An Estimation of Explosive Bubbles in House Prices

This study examines the existence of explosive bubbles in house prices. Even if there is a bubble in the house prices, as long as it is not explosive, we can not conclude that there exists a bubble only based on the fact that house prices soar or are too high because house prices can be restored to a proper level after a certain period of adjustment. Therefore, using Information Error model, we measure the bubble term, which is defined as the gap between the intrinsic value explained by the market fundamentals and the market price, and then check whether there is an explosive bubble by Survival analysis. Housing Sales Price Index as a dependent variable, and explanatory variables such as not only Jeonse Price Index but also Sales-Jeonse Prices ratio, which reflect properly the relation between sales and rental prices, are included for Information Error model. Survival analysis is performed by an exogenous as well as an endogenous bubble model. Survival times of bubbles are assumed to be independent of market fundamentals in exogenous bubbles model while dependent in endogenous bubbles model. Samples are the period from January 1987 to December 2017 and six sub sample periods. Regional ranges are for the four regions of the whole country, the six metropolitan cities, Seoul Gangnam, and Seoul Gangbuk. For the exogenous bubbles model, 'Not' in the whole sample 'nor' any sub-samples, the existence of explosive bubbles was found. For the endogenous bubbles model in which ‘interest rate’ are used as market fundamental variables, shape parameter  is estimated to be     . Values of the shape parameter,  , with larger than 2 are not found in both of exogenous and endogenous bubbles model of Survival analysis, so we can not find any evidence supporting existence of explosive bubbles. It implies that there are no explosive but stable bubbles in any sub-sample, and the bubbles become extinct over time.