Dr Jonathan Reeves

An array of forecasts for Australian house prices ranging from modest rises to dramatic falls currently exist. These large forecast differences are primarily explained by the dataset used by the forecaster.

Most Australian forecasters rely on Australian house price index returns, rather than overseas house price index returns. Whereas, overseas forecasters of Australian house prices, often combine both Australian and overseas house price index returns in their analysis.

The major limitation with the Australian house price indexes is that most commence around the early 1980’s. Providing a relatively short sample of approximately thirty years of quarterly data. Whereas, in the US and Europe house price index data exist for over one hundred years.

Australian inflation adjusted house price indexes display substantial average growth rates over the last thirty years, whereas the US and European inflation adjusted average housing growth rates over the last century are relatively modest.

When Australian house price returns are modeled over the last thirty years, typically correlations are measured between house price returns and factors such as population growth and interest rates. These models have some usefulness in forecasting three to twelve months ahead, though generally are not suitable for longer range forecasting.

Housing for most investors is a long term investment. Thus forecast horizons of three, five and ten years are more relevant than the three to twelve month horizon. For these longer range forecasts, there is simply insufficient historical Australian data available for reliable models to be constructed.

Instead, Australian forecasters should more often use overseas data on the long run behaviour of house prices. In the practice of forecasting, it is often found that the relatively simple models are the most reliable.

In house price modeling with US and European data over the last century, the simple model of long run average house prices being a multiple of four times the average gross household salary, has proved to be a reliable guide. This multiple of four is the average over many decades, with periods where the multiple is above and below.

For example, before the global financial crisis, the multiple in the US was over six and now it is currently three. The economic rational behind long run house prices not staying above a multiple of four for long periods, is that households simply can not afford this through all phases of the business cycle.

Currently in Australia the multiple is over six and thus likely to fall. Otherwise, by current comparisons to other developed countries and by comparisons to the historical record of house prices in many countries, Australia would be exceptional.

Some forecasters consider there to be a shortage of land in Australia. However, this is at most a short run factor as governments have the ability to rezone land for higher density housing.

The magnitude of a fall in Australian house prices will most likely be influenced by overseas events. The most immediate threat may come from the European sovereign debt crisis and the impact this could have on raising mortgage interest rates in Australia. Business cycles transmitted by China are also a high concern, along with the performance of the US economy.

In addition, falling house prices in New Zealand, with a weak New Zealand economy, may have substantial impacts on Australian banks operating in New Zealand.

In contrast to other developed countries, Australian house prices rose in 2009 and 2010. Australia’s rise was due to the government stimulas package that included specific housing stimulas. With the stimulas effects fading in 2011, Australian house prices have started to decline. It is likely that this is the beginning of a substantial housing correction.

Dr Jonathan Reeves is a financial economist at the Australian School of Business, University of New South Wales.

A version of this opinion piece appeared in the Australian Financial Review on the 19 July 2011.