UC San Diego’s new model predicts house prices to fall by up to 18% this year

A new home price forecasting model based on consumer demand projects home prices will fall 5% nationwide and 12% in San Diego County by the end of this year. The model, which highlights online search activity, was recently published in a new study from the University of California San Diego’s Rady School of Management.

The model’s predictions have been shown to have an accuracy rate of up to 70% and are unique to other price predictors — like Zillow, Goldman Sachs, and Redfin — because they take into account a variety of factors, including interest rates, wage growth, unemployment, and housing supply. The Home Search Index, created by Allan Timmermann of the Rady School and collaborators at the University of Arhus in Denmark, focuses on consumer demand by tracking the rate at which potential buyers are using the internet to search for a home.

“It’s one of the purest measures of potential demand you can get because the first thing you do when you’re looking for a home, or interested in buying a home, is go on the internet and look what is available.” said Timmermann, a distinguished finance professor at the Rady School. “Those involved in the home buying market leave a large footprint with their online search activity as it takes time – often several months – to find something suitable.”

Cities like San Diego have house prices falling more than the national average because that’s where the market has been overheated the most during the pandemic, Timmermann said.

“What you saw after the March 2020 lockdowns was that sunshine and suburbs became a big thing,” said Timmermann. “People shifted to working from home so they didn’t have to be close to work and then they might move out of their area altogether and choose to live somewhere with more space and better weather. San Diego has lots of suburbs and of course desirable weather.”

Those features and limited supply sent prices skyrocketing across the country, but the market has cooled by 2.5% since May 2022, when prices peaked.

“Many households were priced out of the market, so we are now seeing that the level is adjusting,” said Timmermann.

But house prices are also likely to fall further in other cities. Phoenix, AZ is expected to see the largest drop at 18%. Other metro areas where prices are forecast to decline include Stockton-Lodi, CA (down 13%), Las Vegas, NV (down 13%), followed by San Diego and Tucson, AZ. Among the cities with the most price stability are the metropolitan area of ​​Scranton-Wilkes-Barre-Hazleton, PA and Kansas City, MO, both of which are expected to increase by 2%. Other cities forecast for stable prices include Hartford, CT, Harrisburg, PA and Omaha, NE.

Timmermann added that the predictive power of web searches is usually a reliable indicator of where the market is going in the short to medium term, as fluctuations in demand are more important than changes in supply, which tends to be quite stable over shorter periods of time.

A key difference between UC San Diego’s home price forecasting model and other commercial price predictors is that the data underlying the Home Search Index is not proprietary. The methodology is fully transparent and reproducible as the study published in management scienceis public so anyone can see how it works.

The formula starts by tracking keywords like “buy a house” and related search terms on Google Trends – a free website that analyzes the popularity of the top searches on Google Search. This data is compared to data on home visits and written offers, allowing researchers to forecast short- and long-term prices.

“Your time investment, search intensity, and number of people searching really reflects the underlying interest in home buying,” Timmerman said. “Ultimately, the higher the demand, the higher home prices tend to be.”

The co-authors of the Management Science paper include Stig Møller, Thomas Pedersen and Christian Schütte from the University of Arhus.