Abstract
Using a data set from a Central Virginia MLS during 2001–2015, we find evidence that the uncertainty in the housing market before an election affects the number of listings and prices, though there is no statistically significant evidence for changes in number of transactions and time on the market. Although the number of listings increases before an election, both the direction and magnitude of price changes depend upon the type of election, in particular on incumbency effects and polling spreads. For presidential elections without an incumbent running and candidates equal in the polls, we find that on average election uncertainty is associated with a 3.1% ($6,720) decrease in price, whereas with an incumbent price as an increase by about 2.7% ($6,050). Additionally, homes in gubernatorial elections in Virginia on average experience 3.2% increase in price before an election ($7,100), suggesting substantial heterogeneity across different types of elections. Our model explains these changes in price and listings as a market characterized by relatively inelastic housing supply that experiences either a reduction in the future price level or an increase in the variance in future price levels.