Review of the Bond Purchase Program

Overview

In November 2020, the Reserve Bank Board introduced a bond purchase program (BPP) as part of a second package of monetary policy measures designed to lower the structure of interest rates in Australia. This package was implemented to provide additional support to job creation and the recovery of the Australian economy from the COVID-19 pandemic, thereby providing extra insurance against the ongoing risk of very bad economic outcomes. The BPP ultimately involved the purchase of a total of $281 billion of Australian, state and territory government bonds between November 2020 and February 2022.

This review examines the experience with the BPP and draws some key lessons.

The key points are:

• The BPP, together with the other monetary policy measures put in place during the COVID-19 pandemic, contributed to the strong recovery of the Australian economy from the pandemic, with unemployment now at its lowest rate in almost 50 years. Together, these measures further lowered the whole structure of interest rates in Australia, and supported confidence in the economy in the face of serious downside risks. Given the reinforcing nature of the policy measures and the fact that the BPP was a new policy tool that works through different channels, it is difficult to isolate the specific effect of the BPP on the economy. A key benefit of the package of policy measures was to provide insurance against the serious downside risks the economy was facing during the pandemic.
• The BPP is estimated to have lowered government bond yields, but by somewhat less than generally suggested by international studies for a program of this size. This reflects the fact that the BPP was introduced at a time when other policy measures were already providing strong signals about future policy and bond market conditions were no longer strained so liquidity premia were low. The Bank's earlier bond purchases helped to restore good function in government bond markets, which were under considerable strain at the onset of the pandemic. These purchases were conducted as required to support market function (with no specified amount or timing) and were not part of the BPP considered in this review.
• Consistent with international experience, the effect of the BPP on government bond yields occurred around the time of the initial announcement of the BPP, reflecting the forward-looking nature of financial market participants. The reduction in government bond yields contributed to lower borrowing costs across the economy and a lower exchange rate than otherwise, at a time when the cash rate was judged to be at its effective lower bound. International studies highlight the importance of central banks acting decisively when policy rates are near the effective lower bound.
• The BPP has affected the public sector balance sheet in several ways. For the RBA, the purchased bonds pay a fixed return, while the interest paid on the Exchange Settlement (ES) balances created to pay for the bonds varies with monetary policy settings. As interest rates increase there is a financial cost to the RBA from this. The ultimate cost will be known only once the last of the purchased bonds matures in 2033, with various scenarios presented in this review. Under most scenarios, the Bank will not be in a position to pay dividends to the government for a number of years.
• For the broader public sector balance sheet there are offsetting financial benefits. The stronger growth and inflation as a result of the package of monetary policy measures will add to the seigniorage the Bank receives from issuing banknotes. The reduction in yields also lowered the cost of government debt issuance, and stronger economic activity than otherwise has increased tax revenues and reduced government support payments. In addition, the debt-to-GDP ratio is lower than otherwise as a result of the boost to nominal economic activity. These benefits to government finances are material, although they are difficult to quantify.
• The design and implementation of the BPP worked broadly as intended, without materially affecting market functioning. The flexibility built into the BPP allowed for adjustments in purchases in response to market conditions, thereby reducing the risk of compromising market functioning. Moreover, the fairly short timeframes of each of the various stages of the BPP to which the Board committed facilitated a timely and relatively smooth end to purchases under the BPP. These are important features to consider in the design of any future bond purchase program.
• The Board's communication of the considerations around adjustments to the BPP prior to the decision points, which were themselves before the implementation dates, helped to avoid an unhelpful, disruptive adjustment as bond purchases were wound down.
• As noted in the Review of the Yield Target (RBA 2022), in light of the experience of policies adopted in response to the pandemic, the Board has agreed to strengthen the way it considers a wide range of scenarios when making monetary policy decisions in future, especially where they involve unconventional policy measures. In relation to any future BPP, this scenario analysis would include the benefits and potential costs of the policy, recognising the difficulties in measuring these outcomes.
• The Board remains of the view that it is appropriate to consider use of unconventional monetary policy tools only in extreme circumstances, when the usual monetary policy tool – the cash rate target – has been employed to the full extent possible. As noted in the Review of the Yield Target, if the use of unconventional monetary policies was being considered in future, the Board has not ruled out the use of either a BPP or a yield target. Compared with a yield target, a BPP would provide more flexibility to respond to evolving economic circumstances. However, a BPP could entail larger financial costs than a yield target, which would need to be carefully considered in the circumstances.

The review is structured as follows. Section 1 provides an overview of the BPP and the main decision points. Section 2 explores the deliberations behind the various decisions in greater detail. Section 3 reviews the effect of the BPP on government bond yields and the exchange rate, and the broader implications of this for the economy. Section 4 considers the financial implications of the BPP for both the Bank and the broader government balance sheet. Section 5 assesses the BPP experience and draws some lessons.

1. Overview of the bond purchase program

On 3 November 2020, the Reserve Bank Board introduced a BPP as part of the second package of monetary policy measures implemented by the Bank in response to the COVID-19 pandemic. The BPP initially involved the purchase of $100 billion (face value) of government bonds of maturities of around 5–10 years. Purchases were made over six months at a pace of $5 billion per week, split 80/20 across nominal (fixed-rate coupon) bonds issued by the Australian Government, and by the states and territories.

The BPP involved purchasing a given quantity of government bonds in order to further lower yields at the 5–10 year part of the yield curve. This quantity target further out the yield curve was introduced to complement the price-based, three-year yield target introduced in March 2020, along with the longstanding overnight cash rate target, which forms the anchor point for the risk-free term structure. Government bonds are the benchmark fixed-income securities in Australia, underpinning the risk-free interest rate embedded in the prices for most other assets. In combination, this second package of monetary policy measures was designed to further lower the whole structure of interest rates in Australia and thereby support the economy by lowering borrowing costs, continuing to encourage the supply of credit (especially to businesses) and delivering a lower exchange rate than otherwise.

The BPP was extended several times. In February 2021, the program was extended to incorporate the purchase of a further $100 billion of government bonds from April to September 2021, at the same pace of $5 billion per week. In July 2021, the Board announced that purchases would continue beyond September 2021 until at least November 2021, but at a lower rate of $4 billion per week. In September 2021, the Board announced purchases would continue at the rate of $4 billion per week until at least mid-February 2022.

In February 2022, the Board decided to cease purchases under the BPP. In May 2022, the Board announced that it did not plan to reinvest the Bank's government bond holdings as they matured, nor did it have plans to sell its holdings.

Upon its completion, a total of $281 billion of government bonds had been purchased under the BPP – $224 billion issued by the Australian Government and $57 billion issued by the state and territory governments.

A summary of the key decision points and actions is provided in the table below and discussed in detail in the following section.

2. The Bank's deliberations

November 2020: Bond purchase program announced

In late 2020, although the Australian economy had begun to recover from the sharp economic downturn associated with the pandemic, there was still significant uncertainty around the economic outlook. In particular, the availability and efficacy of vaccines was not assured. Moreover, the economy remained a long way from the Bank's goals of full employment and inflation within the target range, and the recovery was expected to be extended and bumpy. It was a similar situation in the global economy; renewed virus outbreaks, in Europe in particular, posed downside risks to the economic outlook. Central banks' monetary policy settings remained extremely accommodative, with a number indicating that new stimulus policies were under consideration.

As shown in the forecasts published in the November 2020 Statement on Monetary Policy, it was expected to take some time for output to reach its pre-pandemic level, and an extended period of high unemployment was in prospect. Given the substantial spare capacity expected in the economy over this time, wages growth and inflation were projected to remain subdued. This outlook was broadly in line with the average market forecasts for 2020 and 2021.

In this environment, addressing the high rate of unemployment was viewed as a national priority by the Australian Government and the Reserve Bank Board. Following the initial fiscal and monetary support measures implemented in response to the pandemic, the Bank had extended and expanded the TFF in September 2020 and the Australian Government budget released in October 2020 provided further support to jobs and economic growth. Further fiscal support was expected in the forthcoming state budgets. While the Board considered that additional stimulus would need to come mostly from fiscal policy, it judged that the risks were such that further monetary response was appropriate to provide insurance against very bad economic outcomes. The Board was also of the view that having the various arms of policy all taking steps in the same direction would deliver a greater impact than the sum of the individual parts. International studies also highlight the importance of acting decisively when policy rates are near the effective lower bound.^[1]

Accordingly, in November 2020 the Board introduced a further package of monetary policy measures, including cuts to the cash rate target, the three-year yield target and the TFF rate, as well as a BPP at the longer end of the yield curve.

As noted in the Review of the Yield Target, the Board had decided to include a yield target rather than a BPP as part of its initial package of monetary policy measures in March 2020 because a yield target was seen as a more direct way of achieving the objective of low funding costs in Australia. It also reinforced the forward guidance that the cash rate was expected to remain at a very low level for some years.^[2] The three-year yield target, rather than bond purchases across the yield curve, was judged to be more appropriate for Australia given the prominence of intermediated lending at mostly variable rates or on fairly short fixed terms.

