Buried a bit deep in Wednesday’s inflation report from the Bank of England — a little 
from BoE governor Mervyn King to the UK’s Office for National Statistics (ONS).
The central bank is now claiming that CPI inflation may have been “biased down” between 1997 to 2009, because of a 2010 change in the way the ONS interprets clothing and footwear prices. So basically, the BoE says that measured inflation is now running 0.3 per cent higher than otherwise because of the ONS — convenient for a central bank under the kosh for failing to meet its inflation target (for ages).
Anyway, here’s economist Allan Monks with JPMorgan’s take in a Thursday note titled “BoE rolls hand-grenade under the accuracy of ONS inflation estimates”:
We first drew attention to the issue of methodological changes to clothing prices last year, when we were investigating the relative differences in RPI/CPI. Though the changes made by the ONS had clearly increased the gap between these two measures, it was not initially clear whether the CPI had been biased down, or the RPI biased up. Doing so involves making assumptions about how prices would have behaved in the absence of methodological switches – which unfortunately we will never know. But in [Wednesday's] Inflation Report, the BoE took a stab at this by assuming clothing and footwear prices in the CPI over 1997-2009 would have shown behaviour similar to that seen in other clothing prices measured as imports into the UK, and those available to Euro area consumers – in other words, broadly flat. We would make two points about this:
- * There is another reason why consumer clothing and footwear prices could have fallen more quickly than suggested by imported prices. To the extent that retailers substituted between relatively more expensive domestically produced goods, and toward cheaper imported products, this will have created additional downward pressure on inflation (irrespective of what PPI and import inflation did over the same period). In previous analysis, we found some evidence that such price level switching did occur in goods sectors outside of clothing, in which there have been no comparable methodological changes. Remember that while the UK core goods CPI averaged deflation of 2-3% from 1997, PPI and imported inflation ran higher.
- * The Bank of England argues that the impact on RPI inflation has been larger than the CPI. To be clear, these two inflation measures use different averaging methods (arithmetic and geometric respectively) and the gap between the two does tend to increase as the degree of dispersion of prices in the underlying sample increases (which it has done in this instance). This is a statistical fact. But unless one assumes that there has been an asymmetric bias in the increase in the sample of clothing used by the ONS, then the CPI should have been biased down relative to the RPI, as opposed to the RPI being biased up. This could clearly make a difference to estimates of the net effect of methodological changes to the CPI.
We agree that the changes made by the ONS are likely to push up the average level of CPI inflation looking forward (and probably mean that inflation has been biased downward in the past). But given the two points made above, we think the BoE’s estimate of the potential impact of these changes on the CPI is too large. High inflation last year can be better explained by other factors, that both we, and the BoE, have already drawn attention to.
Perhaps the grenade was less explosive than the BoE would have liked?
Related links:
UK inflation was really 0.3pp higher a year ago – Money supply
The Bank of England’s big dilemma - FT Alphaville