Asset Purchase Programs in European Emerging Markets

William Lindquist

doi:10.5089/9781513593753.087.A001

Jump to Content Jump to Main Navigation

Asset Purchase Programs in European Emerging Markets

Author:

William Lindquist

0000000404811396

https://isni.org/isni/0000000404811396 International Monetary Fund

Search for other papers by William Lindquist in
Current site
Google Scholar
PubMed

Publication Date:: 24 Sep 2021

ISBN:: 9781513593753

Language:: English

Keywords:: Monetary policy; Emerging markets; COVID-19 pandemic; Europe; APP announcement; asset purchase; EM monetary policy; market dysfunction; EM equity markets; Securities markets

Download PDF (1.3 MB)

Several emerging market central banks in Europe deployed asset purchase programs (APPs) amid the 2020 pandemic. The common main goals were to address market dysfunction and impaired monetary transmission, distinct from the quantitative easing conducted by major advanced economy central banks. Likely reflecting the global nature of the crisis, these APPs defied the traditional emerging market concern of destabilizing the exchange rate or inflation expectations and instead alleviated markets successfully. We uncover some evidence that APPs in European emerging markets stabilized government bond markets and boosted equity prices, with no indication of exchange rate pressure. Examining global and domestic factors that could limit the usability of APPs, in the event of renewed market dysfunction we see a potential scope for scaling up APPs in most European emerging markets that used APPs during the pandemic, provided that they remain consistent with the primary objective of monetary policy and keep a safe distance from the risk of fiscal dominance. As central banks in the region move towards monetary policy tightening, the tapering, ending, and unwinding of APPs must also be carefully considered. Clear and transparent communication is critical at each step of the process, from the inception to the closure of APPs, particularly when a large shock hits and triggers a major policy shift.

Abstract

Chapter 1 Introduction

Governments around the world have taken unprecedented policy actions to meet the challenge posed by the coronavirus disease (COVID-19) pandemic, ranging from imposing nationwide lockdowns to extending lifelines to individuals. Broadly, the policy actions have been larger in scale and more comprehensive in scope than in any crisis in living memory, including the global financial crisis (GFC) more than a decade ago. To cope with a virus that is blind to national or other boundaries, staple economic policies on fiscal and monetary fronts have been significantly enlarged and transformed in advanced economies (AEs) and emerging market economies (EMs) alike.

A particularly novel development in monetary policy has been the widespread adoption of asset purchase programs (APPs) by EM central banks, including in Europe. When international financial market turbulence led to financial outflows and upward pressure on bond yields in the weeks following the outbreak of the pandemic in March 2020, some 20 EM central banks launched APPs, most of them for the first time. In Europe, six EM central banks (Croatia, Hungary, Poland, Romania, Serbia, Turkey) began engaging in APPs during the March to May 2020 period. The stated primary goals were to alleviate the dysfunction in the financial markets, furnish liquidity, and repair impaired monetary transmission mechanisms. In the cases of Hungary and Poland, where APPs have continued beyond the initial financial market turbulence in early 2020, the objectives of APPs have since been tilted more toward supporting monetary policy transmission over a longer period. We observe that at an extraordinary time when government bond markets exhibited significant stress and large fiscal financing needs loomed, the APPs to some degree contributed to the smooth financing of government expenditures. In contrast, the APPs that were introduced in AEs following the GFC were mainly aimed at providing additional stimulus, as policy rates had reached the effective lower bound.

Until the pandemic, APPs had largely been the domain of reserve currency central banks, mainly because it was feared that APPs in EMs could de-anchor inflation expectations and trigger exchange market pressures. The global crisis, caused by a virus rather than policy mishaps, however, led EM central banks to deploy APPs successfully. To date, APPs in EMs have not resulted in the feared disruption and, instead, have stabilized markets since the initial lockdowns were put in place and improved monetary transmission channels (IMF 2020a).

This paper attempts to understand what has made the so-far successful use of APPs in European EMs possible and ascertain whether these “new” tools can be extended beyond the initial pandemic response. A close examination of country and global factors as well as institutional aspects, which allow for a case-study-type approach, distinguishes this paper from several other recent econometrics-based studies of EM APPs.

The paper asks the following questions and offers some preliminary answers:

What distinguishes APPs in EMs from those in AEs? The scale of APPs introduced in European EMs ranges from small (Romania, Serbia, Turkey) to more sizable (Croatia, Hungary, Poland), though all have been smaller in magnitude compared to those conducted by the Federal Reserve and the European Central Bank (ECB). The APPs have generally been fully or partially sterilized by various means. The generally smaller size and the use of some sterilization have set these APPs apart from the large-scale asset purchases, or quantitative easing (QE), adopted by the larger central banks since the GFC.
Have APPs achieved their aims in European EMs? They were successful in alleviating market dysfunction in the immediate aftermath of the global shock in March and April of 2020—a key shared reason for their deployment. Using event studies, we find some evidence of easing of liquidity pressures and a reversal in surge in term spreads, with no indication of foreign exchange market pressures. APPs seem to have had some economy-wide effects, with positive spillovers into equity markets.
Will EMs face tighter limits to their use of APPs? If so, what will determine such limits? EM monetary policy is influenced by global conditions, and APPs are no exception. Intertwined with global conditions, domestic factors will affect the extent to which APPs can be utilized as a policy tool in European EMs. For the EU members, the structure and discipline provided by membership make the oft-stated fiscal dominance risk less acute. Monetary policy frameworks in these countries enjoy credibility, and fiscal policy is anchored in medium-term sustainability.
When should APPs be terminated? How should the exit be managed? Slowing down and exiting from APPs remain a challenge, with little guidance available from the existing experience of the large central banks. APPs should stop when initial objectives have been met, are no longer relevant, or when their limits have been reached. The exit, especially the unwinding of purchased assets, should attempt to minimize market disruption and would likely be a long-term process.
What is the role of communication in using APPs? The novel nature of APPs in EMs makes effective communication particularly important. Central banks should communicate clearly at the inception of APPs, during operation, and at the time of exit and unwinding.
What should be the role of APPs in these European EMs going forward? With market conditions quite different than at the time of their introduction and success at alleviating market dysfunction in 2020, the initial goals of APPs have been met and several APPs have concluded. For the APPs that continue, it will be important to consider the role of APPs going forward and their interaction with eventual monetary policy tightening. It would be appropriate to complement an increase in policy rates with a tapering or discontinuation of APPs, which would support the transmission of higher policy rates to the longer end of the yield curve, supporting the monetary policy transmission mechanism.

The rest of the paper is organized as follows. Chapter 2 provides an overview of the characteristics, goals, and operations of APPs in EM Europe. Chapter 3 analyzes the effectiveness in meeting the stated goals. Chapter 4 studies the factors that could constrain the scope for APPs, including the risk of fiscal dominance. Chapter 5 tackles the questions of how to exit APPs and the importance of communication. Finally, Chapter 6 summarizes the lessons and offers some thoughts on how APPs might be used in future as part of the monetary policy toolkit for EM central banks in Europe.

Chapter 2 Characteristics of Asset Purchase Programs in European Emerging Markets

The widespread implementation of APPs in EM economies has been a novel aspect of the policy response to the COVID-19 crisis. Large-scale APPs became more common in AEs in the aftermath of the GFC in 2008, including in the form of QE, when policy rates already were close to the effective lower bound. While EMs did use some unconventional policy measures to provide liquidity after the GFC and the Taper Tantrum in 2013, the COVID-19 pandemic has led to the first broad EM foray into APPs in local currency bond markets (Arslan, Drehmann, and Hofmann 2020; Sahay and others 2014).

EMs implemented APPs to alleviate market dysfunction at the beginning of the pandemic. During the initial weeks of the pandemic, bond spreads spiked, and investors offoaded local currency EM instruments (Figure 1). It was widely believed that liquidity stress in the banking system could affect nonfinancial corporations, households, and the government, which would be facing shortfalls in income and revenue but were still expected to meet expenditures. In the event, some EM central banks began purchasing local currency bonds in a bid to restore the functioning of domestic government bond markets, as APPs are believed to be able to ease market liquidity (Christensen and Gillan 2019). Importantly, in contrast to APPs in AEs, EM APPs were generally not aimed at providing macroeconomic stimulus (IMF 2020a; Hofman and Kamber 2020; Ha and Kindberg-Hanlon 2021).