However, by November 2020, the Board judged that other monetary policy measures had already been employed to the extent it made sense to do so, and so implementing a BPP was appropriate at that time. The cash rate target, yield target and interest rate on the TFF were reduced to very low levels, and negative interest rates were not considered to be a sensible policy option given the experience in other economies.^[3] Even though the BPP's effect on yields was uncertain and it would affect yields further out the yield curve – which were less important for private-sector funding costs in Australia than in many other countries where BPPs had been employed – the additional stimulus was still judged to be helpful in the circumstances.

An important consideration in introducing the BPP was that almost all other advanced economies had done so. The fact that Australia did not have such a program was contributing to Australian yields being higher than elsewhere and putting upward pressure on the exchange rate; the spread between 10-year yields for AGS and US Treasury bonds had risen to be around 25–30 basis points in mid-2020, from around –70 basis points in 2019. Given that unemployment was expected to remain high and inflation subdued for an extended period, this was seen as unhelpful.

The BPP initially involved the purchase of $100 billion of government bonds of maturities of around 5–10 years over the following six months. Purchases were made at a rate of $5 billion per week. Purchases were split 80/20 across nominal bonds issued by the Australian Government and by the states and territories, guided by the value of bonds outstanding.

The inclusion of state and territory government securities (semis) in the BPP was expected to lower the spread of semis yields to AGS yields. This was in addition to the decline of semis yields in line with AGS yields, given that semis are priced relative to AGS, which constitute the benchmark yield curve in Australia. The allocation of bond purchases across AGS and semis was weighted a little more to AGS than semis, relative to the value of bonds outstanding, which was 70/30 at the time. This reflected the fact that semis markets are smaller and less liquid than AGS markets, so there was a greater risk of purchases under the BPP adversely affecting the functioning of the semis market. At the same time, it was recognised that these features of semis markets also meant that regular purchases under the BPP had the potential to support smooth functioning of these markets. The allocation of bond purchases across the states and territories was guided by the stock of debt outstanding and relative market pricing, so as to avoid distorting yield relativities between states and territories and to support market functioning.^[4]

The bonds were purchased in regular auctions conducted by the Bank in the secondary market, not directly from the government, thereby maintaining the separation between monetary policy and fiscal financing. Purchases of AGS with residual maturity of around 5–7 years were held on Mondays, and purchases of AGS with residual maturity of around 7–10 years were held on Thursdays. Semis purchases were held on Wednesdays.

While the BPP did lower the cost of longer term finance for the government, along with all borrowers, the Australian Government and the states and territories continued to fund themselves in the market as usual. Indeed, the governments' bond auctions had been consistently well subscribed ahead of the introduction of the BPP. Even so, by adding to overall demand for government bonds, the BPP supported market functioning as the market absorbed the ongoing rapid issuance. Semis markets, in particular, can be thin at times, so the effect on market functioning can be an important consideration for the nature and timing of issuance.

In announcing the BPP, the Board noted that the Bank would closely monitor the impact of purchases on market functioning and adjust the auctions if necessary, including their size, composition and timing. The smooth functioning of government bond markets is important, not only for the ability of governments to fund themselves, but because government bonds are the benchmark fixed-income securities in Australia.

At $100 billion, the announced purchases amounted to 5 per cent of GDP or close to 20 per cent of government bonds outstanding in the maturity range of the program. Based on international evidence, this was expected to reduce the 10-year bond yield by around 30 basis points. The size of the BPP as a share of GDP was smaller than in other economies, although the purchases occurred at a relatively rapid pace relative to the stock of bonds outstanding. This reflected the smaller size of the government bond market in Australia compared with most other economies. The total size of the BPP after the final purchases in February 2022 remained below those of other central banks as a share of GDP, but within the range of others as a share of the stock outstanding.

The Board was well aware that the BPP, if successful, would come at a financial cost. The Board had been considering the costs and implications for the Bank's balance sheet associated with the unconventional monetary policy measures, including bond purchases, since the introduction of the first package of measures in 2020. This is discussed further in section 4 below. The fact the BPP was confined to bonds of maturities of around 5–10 years limited the Bank's exposure to interest-rate risk; this was in comparison with some other central bank asset purchase programs that included purchases at maturities beyond 10 years.^[5]

Early February 2021: Announcement of second $100 billion of bond purchases

In early 2021, the global economic outlook had brightened – positive advancements had been made on COVID-19 vaccines and the domestic economic recovery had progressed better than earlier expected. However, Australia's vaccination program had only just begun and the health and economic outlook remained very uncertain, with a continuing possibility of very bad economic outcomes. Further, the Australian economy remained well below the Bank's full employment and inflation goals.

It was against this background, and with the cash rate judged to be at its effective lower bound, that the Board decided at its February 2021 meeting to purchase an additional $100 billion of bonds when the existing program was completed in mid-April. These purchases would be made at the same rate of $5 billion a week.

In coming to this decision, the Board considered three factors, as recorded in the minutes from the February 2021 meeting:

1. the effectiveness of the bond purchases
2. the decisions of other central banks
3. the outlook for inflation and employment.

In terms of the BPP's effectiveness, the Board assessed that the BPP had helped to lower interest rates and had contributed to a lower exchange rate than otherwise. It also observed that Australia's government bond markets continued to function well and further purchases were not expected to be a source of market dysfunction.

The fact that many other central banks had recently announced extensions of their BPPs until at least the end of 2021 was another factor in the Board's decision. Given this – along with a widespread expectation among market participants that the Bank's program would be extended in some way – the Board judged that, if the Bank were to cease bond purchases upon completion of the initial $100 billion of purchases, this would likely lead to upward pressure on bond yields and the exchange rate, which would not have been helpful in the circumstances.

In regards to the third factor, it was expected to be some years before the Bank's goals for inflation and unemployment would be achieved.

In light of these three considerations, the Board decided that an extension of the BPP was appropriate to continue to provide further monetary support for the Australian economy.

The Board made and communicated the decision in relation to the future of the BPP two months ahead of completion of the first set of purchases in order to reduce uncertainty and unhelpful speculation regarding the Board's intentions, which could place upward pressure on bond yields and the exchange rate. Other central banks had provided early signals about the future course of asset purchase programs for similar reasons. At the same time, given the uncertainty about the outlook, the Board was careful not to make an unduly long-term commitment in advance regarding the BPP.

In the event, the announcement that the Bank would purchase an additional $100 billion of bonds after the initial purchases were completed led to a very modest reduction in yields – this was consistent with the extension having already been largely reflected in market pricing.

Late February to March 2021: Adjustment to purchases and communication amid change in market outlook

In late-February and early-March 2021, global yields increased as the global economic outlook improved following surprisingly good news on vaccination developments, a decline in COVID-19 cases in most advanced economies and a large US fiscal stimulus package. The associated bond-market sell-off and adjustment in yields was accompanied by an increase in the volatility of yields and a decline in bond market liquidity, including in Australia.

To support market functioning, the Bank brought forward some purchases under the BPP, increasing the size of the AGS auction on 1 March 2021 to $4 billion, compared with the scheduled $2 billion. The Bank also adjusted the format of the semis auctions in response to declining coverage in March 2021; the weekly semis auction was expanded to cover the 5–10 year tenors, rather than alternating between the 5–7 year tenors and 7–10 year tenors every other week as had been done up to that point. While the Board stated that the Bank was prepared to make further adjustments to its purchases in response to market conditions, this was not required in the event.

Longer term Australian Government bond yields rose in line with yields globally and, while the shorter end of the Australian yield curve remained fairly well anchored, there was some upward pressure on the target three-year yield. To support the yield target, the Bank made bond purchases (separate from the BPP) and publicly reiterated the Bank's commitment to the yield target. To assure markets of ongoing accommodative monetary conditions, the Board's preparedness to continue bond purchases beyond the announced extension of the BPP if required to reach the Bank's goals was also highlighted in public communication. This message was maintained in following months.

July 2021: Announcement to continue bond purchases until at least November 2021, but at a slower pace

The Governor's statement following the May Board meeting indicated that at the July meeting the Board would make a decision on the future of the BPP beyond September 2021, when the second $100 billion of purchases under the BPP would be completed. This continued the Board's approach of announcing its intentions regarding the future of the BPP well ahead of time to reduce uncertainty and unhelpful speculation. It was also indicated that a decision whether to retain the April 2024 bond as the target bond for the three-year yield target or to shift to the next maturity (the November 2024 bond) would likewise be made at the July meeting.

In preparation for these decisions, the Board discussed the key considerations at its June meeting. As recorded in the minutes from that meeting, the Board discussed options for the BPP, including:

• ceasing purchases in September;
• repeating $100 billion of purchases for another six months;
• scaling back the amount purchased or spreading the purchases over a longer period; or
• moving to an approach where the pace of the bond purchases was reviewed more frequently, based on the flow of data and the economic outlook.

The Board noted that a key consideration for the decision in July would be the progress made towards the Bank's goals for employment and inflation. In light of this, at the June meeting the Board agreed that it would be premature to consider ceasing the BPP in September, given that it judged that achievement of the Bank's goals remained a fair way off and the BPP had been one of the factors underpinning the accommodative conditions necessary for the economic recovery.

