Don't believe the house-price gloom
Publish a “list” these days that offers simple rankings on any high-visibility subject — best dressed, worst dressed, richest, most admired, smartest, sexiest and most innovative — and you are bound to capture attention. Lists appeal to our hedonistic need for instant gratification. Unfortunately, lists are rarely meaningful and can be downright misleading.
This principle is also true of any list purporting to inform us about the relative affordability of global housing markets. If you are an urban planning consultant and lobbyist for the roadway (read: pro-development) industry, such as Wendell Cox, who also happens to own the widely quoted “Demographia” website, the instant media fame and implied authority garnered from publishing a list of whether house prices across different countries are over- or under-valued can be exceedingly lucrative. And don’t for a minute think that the media would be concerned by the fact that Demographia is actually the Wendell Cox Consultancy, or that your co-author in this endeavour is a New Zealand property developer, Hugh Pavletich of Pavletich Properties.
And the rankings are all the better if they show that the most sought after cities – the Londons, New Yorks, Sydneys, Dublins, Melbournes, Aucklands, San Franciscos and so on – are, surprise, the most expensive conurbations and lend credibility to your core and oft-repeated argument that governments should not impose “prescriptive planning policies (such as ‘rationing’ of land)”.
Even better if they support your argument that the true role of urban planning is “to facilitate lifestyles as revealed in household preferences (the market)” — or, put differently, to accommodate as much development as is allowed by the untrammeled whims of the market.
No one is likely to pick up on the fact that Wendell Cox is a self-described “opponent of smart growth, especially urban growth boundaries, impact fees, and large lot zoning” or that his critics claim that he, as a paid consultant, has “authored studies for the American Highway Users Alliance, a group that lobbies for more highways [and]… has been employed by various conservative and road building groups over the years.”
Never mind that the reason exclusionary zoning laws were introduced in the first place was to protect the community from the “negative externalities” associated with unconstrained development (which have grown as our understanding of the economic and environmental costs of urban agglomeration has improved over time).
And you can rely on our headline-obsessed, due diligence free media to not in any way question — just for a moment — either the usefulness of the rankings you publish or the interests of the individuals responsible for producing them. After all, rankings are rankings — they must be right. And with a fancy name like “Demographia”, you might as well be the World Bank (more on the World Bank later).
No-one will draw attention to the fact that in your ranking of some of the world’s 32 most unaffordable metropolitan markets you have listed, alongside the obvious culprits, Bundaberg, Newcastle, Wollongong, Cairns and Hobart. In fact, according to Demographia, housing in Bundaberg is less affordable than New York and London!
All of the Demographia International Housing Affordability Survey findings are based on one extremely simple metric known as a “house price-to-income” ratio. In short, Cox & Co. assume that you can “value” housing markets using this crude measure.
The first problem here is that there is no statistical evidence that there is any stable or predictable long-term relationship between median incomes and house prices. Notwithstanding this, Wendell Cox claims, “The [house price to income ratio] in Australia is 6.0, double the 3.0 historic maximum norm and well above levels of just a decade ago.” In fact, the median incomes used for the purposes of creating these house price-to-income ratios are likely to be quite different to the incomes associated with the marginal home buyer (where the median income used in their ratio includes much lower income and higher credit risk households).
The RBA has recently published some long-term analysis on the subject across a variety of countries. It is useful noting upfront that the RBA’s estimate of Australia’s house price-to-income ratio in 2007 of 5.5x is lower than the Demographia finding. The second interesting point deriving from the RBA analysis is that contrary to some of Demographia’s claims, Australia’s house price-to-income ratio has actually declined since 2003 when it peaked at 5.9x (see chart below). Given that house prices in 2008 have decreased slightly while nominal disposable household incomes rose by a strong 9.5 per cent in the year to end September 2008 alone, our house price-to-income ratio should have declined quite substantially.
The next chart compares Australia’s long-term house price-to-income ratio with other countries. A few things are clear from this data. Observe that since 1985 the Australian ratio has always been higher than the other countries included in the analysis. This is because Australia has always experienced rates of capital growth that have been higher than most other nations since 1971 (according to IMF analysis) for idiosyncratic reasons related to the demand and supply drivers of housing in this country.
The other thing that you notice about these long-term ratios is that they do not appear to be stable; that is to say, in contrast to Cox’s claims, there is no evidence that a house price-to-income ratio of three can be classified as a “maximum historic norm”.
Perhaps the most important flaw in this analysis, however, is that incomes are but one variable that determine long-term changes in house prices. House prices are determined by the intersection of a range of demand- and supply-side variables including interest rates, incomes, immigration, organic population growth, employment, and the supply of homes (or housing starts).
