Projection bias in the residential property market : how do weather and shelter affect the housing price in Hong Kong?

Abstract

From prehistoric caves to modern residential towers, one of the primary purposes of a home has been to provide a shelter from undesirable weather conditions. Since ancient times, it has been recognized that weather can influence the human mind. In the modern era, researches on such effects in finance and real estate have been conducted in the frameworks of the Efficient Market Hypothesis (EMH), which assumes that human beings can make rational decisions, and behavioural economics, which challenges this assumption based on cognitive limitations. One of the major challenges is projection bias, which refers to the psychological tendency to overpredict future preferences based on the current preferences. Projection bias can explain the irrationality when the decision to buy a property is influenced by the weather at the time of decision-making. Its effects have been studied in other countries, but it has not been studied in the context of covered walkways in Hong Kong, with its unique features of subtropical climate and high reliance on public transport. Besides, most previous researches focus on the individual effects of variables. Hence, this paper would like to fill in these two research gaps by investigating the interaction effects between weather and shelter on the housing price. “Shelter” refers to covered walkways e.g. footbridges or underground passage. Therefore, the main objective of this research is to examine the interaction effects of weather on the premium of the presence of covered walkways to an MTR station for the market value of a residential unit. Apart from filling in research gaps, the contributions of this study include helping different stakeholders to make better decisions and gaining a deeper understanding of the Hong Kong property market through a multi-disciplinary approach. In the literature review, international peer-reviewed academic papers were important for a basic understanding of the concepts of projection bias, behavioural economics, EMH, rational choice theory, bounded rationality, transit-oriented development, etc. By reviewing a wide range of previous studies across different countries, it can be concluded that projection bias has been observed in real estate, finance and various aspects of daily life. Based on the literature review, a hypothetical model was formulated: a property with covered walkways to the nearest MTR exit has a price premium over a property without them, holding other factors constant. Under a rainy (hot) weather, the coverage (air-conditioning) of a property will have a larger positive effect than under a non-rainy (non-hot) weather. The methodology consists of both the qualitative methods of personal interview and site inspections as well as the quantitative method of secondary data analysis. The sources of data include the EPRC database, peer-reviewed academic papers, government statistics, etc. In the personal interview, the interviewee was a licensed property agent with over 4 years of experiences in the transactions of residential properties in Taikoo. The interview results provided insights into transaction procedures, housing attributes, property valuation, etc. In the site inspections, personal visits to Lai Chi Kok and Taikoo clarified several issues. In the secondary data analysis, the hedonic pricing model was adopted with hierarchical multiple regression analysis and the ordinary least squares (OLS) approach. Diagnostic tests were conducted to confirm that the data analysis complied with the assumptions. The proposed model include both basic attributes and interaction terms. To ensure a representative sample, 3 data sets covering different regions, different types of properties, different forms of covered walkways, and different types of comparisons were selected around the MTR stations of Tsuen Wan, Lai Chi Kok and Taikoo, with justifications. Regarding the results, the Tsuen Wan dataset showed no statistical significance of either the existence of covered walkways or the interactions, while Lai Chi Kok only had the individual term and Taikoo only had the interaction term as being statistically significant with a p-value below 0.05 in their respective final models. There were three possible explanations. Firstly, the Tsuen Wan case might be attributed to the high efficiency of the Hong Kong property market according to the interviewee. Nevertheless, this did not explain the statistical significances of the Lai Chi Kok or Taikoo cases. The second explanation was that the selected Tsuen Wan properties were more heterogeneous, so covered walkways were a defining feature in the other two to distinguish different flats but only as one of the many factors in Tsuen Wan. The third explanation accounted for the differences between Lai Chi Kok and Taikoo that the real-life interactions were consistent with the mathematical one. It is believed that the findings can help governments, developers, property agents and individuals to make better decisions. Limitations of this study include an insufficient sample size and lack of further analysis due to time constraints. Future researches may apply the methodology to a larger sample covering more districts and more types of interactions.