top of page

Is there a cyclical relationship between dwelling values and GDP In Australia 2020?

To see if there is a relationship between house prices and GDP the most obvious starting point is to compare the Australian property cycle and real business cycle. As each region in Australia is at a slightly different point in the property cycle, but shares the same national GDP figure, Sydney was used as the base economy, as it roughly accounts for a quarter of national GDP (24.4% from 2018-2019), to increase accuracy. It is evident that Sydney is currently going through a correction phase of the property cycle, given dwelling prices are down 13.2% from their July 2017 peak (Commbank, 2019). In comparison Sydney's GDP has continued to increase year on year, despite house prices now falling. However studying Sydney's annual GDP growth rate, as in Figure 0, we can see that in 2016 Sydney's GDP growth rate peaked at 4.2% (Figure 0) and has relatively declined since - a growth theory recession. This is the same pattern as the property cycle, just 1 year earlier. This begs the question; is GDP and the annual growth rate of GDP a predictor for dwelling values and if so how highly correlated are they?

Long-run data for Australia show that house prices have increased more than tenfold since 1870 in real terms, and more recently, between 1991 and 2012 Australian real house prices have nearly doubled (Knoll et al, 2017). The time path of the trend, follows a “hockey stick” pattern; real prices were stable until the 1960s and increased strongly since then (Knoll et al, 2017).

The Australian GDP growth rate has moved far less than the Australian Real House Price Index, as above. The GDP growth rate has relatively declined from upwards of 4% in the 1990s, towards a growth rate of 2.3% at end 2018 (Trading Economics, 2019). Yet from 1993-2018, median dwelling values increased by 412% (Aussie, 2018). In our own analysis, the below results were obtained, signifying positive growth in each variable (Figure 3).

To further analyse the questions hypothesis, we used GDP, House Values and Total Returns data to obtain a growth cycle, consistent with Lucas (1977). The HP filter allowed us to remove the long term growth trend by filtering out low frequency components, to reveal the secular growth trend. It appears there is some movement of the cycles together, however the fluctuations in housing values are much more volatile, while GDP hovers around the average growth rate.

GDP appears to be a leading indicator for the housing cycle. A higher GDP in a previous period appears to drive higher house prices in the following period, yet the fluctuations are much larger in both returns and values than GDP. Our calculations show housing values and total returns are 99.4% correlated. This shows that yield and price move together. From 1980-2018, the series data shows GDP and house values were 38.9% correlated. However note, from 2005-2018, GDP and both of the other values were -0.12% correlated. This needs to be explored further to determine if this is as a results of unforeseen economic shocks such as the Global ‘09 Financial Crisis, or if there is another causation/correlation between GDP and Dwelling values, although we will not be focusing on this specifically in this report.

The less volatile standard deviation results (above) makes sense, as the largest criticism of the log difference method for detrending is that it removes too much information.

Thus, as evidenced by Mohler and van der Merwe (2015), cyclical factors appear to drive dwelling price growth over shorter periods, yet there are other factors at play (Mohler and van der Merwe, 2015). Monetary policy has a heavy influence on the housing market, both in the established market through higher activity (intertemporal substitution effect) and in the new dwelling market by encouraging dwelling investment. Otto (2007) stated low mortgage interest rates can explain most of the Australian capital city dwelling price growth prior to 2017 (Mohler and van der Merwe, 2015). Lower interest rates increase GDP (classical IS-LM analysis) and according to Otto, house prices as well. This implies that from a given point, enacting expansionary monetary policy would increase GDP and dwelling values, thus implying a correlation and causation, which will be further explored in Sections 2 and 3.