Conceptual Underpinnings

The theoretical underpinnings of APPs are vague and still evolving (see Annex 1). While conventional monetary policy is effective in the short term because it exploits near-term price rigidities, APPs are viewed to be effective by exploiting segmentation of the yield curve. This could in turn be attributed to asymmetric information after a shock and imperfect arbitrage, until consensus about a new equilibrium has emerged. Generally, APPs can (1) influence quantity by easing liquidity constraints—depending on the degree of sterilization—and by reducing segmentation among institutions and along the yield curve; (2) smooth the yield curve, including by mitigating the overshooting in a highly segmented yield curve; and (3) anchor expectations, mainly by conveying information about intended policies.

APPs can influence agents’ ability and willingness to borrow and lend, and thus help stabilize aggregate demand in the short term. How effectively APPs can influence the yield curve hinges on country- and shock-specific factors. What would matter initially is whether markets believe that APPs can effectively alleviate the shock by easing the aforementioned constraints. If APPs are mainly used to finance higher public spending, the efficiency of such spending, and the degree to which it is appropriately sterilized, will matter.¹ Otherwise, APPs could trigger macroeconomic instability, if inflation expectations become de-anchored. Over the long term, monetary policy is considered neutral, and it is difficult to persistently influence the yield curve, as deviations from market perceptions will in principle be arbitraged.

Scale and Timing

The APPs implemented by European EMs during the pandemic can mostly be characterized as small-scale, in line with other EMs. In total, some 20 EM central banks adopted APPs, including six in EM Europe: Croatia, Hungary, Poland, Romania, Serbia, and Turkey (IMF 2020a). Among EMs, the scale of assets purchased from March 2020 to June 2021 varies widely (Figure 2). Some European EMs implemented smaller-scale APPs (Romania, Serbia, Turkey), while others (Croatia, Hungary, Poland, 5 to 7 percent of GDP) implemented APPs that were at the higher end of those of EMs, globally.² More broadly, the size of APP in EMs has been significantly smaller than that conducted by the Federal Reserve and ECB, which purchased 23 and 14 percent of GDP in assets, respectively.³

The bulk of EM central bank asset purchases in Europe had paused by the mid-2020, Hungary and Poland being notable exceptions. Asset purchase activity peaked in April and May 2020 and then fell of sharply. By June/July 2020, purchases in Croatia, Serbia, and Turkey are known to have ceased.⁴ The National Bank of Poland (NBP) has continued asset purchases in 2021, initially at a very slow pace after July 2020. In March to May 2021, however, the NBP again increased the pace of asset purchases in the context of a global increase in long-term bond yields. Romania’s central bank also made small additional purchases in March 2021 after having stopped in August 2020. In contrast, the Hungarian National Bank’s (MNB’s) purchases accelerated in the second half of 2020 and remained at a high level in 2021 (Figure 3).⁵

Goals

A common objective of European EM APPs was the mitigation of financial market dysfunction, provision of liquidity, and repairing impaired monetary transmission mechanisms early in the pandemic. During the financial market shock at the beginning of the pandemic, European EM central banks recognized a liquidity squeeze, likely to be compounded by capital outflows and higher long-term yields. They quickly provided additional liquidity, including through intervention in government securities markets. Once calm was restored to financial markets, most regional central banks paused or significantly scaled back APPs.

The Polish APP has aimed to support the monetary policy transmission mechanism beyond the initial pandemic financial market shock. Following large purchases during the first months of the pandemic, the central bank purchased additional assets at a slow pace from August 2020 to February 2021. However, from March to May 2021, the central bank significantly increased the scale of APPs. As this period coincided with a global increase in long-term bond yields driven by expectations of economic recovery, particularly in the United States, the increased purchases appear to have been an effort to prevent a premature tightening of long-term yields in Poland. In other words, the purchases aimed to contain long-term bond yields and bolster the transmission of monetary easing. These actions were consistent with the initially stated goals but were unique from most other EM European APPs in that they occurred well past the period of financial market turbulence.

The MNB has also revised the goals of its APP (Box 1). Hungary was unique in that the MNB had implemented APPs even before the pandemic, particularly for private securities, with the goal of promoting market development. At the start of the pandemic, the MNB initially cited market dysfunction and a desire to lower long-term interest rates upon introducing the APPs. The pace of asset purchases was then increased later during the pandemic.

The purchases of government securities may also have eased financing costs and lowered refinancing risks of the government. Initially, the long-term collateralized credit facility introduced at the beginning of the pandemic played a relatively large role in promptly providing liquidity. Later, the APP partially took over this role, as the MNB found that purchases of government securities were more efficient in influencing the transmission to long-term interest rates.

APPs also have facilitated the smooth financing of anticrisis fiscal expenditures. There appears to be some correlation between the size of the increase in budget deficits and size of APPs (Figure 4). It may be fair to conclude that central bank demand for domestic Treasury instruments may have eased the placement of new securities in the markets, lowering both funding costs and rollover risks. For example, in Poland, the NBP cumulatively purchased about 36 percent of the combined issuance of domestic Treasury bonds/ bills and state-guaranteed debt between March 15 and end-2020 (Figure 5). In Croatia, the central bank made five large bond purchases early in the pandemic, and markets quickly calmed thereafter, assuaged by the EUR 2 billion swap agreement with the ECB in mid-April 2020, and the authorities’ strong commitment to the exchange rate anchor and envisaged ERM II membership (which happened on July 10, 2020).

Operational Aspects

As in other EMs, assets purchased in EM Europe mostly consist of local Treasury bonds. In Poland, the central bank’s purchases have been evenly split between local Treasury securities and government-guaranteed securities issued by Polish development banks.⁶ In Hungary, the MNB has also purchased a sizeable amount of mortgage-backed securities and corporate bonds (Figure 6). These purchases of private securities affect long-term yields and can thus enhance the monetary policy transmission mechanism.⁷ Such purchases, however, pose additional challenges for central banks, including assessing the perceived maturity and credit risk. They may also create governance risks if the central bank is seen as favoring some corporates over others, particularly if securities are acquired in the primary market.

Central banks have purchased government securities primarily at the long end of the yield curve. Granular information on bond-by-bond purchase of Treasuries is available for Poland and Hungary only. The data show that the NBP concentrated its purchases on 5- to 10-year bonds, while the MNB overwhelmingly purchased Treasury securities with greater than 10-year maturity (Figure 7).

Most purchases were in the secondary market. EU central banks with APPs are required to buy government securities only in the secondary market.⁸ In principle, such practices minimize the impact of central bank participation on price discovery in the market. In Hungary, the MNB bought mortgage bonds and corporate bonds, initially in both the primary and secondary markets. It later decided to pause purchases of mortgage bonds and announced that it would only buy green mortgage bonds.

Purchase methods have varied, but a quantity-based approach was preferred. In Croatia, the central bank conducted five large ad hoc tenders. In Hungary, regular auctions are conducted, but a smaller part are over-the-counter transactions depending on market conditions. In Poland, biweekly or monthly tenders are conducted. In Romania, only targeted bilateral purchases are conducted with a view to alleviate the liquidity position of specific counterparts. In line with good practices (Adrian and others, forthcoming), the APPs were generally quantity based to influence interest rates rather than price based, which could effectively result in interest rate caps.

Interaction of Asset Purchase Programs with Other Monetary Policy Tools

APPs have been implemented alongside conventional monetary easing, though the scope for conventional interest rate easing may not have been fully exhausted. In the aftermath of the GFC, several AE APPs (such as the Federal Reserve’s Large-Scale Asset Purchase Program) aimed to further ease monetary conditions once conventional monetary tools had been largely exhausted (that is, with rates near the effective lower bound). In contrast, the implementation of APPs in EM Europe preceded the full use of conventional monetary policy space (Figure 8). In Hungary, monetary operations were revamped at the beginning of the pandemic to effectively increase the money market rates to support the currency, while initially maintaining the policy rate, which later was lowered. In Poland, the NBP reduced its policy rate to near zero, but only after initiating the APP. In Turkey, asset purchases were initially accompanied by policy rate reductions, but the policy rate increased later in 2020 in response to currency pressures and declining foreign exchange reserves. In June and July 2021, the Hungarian Central Bank increased its policy interest rate and began to phase out some of its crisis measures but continued its purchases of government securities and commercial bonds. Upon raising rates again in August 2021, the Hungarian Central Bank also announced that it would begin to gradually reduce its purchases of government bonds (see Box 1).

