2.3 Main Real-Estate Indices Worldwide
2.3.6 HALIFAX HOUSE PRICE INDEX
The Halifax House Price Index (HHPI) is constructed from the monthly val- ues recovered from the mortgage database of the Halifax which is part of Lloyds Banking Group. Typically the HHPI covers an average of 15,000 house purchases per month. The index is constructed using the hedonic approach whereby a “standardized” house price is derived and seasonal adjustments are applied. The seasonal factors are updated monthly. In a similar fashion, a quarterly index series is also produced. The methodology for establishing this index is based on the research carried out by Fleming and Nellis (1981).
The main idea in constructing this hedonic index was not to use averages but to take into account the property characteristics, taking advantage of the large pool of mortgage data managed by Halifax. In a nutshell the price of a house is estimated using a regression model that includes both qualitative and quantitative variables that include location (region), type of property (house detached, semi-detached or terraced bungalow, flat), age of property, tenure (freehold, leasehold, feudal), number of rooms (habitable rooms, bedrooms, living-rooms, bathrooms), number of separate toilets, central heating (none, full, partial), number of garages and garage spaces, garden, land area if greater than one acre, road charge liability.
The regression model used for calculation is specified as
ln(Pi)=β0+β1X1i+β2X2i+. . .+βjiXji+εi (2.4) The average house price within a particular period is influenced by the number of observations on each characteristic in the same period. Therefore, the house price could rise simply due to an increase to the number of houses sold relative to total house number. One way to solve this problem is to keep the mix of
0 100 200 300 400 500 600 700
1983Q1 1983Q4 1984Q3 1985Q2 1986Q1 1986Q4 1987Q3 1988Q2 1989Q1 1989Q4 1990Q3 1991Q2 1992Q1 1992Q4 1993Q3 1994Q2 1995Q1 1995Q4 1996Q3 1997Q2 1998Q1 1998Q4 1999Q3 2000Q2 2001Q1 2001Q4 2002Q3 2003Q2 2004Q1 2004Q4 2005Q3 2006Q2 2007Q1 2007Q4 2008Q3 2009Q2 2010Q1 2010Q4 2011Q3 2012Q2 2013Q1 2013Q4 2014Q3
Halifax UK Index
Figure 2.10. The Halifax House Price Index, quarterly between Q1 1983 and Q3 2014.
Notes: The historical price of Halifax All Property index, quarterly between Q1 1983 and Q3 2014, all houses, all buyers, seasonally adjusted.
house characteristics constant across time, by choosing a representative set of weights based on the number of observations of each characteristic for a chosen time period. The set of weights are fixed to the base period of 1983. The numerical value of this is derived by comparing the current mix-adjusted price to the base period weighted average price.
The evolution of the quarterly Halifax All Property SA is shown in Figure 2.10. The monthly version of the same index is illustrated in Figure 2.11.
As can be seen by comparing the two figures, there is very little difference between the two indices, confirming the known empirical characteristics of the housing markets regarding price stickiness, inertia, strong short-term positive autocorrelation, and so on. One possible explanation is that it takes on average three calendar months to complete a property sale. Therefore, the speed of transactions in the housing market is very slow and structural changes to property prices may be observed only after periods longer than three months.
Another possible explanation is based on the methodology used to construct the index. The hedonic approach is essentially a regression type approach. The variables used to construct the quarterly and the monthly indices, respectively, are the same and they are likely to have the same or very close values for all months within a quarter. This may lead to roughly the same path evolution of HHPI.
As with the IPD indices, the graphs illustrated in Figure 2.11 indicate that there is positive autocorrelation of returns up to about three years and then negative autocorrelation after five years. Although the Halifax UK SA monthly index levels show a similar evolution to all other real-estate indices,
–100 0 100 200 300 400 500 600 700
Halifax UK SA Index Feb 83 Feb 86 Feb 89 Feb 92 Feb 95 Feb 98 Feb 01 Feb 04 Feb 07 Feb 10 Feb 13
−0.04
−0.03
−0.02
−0.01 0 0.01 0.02 0.03 0.04 0.05
Returns Feb 83 Feb 86 Feb 89 Feb 92 Feb 95 Feb 98 Feb 01 Feb 04 Feb 07 Feb 10 Feb 13
0 20 40 60
−0.5 0 0.5 1
lag
Sample Autocorrelation
0 20 40 60
−0.5 0 0.5 1
lag
Sample Partial Autocorrelation
Figure 2.11. The Halifax House Price Index monthly between January 1983 and December 2014.
Notes: The historical price of Halifax All Property index, quarterly between Q1 1983 and Q3 2014, all houses, all buyers, seasonally adjusted. The upper graphs show the index levels and returns for monthly series between January 1983 and December 2014, all houses, all buyers, seasonally adjusted. The lower graphs show the ACF and PACF plots.
the percentage returns series seems to be a lot more volatile than the other property indices, but otherwise quite stationary. The Ljung-Box Q-test strongly rejects the absence of no autocorrelation at 6, 12, 30, and 96 monthly lags.
Hence, this UK housing price index has similar stylized features as the US commercial and housing price indices. There are also 12 regional indices as well as a subindex dedicated to first time buyers, all calculated with the same hedonic methodology.