This weeks post is a little off topic, however, the subject of Australian house prices is getting quite a bit of press, and I thought I would post some of my own analysis.
Considering that a significant proportion of home lending in Australia is financed from overseas lenders (this article suggests 30% http://www.smh.com.au/business/banks-pave-the-way-for-surplus-rate-rises-20101001-1612p.html), I thought I would approach the analysis by taking a look at Australia's current account deficit (CAD).
Treasury defines the CAD as :
The CAD measures the extent to which Australia draws on foreign savings to fund that portion of national investment that is not funded by domestic national savings.
http://www.treasury.gov.au/documents/1087/PDF/02_ABE_Keynote.pdf
CAD is influenced by exports and imports, particularly the price and volume of resources that Australia exports relative to the price and volume of goods and services that are imported. The surplus of imports over exports essentially has to be funded by overseas lending or equity investments - capital inflows.
This analysis hypothesises that this overseas lending is a key source of funds for home lending, and that the CAD might have some predictive value in determining future house prices.
You can download the CAD history from the Australian Bureau of Statistics
(http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/5302.0Jun%202010?OpenDocument).
You can download the Australian historical house price index (HPI) from the same site
(http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/6416.0Jun%202010?OpenDocument).
So, taking these two data sets (and setting to a common base) for the period June 1986 to June 2010 I have used excel to do a basic linear regression analysis.
So we have a pretty good match with an R squared of 73.6%. So there seems to be at least some sort of relationship between the two data sets.
Now if we exclude the outlier in the bottom right hand corner of the chart and run the analysis again we get 80.6%.
So what is the significance of the outlier - we cant just exclude data from our analysis because it doesn't fit our expectations. Well the significance of the outlier is that it is the June 2010 result, so all we have really done by excluding the outlier is run the analysis as at March 2010.
So lets have a look at the data from another perspective. The chart below looks at the two data sets over the timescale (I have reset both index's as at 1986) :
So the chart above certainly suggests there is some relationship between the CAD and the HPI over the period of analysis - and in fact we have shown with our regression that R squared is 73.6% (or 80.6% if we exclude the last quarter's data).
Treasury itself in 2005 even alludes to this relationship :
Over the next several years, however, the share of Australian GDP devoted to dwelling investment may decline gradually as the house-price boom continues to wane, reducing this contribution to the CAD.
http://www.treasury.gov.au/documents/1087/PDF/02_ABE_Keynote.pdf
The Treasury paper, however, considers house prices as a determinant of CAD not CAD as a determinant of house prices. Our analysis, however, does not indicate causation and hypothesises that CAD may be partially at least, determinant of HPI.
Interesting also, that Alan Greenspan also makes reference to the strong correlation between home mortgages in the US, and the size of the US CAD :
Interestingly, the change in U.S. home mortgage debt over the past half-century correlates significantly with our current account deficit.6
http://www.federalreserve.gov/boardDocs/speeches/2005/20050204/
So if we continue this line of argument, the next question is where is the CAD heading ? If we know whether the CAD is increasing or decreasing, we can hypothesise the effect on the HPI.
Fortunately, Treasury gives us some guidance in the 2010/11 budget :
The current account deficit (CAD) is expected to narrow from 4¾ per cent of GDP in 2009‑10 to 3¾ per cent in 2010‑11, with the trade balance moving temporarily into surplus. Export incomes are expected to rise substantially in 2010‑11, largely driven by increased commodity prices, but also by higher commodity export volumes. A wider net income deficit is expected to offset the improved trade balance somewhat, with a significant proportion of increased mining profits flowing to overseas investors. The CAD is expected to widen to 5 per cent of GDP in 2011‑12, a0arial, helvetica, cleank into deficit, as commodity prices fall slightly and import volumes continue to grow strongly.
http://www.budget.gov.au/2010-11/content/bp1/html/bp1_bst2-03.htm
So we can expect a 20% decrease in CAD during 2010/11 according to Treasury's budget paper. The June 2010 actual CAD indicates that Treasury's estimates might be on the conservative side. In any event, utilising the estimate of 20% decrease in CAD, what does this equate to in terms of expected HPI.
Referring back to our linear regressions above the gradient in each of the regression calculations is around 0.8 - in other words Δ CPI = 0.8 Δ HPI
Rearranging the formula, Δ HPI = 1.25 Δ CAD
So a 20% decline in CAD, according to our hypothesis, results in a 25% decline in HPI.
So with the last quarter's CAD declining by 60% as shown on our chart above, and current analysis of the news suggesting that this trend might be maintained, this blog hypothesises that we may soon see some correction in the HPI. We will watch and wait for the September quarter results of both the CAD and the HPI and post a follow up.
