Paper 141

Forecasting the U.S. Home Vacancy Rate using ARIMA

September 24, 2025: 12:30 PM - 1:00 PM

Careers, Training, and Education (CTE), Breakout Room #7

Authors	Abstract
Joseph Maina, Abdelmonaem Jornaz	This study forecasts the U.S. home vacancy rate, a key indicator of housing market health, using Federal Reserve Bank of St. Louis (FRED) data from 1986 to 2024. The current homeowner vacancy rate stands at 1.1% in Q1 2025. The ARIMA model captures temporal patterns, while machine learning algorithms such as regression analysis, random forests, and gradient boosting models non-linear relationships with economic, housing, and demographic factors. Integration of these models enhances predictive performance, optimized using the Akaike Information Criterion (AIC) and Root Mean Square Error (RMSE). This provides insights on economic indicators, housing supply and demand, and demographic factors, through the performance optimization of the models. The forecasts enable data-driven decisions for policymakers, real estate professionals, and urban planners, supporting housing development, supply chain planning, and related sectors like Heating, Ventilation, and Air Conditioning (HVAC). This research aims to promote economic stability through data-driven decisions.

Joseph Maina, Abdelmonaem Jornaz

This study forecasts the U.S. home vacancy rate, a key indicator of housing market health, using Federal Reserve Bank of St. Louis (FRED) data from 1986 to 2024. The current homeowner vacancy rate stands at 1.1% in Q1 2025. The ARIMA model captures temporal patterns, while machine learning algorithms such as regression analysis, random forests, and gradient boosting models non-linear relationships with economic, housing, and demographic factors. Integration of these models enhances predictive performance, optimized using the Akaike Information Criterion (AIC) and Root Mean Square Error (RMSE). This provides insights on economic indicators, housing supply and demand, and demographic factors, through the performance optimization of the models. The forecasts enable data-driven decisions for policymakers, real estate professionals, and urban planners, supporting housing development, supply chain planning, and related sectors like Heating, Ventilation, and Air Conditioning (HVAC). This research aims to promote economic stability through data-driven decisions.