At its meeting in July 2021, the Board decided that bond purchases would continue beyond September until at least November 2021 at a purchase pace of $4 billion a week, down from the preceding pace of $5 billion a week. (The Board also decided to maintain the yield target for the April 2024 bond, rather than extending the target to the November 2024 bond.)

The decision on the BPP was made according to the previously agreed framework for making decisions about bond purchases, taking into account:

1. the progress towards the Board's goals for inflation and employment
2. the effectiveness of the bond purchases to date
3. the decisions of other central banks.

Continuation of bond purchases was deemed appropriate given:

• the Bank was likely to be short of its full employment and inflation goals for some time and the outlook remained uncertain; setbacks to the vaccine rollout and fresh COVID-19 outbreaks clouded the outlook
• the BPP had been successful in lowering risk-free rates across the yield curve, thereby reducing borrowing costs for households and businesses, supporting asset values and contributing to a lower exchange rate than otherwise
• most central banks in major advanced economies had maintained the pace of their government bond purchases, although some had indicated that a slowing in the pace of purchases would be likely in the period ahead.

Given the high degree of uncertainty about the economic outlook, the Board announced a weekly purchase pace over three months, rather than announcing purchases over a longer period. This would provide flexibility to respond to economic and financial developments in a timely way, by either increasing or reducing weekly bond purchases.

As the economy was on a better path than earlier expected and the outlook had improved, the Board decided to reduce the pace of weekly purchases to $4 billion – thereby reducing the stock of bonds purchased over any given period. The Board also recognised that a tapering in the pace of purchases would be helpful in signalling that the Board was moving towards an eventual exit from the BPP as conditions improved. This would help to avoid ever-growing expectations of additional bond purchases, which would make a smooth cessation of purchases more difficult to achieve.

The continuation of purchases at a weekly pace of $4 billion was within the range of market economists' expectations at the time, suggesting that such an outcome was likely to be at least partly already priced in by markets. Even so, in making this decision the Board was mindful of the risk that it could have been misinterpreted as a withdrawal of monetary support, resulting in an unhelpful tightening in financial conditions. Accordingly, in communicating the decision the Governor noted the empirical evidence that central bank bond purchases had their effect through the total stock of bonds purchased, not the flow of those purchases. The additional bond purchases, albeit at a slower pace, would continue to add to the support provided to the Australian economy. In the event, there was only a very slight increase in yields following the announcement of the decision, consistent with it being within the range of market expectations.

September 2021: Announcement to taper bond purchases as planned, but extend the purchase period until at least February 2022

By early September 2021, it was evident that the containment measures in response to the Delta outbreak that had occurred in Australia since July had interrupted the economic recovery and that the near-term outlook was highly uncertain and dependent on health outcomes. As a result, progress towards the Bank's goals was likely to take longer and was less assured. Given this, the Board announced that the period over which bond purchases would continue at the rate of $4 billion per week would be extended beyond November to at least mid-February 2022. This decision was made to add to the support provided to the economy during the recovery phase, providing some additional insurance against downside scenarios. It was again announced well ahead of time to reduce uncertainty and unhelpful speculation.

The Board had also considered the option of maintaining the rate of $5 billion of purchases a week (rather than tapering to $4 billion per week in September) and reviewing the program again in November as planned. At the time, a little over half of surveyed market economists expected the Bank to maintain the pace of bond purchases at $5 billion per week, or even increase this pace. However, the Board assessed that it remained appropriate to taper purchases given that:

• the shock to the economy was expected to be temporary, with containment measures expected to be eased over subsequent months
• the fiscal support being provided by the government was likely to be effective in supporting household and business balance sheets in the face of the temporary and sharp reduction in private sector incomes
• a number of other central banks were tapering their bond purchases, and – even after tapering from $5 billion a week to $4 billion – the Bank's bond purchases would be expanding faster relative to the stock of bonds outstanding than that of many other central banks
• proceeding with tapering in this way would signal that the Board was still moving towards an eventual cessation of purchases as conditions improved, supporting a smoother wind-down of purchases.

In the event, bond yields declined slightly and the Australian dollar exchange rate depreciated slightly following announcement of the Board's decision.

Bond yields subsequently rose sharply over October, reflecting rising inflation expectations and expectations that central banks would begin to reduce stimulus earlier than previously expected, including in Australia. As discussed in detail in the Review of the Yield Target, at its November 2021 meeting the Board decided to discontinue the yield target as it had become less effective and was inconsistent with the economic outlook.

February 2022: The ceasing of further bond purchases

In early February 2022, the Board announced its decision to cease further purchases under the BPP in mid-February. The Board had already flagged in September 2021 that it would review the future of the BPP at the February Board meeting.

At the December 2021 meeting, the Board had noted that the decision would depend on the previously agreed criteria.

Given the expectation that the economy would continue to bounce back following the outbreak of the Delta variant and that the recovery would not be derailed by the emergence of the new Omicron variant, the Board discussed three options:

1. reduce the pace of purchases from mid-February with an expectation of a likely end point in May 2022;
2. reduce the pace of purchases and review it again in May 2022; or
3. cease purchases altogether in mid-February.

The Board also recognised that a different set of options would need to be considered in February if there was another serious economic setback.

In early February, the Board decided on the third option – to end purchases under the BPP in mid-February. This was based on the following key considerations:

• It was clear that progress on the Board's employment and inflation goals had continued and was better than expected.
• Most other central banks had already completed their bond-buying programs, or would do so soon.
• The Bank staff's assessment was that, although the Australian bond market could accommodate further bond purchases by the Bank, eventually additional strains would be likely to emerge. Indeed, there had already been a deterioration in measures of market functioning in late 2021 alongside the rise in global government bond yields and the discontinuation of the yield target, although conditions had improved since then and increased stock lending from the Bank's bond holdings was supporting market functioning.

In announcing the decision, the Governor emphasised that:

• the stimulus associated with the stock of bonds already purchased in the program would continue to provide support to the economy for some time
• the decision to cease bond purchases did not imply a near-term increase in interest rates
• the Board remained committed to maintaining highly supportive monetary conditions to achieve its objectives of a return to full employment and inflation consistent with the target.

The Governor also flagged that the Board would consider the issue of the reinvestment of the proceeds of future bond maturities at its meeting in May, with the key considerations being the state of the economy and the outlook for inflation and unemployment.

May 2022: Announcement that the Board is not planning to reinvest on maturity or sell purchased bonds

In early May 2022, the Board increased the cash rate target and interest rate on ES balances by 25 basis points. With unemployment having declined to 4 per cent, inflation having picked up significantly and growing evidence that wages growth was also picking up, the Board decided it was time to begin withdrawing some of the extraordinary monetary support that was put in place during the pandemic. Consistent with this, the Board announced that it did not plan to reinvest the proceeds of its maturing government bond holdings or to sell the government bonds purchased during the pandemic. This meant that the contribution of the Bank's bond holdings to lower bond yields would slowly diminish over time as the bonds mature. For further details, see Kent (2022).

3. Effects of the bond purchase program on financial markets and the economy

The theory and international experience

There are three key channels through which central bank purchases of longer term government bonds can lower bond yields:

• Portfolio rebalancing – The accumulation of government bonds by the central bank bids up their price as other assets are only imperfect substitutes, removes interest rate risk from the market, reduces term premia, and induces investors to buy other assets (including to replace the bonds they sold).
• Signalling – Bond purchases underline the commitment of the central bank to hold policy rates lower for longer (including because policy rates are unlikely to be raised while bond purchases are ongoing) and so reinforce expectations for a low policy rate.
• Reducing liquidity premia – Steady central bank buying reduces the risk of investors being unable to sell bonds at a reasonable price, and increases commercial banks' reserve balances.

The portfolio rebalancing and signalling channels operate in the main via an ‘announcement effect’. That is, they cause bond yields to change when bond purchases are announced – or even earlier if such an announcement is widely anticipated by market participants – rather than when purchases are actually made. This is because market participants are forward looking, and so will price in the effects immediately. The liquidity premia channel will have a larger implementation component associated with actual bond purchases as and when they are made (this is especially the case in government bond markets, when liquidity premia are likely to be elevated only in stress conditions).

There is no consensus about the relative importance of each channel, but it is generally accepted that the liquidity premia channel is most important during periods of market stress; conversely, in more normal times, when government bond markets are liquid and well-functioning, liquidity premia are already low and so that channel is less important.

The empirical literature on bond purchases, based on experiences in other countries, suggests that around the time of the initial purchase program announcement each 1 per cent of GDP worth of purchases sees yields decline by approximately 5–7 basis points on average, although the range of estimates is wide. Initial BPPs also tend to have larger apparent effects than the announcement of subsequent programs or extensions of the initial program.^[6] This is because additional rounds of bond purchases are often anticipated by markets and so are at least partially priced in at the initial announcement of a program; it is therefore difficult to disentangle these pre-existing expectations from the new information in an announcement of a program extension. Also, many early BPPs were initiated during a period of market stress, when the liquidity premia channel of bond purchases was relatively important, whereas subsequent programs were often implemented in more settled markets when liquidity premia were low.

