Skip to content

Vector Autoregressive Model with 20 variables - Australian Housing Data.

License

Notifications You must be signed in to change notification settings

virajvaidya/VARAustraliaHousingForecast

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

29 Commits
 
 
 
 
 
 

Repository files navigation

🏡 Australia Housing Data Forecast - Vector Autoregressive(1) Model 📈

This project was completed as a group project with Mr. John Dundas for the ECMT6003 Applied Business Forecasting Unit as part of my Master of Economics program at The University of Sydney.

The model is a Structural Vector Autoregressive Model modelled with the following twenty variables sourced from the Australian Bureau of Statistics (ABS) and the Reserve Bank of Australia's Statistical Tables.

Variable name Variable description ABS/RBA Description Unit
batotalvol Building Approvals Total value of building jobs; Chain Volume Measures; Total Sectors; Total Residential; New AUD Thousands
ctotalvol Building Commencements Value of work commenced; Chain Volume Measures; Total Sectors; Total Residential; New AUD Thousands
dwell_complete Dwelling completions Dwelling units completed; Total Sectors; Total (Type of building); Total (Type of work) Number
ipd_natdi Investment deflator Private; Gross fixed capital formation - Dwellings - Total Index
natdi Dwelling investment Private; Gross fixed capital formation - Dwellings - Total AUD Thousands
wdnewvol Work done, new and used Value of work done during quarter; Chain Volume Measures; Total Residential; Private Sector; New AUD Thousands
cash Reserve Bank of Australia Interbank Overnight Cash Rate Quarter average Percent
inf_exp Inflation expectations Market economists' inflation expectations one year ahead Percent
rinc Real disposable income Total income divided by the Consumer Price Index (CPI) Percent
rinc_per_capita Real disposablew income per person (age 15 years and above rinc divided by wap(working age population) AUD
i_10y_bond 10 year Australian Government bond yield Commonwealth Government 10 year bond Percent
house_price_index Real CPI House Price Residential Property Price Index; Weighted average of eight capital cities Index
rrent Real CPI rents Index Numbers; Rents; Australia Index
stock Housing stock bHousing stock; Chain Volume Measures AUD Thousands
ur Unemployment Rate Unemployment Rate; Persons Percent
mr Variable Mortgage Rate Lending rates; Housing Loans; Variable; Standard; Owner-Occupier Percent
wap Working Age Population Civilian population aged 15 years and over; Persons Thousands
under Underemployment rate Underemployment rate expressed as proportion of Labour Force population; Persons Percent
gdp_aggregate_milaud Aggregate Gross Domestic Product Aggregate GDP; Chain Volume Measures AUD Millions
gdp_per_capita Per capita Gross Domestic Product gdp_aggregate_milaud divided by total population AUD

The data is quarterly - starting from March 1987 to December 2020, spanning the observations of the twenty variables over 136 quarters. The model computes a Vector Autoregressive forecast for the following 4 quarters, i.e. for the year 2021.

Some limitations of this model to be kept in mind however; the model assumes all the variables specified above are exogenous, which is highly unlikely. There is a high chance that the model suffers from overfitting.

The ForecastFile.ipynb file might have some issues loading on some devices due to its size.

In the event that this happens, please click this link for an off-GitHub version of the Python code at https://colab.research.google.com/drive/15blME9jemV8DiQYXmJWgNDg-LwSWR7tA?usp=sharing