House Price Market - How Should that be measured?

Forums › Home Workshop DIY › House Price Market - How Should that be measured?

tdgeek

29753 posts

Uber Geek

Trusted

Lifetime subscriber

#237546 7-Jun-2018 08:56

Today there was an article in Stuff about the housing market. People latch onto this, but I can't see that its accurate at all. Say there was a spate of super expensive house sales in AKL or elsewhere. Say there was a spate of no or low sales of expensive houses. Say many young ones are buying lower priced homes due to affordability. These affect the average or median house price, but they do not reflect if house prices are going up, or down, or flat. If the sell prices were based on comparison to GV then at least there is a line in the sand, it doesnt matter if GV are not real, they are based on the same factors, although they can exclude some improvements

Fred99

13684 posts

Uber Geek

#2031219 7-Jun-2018 09:08

You can be certain of only one thing:

Depending on the reason for pushing a barrow, the person or organisation pushing it will present selected data that supports their bias.

Be that "average" or "median" sale price, select a time period to "show a trend", a correlation between something like immigration and sales numbers or whatever.

tdgeek

29753 posts

Uber Geek

Trusted

Lifetime subscriber

#2031225 7-Jun-2018 09:12

Fred99:

You can be certain of only one thing:

Depending on the reason for pushing a barrow, the person or organisation pushing it will present selected data that supports their bias.

Be that "average" or "median" sale price, select a time period to "show a trend", a correlation between something like immigration and sales numbers or whatever.

Muldoon was good at that, nicknamed the numbers game, showing a graph, and managing the axes to give the chosen end result.

House prices is a big national topic, its a pity that its fudged

MikeAqua

7785 posts

Uber Geek

#2031227 7-Jun-2018 09:14

A run of high or low sales does not affect the median that much, in large data sets. That's the benefit of a median. As the middle value in a numerically ordered data set and not very sensitive to values at either end of the distribution.

A mean (what most people call an average) is much more sensitive to skewing.

The other option would be to use the mode, which is the most common value. For example: Round all sale prices to the nearest $10k and report the most common sale price. I don't think that would be much different to the median.

Mike

tdgeek

29753 posts

Uber Geek

Trusted

Lifetime subscriber

#2031237 7-Jun-2018 09:22

MikeAqua:

A run of high or low sales does not affect the median that much, in large data sets. That's the benefit of a median. As the middle value in a numerically ordered data set and not very sensitive to values at either end of the distribution.

A mean (what most people call an average) is much more sensitive to skewing.

The other option would be to use the mode, which is the most common value. For example: Round all sale prices to the nearest $10k and report the most common sale price. I don't think that would be much different to the median.

But if the median was based on more volume of low priced houses and less of high priced houses, it will reduce, even if the prices were increasing

MikeAqua

7785 posts

Uber Geek

#2031262 7-Jun-2018 09:46

tdgeek:

But if the median was based on more volume of low priced houses and less of high priced houses, it will reduce, even if the prices were increasing

Not necessarily. There tends to be most similarity among the numbers in the middle of data set. So if you pad one end of the dataset you may push the median onto another similar value.

Of course there is limit to this - a signification proportion of outlier values will move the median. But at that proportion they aren't really outliers and the median should move.

Mode (most common value) may be less sensitive than median to outlier data. I'm not sure. I've never used it other than in Uni assignments. Many, many years ago

Mike

Aredwood

3885 posts

Uber Geek

#2031272 7-Jun-2018 10:10

Different councils do their rating valuations at different times. So even using % Over GV is still no good for comparing say Auckland house prices to Hamilton house prices.

tdgeek

29753 posts

Uber Geek

Trusted

Lifetime subscriber

#2031279 7-Jun-2018 10:16

Aredwood: Different councils do their rating valuations at different times. So even using % Over GV is still no good for comparing say Auckland house prices to Hamilton house prices.

I thought that might be the case. I'm looking more at a % change relative to GV