Collateralized lending arrangements have also been employed alongside APPs. In Poland, the NBP provided repo transactions to provide liquidity to banks and also introduced a discount credit facility to refinance bank loans extended to firms, though both facilities were used only sparingly. The Hungarian Central Bank has been active in expanding lending facilities, introducing a long-term collateralized lending facility with a maximum five-year maturity to support banks’ liquidity management as well as broadening the scope of eligible collateral.

Asset purchases are not the only factor explaining the expansion of EM Europe central bank balance sheets. Earlier in the pandemic (end-June 2020), the overall size of the balance sheet of the Croatian and Romanian central banks had stayed broadly stable, as domestic asset purchases were offset by the sale of foreign exchange reserves (Figure 9). By end-May 2021, these central banks’ foreign exchange reserves had recovered, and asset purchases did coincide with overall balance sheet expansions.⁹ In the case of Hungary, significant central bank collateralized lending to the banking sector was also a major factor in balance sheet expansion. From the beginning of the pandemic to end-May 2021, the expansion of MNB’s balance sheet has also been large (20 percent of GDP), at par with the expansion in the Federal Reserve and ECB (18 and 26 percent of GDP, respectively). Only about one-third of MNB’s balance sheet expansion can be tied to securities purchases.

For the central banks in EM Europe, which do not issue reserve currencies, sterilization of APPs has been a prominent feature, ostensibly helping with anchoring inflation expectations. In contrast to QE by reserve currency issuing central banks, most of the APPs in EM Europe have been accompanied by some sterilization—through which the additional bank reserves resulting from purchases are drained through other operations. Sterilization has taken several forms. The purchase of new domestic assets (that is, bond purchases) was offset by the sale of foreign assets in Croatia, thus preventing the central bank’s balance sheet from growing. Alternatively, some central banks issued their own securities (Poland; Figure 10) or used a deposit facility to drain bank reserves. In these cases, the central bank’s balance sheet grows, but there is a substitution of liabilities to reduce bank reserves.¹⁰ Sterilization is consistent with the narrow goals of the APPs, which are to alleviate market dysfunction but not to ease overall monetary conditions. Sterilization may also be intended to signal a continued commitment to primary monetary policy objectives, including medium-term price stability, which can be important for EM central banks employing new tools. This point underlines that the purpose of APPs is not always to increase the provision of overall liquidity, but rather to overcome segmentation in the money market or repair the impaired transmission mechanism.

Impact on Target Markets

APPs have greatly expanded central bank holdings in domestic government bond markets, cushioning outflows from investment funds in some cases. Prior to the pandemic, central banks in the region played a limited role in local Treasury markets. They generally held little or no Treasury securities on their balance sheets, except as collateral for underlying instruments for repos. By May 2021, however, their market share of outstanding local Treasury securities had increased to between 6 percent in Turkey and 9 percent in Poland (Figure 11). These purchases played an important role in stabilizing markets in the context of a large increase in new issuance during the pandemic. In Poland and Turkey, the scale of central bank purchases broadly offset outflows from foreign investors. Central bank purchases supplemented continued strong demand from domestic banks in Serbia and other domestic investors in Hungary (for example, special retail bonds for households; Figure 12).

APPs may affect the market liquidity of the targeted securities. The liquidity premium of a financial asset, like a bond, reflects how easily it can be traded without large price movements not merited by new information. The higher daily trading of a bond, the more liquid it typically is. If central banks purchase large amounts of a particular bond, they could potentially become less liquid and thus instead increase funding costs of the government. Both Croatia and, initially, Hungary stated that they would not buy more than a third of each bond series,¹¹ but Hungary later abandoned this limitation. On the other hand, the presence of a committed large buyer (like the central bank) improves “price discovery,” which may be reflected in smaller bid-ask spreads.

^{Box 1.}

The Evolution of the Asset Purchase Programs of Hungary

Hungary is particularly interesting because the central bank (Hungarian National Bank [MNB]) has continued to expand its APPs. Hungary is also unique in that it had employed asset purchase programs (APPs) prior to the pandemic, although not in government securities. The communication has evolved from a cautious beginning with a view to mainly mitigate market dysfunction and lower long-term interest rates,^¹ to more focus on funding costs and refinancing risks of the government, as financial markets stabilized.

On April 7, 2020, the MNB announced that it would (1) introduce an APP for government securities, (2) reactivate its mortgage bond purchase program used in 2018, and (3) ease conditions of its prepandemic purchases of corporate bonds. The main objectives of the APPs were “to restore the stable liquidity position of the government securities market” (MNB 2020b) and “to strengthen monetary policy transmission” (MNB 2020d).

In August 2020, the MNB explicitly mentioned “the higher government financing needs” and that it would “continue to make purchases in the long segment to support an extension in the maturity structure of government debt” (MNB 2020c). The MNB has clearly communicated and done so-called technical reviews before the envelopes of the APPs were expanded. While the MNB launched its APPs of government securities later than peers, it has continued and intensified its purchases, while others have slowed them.

In January 2021, the MNB stated that it would gradually reduce its long-term collateralized lending facility and rely more on its APP in government securities—as its transmission appeared to be more effective—as well as to buy government securities with less than 10 years maturity (MNB 2021a). In April 2021, the MNB announced that it would include so-called green bonds issued by the government in its asset purchases. In June and July 2021, the Hungarian Central Bank increased its policy interest rate and began to phase out some of its crisis measures but continued its purchases of government securities, which is considered “crucial in its set of monetary instruments” (MNB 2021b). The next review of the APP of government securities will take place when the current envelope is close to HUF 3000 billion (about 5.7 percent of projected GDP), which is likely to happen in mid-August 2021. On August 24, 2021, the Monetary Council decided to “begin gradually withdrawing the government securities programme while considering aspects of maintaining market stability,” noting that it would not set a limit applicable to the entire stock purchased but rather “set a target amount for weekly purchases” (MNB 2021c).

^¹ The MNB tried to lower long-term yields in 2017–18. In 2018, it introduced unconditional interest swaps—by swapping long-term interest rate risk of banks with short-term interest rate risk. The announcement in the fall of 2017 had an immediate impact: long-term interest rates of Hungarian government bonds declined vis-à-vis peers. The impact gradually evaporated, as auction procedures were amended to restrict participation to Hungarian licensed banks. The program was stopped at end-2018.

Chapter 3 Effectiveness of Asset Purchase Programs

Did the APPs implemented by European EMs succeed, in line with their goals of alleviating financial market dysfunction and improving market liquidity and monetary policy transmission? This section studies effectiveness by assessing the impact of APPs on domestic financial markets in early 2020 by focusing on government bond yields, term spreads, and liquidity conditions. We also assess the extent of spillovers from these programs into other asset classes, including exchange rate, equity markets, and corporate credit default swaps. The empirical analysis uses event study and local projections analysis to assess the presence of systematic responses across the sample countries. The assessment is limited to quantifying overall impacts rather than identifying specific transmission channels that could explain the findings.

Related Literature

The emerging literature on APPs in EMs in response to the COVID-19 crisis finds that the programs had an immediate positive impact. Central banks’ purchases of government bonds reduced long-term government bond yields immediately after APP announcements (see Table 1). Some of the papers find that APPs also reduced term spreads. Regarding spillovers to other asset prices, some papers find that APPs did not induce exchange rate depreciations, and one paper finds that equity markets in EMs improved a few days after the announcements. These recent studies build on a large body of literature that has analyzed the effectiveness of unconventional monetary policies (UMP) mainly focusing on the experience of AEs (Bhattarai and Neely 2016, forthcoming) over the last decade.

^{Table 1.}

Summary of Recent Studies of Asset Purchase Programs’ Impact in Emerging Markets

Source: Authors. Note: APP 5 asset purchase program; bps 5 basis points; EM 5 emerging market; EMBI 5 Emerging Market Bond Index.

