## Annex 1. Conceptual Underpinnings of Central Bank Asset Purchase Programs ## Source details **Canonical URL:** [Annex 1. Conceptual Underpinnings of Central Bank Asset Purchase Programs](https://www.imf.org/-/media/files/publications/dp/2021/english/appeemea.pdf) ## Other formats - [Markdown version](/-/media/files/publications/dp/2021/english/appeemea.pdf.md) - [Structured JSON version](/-/media/files/publications/dp/2021/english/appeemea.pdf.json) --- ### Introduction - Widespread adoption of asset purchase programs (APPs) by emerging market (EM) central banks in response to the COVID-19 pandemic; some 20 EM central banks launched APPs following March 2020 market turmoil. - In Europe, six EM central banks engaged APPs during March–May 2020: Croatia, Hungary, Poland, Romania, Serbia, Turkey. - Stated primary goals: alleviate financial market dysfunction, furnish liquidity, and repair impaired monetary transmission mechanisms. - Distinction from AE APPs: AE APPs after the GFC mainly aimed to provide additional stimulus at the effective lower bound; EM APPs during the pandemic focused on market functioning and liquidity. ### Conceptual underpinnings - Theoretical foundations described as "vague and still evolving." - Mechanisms through which APPs operate: - Influence quantity by easing liquidity constraints—depending on the degree of sterilization—and by reducing segmentation among institutions and along the yield curve. - Smooth the yield curve, including by mitigating overshooting in a highly segmented yield curve. - Anchor expectations by conveying information about intended policies. - Channels summarized: - Portfolio rebalancing channel: changes relative prices and duration risk, prompting investor reallocation. - Liquidity channel: provides liquidity depending on sterilization; can ease slope and bumpiness. - Expectation—signaling/confidence channel: alters information sets and reduces uncertainty. - Intermediation channel: affects intermediaries’ on-lending capacity and capital constraints. - Market microstructure effects: committed large buyers can improve price discovery and narrow bid-ask spreads. - Long-run neutrality caveat: persistent influence on the yield curve is difficult because deviations are arbitraged away. ### Scale, timing, and key magnitudes - APPs in European EMs were mostly small-scale relative to large central banks; some EMs implemented APPs at the higher end globally ("5 to 7 percent of GDP" among some European EMs). - Specific magnitudes: - Croatian National Bank market value of securities purchased: 5.5 percent of GDP. - Federal Reserve purchased: 23 percent of GDP (assets purchased since the beginning of the pandemic through the end of June 2021). - ECB purchased: 14 percent of GDP (same reference period). - 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. - MNB balance sheet expansion from start of the pandemic to end-May 2021: 20 percent of GDP. - Federal Reserve balance sheet expansion comparable figure: 18 percent of GDP. - ECB balance sheet expansion comparable figure: 26 percent of GDP. - Time pattern: - APP activity peaked in April and May 2020 and fell off sharply thereafter. - By June/July 2020 purchases in Croatia, Serbia, and Turkey had ceased. - NBP continued purchases into 2021, slow after July 2020 then increased March–May 2021. - MNB purchases accelerated in second half of 2020 and remained at a high level in 2021. ### Operational aspects and interactions with other tools - Assets purchased: - Mostly local Treasury bonds; country-specific deviations include mortgage-backed securities and corporate bonds (MNB) and government-guaranteed securities (Poland). - Poland: purchases evenly split between local Treasury securities and government-guaranteed securities; NBP purchased about 36 percent of combined issuance (March 15–end-2020). - MNB: sizeable purchases of mortgage-backed securities and corporate bonds; by end-June 2021 MNB had bought 7.2 percent of GDP in securities. - Market segment and maturities: - Most purchases conducted in the secondary market (EU requirement). - NBP concentrated on 5- to 10-year bonds; MNB overwhelmingly purchased >10-year Treasury securities (later planned to buy <10-year). - Purchase methods: - Quantity-based approaches preferred (auctions, regular tenders, bilateral purchases) over price-based methods. - Croatia: five large ad hoc tenders; Croatian purchases recorded as 4.9 percent of GDP (nominal value) and 5.5 percent of GDP (market value). - Poland: biweekly or monthly tenders; total purchased government bonds and guaranteed bond purchases as of end-June 2021: 5.9 percent of GDP. - Romania: targeted bilateral purchases; program size about 0.4 percent of GDP. - Turkey: purchases amounted to 1.6 percent of 2020 GDP; no additional purchases since July 2020. - Serbia: central bank holdings of central government securities increased from zero before the pandemic to 1.7 percent of GDP. - Interaction with conventional policy: - APPs often implemented before conventional policy space was exhausted. - Hungary revamped operations early in the pandemic to support the currency while initially maintaining the policy rate; later lowered policy rate and continued APPs. - Poland reduced policy rate to near zero only after initiating the APP. - Turkey initially paired asset purchases with rate reductions, but policy rate later increased in 2020 in response to currency pressures. - Sterilization and balance sheets: - Sterilization was widely used (deposit facilities, issuance of central bank securities, sale of foreign assets) to anchor inflation expectations and signal commitment to price stability. - Croatia offset purchases by sale of foreign assets (preventing balance sheet growth). - Only about one-third of MNB’s balance sheet expansion tied to securities purchases. ### Country-specific objectives and cases - Common objective: mitigate market dysfunction, provide liquidity, repair transmission. - Country examples (selected): - Croatia: first auction March 13, 2020; objectives