As highlighted in Atkin, Hartstein and Jääskelä (2021), there is a broad consensus in the international literature that the lower yields from BPPs in turn contribute to the exchange rate being lower than otherwise. However, the maturities of affected interest rates will be different compared with a conventional policy easing via the cash rate. Changes in conventional policy are reflected in policy rates and shorter term interest rates, which are typically estimated to have a larger effect on the exchange rate than longer term interest rates where BPPs have most of their effect. Even so, there is not a clear consensus around the magnitude of the exchange rate effect of BPPs compared with conventional policy measures.^[7]

Event studies are widely used to examine the effect of interest rates, and BPPs, on the exchange rate (Ferrari, Kearns and Schrimpf 2017; Glick and Leduc 2018; Rogers, Scotti and Wright 2014). A range of these studies suggest that an announcement that reduces long-term bond yields by 100 basis points typically results in a 3–9 per cent depreciation of the exchange rate. Some studies have also indicated that the sensitivity of the exchange rate to changes in interest rates has increased over time as policy rates have declined to low levels.

The Australian experience

The BPP was successful in lowering government bond yields, which flowed through to lower funding costs across the economy and a lower exchange rate than otherwise. This contributed to the strong recovery in the Australian economy following the pandemic. This section discusses the channels through which the BPP affected the economy, the observed effects in financial markets and the economy, and provides some empirical estimates of the effect of the BPP. However, these estimates understate the true effect. This is because it is difficult to estimate the effects of policy measures, such as the BPP, that have not been used before and to capture their reinforcing effects as part of a package of measures. While difficult to quantify, it is clear that the BPP and other monetary policy measures reinforced each other and together provided significant policy stimulus leading to the strong economic recovery and the unemployment rate now being at its lowest level in almost 50 years.

Effect on government bond yields

Reserve Bank staff have used a number of different methods to estimate the effect of the BPP on government bond yields in Australia. Finlay, Titkov and Xiang (2022) provide details of the results, which are summarised below.

Overall, this work suggests that the announcement effect of the BPP was to lower longer term AGS yields by around 30 basis points and to lower the spread of semis yields to AGS yields by around 5–10 basis points. As discussed below, taking into account the ultimate size of the BPP, much of which appears to have been priced in by markets at the outset, this is a smaller effect than generally suggested by international studies. However, that is not particularly surprising given the BPP was introduced at a time when other policy measures were already providing strong signals about future policy and market conditions were not strained (so liquidity premia were low). The Bank had purchased bonds in support of market function between March and May 2020, when government bond markets were under considerable strain at the outset of the pandemic. These purchases were conducted as required to support market function with no specified amount or timing.

Moreover, as highlighted by the international literature, estimating the effect of BPPs is very difficult. Adding the BPP to the broader package of policy measures provided further weight to the Board's commitment to do what it could to support the economy's recovery from the pandemic, particularly in the circumstances wherein almost all other advanced economies had bond purchase programs. In this context, and with widespread expectations that a BPP would be introduced, it seems likely there would have been a material increase in yields had a BPP not been adopted. However, it is difficult to isolate the contribution of the BPP in this regard, given the other elements of the package were working in the same direction.

Estimates of announcement effects using event study

As noted above, in normal bond market conditions – such as prevailed in Australia in late 2020 when the BPP was introduced – most of the effect on yields from central bank BPPs occurs when expectations are formed, rather than when purchases are made. Accordingly, an event study of the announcement effects of the BPP – where key dates in the lead-up to and announcement of the program are identified and the yield change that occurs on those dates is assessed – is one way to measure the overall effect of the BPP.

Finlay et al (2022) employed this method, identifying nine events in the two months leading up to the announcement of the BPP by the Board that affected market expectations. Examination of movements in bond yields around these events suggests a cumulative decline in 10-year AGS yields of around 30 basis points as market participants priced in the effect of the BPP. Given that the spread of 10-year AGS yields to overnight indexed swap (OIS) rates fell by a similar amount, it is likely the decline in 10-year AGS yields was for the most part driven by falls in AGS term and liquidity premia, and most likely the former (because outside of periods of market dysfunction, liquidity premia are typically low in the AGS market). OIS rates provide a measure of market expectations for the evolution of the cash rate, and so can be used as at least a partial control for any other macroeconomic or financial market news that was unrelated to bond purchases but affected cash rate expectations.

By contrast, falls in the spread of shorter dated AGS yields to OIS rates out to about five years were more limited around the identified events. This suggests that for these shorter dated maturities most of the observed fall in AGS yields was due to lower cash rate expectations, which are likely to in part have reflected the signalling channel of the BPP, which reinforced the Board's forward guidance about the cash rate.

The event study also suggests that the BPP led to a decline in the spread of semis yields to AGS yields – as was expected given the inclusion of semis in the BPP – of around 5–10 basis points.^[8]

While the Reserve Bank initially announced $100 billion of bond purchases, many market participants were likely to have expected from the outset that further extensions to this program would be announced in time. This implies that the 30 basis point fall in longer term AGS yields that was observed may be better explained by a larger total expected stock of purchases than the size of the initial announcement. It is difficult to be precise about how large the total expected stock of purchases might have been, but it seems reasonable to think that expectations might have been in the order of $200 billion to $300 billion, with later forecasts made by market economists over the first half of 2021 tending to fall within that range. This would imply that each $10 billion (or roughly ½ percentage point of GDP) of expected purchases over the life of the BPP resulted in a fall in longer term AGS yields of around 1–1½ basis points. While this is smaller than the size of the effect generally suggested by international studies of central bank BPPs (i.e. 1 per cent of GDP worth of purchases results in 5–7 basis point fall in yields), there is a range of estimates in these studies and it is difficult to measure the extent of bond purchases priced into yields by market participants. Further, a smaller effect might be expected in the case of the Bank's BPP given it was introduced at a time when forward guidance and the three-year yield target were already providing a powerful signal regarding future policy and liquidity premia were already very low.^[9] The evolution of market economists' forecasts and market reactions at the time of the BPP extension announcements is discussed in detail in Finlay et al (2022).

Estimates of effects of actual purchases

Finlay et al (2022) used three different methods to assess the ‘implementation effects’ of the BPP (i.e. the effects beyond the initial announcement effect of the BPP) as the purchases occurred. The details of this analysis are set out in the full Research Discussion Paper and include time-series and cross-section analyses examining yield movements in relation to the amount of bonds purchased at auction and whether the bonds were eligible for purchase in an auction. Overall, the implementation effect of the BPP was found to be small and temporary.

Estimates based on counterfactual scenarios

Finlay et al (2022) constructed counterfactual scenarios for how AGS yields might have moved in the absence of the BPP. The counterfactual is very difficult to know; however, one approach is to construct a model of AGS yields that controls for a handful of domestic and international factors but for the most part does not capture the effects of the BPP. The implied path of AGS yields resulting from this model is then used as a counterfactual against which to measure the yield impact of the BPP. This method provides only a rough approximation, given the difficulties in modelling yields. Subject to this caveat, this method suggests the BPP reduced long-term AGS yields by around 20 basis points on announcement, with this effect persisting through most of 2021. An even rougher approximation of the counterfactual is to assume that, in the absence of the BPP, longer-term AGS yields would have moved in line with US Government bond yields. The difference between the observed yield changes and this counterfactual also suggests that the announcement of the BPP led to a fall in longer term AGS yields of around 30 basis points, with no material additional effect as the purchases actually occurred.

Effects on the exchange rate and broader economy

In the months leading up to the announcement of the BPP in November 2020, AGS yields declined as market participants anticipated further monetary policy easing, including through a potential BPP. This flowed through to lower funding costs across the economy, as well as a lower exchange rate. The Australian dollar depreciated by around 5 per cent on a trade-weighted (TWI) basis over the period from early September to early November. Over the same period, commodity prices were little changed, while US equity prices declined amid a decline in global risk sentiment; this is also likely to have played a role in the depreciation of the exchange rate. (See below for model estimates of the effect of the BPP on the exchange rate.)

The lower government bond yields flowed through directly to lower government borrowing costs (discussed in section 4 below) and were associated with a substantial increase in infrastructure spending by the state and territory governments. The BPP also lowered borrowing rates for households and businesses through a number of channels. AGS yields act as the benchmark yield curve in Australia, with other fixed-income securities typically priced relative to AGS yields or swap rates (which embody market expectations for the cash rate and a term premium, and are closely correlated with AGS yields over time).

Accordingly, the BPP exerted downward pressure on banks' funding costs by lowering the cost of issuing long-term wholesale debt, such as bonds. In addition, portfolio rebalancing by some investors contributed to increased demand for bank bonds. Bank bond spreads (and yields) declined to very low levels relative to history, reflecting the combined effects of the Bank's package of policy measures.

However, low bank bond issuance suggests the overall direct effect on bank funding costs through lower long-term bank bond yields was likely small during 2021. The share of the major banks' funding drawn from long-term wholesale debt declined over 2020 and 2021, partly reflecting increased use of TFF and deposit funding, which displaced long-term debt issuance. This decline in issuance contributed to the decline in bank bond spreads. Bank bond issuance picked up in early 2022 as banks responded to the wind-down of the Committed Liquidity Facility over 2022 and TFF maturities from 2023.

