Capital city price correlations

[Twitch]

I've fired up excel again, cranked the handle a couple of times and come up with even more meaningless statistics on capital city prices

This time it's the turn of correlation, as in the correlation of the prices and price movements between the cities, see the attached table for the gruesome detail. My 2c worth of observations are:

• All cities are pretty highly correlated, no surprise given they are all Australian cities and capital is portable between the states.
• The 2 cities with the highest correlation are Adelaide and Canberra.
• The 2 cities with the lowest correlation are Darwin and Sydney.
• The cities that most closely track Brisbane price movements are Hobart followed by Canberra.
• The cities the most closely track Melbourne price movements are Sydney followed by Adelaide.
• The cities that most closely track Perth prices are Darwin, then Brisbane.

Here’s a brief explanation of what correlation is:

• A correlation of 1 between two cities would mean their prices move together in lock step.
• A correlation of 0 would mean the price movements of two cities bear absolutely no relation to one another at all.
• A correlation of -1 means the two cities prices would move in perfect opposition to one another.

Finally, a high correlation correlation does not indicate dependence, neither does a low correlation indicate independence.

Acknowledgement: All the raw data is from Residex, the "value add" is from me

Enjoy

EDIT: Darwin and Sydney have the lowest correlation, not Perth and Sydney (although they are the second lowest correlated)

 

[alexlee]

Logically, it makes sense since Perth and Sydney probably have the most different economic characteristics amongst all the cities.

But Twitch, how does timing come into this? I mean, can you see a pattern to the cycle, in terms of which city booms after which one?
Alex

 

[Twitch]

But Twitch, how does timing come into this? I mean, can you see a pattern to the cycle, in terms of which city booms after which one?
Alex

I'm still working on the best method for determining this ....

 

[alexlee]

I'm still working on the best method for determining this ....

Will just graphing all the prices with different colours for different cities give us an idea? I'm picturing a series undulating, colourful waves.
Alex

 

[Ajax]

Alex,

do you mean like this?

http://www.somersoft.com/forums/gallery/showphoto.php?photo=2014&what=allfields&name=Ajax

 

[Peter Parker]

An interesting excercise, but if we apply a Terra Nullius theory of value since 1788 (to take a longer-term perspective) then the difference even between Sydney and Hobart is small (ie taken over 200 years to achieve a 2:1 difference). And since Perth and the other cities were settled later than Sydney, if you use settlement dates as your baseline, then other cities have beaten Sydney since they're younger and almost as dear!

A problem with graphs like the above is that they emphasise recent changes (eg Perth's almost exponential growth) and downplay those near the start of the graphing period (since all the lines are close together and obscured).

There were certain times when particular cities were particularly cheap or had big growth spurts.

Eg:

* Melbourne was very sluggish for a long time in the 1990s, whereas Sydney (which lacked the state financial crises that Vic, SA & WA had) moved faster (the direct opposite to 2007).

* Adelaide had a boom around 1985 that the other states didn't have.

* Perth, Sydney and Melbourne grew faster than some of the others (eg Adelaide & Hobart) in the late 80s boom.

* Canberra was insulated from the early 90s recession but had a slow period in the late 90s.

* Perth was cheap compared to the east in 2002, as was Brisbane compared with Sydney for a period in the 1990s/early 2000s.

If you take out these few state-specific things then growth rates are suprisingly similar. Early-ocurring or particuarly hot booms are generally met with a slower than average growth later.

And note that figures for any one city are composites of inner and outer areas, and as we're seeing at the moment in Melbourne there are huge differences between them, so the average doesn't mean that much.

Peter