The CPI data were released a couple of days ago. There was, inevitably, a lot of commentary around higher petrol prices, although most commentators noted that the Reserve Bank was likely to “look through” what we are seeing, and not adjust monetary policy just because of higher petrol prices. That would, indeed, be consistent with the Bank’s mandate – and practice – over almost thirty years of inflation targeting.
One can have all sorts of debates about what sorts of effects should be “looked through”. We used to have lengthy discussions attempting to distinguish between petrol price effects themselves, indirect effects (eg higher airfares or courier costs directly resulting from higher fuel prices) and second-round effects – the real worry, if changes in oil/petrol prices came to affect the entire inflation process, including medium-term expectations of inflation. Those risks were real, and realised, back in the 1970s oil shocks, and that set the scene for much of the subsequent discussion and precautionary debate.
SNZ only has a CPI ex-petrol series back to 1999. In this chart, I’ve shown the headline CPI inflation rate, the CPI inflation rate ex-petrol, and the Reserve Bank’s preferred core inflation measure, the sectoral factor model.
I’ve highlighted four episodes in which petrol price inflation was much higher than overall CPI inflation, and one (quite recent) when it was much lower.
In the first of those episodes – around 2000 – the surge in petrol prices coincided with quite a lift in core inflation. Bear in mind that the economy was recovering from the brief 1998 recession, and the exchange rate had fallen sharply.
In the second episode – 2004 and 05 – the surge in petrol price inflation coincided with no change in core inflation.
In the next episodes – 2008 and 2010 – the surge in petrol price inflation coincided with a fall in core inflation. In the 2008, the Reserve Bank explicitly recognised some of this at the time, and talked of scope to cut the OCR soon, despite the high headline inflation.
And in the recent episode when petrol price inflation was very low, there was no fall in core inflation – if you look hard enough, it may actually have increased very slightly.
There is talk that, if oil prices persist, headline inflation could get as high as 2.5 per cent before too long. The experience of the last couple of decades suggests that will tell us nothing useful about underlying/core inflation trends, or about the appropriate stance of monetary policy. And the preferred core inflation measure remains below the target midpoint, as it has been for almost a decade now.
Here are a couple of other series worth looking at.
The blue line is a fairly traditional sort of exclusions-based core inflation measure: excluding volatile items (food and fuel) and (administered) government charges (altho not tobacco taxes), and the orange line is non-tradables inflation excluding government charges and cigarette and tobacco taxes (which, you will recall, have been raised relentlessly each year, in a political non-market process). There is no sign in either of these series of underlying inflation moving higher in the last year or two. Core non-tradables inflation of under 2.5 per cent is not consistent, typically, with core (overall) inflation being at 2 per cent.
Having said all that the financial markets appear to have taken a slightly different view of this week’s inflation data. Here is a chart of the breakeven inflation rate from the government bond market – the difference, in this case, between the 10 year conventional bond rate and the 2030 indexed bond (real) rate. I’ve highlighted the change since the inflation data were released.
At 1.4 per cent, the gap is still miles off the 2 per cent target midpoint (or than the comparable numbers in the US), but the latest change does look as if it is worth paying at least a bit of heed to. Perhaps it will dissipate over the next few weeks, but if not it wouldn’t be a cause for concern, but some mild consolation that – after all these years – there was some sign of market implied inflation expectations edging a little closer to target.
What about a longer run of data? We only have a scattering of inflation indexed bonds, in this case one maturing in September 2025 and one maturing in September 2030. The 2030 bond was first introduced five years ago this month. Creating a rough constant maturity 12 year indexed bond series – the 2025 bond had 12 years to run in 2013, and the 2030 one has 12 years to run now – and subtracting the result from the Reserve Bank’s 10 year conventional bond series produces this (rough and ready) chart.
A clear rebound from the lows of 2016, but implied breakeven inflation rates still much lower than they were five years ago.
There still seems to be quite a long way to go for the Reserve Bank to really convince investors that, over the decade ahead, they will do a better job of keeping inflation averaging near target than they have done this year to date.
Continuing to talk down the risks of the next serious recession, and the limitations of policy here and abroad to act decisively to counter such a recession and the likely deflationary risks, is cavalier and irresponsible. It might (seem to) help confidence in the short-run, but if those risks crystallise – and central banks should focus on tail risks in crisis preparedness – the Bank will bear a lot of the responsibility if the economy performs poorly, and inflation ends up so low as to vindicate (and more) the evident lack of confidence among people putting real money on a view about the average future inflation rate.