Bond purchases also increased deposits in the banking system, which led to lower bank funding costs and put downward pressure on housing and business lending rates. While growth in the stock of deposits reflected a number of factors, the BPP directly created a significant amount of deposits (Fitzpatrick, Shaw and Suthakar 2022). Bonds were purchased from commercial banks via auction, and were paid for with newly created money credited into banks' ES Accounts. Some of these bonds sold by commercial banks would have been purchased from private (non-bank) investors, generating a flow of funds into non-bank investors' deposit accounts. The BPP thereby contributed to the deposit share of banks' total funding increasing to around 60 per cent, from around 55 per cent prior to the pandemic. The share of government bonds held by non-bank investors fell to around 50 per cent by February 2022, from around 70 per cent in the several years preceding the BPP. While the counterfactual is unknown, if we assume that in the absence of the BPP non-bank investors' holdings share had remained at its pre-pandemic level, much of the Bank's bond purchases under the BPP may have been ultimately sourced from the non-bank sector.

Most of the new deposits created over the past two years flowed into at-call accounts held by households and businesses, while the volume of term deposits decreased over this period. Both deposit growth and the changing composition of deposits contributed to a decline in deposit rates, which lowered banks' funding costs.^[10] Deposits created by previous bond purchases will continue to put downward pressure on deposit rates until they are drained by coupon payments to the Bank and repayments to the Bank upon maturity of the bonds.^[11]

The shift lower in benchmark yield curves and favourable pricing for RMBS issuance also lowered funding costs for non-bank lenders over 2021. From late 2020 to late 2021, on top of the decline in AGS yields, spreads on RMBS declined by around 60 basis points; this was more than the previous four monetary policy easing phases, despite the magnitude of the decline in the cash rate during those episodes being larger than during the pandemic. Non-banks responded by raising funds at the fastest pace since before the global financial crisis. One factor supporting the RMBS market in this episode was the reduction in the supply of bank bond issuance partly reflecting the funding banks obtained from the TFF. Another factor was the public sector purchases of RMBS through a complementary program created by the Australian Government – the Structured Finance Support Fund, implemented by the Australian Office of Financial Management. It is also possible that the portfolio rebalancing channel of the BPP played a role, although this channel is strongest for assets that are most similar to the bonds purchased by the Bank (Krishnamurthy and Vissing-Jorgensen 2011).

The effects of the BPP, in conjunction with the other elements of the Bank's policy package, saw lending rates offered by banks and other intermediaries on loans to households and businesses fall to historical lows. This was evident in substantial declines in rates on new and outstanding loans, for both housing and business lending. However, it is difficult to disentangle the effects of the BPP from that of the other measures, because the cash rate target, the yield target, and the interest rate on TFF drawings were all reduced alongside the BPP announcement. The TFF and yield target were particularly relevant for funding costs around the three-year maturity and the rapid growth in fixed-rate mortgages (as discussed in the Review of the Yield Target).

Low funding costs supported the supply of credit to the economy. Both housing and business credit growth picked up. The BPP helped to reinforce the Bank's forward guidance around the cash rate (the signalling channel), supporting demand for credit from households and businesses. International studies suggest the announcement of the BPP may also have lowered uncertainty and reduced the perceived risk of the most negative ‘tail’ outcomes, thereby supporting credit demand.^[12]

The BPP also contributed to lower borrowing costs for larger businesses raising funds directly from financial markets. Non-financial corporate bond yields and spreads declined in late 2020 to their lowest levels since at least the global financial crisis and bond issuance in 2020 and 2021 was above average, particularly for bonds with maturities of around 5–10 years. As well as contributing to a shift lower in benchmark yield curves and lower spreads as a result of any portfolio balancing effect, international evidence suggests the BPP may also have lowered corporate bond spreads through reducing uncertainty and improving the outlook for the economy – and therefore lowering default risk premia (Kaminska and Mumtaz 2022).

These effects of the BPP, together with the other elements of the policy package, contributed to the strong recovery of the Australian economy following the pandemic, with the unemployment rate having now fallen from a peak of around 7 per cent in 2020 to 3.4 per cent in July 2022 – its lowest level in almost 50 years.

Model estimates of the effect on the economy

This section provides estimates of the effect of the BPP on the economy based on event studies (discussed above) and the Bank's macroeconometric model (Ballantyne et al 2019). The Bank's macroeconometric model is designed to capture the key economic relationships relevant for the conduct of monetary policy in Australia; as with all macroeconometric models, it is a simplification of the real world. Using this model to assess the effects of the BPP is particularly challenging because there is no historical experience of this policy tool. Notably, the BPP works through novel channels, including the signalling and portfolio rebalancing channels, which are not captured in the model. The model also does not account for the fact that the BPP was implemented as part of a package of policy measures that had reinforcing effects, including through supporting the confidence of households, businesses and governments to continue spending. Accordingly, the estimated effects will be understated by the results from this model.

As discussed in Atkin, Hartstein and Jääskelä (2021), estimates from a range of exchange rate models and event studies around the Reserve Bank's policy announcements suggest that a decline in both shorter term and longer term AGS yields results in the Australian dollar exchange rate being lower than otherwise.^[13] Based on these estimates, the 30 basis point decline in long-term bond yields (and yield differentials all else equal) that occurred as a result of the BPP would typically suggest around a 1–2 per cent depreciation in the exchange rate based on historical relationships. However, similar to international studies, there is a high degree of uncertainty around the estimated effects of quantitative easing measures on the exchange rate in Australia.

Estimates from the Bank's macroeconometric model suggest that the 30 basis point decline in long-term bond yields and the 1–1½ per cent depreciation of the exchange rate associated with the BPP resulted in stronger economic activity and inflation than otherwise. After the first three years alone, model estimates suggest there would be an extra $25 billion of nominal GDP in cumulative terms.^[14] Because the effect of the BPP will persist for some time – in line with the gradual reduction in bonds outstanding as they gradually mature – the positive economic effect will also persist. However, as noted above, this understates the effect of the BPP given the limitations of the model.

The Bank's macroeconometric model connects inflation and unemployment through a Phillips Curve relationship. However, some commentators have pointed to an alternative approach considering the identity that links nominal spending to the quantity of money and the velocity of money.^[15] If the velocity of money is assumed constant, this leads to the Quantity Theory of Money, which suggests that a large increase in the money supply owing to the BPP would lead to a sharp increase in inflation.

While inflation has increased following the BPP, it is not clear that this can be explained by the Quantity Theory. In particular, there has not been a stable relationship between money aggregates and economic activity or inflation in Australia for several decades.^[16] The velocity of money is not constant and has declined over time, falling sharply during 2020. Moreover, different components of the money supply can move independently over time. While the BPP led to a sharp increase in ES balances and thus ‘base’ money, the increase in the broader money supply, which is relevant for nominal expenditure in the economy, was not as large.^[17]

Effects on market functioning

The international experience suggests that central bank purchases – up to a point – can support good bond market function and lower liquidity premia, particularly in times of market stress. This was in fact the aim of the purchases made by the Bank between March and May 2020. These purchases helped to restore good function in government bond markets, which were under considerable strain at the time (Finlay et al 2022). However, purchasing a very high share of outstanding government bonds could have negative effects on these bond markets. Too much scarcity of particular bonds could lead to bond dealers becoming reluctant to offer to sell that bond to their clients for fear they would be left short. This could result in reduced bond market liquidity, and could see bid-offer spreads widen and pricing anomalies emerge. Eventually, this could diminish the attractiveness of government bond markets for investors and contribute to larger liquidity premia for government bonds in Australia. As discussed above, the risk of this occurring was a consideration in the Board's decisions around the BPP. The Bank has also been prepared to lend its bond holdings back into the market for short periods to alleviate any strains that might occur.

One measure of market functioning is bid-offer spreads. The deterioration in market conditions in early 2020 at the onset of the pandemic was reflected in a sharp increase in bid-offer spreads (among other indicators) (Finlay et al 2022). Bid-offer spreads then declined noticeably as the Bank purchased bonds through March to May to support market functioning. Bid-offer spreads for shorter term AGS rose through 2021, spiking in particular for the April 2024 AGS around the discontinuation of the yield target in late 2021, suggesting that the Bank's substantial holdings of the April 2023 and April 2024 AGS – purchased in support of the yield target – resulted in some deterioration in market function around the three-year part of the yield curve. Lending of three-year bonds by the Bank increased sharply in late 2021, helping to support market functioning; demand to borrow government bonds from the Bank has remained high by historical standards over 2022 to date. By contrast, bid-offer spreads remained very low for longer term AGS and for semis from late 2020 when the BPP was announced and through 2021. Indicators of bond futures market functioning – including bond futures market depth and spreads between bonds and futures – tell a similar story (Finlay et al 2022).

Regressions conducted by Finlay et al (2022) estimating the effect of BPP purchases on bid-offer spreads for AGS suggest that the Bank's purchases led to only a small rise in bid-offer spreads. Each 10 percentage points of purchases of the free float of AGS was associated with bid-offer spreads being 0.2 basis points wider; similarly, each 10 percentage points of holdings of the outstanding stock was associated with bid-offer spreads being 0.2 basis points wider. By contrast, for semis purchased under the BPP, there was no statistically significant effect of purchases on bid-offer spreads. Taken together, these results suggest that in the liquid AGS market, the Bank removing a sizeable share of bonds from the market outside of periods of stress led to a modest rise in bid-offer spreads as bonds become more difficult to source. Conversely, for the less liquid semis market, it appears that regular purchases of modest amounts of these bonds by the Bank offsets any negative impact from less bonds being available to the wider market.