included exchange rate stability, domestic liquidity, and government securities market stability; purchases: 4.9 percent of GDP (nominal), 5.5 percent of GDP (market). - Hungary (MNB): pre-pandemic APP-like measures existed; April 7, 2020 APP announced; by end-June 2021 MNB purchases: 7.2 percent of GDP; envelope close to HUF 3000 billion (about 5.7 percent of projected GDP) was a review trigger; August 24, 2021 decision to begin gradually withdrawing the government securities programme and set a target amount for weekly purchases. - Poland (NBP): announced March 16, 2020; rapid purchases through July 2020 then modest purchases; pace increased March–May 2021; total purchases as of end-June 2021: 5.9 percent of GDP; NBP sterilized purchases largely. - Romania (NBR): announced March 20, 2020 package including leu-denominated government securities purchases; purchases mainly early and effectively stopped August 2020; total program size about 0.4 percent of GDP. - Serbia (NBS): did not announce APP but performed outright bilateral transactions; holdings rose to 1.7 percent of GDP and represent about 9 percent of outstanding local government securities. - Turkey (CBRT): announced March 31, 2020; purchases in secondary market including from Unemployment Insurance Fund; total purchases 1.6 percent of 2020 GDP; no purchases since July 2020. ### Evidence on effectiveness — event-study and local projections - Event-study (11 announcement events across five countries; 3-day window around announcements): - Prior to initial announcements: 10-year yields increased by 0.20 percentage points in the three days preceding APP announcements. - Median findings following announcements: - Long-term yields fell by 0.3 ppts by t = 1 and remained below t = 0 levels subsequently (median decline by maturity: 1-year fell by 0.1 ppts; 10-year fell by 0.3 ppts). - Yield curve flattened after initial APP announcements. - Term spreads declined following initial announcements but decline was short-lived and did not extend to subsequent announcements. - Liquidity (bid-ask spreads) tightened prior to initial announcements and temporarily improved after announcements. - Exchange rate: APP announcements did not induce a significant exchange rate response; median exchange rate response broadly stable. - Equity markets: equity indexes rose gradually, reaching about 2 percent by day four following announcements; initial announcements showed a median equity response of 4 percent by day four. - Local projections (controls for domestic policy rate and global factors—VIX or Fed APP announcement): - APP announcements associated with drops in bond yields after controls. - Statistically significant impact on 10-year government bond yield of around –35 basis points one to three days after the announcement. - No systematic impact on term spreads (consistent with event study). - Equity prices: estimated response 1–3 percentage points over three-day horizon; statistical significance sensitive to global control used (not significant when VIX used). ### Literature synthesis (selected findings) - Hartley and Rebucci (2020), 13 EMs: [20.28, 20.43] bps impact on 10-year government yields, window 1–3 days. - Arslan, Drehmann, and Hofmann (2020), 9 EMs: [210, 250] bps impact on 10-year government yields, window 1–5 days. - Sever and others (2020), 11 EMs: 235 bps impact on 10-year government yields in the next few days; median-term premiums reduced by almost 20 bps in a week; limited currency impact; EM equity markets improved few days post announcements. - Fratto and others (2020), 15 EMs: average statistically significant effects broadly consistent with above findings; heterogeneous exchange rate impacts and positive EMBI responses. ### Limits, risks, and mitigation factors - Principal constraint: fiscal dominance risk—APPs could facilitate financing of deficits and threaten central bank independence if fiscal sustainability is weak. - Factors mitigating risks and expanding APP scope: - Fiscal sustainability, well-developed domestic bond markets, credible monetary policy frameworks, legal prohibitions on direct financing, strong growth prospects, absence of large current account deficits, robust FX reserve coverage. - Additional risks from APP design: - Increased exposure to maturity and interest rate risk; credit risk when purchasing corporate and mortgage bonds. - Potential reduction in liquidity of specific bond series if central banks become large holders. - Sterilization and communication as safeguards: - Sterilization can preserve credibility and limit inflationary pressures. - Clear, transparent communication reduces perceptions of monetary financing and supports credibility. ### Exit strategies, sequencing, and communication - Conceptual exit cycle: 1. tapering (slowing pace of purchases); 2. stopping purchases; 3. maintaining stock by rolling over purchased assets; 4. reducing stock passively or actively. - Guidance and practice: - Exiting should minimize market disruption and may be a long-term process. - Unwinding principles: preserve central bank independence, understand fiscal/financial risks, coordinate with Treasury, prioritize passive exits (maturity rolls) but use repos or sales if needed. - Examples: Bank of England tied balance sheet reduction to bank rate reaching about "1.5 percent" in its practice. - Central banks should be ready to "pause" unwinding if major central bank spillovers are abrupt. - Communication best practices: - Clear communication at inception, during operation, and at exit reduces speculative panics and limits perceptions of monetization. - Croatia communicated APPs would be reactivated only if similar stress reemerged; Hungary communicated gradual increases and envelope motives. ### Policy conclusions and recommendations - APPs during the pandemic generally met immediate market-functioning objectives with limited fiscal dominance risk given their relatively small and time-bound scale. - Recommendation for normalization: - Complement an increase in policy