A scientific method to compare the value of suburbs

[inv99]

One bit i think is missing is weighting for each area. In the above model, each of the 6 scores is given weight - 16.6%.

Just to add another spanner in the works which will make the model more accurate is to adjust scores to put more emphasis on differing factors, eg cbd might be more important than proximity to water.

This can either be worked out by yourself to fit your own personal criteria, or you could use trial + error by using known values to find which fits your pre-determined answer the most accurately.

eg. Changing:
CBD - 17%
Water - 17%
Trees - 17%
Schools - 17%
Transport - 16%
Local Facitilities - 16%

to:

CBD - 25%
Water - 10%
Trees - 15%
Schools - 15%
Transport - 19%
Local Facitilities - 16%
(numbers totally made up).

 

[Peter Parker]

One bit i think is missing is weighting for each area. In the above model, each of the 6 scores is given weight - 16.6%.

Just to add another spanner in the works which will make the model more accurate is to adjust scores to put more emphasis on differing factors, eg cbd might be more important than proximity to water.

Yep, fair point.

I implied some weighting in my initial post (by seperating the three biggest and three lesser contributors to suburb value) but didn't carry this through in the previous post.

This was partly due to slackness and partly due to wanting to see how it would work with equal weighting. Then looking how the results change by introducing weighting factors.

This can either be worked out by yourself to fit your own personal criteria, or you could use trial + error by using known values to find which fits your pre-determined answer the most accurately.

Depends on the purpose. The reason for the formula in the hands of the buyer is to (i) identify areas of relative good and poor value and (ii) as a tool to compare similar suburbs.

(ii) could be useful, especially for roughly similar suburbs. Everyone knows that Toorak and Melton are chalk and cheese so there's no point in messing with numbers to prove it. But it could be useful for roughly similarly priced areas, eg Glen Waverley versus Blackburn North or Rowville or Melton versus Hoppers?

I would prefer not to have a pre-determined answer; the formula is intended to illuminate rather than justify.

However we all have mental price 'anchors' for particular suburbs. It becomes more real and less abstract if we multiply our score to a figure that we can compare to varius areas, eg $200k, $400k or $800k.

 

[Peter Parker]

Equal weighting of all six factors has given the following figures which were both hastily done and rubbery.

Broadmeadows $400,000
Brunswick $560,000
Carnegie $560,000
Caroline Springs $240,000
Chelsea $520,000
Cranbourne $320,000
Croydon $440,000
Dandenong $360,000
Doncaster $360,000
Doveton $160,000
Eltham $480,000
Epping $360,000
Essendon $680,000
Footscray $520,000
Frankston North $280,000
Frankston South $560,000
Glen Waverley $640,000
Greensborough $440,000
Heidelberg $720,000
Kew $800,000
Laverton $320,000
Melton $200,000
Mentone $760,000
Moonee Ponds $720,000
Richmond $680,000
Sandringham $680,000
South Yarra $840,000
Sunshine $400,000
Templestowe $480,000
Toorak $920,000
Wantirna $280,000
Werribee $320,000
Williamstown $640,000

Notes:

* Others might have different assessments so figures will vary somewhat. Eg - does a 'good school' have to be in the suburb, or is the next suburb or on the same tramline good enough? What about well-regarded primary schools.

The definition of a supermarket may vary (I've preferred to include larger Coles/Safeway only) and it's unfair to rate an area with 4 tramlines but no trains less than a 4/4 for transport. In contrast country trains are infrequent so were disregarded.

* A problem with a small number of groups is the coarseness of categories. A better model would use actual numbers (eg distance from the beach) and a formula instead of crude groups.

* There is a ratio of arough 6:1 which is less than average house price differences between the poorest and richest suburbs. A higher multiplier for some would increase this. Also posh suburbs may have a '**** factor' and, from a value perspective, one reaches a point of diminishing returns in terms of basic services and facilities once one exceeds 'upper-middle class', or around $700k.

* As mentioned before the above scoring ignores size, style and quality of housing. This would favour the leafy suburbs. And because of the bias towards facilities and transport, a 'good' suburb like Templestowe scores poorly.

 

[Peter Parker]

Here's a repetition of the above but with weighting.

Proximity to CBD, proximity to water and tree were given double importance with a x2 weighting.

Broadmeadows $346,671
Brunswick $533,340
Carnegie $506,673
Caroline Springs $213,336
Chelsea $506,673
Cranbourne $240,003
Croydon 2 $400,005
Dandenong $293,337
Doncaster $346,671
Doveton $133,335
Eltham $453,339
Epping $293,337
Essendon $613,341
Footscray $480,006
Frankston North $346,671
Frankston South $533,340
Glen Waverley $560,007
Greensborough $400,005
Heidelberg $773,343
Kew $746,676
Laverton $293,337
Melton $160,002
Mentone $720,009
Moonee Ponds $640,008
Richmond $640,008
Sandringham $666,675
South Yarra $853,344
Sunshine $346,671
Templestowe $560,007
Toorak $906,678
Wantirna $266,670
Werribee $240,003
Williamstown $666,675

Notes:

* Because the figures generated a score out of 36 (instead of 24) the multiplier factor was lowered to provide an equivalence with the earlier table. In other words perfect score on each factor gave an area worth of $960k.
* The weighting provided a greater dispersion between cheap and expensive areas. This is probably closer to actual property values.

 

[DavidMc]

Frankston North seems like good value.

 

[DavidMc]

Doncaster seems off. I know the area is favoured by people with an Asian background, is it overvalued or are those city views really that great?

 

[DavidMc]

Footscray seems about right, not as amazing value as we thought it all. Sure it's close, I guess it's the lack of tree coverage that kills it.

I've always liked the leafy streets too... don't know why people just go around and plant some more trees. Sure it takes a few years to get going, but such an improvement for the cost.

Interesting thing out of this, if it's shown what such an effect tree coverage can have then it might be a little hint to look for areas which aren't considered tree suburbs but have little pockets of tree lines streets. Changes are these may have been already reflected though in prices.

 

[Peter Parker]

Interesting thing out of this, if it's shown what such an effect tree coverage can have then it might be a little hint to look for areas which aren't considered tree suburbs but have little pockets of tree lines streets. Changes are these may have been already reflected though in prices.

DavidMc, Of the six factors, tree coverage is the 'odd man out'.

I wanted some factor to explain why when all other facts are equal, the eastern suburbs are dearer than those to the north or west. Eg a suburb 30km east might be a similar price to one 20km west of the CBD.

Tree coverage, rainfall or undulating topography all seem to be associated with the more affluent suburbs; the west and north are known as flat and featureless plains.

While there appears to be an association, I cannot claim it to be any form of cause and effect (though it might be as people prefer certain types of landscapes and will pay more for them). Nevertheless if one was planning to buy in a new suburb, it would not be silly to buy in one with wide verges, trees planted (and few vandals) so that in time the streetscape becomes less raw and more like a high status suburb, both of which can't harm capital growth!

Another explanation of the east being dearer could be that more people live on the east side than the west side and the geographic centre of Melbourne is in the south-eastern suburbs. However this does not explain why with other cities (eg Sydney) where the geographic centre is also skewed (in this case to the west) yet their more populous western suburbs are cheaper, not dearer, than the east.

 

[Peter Parker]

Frankston North seems like good value.

While Caroline Springs, Doveton and Wantirna look like duds.

Doveton was probably harshly treated because its proximity to Dandenong was not factored in while Wantirna was disadvantaged by the transport formula only being for public transport (and not roads like Eastlink). Unless there are some really good schools that I didn't count, I can't see any redeeming features of Caroline Springs, although its average incomes are much higher than somewhere like Sunshine or St Albans.

On these calculations Melton also scores low, but it's also a cheap area, so it's value is not necessarily that poor.

 

[ILoveProperty]

Spiderman
Enjoying the thread. But looking at the weighted numbers - Heidelberg > Kew?
I wish it were true, but can't quite see it - How'd you arrive at this?

 

[VCDenton]

With broadmeadows and doncaster, would you be able to give a more breakdown analysis between these two suburbs as its quite surprising that they are of the same value.

I can think of

Broadmeadows=train, activity center(private money?), relatively easier access to the airport(not sure whether it is a good thing to a suburb), population increasing

Doncaster=activity center(westfield SC?k), close to some good schools, easy access to the city via eastern freeway(well at least off peak...)

 

[Peter Parker]

Heidelberg vs Kew (weighted): The factor that did it for Heidelberg was tree cover, though on reflection I was probably too harsh on Kew. On every other factor Kew was equal or superior to Heidelberg.

The above indicates that doing things by suburb is really not precise enough, especially in older areas where there's more diversity. Eg how does one rate suburbs where some streets are 'green' with old houses but others have only stunted trees with more units.

Broadmeadows vs Doncaster (weighted): Both Broadmeadows and Doncaster scored zero for 'good schools'. However the latter is probably unfair for Doncaster which is not too far from Balwyn High. It's not in the same suburb but it's relatively close. Taking a punt one might guess that the 'average' school in Doncaster is probably 'better' than the 'average' school in Broadmeadows due to differing catchments but my figures don't take that into account.

Future developments like Transit Cities or private investment were not included in the calculations - I stuck to strictly the six critieria and nothing else. Also I don't see a strong link between population growth and capital gain, even though there is certainly a link between population decline and sluggish growth. Take Toorak. Its price growth is driven by scarcity value because they're not making any land but everyone wants to live there. But in recessionary times fewer can justify or afford boom prices and it's not a necessity to live there, so prices get hit. Those boring 'bread and butter' or 'commodity' suburbs that are reasonably convenient tend to be more stable as everyone has to live somewhere and more people can afford them.

The big reason for Doncaster vs Broadmeadows the difference is public transport; central Broadmeadows has a direct train to the city whereas Doncaster has freeway buses which provide a good service during peak hour but less frequent at other times. There are plans for big bus improvements for Doncaster (DART). After these happen Doncaster will lift from being a 1/4 to a 2/4 for transport because it will clearly will fit into the 'good bus' category.

