers whose costs and incentives are separate from their own mortgages. That makes the declines lumpy and local, and tied to builder inventory and investor exits.
The second objection is that falling rates could thaw the lock-in and bring a flood of resale listings everywhere at once, which would make the boomtown declines a preview of a national adjustment rather than a regional one. That is a real risk. The answer depends on how far rates fall. A modest decline would release some sellers and some buyers together. Only a large move would release supply faster than demand.
What Would Confirm or Break the Read
The supply-side reading makes three predictions that can be tested.
- Builder margins should keep absorbing incentive costs in the declining markets until finished-home inventory falls. If margins recover while the 26% markets keep falling, builders are not the price-setters and the thesis is wrong.
- Resale listings in the declining markets should keep rising faster than in locked-in coastal metros. If inventory converges nationally, the lock-in has broken and declines will spread.
- San Francisco's strength should track AI equity liquidity, not mortgage rates. A drawdown in AI valuations, or a pause in tender offers and secondary sales, should cool Bay Area bids even if rates fall. That is the cleanest test of whether a wealth shock, rather than credit, is driving the market.
The thesis breaks if price declines spread evenly into markets with tight inventory and little new construction. Until that happens, the cleaner reading of the evidence is that America does not have a housing market that is falling. It has a set of markets where the marginal seller is no longer locked in, and those are the only ones where prices have been allowed to move.
