The Q3’21 CPI print was higher than most people expected, with trimmed mean CPI printing at +0.7%qoq v. mkt +0.5%, taking the through-the-year measure to 2.1%. If you trace the line in the August SOMP, it suggests the RBA was expecting something like 0.35%qoq and a touch over 1.6%yoy … so there’s a lot more inflation than the RBA expected. Particularly at the core level.
While faster-than-expected inflation in Q3’21 will likely pull forward the normalisation of monetary policy, it hasn’t changed terminal rate estimates. The Aussie curve is probably headed for a period of bear flattening as the hikes come forward — but with 2x5x10 near 40bps and 3y2y at 2% it’s looking fully priced in the belly. It’s mostly a monetary policy trade now. And while I do think the RBA will dump YCC and halve QE, I don’t think they are in a hurry to hike.
CPI and monetary policy
Trimmed mean CPI is the one that goes into the policy rule, and that was +0.7%qoq (v. RBA +0.35%qoq); taking the through-the-year measure to 2.1% (v. RBA 1.65%yoy). There’s some nerdy
interesting details about the housing subsidy and its effect on core inflation, but the simple fact is that inflation was faster and broader than the street and the RBA expected in Q3’21. The weighted median print (+0.7%qoq and 2.1%yoy) is probably the best measure of inflation just now.
Given the importance of housing it’s worth noting that ex-housing inflation actually slowed ~20bps in Q3, to +0.6%qoq. While there’s probably some catchup inflation to come over the next few quarters, as the housing subsidy bleeds away, ex-Housing inflation has almost certainly peaked.
The RBA has had trouble with too-low inflation, so I don’t think they’re incentivised to minimise the inflation print. Also, I don’t think the RBA enjoys non-standard policy very much. Higher inflation gives them a way out of both problems.
The November SOMP will almost certainly contain upgrades to their inflation track, and my guess is that the GDP path will be a little better than they forecast in the August SOMP. The RBA started to create some space for an upgrade narrative based on faster-than-expected re-opening in their October minutes, and I expect that they’ll use the opening they created in November. They won’t want to tighten just because inflation is higher. It has to be about growth too.
A change to forward guidance that links low rates to the economy (and not the calendar) seems like the most obvious change. This could come as soon as November.
Gov Lowe made a step in this direction in the Anika speech, where he said “we want to see inflation around the middle of the target range and have reasonable confidence that inflation will not fall below the 2–3 per cent band again” before hiking. The early acceleration of trimmed mean CPI (2.1%yoy), and the still-low level of WPI (1.7%yoy), means that we’ll probably hear more about the necessity of faster wages to make in-target CPI sustainable. This is an old theme for Gov Lowe — his second (2017) Anika speech, was on this topic. I recall walking out of the theatre and summarizing the speech as “Lowe will get moving on rates when wages get moving”. At that time private sector wages were running around 2%yoy.
The maturities of the YCC bonds (April 2023 and April 2024) are so far away that a change to outcomes based forward guidance in November 2021 would probably mean the end of YCC. The RBA has consistently characterised the yield target as a tool to reinforce the Board’s forward guidance on the cash rate. If forward guidance lost the calendar aspect, it wouldn’t make sense to target the bonds.
Deputy Gov Debelle carefully made space for dumping YCC in his May 2021 review of monetary policy during COVID.
In March 2021, the Board agreed that it would not consider removing the yield target completely or changing the target yield of 10 basis points when reviewing the yield target bond later in the year.Deputy Gov Debelle, May 2021, Monetary Policy During COVID
Saying they weren’t considering removing the target at the current review made it clear that this was possible.
Ending YCC is a bigger step than changing forward guidance, and the market probably needs a bit more preparation before they do it. My guess is that the RBA will use the November meeting to change forward guidance to be outcomes based, and announce a more fulsome review of YCC at the December meeting. There’s a risk they announce the review in November and decide it all together in December, but that makes the outlook statement at the November meeting / SOMP messy. Gov Lowe could use his 16 November speech to prepare the market for the change to YCC.
