How tight can semi-spreads go post QE?

Authored by Dr James Yang, Senior Data Scientist, Christopher Joye, Chief Investment Officer, Matthew Coleman, Senior Consultant, Coolabah Capital Investments.

This brief note seeks to provide some evidence on how tight state government (semi) bond spreads in Australia could go in an environment in which the RBA is seeking to drive the jobless rate back to its full-employment (NAIRU) level. Pre-covid-19 estimates of the NAIRU published by the RBA were around 4% to 4.5% (see Ellis (2019)) while Treasury has recently parameterised the NAIRU at 4.75% to 5.0%. 

Coolabah’s Chief Macro Credit Strategist, Keiran Davies, published a research report in October on the likely contours of the RBA’s next monetary policy stimulus program, which you can read here. Kieran also provides a sensitivity table that summarises the amount of QE required by the RBA in order to get the jobless rate down to different NAIRU estimates. He writes:

Given that a key input into this [QE] calculation is the NAIRU, we explored the impact of alternative scenarios on the estimate of QE. Assuming that the NAIRU is unchanged from the RBA’s pre-virus estimate of 4.5% – or, more plausibly, it may have been lower than the RBA’s figure – the RBA would have to do about twice as much QE as we have estimated and/or call on additional government support. Alternatively, if the NAIRU is closer to 5.5%, the analysis suggests that the RBA can broadly rely on existing policy settings, including a full drawdown of the Term Funding Facility, to eventually achieve full employment. 

At the RBA’s pre-COVID-19 estimated range for the NAIRU, the implied QE requirement is very large at between $272bn to $404bn. Note, however, that Kieran’s forecast range is lower at $115bn to $180bn given he has adopted a higher 5% point estimate for the NAIRU.

In the analysis that follows we provides an historical comparison between spreads on bonds issued by KFW, a German state-owned development bank, and the yield on German (government) bunds. We also examine the spreads on NSW Treasury Corporation (NSWTC) bonds relative to (1) Australian Commonwealth government bond (ACGB) yields (known as the “G-spread”) and (2) the swap rate on a matched maturity basis for 5 year and 10 year maturities (called the “I-spread”). 

What is evident from the data is that the ECB’s QE initiative via the public sector purchase programme (PSPP) in March 2015 initially drove the 5- and 10-year KFW-Bund spread into negative territory. 

There is a significant risk that further QE from the RBA, in addition to continued demand from bank balance sheets fuelled by increases in the size of the RBA’s Term Funding Facility, the winding down of APRA’s $223bn Committed Liquidity Facility, and excess ES (cash) balances held at the RBA could accelerate the compression of the 5- to 10-year semi vs ACGB spread towards zero.

In fact, it’s possible semis could trade on negative G-spreads given I-spreads (or spreads above swap) remain at historically very attractive levels for bank balance-sheet buyers and the fact that the supply of semis is going to be relatively constrained compared to the supply of Commonwealth government bonds.

Please click here to download the full report.