Qld property market crashes !! ;)

[Mike]

Hi Les,

Just wondered whether you thought of subscribing to:

Queensland Property & Lifestyle is the research publication of The Real Estate Institute of Queensland (REIQ) - produced quarterly, the 64 page magazine includes median sales data, vacancy rates, and rental data for nearly every suburb in the state. http://www.propertylifestyle.com.au/

Don't discount the importance of sales data to check whether current demand is up or down.

Steve McKnight has a good article on median prices at:
Making ¢ents of the $tatistics
http://www.propertyinvesting.com/article11_03

A property database:

QUEENSLAND VALUATIONS AND SALES DATA
http://www.abr.com.au/home/news.php3?ARTICLE=news.12

ERSIS Australia, in partnership with Australian Business Research, has produced three new datasets. The datasets are based on the Queensland Department of Natural Resources' real estate valuation and sales data. They include an Address Point File, a Property Database, and a Sales Database.

More than 1.6 million property records are contained in each database, which addresses over 98% of all Queensland property parcels.

The address point file database contains property addresses from the Queensland valuation database, with latitude and longitude for each address derived from the digital cadastre database.

The property database includes the parcel details and address, name and address of owner, and property valuation.

The sales database includes all the fields from the property database in addition to property sales history since 1990. It also contains the name and address of the vendors and purchasers and the purchase prices of the property.

Another database:

RP Data On-Line Property System
http://www.rpdata.net.au/home/brochures/online.html

A good article on factors affecting property values:

WHAT IS YOUR PROPERTY WORTH?
http://www.rainehorneqld.com.au/pages/pricing.html

Home Price Guide
Property Prices* 12 months to 31 March 2003 for postcode 4160

.        Houses 4160   Brisbane East 
Average   $291,972      $295,780 
Median    $272,750      $257,500 
Median change over last 12 months
           +13%             +21% 
  
         Units 4160    Brisbane East 
Average   $183,797      $227,689 
Median    $155,500      $195,500 
Median change over last 12 months
           -21%             +5%

 

[XBenX]

les,

i dont work for residex so i dont know how they calculate the growth figure - id assume that they exclude sales such as the one above (im pretty sure this is also a well protected secret)

the figures will only work where the true market value sale occurs id assume the residex method seeks a way to exclude the sales which are not at market value....

ive done some work calculating growth for suburbs - i did it manually with sales data .... when doing it manually you can tell when sometime doesnt seem quite right... say 200% growth in a year and exclude these outliers

its also impossible to quantify how much more accurate .... but changes in what is being sold is more likely to occur than sales occuring at prices that dont represent true market value ... but thats just my opinion as to why medians are not as good.

Originally posted by Les
Spill the beans, XBenX - just HOW does this growth figure get calculated ??????? The current trend seems to be to say it is "more accurate" than medians - fair enough, but HOW MUCH more accurate?

Can you help?

Regards,

 

[Kevmeister]

Even if you were to assume that all the sales figures were at market value, unless you can show the same house bought and then sold in a 3 month period, how can you reasonably say what the growth was on that house for that 3 month period.

Sure, the $120K property XBenX takes as an example which sells for $170K shows 5.5% growth over 7 years. But that says nothing about the growth which has occurred over the last quarter. That same house might also have sold for $170K one year previously, implying no growth in the last year.

So, because houses are not bought and sold that often, the growth is averaged over a longer term, which reduces the short-term statistics.

(Well, renos are usually bought/sold quite quickly, but they suffer from their own problem to use them for measure cap. growth, because they've had $$$$ invested in them, then sold, in which case the buy/sell price does not reflect true growth anyway, since a proportion of that $ growth is recouping the improvement costs).

Where's Mr Residex (JohnE) gone to ?

PS: Having an organisation like Residex say "trust us, the numbers are right" without describing the manner in which the numbers are calculated would seem a bit dodgy. I'm not saying they do this, just alluding to XBenX's comment about it possibly "being a well protected secret". If it isn't transparent then I don't even know if you can trust the numbers.

 

[XBenX]

gotta go a bit easy Kev - it might take him a bit of time to respond - not everyone is an addict like us

I dont know about the 3month rates ....

 

[JohnE]

Hi, Time is short in a 12 hour plus work day. The matters are very complex and very mathematical but in simple terms we record all sales that occur and have such data in most states for more than 25 years. That means we can calculate the return on each property that has been sold and resold over the last 20 plus years. Then using statistical techniques (regression techniques etc) we are able to break down this large mass of individual returns into period returns for segments of the market. To do the calculation there is no need to have properties buying and selling in the same three month period. For those interested in the very basic mathematical technique please see our original release in an Actuarial paper in 1992. Am sure the Institute would be prepared to provide a copy. We don’t have an electronic version in the office. Be warned it will be heavy reading for most.

Regards John E

 

[Bill.L]

Hi JohnE

I think you would be surprised by the number of people who would not only read your actuarial paper and understand it, but also question parts of its relevance and offer suggestions that haven't been thought about. If an electronic version could be provided somewhere, maybe pdf format, shouldn't be to hard to do.

P.S. I have some units of tertiary statistics in my background, and I'm sure there are more qualified people than me lurking here.

bye

 

[Kevmeister]

Which Institute are we talking about that might provide a copy - I'm interested in taking a look, though I don't know whether I will fathom it.

 

[Les]

G'day all,

Some good points made - thanks...... But I'm left with HOW is this "growth" figure arrived at? And, how much more valid is it than a change in medians? Can a more accurate figure be obtained by taking a "half-way point" between median change and growth % ????


e.g. even if "growth" is calculated ONLY from properties sold within a 12 month period - doesn't this limit it to an unadvisedly SMALL sample?? There must be more to it.

JohnE - can you provide a link to this "original release in an Actuarial paper in 1992" - I, for one, would be interested in reading just where this growth figure originates from, even if it takes a few days

Regards,

 

[Mike]

Hi Les,

Other reports I have seen use median change as the growth figure. For a start, it makes sense. I think Residex is confusing this issue by calculating growth some other way. Are they saying their growth calculation is more accurate than median change. If so, then that casts doubt on the stated median price. If the median price is doubtful then don't publish it.

If you are interested in the maths behind House Price Indexes, you can find an example here:

How is the Halifax House Price Index constructed?
http://www.hbosplc.com/help/faqs/housepriceindex.asp

For a broader discussion use the keywords hedonic house price in a Google search.

Regards, Mike

 

[Les]

G'day Mike,

If you are interested in the maths behind House Price Indexes, you can find an example here

WELL, I THOUGHT I was interested in the maths, but after looking at that, I lost interest real quick (but thanks, anyway, Mike... )

Are they saying their growth calculation is more accurate than median change

Yeah, Mike - from JohnE's posts I think that is exactly what they are saying. Unfortunately, for me, this just introduces yet another unknown variable.

I'd like to hear more from JohnE to clarify this .........

Regards,