Overview of the RP Data-Rismark Indices
 The
RP Data-Rismark Indices can be produced for various geographic demarcations
from the suburb or postcode to entire regions, states or nationally.
They can also be produced across all properties or divided between
property types, such as units and houses.
There are three “classes” of RP Data-Rismark Indices:
The first class of indices RP Data-Rismark have produced
are based on median and “stratified” median price series.
Stratification is a process for creating subsets of houses which
are qualitatively similar. Unique price series are created for these
subsets which are then aggregated to estimate quality-adjusted price
movements in the overall market. The strata definitions used to
classify properties into subsets are based on price, geography,
landsize, and interactions of these variables. The stratified median
index that RP Data-Rismark produces for units and apartments groups
suburbs by their long term median transactions price. The stratified
median index that RP Data- Rismark produces for houses, groups suburbs
by their long term price to land size ratio.
This set of median and stratified indices are the benchmark class
of index that will be made available to commercial clients since
they are conceptually straightforward while retaining a number of
advantages of the more complex models. The more advanced indices
available from RP Data-Rismark are discussed below.
The second type of index estimates the performance of
the market by analysing the returns on individual properties that
sell at least twice. This is called the “repeat-sales”
index. RP Data-Rismark produce five different repeat-sales indices.
The first is called the “linear” weighted repeat-sales
model (see Case and Shiller (1987)). This was the first repeat-sales
model to identify that the dispersion of price appreciation of properties
is likely to be related to the time between sales and makes explicit
adjustments that mitigate the biasing effect of this. The second
two repeat-sales models, developed by Calhoun (1996) and Webb (1988),
extend the Case and Shiller model to allow for “non-linearities”
in the relationship between time and price appreciation dispersion.
The fourth repeat-sales model of Goetzmann and Spiegel (1995) is
motivated by the fact that in many cases the features of properties
are not constant through time. Be it a fresh layer of paint, or
the installation of air-conditioning, most houses undergo some level
of revamping, often just prior to sale. The Goetzmann and Spiegel
model controls for elements of price appreciation that are not related
to the time between sales, and thus is ideal for the customer looking
for a “pure” estimate of price growth in residential
real estate. The fifth repeat-sales model is based on the seminal
paper of Bailey, Mourse and North (1967).
The third class of index, known as the “hedonic”
model, has not previously been commercially produced in Australia.
This index utilises comprehensive information on the attributes
and characteristics of residential properties, such as location,
land size, and bedrooms, to measure “quality-adjusted”
changes in property value over time. Two alternative methods for
constructing hedonic models have been designed by RP Data-Rismark.
The first pooled hedonic index combines data from all time periods
in the one estimation procedure to obtain index estimates. A potential
caveat to this technique is that implicitly the values of housing
attributes are held constant over time. That is, the value that
an additional bathroom adds, for argument’s sake, to the total
value of a house is the same today as when the house was bought.
Few, however, would dispute the empirical result that the value
of a bathroom as a proportion of the value of a house is greater
today than yesteryear. The adjacent-period approach that RP Data-Rismark
have developed combines data from consecutive time periods to derive
an index which allows for the implicit value of property attributes
to vary over time. This model also avoids the issue of revisability.
That is, as time moves forward and more data becomes available,
historical estimates of the index are likely to change when data
from all time periods is pooled. As a result of the theoretical
and practical advantages of the adjacent period hedonic model over
the pooled hedonic index, it is the preferred hedonic index.
|
 |