4. Financial impacts of the bond purchase program

The BPP affects the overall public sector balance sheet in a number of ways. The BPP was expected to have a direct impact on the Bank's balance sheet – via the size and cost of the associated liabilities (ES balances) – although the extent could not be known at the outset. There were also likely to be some indirect effects associated with changes in seigniorage received by the Bank.^[18]

Conceptually, the yield on a government bond is equal to the average cash rate expected to prevail over the life of the bond, plus a term premium. If a BPP is successful in lowering bond yields upon its announcement, and supports a faster economic recovery than would have otherwise been the case, then it would contribute to a higher-than-previously expected path for policy rates. In that case, the fixed return earned on the purchased bonds will turn out to be less than the interest paid on the corresponding ES balances. The eventual direct financial cost to the Bank of the BPP will depend on the path of policy rates over the life of the purchased bonds. Some scenarios are presented below to demonstrate how these direct costs to the Bank vary under different paths of policy rates.

The stronger economy as a result of the BPP has also had a positive effect on the general government balance sheet. The support provided by the BPP resulted in an increase in economic activity and tax revenue and a decrease in the government's financing costs, compared with otherwise. Higher nominal output (via real GDP growth and inflation) also lowered the ratio of government debt to GDP compared with otherwise. By itself, higher nominal GDP increases the ability to service a given value of debt, since taxation revenues rise with the value of nominal economic activity. These benefits to government finances are material, although difficult to quantify.

Direct effects on the Reserve Bank's balance sheet and earnings

The Bank's balance sheet increased substantially as a result of the BPP. The increase in the Bank's domestic bond holdings on the assets side was matched by an increase in liabilities in the form of ES balances paid to the dealer banks that were selling the bonds. (The size of the balance sheet will decrease in future as bond maturities reduce both the Bank's assets and the level of ES balances in the banking system.^[19])

The BPP has made the Bank's earnings more sensitive to interest rate movements. This is because the interest income received on the bonds in the program is fixed, while the interest paid on the ES balances created to purchase the bonds varies according to the cash rate target. The Bank's net interest income – the income earned on the bonds less the interest paid on ES balances – is now much more exposed to a rise in the cash rate than prior to the BPP. Given this, an estimate of these ‘earnings-at-risk’ was added to the Bank's capital target in 2020/21 (RBA 2021).

Interest rate movements also affect the value of bonds recorded on the Bank's balance sheet, which will feed through to the Bank's accounting profit or loss. The Bank reports net profit in accordance with Australian Accounting Standards, under which bond holdings are measured at fair value through profit or loss (i.e. marked to market).^[20] If yields increase, the market prices of bonds decline and the Bank records a valuation loss (or a valuation gain if yields decline and the market prices of bonds increase). However, given the Board plans to hold the BPP bonds to maturity, the Bank will receive the face value of the bonds at that point, irrespective of any mark-to-market gains/losses in the interim.

The accounting treatment of BPPs varies across different central banks. Some, like the Bank, report the market value of their programs – this includes the Riksbank, the Bank of England and the Reserve Bank of New Zealand. However, for the latter two central banks, government indemnities mean their profits are unaffected by mark-to-market gains or losses, because they are offset by a matching entry that reflects the value of their indemnity.

The Reserve Bank Board considered whether to seek a government indemnity for the BPP, but decided it was not necessary. The Board recognised that an indemnity would insulate the Bank's balance sheet from the effects of the BPP, but it would have no effect on the overall public sector balance sheet; the impact of the BPP would simply be transferred from the Bank's balance sheet to the government's balance sheet. Given this, and the risk that an indemnity could reduce the Bank's policy effectiveness if it weakened perceptions of the Bank's independence, the Board decided not to seek an indemnity. The decision also reflected the fact that a large loss – that resulted in a negative equity position – would not affect the ability of the Bank to do its job. A number of central banks in other countries have operated for extended periods with negative equity. Several other central banks are, or are likely to be, in a similar position over coming years.

By contrast, the Federal Reserve, the Bank of Japan and the European Central Bank account for their programs at amortised cost. The annual earnings of central banks following this approach will adjust over time, depending on the net interest income from their BPPs and a constant amortisation of the purchase premiums paid on bonds.

While the accounting treatment that is used will affect when a central bank's gains or losses are recognised, it will not alter overall financial gains or losses – the bonds will still mature at their face value or be sold at the prevailing market yields.

Projections of the direct effect on the Bank's earnings

The direct effect of the BPP on the Bank' earnings will be determined by:

• the (known) total return on the bonds purchased – the total coupon income on the bonds less the purchase premium the Bank paid on the bonds (the total purchase price of the bonds was around $300 billion whereas the total face value of the bonds, which the Bank will receive when they mature, is $281 billion)^[21]
• the (unknown) cost of funding the program – the rate the Bank pays on the ES balances created to purchase the bonds, which will depend on future monetary policy decisions of the Board.

Four scenarios are presented below to illustrate how the direct effect on the Bank's earnings varies with different paths of policy rates. For all scenarios, the ES rate is projected to evolve in line with the future path of the cash rate implied by prices for OIS contracts out to mid-2023, assuming the margin between the cash rate target and the ES rate remains constant at its current level of 10 basis points. The projections beyond 2023 assume the ES rate converges to either 1½ per cent, 2½ per cent, 3½ per cent or 4½ per cent.

The first graph below shows these projected paths for the ES rate, along with the weighted average (net) yield on the outstanding bonds held by the Bank over time (i.e. coupon payment less purchase premium amortisation^[22]). The yield increases over time as the shorter dated bonds – which tend to have lower yields – mature before the longer dated bonds. The second graph below shows the stock of outstanding bonds held by the Bank. This is the stock on which the Bank earns the yields shown in the first graph over time.

The graph below shows the net financial return of the BPP to the Bank (the yield received on the Bank's bond holdings less the compound interest paid on the corresponding ES balances) for each scenario on an annual basis.^[23] The net earnings flow is initially positive, but turns negative once the ES rate begins to rise. Overall, the cumulative financial cost of the program to the Bank over the period to 2033 is $35 billion, $42 billion, $50 billion and $58 billion under the scenarios from lowest to highest ES rate path. This is the cumulative effect on the Bank's balance sheet in principle; however, in practice, the actual cumulative interest paid on ES balances may be less than the amount included in this calculation. In particular, the actual stock of ES balances may decrease as new banknotes are issued in exchange for ES balances. Moreover, there will be other effects on the Bank's balance sheet associated with rising yields both in Australia and offshore. Most notably, the Bank's income from seigniorage will rise given the return on the Bank's assets will increase – particularly on its foreign exchange reserves – while the Bank's banknote liabilities do not incur additional costs.

The Bank's reported profits under the mark-to-market accounting treatment used in the Bank's financial accounts are likely to be volatile over the years ahead as bond yields move. In the 2021/22 financial year, a large loss was recorded because the increase in bond yields (consistent with the higher expected path of the ES rate) caused the market value of bonds in the BPP to fall below the purchase price, and indeed below their face value.^[24] By the time the bonds mature at their face value, the Bank will have recorded a net valuation gain from the current value of the BPP bonds.

Effects on the general government balance sheet

The decline in government bond yields as a result of the BPP purchases in the secondary market led to government debt being issued at a lower (fixed) cost than otherwise. Based on the estimated effects on government bond yields from the event study in Finlay et al (2022), the federal and state governments' overall costs on debt issued between November 2020 and February 2022 are estimated to have been around $7 billion dollars lower (in nominal terms) than they would have been in the absence of the BPP.^[25] The Bank's holdings of government bonds under the BPP continue to contribute to lower yields, although this effect will diminish as the bonds mature over the next 10 years. While the ongoing effect on yields is difficult to quantify, it will contribute to lower costs of issuance (that occurs beyond February 2022) than otherwise, hence the total benefit associated with lower issuance costs for governments will be larger than the estimate above. In addition, the higher nominal GDP as a result of the stimulus from the BPP reduces the outstanding government debt when expressed as a ratio to nominal GDP.

By contributing to stronger economic outcomes than otherwise, the BPP – together with the other measures in the policy package – has also contributed to an improved budget position, with higher tax revenue and lower support payments to households and businesses than would otherwise have been the case. These benefits to government finances are material, although difficult to quantify.

5. Assessment

The BPP was one of a range of extraordinary policy measures implemented by the Bank in response to the pandemic. It was included in the Bank's second package of policy measures, which was implemented at a time when it was judged the cash rate was at its effective lower bound and other unconventional monetary policy tools had already been employed to the extent it made sense to do so in the circumstances. The Board sought to insure against the extreme downside risks in the circumstances, recognising the importance of decisive action when there was limited scope for further policy action if the downside prevailed. In assessing the BPP, four aspects have been considered:

1. its success in lowering bond yields, funding costs more broadly and the exchange rate
2. its effect on economic growth and jobs
3. whether it was executed without materially impacting market functioning
4. its effect on the public sector balance sheet.

