Supply sceptics beware: without more housing, it won t be affordable

Supply sceptics beware: without more housing, it won t be affordable Brendan Coates, Fellow, Grattan Institute Australian Conference of Economists, Canberra 11 July 2018

Without more housing, it won t be affordable Housing has become less affordable, especially for low-income earners Housing construction hasn t kept pace with additional demand Population growth jumped, and dwelling construction did not respond for almost a decade Mismatch between demand and supply ultimately clears through larger household sizes Even current record housing construction is barely keeping up with population growth Planning policy is a problem Rising house prices mainly reflect land values, not dwellings The zoning premium in Australia is large We re not building the housing people want, where they want it Building more homes would make housing more affordable Building an extra 50,000 homes a year for a decade could see house prices 10-15% lower Other policies alone are unlikely to be enough, or produce collateral impacts (Premature) interest rates hikes have economic costs Macro-prudential rules have short-lived impact Tax reforms would make housing modestly more affordable Slower migration would improve housing affordability, but has collateral impacts 2

Without more housing, it won t be affordable Housing has become less affordable, particularly for low-income earners Housing construction hasn t kept pace with additional demand Population growth jumped, and dwelling construction did not respond for almost a decade Mismatch between demand and supply ultimately clears through larger household sizes Even current record housing construction is barely keeping up with population growth Planning policy is a problem Rising house prices mainly reflect land values, not dwellings The zoning premium in Australia is large We re not building the housing people want, where they want it Building more homes would make housing more affordable Building an extra 50,000 homes a year for a decade could see house prices 10-15% lower Other policies alone are unlikely to be enough, or produce collateral impacts (Premature) interest rates hikes have economic costs Macro-prudential rules have short-lived impact Tax reforms would make housing modestly more affordable Slower migration would improve housing affordability, but has collateral impacts 3

Home ownership is falling particularly fast among younger low-income earners Home ownership rates by age and income, 1981 and 2016 Age group 90 25-34 35-44 45-54 55-64 1981 80 2016 70 60 50 40 30 20 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Equivalised household income quintile Notes: This graph updates Burke et al 2014 using Census data obtained from the ABS. Limitations in Census calculations of household incomes means that changes in home ownership rates by age and income are indicative and small changes in ownership rates should be ignored. Excludes households with tenancy not stated (for 2016) and incomes not stated. Uses age of household reference person and equivalised household income quintiles. Source: ABS Census; Burke et al 2014 Generational change in home purchase opportunity in Australia ; Grattan Institute 4

There are three hurdles to home ownership 1. Can you save the deposit? Saving the deposit is harder than the past 2. Can you afford to service the mortgage? Mortgage costs are affordable now Paying off a loan is likely to be harder with inflation, wages lower in future 3. Are you comfortable with the level of risk you re taking on? What if interest rates rise? Few new buyers are taking out high LVR loans 5

(Quality-adjusted) rents have grown in line with wages Nominal, index, 1997 = 100 450 400 350 300 250 200 150 100 Wages House prices 50 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 CPI Rents Notes: Nominal house price growth; All Groups CPI; Wage price index (excluding bonuses; private and public) Source: ABS 6401.0 Consumer Price Index, Australia; ABS 6345.0 Wage Price Index; OECD Affordable Housing Database 6

Rents have risen fastest for low-income households, well above inflation Rent prices and CPI, indexed to December 2000 250 200 Rent index, low income households All rents 150 Inflation 100 2000 2003 2006 2009 2011 2014 2017 Source: Productivity Commission (2018), Reforms to Human Services: Social Housing in Australia, Figure 6.1 7

In countries that built more housing, prices grew more slowly Change in real house prices, 1990-2015, 200% NOR 150% AUS 100% UK DNK SWE FRA IRE 50% USA CAN ESP 0% ITA DEU FIN -50% PRT JPN -5% 0% 5% 10% 15% 20% 25% 30% Change in housing stock per inhabitant aged 20+ (per cent) Note: Australia data to 2015. Includes all dwellings 9

Population growth jumped, but construction did not Population increase per year, 000 500 Dwellings completed per year, 000 250 400 300 200 100 2005 to 2016 average 1990 to 2004 average 0 1990 1995 2000 2005 2010 2015 200 SA Other 150 WA 100 Qld 50 Vic NSW 0 1990 1995 2000 2005 2010 2015 Source: ABS 3101.0 - Australian Demographic Statistics 10

