Economic Variable Importance and Valuation Over Time in County-Level Real Estate Markets

Residential real estate markets are the biggest asset class in the world. They are interlinked with policy decisions, individual wealth, economic decisions, and a web of cause-and-effect relationships with moving parts of our country and the world. The vast amount of information associated with housing markets makes identifying existing relationships and patterns in our housing markets a surefire way to improve United States citizens’ lives and economic efficiency. To do so, diverse data associated with the HPI of a county in any way was aggregated for the United States, and machine learning models were trained on these datasets to accurately analyze their effect on HPI. These models show relevant patterns and trends, relationships between variables, which variables have the greatest impact, interactions between variables, and other variables for different granularities, allowing for an in-depth analysis of HPI and related socioeconomic variables and conditions. Additionally, by predicting HPI against real HPI, we form a Valuation Index showing overvalued and undervalued markets, which illustrates connections between this and other economic phenomena when comparing. By applying results from these models to the current socio-economic situation in the United States, patterns, trends, and relationships to existing research were identified regarding the housing market bubble, coastline premiums, artificial real estate inflation, and other variables.