Now Listen Up Kev – What about this housing bizzo?

In an act of impeccable timing, George Meg over at the The Oz on Saturday had a great article on housing affordability (which has now turned up on his blog) and is a must read.

As happens disturbingly often with Mr Meganomics ( and why the timing is impeccable), I’ll often be planning to write about some particular thing, having a number of ideas floating around my head, only to be beaten to the punch by George not only on the topic, but often on the content and sometimes on the data as well. It gets a bit tin foil hattish at times.

There is very simple root cause to the lack of housing affordability in Australia, but it is also a cause that has virtually no easy solution – Australia simply has too few large cities for the proportion of our population that chooses to live in an urban environment. As a result, the supply of desirable urban locations to live – be they inner city or hilltop or hinterland or bayside locations relatively near a city centre – is swamped by the enormous demand. Desirable locations lead the charge in house price growth which pulls the price of all suburbs up with it, generally in decreasing amounts the further away one gets from the suburbs with a high desirability premium.

But since we can’t really change this situation over any time frame shorter than the very long term, even if we were to attempt to, the policy options left to address housing affordability issues become more complex as a result of our policy options being forced to try and affect the affordability ends from policy angles that aren’t necessarily the root cause of the affordability means.

As always, we should probably look at the data first and for those non-economists reading I’ll walk you through what I’ve done each step of the way. I’ll divide the posts into Part 1 and Part 2. Part 1 will develop a ratio of the real median house price to the real annual average male full time ordinary earnings, where both are adjusted for State based (rather than national) inflation, and both denominated in 2007 dollars so we can see how the real state of housing affordability has changed over the last 25 years.

In Part 2, the next post, we’ll get to the policy options.

PART 1

The two key metrics that are the foundation of housing affordability are income and house prices, with interest rates floating around as a third lesser order, though still important issue.

One of the most useful indicators of housing affordability is what’s called the median ratios – the ratio of the median house price to the median wage. What it tells us is how many times the median wage it costs to buy the median house. But there are also two things that should be taken into consideration when measuring housing affordability over time – the spatial distribution of wages and house prices (such as how the relationship between the two in different states plays out) as well as adjusting both for inflation to give real values, which in turn provide more meaningful results.

Unfortunately, I couldn’t get access to median wage data on a state by state basis over a decent time span in a timely fashion (which will become part of a larger bitch and moan I’ll have in another “Now listen up Kev” post in the near future – better national organisation of data sources to assist productivity growth), so instead of using median wages, we’ll use average full time male ordinary earnings from the ABS. The problem with this measure is that it gives an impression that the “average person” earns more than they actually do compared to using the median wage measure. This happens because high income earners skew the average upward. This is IMPORTANT to keep in mind. So while not perfect, it will still give us a consistent feel for the data over time even though “average” is slightly higher than the median (which is the income amount that exactly 50% of the population earns more than and 50% earns less than)

If we have a look at the annual average full time male ordinary earnings over the period since 1983, you might notice something that makes it a pretty useless measure in its current state:

A lot of the useful information we need here gets washed out by inflation. So what we need to do is deflate these state based nominal wage measurements by state based CPI to give us a measure of what the real wage levels in the states over time have been, but “real” in terms of local CPI rather than the national aggregate of CPI.  This is a bit of an unusual approach mind you, but IMHO results in a more accurate measure of “real wages” for the purpose of looking at state based housing affordability than using simple national measures of average wages and consumer price inflation. Once we do this, we end up with a dataset that measures the average wage over the last 25 years, but which is denominated in today’s dollar terms.

So what this allows us to do is to say, for example, that in 1998, the real wage in NSW in today’s dollars was $55000 a year. It lets us see how real wages have changed over the last 25 years now that the average wage has been adjusted for inflation.

What is interesting here is how real wages declined during the 80’s as a result of the Accord being used to control inflation, as well as how the long term benefits of that started to flow through after the 91 recession when real wages took off. Also interesting is how up until 2000, WA and NSW real wages walked virtually in lockstep with each other once they were adjusted for the inflation levels that were occurring in each State.

We can do the same thing for state based median house prices and adjust them for State based CPI to give us a “real” value of how house prices have changed over time by State, as well as measure the whole thing in terms of today’s dollars.

First off though, the nominal, unadjusted data of median house prices, by State, over the last 25 years:

After we adjust the data for State based inflation and place it all in 2007 dollar terms we get:

 

What is really interesting here in terms of housing affordability is that real house prices remained virtually frozen over the period from 1990 through to 2000. It wasn’t until Howard started stuffing around with halving the capital gains rate and things like the first home buyers grant that real house prices started to accelerate.

Now we can create our ratio by simply dividing the real house price by the real average wage, by state, and we get:

This real ratio tells us by how much the price of housing has increased in real terms. It also highlights in real terms just how much the NSW market has dropped over the last couple of years.

Next up – Part 2, policy options.