Home' API Magazine : October 2014 Contents 48 n APIMAGAZINE.COM.AU n OCTOBER 2014
AUSTRALIANS HAVE A STRONG HISTORICAL COMMITMENT TO BOTH
home ownership and residential investment. Home ownership
rates have hovered around 70 per cent for decades with residential
investors typically accounting for up to half of all home sales. There
are numerous drivers of high home ownership rates and investment
levels but the long-term solid growth in Australian house prices is a
key to resilient demand for housing. This financially-focused demand
is enhanced though specific taxation advantages such as tax-free
capital gains for homeowners, and negative gearing and depreciation
allowances for investors.
The importance of housing as a significant capital asset and
investment vehicle ensures the need and demand for effective models
of house price growth that accurately reflect underlying levels of
House price modelling however remains problematic, particularly
over the shorter-term, given the discrete nature of home sales. Every
property transaction remains essentially unique geographically as do
the circumstances and motivations of individual buyers and sellers.
Markets also remain unique in their local demand and supply
drivers, not only between capital cities, regions and
suburbs but also between adjacent streets, within
the same street and the same building.
Examples of localised price disparity are revealed
on auction-based TV reality shows such as The Block
where similar properties in the same building can
be sold for widely varying prices over a period of
mere minutes. This clearly reflects the random price
diversity of the micro-market at its most granular.
The limited number of transactions recorded
over a given period of time is another constraint to effective house
price modelling with around six per cent of housing stock typically
sold in a year. Relatively small sale numbers provides a low base
for modelling underlying price growth levels and revealing the
aggregated supply and demand energy of the total population
The unique nature of property sales also impacts effective house
price modelling through the price mix of properties typically sold
over a given period. A greater proportion of higher-priced property
sales can skew prices upwards compared to the previous period with
the reverse impact if a greater proportion of lower-priced properties
are sold between periods. This price-mix disparity is further
complicated when there are low numbers of sales in a period.
Another impediment to house price modelling is the impact of
multi-speed market segments at the regional or capital city level,
where some sections of the market may be performing at significantly
different levels to others.
Data collection lags can also inhibit effective and timely house
price modelling through delays with the collection and processing of
official dwelling sales records from the actual date of sale.
Some seasonal effects due to higher levels of activity in the autumn
and spring selling seasons, which skew towards higher-priced home
sales, complicate the issue further.
A number of methodological approaches are typically used to
model underlying house price activity. These include median house
price models that measure the mid-point of a series of sales within
a suburb, region or city. Hedonic models relate the price of a house
to common elements such as number of bedrooms, number of
bathrooms, size of garage, land size and floor area. The repeat-sales
method compares the sale price of individual houses to the
previous sale price.
A variation of the simple median house price methodology is the
stratified or compositionally-adjusted model. Sales data from a region
is divided into various sectors to modify the effects of an over-
representation of high or low value sales in a given period.
The median price for each of these sectors is calculated, then all the
various medians are combined to estimate an overall median house
price for the region. The Australian Bureau of Statistics (ABS), in its
quarterly house price series, uses this methodology. In addition to
moderating the impact of compositional bias in median house price
measurements, mix-adjusted models can also neutralise the seasonal
impacts of house price data.
Australia’s established published house price models use different
methodological approaches or variances of specific methodologies.
Models are also differentiated according to particular dwelling types
or market segments.
Despite these different approaches, and some marginal differences
in period-on-period reported results, most published models produce
similar outcomes in tracking the underlying trend of the marketplace
over the medium and longer-term – whether prices are falling, steady
or growing overall in a particular market.
The inherent constraints to effective house price modelling however
ensure that robust and reliable results for major markets are typically
achieved with data models measured with a minimum of three
months sales activity – as used by the ABS. The higher the volume of
sales data used, the less the volatility in modelled price movements,
regardless of the methodology.
And at the end of the day, stable and reliable measures of house
price movements that accurately reflect the underlying supply and
demand dynamics of housing markets are the key insights required
by both homeowners and investors. API
WHICH HOUSE PRICE
DATA MODEL IS BEST?
“ A greater proportion of
higher-priced property sales
can skew prices upwards.”
n DR ANDREW WILSON is the Domain group senior economist. Follow him on
Twitter @DocAndrewWilson or listen to him on The Property Hour on Radio 2UE
on Saturdays from 2pm to 3pm.
INVESTORS’ MARKET VIEW n Dr. Andrew Wilson
Links Archive November 2014 September 2014 Navigation Previous Page Next Page