Housing Supply Elasticities: A Structural Vector Autoregression Approach

We estimate the first time‐varying, county‐level index of housing supply elasticities across the United States, using a structural vector autoregression model with sign restrictions to identify the effect of a positive demand shock. This index reveals four features of the supply curve in different local housing markets. First, supply elasticities are low, on average, generating faster price growth than housing unit growth for most of the country. Second, there is significant variation in elasticities across counties, both urban and rural. Third, elasticities have declined in more counties than they have risen over time, particularly in large urban markets. Fourth, these new estimates predict housing quantity growth better than previous models in the literature.