interest rates with tapering or discontinuation of APPs to support transmission of higher policy rates to the longer end of the yield curve. - Where APPs have effectively ended (Croatia, Romania, Serbia, Turkey), conventional instruments should drive tightening. - For ongoing APPs (Poland, Hungary), tapering/discontinuation alongside short-end tightening is appropriate to support transmission. - Pace and sequencing: - Tapering pace should be tailored to country circumstances to avoid excessive tightening. - Unwinding can be gradual (let portfolios mature) or active if necessary to absorb liquidity, with careful communication. - Caveat: - Raising interest rates does not preclude further exceptional asset purchases if exogenous, significant, and temporary market disruptions occur. - Complementary measures: - CFMs, macroprudential tools, reserve requirements, broadened collateral, and lending facilities can complement APPs to alleviate liquidity stress. - Structural and governance reforms can improve APP effectiveness and reduce fiscal costs (example finding: improving governance to top 10 percent could, on average, save about "0.3 percent of GDP per year"). *Source: https://www.imf.org/-/media/files/publications/dp/2021/english/appeemea.pdf* ### Annex 1. Conceptual Underpinnings of Central Bank Asset Purchase Programs .........49 ### Annex 1. Conceptual Underpinnings of Central Bank Asset Purchase Programs .........49 ### Introduction - Governments enlarged fiscal and monetary policies in response to the COVID-19 pandemic; a novel monetary policy development was the widespread adoption of asset purchase programs (APPs) by emerging market (EM) central banks, including in Europe. - When international financial market turbulence followed the outbreak in March 2020, some 20 EM central banks launched APPs, most 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. - Stated primary goals: alleviate financial market dysfunction, furnish liquidity, and repair impaired monetary transmission mechanisms. - Distinction from AE APPs: APPs introduced in AEs after the GFC were mainly aimed at providing additional stimulus, as policy rates had reached the effective lower bound. ### Conceptual underpinnings - Theoretical foundations of APPs are "vague and still evolving." - Mechanisms through which APPs operate: - Influence quantity by easing liquidity constraints—depending on the degree of sterilization—and by reducing segmentation among institutions and along the yield curve. - Smooth the yield curve, including by mitigating overshooting in a highly segmented yield curve. - Anchor expectations, mainly by conveying information about intended policies. - APPs can influence agents’ ability and willingness to borrow and lend, stabilizing aggregate demand in the short term. - Effectiveness depends on country- and shock-specific factors, market beliefs about APP efficacy, and whether APPs are used to finance higher public spending (with the efficiency of such spending and degree of sterilization mattering for outcomes). - Over the long term, monetary policy is neutral and persistent influence on the yield curve is difficult because deviations will be arbitraged. ### Scale and timing - APPs in European EMs during the pandemic were mostly small-scale relative to large central banks. - Some European EMs implemented smaller-scale APPs (Romania, Serbia, Turkey), while others (Croatia, Hungary, Poland, "5 to 7 percent of GDP") implemented APPs at the higher end among EMs globally. - Specific numeric points: - The market value of securities purchased by the Croatian National Bank is 5.5 percent of GDP. - The Federal Reserve and ECB purchased 23 and 14 percent of GDP in assets, respectively (figures refer to assets purchased since the beginning of the pandemic through the end of June 2021). - Time pattern: - APP activity peaked in April and May 2020 and then fell off sharply. - By June/July 2020, purchases in Croatia, Serbia, and Turkey are known to have ceased. - The National Bank of Poland (NBP) continued purchases in 2021, initially at a very slow pace after July 2020, then increased the pace in March to May 2021. - Hungary’s National Bank (MNB) purchases accelerated in the second half of 2020 and remained at a high level in 2021. ### Goals and country-specific objectives - Common objective: mitigate financial market dysfunction, provide liquidity, and repair impaired monetary transmission mechanisms early in the pandemic. - APPs were generally not aimed at providing macroeconomic stimulus in EMs, unlike many AE APPs. - Country examples: - Poland: initial large purchases during early pandemic months; slow pace from August 2020 to February 2021; a significant increase in purchases in March to May 2021 to prevent a premature tightening of long-term yields and bolster monetary easing transmission. - Hungary (MNB): had implemented APPs even before the pandemic (including for private securities) to promote market development; at the start of the pandemic, cited market dysfunction and lowering long-term interest rates; later increased the pace of purchases and found purchases of government securities more efficient than the long-term collateralized credit facility for influencing long-term rates. ### Effects on government financing and fiscal interaction - Purchases of government securities may have eased financing costs and lowered refinancing risks of the government. - APPs facilitated the smooth financing of anticrisis fiscal expenditures. - There appears to be some correlation between the size of the increase in budget deficits and the size of APPs (visual correlation noted in the source figures for 2020). - Bond and equity outflows from end-February to end-April 2020 generally reversed subsequently. ### Effectiveness in markets and transmission - Event-study evidence cited indicates APPs were successful in alleviating market dysfunction in the immediate aftermath of the March–April 2020 shock: - Some evidence of easing liquidity pressures. - A reversal in the surge in term