Spatio-Temporal Effects of an Urban Development Announcement and Its Cancellation on House Prices – A Quasi-Natural Experiment

Abstract

We use a spatio-temporal autoregressive difference-in-differences (SDID) framework to assess the strength of the effect, distributed across space and time, of announcements associated with urban development on house prices. We use a quasi-natural experiment of a large-scale urban redevelopment project in Seoul, South Korea, first announced in 2007 and cancelled six years later without any construction having taken place. Using a rich dataset comprising of 21,200 apartment transactions between 2006 and 2015, we find that the development announcement effect accounts for an increase in apartment transaction prices by up to 7.3% depending on the location of the property from the project site. Prices decrease by up to 5.2% following the cancellation of the project. The SDID captures significant spatial autoregressive effects accounting for up to 15% of the changes in transaction prices, reducing the announcement effects from a standard DID model.

Keywords: Announcement effect, spatial spillover, spatio-temporal autoregressive model, difference-in-differences, apartment transaction prices.

JEL Classification: C21, R15, R31, R53