As a new official House Price Index is set to launch next week, EYE takes a look at the issue with indices.
House price data can be a bane for agents as buyers and sellers come armed with information that may be incorrect or irrelevant.
Average house prices can vary across indices on the same month, for example Nationwide has values currently at £204,368, while LSL is higher at £298,030.
These disparities are due to the various methodologies and data used, but many buyers or sellers won’t appreciate these differences.
There are two different methodologies.
First, there is the hedonic regression model, used predominantly by Halifax and Nationwide. This assigns a value to different aspects of certain types of property to put a figure on a typical rather than average home.
How are average house prices calculated?
Some also use mixed adjustment, which divide the property market into different cells based on similar characteristics such as size and then takes the average of each to get an overall figure.
A new tool is being launched next week when the Land Registry and ONS data is replaced with a new tool that provides quarterly figures in a UK summary, one for England, one or Scotland and one for Wales.
As well as average property prices it will also provide a breakdown for first-time buyers and new builds as well as by postcode giving a more exact reflection of your area.
The Land Registry does technically provide a bigger picture as all sales are registered with it and it will include cash purchase, unlike indices from lenders like Halifax that are based on their own business.
The trouble with the Land Registry and the ONS
Dating back to 2005, all properties in England and Wales are registered with the Land Registry so there should be a big bank of data.
The trouble is that it can take months for transactions to be processed and set live so the figures may be out of date.
The ONS figures were based on mortgage completions collected by the Council of Mortgage Lenders. This doesn’t include cash buyers and suffers from a time lag/
But the other indices also have their faults. So how big are the differences with the other indices?
Halifax
Current average house price £213,472
Halifax, part of the Lloyds Banking Group, makes up the biggest lender in the UK. Its figures go back to January 1983.
Using a hedonic regression model, the Halifax house price index is based on its transactions taken from mortgage approvals.
This can also be broken down into first-time buyers and home-movers using its own customer data.
This means it may not be totally accurate if some transactions fall through.
It can, however, provide data earlier than other that wait for complete deals based on its approvals giving an idea of where the market is going.
Nationwide
Current average house price: £204,368
Nationwide is the third biggest lender in the UK
It takes its house price information from its own mortgage approvals.
Its data stretches back to 1952 on a quarterly basis and 1991 on a monthly basis and doesn’t include buy-to-let.
Its methodology is similar to Halifax but it also says it varies its definition of the typical house each year.
LSL
Current average price: £298,030
LSL, which owns Your Move and Reeds Rains estate agents, looks at the average price of properties purchased and registered with the Land Registry over the past month.
It takes Land Registry data that is given to its partner Academetrics. The data is then mix adjusted to provide an average price that isn’t affected by the amount of sales each month.
Because it uses the Land Registry it will also include cash sales as well as mortgages
Sales figures can often be revised by the Land Registry later on, but by then they will have already appeared in the LSL data.
Rightmove
Average asking price: £308,151
Rightmove takes the asking prices advertised on its listings and works out the average prices each month
The data is mix adjusted so it isn’t skewed by more listings one moth
The data goes back to 2002
Obviously there will be a difference between the asking price and eventual sale price, which is a drawback, but it gives buyers and sellers an idea of the direction of the market.
You’ve missed one Marc but only because it’s not widely distributed or known about; the Rummage4 data set is being described as “best in class data”. In blind tests against experienced, local, qualified (FRICS) valuations it typically hits +/- >0.5% accuracy ( that is less than £1400 out on the average property)
I have deliberately targeted some seriously knowledgeable, critical, robust and respected industry figures to find weakness in our data, to date not a single blemish has been found in the system other than “the output is not snazzy enough”.
When we first noticed and questioned Land registry about the inconsistencies in their data we were dismissed as fools, we have proven we have a better understanding of the data than Land Registry* and have developed a methodology that is simple , honest and correct. There are no fancy algorithms, certainly no fancy names for overly complicated mathematical calculations, just simple facts that agents, vendors, and investors can understand and data which is specific to property types and location rather than generalised and therefore invalid generalisations.
Each of the indices mentioned above are limited by their data set; not many people are aware that the Land registry index is restricted to just 36% ( close estimate) of annually transacted properties, the data is skewed by what we call ‘type frequency’
Our data isn’t designed to grab headlines or form the basis of a monthly, quarterly, annual report. We have designed a system to help local agents win profitable instructions away from passive intermediary internet listers who rely on looking up inaccurate guestimates of value on the internet. As one agent said, “Robert why didn’t I run into you 30 years ago? this is brilliant”
* the adjustment of ONS numbers followed on from us identifying over 140,000 LR errors/anomalies which were affecting averages by as much as £2,000,000 for a single outcode.
You must be logged in to like or dislike this comments.
Click to login
Don't have an account? Click here to register