spreads. - No indication of foreign exchange market pressures in the evidence presented. - Positive spillovers into equity markets noted. - The initial long-term collateralized credit facility played a relatively large role in promptly providing liquidity; later, APPs partially took over that role in some cases (MNB finding purchases of government securities more efficient for affecting long-term rates). ### Limits, exit, and communication - EMs may face tighter limits to APP use, influenced by global conditions and domestic factors. - For EU members, membership structure and discipline reduce the fiscal dominance risk; monetary policy frameworks enjoy credibility and fiscal policy is anchored in medium-term sustainability. - Termination and exit: - APPs should stop when initial objectives have been met, are no longer relevant, or when limits have been reached. - Exiting and unwinding purchased assets should attempt to minimize market disruption and would likely be a long-term process. - Slowing down and exiting from APPs remains a challenge with little guidance from large central bank experience. - Communication: - Effective communication is particularly important given the novel use of APPs in EMs. - Central banks should communicate clearly at inception, during operation, and at exit/unwinding. ### Role of APPs going forward - With initial goals largely met and market conditions different than at introduction, several APPs have concluded; for those that continue, it is important to consider their role going forward and interaction with monetary policy tightening. - Policy recommendation: complement an increase in policy rates with a tapering or discontinuation of APPs to support transmission of higher policy rates to the longer end of the yield curve. *Source: https://www.imf.org/-/media/files/publications/dp/2021/english/appeemea.pdf* ### 2. Asset Purchase Programs and ETF and Mutual Funds Bond and Equity ### 2. Asset Purchase Programs and ETF and Mutual Funds Bond and Equity ### Operational aspects of APPs in European emerging markets - Assets purchased: - Mostly local Treasury bonds. - Poland: purchases evenly split between local Treasury securities and government-guaranteed securities issued by Polish development banks; 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. - Hungary (MNB): sizeable purchases of mortgage-backed securities and corporate bonds in addition to Treasury securities. - Market segment and maturities: - Most purchases conducted in the secondary market; EU central banks with APPs are required to buy government securities only in the secondary market. - NBP concentrated purchases on 5- to 10-year bonds. - MNB overwhelmingly purchased Treasury securities with greater than 10-year maturity. - Purchase methods: - Quantity-based approaches were generally preferred over price-based methods to influence interest rates rather than cap yields. - Croatia: five large ad hoc tenders (market value of securities purchased by Croatian National Bank is 5.5 percent of GDP). - Hungary: regular auctions and some over-the-counter transactions; initially purchased in primary and secondary markets for private securities, later paused mortgage bond purchases and announced it would only buy green mortgage bonds. - Poland: biweekly or monthly tenders. - Romania: targeted bilateral purchases to alleviate specific counterpart liquidity positions. ### Interaction with other monetary policy tools and balance sheets - Timing relative to conventional easing: - APPs in EM Europe were implemented before conventional monetary policy space was fully exhausted. - Hungary: revamped monetary operations early in the pandemic to support the currency while initially maintaining the policy rate, which was later lowered. - Poland: NBP reduced its policy rate to near zero, but only after initiating the APP. - 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. - June–July 2021: Hungarian Central Bank increased its policy interest rate and began to phase out some crisis measures but continued purchases of government securities. - Collateralized lending and other facilities: - Poland: repo transactions and a discount credit facility to refinance bank loans (used sparingly). - Hungary: long-term collateralized lending facility with maximum five-year maturity; broadened eligible collateral. - Balance sheet and sterilization: - APPs were not the sole driver of balance sheet expansions; early in the pandemic (end-June 2020) Croatian and Romanian central bank balance sheets stayed broadly stable as domestic asset purchases were offset by foreign exchange reserve sales. - By end-May 2021, foreign exchange reserves had recovered and asset purchases coincided with overall balance sheet expansions in those central banks. - MNB’s balance sheet expansion from the start of the pandemic to end-May 2021 was 20 percent of GDP; comparable expansions: Federal Reserve 18 percent of GDP, ECB 26 percent of GDP. - Only about one-third of MNB’s balance sheet expansion can be tied to securities purchases. - Sterilization was a prominent feature: - Croatia: purchase of domestic assets offset by sale of foreign assets, preventing balance sheet growth. - Poland: issuance of central bank securities to drain reserves (substitution of liabilities). - Deposit facilities also used to drain bank reserves. - Sterilization purpose: to anchor inflation expectations and signal commitment to medium-term price stability; APPs intended to alleviate market dysfunction rather than increase overall liquidity. ### Impact on targeted markets and investors - Central bank holdings and market stabilization: - By May 2021, central bank market shares of outstanding local Treasury securities increased to between 6 percent in Turkey and 9 percent in Poland. - Central bank purchases helped stabilize markets amid increased issuance during the pandemic and in Poland and Turkey broadly offset outflows from foreign investors. - In Serbia and Hungary, central bank purchases were supplemented by continued strong