RDP 2022-02: The Yield and Market Function Effects of the Reserve Bank of Australia's Bond Purchases 1. Introduction

This paper focuses on three Reserve Bank of Australia policy measures that were implemented in the wake of the COVID-19 pandemic and involved purchases of government bonds.^[1] In particular, there were: purchases to support government bond market function; purchases to support a yield target for 3-year Australian Government Securities (AGS); and purchases under the bond purchase program designed to lower yields across the yield curve. Our focus is on the extent to which these measures led to lower government bond yields, and also the effect of bond purchases on various measures of government bond market function.

Purchases under the bond purchase program were clearly delineated from other purchases, while the distinction between purchases to support market function and purchases to support the yield target were less clear. For the purposes of this paper we define purchases to support market function as the bonds purchased at auction between 20 March 2020 and 6 May 2020, which consisted of $51.3 billion in purchases of AGS and securities issued by the states and territories (semi-government securities, or semis) with residual maturity between 1 and 10½ years. While these purchases also supported the yield target, and included the yield target bond, their primary focus was to restore good function in government bond markets, which were under considerable strain at the time. We define purchases to support the yield target as the purchases made at auction between 5 August 2020 and 22 October 2021 that were not made under the bond purchase program; they totalled $29 billion, and all were of either the November 2022, April 2023, or April 2024 AGS.^[2]

In Section 2 we briefly discuss how central bank government bond purchases can work to ease financial conditions, and review some of the international experience of large-scale bond purchases. This includes both purchases in aid of market function, and purchases designed to lower longer-term yields, commonly known as quantitative easing, or QE. In theory, bond purchases could affect financial conditions when the purchases are announced, when the purchases are implemented, or at both announcement and implementation. We assess both the announcement effects of the Reserve Bank's three measures (Section 3) and the implementation effects – that is, any additional effects that occurred as and when purchases were conducted (Section 4).

For central bank bond purchases announced and conducted during periods when markets are well-functioning, where the amount of future purchases is reasonably predictable, and where the central bank announcement is credible, we would expect most of the effect on yields to occur in anticipation of or at the announcement of the measure, and for there to be relatively little additional implementation effect on yields as purchases are actually made. This follows from the observation that government bond market participants are forward-looking, and so will trade on news regarding future central bank purchases, causing prices to adjust as that news is anticipated or announced.^[3]

The above description of the announcement of a credible and predictable purchase amount in a well-functioning market fits well with the Reserve Bank's bond purchase program, which was announced and carried out during a period over which markets were, for the most part, calm. Reflecting this, we find that most of the yield effect of the bond purchase program occurred as market prices adjusted in anticipation of the announcement, and our key results come from an event study covering the period leading up to the announcement. In particular, from September 2020 financial markets were increasingly pricing in the possibility that the Reserve Bank would conduct a bond purchase program focused on longer-term bonds, with these expectations confirmed by the Board announcement on 3 November 2020. We identify the key events that led financial markets to reassess the likelihood that the Reserve Bank would conduct a bond purchase program, and measure the change in government bond yields around these dates. In sum, the yield on the 10-year AGS declined by around 30 basis points. The events include public communications by the Reserve Bank, newspaper articles and market economist reports. In contrast, we find relatively little additional effect on yields from the purchases themselves. For bond purchases to restore market function we find a smaller announcement effect and on balance a larger implementation effect. This is unsurprising as these bond purchases were announced and conducted during a period of market stress, and the amount of purchases was not specified nor predictable at the outset. For the yield target, we find a substantial announcement effect and moderate implementation effects. These findings are discussed in Sections 3 and 4.

In Section 5 we also consider an alternative approach to measuring the effect of the bond purchase program – we construct a counterfactual scenario of what bond yields might have been in the absence of the program. In particular we consider two variants: the first assumes that AGS yields would have moved in line with those of US Treasury bonds; and the second constructs a counterfactual based on the historical relationship between AGS yields and a handful of financial market factors, both domestic and international. The counterfactual scenarios suggest that the bond purchase program reduced yields by somewhere between 20 and 30 basis points, broadly in line with the results from our event study. While longer-term bond yields rose subsequent to the bond purchase program being introduced, this does not imply that the effect of the program was transitory: many other factors influence bond yields, and the evidence suggests that the accumulation of bonds by a central bank holds yields lower than they would have otherwise been over an extended period.^[4]

Bond yields can change due to changing expectations of future short-term interest rates or changes in term premia, and in Section 6 we discuss the results of a model that seeks to decompose observed bond yields into these two components. In Section 7 we assess how bond purchases affected government bond market functioning (helping for purchases in support of market function, but hindering, to some degree, for the yield target and bond purchase program). Section 8 concludes.

Footnotes

See Debelle (2021) for a discussion of all of the Reserve Bank's policy measures during the pandemic. [1]

From its inception on 19 March 2020 until 6 July 2021, the yield target was for the ‘3-year AGS’, specified in terms of the AGS with residual maturity closest to three years, being the April 2023 AGS until 20 October 2020, and the April 2024 AGS thereafter. From 6 July 2021 until it was discontinued on 2 November 2021, the yield target shifted from applying to the 3-year AGS – which changes over time – to applying specifically to the April 2024 AGS. The level of the yield target was changed once, in line with a change in the cash rate target, being around 25 basis points between 19 March 2020 and 3 November 2020, and around 10 basis points from 4 November 2020 until 2 November 2021. The bond purchase program was initiated at the November 2020 Board meeting when $100 billion of purchases were announced, with an 80/20 split between AGS and semis. In February 2021, the Board extended the program by announcing the purchase of an additional $100 billion of bonds after the completion of the initial purchases in mid-April 2021. In July and September 2021, the Board further extended the program, to the effect that purchases would proceed at a pace of $4 billion per week until at least mid-February 2022. The bond purchase program ended on 10 February 2022. Overall, the Reserve Bank purchased $40.3 billion of AGS and $11.1 billion of semis ($51.3 billion in total) to support market function between 20 March 2020 and 6 May 2020; $29 billion of AGS to support the yield target between 5 August 2020 and 22 October 2021; and $223.7 billion of AGS and $57.0 billion of semis ($280.7 billion in total) under the bond purchase program between 5 November 2020 and 10 February 2022. Purchases under each measure are detailed in Statistical Table A3 ‘Monetary Policy Operations’ – see <https://www.rba.gov.au/statistics/tables/xls/a03.xls>. [2]

A thought experiment is useful in affirming this: imagine that a central bank could credibly commit to a set amount of bond purchases, and no more, and then implemented this policy. If yields responded to the flow of these purchases, rather than the announced stock, yields would fall when the central bank commenced purchasing bonds but then spring back as soon as purchases stopped. But such a predictable course of events would present a large arbitrage opportunity to bond traders, who would exploit and therefore remove it. See also D′Amico and King (2013), Arrata and Nguyen (2017) and De Santis and Holm-Hadulla (2020). [3]

Here we are referring to the portfolio rebalancing effect of bond purchases, discussed in Section 2; the liquidity effect is related to the flow of bond purchases and so is likely to exist only while purchases are being conducted, while the signalling effect will be affected by any future signals from the central bank. See, for example, Ihrig et al (2018) and Eser et al (2019) for further discussion. [4]