It sounds silly, but I cannot see how from the six factors Doncaster can be any more than slightly ahead of Broadmeadows, and this is due to better schools and tree cover (I actually gave Doncaster 4/4 which was being generous). Improved transport for Doncaster will increase the margin a bit more.

However beyond that, based on my six utilitarian measures of value, and not on 'snob factor' perceptions, Broadmeadows is better value, (even though I understand that if polled most people would likely pick Doncaster as the better place to live).

 

[Piston Broke]

Spiderman, while all this analysis is interesting, your still not accounting for one thing:
Buyers do not buy property by doing the analysis your doing.
They are obviously using a different yardstick or factors.
So if your trying to follow or pickup a trend, imo you wont.

Just like any market, the property market is made of people, which means it's irrational. Some people do all these calcs & when a boom comes along, show friends & relatives how much of a genius they are, but in effect it makes no difference. Sometimes even worse because it gives a false sense of security and narrows the options available.
I've seen 3 booms, and it's all been the same.
When a boom is underway, there are always people coming out with the "beaut" formula or secret to finding the best suburbs (or stocks).
Knowing which suburb will give better CG is more psychological than mathematical.
Typical case is Denham Court NSW, it's terrible in every sense but prices.
But those who live there, don't know better nor could care less.

 

[Peter Parker]

What you get is as important as what you pay

Spiderman, while all this analysis is interesting, your still not accounting for one thing:
Buyers do not buy property by doing the analysis your doing.
They are obviously using a different yardstick or factors.
So if your trying to follow or pickup a trend, imo you wont.
Just like any market, the property market is made of people, which means it's irrational.

All true.

However the object of all this is to arrive at a tool for identifying and comparing value of places based on a small number of easily ascertainable factors.

It is not necessarily a method to predict what's going to boom in the next year (or five).

Irrational markets present more opportunities than rational markets, especially if the value of the underlying assets is carefully studied.

At various times some suburbs are dearer relative to their amenity and others are cheaper relative to their amenity.

Because of people's work, schools and family ties etc, the residential market is 'sticky' and capital is less mobile. Therefore house price movement is slower than for shares which tend to be more volatile.

Some people hold the theory that asset prices in the long run are rational. While the price of a particular asset may have cycles of being over and undervalued, in the long run it is fairly priced. Variations tend to favour a reversion to the mean unless the asset is poor quality in which case it tends towards worthlessness. Identifying value is an important part of asset selection, even if it is only to avoid over-priced areas (which can nevertheless still appreciate due to non-value factors).

But even if you don't subscribe to this theory, it is probably fair to claim that relatively good value suburbs are likely to be more resilient than poor value areas, and since the former cost no more, why bother buying the latter?

 

[Piston Broke]

Lemme confess first that I'm an analytic type person by nature, i like numbers maths etc. It's just that it does'nt agree with my experiences be it RE, equities or derivatives.

"Irrational markets present more opportunities than rational markets, especially if the value of the underlying assets is carefully studied."
Oh yeah!

My point is that most people won't use your "method" for picking suburbs, streets & houses to live in.
Do you really think majority of people care about trees?
Not many many will admit it though.
What I'm saying is that generally imo it's the intagibles that people use to decide where to buy.
Just like someone who buys a holden or a ford.
There is no rational reason to a "holden or ford person".
People tell their kids "don''t be silly" when they do they same thing.
They buy their holden ford or wateva rust box, and only after
will find the reasons to justify it.
The trees only matter after the decision.
Even schools & facilities seem not to all that important anymore.
And of course all of this changes for different price categories.
So to answer your last question: imo nope.
Infact I'm seeing the opposite. The cheaper the house, the more it's retaining value.

 

[inv99]

I would prefer not to have a pre-determined answer; the formula is intended to illuminate rather than justify.

However we all have mental price 'anchors' for particular suburbs. It becomes more real and less abstract if we multiply our score to a figure that we can compare to varius areas, eg $200k, $400k or $800k.

The thing is you need to confirm the formula first before using it i think. So if you have a few suburbs that you know of, you use the formula to test those suburbs and get the answer you're expecting.

Once that's done, you use it to identify other places.

Idea is to increase the accuracy of the model by using things you know, and then testing it on other areas.

 

[Peter Parker]

The thing is you need to confirm the formula first before using it i think. So if you have a few suburbs that you know of, you use the formula to test those suburbs and get the answer you're expecting.

Once that's done, you use it to identify other places.

Idea is to increase the accuracy of the model by using things you know, and then testing it on other areas.

That sounds like a reasonable idea, although I don't really want it to accurately reflect current values across all suburbs as its usefulness would then be limited.

Instead I'd prefer it as a tool to identify possibly under and over-valued areas in an imperfect market.

So in this case differences are more a reason for further study (and possible buying opportunities!) rather than an indication that it needs to be tweaked to the nth degree.

It would be great if it had a growth dimension, but that requires time and hindsight. It's no good if you had a great method to identify value but didn't put it to use by buying good value properties at the time.

A 'magic formula' (which may not exist) may not necessarily remain valid over several property cycles so spending too much time on developing one based on the last boom (while others are buying) might not be a good use of time.