Latest RoS, Scotland Prices Rising Strongly......

HAMISH_MCTAVISH

Generali wrote: »

That can be seen by the extreme volatility of the figures - Aberdeen is up 10% on the year but down something like 7 or 8% on the quarter in the last quarterly figures,

But in fairness, most of Scotland shows seasonal volatility, with 10% differences between the summer/autumn peak and mid winter being common. The quarterly stats can reflect that.

That house prices in Aberdeen are up 12% on the year feels right: it's a bit more than the average across the UK and the dominant industry in the city is doing well. Using RoS figures to measure that is a poor way to measure it IMO.

If you look at the monthly stats in graphical form, (available here... http://www.housepricecrash.co.uk/forum/index.php?showtopic=110165&st=135 )

Aberdeens trends match up almost perfectly with the trends in other Scottish cities. If the results were unreliable due to low volume, we'd surely expect to see far more randomness between the major cities than we do.

Which leads me to believe that the sample size, whilst smaller than it was at peak, is still large enough to be statistically significant.

Doctor_Gloom

DaddyBear wrote: »

But posting at 3am on a bank holiday weekend when you have no VI in what you are posting is just plain weird.

I think everyone can agree that Hamish seriously needs to get a life.

Graham_Devon

Doctor_Gloom wrote: »

I think everyone can agree that Hamish serriously needs to get a life.

Probably not the best thing to say when you, I, and everyone else is on here too!

Generali

HAMISH_MCTAVISH wrote: »

But in fairness, most of Scotland shows seasonal volatility, with 10% differences between the summer/autumn peak and mid winter being common. The quarterly stats can reflect that.



If you look at the monthly stats in graphical form, (available here... http://www.housepricecrash.co.uk/forum/index.php?showtopic=110165&st=135 )

Aberdeens trends match up almost perfectly with the trends in other Scottish cities. If the results were unreliable due to low volume, we'd surely expect to see far more randomness between the major cities than we do.

Which leads me to believe that the sample size, whilst smaller than it was at peak, is still large enough to be statistically significant.

It's not so much that the sample size is too small although it is on the small side, it's that there is no mix adjustment so you may well be comparing this month's chalk with next month's camembert.

Doctor_Gloom

Graham_Devon wrote: »

Probably not the best thing to say when you, I, and everyone else is on here too!

I don't set my alarm clock for 3 a.m. to get up and post provocative nonsense on a rather obscure internet forum though.

HAMISH_MCTAVISH

Generali wrote: »

It's not so much that the sample size is too small although it is on the small side, it's that there is no mix adjustment so you may well be comparing this month's chalk with next month's camembert.

But again, that has always been the case. And if there was volatility around the mix changing from month to month, you would expect such volatility to be randomised across all the major Scottish cities.

The fact that they all track quite consistently with each other, IMO, shows that it isn't a problem. Unless you believe it is likely that the mix somehow changes identically in all the major cities from month to month throughout the year? (which seems highly improbable)

Actually, now that I think about it some more, there probably are some seasonal trends with mix that are consistent from year to year, ie, people don't tend to sell big family houses in December and move over Christmas, and mid summer sees a higher number of student flats sold every year than mid winter, but those types of changes would remain consistent to previous years as well.

HAMISH_MCTAVISH

Doctor_Gloom wrote: »

I don't set my alarm clock for 3 a.m. to get up and post provocative nonsense on a rather obscure internet forum though.

Neither do I.;)

Well, not the alarm clock bit anyway. Can't deny posting provocative nonsense on obscure internet forums.:D

Generali

HAMISH_MCTAVISH wrote: »

But again, that has always been the case. And if there was volatility around the mix changing from month to month, you would expect such volatility to be randomised across all the major Scottish cities.

The fact that they all track quite consistently with each other, IMO, shows that it isn't a problem. Unless you believe it is likely that the mix somehow changes identically in all the major cities from month to month throughout the year? (which seems highly improbable)

Actually, now that I think about it some more, there probably are some seasonal trends with mix that are consistent from year to year, ie, people don't tend to sell big family houses in December and move over Christmas, and mid summer sees a higher number of student flats sold every year than mid winter, but those types of changes would remain consistent to previous years as well.

It's interesting. Clearly, the drivers behind property supply and demand in say Aberdeen, Edinburgh and the Orkneys are likely to be very different yet, on the face of it, the prices move in lockstep. It would be interesting to do an analysis to see how changes in price correlate. I think that's one to do at work tomorrow while on double time.

HAMISH_MCTAVISH

Generali wrote: »

It's interesting. Clearly, the drivers behind property supply and demand in say Aberdeen, Edinburgh and the Orkneys are likely to be very different yet, on the face of it, the prices move in lockstep. It would be interesting to do an analysis to see how changes in price correlate.

It is indeed interesting. Of course there are small random variations where, on the face of it, mix change is as good an explanation as any. But the general trend is that of all the major cities matching up with a surprising degree of reliability. Which is why I don't buy into the whole mix or volume led innacuracy arguments.

The sample size across Scotland as a whole is more than big enough to smooth out any inconsistencies. And the match between areas would not be so close if there really were random mix changes skewing the data.

I think that's one to do at work tomorrow while on double time.

:rotfl:

Fair enough. I'd be very interested to see your results if you get around to it.

Generali

I'll need to remind myself of the whole stats thing regarding correlation and then get the data and do the sums. That could easily take a couple of hours.