^{Table 1.}

Summary of Recent Studies of Asset Purchase Programs’ Impact in Emerging Markets

Author(s)	Sample	Impact on 10-year Government Yields	Window of Time		Other Asset Prices	Cont	rolling for Other Variables
Hartley and Rebucci (2020)	13 EMs	[20.28, 20.43] bps	1–3 days	No		No
Arslan, Drehmann, and Hofmann (2020)	9 EMs	[210, 250] bps	1–5 days	Yes	The announcements appear to have shored up the exchange rate.	Yes	Estimated immediate announcement effect on yields is very similar while the estimated negative effect after five days is somewhat smaller, at 25 basis points. Exchange rate results also hold in a panel regression setup.
Sever and others (2020)	11 EMs	235 bps	Next few days after the announcement	Yes	Median-term premiums reduced by almost 20 bps in a week after the announcement. The impact on currencies was relatively limited. EM equity markets also improved a few days post the announcements.	Yes	Yield was 35 bps lower than the preannouncement level, even on the sixth day after the announcement. APP announcements were not followed by significant depreciations in local currencies. Domestic APP announcements did not have much impact in the very short term but started to have somewhat a statistically significant positive effect in later days.
Fratto and others (2020)	15 EMs	On average, the estimated effect is statistically significant and broadly consistent with that found in the papers listed previously. Results continue to hold when the authors exclude from the sample the announcements that coincide with a policy rate cut.		Yes	Impact on the exchange rate depends on whether the APP announcement was made within a few days of a policy rate cut. APP announcements have predominantly a positive and statistically significant effect on the EMBI, although with a significant heterogeneity across the sample.	Yes	The event study with control variables as well as the panel regression, presented in Annex 4, broadly confirm the previous findings.

Source: Authors. Note: APP 5 asset purchase program; bps 5 basis points; EM 5 emerging market; EMBI 5 Emerging Market Bond Index.

^{Table 1.}

Summary of Recent Studies of Asset Purchase Programs’ Impact in Emerging Markets

Author(s)	Sample	Impact on 10-year Government Yields	Window of Time		Other Asset Prices	Cont	rolling for Other Variables
Hartley and Rebucci (2020)	13 EMs	[20.28, 20.43] bps	1–3 days	No		No
Arslan, Drehmann, and Hofmann (2020)	9 EMs	[210, 250] bps	1–5 days	Yes	The announcements appear to have shored up the exchange rate.	Yes	Estimated immediate announcement effect on yields is very similar while the estimated negative effect after five days is somewhat smaller, at 25 basis points. Exchange rate results also hold in a panel regression setup.
Sever and others (2020)	11 EMs	235 bps	Next few days after the announcement	Yes	Median-term premiums reduced by almost 20 bps in a week after the announcement. The impact on currencies was relatively limited. EM equity markets also improved a few days post the announcements.	Yes	Yield was 35 bps lower than the preannouncement level, even on the sixth day after the announcement. APP announcements were not followed by significant depreciations in local currencies. Domestic APP announcements did not have much impact in the very short term but started to have somewhat a statistically significant positive effect in later days.
Fratto and others (2020)	15 EMs	On average, the estimated effect is statistically significant and broadly consistent with that found in the papers listed previously. Results continue to hold when the authors exclude from the sample the announcements that coincide with a policy rate cut.		Yes	Impact on the exchange rate depends on whether the APP announcement was made within a few days of a policy rate cut. APP announcements have predominantly a positive and statistically significant effect on the EMBI, although with a significant heterogeneity across the sample.	Yes	The event study with control variables as well as the panel regression, presented in Annex 4, broadly confirm the previous findings.

Source: Authors. Note: APP 5 asset purchase program; bps 5 basis points; EM 5 emerging market; EMBI 5 Emerging Market Bond Index.

Findings of Event Study Analysis

Our event study analysis documents the high-frequency evolution of the variables of interest around the APP announcement dates in European EM countries. Event study methodology has been commonly used to study the impact of UMP in AEs and EMs alike. We examine all publicly available APP announcements in European EM economies from March 2020 onwards (see Annex 3 for details). The content of announcements differed across countries and over time (see Table 2). Initial APP announcements were commonly a part of a broader policy package with policy rate cuts, new and/or expanded liquidity facilities, and foreign exchange interventions as accompanying measures.¹ There were also differences in the scope of the announced APPs (see Chapter 2). The event window is defined as three days around the announcement date. Impact on the government bond market is assessed by studying bond yields with maturities of 1, 3, 5, and 10 years. For spillovers into other asset prices, we focus on exchange rate and equity price channels. Findings are reported in terms of the median response and corresponding interquartile range for the 11 announcement events. In addition, the median response based on the initial announcement event in each country is highlighted separately to assess the difference between initial versus all announcements.²

^{Table 2.}

Initial Announcement Dates for Sample Emerging European Countries

Sources: Croatian National Bank; National Bank of Poland; National Bank of Romania; Hungarian National Bank; and Central Bank of the Republic of Turkey. Note: For Hungary, 3M money market rate used to assess policy rate. For Serbia, no data on announcement or implementation dates were available. The table uses International Organization for Standardization (ISO) country codes.

^¹The base rate was later reduced in two steps (June 24, 2020, and July 22, 2020) from 0.90 to 0.60 percent.

^{Table 2.}

Initial Announcement Dates for Sample Emerging European Countries

	Initial APP announcement date	Policy rate cut	Expanded liquidity facilities	Reduction in bank reserves	Macropru loosening	FXI
HRV	13-Mar	no	yes	yes	no	yes
POL	16-Mar	yes	yes	yes	yes	no
ROM	20-Mar	yes	yes	no	no	yes
HUN	7-Apr	no^¹	yes	no	no	no data
TUR	31-Mar	no	yes	yes	no	no data

^¹The base rate was later reduced in two steps (June 24, 2020, and July 22, 2020) from 0.90 to 0.60 percent.

^{Table 2.}

Initial Announcement Dates for Sample Emerging European Countries

	Initial APP announcement date	Policy rate cut	Expanded liquidity facilities	Reduction in bank reserves	Macropru loosening	FXI
HRV	13-Mar	no	yes	yes	no	yes
POL	16-Mar	yes	yes	yes	yes	no
ROM	20-Mar	yes	yes	no	no	yes
HUN	7-Apr	no^¹	yes	no	no	no data
TUR	31-Mar	no	yes	yes	no	no data

^¹The base rate was later reduced in two steps (June 24, 2020, and July 22, 2020) from 0.90 to 0.60 percent.

We find that APP announcements significantly improved conditions in the government bond markets. The period leading up to initial announcements saw a systematic increase in long-term bond yields.³ Liquidity conditions in the bond market, as proxied by bid-ask spreads on government bonds, also tightened prior to initial announcements (Figure 13, panel 4). Following the announcements, long term yields fell for the median response by 0.3 ppts by t=1 and remained below the t=0 levels in subsequent days (Figure 13, panel 1). Given that APPs tended to target bonds with longer maturities, term spreads and changes in the slope of the yield curve can be more directly linked to the impact of APP purchases. Consistently, the yield curve fattened following the initial APP announcements (Figure 13, panel 2), while bond yields declined across the yield curve—the median decline in yields for the 1-year maturity was 0.1 ppts, compared to the 0.3 ppts decline for the 10-year bond. Term spreads and bid-ask spreads also declined (Figure 13, panels 3 and 4), although the gains in this regard were smaller in magnitude and less persistent. Broadening the analysis to all announcements confirms the decline in yields, the fattening of the yield curve and a temporary improvement in liquidity conditions. However, the size of the median response shrinks and the impact on term spreads disappears, consistent with the heightened role played by the initial announcement in each country. Additional results from the impact of APPs on government bond markets are provided in Annex 3.

APPs did not induce a significant exchange rate response. Following the announcements, the median exchange rate response, including for initial announcements, stayed broadly stable (Figure 14).⁴ This finding could partly reflect the foreign exchange interventions and other forms of monetary policy sterilization that mitigated depreciation pressures around the APP announcement dates.⁵ Lack of exchange rate depreciation is more notable, given the accompanying reductions in policy rates in the case of Poland and Romania (see Table 2).

