Abstract/Summary

This thesis investigates herding behaviour across US metropolitan housing markets. Herding, the conscious or unconscious copying of behaviour by individuals on a large scale, is a phenomenon commonly observed across markets and locations. Firstly, the presence of herding and reverse herding, the latter a rarely discussed phenomenon, is identified across major US urban housing markets. There is broad spatial and temporal variation, and reverse herding is found to be far more prevalent than herding. Market inefficiencies and the cost of obtaining private information, along with innate homeowner overconfidence, are presented as the theoretical context for these results. Secondly, the commonly discussed connection between herding and price bubbles is estimated using several measures of bubble activity to ensure robustness. There is evidence that extreme price growth may be a significant determinant of herding behaviour, with stronger effects observed in larger gateway cities that draw investment. The results indicate that herding behaviour may be triggered by speculation rather than information asymmetries. Lastly, the determinants of herding and reverse herding behaviour at the urban market level are estimated using a discrete choice model. Various social and housing factors are significant determinants of rational behaviour, as opposed to the overall economic conditions. The scale and role of housing demonstrates the significance of these findings. Whilst some results are in line with previous literature, there are sufficient variations to illustrate the importance of market structure and location. It also illustrates that further work is required to disentangle rational and irrational dynamics, and that leading indicators could be constructed for practical use in risk management and portfolio construction, in addition to any value for policy making.