Setting Alan Kohler straight on housing policy
by Brendan Coates, Joey Moloney
Journalist Alan Kohler recently published a Quarterly Essay on Australia’s housing crisis. Our housing experts Brendan Coates and Joey Moloney took issue with some of what he said. They wrote a critique, which has now been published by Quarterly Essay, along with Alan Kohler’s response to our response.
Here’s what Brendan Coates and Joey Moloney wrote:
Alan Kohler’s The Great Divide is a compelling account of Australia’s housing calamity and how it threatens to tear our society apart. Within living memory, Australia was a place where housing costs were manageable and people of all ages and incomes had a reasonable chance to own a home with good access to jobs. But the great Australian dream of home ownership is rapidly turning into a nightmare for many young Australians, while the growing divide between the housing “haves” and “have nots” risks returning Australia to the Jane Austen world of the late-eighteenth century.
Kohler correctly diagnoses the core driver of unaffordable housing: it’s too hard to build more homes in established suburbs where people want to live. But having done so, he veers regrettably off-course to propose solutions that have little chance of working, and which simply act to distract his readers from the main game of building more homes. Kohler misses the moment. The political mood is changing. There is a growing groundswell of support for more density, and a growing awareness of the costs of locking up vast tracts of our cities from development. With momentum building and much more still to do, Kohler’s misfire is particularly unfortunate.
Historically, Australia has not built enough housing to meet the needs of its growing population. Heading into the COVID-19 pandemic, Australia had just over 400 dwellings per 1000 people, which was among the least housing stock per person in the developed world. Australia had also experienced the second-greatest decline in housing stock relative to the adult population over the twenty years leading into COVID, and Australian cities are some of the least dense in the developed world.
The reason is simple. The frameworks and processes that dictate what gets built where are hugely biased against change. Older and wealthier residents of well-located suburbs – those who prefer their neighbourhoods to stay the same – get an outsized say. Prospective residents, who might live in new housing in desirable suburbs were it to be built, find themselves effectively unrepresented.
The result is “missing middles”: hectares of prime inner-city land, close to jobs and transport, rising barely taller than two stories. The flow-on effect is high prices and rents, a stagnating economy because fewer people can live close to jobs, and expensive and environmentally damaging sprawl into farmland and floodplains.
If the problem is not enough homes in established suburbs, surely any meaningful solution must involve building more homes in said suburbs? But Kohler is unduly pessimistic, arguing that more medium-density housing is “going to be difficult, if not impossible,” “won’t work,” and “will never actually happen.”
Kohler contends that addressing the supply problem directly is too hard, and instead searches for alternatives that have little prospect of succeeding. He does this because he judges that the obvious answer – building more housing in the inner- and middle-ring suburbs of Australia’s major cities where most Australians still want to live – is politically unworkable.
Kohler frames NIMBYism and heritage restrictions as “natural barriers” to greater density. But there’s no natural law that says we must let the aesthetic preferences of existing residents for Victorian terraces or Californian bungalows trump the needs of their fellow Australians to have somewhere to live. The restrictive zones in desirable suburbs are not unalterable commandments handed down like ancient laws. Building denser cities is a political decision, and Kohler misses that the political tide is starting to turn.
Until recently, supply-side reform was an obsession for a passionate few, but largely absent from broader political discourse. But in recent times, the political clout of renters has grown and the YIMBY movement has gained momentum. Sacred cows are slowly being slaughtered. The Minns government in NSW has plans to up-zone large amounts of well-located land, including overriding heritage controls where they conflict with more density. Victoria is aiming to build 800,000 homes over the next decade, with at least 70 per cent in established suburbs. The Albanese government has put $3.5 billion of federal money on the table, mirroring a Grattan Institute recommendation, to push the states to help build 1.2 million homes over the next five years. This isn’t just an Australian phenomenon. Similar shifts are taking place in the United States, Canada, the United Kingdom, and New Zealand. Kohler has misread the political winds.