Stock markets responded positively to APP announcements. The response of equity prices to the announcement events is gauged by examining the dynamics of main stock market index in each country. Results reveal a gradual increase in equity price indexes, reaching 2 percent by the fourth day following the announcement. For the initial APP announcements, the median response was more pronounced at 4 percent. Analysis of corporate credit default swaps (not reported) did not reveal any systematic impact from APP announcements, which could be due to the novelty of APPs in EMs and the uncertainty about the transmission mechanism of APPs. Annex 3 reports sectoral results for the stock market response to the crisis, zooming in on the financial sector and small-cap companies.

Local Projections Analysis

Local projection analysis allows us to assess if the findings of the event study are robust to potential influences from other domestic or external factors. A key shortcoming of the event study analysis is its lack of control for factors other than the APP announcements in explaining the APP impact. In this section, we control for an additional domestic variable—domestic policy rate cuts—and global factors—namely, the Chicago Board Options Exchange Volatility Index (VIX) or the APP announcement by the Federal Reserve (see Arslan, Drehmann, and Hofmann 2020; Sever and others 2020). Details of the econometric specification are provided in Annex 3.

The results suggest that after controlling for other factors, APP announcements in European EM countries were associated with drops in bond yields, though we did not find systematic impact on term spreads. There was a statistically significant impact on the 10-year government bond yield one to three days after the announcement of around –35 basis points (Figure 15). The results are similar when using the Federal Reserve APP announcement. We also used the model to assess the effect of domestic APP announcements on term spreads, used as a proxy for term premiums—the dependent variable in this case is the cumulative change in the difference between the 10-year government yield and the 1-year government yield.⁶ This finding for the sample of 11 announcement events is consistent with the event study results, which revealed no systematic deviations in the term spread following the announcements (Figure 16; see Figure 13, panel 3).

The estimated impact of APP announcements on equity prices is consistent with the event study results, but its statistical significance varies with the control used (Figure 17). Our local projection analysis uses cumulative change in the country’s equity price index as the dependent variable. The magnitude of the estimated response—at 1–3 percentage points over the three-day horizon—and its lagged nature—with the impact increasing in days two and three—are consistent the results from the event study. However, the response is not statistically significant when we employ VIX as the control for global factors.

Chapter 4 Limits to Emerging Market Asset Purchase Programs

As EMs have employed APPs since the outbreak of the pandemic, a natural question arises regarding the limits of APPs in EMs. In the aftermath of the GFC, similar questions were frequently raised in AEs, including the extent to which central banks could expand balance sheets without risking a large future increase in inflation. Such fears have not materialized. Unlike AEs, EMs do not issue reserve currencies or benefit from global demand for reserve assets, and they do face the risk of currency substitution. Given the mixed past experiences of some EMs with monetary financing—for example, the instability associated with prices, exchange rate, and economic stability— the risks and limits of APPs in EMs are worth exploring.

Rather than attempting to quantify specific limits to APPs in EMs, this chapter analyzes the factors that may act as constraints on individual central banks’ ability to employ this tool. The constraints on the utility of APPs as a monetary policy tool are likely to vary over time and across countries. The nearly universal easing of monetary policies during the deep, synchronized global recession resulting from the pandemic created more conducive market conditions for EMEs to employ such tools than during normal times. For example, asset purchases that reinforce a synchronized push to lower global bond yields are not likely to elicit negative market reactions such as currency depreciation. Similarly, individual country circumstances are likely to affect market perceptions of APPs, making them a more feasible tool for some countries than others. The apparent effectiveness of APPs in EMs during the pandemic—a relatively short period with somewhat exceptional circumstances—does not guarantee a similar experience on an open-ended basis or under different conditions.

Fiscal Dominance as a Constraint on Asset Purchase Programs

The risk of fiscal dominance acts as a general constraint on APPs in EMs. Broadly defined, fiscal dominance refers to pressure on the central bank to subordinate its objectives to those of the government (Adrian and others, forthcoming). In the extreme, fiscal dominance can take the form of a central bank financing the government directly over an extended period, but it can also refer to explicit or implicit limitations on a central bank’s ability to adjust monetary policy to meet its objectives. APPs could facilitate fiscal dominance if they help support fiscal deficits that cannot be met by conventional means, including borrowing at market rates, growth in economic activity, and/or future fiscal reforms.

The past experiences of macroeconomic instability in some EMs during episodes of monetary financing of governments serve as a reminder of how APPs can run astray. Large fiscal deficits in several large EMs in the 1980s were financed in part by their central banks and eventually led to high inflation, contributing to a prolonged period of macroeconomic instability and depressed growth (Ha and Kindberg-Hanlon 2021).

The Importance of Macroeconomic Fundamentals

The significant improvement in macroeconomic fundamentals in EMs over the past couple of decades has helped lower the risk of APPs supporting fiscal dominance. The instability of the 1980s and 1990s, including high levels of inflation, had largely been surpassed by the early 2000s, partly because of improved macroeconomic policies and institutions (IMF 2001). The enhanced track record of macroeconomic stability and strengthened economic institutions have bolstered the scope for APPs in many EMs. Specific factors include the following:

Fiscal sustainability: Fiscal sustainability is key to achieving and maintaining macroeconomic stability. Large and persistent fiscal deficits in combination with high levels of public debt have often preceded debt monetization.
Domestic financial market development: EMs with deep, broad, and well-functioning domestic bond markets provide central banks with a larger pool of securities to purchase, potentially expanding the scope for secondary market APPs. On the other hand, a sizable presence of state-owned banks may undermine the scope for market-based APPs, raising concerns about coordination between the finance ministry, state banks, and the central bank to bypass market-based placement of debt in the primary market and subsequent secondary market asset purchases.
Monetary policy frameworks and credibility: Well-anchored inflation expectations, underpinned by a credible monetary policy regime, expand the room for central banks to engage in APPs without de-anchoring inflation expectations or triggering capital outflows/currency depreciation. Specific legal provisions in the central bank mandate, such as prohibitions on direct financing of governments, can further bolster credibility. At the same time, painstakingly established credibility can be lost, if APPs are perceived as conflicting with well-established monetary policy objectives.
Growth prospects and external balance: Strong economic growth prospects facilitate the stabilization or reduction of government debt without the need for leaning on the central bank for support. The absence of large current account deficits and strong foreign exchange reserve coverage also reduce vulnerability to capital flow volatility. Such strong macroeconomic fundamentals may increase market participants’ confidence that APPs will not be associated with excessive pressures on the exchange rate, thus increasing the scope for APPs.

Macroeconomic Fundamentals in European Emerging Markets with Asset Purchase Programs

Underlying fiscal strength varies among the European EMs that have implemented APPs during the pandemic. European EMs are benchmarked against a sample of 22 EMs worldwide and the euro area average, where applicable (Figure 18). Prior to the crisis, Romania and Turkey stood out with relatively large fiscal deficits in 2019. On the other hand, the relatively low level of general government debt in Turkey and Romania, compared to peers, partly attenuates concerns about fiscal sustainability. While general government debt is relatively high in Croatia and Hungary, it is projected to decline over the medium term. EU membership, with its fiscal rules and institutions, as well as the availability of additional fiscal resources to members, is an additional factor that bolsters fiscal sustainability.

The level of domestic financial market development may also affect the scope for APPs. Domestic government securities markets are sizable in Croatia, Hungary, and Poland relative to EM peers, whereas they are relatively thin in Romania, Serbia, and Turkey (Figure 19), suggesting a limited scope for APP expansion in the latter set of countries. However, with the implemented size of APPs reaching only a fraction of outstanding domestic government securities, none of these countries currently faces a binding constraint from the supply of domestic securities available to purchase.¹ On average, the degree of state ownership of the banking sector in the sample of EM European countries is limited and compares favorably with the EM benchmark. Nevertheless, the relatively heavy state ownership of the banking sector in Poland and Turkey could trigger unease among market participants should APPs expand significantly, if the involvement of state banks in the APPs was perceived as less than transparent.