For its population, Australia has relatively little housing stock (and it hasn t grown) Dwelling stock per 1,000 people Portugal Bulgaria France Finland Switzerland Austria Germany Cyprus Sweden Denmark Netherlands Hungary Norway Lithuania Ireland UK USA Luxembourg Australia New Zealand Poland Chile 2000 2015 2016 for Australia 250 300 350 400 450 500 550 600 Note: Figures are for total dwellings, not just private dwellings, and include unoccupied dwellings. Figures for Australia updated to 2016 using ABS Census. Source: OECD Affordable housing database; ABS Census; ABS Demographic Statistics 11

Housing construction lagged population growth for much of the 2000s, but picked up recently Dwelling completions per additional thousand people Between 450 and 550 1,500 1,000 500 0 1,500 1,000 500 Victoria NSW Queensland Australia new homes are needed per 1,000 new residents 0 1985 1990 1995 2000 2005 2010 2015 1985 1990 1995 2000 2005 2010 2015 Notes: Does not take into account demolitions. The Victorian series spikes at 3,500 in 1993 (cut off to improve readability). Higher rates of home building per additional resident in the 1990s in part reflect declines in average household size among the existing population. Average household size fell from 2.8 people per household in 1991 to 2.6 by 2001. Average household size has been flat at around 2.6 people per household since 2001. Source: ABS 8752.0 Building Activity, Australia; ABS 3101.0 Australian Demographic Statistics; Grattan analysis 12

Housing construction in Australia has been strong, but so has population growth Dwelling completions in 2015 5 4 3 2 1 0-1 Completions per additional person (2014 to 2015) Completions as percentage of the existing housing stock Notes: Estonia, Japan and Poland not shown due to large negative numbers. Dwelling construction data for 2015 or closest year. Source: OECD Affordable Housing Database; United Nations population database 13

Household size flat-lined from 2000 due to worsening affordability Average household size, actual and projected 4.0 3.5 Actual 3.0 2.5 2.0 ABS 1999 ABS 2015 ABS 2004 McDonald & Temple 2013 1.5 1966 1976 1986 1996 2006 2016 2026 2036 Notes: ABS projections are series II/B projections. ABS 1999 projection for 2021 is for average household size to be between 2.2 and 2.3 Source: ABS; McDonald and Temple (2013) 14

Recent construction does not meet housing targets, let alone actual population growth Average annual net housing construction 50,000 40,000 2017 To meet Plan target No 2017 data for Melbourne and Brisbane 30,000 20,000 2006-2011 2011-2016 10,000 0 Sydney Melbourne Brisbane Notes: Draft Greater Sydney Region Plan: 725,000 additional dwellings over 2016-2036 (excludes the Central Coast). Plan Melbourne 2017: 1,550,000 additional dwellings over 2015-2051 (based on Victoria in Future projections). For 2006 to 2016 data, growth in dwelling stock is calculated using 2016 Greater Capital City Statistical Areas. Data for 2017 dwelling completions in Sydney from NSW Department of Planning and Environment (2018). No 2017 completions data available for Melbourne Sources: Greater Sydney Commission (2016); Victorian Government (2017); NSW DPE; Queensland Government (2017). 15

Victoria and NSW populations are growing much faster than in the past or projections Net population growth, four-quarter rolling sum, thousands 160 140 120 100 80 60 40 20 0 1990 1995 2000 2005 2010 2015 2020 2025 Average annual long-run projections Victoria NSW Qld WA SA Tas, NT, ACT Notes: Annual average population projections from state government departments: NSW (2016) 2011-2031; Vic (2016) 2011-2031; Qld (2015) 2011-2036. Source: ABS 3101.0 - Australian Demographic Statistics; state government population projections 16

Housing prices increased mainly due to higher land values, although buildings are also better Real market value of Australian property, $2016, trillions 7 6 5 4 3 Residential improvements (3% Compound Annual Growth Rate) 2 1 Residential land (6% CAGR) 0 1990 1995 2000 2005 2010 2015 Notes: Residential improvements consists of the value of the stock of dwelling construction; historical figures are deflated by the Consumer Price Index to $2016. Source: ABS 5204.0; ABS 6401.0; Grattan analysis. 18