demand from domestic banks and other domestic investors (for example, special retail bonds for households). - Market liquidity considerations: - Large central bank purchases could reduce liquidity of specific bond series, potentially increasing government funding costs if bonds become less liquid. - Some central banks (Croatia and initially Hungary) stated limits not to buy more than a third of each bond series; Hungary later abandoned this limit. - A committed large buyer can improve price discovery, potentially narrowing bid-ask spreads. ### Hungary case study (Box 1): evolution of MNB APPs - Pre-pandemic activity: - MNB had employed APP-like measures before the pandemic (mortgage bond programs and unconditional interest swaps in 2017–18) aimed at lowering long-term yields; the 2018 program was stopped end-2018. - Pandemic-era actions: - April 7, 2020: MNB announced (1) an APP for government securities, (2) reactivation of mortgage bond purchase program, (3) easing conditions for corporate bond purchases. Objectives: “to restore the stable liquidity position of the government securities market” and “to strengthen monetary policy transmission.” - August 2020: MNB cited “the higher government financing needs” and committed to continue purchases in the long segment to extend government debt maturities. - January 2021: MNB to gradually reduce its long-term collateralized lending facility and rely more on APP in government securities; planned to buy government securities with less than 10 years maturity. - April 2021: MNB to include so-called green bonds issued by the government in its asset purchases. - June–July 2021: MNB increased policy interest rate while continuing government securities purchases; declared APP crucial in its monetary instrument set. - Envelope and withdrawal plans: - Next review of APP when envelope close to HUF 3000 billion (about 5.7 percent of projected GDP), likely mid-August 2021. - August 24, 2021: Monetary Council decided to “begin gradually withdrawing the government securities programme while considering aspects of maintaining market stability,” and to “set a target amount for weekly purchases” rather than a limit on total stock purchased. *Source: IMF — 2. Asset Purchase Programs and ETF and Mutual Funds Bond and Equity (appeemea - 2. Asset Purchase Programs and ETF and Mutual Funds Bond and Equity).* ### Box 1. The Evolution of the Asset Purchase Programs of Hungary (continued) ### Box 1. The Evolution of the Asset Purchase Programs of Hungary (continued) ### Related literature - Emerging literature on APPs in EMs during the COVID-19 crisis finds immediate positive impacts of central bank purchases of government bonds on long-term government bond yields. - Selected findings from recent studies (as reported): - Hartley and Rebucci (2020), 13 EMs: [20.28, 20.43] bps impact on 10-year government yields, window 1–3 days; no assessment of other asset prices or controls reported. - Arslan, Drehmann, and Hofmann (2020), 9 EMs: [210, 250] bps impact on 10-year government yields, window 1–5 days; yes on other asset prices (exchange rate appeared shored up); yes controlling for other variables. - Sever and others (2020), 11 EMs: 235 bps impact on 10-year government yields, next few days; yes on other asset prices (median-term premiums reduced by almost 20 bps in a week; limited currency impact; EM equity markets improved few days post announcements); yes controlling for other variables. - Fratto and others (2020), 15 EMs: average statistically significant effects broadly consistent with above findings; exchange rate impact depends on proximity to policy rate cuts; APP announcements predominantly positive for EMBI with heterogeneity; event study and panel regression broadly confirm findings. ### Effectiveness of Asset Purchase Programs — Event study findings - Sample and methods: - Examined all publicly available APP announcements in European EM economies from March 2020 onwards. - Event window defined as three days around the announcement date. - Government bond maturities analyzed: 1, 3, 5, and 10 years. - Spillovers studied: exchange rate and equity price channels. - Findings reported as median response and interquartile range for 11 announcement events; initial announcement event per country highlighted separately. - Government bond market: - Prior to initial announcements: 10-year bond yields increased by 0.20 percentage points in the three days preceding APP announcements. - Following announcements: - Long-term yields fell for the median response by 0.3 ppts by t = 1 and remained below t = 0 levels in subsequent days. - Median decline in yields by maturity: 1-year maturity fell by 0.1 ppts; 10-year bond fell by 0.3 ppts. - Yield curve flattened after initial APP announcements. - Term spreads declined following initial announcements but the decline was short-lived and did not extend to subsequent announcements. - Liquidity (proxied by government bond bid-ask spreads) tightened prior to initial announcements and temporarily improved following announcements. - Broadening to all announcements: decline in yields and yield curve flattening persist but median response size shrinks and impact on term spreads disappears—consistent with a stronger role for initial announcements. - Exchange rate and equity spillovers: - APP announcements did not induce a significant exchange rate response; median exchange rate response stayed broadly stable around announcement dates. - Equity markets: - Equity price indexes showed a gradual increase, reaching 2 percent by the fourth day following the announcement. - For initial APP announcements, the median equity response was 4 percent by day four. - Corporate credit default swaps: no systematic impact detected in the (not reported) analysis. ### Local projections analysis - Purpose and controls: - Addressed potential confounders omitted in event study by controlling for domestic policy rate cuts and global factors: either VIX or the Federal Reserve’s APP announcement. - Key results: - Domestic