Inflation expectations remain well anchored in the region. End-of-period inflation in 2020 fell within target ranges and was below the EM average, except for Turkey, where inflation significantly exceeded the target range at 14.6 percent (Figure 20). As of end-2020, there is no systematic evidence that long-term inflation expectations have increased in the wake of APPs, relative to prepandemic levels. European EM central banks’ decisions to sterilize APPs may have helped maintain the credibility of monetary policy frameworks. The central banks of EU members in the sample, along with Serbia, may have also benefited from implementing institutional safeguards, such as de jure central bank independence and prohibitions on direct financing of government entities.²

For most of the European EMs that have undertaken APPs, good medium-term growth prospects and strong external positions point to low precrisis vulnerabilities. Economic growth prospects over the medium term for these countries are sound, with lower projected growth in Hungary and Poland correlated with their higher income levels (Figure 21). External imbalances and associated vulnerabilities vary. While Romania and Serbia (and Turkey prior to 2019) exhibited relatively large prepandemic current account deficits, financing sources differed. Romania and Turkey depended more on portfolio inflows, whereas Serbia received large other investment inflows. Both sources can be volatile and subject to reversal during crises. Foreign reserve coverage is robust, though foreign reserves in Turkey fall short of the IMF’s reserve adequacy metric. EU member states may also benefit from potential access to swap or repo lines with the ECB. Turkey’s vulnerabilities have been apparent through the depreciation of its currency during the pandemic, in contrast to other EMs in the region.

In summary, the degree to which macroeconomic fundamentals likely constrain the scope for APPs varies within the region (Figure 22). Based on a combination of external vulnerabilities, precrisis fiscal weakness, and relatively high inflation, Turkey is likely to find its use of APPs as an ongoing policy tool more constrained than others in the region. In comparison to most other EMs, several European EMs have more developed policy frameworks and institutions, bolstered by EU members. Inflation-targeting frameworks also have a strong track record. Nevertheless, the risk of fiscal dominance is not completely eliminated and should be monitored.

Chapter 5 Complementary Policies and Communication

Additional Unconventional Monetary Tools

Adjusted Lending Operations

Beyond APPs, liquidity constraints can also be effectively mitigated through lower reserve requirements, broadened eligible collateral, and adjusted lending operations. European EMs that implemented APPs also took other complementary measures to overcome liquidity stress. Some lowered the reserve requirement ratio (Croatia from 12 to 9 percent and Poland from 3.5 to 0.5 percent), or temporarily waived sanctions for noncompliance with reserve requirements (Hungary from March to September 2020). The MNB also promptly broadened eligible collateral (by almost 5.5 percent of GDP) to include performing corporate bank loans with a standard haircut of 30 percent, irrespective of maturity and currency. To the extent the concern was lack of adequate long-term liquidity, Croatia conducted a five-year structural repo transaction, Poland introduced a credit rediscount facility, and Hungary began to conduct regular tenders of collateralized long-term loans with maturities of 3, 6, and 12 months, and three and five years at the policy rate. However, APPs differ from most other liquidity providing tools by the central bank becoming more prone to maturity and interest rate risk due to the longer maturities of the purchased assets, and even credit risk for mortgage bonds but particularly for corporate bonds.

Negative Policy Rates

Negative interest rate policies (NIRPs) aim to lower the level of the yield curve but may have limited use in European EMs. The first AEs to introduce NIRPs did so to mitigate appreciation pressures (Brandão-Marques and others 2021). Hungary has long maintained a negative overnight deposit rate (–5 basis points), Croatia at zero, whereas Poland lowered the deposit rate to 0.1 percent in late May 2020. Interestingly, most of our sample countries have had rather low or even negative real interest rates (Figure 23). A host of factors argue for there being a higher risk premium in EMs than in reserve currency issuing AEs and thus a more limited role for NIRPs without adversely affecting the exchange rate. EMs are typically reliant on foreign funding and their households and even firms tend to have a stronger preference for currency in circulation compared to bank deposits. Many EMs typically have a relatively shorter track record of good macroeconomic policy implementation compared to AEs.

Forward Guidance

Forward guidance is now an intrinsic part of monetary policy, but there are limits. Forward guidance firms up expectations by reducing the noise about both the future discount factor as well as expected future cash flow. This is particularly relevant at low interest rates where there is little scope for conventional tools. However, EM reliance on volatile capital flows and the attendant exchange market pressures tend to limit the effectiveness of forward guidance as a commitment device, mainly because the central bank may need to break “its promise” (the “time-inconsistency” challenge) on adjusting the interest rate.

APPs can be seen as akin to forward guidance, but one where the central bank puts its money where its mouth is. APPs may be more credible than forward guidance, as they alleviate the time-inconsistency challenge that central banks face. If a central bank, despite its previous guidance, changes policy in such a way that long-term rates increase, it will in principle realize a loss (for example, Clouse and others 2003; Bhattarai, Eggertsson, and Gafarov 2015), depending on how the securities are booked.¹ Evidently, the better the communication, the more APPs can reduce risks and uncertainties and ultimately smooth the adjustment to the new equilibrium. The European EMs that employed APPs have generally provided both sound guidance and information about the APPs, including in the form of press releases and inflation reports.

Complementary Policies

Capital flow management (CFM) and prudential measures could complement APPs in alleviating market dysfunction, but there are limits to deploying them. In principle, capital will flow to where it is most needed (highest expected returns). Moreover, free capital mobility is really the litmus test of whether the policy mix is sustainable. Nevertheless, special conditions—such as an uneven playing field or asymmetric shocks—could validate temporary CFM measures, provided that they are temporary, transparent, and do not discriminate by residency status (1) to contain capital inflows, when options for macroeconomic policy adjustments are limited; and (2) to contain disruptive outflows in case of crisis and imminent risk of a crisis (but without delaying any needed macroeconomic policy adjustments) (IMF 2012, 2018). Macroprudential measures are distinct to CFMs, as they are motivated by ensuring financial stability to mitigate capital flow cycles. Temporary release of buffers to absorb losses or to tighten limits on foreign exchange exposures are appropriate (IMF 2017; Nier and Olafsson 2020). Hungary, for instance, tightened foreign exchange funding limits of banks in March, which were reversed in September 2020. Regardless, the use of CFMs in European EMs is circumscribed—EU members are prohibited from deploying capital controls except in case of a national emergency.²

APPs and fiscal policies are interdependent. First, to the extent that APPs swiftly and successfully stabilize dysfunctional markets, and lower long-term yields of government securities, they obviously reduce funding costs and rollover risks of the government. Second, in principle, this permits a larger short-term fiscal boost than would otherwise be the case, provided the economy is below capacity. For instance, the government can accept risks that individuals may not be willing or able to take at times of distress. In principle, if the spending is perceived as being efficient, a larger deficit should be easy to finance, either with higher taxes³ or debt without increasing yields. In contrast, if spending is perceived as inefficient, then the risk of fiscal dominance becomes more important and limits the scope of APPs (as discussed in Chapter 4).

In principle, structural and institutional reforms also matter, both for the need and for the effectiveness of APPs. Good governance practices, which ultimately are reflected in the trust in the state, tend to reduce the risk of market dysfunction and noise in the financial transmission mechanism.⁴ Country-specific factors are thus critical for the choice of APPs, their potential effectiveness, and the needed communication. Structural and particularly governance reforms are often ignored because their positive impact is gradual, and their benefits are widely dispersed and difficult to internalize.

Communication

A transparent and well-informed approach can better mitigate and prepare for any potential losses associated with the exit process and potential “taper tantrums.” In the initial stage of implementation of APPs in European EMs, the communication of central banks and the resulting benign market perception have helped in containing perceived risks of bond yield volatility, currency weakness, or, more broadly, fiscal dominance. Nevertheless, APP announcement communication has been varied when it comes to reporting on purchases, explanation of how potential adverse effects will be handled, and the timely reporting of amendments of APPs. For instance, Croatia has explained that they only intend to reactivate their APP if similar stress reemerges. Hungary, which has gradually increased purchases, has clearly communicated any changes and the motivation for such changes, including when the envelope has been increased.

Central banks should clearly communicate at the introduction, during operation, and the exit of APPs with a view to reduce speculative panics, enhance trust, and facilitate normalization. APPs could carry substantial risks, including for central bank credibility and independence. Worse, they could create the perception of monetary financing, particularly if the sterilization strategy is not clearly conveyed. Honest and transparent communication, including the publication of conditions governing future actions, are key for a smooth exit.⁵ The IMF’s “Central Bank Transparency Code” (2020b) offers relevant guidance for the setup of the communication strategy. As Unsal and Garbers (forthcoming, 1) note: “Exceptional times need exceptional communication to safeguard credibility.”

