Regressions, digressions, means and growth

[Peter Parker]

Spiderman - there is stat software that does this analysis and much more - Im struggling to remember it SPSS I think - google confirms that - its rapidly replacing parts of my memory =)

I dont have a copy anymore and Im sure there is something better being used now...

By far the the hardest thing would be getting reliable data going back a sufficient time (50 - 100 years). Then there's the painstaking duty of culling it to only include established homes of a particular type (off-the-plan building having existed at least since the 1880s).

Once you have the 50-100 years worth of data and have done averages for each year (suburb of interest AND metropolitan average), then the table and graph should be a piece of cake even if done by hand.

With a red pencil ('cause you can't afford textas until you've made money!) colour in the part of the graph where the suburb's growth exceeded the metropolitan average (colour between the line and the zero on the vertical (Y) axis).

Then with a dark pencil shade the portion where the suburb grew more slowly than the city average (again colour between the line and the zero on the vertical axis).

If you see a bigger area of red the suburb has mostly outperformed the metro average. If you see more black then it's been a laggard. A bigger area of both colours combined means high volatility with respect to the metro average; a smaller area means low volatility, ie it's a very stable, 'average' suburb.

Note that inaccuracies in data can affect percentages more than actual amounts. A house costing 500 pounds a century ago and now worth $500k would have a markedly different growth rate compared to if it actually only sold for 250 pounds. But the actual difference in gains would be imperceptible and swamped by the error margin, particularly in the earlier years.

Regards, Peter

 

[Aceyducey]

I used to use SPSS - frankly you can do most of this using Excel if you know how to write formulas.

Cheers,

Aceyducey

 

[NigelW]

Yes, but only if the extra value is more than what it costs you.

Rgds, Peter

If you weren't pretty sure on that point then you never should've done the reno in the first place (not forgetting the value of your time too)! If you're not getting at least 2-3 times your investment in revaluation and increased rental then you're wasting your time and money.

I do think you've raised a VERY interesting issue. Surely as property prices increase the % increase must start to flatten out... a 10% increase on a $1m property is equal to a 33.3% increase on a $300k one etc... particularly if there's a correlation (however elastic) between AWOTE and prices...

I've said before I think it's easier to bring a below average property up to average or above than the make a property truly exceptional through reno. I think you get to a point of diminishing returns (perhaps unless you're operating in the luxury/ultra-pricey space...which would have its own risks).

Equally though, does the regression to mean analysis break down at the extreme upper end of the market??? (probably academic for all of us I guess! )

Cheers
N.

 

[Peter Parker]

I do think you've raised a VERY interesting issue. Surely as property prices increase the % increase must start to flatten out... a 10% increase on a $1m property is equal to a 33.3% increase on a $300k one etc... particularly if there's a correlation (however elastic) between AWOTE and prices....

Nigel, just to clarify I assume that when you say 'as property prices rise' you mean you're going towards up market properties (ie from median value to 2, 3, 4 times etc) rather than referring to normal capital growth.

I've said before I think it's easier to bring a below average property up to average or above than the make a property truly exceptional through reno. I think you get to a point of diminishing returns (perhaps unless you're operating in the luxury/ultra-pricey space...which would have its own risks).

As a first guess you could look at Toorak house values relative to the metro average. Then hone in on only the dearest quarter of Toorak houses, to get the 'luxury house' figures. My gut feeling is that prices may be volatile but not necessarily outperform the market.

Who would rent a luxury house? By definition it would be those who can afford to buy but don't as they're only itinerants (eg CEOs of multinationals) want to impress others with a premium address. With yields near 2%, you'd barely cover holding costs even if you paid 100% cash.

And who would buy it? It would be a very selective market. Melbourne may be different, but the high profile nouveau riche in Perth have been vain enough to want to have a new home custom-built rather than move into a secondhand house.

As investments they don't look very good to me, and may be more like items like art, antiques and wine, where investment is just one reason for posessing.

Equally though, does the regression to mean analysis break down at the extreme upper end of the market??? (probably academic for all of us I guess! )

I would guess not. But rather being tied to average incomes, it may have more to do with:

1. Wealth of entrepeneur as a proportion of national wealth (ie is there a small class of super-rich?)
2. Proclivity of entrepeneurs to display their wealth through flashy houses

A book I read (forgot the title) on the history of entrepeneurs in Australia showed that there were some impressive fortunes (relative to total national wealth) held by some entrepeneurs in the late 19th century. This declined for much of the 20th century, but made a comeback from the 1970s.

In relation to 2. JK Galbraith mentions that for most of the 20th century, ostentatious displays of wealth by the rich were regarded as gauche, distasteful or likely to promote ferment amongst the proletarians, but such social inhibitions loosened since the 1970s.

What people are willing to pay for luxury houses can't be divorced from these trends which affect both the number of people with the capacity to pay and their willingness to do so.

Regards, Peter

 

[Peter Parker]

possible example of regression towards the mean

from the Herald Sun by Kylie Hansen 06 June 2005

MEET our new breed of fringe dwellers – cashed up established families fleeing the mortgage belt.

These middle Melbourne families are selling their expensive suburban homes in exchange for larger, newer and cheaper homes in outer rim estates.

'Here you can still commute to the city but instead of living somewhere like Caulfield where you have a half million dollar mortgage, (here) it's half that with all the facilities'

more at:
http://www.heraldsun.news.com.au/common/story_page/0,5478,15518766%5E2862,00.html

Now this may not necessarily be indicative of a mass trend. After all it's easy enough to get a photo of a relocated family, a quote from an academic and a couple of grabs from the local real esate agent and presto, you have an article.

One if not two of the fastest three appreciating suburbs (Hawthorn, Williamstown and Cranbourne) could be described as 'inner'. And the article glosses over things like fuel and transport costs from the more distant suburbs (especially in light of 'peak oil').

BUT if the trends in the article are substantial and sustained, it may result in a partial closing of property price gaps between inner and outer suburbs. For this to happen, outer areas will need to appreciate at a faster rate than the inner areas (automatically favoured by some investing gurus) for the next 5-10 years.

Also the article draws out three recurring themes of this series of posts.

These are:

i. The idea that a suburb obtaining consistently above average capital growth over a long term is unlikely. The likelihood drops as the term is increased.
ii. It may be more probable that capital growth rates of various suburbs regressing towards a mean over the longer term.
iii. Property values are socially determined. Fashion drives demand. Demand drives prices.

Regards, Peter