Affordability of housing has become a key challenge for the local authorities to provide a better living quality and liveability of many cities. Affordable housing provides positive externalities in terms of social stability, public welfare and economic development of a city. As in Taiwan, the median house price to median annual household income ratio is 15.7 times in 2014, which is at the seriously unaffordability level. House price index in Taiwan averaged 186.27 index points from 2001 until 2017, reaching an all-time high of 297.78 index points in the second quarter of 2014. The house price movements are affected by both macro-economic determinants such as loan to housing, unemployment rate, per capital income and inflation rate; and micro-perspective like build-up area, number of bedrooms, and distance to the nearest central business district. Most studies applied time-series models or regression models to predict the housing price with a set of macroeconomic determinants or a set of micro-perspective factors. Nonetheless, none of the studies considered the explanatory variables subject to errors, i.e. autocorrelated errors. This study aims to investigate the relationship between the housing price index and a set of macroeconomic determinants in Taiwan. An autocorrelated unreplicated linear functional relationship model is developed to identify changes in the Taiwan’s residential market in terms of significant macroeconomic determinants and their corresponding effect on house values. The main contribution of this study is that the proposed model considers both the housing prices and the macroeconomic determinants are all subject to autocorrelated errors, and hence it leads to a combination of the time-series ARIMA(r,d,q) model and the multiple unreplicated linear functional relationship model (ULFR).
This research was financially supported by MOFA Taiwan Fellowship 2018. This talk is based on the join work with Looi Sing Yan.