Why San Francisco's housing market confounds the forecasters

THE QUESTION of what a home in the San Francisco metropolitan area is worth is less interesting than the reason the answer keeps changing. In September 2025, Zillow, a property-technology firm, published a 12-month forecast predicting that the San Francisco-Oakland-Berkeley metro would see home values fall by 3%, one of the steepest declines among America's 300 largest housing markets. By June 2026, Zillow's own data showed that the average home value in San Francisco proper stood at $1.4m, up by 9.5% over the previous year. A prediction of decline has been overtaken by a resurgence.

The discrepancy is not a failure of modelling so much as a failure of the model's assumptions about what drives the Bay Area's housing market. Zillow's forecast was built on national patterns: cooling demand, elevated mortgage rates keeping buyers on the sidelines, and a softening Sun Belt dragging the average. It did not capture what happens when an exceptionally tight supply constraint meets a recovering local employer base.

San Francisco's housing market operates under structural conditions that differ from those in most of America. The Bay Area is landlocked on three sides and hemmed in by one of the world's most aggressive growth-management regimes. Between 2000 and 2024, the value of a typical Bay Area home rose by 75%, to $1.18m, according to the Metropolitan Transportation Commission, which tracks regional economic data. Values peaked at $1.29m in 2022, fell as mortgage rates rose, and have since climbed back. The underlying mechanism is simple: there are far fewer houses than people who want to live near high-paying jobs in Silicon Valley and San Francisco.

The incentives at play reinforce the scarcity. Homeowners who bought during the 2022 peak face negative equity if they try to sell into a softening market, so they hold. Developers face years of environmental review, neighbourhood opposition and construction costs that make new housing economically marginal unless prices remain high. Municipalities have little political incentive to approve density that would dilute current property values and increase pressure on schools and infrastructure. The result is a supply curve that is almost vertical.

To be sure, the Bay Area is not immune to economic pain. The tech sector's post-pandemic reckoning - mass layoffs in 2022 and 2023, followed by a cautious rehiring - did dent demand. Remote work reduced the premium on living within commuting distance of an office. A blog run by a local real-estate data site, FirstTuesday, argued in August 2025 that San Francisco's housing market would not bottom until 2027 or 2028. That forecast has already been disproved by the data.

The reason is not hard to see. Remote work turned out to be less of a permanent displacement than forecasters feared. Companies that pulled back on office mandates in 2024-25 found that productivity and collaboration suffered, and began requiring employees to return. Meanwhile, artificial-intelligence investment has concentrated spending and talent in the Bay Area. The region's employers are hiring again, wages are recovering, and buyers with savings - often those who were priced out during the pandemic frenzy - are returning to the market. Demand was never the binding constraint. Supply was.

The numbers bear this out. March 2026 saw the median sale price of single-family homes in San Francisco city reach $2.2m, up by 18% year on year, according to market reports cited by local real-estate observers. Sales volume remains below pre-pandemic levels, which is consistent with a market where high prices ration demand rather than a market in structural decline.

This matters for three reasons. First, it shows that national housing models systematically underweight local supply constraints. A forecast that treats San Francisco as another metro area with cooling demand and rising rates misses the fact that its prices are governed by land scarcity and regulatory friction, not by mortgage affordability alone. Second, it reveals a distributional problem: the people who benefit from scarcity-driven appreciation are those who already own. Everyone else - renters, younger buyers, and lower-wage workers - pays the cost in the form of ever-higher housing expenses and displacement. Third, it exposes the political impasse. Local governments in the Bay Area have occasionally passed ambitious housing targets, but the mechanisms to enforce them - state pre-emption of local zoning, streamlined permitting, density bonuses - are applied unevenly and opposed ferociously.

The broader lesson is that housing markets do not respond uniformly to macroeconomic shifts. Rate hikes suppress demand everywhere, but only where supply can expand do prices actually fall. In the Bay Area, the supply side is structurally locked. Until the regulatory barriers to building are lowered - a change that would require state-level political courage and a willingness to alienate current homeowners - the market will continue to confound the forecasters who expect normal behaviour from an abnormal system.

That bargain is breaking: the Bay Area wants both unlimited tech-sector growth and limited housing supply. The arithmetic suggests it can have only one.