Chapter 6 Strategies for Exiting Asset Purchase Programs

Exit Strategies

Exiting APPs is a process that involves a number of steps, the full cycle of which has rarely occurred in practice. Conceptually, the exit would start with tapering (slowing the pace of purchases), progressively moving to stopping purchases, maintaining the stock of purchased assets by rolling them over, and finally reducing the stock in a passive or active manner. Thus far, however, exit experiences are only available for the case of large-scale APPs in AEs, and even many of those have not been completed. Some evidence from Japan argues for the following steps, although not necessarily in this sequence: (1) halting extraordinary interventions; (2) downsizing and unwinding the central bank balance sheet; (3) selling the purchased assets, if necessary, while keeping country-specific factors in mind; and (4) raising short-term interest rates (Yamaoka and Syed 2010; Agostini and others 2016). The Bank of England (2018), for instance, announced that it did not intend to start reducing the stock of purchased assets until the bank rate reached about 1.5 percent, a level from which it could be cut materially if needed to react to shocks. The Bank of England also stressed that any reduction in the stock of purchased assets would be conducted at a gradual and predictable pace. More recently, the Bank of England governor has indicated that unwinding QE might instead be a first tightening step (House of Lords 2021).

The decision of when to stop or taper purchases should be driven by several factors—not least the impact on central bank credibility in achieving its objectives, tasks, and functions. There may be key differences between large-scale APPs in AEs and smaller-scale ones in EMs, as their primary purposes often differ. The former often issue a reserve currency and have a longer credible policy track record than the latter, hence enjoying more flexibility. In contrast to large-scale APPs in AEs that aimed at significant monetary stimulus to achieve the inflation target, the EM APPs have typically been smaller in scale and primarily aimed at alleviating market dysfunction. Tapering could thus be a simpler process for EMs, but given their typically thinner markets, smaller amounts could have a relatively larger impact. Consideration should be given to stopping purchases when:

The initial condition(s) causing the APPs are no longer relevant. Thus, if the market dysfunction, originally caused by exogenous factors, dissipates and normalcy is restored, then continuing with APPs may raise questions about their need and effectiveness.
The initially stated objective(s) have been achieved. If an APP continues after the stated objective of stabilizing dysfunctional government bond markets has been achieved, but with a view to serving other objectives, say, to lower government funding costs in a nontransparent manner, then credibility may be eroded quickly.
Excess limits are in danger of being reached. Going further may endanger both de jure and de facto central bank independence as well as statutory and real fiscal limits (see Chapter 4).

Stopping or scaling down purchases (that is, tapering) requires careful management. A firm understanding of all the financial and fiscal risks associated with prematurely stopping purchases and scaling down is critical. For instance, a narrow focus on inflation—although well-intended as the primary objective of monetary policy—could still trigger temporary market turbulence. A combination of carefully calibrated and clearly communicated unconventional tools may help smooth an otherwise segmented transmission mechanism until the new information is fully disseminated. The latter may be even more important in EMs with shallower financial markets. Hence, a thorough understanding should also help in the calibration of communication and reduce the risk of market overreaction.¹ A particularly important issue arises if the stopping signals a major policy shift. In such cases, an “announced” stop accompanied by a coherent and well-communicated policy package may be necessary.

The unwinding of central bank asset purchases should be flexible and viewed as a long-term objective. The perceived steady-state level of central bank holdings is specific to the country, its policies, and the situation it faces. Nevertheless, within the monetary anchor, two aspects of unwinding should always be kept in mind: preserving central bank independence and reducing the risk of excessive market fluctuations. First, a firm understanding of the fiscal and financial risks should be a precondition for any unwinding (Agostini and others 2016). Second, the unwinding should be transparent and undertaken in coordination with the Treasury. In principle, central banks should solely focus on their stipulated objectives, tasks, and functions instead of maximizing profits. In practice, however, most central bankers feel that remaining solvent and especially being able to regularly transfer profits to the government help ensure their autonomy. Although depending on the accounting of the purchased securities—whether held for trading or kept until maturity—a transparent and well-informed approach can better mitigate and prepare for any potential balance sheet losses associated with an increase in long-term interest rates. In most cases, a passive exit may be appropriate. Subject to market conditions, the central bank should consider letting its portfolio mature without rolling it over (as Japan did in 2006). However, it should not hesitate to use its government securities for repos or outright sales if it needs to tighten.

Spillovers from the actions of major central banks can confound exit, and “pausing” should be an option. The QE implemented by the Federal Reserve and ECB has spilled over to EMs, and such actions affect capital flows, asset prices, and the exchange rate. Whether these spillovers have been stronger than the ones related to conventional monetary policy during the pre-COVID period is still debatable.² While spillovers to EMs are highly likely, their magnitude is difficult to ascertain ex ante. The extent to which markets react to major central bank actions, or “tantrum episodes,” would be key. Whether European EMs choose a passive or an active (exit) unwinding strategy, their central banks should stand ready to “pause” if spillover effects are deemed to be abrupt, especially for the exchange rate.

Exit Options in European Emerging Markets

Circumstances are maturing for an exit from APPs in European EMs. The region’s APPs were generally deployed with the primary goal of alleviating financial market dysfunction. Despite the severity of the economic shock, financial market dysfunction quickly abated after the initial weeks of the pandemic. An economic recovery appears to be well underway, and relatively elevated levels of inflation in the region have sparked discussion about the timing for the exit from policy stimulus and monetary policy tightening.

Where the goals of APPs remained relatively limited to alleviating market dysfunction, APPs have stopped, while the unwinding, if needed, can be managed quite flexibly. Barring another episode of exogenously driven market dysfunction, there is no need to resume APPs. The relatively limited holdings that remain on central bank balance sheets may be maintained for a prolonged period until they mature; if needed, they can be unwound gradually with a view to limit the market impact in the process. However, the monetary tightening phase may provide an expedient opportunity to unwind the holdings.

In countries where prolonged APPs have tackled more diverse objectives, exiting from APPs should be weighed against those expanded objectives and the interaction with conventional monetary policy adjustment. If the expanded objectives warrant continuation of APPs, an efficiency comparison should be made between APPs and other potentially more adequate means to meet the expanded objectives. For example, the Hungarian Central Bank later decided to scale back its long-term collateralized credit facility and partially substitute it with continued purchases of government securities, arguing that the APP may ensure more effective transmission to long-term interest rates. Where APPs have continued throughout the pandemic with the objective of supporting monetary easing through transmission to the longer end of the yield curve, it would be appropriate to taper or stop APPs, as central banks move toward increasing policy interest rates. Given the heterogeneity, country-specific conditions, and clearly communicated changing states objectives of APP, any action should be state-contingent and specific to each country. However, outright unwinding of acquired assets should be managed (and paced) in consideration of the size of outstanding positions and their likely effects on the objectives for which APPs were deployed.

Chapter 7 Conclusions and Policy Lessons

APPs have served several European EMs well during the pandemic. Their main goal was to alleviate market dysfunction, and in that regard, they have been quite distinct from QE used by the major AE central banks following the GFC. In Poland and Hungary, APPs have also aimed to support the monetary policy transmission mechanism beyond the initial period of market turbulence, as evidenced by their prolonged use compared to others in the region. Focusing on the initial period of market turbulence, we find evidence of a stabilizing impact on government bond markets. APP announcements had an immediate positive impact on the functioning of the government bond markets in European EM countries that pursued such programs. APPs reduced 10-year government bond yields and compressed bid-ask spreads for government bonds. Initial announcements in March and April 2020 also halted the increase in term spreads. In doing so, APPs also helped facilitate smooth fiscal financing at an extraordinary moment. At the same time, we find little evidence of systematic currency depreciation. In terms of spillovers to other asset prices, equity prices increased by around 2 percent two to three days after the APP announcements.

