Exploring Housing Market Dynamics through Google Search : A Case of Taiwan

Sammanfattning: To capture house price fluctuations, it is important to combine appropriate factors into the price forecasting model. Fundamental macroeconomic variables have been considered quite completely in many housing price models. However, the ability of these models to predict housing prices are still limited. According to Shiller (2007), it is the psychological factors that cause fluctuations in the housing market. To fully present the current psychological state of the market, researchers need a more powerful database. To solve this problem, Google Trends seems to be a useful tool that provides us the Google search engine indices of specific terms or predetermined categories within a specified time and region. By tracking the search intensity on the Internet, this study aimed to uncover the intentions of potential buyers and used this information to analyze the dynamics of the housing market. This study performed an empirical analysis with data from Taiwan. The purpose is to detect the interaction between the Internet search index and house price and transaction volume in Taiwan and the six main Taiwanese cities. The Internet search information provided by Google Trends is presented in an index form. The data is anonymous, free of charge, and frequently updated. The VEC models were performed to measure the explanatory power of the search volume indices on house prices and sales volume. Moreover, variance decomposition and impulse response were used to examine the dynamic of the variables in the model. The findings reveal that, in Taiwan and its five out of six cities, the Google indicator using the names of estate agencies as the search query could serve as an indicator of transaction volume but not for house price. By proving that the Internet search volume could capture the market sentiment for transaction volume in Taiwan and its cities, it would help the local government and decision-makers in related companies to make more precise predictions of housing market based on market sentiment with a lower cost.