**Sigh** The dreaming goes on. Where are the 40% falls??? Everytime I come back here I get the same renters dreaming of the great times to come that Keen has promised them.
Ah well, keep waiting and good luck with it. The article below is well written and should help understand differences in reporting figures.
From http://www.businessspectator.com.au/bs.nsf/Article/Which-way-for-house-prices-pd20090504-RQ7AZ?OpenDocument&src=is&is=Property&blog=Concrete%20Detail
Posted 4 May 2009 2:39 PM
Which way for house prices To make life a little confusing today, the ABS established house price index came out with a negative result (-2.2 per cent) for the March 2009 quarter.
This outcome is bizarre because it conflicts starkly with what two other major house price index providers concluded using broadly similar data (including one that has historically provided more conservative estimates of capital growth than the ABS).
According to Australian Property Monitors (APM), Australian house prices rose by 0.1 per cent in the first quarter of 2009, while unit prices increased even more strongly by 0.5 per cent. This is consistent with strong housing finance and auction clearance rates data that have been coming through since the end of 2008.
Including all dwellings, APM’s data suggests that Australian house prices grew by just under 0.2 per cent over the quarter (i.e. nothing remotely like the ABS numbers).
APM use what is known as a “stratified median price index” method to measure house price changes through time.
This is a simple way of controlling for the notorious “compositional biases” that adversely affected the old median house price indices in days gone by.
In short, if you have more, say, first time buyers purchasing lower-end properties during a period (which is what has been happening), a median price index would imply that house prices had fallen simply because there had been a larger volume of cheap properties trade.
In truth, though, house prices would not necessarily have changed at all and could, in fact, have been rising.
The ABS house price index also uses a stratified median price index method, which should in theory give very similar results to the APM index.
Yet according to the ABS, Australian established house prices actually fell by 2.2 per cent over the first quarter of 2009.
This is a dramatically different outcome to the APM findings.
One large source of bias within the ABS numbers is the fact that the ABS only considers "detached” houses. The ABS index does not include “attached” forms of housing, such as apartments, terraces and semi-detached properties.
This results in the ABS index being biased towards both more valuable properties (since the median house price is around $480,000 whereas the median unit price is $390,000) and, by definition, just houses, which are less likely to be purchased by first-time buyers because of both their more expensive price and their larger size.
The final major index provider is RP Data-Rismark.
The ASX-listed RP Data has Australia’s largest property information database and uses a “hedonic” regression approach to precisely control compositional biases in the observed property sales data over time.
This is a much more sophisticated statistical method than that which is used by either the ABS or APM and has historically generated lower estimates of capital growth and volatility (or 'noise').
In brief, the RP Data-Rismark Hedonic Index measures the actual changes in dwelling values through time rather than changes arising as a function of compositional bias (e.g. more western Sydney property trading than homes in the eastern suburbs).
Worryingly for the ABS, both the APM stratified median price index and the RP Data-Rismark Hedonic index arrived at nearly identical results.
On a quarterly basis, the RP Data-Rismark Index was up 0.1 per cent in the period to end March 2009 (which compares well with the APM result of circa +0.2 per cent).
Since most of the raw data for all three index providers is sourced from the same Valuer Generals offices, the ABS numbers are hard to understand.
Even when RP Data-Rismark computes a stratified median index (which is what is used by both APM and the ABS) using only data for houses (i.e. not apartments etc) this index is still down just 0.4 per cent for the first quarter of 2009 – nothing like the ABS numbers.
At the end of the day, any median price index is going to be subject to compositional bias that gives rise to a lot of noise (or error). In the ABS’s case, this bias has been exacerbated by relying only on data pertaining to houses and ignoring altogether the semi-detached, terrace and unit market (effectively up to 30 per cent of all sales in the current market).
As numerous academic research papers have found, and Moody’s Economy.com recently concluded in an independent review, hedonic house price indices afford by far the most accurate measure of dwelling price changes through time.
Unfortunately, the more volatile median price indices tend to overstate both the ups and downs, which is probably why we have this odd ABS result.