The success of APPs so far in EM Europe comes with some caveats. First, our results are silent about the potentially important longer-term impacts of the APPs, identification of which should be pursued when sufficient time has elapsed since initiation. Second, due to data limitations, our high-frequency analysis does not control for all domestic policy actions around APP events, including the impact of foreign exchange intervention on exchange rates. Finally, our analysis focused on direct impacts of the APP announcements but did not attempt to construct the counterfactual case of no APP announcements, which could potentially modify the quantitative findings.

In case of renewed market dysfunction, there is potential scope to scale up APPs in most of the European EMs we examined, provided they remain consistent with primary monetary policy objectives with the risk of fiscal dominance well contained. The macroeconomic fundamentals of this group of countries are generally strong, though with some variation that may affect the feasible scope for APPs in each country. The relatively small scale of APPs as a generally time-bound tool during the pandemic has posed little risk of fiscal dominance. Nevertheless, the availability and usefulness of APPs as a larger-scale tool is likely to vary among countries, including based on the perceived risks of fiscal dominance. EM central banks should be mindful of the limits for their APPs, the breach of which may endanger both de jure and de facto central bank independence as well as statutory and real fiscal limits.

Central banks need to plan for the role of APPs in eventual monetary policy tightening. Inflation has recently increased in many European EMs, in large part due to temporary factors, including base effects, pent-up demand, and supply bottlenecks. Setting aside the question —which is beyond the scope of this paper—of whether the current increase in headline inflation merits immediate policy tightening, as national vaccination programs proceed and a durable recovery appears increasingly possible, central banks should begin to plan for monetary policy normalization, including the future of APPs.

An increase in policy interest rates should be accompanied by a tapering or discontinuation of APPs. Where APPs were concentrated at the beginning of the pandemic or have effectively ended (Croatia, Romania, Serbia, Turkey), these APPs would play little role in monetary policy tightening, which should be done through conventional instruments. The Polish and Hungarian APPs have continued, arguably influencing the yield curve over a longer period. When these central banks tighten monetary policies, it would be appropriate to complement tightening at the short end of the curve (policy rates) with a tapering or discontinuation of APPs to support the transmission of higher policy rates to the longer end of the curve. Furthermore, when the government securities market has largely regenerated following a major shock and is thus no longer illiquid or segmented, continued central bank intervention in this market could be perceived as monetarization of budget deficits and ultimately increase risk premiums. The pace of tapering should be tailored to country circumstances, avoiding a greater-than-desired tightening of overall monetary conditions. Unwinding of purchased securities could be done gradually, as they mature or to absorb liquidity as part of a tightening, with a view to limit market disturbances. However, raising interest rates does not necessarily preclude further exceptional asset purchases during circumstances of exogenous, significant, and temporary market disruptions.

Whether central banks use conventional or unconventional monetary policy tools, clear and transparent communication at each step of the process is critical. This is important for the central banks to remain credible and is particularly important when a large shock hits and triggers a major policy shift. Effective communication will also facilitate information dissemination, thereby reducing asymmetric information and ensuring a faster and smoother adjustment to a new equilibrium.

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/087/2021/021/article-A001-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Collapse
Expand

Asset Purchase Programs in European Emerging Markets

Author:

Mr. Marco Arena

Mr. Rudolfs Bems

Mr. Nadeem Ilahi

Mr. Jaewoo Lee

William Lindquist

, and

Mr. Tonny Lybek

Volume/Issue:: Volume 2021: Issue 021

Publisher:: International Monetary Fund

ISBN:: 9781513593753

ISSN:: 2616-5333

Pages:: 59

DOI:: https://doi.org/10.5089/9781513593753.087

Keywords
Chapter 1 Introduction
Chapter 2 Characteristics of Asset Purchase Programs in European Emerging Markets
Chapter 3 Effectiveness of Asset Purchase Programs
Chapter 4 Limits to Emerging Market Asset Purchase Programs
Chapter 5 Complementary Policies and Communication
Chapter 6 Strategies for Exiting Asset Purchase Programs
- Exit Strategies
- Exit Options in European Emerging Markets
Chapter 7 Conclusions and Policy Lessons

View raw image

^{Figure 1.}

Emerging Market Bond Fund Flows

(Billions of USD)

View raw image

^{Figure 2.}

Asset Purchases by Central Banks During Pandemic

(Percent of 2020 GDP; end-February 2020 to end-June 2021)

View raw image

^{Figure 3.}

Asset Purchases in Emerging Market Europe by Month

View raw image

^{Figure 4.}

Asset Purchase Programs, Capital Flows, and Budget Deficits, 2020

View raw image

^{Figure 5.}

Poland: Cumulative Issuance of Securities and Central Bank Purchases

(Billions of PLN)

View raw image

^{Figure 6.}

Asset Purchases by European EM Central Banks

(Percent of 2020 GDP; from end-February 2020 to end-June 2021)

View raw image

^{Figure 7.}

Maturity Composition of NBP and MNB Asset Purchases

View raw image

^{Figure 8.}

Central Bank Policy Interest Rates

View raw image

^{Figure 9.}

Change in Central Bank Assets

View raw image

^{Figure 10.}

Poland: Central Bank Bills and Asset Purchases

(PLN billion; cumulative from March 15, 2020, to July 2, 2021)

View raw image

^{Figure 11.}

Central Bank Holdings of Local Treasury Securities

(Percent of Treasury securities outstanding)

View raw image

^{Figure 12.}

Change in Domestic Treasury Bond Holdings

(Cumulative change; percent of 2020 GDP)

View raw image

^{Figure 13.}

Government Bond Market Response to the Asset Purchase Program Announcements

View raw image

^{Figure 14.}

Spillovers from Asset Purchase Program Announcements into the Broader Macroeconomy^¹

View raw image

^{Figure 15.}

Impact of Asset Purchase Programs Announcements on Domestic Government Bond Yields

View raw image

^{Figure 16.}

Impact of Asset Purchase Programs Announcements on Domestic Term Spread

View raw image

^{Figure 17.}

Impact of Asset Purchase Programs Announcements on Equity Prices

View raw image

^{Figure 18.}

Indicators of Fiscal Sustainability

View raw image

^{Figure 19.}

Indicators of Domestic Financial Market Development

View raw image

^{Figure 20.}

Inflation and Expectations

View raw image

^{Figure 21.}

Indicators of Medium-Term Growth and Macroeconomic Stability

View raw image

^{Figure 22.}

Indicators of Scope for Asset Purchase Programs^¹

View raw image

^{Figure 23.}

Real Interest Rates, 2015–21

(Percent, three-month moving average: one-year government bond yield deflated by CPI^¹)

Foresee

^{Table 2.}

Initial Announcement Dates for Sample Emerging European Countries

	Initial APP announcement date	Policy rate cut	Expanded liquidity facilities	Reduction in bank reserves	Macropru loosening	FXI
HRV	13-Mar	no	yes	yes	no	yes
POL	16-Mar	yes	yes	yes	yes	no
ROM	20-Mar	yes	yes	no	no	yes
HUN	7-Apr	no^¹	yes	no	no	no data
TUR	31-Mar	no	yes	yes	no	no data

^¹The base rate was later reduced in two steps (June 24, 2020, and July 22, 2020) from 0.90 to 0.60 percent.

Topics

Countries

Series

About

Resources

Abstract

Chapter 1 Introduction

Chapter 2 Characteristics of Asset Purchase Programs in European Emerging Markets

Conceptual Underpinnings

Scale and Timing

Goals

Operational Aspects

Interaction of Asset Purchase Programs with Other Monetary Policy Tools

Impact on Target Markets

Chapter 3 Effectiveness of Asset Purchase Programs

Related Literature

Findings of Event Study Analysis

Local Projections Analysis

Chapter 4 Limits to Emerging Market Asset Purchase Programs

Fiscal Dominance as a Constraint on Asset Purchase Programs

The Importance of Macroeconomic Fundamentals

Macroeconomic Fundamentals in European Emerging Markets with Asset Purchase Programs

Chapter 5 Complementary Policies and Communication

Additional Unconventional Monetary Tools

Adjusted Lending Operations

Negative Policy Rates

Forward Guidance

Complementary Policies

Communication

Chapter 6 Strategies for Exiting Asset Purchase Programs

Exit Strategies

Exit Options in European Emerging Markets

Chapter 7 Conclusions and Policy Lessons

Share Link

Table of Contents

Headings

Figures

Metrics

Foresee