Determinants of Housing Prices in Hong Kong: A Box-Cox Quantile Regression Approach

This paper analyzes the determinants of housing prices in Hong Kong by using property transaction data of condominium units from Taikoo Shing, one of the largest real estate properties in Hong Kong. We use a hedonic pricing model for the empirical analysis and estimate the model by using the Box-Cox quantile regression method. The empirical results show that this method provides a more comprehensive description of housing price determinants. Housing prices and characteristics have a nonlinear relationship, and this relationship varies across all quantiles. In addition, the response of housing prices to various housing characteristics varies across quantiles. For example, an increase in the size of the gross floor area is more valued at higher quantiles. Other variables have differential effects on housing prices across the distribution of housing prices. We also perform a simple simulation for model predictability and show that our model outperforms other models which have been frequently used in the previous studies.