Housing and behavioural factors

Abstract

The field of behavioural finance is well established. Studies of housing bubbles are less common than studies of equity overvaluation and studies of the behavioural aspects of real-estate pricing are rare. This article examines the housing market from 1987 to the present with particular attention to the 2003 to 2007 bubble. Both behavioural and traditional variables are considered. Models with behavioural variables are seen to provide a better fit during the housing bubble. A state-space model shows that the coefficients on those behavioural variables vary in ways consistent with the emergence and dissipation of the recent housing bubble.

Keywords:

• behavioural
• housing
• bubbles
• markets
• Robert Shiller

JEL Classification:

• R31
• R21

Notes

¹ Data for all variables were obtained from the Federal Reserve Bank of St. Louis.

² 2011 Annual Table 24. Selected Reasons for Opinions About Buying Conditions for Houses.

³ Following Leamer, the lag structure is limited to a single period. Adding additional lagged terms increases the values of the Akaike, Schwarz and Hannan–Quinn criteria for lag length selection.

⁴ The result is also in opposition to the general views of Akerlof and Shiller (2009), who describe the money illusion principle in favourable terms.

⁵ Following Leamer, the lag structure is limited to a single period. Adding additional lagged terms increases the values of the Akaike, Schwarz and Hannan–Quinn criteria for lag length selection.

⁶ Since this is a recursive process, the initial estimates of the coefficients are based on very few observations and thus are often volatile. Including them in the graph changes the scale to match the early extreme values and makes later, more gradual changes in the coefficients difficult to see. For that reason, the graphs begin in 1992.