News media coverage and the predictability of house prices

Abstract This article introduces new housing‐media‐attention indices for the 50 U.S. states based on the Bloomberg Terminal News Trends (NT) function, which collects articles from various news and social media sources and identifies their content using artificial intelligence tools. The results indicate that the state‐level housing‐media‐attention index explains a significant portion of the total variation in future house prices, even when economic fundamentals are considered. Additionally, I find that the impact of housing media attention on future house prices is stronger in states with non‐recourse mortgage laws, greater land‐use regulations, and higher social connectedness among individuals of both low and high socioeconomic status. Out‐of‐sample forecasting results further suggest that housing media attention has the potential to act as an early indicator of the direction of the housing market.