Planning restrictions have pushed up the price of houses in Australia s major cities Contribution to the price of an average detached house, $000s Sydney Melbourne $1,200 Zoning effect $1,000 Physical land $800 Structure $600 $400 $200 $0 $1,200 Brisbane $1,000 $800 $600 $400 $200 $0 2000 2004 2008 2012 2016 Zoning effect, % of price (RHS) Sydney Melbourne Brisbane 2000 2004 2008 2012 2016 Notes: Zoning effect as % of price is negative for Brisbane from 2000 to 2001, so not shown. Zoning effect is the sale price less structure and land costs. Source: Kendall and Tulip (2018) The Effect of Zoning on Housing Prices, Reserve Bank of Australia. 50 40 30 20 10 0 19

There s a shortage of medium density housing compared to what people want Per cent of housing stock, actual and preferred 100 80 60 40 20 4 storeys and above Up to 3 storeys Semidetached Detached house 0 2006 actual 2016 Preferred actual stock Melbourne 2006 actual 2016 actual Sydney Preferred stock Notes: Preferred stock is from the survey of 700 residents about housing preferences from Grattan s 2011 report, The housing we d choose. Data may not sum to 100 due to rounding. Excludes dwellings listed as Not stated and Other dwellings, such as caravans. Source: Census; Housing we d choose; Grattan analysis 20

Building more homes could substantially improve affordability, in the long-term Housing demand is price inelastic Abelson (2005) a 1% increase in housing stock per capita led to a 3.6% fall in real house prices, based on 1970-2003 in Australia Girouard et al (2006) an analysis of 20 studies across 12 countries found that elasticity of real house prices to housing stock is ~3.1% Building more homes would make housing cheaper Building an extra 50,000 homes a year = 500,000 extra homes in a decade Roughly equivalent to an extra 4% of the existing housing stock Given observed elasticities, house prices could be 10-15% lower after a decade 22

Without more housing, it won t be affordable Housing has become less affordable, particularly for low-income earners Housing construction hasn t kept pace with additional demand Population growth jumped, and dwelling construction did not respond for almost a decade Mismatch between demand and supply ultimately clears through larger household sizes Even current record housing construction is barely keeping up with population growth Planning policy is a problem Rising house prices mainly reflect land values, not dwellings The zoning premium in Australia is large We re not building the housing people want, where they want it Building more homes would make housing more affordable Building an extra 50,000 homes a year for a decade could see house prices 10-20% lower Other policies alone are unlikely to be enough, or produce collateral impacts (Premature) interest rates hikes have economic costs Macro-prudential rules have short-lived impact Tax reforms would make housing modestly more affordable Slower migration would improve housing affordability, but has collateral impacts 23

Other housing policies are likely to prove unattractive, or only modestly effective Interest rate hikes Tighter macroprudential rules Abolish negative gearing + 25% CGT discount Home above $500k in pension assets test Stamp duty land tax swap Curb migration (i.e. 50,000 less a year) Housing Economy Budget Would substantially reduce house prices (if they rise) House prices fall short-term, little impact long-term Prices ê ~2% overall, more at bottom Rents won t rise Prices ê ~1% overall Few pensioners downsize Better use of housing stock (i.e. spare bedrooms) Prices ê ~6% House prices 6-7% lower than otherwise after a decade Hiking rates prematurely has large economic costs Positive if used to mitigate financial risks; otherwise not Reduces over-investment in housing Big economic benefits (~$17b / yr) Small reduction in GDP per capita $5 bn / yr $1-2 bn / yr Budget neutral swap Hits budget balance Notes: Excludes policies explicitly intended to improve affordability for low income earners via housing subsidies, such as more social housing, or increases in Commonwealth Rent Assistance. 24

All the important reforms are difficult; all the easy reforms are cosmetic Social, economic and budgetary impacts Home in pension Positive assets test Congestion charging Improve renting conditions Neutral Negative SMSF borrowing Deposit schemes Tax empty dwellings Social housing bond aggregator Shared equity schemes Political difficulty: Foreign investor limits/ taxes stamp duty for downsizers Downsizers keep pension / Regional exempt from development super rules FHB grants / concessions Minimal Small Medium Easy Medium Difficult Abolish stamp duty Boost density in Improve transport project middle suburbs selection Reform state land taxes CGT discount Boost density along Negative transport corridors gearing Macro-prud. greenfield land supply rules CGT on primary residence Reduce immigration Large Impact on housing affordability Very large Source: Grattan analysis. Notes: Prospective policies are evaluated on whether they would improve access to more affordable housing for the community overall, assuming no other policy changes. Assessment of measures that boost households purchasing power includes impact on overall house prices. Estimates of the economic, budgetary or social impacts should not be treated with spurious precision. 25

For more 26