House price cycles in Australia's four largest capital cities

The aims of this paper are twofold. First, we propose a threshold cube equation, augmented with the optimal number of autoregressive terms, to model the likely future path of house prices in Australia's four major capital cities. The proposed framework can be adopted to simulate house prices in other cities experiencing similar price cycles (spirals) such as London and Vancouver. Second, using a proprietary dataset not publicly available (1995m12-2015m10), we use the proposed model to simulate house prices in Brisbane, Melbourne, Perth and Sydney, which have been on a steep upward trajectory since 2013. To check the sensitivity and robustness of our results, we evaluate the model in terms of the out-of-sample accuracy for two separate 12-month periods. We find that the forecasting performance of the model appears to be reasonable. Then we use the model to provide the ex ante future path of house prices during the next twelve months (2015m11-2016m10). Although we do not seek to forecast all see-saw changes in property prices, based on the historical length of boom and bust cycles in the past, some interesting overall price paths are detected. It is observed that during the next year house prices in all capital cities will start plateauing with some likely moderate falls. This downward price adjustment is particularly more noticeable in Perth.