|
Louis Christopher SQM cocks up his attack on RP Data
|
|
Topic Started: 4 Jul 2012, 10:14 AM (2,839 Views)
|
|
TheRazorsEdge
|
5 Jul 2012, 04:01 PM
Post #46
|
- Posts:
- 405
- Group:
- Gold Member
- Member
- #2,160
- Joined:
- 26/06/2012
|
- Aussiehouseprices
- 5 Jul 2012, 03:53 PM
Where did you think it was grown??
Macquarie St. Sydney.
|
|
|
| |
|
earthsta
|
5 Jul 2012, 06:00 PM
Post #47
|
- Posts:
- 4,545
- Group:
- Diamond Member
- Member
- #31
- Joined:
- 02/11/2010
|
- Aussiehouseprices
- 5 Jul 2012, 01:52 PM
- miw
- 4 Jul 2012, 08:33 PM
Is this what you were looking for under "independent variables"? (They call these the "hedonic variables" in the methodology paper.
Suburb
Land size (m2) if property type = house
Street classification (highway, main road, cul de sac etc.)
Property type (eg. free standing house, semi etc. for houses)
Number of bedrooms
Number of bathrooms
Number of car spaces
Pool (Y/N)
Waterfront (Y/N)
Air Conditioning (Y/N)
View (Y/N)
If you have a view of a sewage farm, would you get a "Y" or "N"?
Ask Frank. His PPOR overlooks Luggage Point 
|
|
|
| |
|
TheRazorsEdge
|
5 Jul 2012, 06:49 PM
Post #48
|
- Posts:
- 405
- Group:
- Gold Member
- Member
- #2,160
- Joined:
- 26/06/2012
|
- earthsta
- 5 Jul 2012, 06:00 PM
Ask Frank. His PPOR overlooks Luggage Point
Frank lives on Nudgee Golf Course??? No wonder he is so relaxed.
|
|
|
| |
|
miw
|
5 Jul 2012, 07:55 PM
Post #49
|
- Posts:
- 3,903
- Group:
- Diamond Member
- Member
- #1,917
- Joined:
- 22/04/2012
|
- Aussiehouseprices
- 5 Jul 2012, 01:46 PM
What about age of house / how recently renovated?
Say I buy an old place, renovate every room and sell for a higher price, because none of the variables listed have changed, it will help to boost the index.
I would have thought that the point of a hedonic index is that it shouldn't.
There are bunch of things that I would have thought were important that aren't in there. Proximity to public transport has already been mentioned.
Some of them are no doubt not there because the data just has not been collected for a big enough percentage of the stock - age of building and current state of renovation would come into that category I guess, but proximity to transport woud be easily calculable and is quite significant.
Whether it has any huge impact on the index I wouldn't know.
|
Last time I followed my dreams it led me to a casino, a bar, a gun shop, then a bank. Never again!
AREPS™
|
| |
|
Perthite
|
5 Jul 2012, 08:30 PM
Post #50
|
- Posts:
- 1,064
- Group:
- Platinum Member
- Member
- #229
- Joined:
- 07/01/2011
|
With regards to the accuracy of RPData. What's with the voilent fluctuations in Perth?
http://www.rpdata.com/research/back_series.html
|
|
|
| |
|
miw
|
5 Jul 2012, 09:08 PM
Post #51
|
- Posts:
- 3,903
- Group:
- Diamond Member
- Member
- #1,917
- Joined:
- 22/04/2012
|
- Perthite
- 5 Jul 2012, 08:30 PM
The whole index seems somewhat borked across the change in financial years, actually. I've been watching Brisbane and either armageddon is here, or there is severe short-term bogosity in evidence.
|
Last time I followed my dreams it led me to a casino, a bar, a gun shop, then a bank. Never again!
AREPS™
|
| |
|
TheRazorsEdge
|
5 Jul 2012, 09:25 PM
Post #52
|
- Posts:
- 405
- Group:
- Gold Member
- Member
- #2,160
- Joined:
- 26/06/2012
|
- miw
- 5 Jul 2012, 09:08 PM
The whole index seems somewhat borked across the change in financial years, actually. I've been watching Brisbane and either armageddon is here, or there is severe short-term bogosity in evidence.
If one could determine the weakness in the methodology, it could provide arbitrage. I think the independent variables list is probably not resilient enough to measure exogenous effects such as stamp duty extensions/FHOG, sudden changes to local employment (such as mass layoffs), floods, bushfires, zoning corruption scandals in local government.
Still, betting against a Goldmanite is more dangerous than dancing with the devil, so best just give it a wide berth.
|
|
|
| |
|
kennyjaiz
|
5 Jul 2012, 10:06 PM
Post #53
|
- Posts:
- 1,149
- Group:
- Platinum Member
- Member
- #11
- Joined:
- 29/10/2010
|
- miw
- 5 Jul 2012, 07:55 PM
There are bunch of things that I would have thought were important that aren't in there. Proximity to public transport has already been mentioned.
