In the run-up to the 2016 renewal of its inflation-targeting agreement with the Government of Canada, the Bank of Canada (BoC) has been openly musing about whether it should continue the practice of identifying one preeminent measure of inflation as its operational guide, and whether that measure should continue to be the core consumer price index (CPIX). In 2013, the BoC published a working paper proposing an alternative measure of inflation it called the common component of CPI (CCCPI).1 In a speech delivered last November, its Deputy Governor suggested the CCCPI will be the Bank’s preferred operating guide going forward.2
Both the BoC’s research and its Deputy Governor’s statements emphasise the common component’s supposed advantages relative to CPIX, but not its limitations. However, comparing the two measures and taking a closer look at the CCCPI, the new measure’s supposed advantages are less than obvious, and its limitations seem significant.
Contrasting CPIX, CCCPI
Both the CPIX and CCCPI are based on the all-items CPI basket. While somewhat beyond the scope of this discussion, it’s worth noting that if the BoC wished to target underlying inflation, its measures should be based on the CPI excluding the effect of indirect taxes (CPIXT).
Whereas the CPIX explicitly excludes eight CPI basket items, the CCCPI supposedly excludes only the unique price movement of each item that does not strongly correlate with general price movements over time (as determined by a ‘common component’ of correlated price comovements). While no items are explicitly excluded from the CCCPI, the statistical method used effectively excludes nearly one third of CPI basket items.
Significantly, several items excluded from the CPIX – namely mortgage interest, tobacco products, fruits and nuts – are highly correlated with the CCCPI, meaning they have a greater impact on changes to the latter. A 2001 BoC research paper notes, “Since the Bank’s policy instrument—the target overnight rate of interest—has a very direct effect on mortgage rates at shorter maturities, this gives a misleading signal of the future trend in inflation.”3,4 As such, it’s unclear how the CCCPI could “provide a clearer signal of the underlying trend in inflation” than CPIX, as the 2013 BoC working paper suggests.
Given these difference, it should come as no surprise that the two measures’ trends tend to occasionally diverge, sometimes rather markedly.
CCCPI and ‘sector-specific price movements’
The BoC CCCPI working paper repeatedly singled out one period of divergence “in 2002–03, when the common component remained close to 2.0 per cent, while both total CPI and CPIX inflation increased significantly following the run-up in automobile insurance premiums.” The anecdote presumably highlights one of the two key advantages the BoC posits for the CCCPI: That it “is not affected by sector-specific price movements that can distort the signal in many other measures of underlying inflation” like the CPIX.
However, the same paper glances over the divergence between the CPIX and CCCPI during a more critical point: the 2008-2009 recession. The BoC went so far as to avoid directly comparing the two measures during the recession.
In fact, the entire working paper contains just one instance of the R word: “Following the recent recession…” Curiously, the Deputy Governor’s speech extolling the virtues of the CCCPI is titled, “Inflation targeting in the post-crisis era”.
It seems the BoC would rather not draw attention to the CCCPI trend during the previous recession.
That’s because the CCCPI showed a remarkable rise during that period, peaking at 2.8 percent – near the upper bound of the BoC inflation target range – from December 2008 to April 2009 and remaining at or above the BoC’s target rate until January 2010, after the recession had officially ended.5
By contrast, the CPIX inflation trend showed a notable decline over the same period, dropping to 2.0 percent by March 2009 and continuing well below the BoC’s target rate until it reached 0.9 percent in February 2011, well into the shaky recovery that followed the recession.
It would appear that the CPIX trend during the recession intuitively made more sense, especially so if the Bank is supposedly conducting monetary policy with a view to addressing the so-called output gap.
CCCPI and the output gap
The economic theory behind it (the Phillips curve) aside, accurately forecasting inflation based on the concept of an output gap has proven less than reliable. In part that has to do with the remarkably imprecise science of estimating an economy’s true potential output at any given point in time based on reported labour market conditions.
Nevertheless, the BoC believes it’s found the magic bullet in the CCCPI, noting in its paper: “Contrary to previous findings for Canada, we do find evidence of an inflation-output relationship.”
However, the supposed CCCPI-output gap relationship does not appear as strong as suggested. The BoC concedes that “the estimated coefficient on the output gap is quite small and therefore consistent with the broader finding of a flattening of the Phillips curve under Canada’s inflation-targeting regime.” So apparently the CCCPI-output gap relationship is relatively weaker because the inflation-targeting regime based on the now disfavoured CPIX seemed to have worked quite well, despite itself.
Ad hoc monetary policy
Another possible explanation for the sudden interest in alternative and/or multiple inflation-targeting measures is to take the burden off the BoC having to justify its recent rate decisions. Since adopting fixed announcement dates (November 2000), the Bank had not undertaken cuts to its key rate when the CPIX remained above its inflation target rate for 11 consecutive months or more… until this past year: The CPIX has remained above the inflation target since August 2014, during which time the Bank has twice cut its lending rate.
Conveniently, the Bank can now pivot to the CCCPI, which has remained well below the Bank’s inflation target rate over the same period, to justify its otherwise unjustifiable rate cuts.
Hopefully whoever ends up as Finance Minister next year sees fit to demand clarification from the BoC as to what its new preferred inflation indicator actually measures, whether that measure is appropriate, necessary or even fit for use as an inflation targeting tool and how the Bank intends to make consistent monetary policy decisions going forward. Because as things stand, the BoC appears to have adopted an ad hoc monetary policy, which is really no policy at all.
1. The Common Component of CPI: An Alternative Measure of Underlying Inflation for Canada, Mikael Khan, Louis Morel, Patrick Sabourin, Bank of Canada, October 2013
2. Inflation Targeting in the Post-Crisis Era, Agathe Côté, speech delivered to the Calgary CFA Society, Bank of Canada, November 18, 2014
3. A New Measure of Core Inflation, Tiff Macklem, Bank of Canada, Autumn 2001
4. In fact, the BoC’s rate cuts have directly contributed to the decline in the CPI weight for mortgage interest cost; up until the 2005 CPI basket update, it was the most heavily weighted CPI basket item, since superseded by rent and homeowners’ replacement cost.
5. Notably, the CCCPI remained above the BoC target rate for three straight years, from January 2007 through December 2009; over the same period, the CPIX remained above the target rate for only 10 months, all but two of which occurred in 2007. The BoC has suggested adoption of the CCCPI would require a concurrent rise in its target rate and range.