The Madness of Seasonally Adjusting House Prices
Posted on 8 January 2015 by
Today’s Halifax house price index press release is shamefully misleading. One of its three bullet points says:
“House prices grew by 0.9% between November and December,” which is not what its data shows.
Its data actually shows a fall of 1.1% in December. The increase of 0.9% is a manipulated figure, or “seasonally adjusted” to use the technical term.
Although the table on the last page of its press release admits the figures are seasonally adjusted there is no reference to this earlier, which is unacceptable in this age of transparency.
Nationwide also uses seasonally adjusted figures for its press release headlines but it does make it clear they are seasonally adjusted and it also shows the actual figures in its press release, whereas to find the actual Halifax figures one has to delve into a spread sheet on the Lloyds Bank web site.
It is instructive to compare December’s actual and seasonally adjusted monthly figures from the two lenders as both indices are broadly designed to measure house prices on the same basis and on the same times scale (actually about a week different as Nationwide closes its figures just before the end of the month).
Nationwide reported an actual fall of 0.4% and a seasonally adjusted increase of 0.2%.
Halifax reported an actual fall of 1.1% and a seasonally adjusted increase of 0.9%
Thus for its idea of seasonal adjustment Nationwide increased the actual figure by 0.6%, but Halifax’s idea is an increase of a massive 2%.
However, although a difference in the actual figures from the two lenders is understandable it is much harder to think of a logical reason for such a big difference in the amount of seasonal adjustment, which is as high as 1.4%, a figure larger than the movement in prices in most months!
The lenders like to call the actual figures “not seasonal adjusted,” as if manipulating the figures should be considered normal.
If the seasonal adjustment process was robust it seems reasonable to expect that both lenders would make a similar adjustment. The fact that they use different methods which some months result in very different figures just highlights why the focus should be on actual figures.
After all, although activity in the market does vary throughout the year, buyers and sellers both know that and prices are affected by many other factors as well.
If the logic of seasonally adjusting house price indices was applied to something which really is affected by the seasons perhaps we should seasonally adjust the reporting of temperatures. Reported summer temperatures would then be reduced and winter temperatures increased, resulting in a fairly stable temperature being reported every day!
The views expressed here are those of the author and do not necessarily represent or reflect the views of John Charcol Ltd
The blog postings on this site solely reflect the personal views of the authors and do not necessarily represent the views, positions, strategies or opinions of John Charcol. All comments are made in good faith, and neither Charcol Limited nor Ray Boulger will accept liability for them.