The U.S. housing market softened in 2025 but avoided a sharper downturn that many analysts had anticipated, according to a new analysis from the Federal Reserve Bank of Dallas that uses a real-time model to track home prices.
The research addresses a persistent challenge in housing data: lag. Most official house price indexes are released weeks or months after a quarter ends, limiting their usefulness for policymakers and market participants trying to assess current conditions. The Dallas Fed’s model blends traditional quarterly price data with faster-moving monthly indicators to estimate inflation-adjusted house prices in real time.
Those indicators include real gross domestic product, housing permits, housing starts, new-home sales and average prices of newly built homes. By updating as new data are released, the model provides an estimate of current-quarter price movements rather than relying solely on historical trends.
Applied to 2025, the model showed modest declines in real house prices early in the year, extending a cooling period that followed the rapid appreciation of the pandemic era. However, the analysis indicates the pace of decline slowed and stabilized by midyear, rather than deepening into a broader correction.
That assessment was later supported by official data showing a slight increase in real house prices in the second quarter. The Dallas Fed notes that, historically, the model’s estimates closely track observed price movements and generally outperform simpler forecasting methods.
Housing remains a critical component of the U.S. economy, in part because residential property represents the largest asset for most households. Significant price declines can reduce household wealth, dampen consumer spending and contribute to tighter financial conditions.
The findings suggest those effects did not materialize in 2025. Instead, higher interest rates appeared to slow housing activity without triggering a sharp fall in prices. Limited housing supply, continued demand and the prevalence of low fixed-rate mortgages among existing homeowners likely helped support values.
While the model is not designed to predict long-term trends or account for unexpected shocks, the Dallas Fed said its real-time approach offers a clearer picture of housing conditions during periods of economic uncertainty.