government bond yields: - APP announcements associated with drops in bond yields after controlling for other factors. - Statistically significant impact on the 10-year government bond yield of around –35 basis points one to three days after the announcement. - Results similar when using the Federal Reserve APP announcement as the global control. - Term spreads: - No systematic impact on term spreads; local projection results consistent with event study (no systematic deviations in term spread following announcements). - Equity prices: - Estimated response magnitude: 1–3 percentage points over the three-day horizon, with the impact increasing in days two and three. - Statistical significance of equity response varies with the global control used; the response is not statistically significant when VIX is used as the global factor. ### Limits to APPs in emerging markets - General constraints and risks: - Fiscal dominance is a principal constraint: APPs could facilitate support for fiscal deficits that cannot be met by conventional means, potentially leading to pressures on central bank independence and macroeconomic instability. - Historical precedent: monetary financing of governments in some large EMs in the 1980s contributed to high inflation and prolonged macroeconomic instability. - Factors that can mitigate risks and expand scope for APPs: - Fiscal sustainability: large and persistent fiscal deficits and high public debt precede debt monetization; fiscal sustainability is key to avoiding fiscal dominance. - Domestic financial market development: deep, broad, well-functioning domestic bond markets provide a larger pool of securities for secondary-market APPs; a large presence of state-owned banks may undermine market-based APPs and raise coordination concerns. - Monetary policy frameworks and credibility: well-anchored inflation expectations and credible monetary policy regimes expand room for APPs without de-anchoring expectations or triggering capital outflows/currency depreciation; legal prohibitions on direct financing bolster credibility. - Growth prospects and external balance: strong growth prospects, absence of large current account deficits, and strong FX reserve coverage reduce vulnerability to capital flow volatility and increase market confidence that APPs will not exert excessive exchange rate pressures. - Cautionary note: - The apparent effectiveness of APPs in EMs during the pandemic—a relatively short period with exceptional global conditions—does not guarantee similar outcomes on an open-ended basis or under different conditions. ### Macroeconomic fundamentals in European EMs with APPs - Cross-country variation: - Underlying fiscal strength varies among European EMs that implemented APPs during the pandemic. - European EMs are benchmarked against a sample of 22 EMs worldwide and the euro area average where applicable. - Prior to the crisis, Romania and Turkey stood out with relatively weaker indicators (as reported in the chapter). *Source: Box 1. The Evolution of the Asset Purchase Programs of Hungary (continued), from the chapter “Asset Purchase Programs in European Emerging Markets.”* ### 1. General Government Fiscal Deficit, 2019 ### 1. General Government Fiscal Deficit, 2019 ### Fiscal position and debt - Large fiscal deficits in 2019 noted across the sample. - General government debt is relatively low in Turkey and Romania, compared to peers, partly attenuating concerns about fiscal sustainability. - General government debt is relatively high in Croatia and Hungary, but 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, bolsters fiscal sustainability. ### Asset Purchase Programs (APPs) and market constraints - The level of domestic government securities markets: - Sizable in Croatia, Hungary, and Poland relative to EM peers. - Relatively thin in Romania, Serbia, and Turkey (Figure 19), suggesting limited scope for APP expansion in the latter countries. - Implemented APP sizes reach only a fraction of outstanding domestic government securities; none of the countries currently faces a binding constraint from supply of domestic securities available to purchase. - APP size heterogeneity (alternative metric): ranges from 2.7 percent in Romania to 14.3 percent in Poland. - State ownership definition and implications: - State ownership defined as government owning more than 50 percent of equity in a bank. - 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. - Relatively heavy state ownership in Poland and Turkey could trigger unease among market participants should APPs expand significantly, particularly if involvement of state banks in APPs is perceived as less than transparent. - Note: In Hungary, Poland, and Romania, the state’s share of subscribed banking sector capital has increased since 2016. ### Inflation, expectations, and central bank safeguards - 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. - Institutional safeguards for EU members and Serbia include de jure central bank independence and prohibitions on direct financing of government entities. - Article 21 of the Statutes of the European System of Central Banks and the ECB prohibits direct central bank lending to the government. - The EU framework ensures EU central banks are autonomous, have adequate authority, and are accountable for transparently achieving their mandated objectives, tasks, and functions. - Good practices for APPs include: (1) purchases being done at the initiative of the central bank in order to achieve its mandate(s), (2) central bank responsibility for rate setting, (3) central bank control of its balance sheet, and (4) purchases done in the secondary market to reflect market prices. ### Medium-term growth, external positions, and reserves - Medium-term growth prospects are sound for most of the European EMs; lower projected growth in Hungary and Poland is correlated with their higher income levels (Figure 21). - External imbalances vary: - Romania and Serbia (and Turkey prior to 2019) exhibited relatively large prepandemic current