The dual role of sentiment on housing prices in China

This research investigated the role of sentiment on housing prices in China from 2013 to 2022 within a time-varying framework, highlighting the underexplored liquidity channel. We constructed a housing sentiment index by analyzing over 2.44 million social media posts using the Bidirectional Encoder Representations from Transformers (BERT) technique, which is recognized for its advanced capability in processing natural language. With a Time-Varying Coefficient Vector Autoregression (TVC-VAR) model, we estimated the time-varying impulse response of housing prices to sentiment and liquidity shocks, as well as the response of liquidity to sentiment shocks. Our findings revealed that positive public sentiment not only directly pushes housing prices upwards but also indirectly inflates them through an enhanced liquidity, a byproduct of optimism. Notably, the dual role of sentiment becomes more pronounced during the periods with high uncertainty, such as COVID-19 pandemic period. Moreover, our conclusions survived a battery of robustness checks. The results underscore the importance of integrating psychological factors and market dynamics to understand the complexities of real estate markets, particularly in turbulent times.