---
title: "Real-Time House Price Model: U.S. Housing Market"
authors: "Federal Reserve Bank of Dallas"
source: "Dallas Fed"
source_url: "https://www.dallasfed.org/research/economics/2026/0115"
paper_id: "Dallas Fed Real-Time House Price Model 2026"
harvested: 2026-08-22
year: 2026
keywords: [house prices, mixed frequency, permits, housing starts, real-time, leading indicator]
---

## Abstract

Federal Reserve Bank of Dallas, *Real-time house price model shows U.S. housing market...* (2026).

The Dallas Fed built a mixed-frequency dynamic-factor style forecast model for US house-price
dynamics that combines slower-moving quarterly data with faster monthly indicators to produce
real-time estimates of house price growth — an early-warning tool for policymakers monitoring
systemic risk, and a timelier signal for households and businesses shaping borrowing/saving/
investment decisions.

From 20 candidate indicators (labor-market data, interest rates, construction permits, and
more), the best-performing model kept **five key variables**:
- Real GDP
- Average sale price of new homes
- **Permits for new single-family houses**
- **Housing starts**
- Sales of new single-family homes

**Relevance to StratMac RE-sector work:** the Dallas Fed model independently validates that
**permits and housing starts are load-bearing concurrent predictors** (here of house prices,
not quantity targets). This corroborates StratMac's Tier-1 screen (where CSUSHPINSA, REALLN,
TOTALSL, PRFI are load-bearing for predicting HOUST/PERMIT) and confirms the two RE-series earn
their place in a leading-indicator panel.