Appendix A: Regression Data | RDP 2020-05: How Risky is Australian Household Debt?

RDP 2020-05: How Risky is Australian Household Debt? Appendix A: Regression Data

Jonathan Kearns, Mike Major and David Norman

August 2020

Download the Paper 1,881KB

Table A1: Regression Data
Variable	Source	Unit	Measurement
Household debt-to-income	ABS; Jordà et al (2013); Thomson Reuters; authors' calculations	ratio	Disposable income is measured after tax but before deduction of interest payments for most countries, includes non-profit sector; Australian DTI ratio extrapolated three years earlier than ABS data using the Jordà et al (2013) database
Time-varying
Real income per capita	OECD; World Bank; authors' calculations	US$	Real disposable income per capita; in 2017 dollars, converted to US$ using year-average exchange rate
Real interest rate	Bloomberg; IMF; RBA; central banks	%	Calculated as 10-year nominal sovereign bond yield minus past year's CPI inflation
Inflation	IMF	%	CPI inflation over the prior year
Housing pricing growth	BIS; IMF; authors' calculations	% pa	Growth in BIS long-run housing price series, minus CPI inflation over a three-year period
Max price fall	BIS; authors' calculations	%	Absolute peak-to-trough fall in the BIS long-run housing price series over the prior period; weight on price outcomes is given full weight for the most recent five years, and then gradually declines according to the following formula: $Δ P \times {0.8}^{(t - 5)}; \forall 5 < t \leq 15$ , if prices have not fallen during that time, the value is set to 0
Urban population growth	United Nations	% pa	Two-year annualised rate of change in population for cities with at least 300,000 people in 2018
Urbanisation	United Nations authors' calculations	%	Share of national population living in cities that had at least 300,000 people in 2018
Housing stock per person	United Nations; authors' calculations; national agencies	ratio	Total dwellings divided by population
Younger population	United Nations; authors' calculations	%	Share of total population aged 25 years or less
Older population	United Nations; authors' calculations	%	Share of total population aged 65 years or older
Financial liberalisation	Abiad, Detragiache and Tressel (2008); authors' calculations	index	Based on seven metrics of deregulation (credit controls, interest rate controls, entry barriers, state ownership, capital account restrictions, prudential regulation and securities market policy); we update this to 2018 using various sources of information on these metrics
Income growth expectations	Consensus Economics	%	Forecasts for GDP growth in the year t + 1, made in March of year t
Unemployment rate	IMF	%	Unemployed persons as a per cent of the labour force
Baa-Aaa spread	Federal Reserve Bank of St. Louis	ppt	Difference between yield on Baa-rated bonds and equivalent tenor Aaa-rated bond
‘Time-variant’
Ownership of housing	authors' calculations; national agencies	%	Share of all housing owned by households; generally collected from financial accounts data, where that is unavailable then census data are used
Strength of legal rights	World Bank	index	Ranges from 0 to 12, with 12 being the strongest creditor rights
Urban density	Charting Transport	person/ha	Population-weighted density of main cities in each country, calculated using square km grids
Inequality	OECD	ratio	Gini coefficient
Interest tax deductible	Ellis (2005); authors' calculations	dummy	1 = interest payments are tax deductible
Capital gains taxation	Ellis (2005); authors' calculations	dummy	1 = capital gains on housing are taxed at full rate

Table A2: Sample Period for DTI Ratios across Countries
Country	Sample starts	Country	Sample starts
Australia	1985^(c)	Japan^(b)	1980
Austria^(a)	1999	Netherlands	1980
Belgium^(a)	1999	New Zealand	1980
Canada	1980	Norway^(a)	1987
Chile^(b)	2003	Portugal^(a)	1999
Denmark	1999	South Korea^(b)	2008
Finland^(a)	1999	Spain^(a)	1980
France^(a)	1980	Sweden	1980
Germany	1991	Switzerland	1999
Ireland^(a)	2001	United Kingdom	1980
Italy^(a)	1999	United States	1980
Notes: (a) Not included in narrow regression (b) Not included in narrow or wide regressions (c) ABS data begin in 1988, DTI ratio is backcast a few years based on data published by Jordà et al (2013)