Judith Yates has produced a comprehensive and highly-informed commentary on the Henry Tax Review’s discussion of housing taxes and tax expenditures (subsidies). This is no small feat because, as Professor Yates points out, at least 25 of the Henry Review’s 138 recommendations relate directly to housing and many others have an impact on housing. While this commentary agrees with many of Yates’ observations, it also finds some significant points of difference.
The general thrust of Yates’ chapter is that the Henry Review is strong on efficiency issues but soft on the distributional implications of housing subsidies principally to home owners. Yates alleges that the Review bowed to political pressure in not recommending changes to subsidies such as the non-taxation of imputed rent and “capital gain tax concessions”. Yates argues that these subsidies are not only regressive but also that they undermine housing affordability by encouraging excess demand for housing and pushing up housing prices. However she applauds the Review’s proposed introduction of a general land tax, in addition to local rates, as a means to redress at least partially the inequality and inefficiency of the existing housing tax/subsidy system.
More specifically, as I interpret the chapter, Yates’ main arguments are that:
● Housing affordability is a problem for many households and a growing one
● (Housing challenges in the 21st century).
● This problem is caused to a major extent by excess consumption of housing and high house prices created by excessive housing subsidies, especially by the “concessional capital gains tax” but also by the non-taxation of imputed rent and non-taxation of rental services (The impact of housing taxation on housing).
● Land supply is ultimately limited and therefore the main policies must relate to demand constraint (The contribution made by the AFTS Report recommendations in addressing housing challenges, Housing supply reforms).
● The Review failed to acknowledge the impacts of the tax concessions and did little or nothing to ameliorate these impacts (The contribution made by the AFTS Report recommendations in addressing housing challenges, Housing assistance reforms).
● The Review’s proposed general land tax will produce more equitable and efficient outcomes but if this is left to the States, as proposed by the Review, no serious reform will occur (Conclusions).
Yates makes many other substantive points, but in this commentary I focus on these main points.
Yates bases the notion of declining housing affordability on two main propositions. One is that although 70% of households are still home owners in Australia, as they have been since 1960, this proportion is likely to fall. Yates argues that home ownership has held up to date as a result of the ageing of the population. However she notes that between 1981 and 2006 home ownership rates for households with a reference person aged between 25 and 34 years declined by 10 percentage points to 51%. In her view, fewer young people now own housing because they cannot afford to buy a house. No data on housing user costs are provided to support this contention. Another plausible explanation for this decline could be lifestyle changes with deferred long-term partners and deferred children but this and other possible explanations are not explored.
The second proposition is that housing is increasingly unaffordable for renters with nearly 500,000 households paying rent in excess of 30% of their gross household income. The major source for this estimate is the National Housing Supply Council Report (NHSC, 2010).
The basic premise of the NHSC Report and many other commentators (see Yates, 2008) is that housing is unaffordable and households experience housing stress if they are in the lower 40th or 50th percentile and spend more than 30% of their gross income on mortgage repayments and interest or on rents. As the Productivity Commission (2010) points out, the number of households in “unaffordable housing” falls considerably when Commonwealth rental assistance is deducted from both household income and rents paid.
The NHSC Report (2010, p 103) observes that in 2007-08 a total of 1,410,000 private rental dwellings were affordable for the 814,000 private renter households in Australia with incomes below the 40th percentile. However, 1,089,000 of these dwellings were occupied by households in higher income percentiles. Using the 30% housing expenditure criterion, the NHSC Report then concludes that, instead of a surplus of affordable housing, there was actually a shortfall of nearly 500,000 affordable and available dwellings for those in the lower two household income quintiles.
In my view, as argued in Abelson (2009), there are serious weaknesses in the widely adopted definition of housing affordability cited above. The definition is not based on any accepted principle of housing user costs, which for homeowners would exclude all mortgage repayments as savings and include real capital gains as an offset to housing costs. Most importantly it ignores the fact that housing expenditure (and location) and household size are all choice variables. A plausible explanation for much of the allocation of housing to households that the NHSC Report observes is that households are exercising a significant amount of choice.
Undoubtedly many households find it difficult to afford reasonable housing by current day standards. These include many households who spend less than 30% of their gross income on rents. However it would be hoped that methods for estimating housing affordability problems could be significantly improved and that more realistic estimates of households with housing affordability problems could be achieved.
As Yates shows, the three large homeowner subsidies are the non-taxation of capital gains and the non-taxation of imputed rent and of rental services. According to estimates in Abelson and Joyeux (2007), Yates and the Australian Treasury (2010, as quoted by Yates, The extent of housing taxation and transfers) may have slightly under-estimated the size of the latter two subsidies, but grossly over-estimated the capital gains tax (CGT) subsidy. At between $30 billion and $44 billion per year, the Yates and Treasury estimates of the CGT subsidy represent nearly all the estimated net subsidy to housing whereas Abelson and Joyeux (2007) estimated the CGT concession to be worth $7 billion per year.
The major reason for the difference is the choice of benchmark. Yates and Australian Treasury presumably estimate the subsidy relative to the actual CGT tax on 50% of nominal capital gains (at sale realisation) which was introduced in 1999. This CGT is itself widely regarded as concessional but actually it substantially increased the CGT compared with the regime in place for the previous 14 years, namely a tax on 100% of the real gain in prices. Abelson and Chung (2005) estimated that the real increase in house prices in Australia in the long run was about 2% per year but that about half of this was due to increased quality, especially from renovations, so the real long-run increase in house prices for a constant quality product is 1% per year. With a long-run inflation rate of say 3% per year, nominal house prices would therefore increase by about 5% per year. With a tax on real gains after homeowner expenditure, a homeowner would pay CGT effectively on 1% of the house price per year. With a 50% tax on nominal gains a homeowner would pay CGT on 2.5% of the house price per year.
