Banks are about to launch Aussie QE4

Coolabah Capital with the note:

Folks are starting to turn their minds to what life looks like after the end of the RBA’s bond purchase program (aka quantitative easing or QE). There’s a fairly robust consensus that Martin Place will look to accelerate its QE3 taper from $4 billion/week currently down to $2 billion/week in February 2022. What few investors appear to understand is that there is a much larger round of long-term, quasi QE4 coming via the banking system, which we currently size at $317 billion (with a range of between $250 billion to $450 billion) starting at the end of June 2021 or $430 billion if we begin at the end of December 2021. That is more bond-buying than all the RBA’s three QE programs combined (ie, $100 billion each for QE1 and QE2, with another $102 billion expected for QE3), which will have profound long-term consequences for asset pricing.

Importantly, the mix of assets banks buy is likely to be very different to what the RBA has acquired: whereas the RBA has split its purchases 80:20 between Commonwealth and State government bonds, the banking system is currently skewed 68:32 in favour of State securities because of the fact they pay a positive spread above the swap rate whereas Commonwealth government bonds don’t.

There are 2888 words left in this subscriber-only article.

Get your first month for $1

The chart below shows the change in the composition of HQLA1 over the next three years. Central bank balances refers to the excess cash held by banks on deposit at the RBA, which slowly declines over time as the TFF is repaid. If for some reason the RBA did a hard-stop of QE in February (ie, did not extend to $2 billion/week for another three months), this would materially bring forward the banking system’s demand for HQLA1.

On a year-by-year basis, our central case projects the following HQLA1 buying profile from the banks with the splits based on their current portfolio mix between Commonwealth and State government bonds:

  • CY2022: $121 billion ($82bn of States and $39bn of Cwealth)
  • CY2023: $137 billion ($93bn of States and $44bn of Cwealth)
  • CY2024: $172 billion ($117bn of States and $55bn of Cwealth)

This actually adds to $430 billion, not $317 billion, because the $317 billion figure starts at June 2021 whereas the $430 billion HQLA1 shortfall starts at December 2021 (between June and December 2021, the banks’ HQLA1 demand temporarily falls as an artefact of QE2 and QE3 given the RBA is creating so much digital money).

It is critical to note that this buying is timed in our quarterly financial model so that banks always hit a 130% LCR within the quarter. Since banks know about the HQLA1 shortfalls years ahead of time, they will almost certainly get ahead of these gaps. So buying in 2022 could well be substantially larger than buying in 2023, for instance. In addition, banks might be worried about the quantity of this buying reducing the margins (spreads) on their HQLA1, which could encourage them to bring forward some buying to get ahead of the curve (and their peers). A final factor that has to be considered is that banks are holding over $360 billion of excess cash earning 0% interest at the RBA. They therefore have a strong commercial imperative to try to spend this money on higher-margin HQLA1 that pays a positive spread above the swap rate (and above the 0% interest they earn at the RBA), which minimises the drag on their net interest margins.

There are several risks to this central case, including:

  • The banks’ balance-sheet growth could be stronger, driving higher net cash outflows, especially if the RBA does not hike until mid to late 2023;
  • The RBA could effect a hard stop to QE3 in February, bringing forward more HQLA1 buying demand into 2022;
  • The composition of bank deposits could shift from retail households to wholesale businesses, which attract higher net cash outflow assumptions and would drive additional HQLA1 demand; and
  • APRA could continue to tighten-up the net cash outflow assumptions that banks apply to different types of liabilities, driving higher NCOs (this has been the direction in which APRA has been heading of late).

*Recall that the CLF was originally created by APRA and the RBA as a substitute for banks holding government bonds on the basis of the belief that there was an insufficient quantum of Commonwealth and State government bonds outstanding to support the banking system’s liquidity requirements. But with the RBA forecasting that there will be $1.6 trillion of these securities outstanding by the end of 2022, there is no longer a need for the CLF.

**This is technically known as excess cash held in the banks’ exchange settlement (ES) accounts with the RBA, or excess ES balances.

Houses and Holes

Comments are hidden for Membership Subscribers only.