Should target UK inflation include housing acquisition costs?
A major weakness of the official consumer prices index (CPI) is its omission of owner occupiers’ housing costs. The Office for National Statistics (ONS) publishes a companion measure, CPIH, incorporating such costs using the “rental equivalence” method, which imputes rents on owned homes based on those paid on similar properties in the rented sector. The European Commission, however, has decided to base future changes in its harmonised CPI system on the alternative “net acquisitions” method, which measures the costs of purchasing and maintaining a home. The ONS recently started publishing an owner occupiers' costs series using this method. As explained below, an inflation measure incorporating this series would have provided better signals for policy-makers than CPI or CPIH in recent years.
The first chart shows quarterly data for annual CPI and CPIH inflation since 2006. The latter has been consistently slightly lower because imputed rents are estimated by the ONS to have risen more slowly than the CPI*. The difference has averaged only 0.2 percentage points, however, so monetary policy would probably have been the same if the MPC had focused on CPIH instead of CPI inflation.
The second chart compares CPIH inflation with an alternative housing-inclusive measure incorporating the new ONS owner occupiers’ costs series based on the net acquisitions approach**. The latter series gives a significant weight – currently 26% – to house prices. The alternative measure was higher than CPI / CPIH inflation in 2006-07, lower in late 2008 and 2009, and higher again in 2010 and early 2011. A focus on this measure, in other words, would have encouraged tighter policy in 2006-07, when monetary / credit conditions were excessively loose, while suggesting greater room for easing in late 2008 as the recession was gathering pace – the MPC delayed the launch of QE until March 2009.
The alternative measure would have made a stronger case for withdrawing some policy stimulus in 2010, a change that would have tempered the 2011 inflation spike. The MPC, however, might have chosen to ignore the signal because of concern about the growth impact of the pre-announced January 2011 VAT rise and worsening Eurozone economic conditions.
The alternative measure is giving a less dovish message than CPI / CPIH inflation currently, standing at 1.8% in the third quarter of 2014 (the latest available figure) versus 1.4% in both cases. Indeed, “core” inflation (i.e. excluding energy, food, alcohol and tobacco) on the alternative basis was 2.1% in the third quarter – third chart. Reflecting recent house price strength, owner occupiers’ costs rose by an annual 4.1% in the third quarter using the net acquisitions approach versus only 1.0% on a rental equivalence basis.
*The ONS plans to implement methodological “improvements” that will raise its historical and current estimates of rental inflation. The “National Statistics” status of CPIH was temporarily withdrawn in August 2013 pending these changes.
**The alternative measure was calculated by substituting the net acquisitions series for the rental equivalence series in CPIH while maintaining the same basket weight (currently 16%).
Reader Comments (1)
This is an excellent post by Mr Ward. A CPI series incorporating the new quarterly OOH series calculated by ONS would definitely be a superior target inflation indicator either to the CPI series now used by the Bank of England or the CPIH series based on imputed rents. As Mr. Ward said if such a series had been the target inflation indicator the Bank of England’s monetary policy might have been less dovish in recent years.
The basket share of OOH in the CPIH was 15.6% in 2014, but based on ONS estimates it would be inappropriate to calculate a CPI series including an OOH(NA) component with such a high basket share. The ONS did not publish a basket share for OOH when it published the quarterly OOH series in December. Eurostat only required it to calculate quarterly OOH series, and not an aggregate series that incorporated it so we don’t know what the OOH basket share would be in 2013 or 2014. Basket shares were published for an earlier pilot index. In 2012, based on an earlier ONS pilot series, the OOH(NA) basket share was 5.9%, just over half as large as the 11.6% basket share of OOH in the CPIH aggregate in the same year. This much lower basket share seems somewhat suspect, especially as there is no hint in the methods description by ONS that the OOH(NA) weight might be under-estimated; in fact, it is just the opposite. Similarly one wonders, why, as Mr. Ward notes, the new acquisitions share of the OOH (NA) series is just 26% for 2014, when it is obvious that the same component absolutely dominates the recently published Latvian quarterly OOH(NA) series.
In any case, if one accepts ONS estimates as they are, Mr. Ward has somewhat exaggerated the differences between the CPI inflation rates and those based on a CPI including an OOH(NA) component.