Spatio-temporal effects of an urban development announcement and its cancellation on house prices: A quasi-natural experiment

We use a spatio-temporal autoregressive difference-in-differences (SDID) framework to assess the strength of announcement effects associated with an urban development on house prices over time and across locations. The quasi-natural experimental setting of a large-scale urban redevelopment project in Seoul, South Korea, announced in 2007 and cancelled six years later without any construction having taken place, allows us to cleanly disentangle the announcement affect from the actual development effect. Using more than 21,000 apartment transactions between 2006 and 2015, we find that the development announcement increases apartment prices between 2.4% and 7.3% for properties within 1 km radius of the project site. However, for buildings beyond 1 km, the effect becomes negative suggesting spatial redistribution of housing demand. The cancellation of the project leads to a significantly decline in prices between 3.5% and 5.2% for apartments within 1 km from the project site nearly cancelling the associated positive effects. We find that properties which are located around the project site but are not in direct proximity actually have a significant price cut after each announcement. Overall, we show that announcement and cancellation of development projects have a significant impact on residential property prices near the site but vary considerably in strength and direction across space.