In terms of its effect on yields, the BPP is estimated to have lowered longer term Australian Government bond yields by around 30 basis points. Together with the other policy measures, the BPP was successful in achieving its objectives of further lowering the whole structure of interest rates in Australia, reducing borrowing costs to record lows and contributing to a lower exchange rate than otherwise.

However, it is difficult to isolate the effect of the BPP on the economy. The BPP was a new policy tool that works through novel channels and was implemented as part of a package of policy measures that reinforced each other, supporting confidence in the economy in the face of serious downside risks. Model estimates cannot capture these effects. Moreover, a key benefit of the package was insurance against a downside counterfactual that did not eventuate, at least in some part because the package of policies was successful in supporting the economy in the face of very difficult conditions and extreme uncertainty. Regardless of these measurement difficulties, it is clear that the economy recovered very strongly from the pandemic, supported by the package of monetary policy measures: economic activity is growing strongly and unemployment is now at an almost 50-year low.

In terms of the execution of the program, the design and implementation of the BPP worked broadly as intended, without materially affecting market functioning. The fairly short announced purchase horizons of the BPP – as well as the flexibility to adjust the size, composition and timings of auctions within the BPP in response to market conditions – helped to minimise the effect of the BPP on government bond market functioning. Moreover, the additional demand from the BPP may have supported government bond market functioning in the face of rapid government issuance at a time when fiscal stimulus was critical to supporting the economy in the face of an unprecedented negative shock and extreme uncertainty. This is particularly relevant for the semis market, which tends to be less liquid than the AGS market.

The BPP has affected the public sector balance sheet as a whole in several ways. The BPP is expected to entail a sizeable direct financial cost to the Bank, given that it has contributed to a stronger economy and thus a rise in interest rates. The eventual direct financial cost to the Bank will depend on the path of policy rates over the life of the purchased bonds. The unexpectedly sharp increase in inflation, here and globally, and the need to raise the cash rate to return inflation back to the target has added to the likely direct cost of the program. At the same time, the stronger growth and inflation as a result of the BPP will provide some offset to this cost via the seigniorage the Bank receives from issuing banknotes. The net cost of the BPP to the Bank will reduce the dividends it pays to the government.

In terms of the general government balance sheet, the BPP resulted in government debt being issued at a lower fixed cost than otherwise. There will be an ongoing, albeit diminishing, saving on debt issued over the next 10 years, as the Bank's holdings gradually mature. The economic stimulus from the BPP also contributed to a budget saving for the government via its positive effect on economic activity and nominal income. In addition, the debt-to-GDP ratio is lower than otherwise as a result of the boost to nominal economic activity.

As was noted in the Review of the Yield Target, the Board's focus on providing insurance against very bad outcomes – that ultimately did not eventuate – meant that upside scenarios did not receive as much focus in decision-making. Similarly for the BPP, had more weight been given to the upside scenarios, it may have altered the relative weighting of the benefits against the potential costs of the BPP, and led to different decisions about the BPP. However, it is worth noting that the current unemployment and inflation outcomes – and the resulting increases in policy rate expectations and the associated direct financial cost to the Bank of the BPP – are well beyond the upside scenarios considered during 2021. Indeed, they are beyond what most economists had predicted at that time.

In light of this experience, the Board remains of the view that it is appropriate to consider use of a BPP only in extreme circumstances, when the usual monetary policy tool – the cash rate target – has been employed to the full extent possible. If circumstances were to warrant the use of unconventional monetary policies in future, the following lessons from the experience of the past couple of years would be relevant in considering the use of a BPP:

• In the absence of a yield target for the shorter end of the yield curve, a BPP could be used that includes purchases at the shorter end of the yield curve. Compared with the BPP adopted in November 2020 (which did not include short-end purchases because the three-year yield target was already in place), this would contribute more directly to lower household and business borrowing costs – because these tend to ‘price’ off the shorter end of the yield curve. Such a BPP may also have a greater effect on the exchange rate, given the empirical evidence that shorter term interest rates have a larger effect on the exchange rate than longer term interest rates. Moreover, shorter-maturity bond purchases entail less financial risk compared with purchases of longer term bonds. However, because there is typically less new issuance by the government at shorter maturities, the amount of these bonds that could be purchased before market functioning issues arose might be more constrained. Hence, the extent of stimulus able to be provided through the portfolio balance channel of such purchases may be limited.
• Compared with a yield target, a BPP would provide more flexibility to respond to evolving economic circumstances, although it could entail larger financial costs. These costs would need to be carefully assessed at the outset and as the circumstances evolved.
• The short purchase horizons and flexibility of the BPP were helpful in facilitating timely adjustment of the BPP to evolving market conditions, thereby reducing the risk of compromising market functioning. The relatively short purchase horizons of the BPP also assisted in a timely and relatively smooth end to purchases by allowing for adjustments to be made as the economic outlook evolved.
• The Board's communication of the considerations around adjustments to the BPP prior to the decision points, which were themselves before the implementation dates, helped to avoid an unhelpful, disruptive adjustment as bond purchases wound down. This approach came with the trade-off that decisions were being announced on the basis of less timely information than would be available at the implementation date. However, the Board could, and did, reconsider decisions closer to the implementation date, in light of changing economic conditions.
• As noted in the Review of the Yield Target, in light of the recent policy experience the Board has agreed to strengthen the way it considers a wide range of scenarios when making policy decisions, especially when they involve unconventional policy measures. In relation to any future BPP, this scenario analysis would include the benefits and potential costs of the policy, recognising the difficulties in measuring these outcomes.

Endnotes

See, for example, Lane (2021): ‘While central banks can technically raise nominal interest rates without limits, there is only limited space to lower rates into negative territory, owing to the lower bound on cash. … To address the asymmetry of the effective lower bound constraint, the commitment to a symmetric inflation target requires especially forceful or persistent monetary policy action when the economy is close to the effective lower bound, to avoid negative deviations from the inflation target becoming entrenched. … This implies that, faced with large adverse shocks, our policy response will include an especially forceful use of its monetary policy instruments. In addition, closer to the effective lower bound, it may also call for a more persistent use of these instruments. This may also imply a transitory period in which inflation is moderately above target.' [1]

International work points to the beneficial effect of having a package of policy measures tailored to individual country circumstances (CGFS 2019). [2]

Well prior to the pandemic, the Bank had actively considered potential use of unconventional monetary policy tools in the event of reaching the effective lower bound. Papers on the subject were discussed by the Board in July 2016 and August 2019, and the Governor shared considerations publicly in a speech (Lowe 2019). This work had focused on tools used elsewhere: negative interest rates; funding for lending schemes; forward guidance; and bond purchase programs. It concluded that, on the basis of international evidence, it was not clear that the experience with negative interest rates had been a success. While negative rates put downward pressure on exchange rates and long-term bond yields, they can create strains in parts of the banking system and problems for pension funds that need to fund long-term liabilities. In addition, there is evidence that they can encourage households to save more and spend less; they can also damage confidence in the general economic outlook and make people more cautious. [3]

Semis auctions were conducted by grouping similar states and territories together (three auctions covering: New South Wales and Victoria; Queensland and Western Australia; and the other states and territories), with the proportion of bonds bought from each issuer guided by their stock of bonds outstanding, with allocation across different bond lines guided by yield deviations from mid-market rates. [4]

In addition, the Board's decision to restrict purchases to bonds issued by the Australian Government and state borrowing authorities further reduced the Bank's exposure to credit risk – that is, the potential for financial loss arising from the default of a debtor or issuer, or from a decline in asset values following a deterioration in credit quality. [5]

See, for example, Gagnon (2016); CGFS (2019); Bailey et al (2020); Independent Evaluation Office (2021). [6]

For example, Ferrari, Kearns and Schrimpf (2021) provide evidence that the effect of unconventional monetary policy on the exchange rate in some circumstances is larger than that of conventional monetary policy. [7]

The decline in spreads is estimated to have been around 5 basis points when measured over a one-day event window, and around 10 basis points when measured over a two-day window; as semis are less liquid than AGS, measuring yield changes over a slightly longer period may be appropriate. [8]

Rebucci, Hartley and Jimenez (2022) report a bond purchase announcement effect for Australia of 23 basis points after one day and slightly less than 40 basis points after three days. However, the effect captured is not that of the BPP, but rather the announcement of the monetary policy package on 19 March 2020, which included the purchase of bonds in support of the yield target and market functioning, alongside the reduction in the cash rate target and the introduction of the TFF. [9]

The overall decline in deposit rates has been more important in the decline of outstanding funding costs than the increase in the stock of deposits. This is because, despite the large increase in the stock of deposits, new deposits still represent a relatively small share of banks' total funding whereas reductions in deposit rates affect the entire stock of deposits. [10]

For AGS, maturities and coupon payments made to the Bank are made from government deposits with the Bank, which in turn are sourced by the government from deposits of the non-bank sector (e.g. via taxation or bond issuance). By comparison, for semis, maturities and coupon payments made to the Bank from the state and territory central borrowing authorities drain deposits (and ES balances) directly as these funds are typically held by those authorities at commercial banks. [11]

