Dynamic Search in a Non-Stationary Search Environment: An Application to the Beijing Housing Market

This paper studies how dynamic changes in the search environment affect consumer search and purchase behavior. We develop a dynamic model that incorporates a non-stationary search environment and propose a feasible estimation procedure to estimate its parameters. We apply our model and estimation procedure to the Beijing housing market, utilizing detailed data on consumers' complete search records. We show that accounting for dynamics is crucial for accurately estimating search costs. Additionally, we find that search environment dynamics have a significant impact on consumer decisions and welfare. Housing supply policies that alter search environment dynamics—by increasing the number of new listings and slowing down price increases—benefit consumers, primarily by incentivizing longer searches, more property visits, and ultimately leading to purchases that yield higher utility.