In defence of Demographia

By Leith van Onselen

Today, fellow MacroBusiness blogger, Cameron Murray (Rumplestatskin) argues that the latest Demographia Housing Affordability Survey is wrong in its contention that high prices are the result of artificial land constraints.

Regular readers will know that I disagree with Cameron’s analysis. Here’s why.

First, Cameron argues that Australian housing supply has not been restricted because, if it was, there would have been a universal increase in rental costs relative to income:

But there is a much more crucial problem with the Demographia study. It claims to be seeking evidence of detrimental price effects from planning controls, yet it ignores the rental market completely. If housing development is being constrained, then rents must also rise. Any variation between prices and rents can be explained by asset market behaviour. If BHP shares rise 20% but the iron ore price is constant, we don’t suddenly scream about iron ore shortages. But essentially this is what has happened for a decade now with home prices.

The graphs below show some key metrics for Australia housing markets from the ABS Housing Occupancy and Costs survey. While the most recent ABS data is from 2009, that year coincides with the peak unaffordable year in Sydney and Melbourne according to Demographia.

What are the important points to note in the ABS data?

First, and most critically for this debate, is that rents have been essentially flat compared to household incomes since the early 1990s. Since rents for residential housing are determined by competition for location between renters, it is expected that rents are simply a constant share of income. And they are, at around 20% of gross household income.

This also means that over the long run housing rents should rise by more than the CPI, if incomes are, which is generally the case. Welfare agencies should take note.

Second, the steep rental growth observed during the 2007-2009 period was mostly the result of steep household income growth. Waning income growth since then has led to waning rental growth.

Fair enough. But there is a better way of looking at this. The 2009 ABS Survey on Occupancy and Housing Costs covers 18,000 households. The Australian Census covers the entire population and is more up to date, last executed in 2011.

The 2011 Census paints an a different picture of rental costs across Australia. First, it showed than median rents nationally rose by a whopping 50% between 2006 and 2011 (see next chart).

And because median household incomes grew by only 20% over the same period, according to the Census, the ratio of median rental payments to median household disposable income rose from 18.5% to 23.1% – a rise of 25% (see next chart).

There is also the next chart to consider, which shows a strong relationship between between the mid-to-late-2000s population surge and rising rental costs:

And the next chart showing minimal supply response to Australia’s growing population:

Using the more up to date and deeper data, it is clear that rental costs in Australia have surged since the mid-2000s due largely to a failure of housing supply to keep pace with strong population growth.

The Census can also be used to counter Cameron’s second argument, that Demographia’s Median Multiple (median house prices divided by median household incomes), is distorted:

Third, price-to-income ratios have shot up since 2000, but not quite as much as Demographia would have us believe (see top tight panel, and chart below). Perhaps their methodology, which ignores the apartment market, is generating bias in their results.

Again, the 2011 Census provides us with the answers. The below chart and table, which has been derived from RP Data analysis, shows that Australia’s Median Multiple grew by 40% between 2001 and 2011:

Further confirmation can be found in RBA charts, taken from the December Quarter 2012 Bulletin, which show Australia’s median multiples rising from roughly three times household incomes in the early 1980s to over six times incomes currently:

While these charts aren’t identical to Demographia’s, they are a lot closer than the data produced by Cameron above.

Finally, Cameron argues that the bulk of Australian house price growth has been caused by falling interest rates:

Finally, if one does want to consider price as a measure of affordability, they must at least be adjusted for changes to the interest rate. In the table below, using simple round numbers, it can be seen that a reduction in the interest rate on mortgages, from 10% to 5% means that prices can double while affordability remains constant in terms of the annual interest payment necessary to buy the home (median multiple of 6 instead of 3 in the bottom row).

If you want to be more conservative and consider principle and interests payments over a 30 year loan, then prices can increase by 63% while loan payments remain constant (median multiple from 3 to 4.89 in second last row).

In sum, the survey has little to add to the housing debate, apart from confusion. The price path we have witnessed for housing over the past decade has been the result of speculation, rising household incomes, and declining interest rates.

While no one disputes the fact that falling mortgage rates have contributed to Australia’s strong house price growth, Cameron again misses the point. According to the Census, the ratio of median mortgage interest payments to median household incomes rose from 25.6% to 33.7% between 2001 and 2011, representing an increase of 32% (see next chart). Moreover, variable mortgage rates were actually lower in 2001 (average = 6.8%) than in 2011 (average = 7.7% (standard) / 7.0% (discounted)).

While it is true that mortgage rates have fallen sharply since 2011, thanks to monetary easing by the RBA, the ratio of aggregate mortgage interest payments to household disposable incomes remains highly elevated, still tracking some 30% above the late-1980 highs when mortgage rates peaked at 17% (see next chart).

Of course the RBA chart above does not take into consideration the extra principal repayments that must also be paid arising from the significantly higher home values today compared with yesteryear.

In summary, the growth in Australian rental costs, median multiples and mortgage repayments is only partly explained by rising household incomes and/or falling mortgage rates as well as speculation. Structural factors such as an inability of land/housing supply to respond adequately to increases in demand, accounts for the rest.

[email protected]

www.twitter.com/Leithvo