Use of complementary administrative data sets for Australian house price indexes.

Use of complementary administrative data sets for Australian house price indexes

Household borrowing

Household wealth

Housing

Other

Household Wealth and Wealth Distribution, Australia 2005-2006, ABS cat. no. 6554.0

House purchase and home ownership

65%

85%Policy makers, housing industry, households interested in a timely measure of house price change

House Price Indexes, Eight Capital Cities (ABS cat. no. 6416.0)

•Since 1986•Quarterly•Free-standing houses•Eight capitals•National index – weighted average•Using median prices

Australian house price indexes (HPIs)

x

xxx •Ideally based on

current and historical prices of all dwellings•Only a fraction is sold

Assumption: Median price of dwellings sold is representative of housing stock

HPI construction

Time

HPI construction

Re-price representative house over timeNo two houses are identicalQuality may change

Need to control for compositional effects

House characteristics

Price

Hedonics•Regression •Data not available

Repeat sales•Requires long time series•Quality may change•Peculiar characteristics ?

Stratification•Group dwellings

Controlling for compositional change

Stratification

•Compare houses with similar attributes – next best alternative•Suburbs base unit•Cluster suburbs according to price determining characteristics•Maximise homogeneity but require sufficient sales

Cluster- Suburb base unit

Structuralcharacteristics

Neighbourhoodcharacteristics

Locational characteristics

Current stratification

Socioeconomic index for areas (SEIFA)•High incomes•Tertiary education•Skilled occupations

•Number of bedrooms•Owner occupied•Houses

Distance to •CBD•shops•hospital

Neighbourhoodcharacteristics

Long-termmedian price

New stratification

•Simpler approach•Better at controlling forcompositional change

Cluster- Suburb base unit

•Current stratification – suburbs in same cluster in same statistical sub-division

Cluster location

•Current stratification – suburbs in same cluster in same statistical sub-division•New stratification – this constraint relaxed

Cluster location

Entire housing stock (not just houses sold)Median cluster price used to update cluster stock values

Weights

x

x

x

Dates

Verbal agreement Contract signing

Contract

Exchange of contracts Settlement

Dates

Contract

Exchange of contracts

most closely approximates the time at which the market price is determined

Valuer General Mortgage Lender

Data sources

•Exhaustive coverage•Suburb, date, price

•Slow •Timely•Do not capture exchange date

•Loan approval data•Imperfect coverage

Mortgage Lender

Benchmark series Updates

Valuer General

Time

Current quarter

Used for HPI up to 6 months prior to present quarter

Used to update HPI up to present quarter

Data sources

• Extend scope from detached houses

• Possibly extend to regional cities and rural areas

Future developments