US and Australian inflation decomposition
How much is due to demand and how much to supply? Plus, Janet Yellen as accidental inflation hawk; and the RBA's 'perfect' history
The US May CPI came in at 1% over the month, significantly higher than the 0.3% increase in April and above market expectations for a 0.7% rise. The annual rate of inflation rose to 8.6%, the fastest since December 1981.
The US CPI release allows us to update our Cleveland Fed trimmed mean-based model of Australia’s trimmed mean inflation rate for Q2, although we will need the US June CPI to finalise the model’s forecast. Our model implies a 1.5% q/q and 4.7% y/y rise for the trimmed mean compared to 3.7% y/y in Q1. This is already above the RBA’s May Statement on Monetary Policy forecast of 4.5% for the year to June and the 4.6% expected by the end of the year.
David Beckworth suggests a simple decomposition of the US GDP deflator into demand and supply shocks, where excess demand is represented by the Mercatus NGDP gap and supply shocks are measured as a residual after we subtract demand shocks and the inflation target. The results for the US are as follows:
Source: David Beckworth
While only an approximation, the results are broadly in line with what we would expect. In Q2 2020, we get a large negative demand shock and a positive (in the sense of its contribution to inflation) supply shock. The positive demand contribution only comes in from Q3 2021 and gets larger in the following quarters.
We can do the same analysis for Australia’s GDP IPD, where the NGDP gap is based on the methodology I used for my Mercatus paper on nominal income targeting (see the appendix) and for which the inflation target is given as 2.5%, the mid-point of the RBA’s target range. Like CPI inflation, the GDP deflator inflation rate averages 2.5% over the long-run:
The narrative fit looks similar to the US, with a large negative demand and positive supply shock in Q2 2020 with the onset of the pandemic. The demand contribution turns positive in Q1 2021, although gets smaller with the delta variant lockdowns in Q3 2021, before getting larger again. As at Q1 2022, the supply-side contributed around 1.7 percentage points or 21% of the GDP IPD’s annual growth rate.
Monetary policy strategy in both the US and Australia is predicated on reducing excess demand to meet a constrained supply-side. Supply is always constrained at some level in the short-run, it’s just that these short-run constraints are currently even more binding the usual. From a public policy standpoint, it would be better to focus on liberating the supply-side, but that takes time. There are very few quick hits on the supply-side. Abolishing tariffs would be one, but even that has legislative and implementation lags. As a short- to medium- run policy instrument, monetary policy has to take the supply-side as more or less given in responding to excess demand pressures.
Whether it be housing, education, health care or energy, both the level and rate of change in prices is very much a function of supply-side constraints and regulation, which are public policy choices. I don’t know about you, but my biggest expenses are things for which supply is heavily regulated by government. In the absence of these regulatory constraints, rising prices could be expected to focus entrepreneurial attention on alleviating shortages. But entrepreneurs can only do what is permissible.
RBA Board member Ian Harper gave another interview to James Glynn in which he said that the RBA was seeking to pre-empt a wage-price spiral driven by inflation expectations. As we have noted in this space before, this is a somewhat different account of wage-price dynamics to the one the RBA was suggesting previously, in which inflation would only take-off if wages growth accelerated. But there was never much reason to believe that nominal wages growth would grow at a rate consistent with the inflation target while inflation itself was persistently below target. Whereas the RBA previously looked forward to a pick-up in wages growth to drive a higher inflation rate, it now faces the challenge of containing the passthrough of above-target inflation to wages. I think this more or less concedes that it is inflation that determines nominal wage dynamics, not the other way around.
Yellen as accidental monetary policy hawk
Extracts from an authorised biography of Janet Yellen have seen her deny claims she wanted Biden’s fiscal measures cut by around one-third, implying she anticipated the upside inflation risks. In any event, US policymakers now readily concede that macro policy errors contributed to the upside surprise on inflation. Those claims aside, it will be interesting to see how her biographer handles her record as Fed Chair. Yellen had impeccable dovish credentials going into the role, but presided over some of the lowest inflation rates since the 1950s:
I think this speaks to the extent of the policy error in keeping US monetary policy too tight for too long under her (and Bernanke’s) watch. It is also worth noting that US fiscal policy is now turning contractionary, but the large fiscal transfers from public to private saving will see past stimulus have persistent effects on demand as these saving buffers are run-down.
The RBA’s perfect history
Selwyn Cornish, the RBA’s official historian, and John Hawkins had a piece in The Conversation which tried to argue that the RBA’s record in hitting the inflation target ‘has been near perfect and better than its peers,’ based on an average inflation rate since 1994 of 2.5%. But as we have noted here previously, inflation can average 2.5% despite spending a majority of the time outside the target range, as has been the case since 2014. The RBA’s own benchmark is time spent in the band, not the average inflation rate.
Greg Jericho made the observation that the current price level is consistent with where you would have expected it to be back in March 1993 if you had assumed inflation would grow at an average of rate of 2.5%:
This is a more respectable way of making Cornish and Hawkins’ point, which is to say that inflation targeting has successfully anchored the long-run expected price level. That would be an even more defensible benchmark if the RBA were actually doing price level targeting, but that is not the targeting framework it has been pursuing. At business cycle frequencies, the price level has fallen short of its target-consistent path in recent years, as evidenced by the Lowe Gap, which has only recently closed, at least with respect to the price level implied by the bottom of the target range:
An important task for any review of the RBA will be to define benchmarks that are consistent with what the RBA has actually said and done in implementing its inflation targeting regime.
The RBA’s road not taken
Last week, we reviewed Gross and Leigh’s counterfactual simulations of Australian monetary policy, noting that it would be reasonable to simulate the RBA’s pandemic response between March and November 2020 on the same basis given that the zero lower bound was not a binding constraint over this period. Trent Saunders reminded me that he had performed this simulation in early 2021, with the result that the optimal policy rule implied a negative cash rate:
Since Trent shared his optimal control code with Zac and Andrew, Trent’s simulations should be consistent with their results.
ICYMI
A good piece by my former colleague Shaun Ratcliff on the ecological fallacy found in many post-election narratives.
Your tweets: