Jackanory stories

RHemmings

aderbyshirelad wrote:

Ok RH, point taken, I was being rather extreme. Damn, this always happens when I start reading lypsey's posts as they so rile me.

You are right that there should clearly be some correlation between these figures.

However, how do you (or anyone else - that wasn't a challenge) then explain the recent divergence in the nationwide and halifax figures you quote, as these, I presume, are based on asking prices on a nationwide basis over the same period?

At the moment I don't have an explanation. Which is why I'd like to read the editorial in the Journal of the Royal Statistical Society. Maybe that author does.

RHemmings

nobblyned wrote:

The explanation for the falling LR figures is reasonably simple - they aren't seasonally adjusted.

If you compare Q4 2005 with Q3 2005, or Q4 2004 with Q3 2004 etc, you will see that there is always a fall between the two quarters' numbers.

Because of the non-adjusted numbers the only fair comparison is between Q4 2005 to Q4 2006 which is the year on year number that the LR quotes in the report.

In fact if you look at the numbers you will see that the Q3 - Q4 fall this year was actually less than last year (2.6% vs 2.9%).

But don't let the facts stand in the way of a good 'house price crash' thread!

I think you miss the point of this thread. What is being criticised is not whether house prices went up or down, but why different measures of house prices have diverged. And why the high ones are being reported while those showing drops are not.

nobblyned

RHemmings wrote:

I think you miss the point of this thread. What is being criticised is not whether house prices went up or down, but why different measures of house prices have diverged. And why the high ones are being reported while those showing drops are not.

Because the other indices are seasonally adjusted, so the LR index stands out as different.

Doozergirl

RHemmings wrote:

I think you miss the point of this thread. What is being criticised is not whether house prices went up or down, but why different measures of house prices have diverged. And why the high ones are being reported while those showing drops are not.

I'm with you on the other stuff, but this is surely a leading statement...

lypsey wrote:

The crash is surely begining

RHemmings

nobblyned wrote:

Because the other indices are seasonally adjusted, so the LR index stands out as different.

Rightmove does not seasonally adjust their figures.

nobblyned

RHemmings wrote:

Rightmove does not seasonally adjust their figures.

Very true, but are the only index based on (initial) asking prices in which one wouldn't expect the seasonality to be so pronounced.

AFAIK the seasonality adjustment is supposed to correct for the discount under asking price acheived at various times of the year due to the supply demand balance (according to Nationwide anyway), and so wouldn't be so necessary in an asking price index.

Halifax, Nationwide and Hometrack all seasonally adjust.

Alan_M_2

This is like comparing apples with oranges and Kiwi fruit.

Nationwide and Halifax's data is and can only be based on mortgage applications, nowhere in their data (whether it be in an upward or downward direction) is any information included about property that isn't selling, repossessed properties, auctioned properties, people having handed the keys back and left them to it. The very nature of the index means it will always be drawn towards the cream of property, areas where sales are booming and property sought after.

On the other hand areas finding it difficult to generate sales and less sought after neighbourhoods will naturally be excluded by their very lack of activity.

The Land Registry gives you a more rounded view of events, but even then, if sales aren't happening, statistics aren't being generated.

A crash or downturn can only be called as such historically, in the meantime the rest of us have to get by and make decisions as best we can on anecdotal information.

aderbyshirelad

Alan_M wrote:

This is like comparing apples with oranges and Kiwi fruit.

Nationwide and Halifax's data is and can only be based on mortgage applications, nowhere in their data (whether it be in an upward or downward direction) is any information included about property that isn't selling, repossessed properties, auctioned properties, people having handed the keys back and left them to it. The very nature of the index means it will always be drawn towards the cream of property, areas where sales are booming and property sought after.

On the other hand areas finding it difficult to generate sales and less sought after neighbourhoods will naturally be excluded by their very lack of activity.

The Land Registry gives you a more rounded view of events, but even then, if sales aren't happening, statistics aren't being generated.

A crash or downturn can only be called as such historically, in the meantime the rest of us have to get by and make decisions as best we can on anecdotal information.

Beautifully put Alan M. Thanks.

nobblyned

Alan_M wrote:

This is like comparing apples with oranges and Kiwi fruit.

Nationwide and Halifax's data is and can only be based on mortgage applications, nowhere in their data (whether it be in an upward or downward direction) is any information included about property that isn't selling, repossessed properties, auctioned properties, people having handed the keys back and left them to it. The very nature of the index means it will always be drawn towards the cream of property, areas where sales are booming and property sought after.

On the other hand areas finding it difficult to generate sales and less sought after neighbourhoods will naturally be excluded by their very lack of activity.

The Land Registry gives you a more rounded view of events, but even then, if sales aren't happening, statistics aren't being generated.

A crash or downturn can only be called as such historically, in the meantime the rest of us have to get by and make decisions as best we can on anecdotal information.

Agreed! Well put.

I guess the crux of my message was it's a little early to predict a crash based on a declining Q4 LR number when this has happended in London every year for the last few years...

zar

RHemmings wrote:

There is an editorial on "House Price Indicies" in the January issue of the Journal of the Royal Statistical Society, Series A. I don't have immediate access to it. Does anyone?

Thanks for that - I'm going to give a seminar to my fellow stats students on Tuesday on indices and I was going to use house prices as an example so that's useful for me! :beer:

I won't paste the whole thing here as I don't want to violate copyright. The author is Simon Briscoe, from the financial times. He says, "I am biased but I feel that the FT house price index represents a material step on the road
to the perfect index—i.e. one that includes only completed transactions at the right price, is
timely and is adjusted for both seasonal variations and changes in the mix of housing bought
and sold."

A quick summary and choice quotes:

"we are left with the conundrum why, given the breadth and depth of interest in the
market, do we not have decent data?"

3 difficulties:
1)No two house are exactly the same
2)Until recently only those involved in the sale knew the price. Some prices include fixtures and fittings, others don't
3)house are sold infrequently, with 7% of the housing stock sold each year in the 1990s => house on average changes hands every 14 years

Also what is the best way to compile an index from the raw data - mix-adjusting etc. (this is what I'm "professionally" interested in)

"Indices are sometimes
produced as a media event to gain publicity rather than to inform serious analysis."

<cough> rightmove! </cough>

"But does this flood of indices add to the confusion about the value of property or give people
a sufficiently clear message about house price trends? And should the Government have done
more to deliver good timely data?...A joined up
Government, that we hear so much about, should have been able to get the DCLG, Land Registry,
Office for National Statistics, Treasury and the professionals in the conveyancing business
to co-operate more rapidly and effectively."

"Looking forward,
the ‘ultimate target’ of creating a house price index for Britain might be only 5 years away. The Land Registry’s long-term plans for electronic conveyancing will improve the timeliness of the
data and the proposed development of a national property databank will allow it to be mix
adjusted."

And finally,

"But even a perfect index is no guarantee that the media will use it and that it will be accurately
reported...A Bank of England
study concluded that
‘monthly changes in house price inflation do not individually contain much information about whether
the medium-term inflation rate is rising or falling’.
Persuading users to interpret the ‘perfect’ index correctly might take even longer than creating
the numbers in the first place!"