Dynamic price–volume causality in the American housing market: A signal of market conditions

Abstract

This study analyzes the dynamic price–volume causality in the American housing market using the average price and transaction volume of existing houses in the United States from January 1999 to December 2015. A rolling window sample is used for estimation in the bootstrap Granger causality test. The results reveal that the transaction volume tends to be informative during price rigidity. In particular, the housing price tends to lag the volume when the information on housing price decreases is required for market correction. The housing price tends to be informative during volume rigidity, particularly during the substantial increase in housing price and the reduced transaction volume caused by the sellers’ reluctance to sell. The dynamic causality estimation explains that the price–volume relationship varies according to market conditions. Under normal circumstances, both the price and volume efficiently react to the information without a lead–lag relationship between the two. However, during a housing market boom or downturn, a lead–lag relationship between price and volume exists. This paper infers that the existence of a lead–lag relationship between price and volume can be a signal of housing market conditions.