Some of them are no doubt not there because the data just has not been collected for a big enough percentage of the stock - age of building and current state of renovation would come into that category I guess, but proximity to transport woud be easily calculable and is quite significant.
Whether it has any huge impact on the index I wouldn't know.
I think the index attempts to cater for proximity to public transport by "weighted region", with region focused and targeted enough to detect the difference in value resulting from proximity to amenities and transport. How sensitive this weighted region is, I'm not sure.
But age of building, recency of renovation and quality of renovation are not captured, as you mentioned.
|
|
|
| |
|
miw
|
5 Jul 2012, 10:18 PM
Post #54
|
- Posts:
- 3,903
- Group:
- Diamond Member
- Member
- #1,917
- Joined:
- 22/04/2012
|
- TheRazorsEdge
- 5 Jul 2012, 09:25 PM
If one could determine the weakness in the methodology, it could provide arbitrage. I think the independent variables list is probably not resilient enough to measure exogenous effects such as stamp duty extensions/FHOG, sudden changes to local employment (such as mass layoffs), floods, bushfires, zoning corruption scandals in local government.
Still, betting against a Goldmanite is more dangerous than dancing with the devil, so best just give it a wide berth.
Yes. I think I mentioned this before as well. Even excluding manipulation (which I think is highly unlikely - it would kill the goose as dead as a stone eventually) and poor methodology (I think the methodology is pretty robust given the basic intractibility of the problem) There are a bunch of issues that may make the index an unsuitable underlier for derivatives.
What I *think* we are seeing here in Brisbane is the effect of the stamp duty change. Over the last half of June there would have been a reduction in new contracts as people wait for the new regime, but old data from May, April and even March would still be coming in at the normal rate, thus changing the temporal composition of incoming data and causing a predictable anomaly. If I am right, we should see the index snap back to levels around that of June 25th quite quickly, maybe even overshoot as the deals that were closed at the end of June finally get reported and into the data. If I am wrong, then the much-hailed crash may have hit Brisbane in the last week.
Similarly for natural disasters, etc etc. as you point out. The index is a lagging indicator that takes a while to reflect information that we already know. Most if not all derivative underliers reflect all known information almost instantly, minimising arbitrage opportunities.
This is not to say it can't be used and that what we'll see is futures and options prices (options will effectively use the futures contract as the underlier, not the index) will not adjust and move away from the index when the market knows the index has some catching up to do, but there is a long way to go before the index can earn its stripes. Perhaps 10 years.
|
Last time I followed my dreams it led me to a casino, a bar, a gun shop, then a bank. Never again!
AREPS™
|
| |
|
miw
|
5 Jul 2012, 10:27 PM
Post #55
|
- Posts:
- 3,903
- Group:
- Diamond Member
- Member
- #1,917
- Joined:
- 22/04/2012
|
- kennyjaiz
- 5 Jul 2012, 10:06 PM
I think the index attempts to cater for proximity to public transport by "weighted region", with region focused and targeted enough to detect the difference in value resulting from proximity to amenities and transport. How sensitive this weighted region is, I'm not sure.
But age of building, recency of renovation and quality of renovation are not captured, as you mentioned.
I agree. They probably can get some notion of location desirability in there.
Another thing that is glaringly missing is inside floor area.
Makes me think that their major/only source of data is real estate listings. In Australia, the floor area and the date of build/renovation are not often found in listings. If the floor area is listed it is very often a lie.
Finding out the age of a building is not a trivial exercise, at least in Brisbane - which is surprising because it is such an important detail. You find yourself narrowing down on it by looking at the dates of documents that have to have been dated before start of construction and others that can't be done until after completion, but you usually end up with 4-6-month windows for building start and building completion. You are often reduced to opening up the electrical switchboard which usually has an installation date in it and you know that would have gone in reasonably shortly before completion.
|
Last time I followed my dreams it led me to a casino, a bar, a gun shop, then a bank. Never again!
AREPS™
|
| |
|
kennyjaiz
|
5 Jul 2012, 10:52 PM
Post #56
|
- Posts:
- 1,149
- Group:
- Platinum Member
- Member
- #11
- Joined:
- 29/10/2010
|
- miw
- 5 Jul 2012, 10:27 PM
I agree. They probably can get some notion of location desirability in there.
Another thing that is glaringly missing is inside floor area.
Makes me think that their major/only source of data is real estate listings. In Australia, the floor area and the date of build/renovation are not often found in listings. If the floor area is listed it is very often a lie.
