Housing Supply

With so much discussion about 'fixing' housing affordability in NSW, here's a deeper dive into some of the key factors that influence affordability, drawing from the last 20 odd years of studies, to broaden the debate on solutions.

Land Supply

Many are advocating for land release in NSW, or citing the "fixed amount of land" as the cause for our current affordability woes. This is often based on a simple statement of Sydney's population growth, that needs accommodating.

Population-wise, we are growing, at slightly under 1.3% per annum , a declining trend that will dwindle to just over 1% per annum by 2035.

Yet growth is growth, and more land is needed, we are to understand, to accommodate this growth. Now the urban area of Greater Sydney is 12,367.7?sqkm, a huge arable bowl between encircling mountains, national parks and rivers that spans two climate zones. Our urban footprint within that is about 2,037 sqkm. This urban extent has been increasing over time , for example at a rate of 1.6% p.a. from 1991 (when it was 1,130 sqkm) to 2014 (1,625 sqkm). Over the same period Sydney grew just 0.8% p.a. We have grown by ~1.5m people, and ~900sqkm in 30 years. This is the equivalent of having grown at 16.5 people per ha (6.1 dw/ha) or 22 people/ha net residential density (using Landcom's standard conversion ), or ~8 dwellings per ha (using the Census' standard household size of 2.7). If we had set ourselves the target of achieving low rise suburban growth (22.5dw/ha), we should have achieved at least 45 people per ha - ie grown our urban extent by only 333 sqkm in that time.

Yet the release of land continues apace - the Western Parkland City adds another 5,000ha (50sqkm) of developable land, within which Bradfield City Centre adds 114 ha of high density mixed use.

Greater Sydney, of course, thus remains a low density city despite the rising number of apartments, with an average density of 433 people per sqkm (2021), or 2,552 people per sqkm of urban land (an average of about 9 dw/ha). For comparison, the density of Haymarket (19,500 people per sqkm) is some 50x the metropolitan density and 8x the urban density, itself slightly below the densest parts of Melbourne (Southbank, 21,500 and CBD North, 33,500 - ABS ). Urban density is typically ~20% lower than gross density, of course, as it takes into account regional parks, ports, industry and other low density uses, however for comparison, Los Angeles has an urban density of 2,886 people per sqkm (+10% denser), Minneapolis 3,074 (+20%), and Singapore 8,330 (+225%).

In pure numerical terms, without further growing Sydney's urban area, there should be ample opportunity for growth within a 'gentle density' increase within the urban footprint of about 1 - 2% per annum.

Now 'what is gentle density' I hear you ask - and a whole post is needed to cover this, but let's say for now it is medium density (35-45 dw/ha, or 3 - 6 storeys), in areas that are capable of supporting it in terms of their infrastructure (including transport), or even low-medium density of the interwar North Shore suburbs (27 - 35dw/ha, 2 storeys + local centre apartment clusters). Yet our current approach to transit-oriented development focuses on the first 100m from the station (about 10% of the total catchment), and we funnel our current targets for those stations, those 22 storey towers that we are told are density, that we are NIMBYs for opposing. Yet, if spread across 100% of the catchment, the same number requires no more than 4 storeys to achieve. Amenity-focused development could go further, as well as unlocking home owners and small builders in growing supply. But again, all this is for another post.

It is also worth noting the high cost of an ever expanding urban footprint. Over a decade ago, conservative analysis for the NSW Government demonstrated that the highest return option was to contain growth within the existing urban footprint, due to the increasing transport and infrastructure cost of an expanding area, and the relatively lower benefit vis-a-vis urban infill - see The benefits and costs of alternative growth plans for Sydney (CIE, 2010)

Housing Targets

Whether or not population growth is housed on new land or old, the job of strategic planning is to ensure the growth can be accommodated, by calculating where it should occur and by how much to reach the overall target. The city may grow more quickly than predicted (as the 2021 Census suggests), requiring targets to be adjusted over time, but nevertheless targets represent the 'run rate' for housing growth that is expected. This is a key measure against which both housing approvals and housing completions should be measured. Achieving the target should be an objective measure of whether we are keeping pace of population growth without overcrowding. NSW Housing Targets were last set in 2018 with the Greater Sydney Region Plans (and in Regional Plans in similar years), that is, before the COVID-19 pandemic reduced international migration for 2 years, and additional housing approvals were enabled as part of Planning's recovery package. So, the question should be, have we been approving and building to keep pace with those targets?

Housing Approvals

Most of the focus of housing targets has been on housing approvals. Urban Taskforce has written many articles whenever housing approvals dip, but more interesting is their own analysis of cumulative annual housing approvals by LGA which shows, for example, that the total number of dwellings approved in 2020 and 2021 were both consistent with targets, despite local fluctuation. This is, indeed, a good indicator of approvals supporting growth, and not whether they are sluggish or fast (given what goes in still comes out).

Note also that approvals last for 5 years in NSW and are cumulative, meaning irrespective of current approvals, those approved since 2018 remain valid, including the glut of approvals during the COVID-19 pandemic to incentivise economic recovery. That can be seen in the below graph as a spike in approvals for which there is no equivalent increase in housing completions. This is not unusual as the pipeline is typically greater than completions, and there is typically a lag (in other words, the development market sits on approvals and does not realise all of them - the gap is shown in the two trend lines, about 7,500, or 15%):

Now neither the NSW Government nor industry report on the cumulative total of non-expired housing approvals, but it would appear from the above that there is currently a surplus of unrealised housing approvals, as compared to the targets set by the GSC in 2018 and cascaded through the District Plans and local plans.

