Seller Power and Housing Price Surges: Evidence from U.S. City-Level Data (2018–2025)
Image Source: AI-generated
Role
Skills
time line
Full project
The overview
This semester-long research project examines what drove the $100,000+ surge in average U.S. home prices during COVID-19, investigating whether stay-at-home policies increased seller power. Using 4,386 state-month observations from Zillow and the COVID-19 U.S. State Policy Database (2018–2025), I applied statistical modeling and regression analysis to measure how pandemic policies, rental pressure, inventory, and affordability constraints shaped housing prices. Ultimately, the findings help informing equitable housing policy and crisis preparedness.
The development
1. Literature Review & Hypothesis: Reviewed 20 studies on housing economics and pandemic impacts. Formulated the hypothesis that state-level housing prices rose due to increased seller power as stay-at-home policies boosted demand while constraining supply.
2. Data Preparation: Combined Zillow’s city-level housing indicators with Boston University’s COVID-19 policy data, creating a panel of 4,386 state-month observations (2018–2025).
3. Model Development: Built a linear regression model using STATA with core predictors, such as rental demand, housing inventory, and income requirement, to explain home value changes.
4. Testing & Refinement: Checked multicollinearity, removed weak predictors, and finalized a three-variable model with sound statistical significance (R² = 0.196).
the takeaways
Impact of Rental Demand: Stronger renter demand can serve as a leading indicator for broader housing‐price trends.
Affordability Challenges: Higher income thresholds to purchase a home drive up prices and risk widening wealth gaps.
Supply Dynamics: Expanding overall inventory alone has limited effect; targeted affordability measures for first‐time buyers are more effective.
Policy Implications:
Monitor rental‐market metrics to anticipate housing‐market movements.
Address equity concerns by designing programs (e.g., down‐payment assistance, buyer tax credits) that lower income barriers to homeownership.
Research Limitations: The model's low explanatory power (R² = 19.6%) indicates substantial unexplained variation. Critical factors like interest rates, migration patterns, and zoning regulations were not included. There is a class imbalance issue in the Stay Home Policy variable because the policy was active for only about one year across 2020, while the data spanned 2018-2025. This makes its effect statistically undetectable.
©2026 design