The Non-linear Ripple Effect of Housing Prices in Taiwan: A Smooth Transition Regressive Model

Abstract

Different from past research of regional house prices, this paper employs the smooth transition regression (STR) model to investigate the ripple effects among regional house prices in Taiwan. The aim of this paper is to test whether a smooth transition regression model, which is capable of capturing this non-linear behaviour, shows a better characterisation of regional house prices than a linear model. Our main findings are as follows. First, the results reveal that there is a mixture of linear and smooth non-linear ripple effects in Taiwan's regional house prices. If there is enough time to adjust, then the non-linear behavior in house prices will exist, causing a non-linear ripple effect for the regional house prices. Second, a large portion of the variance in Taiwan's overall house price changes is associated with Taipei's house price changes, and Kaohsiung's house price changes do not effectively impact Taiwan's overall house price changes. Third, strong evidence shows that the STR models capture the smooth non-linear ripple effects of regional house prices in Taipei and Taichung, and several features identify the superiority of the non-linear estimation. When compared with Taichung, there is a faster speed of transition between the regimes in Taipei. Fourth and finally, the estimated transition parameter value, revealing the halfway point between two regimes, is lower in Taipei and implies that a change in Taipei's house prices can more easily cause a non-linear "ripple out" change in Taiwan's overall house prices.