Finding out the age of a building is not a trivial exercise, at least in Brisbane - which is surprising because it is such an important detail. You find yourself narrowing down on it by looking at the dates of documents that have to have been dated before start of construction and others that can't be done until after completion, but you usually end up with 4-6-month windows for building start and building completion. You are often reduced to opening up the electrical switchboard which usually has an installation date in it and you know that would have gone in reasonably shortly before completion.
You may be right. But the information presented on a real estate listing is not systematic (i.e. random missing information) and full of error. Hopefully they at least validate the data integrity with secondary or historical sources (e.g. local council, bank, insurance company). Local council planning office should have close enough information about total building floor area, land area, date of build and date of extension, etc. I would have thought 4-6 months discrepancy to be tolerable, but that's just me.
|
|
|
| |
|
miw
|
5 Jul 2012, 11:17 PM
Post #57
|
- Posts:
- 3,903
- Group:
- Diamond Member
- Member
- #1,917
- Joined:
- 22/04/2012
|
- kennyjaiz
- 5 Jul 2012, 10:52 PM
You may be right. But the information presented on a real estate listing is not systematic (i.e. random missing information) and full of error. Hopefully they at least validate the data integrity with secondary or historical sources (e.g. local council, bank, insurance company). Local council planning office should have close enough information about total building floor area, land area, date of build and date of extension, etc. I would have thought 4-6 months discrepancy to be tolerable, but that's just me.
The data would be better if they validated it. I will have another read of the methodology to see if they do, but seem to remember reading that they put default values in where they don't have the data, and update it as new listings for the same property come in. There was also a protocol for dealing with conflicting listings.
Strata plans have floor areas in them. Not sure if the floor areas of detached houses is recorded in any systematic way.
|
Last time I followed my dreams it led me to a casino, a bar, a gun shop, then a bank. Never again!
AREPS™
|
| |
|
TheRazorsEdge
|
5 Jul 2012, 11:33 PM
Post #58
|
- Posts:
- 405
- Group:
- Gold Member
- Member
- #2,160
- Joined:
- 26/06/2012
|
- miw
- 5 Jul 2012, 11:17 PM
The data would be better if they validated it. I will have another read of the methodology to see if they do, but seem to remember reading that they put default values in where they don't have the data, and update it as new listings for the same property come in. There was also a protocol for dealing with conflicting listings.
Strata plans have floor areas in them. Not sure if the floor areas of detached houses is recorded in any systematic way.
Ultimately, they are trying to build a daily index on an asset that is
1) Completely heterogeneous. Every property is a unique item.
2) Illiquid. Even at it's highest turnover rate, it takes at least 6 weeks from offer to settlement. Do you base the index on offers (that may fall through) or settled contract price?
3) Irrelevant. It really matters what the price of oil is today, next week, next quarter. Property, not so much.
This index is more like entertainment (because the crowd love property) than a financial index with any inherent use or meaning.
|
|
|
| |
|
newjez
|
5 Jul 2012, 11:41 PM
Post #59
|
- Posts:
- 3,128
- Group:
- Diamond Member
- Member
- #547
- Joined:
- 30/04/2011
|
- miw
- 5 Jul 2012, 09:08 PM
- Perthite
- 5 Jul 2012, 08:30 PM
The whole index seems somewhat borked across the change in financial years, actually. I've been watching Brisbane and either armageddon is here, or there is severe short-term bogosity in evidence.
I could have sworn you made those two words up - but no - they are real (at least on google) 
I think Perth fluctates because of (especially at the moment) low volumes. It's like looking at the REIWA suburb data.
I think you have to add 1-2% variance to Perth's data before you can read a significant change into it.
Edited by newjez, 5 Jul 2012, 11:43 PM. |
A plug for my books on kindle Jack - a thin veil of duplicity
and Summer Days
|
| |
|
miw
|
6 Jul 2012, 12:12 AM
Post #60
|
- Posts:
- 3,903
- Group:
- Diamond Member
- Member
- #1,917
- Joined:
- 22/04/2012
|
- TheRazorsEdge
- 5 Jul 2012, 11:33 PM
Ultimately, they are trying to build a daily index on an asset that is
1) Completely heterogeneous. Every property is a unique item.
2) Illiquid. Even at it's highest turnover rate, it takes at least 6 weeks from offer to settlement. Do you base the index on offers (that may fall through) or settled contract price?
3) Irrelevant. It really matters what the price of oil is today, next week, next quarter. Property, not so much.
This index is more like entertainment (because the crowd love property) than a financial index with any inherent use or meaning.
Yup. Unless it can prove itself as a derivative underlier, it might as well come out once a month.
|
Last time I followed my dreams it led me to a casino, a bar, a gun shop, then a bank. Never again!
AREPS™
|
| |
| 1 user reading this topic (1 Guest and 0 Anonymous)
|
8:15 AM May 20