Weale and Wieladek (2016) find asset purchases in the United States and the United Kingdom reduced measures of financial market and household uncertainty, which contributed to an increase in real GDP and inflation. [12]

Ferrari et al (2017) provide estimates of the exchange rate effect from a change in central bank policy decisions using an event study methodology for a range of advanced economies, including Australia. The estimates indicate that a 100 basis point increase in two-year and 10-year yields corresponded to an exchange rate appreciation of around 5½ per cent and 4½ per cent, respectively. Estimating a similar model with more recent data and slightly different time windows around policy announcements in the Australian context provides a wider range of sensitivities of the Australian dollar to interest rate changes. The range of estimates highlights that there is a high degree of uncertainty. However, the updated estimates also suggest that the sensitivity of the Australian dollar to yields appears to have increased over time, which is consistent with Ferrari et al (2017). [13]

The effect of changes in the exchange rate on GDP are estimated to be around 15 per cent lower when international border closures are taken into account (i.e. when exports and imports of travel and passenger transport do not respond to changes in the exchange rate). [14]

Nominal output is equal to the product of the money supply and the ‘velocity’ of money (i.e. the frequency at which money is used in purchases during a given time period). [15]

The financial deregulation of the 1980s is frequently cited as driving the breakdown in the expected relationship between monetary aggregates and nominal income. Cline (2015) shows a similar breakdown in the Quantity Theory relationship in the United States following the 1980s. Sargent and Surico (2011) suggest that the lack of empirical support for the Quantity Theory may be explained by the shift towards a significantly more aggressive anti-inflationary stance. [16]

This largely reflects an increase in banks' demand for liquidity and changes to the Bank's monetary policy implementation discussed in Dowling and Printant (2021). Historically, growth in base money was closely linked to the expansion of bank balance sheets and credit creation through fractional reserve requirements. It has become less relevant in recent decades as these requirements have been removed, and various requirements of the market and the prudential regulator regarding liquidity and capital adequacy partly decoupled banks' balance sheets from central bank liabilities (Doherty, Jackman and Perry 2018). [17]

Seigniorage is the financial benefit (or loss) that a government or central bank receives from issuing currency. For further details, see Wakefield, Delaney and Finlay (2019). [18]

When the Bank's holdings of semis mature, the Bank will receive payments from the commercial banks used by the state borrowing authorities, which will reduce the Bank's liabilities in the form of ES balances. When Australian Government bonds mature, the associated payments will reduce the Bank's liabilities in the form of government deposits. The Australian Government will fund these payments through taxation or bond issuance, which reduces ES balances. [19]

This is because bonds are held to implement monetary policy and may be sold or lent, typically for short terms, under repurchase agreements. [20]

The Bank paid a premium because the coupon rates (which are set when the bonds are first issued) were generally higher than prevailing market yields, which were at historically low levels. The weighted average coupon rate (i.e. the dollar value of coupons paid as a share of face value) was around 2¼ per cent, while the weighted average yield to maturity (i.e. overall rate of return as a share of purchase price) was around 1.1 per cent. The actual composition of the bonds held in the BPP may vary due to the Bank's switch facilities, which allow counterparties to purchase securities held by the Bank in exchange for an offsetting duration-neutral sale of government bonds. These compositional changes are unrelated to the BPP and are ignored in the analysis in this paper. [21]

Assuming the value of the bonds moves in a straight line from their price at purchase towards their face value on the date they mature. [22]

In each scenario, the interest paid on ES balances created by the BPP is added to the outstanding stock of ES balances. This means that the amount of ES balances on which interest is being paid grows relative to the face value of the Bank's bond holdings. [23]

The AGS purchased under the BPP are recorded as assets on the Bank's balance sheet and as liabilities on the AOFM's balance sheet; in both cases they are marked to market, resulting in offsetting movements in the Bank's and AOFM's accounting profit and loss over time. [24]

Between November 2020 and February 2022, the AOFM raised $132 billion in funding through issuance of longer term (nominal) AGS at a weighted average yield of around 1.3 per cent. Under a counterfactual scenario where bond yields are estimated to have been higher owing to the absence of the BPP (based on the estimates in Finlay et al (2022) discussed above), the weighted average yield of the same issuance would have been around 1.5 per cent. As a result, the cost of this borrowing would have been $4 billion higher in nominal terms. The same exercise for state and territory government issuance indicates a nominal saving of around $3 billion in borrowing costs as a result of the BPP. [25]

References

Atkin T, I Hartstein and J Jääskelä (2021), ‘Determinants of the Australian Dollar Over Recent Years’, RBA Bulletin, March.

Bailey A, J Bridges, R Harrison, J Jones and A Mankodi (2020), ‘The Central Bank Balance Sheet as a Policy Tool: Past, Present and Future’, Paper prepared for the Federal Reserve Bank of Kansas City Economic Policy Symposium on ‘Navigating the Decade Ahead: Implications for Monetary Policy’, Jackson Hole, 27–28 August.

Ballantyne A, T Cusbert, R Evans, R Guttman, J Hambur, A Hamilton, E Kendall, R McCririck, G Nodari and D Rees (2019), ‘MARTIN Has Its Place: A Macroeconometric Model of the Australian Economy’, RBA Research Discussion Paper No 2019-07.

Independent Evaluation Office (2021), ‘IEO Evaluation of the Bank of England's Approach to Quantitative Easing’, January.

CGFS (Committee on the Global Financial System) (2019), ‘Unconventional Monetary Policy Tools: A Cross-Country Analysis’, BIS CGFS Paper No 63.

Cline W (2015), ‘Quantity Theory of Money Redux? Will Inflation be the Legacy of Quantitative Easing?’, Peterson Institute for International Economics Policy Brief No PB15-7.

Doherty E, B Jackman and E Perry (2018), ‘Money in the Australian Economy’, RBA Bulletin, September.

Dowling S and S Printant (2021), ‘Monetary Policy, Liquidity, and the Central Bank Balance Sheet ‘, RBA Bulletin, March.

Ferrari M, J Kearns and A Schrimpf (2017), ‘Monetary Policy's Rising FX Impact in the Era of Ultra-low Rates', BIS Working Paper No 626.

Ferrari M, J Kearns and A Schrimpf (2021), ‘Monetary Policy's Rising FX Impact in the Era of Ultra-low Rates', Journal of Banking and Finance, 129, Art 106142.

Finlay R, D Titkov and M Xiang (2022), ‘The Yield and Market Function Effects of the Reserve Bank of Australia's Bond Purchases’, RBA Research Discussion Paper No 2022-02.

Fitzpatrick R, C Shaw and A Suthakar (2022), ‘Developments in Banks’ Funding Costs and Lending Rates’, RBA Bulletin, March.

Gagnon JE (2016), ‘Quantitative Easing: An Underappreciated Success’, Peterson Institute for International Economics Policy Brief No PB16-4.

Glick R and S Leduc (2018), ‘Unconventional Monetary Policy and the Dollar: Conventional Signs, Unconventional Magnitudes’, International Journal of Central Banking, 56, pp 103–152.

Kaminska I and H Mumtaz (2022), ‘Monetary Policy Transmission during QE Times: Role of Expectations and Term Premia Channels’, Bank of England Staff Working Paper No 978.

Kent C (2022), ‘From QE to QT – The Next Phase in the Reserve Bank's Bond Purchase Program’, Speech at KangaNews DCM Summit, Sydney, 23 May.

Krishnamurthy A and A Vissing-Jorgensen (2016), ‘The Effects of Quantitative Easing on Interest Rates: Channels and Implications for Policy’, Brookings Papers on Economic Activity, Fall, pp 215–287.

Lane P (2021), ‘The New Monetary Policy Strategy: Implications for Rate Forward Guidance’, The ECB Blog, 19 August.

Lowe P (2019), ‘Unconventional Monetary Policy: Some Lessons from Overseas’, Speech at the Australian Business Economists Dinner, Sydney, 26 November.

Rebucci A, JS Hartley and D Jiminez (2022), ‘An Event Study of Covid-19 Central Bank Quantitative Easing in Advanced and Emerging Economies’, NBER Working Paper No 27339.

RBA (Reserve Bank of Australia) (2021), ‘Earnings, Distribution and Capital’, Reserve Bank of Australia Annual Report.

RBA (2022), ‘Review of the Yield Target’, June.

Rogers J, C Scotti and J Wright (2014), ‘Evaluating Asset-market Effects of Unconventional Monetary Policy: A Multi-country Review’, Economic Policy, 29(80), pp 749–799.

Sargent TJ and P Surico (2011), ‘Two Illustrations of the Quantity Theory of Money: Breakdowns and Revivals’, American Economic Review, 101(1), pp 109–128.

Wakefield M, L Delaney and R Finlay (2019), ‘A Cost-benefit Analysis of Polymer Banknotes’, RBA Bulletin, December.

Weale M and T Wiedladek (2016), ‘What Are the Macroeconomic Effects of Asset Purchases?’, Journal of Monetary Economics, 79, pp 81–93.