It is hard to see any efficiency or equity in a CGT on nominal gains including gains due to owner renovations (which account for over 40% of all investment in housing and are already taxed via the goods and services tax (GST)) or indeed on home and property asset maintenance (which also bears the GST). 1 There remains a substantial subsidy to home owners but estimates of $30 billion to $40 billion per year appear greatly exaggerated against an efficient and equitable benchmark.
The view that land supply for housing is limited is an important input to the conclusion that subsidies for home owners lift housing prices and crowd out housing for lower income households. If land supply is limited, and the application of capital to land is highly constrained, the opportunity for supply-side solutions is greatly reduced and subsidies principally increase prices rather than housing.
It is true that there has been very little growth in new housing supply in Australia over the last 20 years. New housing over the whole country has averaged around 140,000 houses and apartments per year in five-yearly periods since 1990 (and the total has not increased with population or income growth). This lack of growth has contributed to an increase in real house prices. Abelson et al (2005) estimated that the Australia-wide elasticity of real house prices to housing stock per capita is 3.6% (ie, an increase of 1% in the total housing stock would reduce real house prices by 3.6%).
However the basis for the view that residential land is limited is not clear. There is a large amount of land available for residential development around most Australian cities, including Sydney, and all regional centres. The lack of new housing is more a reflection of planning and infrastructure restrictions than of lack of land (Applied Economics, 2010). As a society we may or may not wish to use this extensive land for housing but this should be resolved by cost-benefit analysis not by assumptions about topographical or other constraints.
Do the tax concessions, especially those to home owners, create overinvestment by home owners and worsen housing affordability for some homeowners and most renters? This is a complex issue. However some key points may be made. Clearly, for any given housing supply the home owner subsidies reduce the supply of rental housing and increase rentals. However, it is not clear that the subsidies create over-investment in housing. The main reason is that in many areas the surplus of housing land values over alternative land use values due to restrictions on new housing is much greater than the present value of the annual subsidy to housing. Abelson and Joyeux (2007) estimated that there is an average Australia-wide surplus of $50,000 per new house constructed in Australia and that the extra housing induced by the homeowner subsidy represents a net social benefit of $187 million per year.
Home owner subsidies also encourage over $20 billion of investment annually in home alterations and additions. Given that GST is paid on these expenditures, the net subsidy may not be large. However more relevant here, if a homeowner spends capital to renovate their house or build a second storey, it is not clear that this expenditure of capital is depriving anyone of an affordable house.
However I agree with the Review and with Yates that insofar as households are using scarce and valuable land they are contributing to a housing affordability problem. This may manifest itself in extra transport costs as much as in higher house prices. But either way there is a case for arguing that consumption of land space does cause some pecuniary externalities and hence equity issues. Hence I agree with the proposal for a general land tax in addition to local council rates (an existing tax on land) with the proviso on equity grounds that the revenue should be hypothecated to the relevant urban area city rather than to the nation.
Moreover, in Sydney and possibly in other cities, it may be remarked that there are significant subsidies to land ownership at the fringe which hold up the release of land for residential development. In Sydney, when land is rezoned for housing use it is revalued. However if the main actual land use is deemed to remain rural, the landowner continues to pay the generally low rural rate on the revised land value rather than the higher urban rate. Secondly, payment of rates on the uplift in value on the rezoned site can be postponed until the land is sold. And further, councils are required to write off any debt older than five years. This means that after five years the landowner faces a zero marginal cost for occupying but not developing rezoned housing land.
As remarked at the outset, Professor Yates has produced a comprehensive overview of both housing economics and the Review’s housing proposals. I agree with many of her observations. For commentary I have selected those points with which I most disagree. In some cases the difference is one of degree, for example about the magnitude of the housing affordability problem or the size of tax concessions especially the CGT. I also agree with the proposal for a general land tax, subject to revenue hypothecation to the areas involved. And arguably the housing tax concessions are regressive. But I do not believe that these concessions are particularly inefficient (given other distortions in the market) or that they significantly reduce housing affordability for low income households. In my view the prime policies for increasing housing affordability are supply-side policies that allow more land for housing and a greater application of capital to land.
Abelson, P and D Chung (2005), “The Real Story of Housing Prices in Australia from 1970 to 2003”, Australian Economic Review 38: 1-17.
Abelson, P, R Joyeux, G Milunovich and D Chung (2005), “Explaining House Prices in Australia: 1970 to 2003”, Economic Record 81: S1-S8.
Abelson, P and R Joyeux (2007), “Price and Efficiency Impacts of Taxes and Subsidies for Australian Housing”, Economic Papers 26: 147-169.
Abelson, P (2009), “Affordable Housing: Concepts and Policies”, Economic Papers 28: 27-38.
Applied Economics (2010), Residential Building Activity in Sydney: An Overview and Seven Case Studies, report prepared for New South Wales Treasury.
Australian Treasury (2010), Tax Expenditures Statement 2009.
National Housing Supply Council (NHSC) (2010), 2nd State of Supply Report.
Productivity Commission (2010), Report on Government Services.
Yates, J (2008), “Australia’s Housing Affordability Crisis”, Australian Economic Review 41: 200-214.