In other words, it is likely that as much or more housing has been approved in NSW as a whole than the housing targets require at this point in time.

Housing Completions

Nevertheless there is a growing shortfall in housing supply in NSW. If the approvals are keeping pace with targets, then completions may be a factor. Note the long term trend of 15% less completions than approvals. Many things may drive this, including economic circumstances of individual developers or the market as a whole, lack of finance and similar causes of 'natural' attrition. Labour and material shortages obviously played a part in the last 12 months. Equally, however, housing may never be supplied at the same pace of approvals, even in good years, because the pace of housing release affects sale prices. Release too much in one place in one year and you risk eating into your own profits.

One way of understanding this would be to look at the total application rate compared to the approval rate. For example, if the market intends to supply 50,000 homes and expects a 15% natural attrition, then developers ought to be lodging applications for 57,500 new homes (or maybe 60,000 to allow for the occasional refusal). If, as a result of rejections this was being reduced to 50,000 homes, then sure, planning approvals are an issue. If, however, 50,000 homes are being applied for, and granted (irrespective of the approval time), but only 42,500 delivered in a given year then the issue is not the approval process but the realisation of these approvals. Something at the construction end is awry.

This may be because, as a finely-tuned business model, new housing is not intended to translate into falling house prices - this doesn't make economic sense for development to compete on price in a market where there is undersupply. It is a 'sellers' market, which means a developers market, for new builds. Smart developers will (and do) attenuate supply to the market absorption rate (that is, at current prices). And this absorption rate issue can be seen in that, despite NSW holding the top 4 supply suburbs in 2021-2, that the number of units per suburb is less than 500 dwellings (and often less than 200) per suburb per year, except for Box Hill, where development has occurred at double the projected rate ... to get ahead of adoption of the contributions plan .

At best, then, rising income rather than falling house prices is required to address affordability unless a non-profit driven supplier, like a community housing provider, or the government, is willing to supply housing and re-base the market. However, the number of social housing declined by 10-15% in the last decade (both in Australia overall, and in NSW - see Table 18A.3 ), and likewise the number of Community Housing Providers also declined (Table 18A.8 in the same spreadsheet). Meanwhile disposable incomes are rising more slowly than house prices, in part due to the perverse effect of rising interest rates as set out below.

Housing Affordability

For theories of the cause of the affordability crisis we are currently experiencing, Saskia Sassen has been compellingly arguing since the Financial Crisis began in 2007 that the root cause is housing-as-investment-class - that is (and I paraphrase here), countries around the world that have incentivised investment in housing have seen, consistent with the objective of return on investment, escalating housing costs as he market decouples from its original purpose (of housing people), and inflates to absorb wealth. Those with spare money are encouraged to channel this surplus into housing, further inflating its cost.

International investment (particularly grey money) further inflates the value - as higher value assets are better safe-houses for capital for those concerned about policies that might threaten domestic accumulation (kleptocracies such as Russia, redistributive economies like China etc). Put differently, house prices are thus set by the availability of capital, not the average household income, and so has become decoupled from the means of ordinary people.

AHURI has likewise been tracking the long-term fall in housing affordability, citing evidence since 1960 and acceleration since the late 1990s. In terms of the impact on interest rate rises, in 2007 they also noted (in this study ) how interest rate rises would also exacerbate housing stress:

"When the Reserve Bank raises interest rates to ease pressures in an over-heated housing market, there is a risk that households will cut back on consumption and that the economy will slow down. As incomes fall, housing stress further intensifies unless housing costs also fall proportionately. However, recent home purchasers and investors are locked into the preceding higher price regime and rents may not fall quickly or far enough to offset falling incomes and employment, especially in the low-cost private rental sector. Increasing housing stress in the declining phase of the cycle reinforces the negative effects on consumption, particularly when low- and moderate-income households are carrying large debts as a legacy of the preceding boom.?"

The 2007 report recommends several measures, including increasing social and affordable housing supply through methods such as developer targets, and helping people access that housing (eg: through concessional loans).

However, government and community housing has, in fact, declined. While 'affordable' market housing can theoretically step into the breach, there is no obligation in NSW to provide this unless you are replacing existing supply - ie numbers are at best stable and, because often these can be 'paid out' to council, in reality in decline. Without an obligation for market-led affordable housing in NSW, beyond a vague recommendation of 5-10% supply being possible in Greater Sydney by the GSC (GSRP, 2018), which has not been implemented, there is therefore no low-end supply to rebalance the market-driven prices of the remainder which, as described above, are set by investors.

By 2015 , the issue of accelerating housing stress since the 1990s was more specifically pinpointed by AHURI to the growth in the investment property market (consistent with Sassen), and conversely the lack of investment in social and affordable housing. This graph is self explanatory in this respect:

Even analysis by the Reserve Bank back in 2003 (under Howard) pointed to these same factors - sluggish completions, poor conditions for borrowing and housing-as-investments as key issues for declining affordability. Conditions that may have origins in the 1985 introduction of capital gains tax, that had intensified after 1999 tax reform (including introduction of capital gains tax discounts, accelerated depreciations etc) and subsequent to the RBA report due to the Financial Crisis of 2007 onward.

The solutions, according to the Grattan Institute are the same as they have been since time immemorial (ie 2003 (RBA), 2007 (AHURI), 2012 (AHURI), 2015(AHURI) and 2021 (Sassen) - improve borrowing conditions for low income earners, and build more social housing.