You cannot, therefore, “value” house prices by taking one simple demand-side factor, such as incomes. It would be akin to trying to value gold prices ignoring the supply of gold and only referencing some proxy for demand—say, the time-series change in the number of per capita weddings. And we know that there is a very wide consensus amongst almost all economists and government agencies (including both the Treasury and the RBA) that Australia’s housing demand far exceeds supply.
Building approvals (ie. supply) in Australia have been in free-fall during the last year with the latest November 2008 data implying that housing starts are running at just 110,000 properties per annum compared with Treasury forecasts for housing demand of closer to 200,000 homes per annum. In NSW, building approvals are at their lowest level since 1958.
In a detailed research report published in September 2008 two Morgan Stanley analysts produced “a regression model to predict 12-month forward house price growth in Australia, based on nine major factors – GDP growth, population growth, [housing] supply growth, average mortgage rates, housing affordability, rental growth, employment growth, real wage inflation and loan approvals.”
Perhaps the most interesting finding in Morgan Stanley’s analysis was that of the nine different economic variables included in their model of Australian house prices, housing supply – and not wages or affordability – had been clearly the most important of all the factors in determining Australian house prices over the last 20 years.
To quote Morgan Stanley: “[Housing] Supply growth is most important…This is good news for the industry, as the number of housing completions is currently coming down and in our bearish economic scenario, this is likely to come down further… We do agree that a slow down in new [housing] supply will mitigate the impact of this drop in demand.”
A second major flaw in the Demographia approach is the attempt to compare house price-to-income ratios across countries. There are, however, significant differences between most countries’ property markets that result in divergences in their relative performance. That is, there is no reason to believe house price-to-income ratios should be constant across nations. These differences include their:
– Relative population growth rates (very high in Australia, but much lower in the US (projected to be less than half Australia’s rate) and UK (currently less than half Australia’s rate) and negative in Japan);
– Mortgage default rates — the 90 day default rate on Australian home loans is about 15 and 30 per cent of the level of equivalent default rates on US and UK loans despite historically higher interest rates;
– Tax treatment of housing (in the US, for example, capital gains tax is levied on owner-occupied housing while mortgage interest repayments are tax deductible; neither applies to Australia).
– Size of their public housing markets (the public housing market share in the UK is far larger than Australia, which in turn has a much bigger private rental market).
– Rates of home ownership (54 per cent in the Netherlands, 42 per cent in Germany, and 35 per cent in Switzerland, but about 70 per cent in Australia).
– Responsiveness of housing supply to changes in demand (low in very supply-constrained countries such as Australia, but higher in many countries such as the US, which has had a problem of excess supply).
– Urbanisation or the share of their population living in their largest cities as measured by the so-called “Zipf curve” (very high in Australia but considerably lower in the US, UK, Russia, Japan, India, Germany and China).
For a final word on both the current state of affordability in Australia and the relative valuation of our housing market, I will defer to the World Bank and the RBA.
In its October 2008 World Economic Outlook, the World Bank concluded: “As noted in the 2008 IMF staff report for Australia, if some country-specific factors, particularly the impact of long-term migration on housing demand, are taken into account, the results do not produce evidence of a significant overvaluation of [Australian] house prices.”
On the issue of affordability, the RBA has recently constructed some more sophisticated measures given the well-documented shortcomings of the public proxies. One in particular effectively ranks the relative affordability of Australian households over time.
More specifically, this measure estimates the inflation-adjusted incomes of households under the age of 39 after they have serviced all of their mortgage repayments. Put differently, the RBA has asked the question, “How has the disposable income of home buyers under 39 years of age changed over the last 25 years after meeting all their mortgage repayments?” Based on what one reads in the media, the popular answer to this question would presumably be that households buying a home today are far worse off than their predecessors.
The RBA analysed households between 1982 and 2007 assuming that they had a 10 per cent deposit (or 90 per cent fully amortising mortgage) and were buying a property in the 30th price percentile (ie, a slightly cheaper home). They also focus on younger households headed by a person aged between 25 and 39 years, which is appropriate given first time buyers are typically around 32.
The RBA’s findings run strongly against the grain of what many would have us believe (refer to the chart below): that is, despite strong growth in real house prices over the last 25 years, the income younger home buyers have left over after paying off their mortgage was actually higher in 2007 than it has been at any other point in recorded history. Of course, affordability has only improved since then with significant reductions in the cash rate combined with no growth in house prices during 2008.
Christopher Joye writes Business Spectator's property blog and is managing director of research group Rismark International which produces the RP Data-Rismark Hedonic House Price Indices. Rismark was also responsible for Australia’s first mass-market “shared equity” finance product.