QE has a purpose as they exit YCC. Being able to buy bonds if things get a bit spicy might be useful. Thus, the currently elevated pace of purchases has a purpose (though this is surely an unintended benefit of the decision to extend at 4bn per week). By February 2022, the RBA will have moved to state-based forward guidance, stopped targeting 10bps for April 2023 and April 2024 bonds, and will probably halve the pace of QE to 2bn per week.
In theory they could taper to zero in May 2022 and hike in August 2022, but the Q2 WPI is released on 17 August, so that timeline is probably a bit too tight. November 2022 seems like the earliest real window.
The long end doesn’t care
The market reaction to higher than expected Q3’21 trimmed mean CPI was consistent with recent experience in other markets. The long end doesn’t want to move through the terminal rate, and rates are pretty low, so the curve is flattening as hikes are priced in. This is keeping long end forward-forward yields down. 5y5y OIS was volatile intra-day, though little changed by the close. The 10y5y OIS rate closed a few basis points lower, at 2.27%. Both are clearly below the peaks made in Q1’21. Folks got burnt betting on higher terminal rates in H1’21 and I don’t see much enthusiasm for doing it again.
Assuming the market believes that the RBA’s target will remain 2.5%, 5y5y and 10y5y OIS around 2.25% suggest RSTAR is around -25bps (inflation target *plus* RSTAR = long OIS). The linker curve says much the same thing, with 5y5y real averaging a bit below -25bps over the past three months.
In contrast, there were very large moves at the short end: for example, the Dec’22 bill future sold off 30bps on the day. This was a rare thing, but was mostly a pull-forward of the hikes. The belly took most of the pain, with 2x5x10 shooting higher — but 2x5x10 now seems a bit cheap to me.
If we assume that trimmed mean CPI goes to 2.5% by Q2’22 and remains there, and that the unemployment rate falls to 4% broadly as expected in the August 2021 SOMP, the OIS market now averages about 25bps below my preferred RBA policy rule (the NCR rule from MARTIN, excluding the speed-of-adjustment term on the unemployment rate = the lightblue line; see this post for more).
I don’t think the market will price all the way to the rule, as there’s always the possibility of something going wrong. So this is probably fully priced. Also I doubt the hikes can start before Nov’22, as it’ll take time to get WPI up.
With 1y1y OIS around 1.2%, 5y1y OIS around 2%, and the terminal rate at 2.25%, there’s not much scope for the belly to keep selling off. In 2009, when the RBA hiked the first time (on 7 October) 2x5x10 had already flattened 40bps from the peak of ~90bps to ~50bps. At that time, the market had terminal OIS around 6%.
This time, 2yr yields are likely to rise when the RBA stops targeting the YCC bonds (as YCC bonds cheapen). Unless something global changes, I don’t see terminal rates rising. This should cause 2×5 flatten from the current 75bps.
Higher than expected Q3 CPI, and an economic upgrade narrative due to faster re-opening, will allow the RBA to upgrade away from date-based forward guidance, to state-based forward guidance, in the November SOMP. I expect heavy emphasis on need for faster wages growth to make in-target inflation sustainable. In practice, getting wages up to 3% might make a hike in 2022 hard to achieve. The Q2’22 WPI release is on 17 August, so November is probably the first real hike window.
With forward guidance being tweaked at the November 2021 meeting, the RBA will need to review YCC (probably at the December meeting). The outcome of that review will likely be to allow the market to set the yield of the bonds. Maintaining QE while they exit YCC will allow them to sooth any ructions that occur. QE will then be halved at their February 2022 meeting, to 2bn per week.
After the sharp post-CPI moves, this is mostly priced. The market is reluctant to price through terminal rates. In Australia, I think terminal is about -25bps real and 2.25% nominal. This means the belly of the curve is probably cheap enough: expect 2×5 to flatten and 2x5x10 to decline as the RBA exits non-standard policy and the market brings forward the hikes.
My hunch is that the next big trade in AUD rates will be from the receive side, as disappointment on some front means the RBA can’t hike in 2022. However, the time for that trade is after they’ve tweaked forward guidance and dumped YCC.