Key to this change is the fact that most residents of Sydney and Melbourne actually want more density if it means being able to live in a better-located suburb. Denser dwellings – townhouses, apartments, etc. – made up 44 per cent of Sydney’s housing in 2016, and 33 per cent of Melbourne’s. Yet a Grattan Institute survey showed that residents say they actually want those numbers to be 59 per cent in Sydney and 52 per cent in Melbourne.
The weight of evidence is becoming impossible to ignore. Take New Zealand: in 2016, Auckland – a city of 1.5 million – rezoned about three-quarters of its suburban area to allow more intensive land use. Researchers found that this led to a doubling of the city’s rate of housing construction. Unsurprisingly, rents in Auckland are lower now – relative to inflation – than they were in 2016, whereas rents across the rest of New Zealand have gone up by 10 to 15 per cent over the same period.
Denser cities don’t just offer cheaper housing. Done well, they also bring amenity, vibrancy, and walkability; certainly, much more so than a satellite suburb fifty kilometres from the CBD. Several cities with similar populations but higher densities – such as Vancouver, Toronto, and Vienna – outrank Sydney on quality-of-life measures.
But Kohler argues land-use planning reform is too hard, preferring an alternative approach: run faster trains to peri-urban and regional areas, massively increasing the commutable distance to our major cities. This would be unfair, costly, and ineffective. Allowing more homes in desirable suburbs would enable more young Australians to live, work and add to the social fabric of these communities. Spending billions on trains from somewhere else tells them they’re only wanted there for their labour, and the preferences of those who got there first matter more.
Denser cities are more efficient cities. The NSW Productivity Commission found it costs up to $75,000 less in infrastructure per home in established suburbs than on the urban fringe. Denser cities are also better for the climate – a sprawling, car-dependent city pumps more CO2 into the atmosphere. And denser cities are better for the economy – allowing more employers to locate closer together increases knowledge spillovers and gives workers more options.
More fundamentally, fast trains simply would not solve the problem in the way Kohler contends they will. To be fast, trains need few stops, and few stops along low-density corridors means longer trips to the train station for commuters. Cutting fifteen minutes off a train ride from Geelong to Melbourne isn’t much help if it’s a forty-minute drive to the station, and a race against the clock to find a park before the train leaves. And at the other end, Kohler appears to believe most if not all workers need to get to the CBD. But Grattan research has found that only about 15 per cent of jobs are there, at least in Melbourne. So even after a trek to the station at one end, fast trains to the city still leave workers with more commuting to do at the other end.
The fast-trains solution would leave workers heavily exposed to one service that takes them a hundred kilometres from home. The denser-cities solution offers workers diversity and options. Some people will walk to work, some can ride their bike, others take the tram or train, and inevitably many will drive. But the key point is that when jobs are closer, it is easier for families to organise their lives – easier for one parent to pick up a sick toddler from daycare, or for the other to take a new job opportunity without upending family arrangements.
Australia’s housing affordability crisis is needlessly compounded by muddled housing policy discourse. The heart of the problem is much simpler than many let on: housing costs are too high because there are not enough houses. Kohler provides an incisive critique of the political obstacles towards remedy. The political tide is starting to turn, but there is much more still to do. Distracting Australians with the superficial solution of building trains instead of houses is an unfortunate wrong turn.
And here’s what Alan Kohler wrote in response:
Tackling Australia’s housing mess in a Quarterly Essay was a daunting task, partly because it’s a big, complicated mess, but also because everyone has a firm opinion on the subject and a lot of smart people have spent their lives studying it, and I’m not one of them. But the responses to the essay have been plentiful, informative, and gratifying, including the ones from those who weren’t impressed with what I had written; as always, you learn more from those who disagree with you than those who agree. I’m grateful for all of them.
Brendan Coates and Joey Moloney at the Grattan Institute didn’t like my suggestion of fast trains to open up regional areas for viable commuting to the city, which they called an “unfortunate misfire” and Peter Tulip thought I overemphasised the impact of tax concessions. I want to deal in detail with each of these two responses because I learnt from them, and they get to the heart of the problem – and the solutions.
Brendan and Joey think I too easily dismissed the potential for densifying the suburbs – that is, building more medium-density housing in good locations that use existing infrastructure. They tell us that, going into the pandemic, Australia had 400 homes per 1000 people, among the least amount of housing stock per person in the developed world, and we have some of the least dense cities. The reason is simple, they say: the processes that dictate what gets built where are hugely biased against change.
The answer, they assert, is equally simple: “If the problem is not enough homes in established suburbs, surely any meaningful solution must involve building more homes in said suburbs?”
“Kohler misses the moment,” they write. “The political mood is changing. There is a growing groundswell of support for more density, and a growing awareness of the costs of locking up vast tracts of our cities from development.”
They are dead right that I’ve missed that. If there is a groundswell of support for more density, it has passed me by, which is clearly a failure on my part. Brendan and Joey say that I am unduly pessimistic, and they are also dead right that I’m pessimistic – about the capacity of Australian politics to deliver difficult solutions about anything, especially denser housing. Unduly so, as they assert? Time will tell. I really hope the men from Grattan are right and I’m wrong, because it’s quite true that “denser cities are more efficient cities,” and that by far the simplest solution to the shortage of housing and high prices would be more medium-density housing close to the city.
To drive home my misfire, Mark Walker persuasively explains the difficulties of fast trains: “It is the convoluted, contour-following nature of the original nineteenth-century track alignment that still largely dictates the speed of trains today. To speed them up, we need to spend big on upgrading the actual line of rail – the embankments, viaducts and cuttings on which the rails are laid . . . The problem is centrifugal force. The faster a train travels, the gentler must be the bends in the track, or the engines and carriages can tip up, and tip over.”
So that seals it: fast trains are too expensive and they won’t be needed because the NIMBYs are in retreat. What I wrote in the essay is wrong, it seems, and I couldn’t be happier about that.
Peter Tulip’s complaint is that I put too much weight on the impact of negative gearing and the capital gains tax discount introduced in 1999 – any weight at all, in fact. “There is no credible research supporting this claim,” he writes.
When I began this project, I decided to investigate and explain the housing problem in three steps: first, what happened to house prices; second, what the effect onAustralia and its citizens has been; and third, when it happened. I thought the “when” would help explain the “why,” and all these things together would provide the solutions. The “when,” I thought, was evident from the two charts towards the front of the essay of house prices against both incomes and GDP. It obviously happened in 2000.
I admire Peter’s work and his expertise, and he says I should have used a logarithmic (log) scale, which would have told a different story: that “an acceleration can be seen [after 2000], but it is not dramatic and it begins before the tax change.”
Log scales are used to show exponential curves because they don’t fit on a graph. I’m not sure why a log scale is needed for house prices. Peter includes a log-scale chart of house prices in his response, which he says shows that house prices started rising before 2000. Well, looking at his chart, it’s clear that prices were broadly flat from 1972 to 1987, jumped sharply between 1987 and 1990, which was the rise “before the tax change” that Peter talks about, were flat again for ten years, and then from 2000 rose rapidly and inexorably for more than twenty years to the present day.
I’m sorry, but I reckon that rise in Peter’s log-scale graph is dramatic, and I just don’t accept that the jump in prices in the late 1980s – which was the property bubble and bust that produced the 1991 recession – rules out the tax reforms of 1999 as an important cause of rising house prices. If anything, Peter’s chart reinforces the point, even without including household incomes or GDP.
Graphs aside, house prices increased at 3 per cent per annum before 2000, the same rate as income, and 6 per cent after 2000, double the rate of income. So I stand by the proposition that the psychological effect of halving the capital gains tax with pre-existing negative gearing deductions had a big impact on demand for houses, and that removing those tax concessions must be an important part of dealing with housing affordability.
But I appreciated the generous efforts of Peter, Brendan, and Joey to set me straight, and of course the kind words of many others, including those responses that couldn’t be printed for space reasons. I particularly valued Judith Brett’s historical insights, Nicole Haddow’s millennial viewpoint and Nicholas Reece’s local council perspective.
The process of researching this subject and then engaging with responses to my essay has confirmed that this is a subject about which a lot of people have been thinking deeply and expertly for a long time, and Australia is well served by them. It’s just a pity they are not listened to more. We are less well served by the politicians and bureaucrats whose job it is to do something about it.