account deficits. - 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 overall, though foreign reserves in Turkey fall short of the IMF’s reserve adequacy (ARA) metric. - EU member states may 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. ### Overall assessment of scope for APPs - The degree to which macroeconomic fundamentals likely constrain the scope for APPs varies within the region. - 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. - Several European EMs have more developed policy frameworks and institutions, bolstered by EU membership; inflation-targeting frameworks have a strong track record. - The risk of fiscal dominance is not completely eliminated and should be monitored. ### Complementary unconventional monetary tools and liquidity measures - Beyond APPs, liquidity constraints can be mitigated through lower reserve requirements, broadened eligible collateral, and adjusted lending operations. - European EMs that implemented APPs also took complementary measures to overcome liquidity stress. Example: Croatia lowered the reserve requirement ratio from 12 to 9 percent. *Source: IMF staff calculations and analysis in "Limits to Emerging Market Asset Purchase Programs" (appeemea - 1. General Government Fiscal Deficit, 2019).* ### 3.5 to 0.5 percent), or temporarily waived sanctions for noncompliance ### appeemea - 3.5 to 0.5 percent), or temporarily waived sanctions for noncompliance ### Asset Purchase Programs (APPs): design features and risks - Central banks broadened eligible collateral during the pandemic, e.g., the MNB 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. - APPs increase central bank exposure to: - maturity risk and interest rate risk due to longer maturities of purchased assets; - credit risk, particularly for mortgage bonds and corporate bonds. - Other long-term liquidity measures used by countries: - Croatia conducted a five-year structural repo transaction. - Poland introduced a credit rediscount facility. - Hungary conducted regular tenders of collateralized long-term loans with maturities of "3, 6, and 12 months, and three and five years" at the policy rate. ### Negative policy rates and real rates in European EMs - Hungary maintained a negative overnight deposit rate of "(–5 basis points)". - Croatia had the deposit rate at "zero". - Poland lowered the deposit rate to "0.1 percent" in late May 2020. - Most sampled countries have had rather low or even negative real interest rates (Figure 23: Real Interest Rates, 2015–21; percent, three-month moving average: one-year government bond yield deflated by CPI). ### Forward guidance and APPs as commitment devices - Forward guidance reduces uncertainty about future discount factors and expected cash flows but may be limited in EMs because central banks face volatile capital flows and exchange market pressures that can force policy reversals ("time-inconsistency" challenge). - APPs can be viewed as akin to forward guidance but more credible because the central bank "puts its money where its mouth is" and can realize losses if long-term rates rise (loss realization depends on how securities are booked). - Better communication enhances APP effectiveness by reducing risks and smoothing adjustment to a new equilibrium. ### Complementary policies: CFMs, macroprudential measures, and fiscal interactions - Capital flow management (CFM) and prudential measures can complement APPs to alleviate market dysfunction, but deployment limits exist: - Temporary, transparent, and nondiscriminatory CFMs may be justified to (1) contain capital inflows when macro options are limited, and (2) contain disruptive outflows in crises (IMF 2012, 2018). - EU members are prohibited from deploying capital controls except in case of a national emergency (EU Treaty Article 63; exemptions have been temporarily granted previously). - Macroprudential measures aim to ensure financial stability to mitigate capital flow cycles; temporary release of buffers or tightening foreign exchange exposure limits can be appropriate (IMF 2017). - Example: Hungary tightened foreign exchange funding limits of banks in March 2020, which were reversed in September 2020. - APPs and fiscal policy are interdependent: - APPs that stabilize markets and lower long-term yields reduce government funding costs and rollover risks, potentially permitting a larger short-term fiscal boost if the economy is below capacity. - If fiscal spending is perceived as efficient, financing a larger deficit may be easier; if perceived as inefficient, fiscal dominance risk increases and limits APP scope. - Structural and governance reforms affect both the need for and effectiveness of APPs; better governance is linked to lower net interest margins and potential fiscal savings (example finding: improving governance to top 10 percent could, on average, save about "0.3 percent of GDP per year"). ### Communication best practices - Transparent communication at APP introduction, during operation, and at exit reduces speculative panics, enhances trust, and facilitates normalization. - Variations in practice: - Croatia communicated that APPs would only be reactivated if similar stress reemerged. - Hungary clearly communicated gradual increases in purchases and motivations when envelopes were increased. - Risks from poor communication include perceptions of monetary financing, especially if sterilization strategies are unclear. - The IMF’s "Central Bank Transparency Code" (2020b) and calls for "exceptional communication" in exceptional times are referenced as guidance. ### Exit strategies and sequencing - Conceptual exit cycle: 1. tapering (slowing pace of purchases); 2. stopping purchases; 3. maintaining stock by rolling over purchased assets; 4. reducing stock passively or actively. - Japan and the Bank of England experiences suggest steps including halting extraordinary interventions, downsizing the balance sheet, selling assets if necessary, and raising short-term interest rates. - Bank of England example: did not intend to reduce stock