More BBC economic spin

Two pieces of egregious spin on economics from the BBC:

Economy shrinks less than thought

On last night's news this was spun as indicating that the recession may soon be over. The upward revision was from a 0.8% quarterly decline to a 0.7% decline, so the difference might have been miniscule because of rounding , but of course the BBC did not go into that level of detail. They simply extrapolated from that small detail that our economic woes are over.

Back to reality. Let's put it in perspective. Inflation is over 0.4% per quarter, so a 0.7% decline in GDP is a morge than 1.1% drop in real terms. In how many quarters since 1990 have we seen a 1.1% real terms increase in GDP, which would be more than 1.5% increase in nominal GDP?

According to the BBC: none.

House price rise hits 5-year high

The is little nugget comes from the latest monthly report from the Land Registry that house prices in England and Wales rose by 1.7% in July compared with June, the biggest monthly leap in value since July 2004.

At first blush this might look encouraging, although it is only monthly figures, and based on substantially lower sales figures and thus a more volatile market where house prices are still 11.7% lower than a year ago. The methodology looks sound because the Land Registry compares the value of houses actually sold against known previous sales to compute an index.

The House Price Index is based on values of houses that were actually bought and sold in the relevant month compared with last known sales of the same property. The Land Registry database contains 15 million entries of which about 5 million are matched pairs giving 2.5 million price ratios between sales which can give a reasonable input to a regression analysis, but there are a number of problems with the method.

First of all the data excludes forced house sales from divorce settlements and reposessions. The latter might be sensible in a fair market, but it creates a material upward bias in the Land Registry report when repossessions are more frequent.

It also excludes new builds by definition, which will constitute a large part of the market at any time and where prices may become very depressed in a recession as developers try to clear their books to repay lenders, which would tend to bias the HPI higher than the real average price when developers are having a bad time.

Third, the index assumes that it is comparing a homogeneous distribution of house sales, but if we look at the price points where sales are ocurring we see that this is not the case. Sales at the upper end of the market are about a half of where they were a year ago. On the other hand sales of properties below £50,000 were very much in demand with numbers of sales up by 50% and no doubt higher prices being paid.

A 1% increase in the value of low cost housing stock would be given equal weight in the computation as a 1% increase in the price of high cost housing stock. It looks suspiciously as though the Land Registry methodology has pushed up the House Price Index simply because of a substantial increase in demand for properties right at the bottom end of the market, notably in Wales and the North of England, whilst the price of higher priced housing stock where the number of sales has actually decreased could well be static or declining, but because fewer of those houses have sold they have less impact on the results.

Not that the BBC would ever want to go into that level of analysis.