Goldman’s housing forecast hides 94 cities bleeding under the 5% threshold while spotlighting a sanitized 15%, the real damage is buried in the model

by

Goldman Sachs says 55 cities will see home prices fall. That’s 15% of the 381 metros they track. But they also say 39% will decline. So why spotlight the softest wound?

“Goldman Sachs expects home prices to fall in 55 of the 381 metropolitan statistical areas it tracks, or about 15%.” https://www.cbsnews.com/news/home-prices-forecast-goldman-sachs-housing-market/

“While we think national home prices will likely avoid a correction in 2023, we expect 39% of metropolitan areas to experience price declines.” https://finance.yahoo.com/news/home-prices-fall-goldman-sachs-expects-104729829.html

“Goldman Sachs has identified the San Jose, Austin, Phoenix, and San Diego markets as particularly ‘overheated’. These markets are predicted to experience ‘peak-to-trough declines’ of more than 25%.” https://www.housingwire.com/articles/home-prices-will-drop-to-2008-levels-in-these-four-cities-goldman-sachs/

“Residential investment should fall 8% in the back-half of the year… multifamily homebuilding will remain depressed… single-family starts will contract.” https://www.cnbc.com/2025/08/04/housing-to-remain-weakest-part-of-economy-in-the-2nd-half-goldman-says.html

They say 55 cities will drop. That’s the public stat. But they model 149 metros for decline. That’s the internal sheet. The 55 are breach cases, 5%+ drops. The other 94 are sub-5% bleeds. Quiet. Undetectable. They spotlight the softest damage. They bury the rest under rebound projections. San Jose, Austin, Phoenix, San Diego flagged for 25%+ collapse. But they call it “overheated.” No stress curve. No timeline. No velocity. East Coast labeled “resilient.” But no modeling of rate stickiness or inventory distortion. They publish optimism. They model erosion. The 15% stat is real but it’s a decoy. The 39% model is buried. The omission is engineered.