Forecasting Australian Real House Price Index: A Comparison of Time Series and Machine Learning Methods

Abstract

We employ forty-seven different algorithms to forecast Australian log real house prices and growth rates, and compare their ability to produce accurate out-of-sample predictions. The algorithms, which are specified in both single- and multi-equation frameworks, consist of traditional time-series models, machine learning (ML) procedures and deep learning neural networks. A method is adopted to compute iterated multi-step forecasts from nonlinear ML specifications. While the rankings of forecast accuracy depend on the length of the forecast horizon, as well as on the choice of the dependent variable (log price or growth rate), a few generalisations can be made. For 1- and 2-quarter ahead forecasts we find a large number of algorithms which outperform the random walk with drift benchmark. We also report several such outperformances at longer horizons of 4 and 8 quarters, although these are not statistically signifcant at any conventional level. Six of the eight top forecasts (4 horizons x 2 dependent variables) are generated by the same algorithm, namely a linear support vector regressor (SVR). The other two highest-ranked forecasts are produced as simple mean forecast combinations. Linear ARMA and VAR models produce accurate 1-quarter ahead predictions, while forecasts generated by deep learning nets rank well across medium and long forecast horizons.