of purchased assets until bank rate reached about "1.5 percent". - Decisions to stop or taper should consider: whether initial market dysfunction conditions have dissipated; whether stated objectives have been achieved; and whether excess limits are in danger of being reached. - Stopping/tapering requires careful management and clear understanding of financial and fiscal risks to avoid market turbulence; coordinated policy packages and clear communication may be necessary when stopping signals a major policy shift. - Unwinding principles: - Preserve central bank independence and limit excessive market fluctuations. - Precondition: thorough understanding of fiscal and financial risks. - Unwinding should be transparent and coordinated with the Treasury. - Passive exits (letting portfolios mature without rolling over) are often appropriate; repos or outright sales may be used to tighten if needed. - Central banks should be ready to "pause" unwinding if spillover effects from major central banks are abrupt, particularly for exchange rates. ### Exit options and country-specific considerations for European EMs - Many European EM APPs were deployed primarily to alleviate market dysfunction; financial market dysfunction largely abated after initial pandemic weeks. - Where APP goals were limited to market dysfunction, APPs have stopped and remaining holdings can be maintained until maturity or unwound gradually; monetary tightening may provide an opportunity to unwind holdings. - Where APPs had expanded objectives (e.g., supporting transmission to long-term rates), exit decisions should weigh those objectives and compare APP efficiency to alternative instruments. - Example: Hungary scaled back a long-term collateralized credit facility and partially substituted it with continued purchases of government securities to ensure more effective transmission to long-term rates. - Outright unwinding should be paced considering outstanding positions and their expected effects on policy objectives. ### Evidence on APP impacts in European EMs - APP announcements in March and April 2020 had immediate stabilizing impacts on government bond markets in countries that pursued APPs: - Reduced 10-year government bond yields. - Compressed bid-ask spreads for government bonds. - Halted increases in term spreads during initial announcements. - Facilitated smooth fiscal financing during a critical period. - Spillovers and other asset price impacts: - Little evidence of systematic currency depreciation following APP announcements. - Equity prices increased by around "2 percent" two to three days after APP announcements. - Caveats and limitations of the analysis: - Results are silent on longer-term impacts that require more elapsed time. - High-frequency analysis could not control for all domestic policy actions around APP events (e.g., foreign exchange intervention). - The analysis focused on direct impacts of APP announcements and did not construct a counterfactual of no APP announcements. *Italic: Source — Selected excerpts from the provided IMF content unit.* ### Conclusions and Policy Lessons ### Conclusions and Policy Lessons ### Role and Limits of APPs in European EMs - APPs have generally been consistent with primary monetary policy objectives with the risk of fiscal dominance well contained. - Macroeconomic fundamentals are generally strong across the group, though variation exists 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. - 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; breaching those limits may endanger both de jure and de facto central bank independence as well as statutory and real fiscal limits. ### Planning for Monetary Policy Normalization and APPs - Central banks need to plan for the role of APPs in eventual monetary policy tightening as national vaccination programs proceed and a durable recovery appears increasingly possible. - Recommendation: 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, it would be appropriate to complement short-end tightening with tapering or discontinuation of APPs to support transmission to the longer end of the curve. - When the government securities market has largely regenerated following a major shock and is no longer illiquid or segmented, continued central bank intervention could be perceived as monetarization of budget deficits and ultimately increase risk premiums. - 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. - Caveat: Raising interest rates does not necessarily preclude further exceptional asset purchases during circumstances of exogenous, significant, and temporary market disruptions. ### Communication and Credibility - Clear and transparent communication at each step of the process is critical for credibility, particularly when a large shock triggers a major policy shift. - Effective communication facilitates information dissemination, reduces asymmetric information, and ensures a faster and smoother adjustment to a new equilibrium. ### Conceptual Underpinnings: Why APPs Can Be Effective - APPs can stimulate aggregate demand and alleviate market dysfunction, complementing conventional policies when the standard transmission is impaired. - Factors that can weaken the standard transmission channel after an unexpected shock: - Heterogeneous financial assets (liquid central bank reserves versus less liquid longer bonds). - Diverse economic agents with specific time preferences or risk tolerances. - Financial frictions: intermediaries cannot or will not arbitrage the yield curve until consensus on a new equilibrium emerges. - APPs can: - Ease liquidity constraints overall (depending on sterilization) and along the yield curve. - Smooth price adjustments and reduce overshooting due to rigidities. - Help anchor expectations. - The shape of the yield curve reflects market expectations, arbitrage, and central bank policies; APPs can influence long rates mainly by informing intended future policies. - Yield curve theories and APP effects: - Expectation theory (Lutz (1940) building on Fisher (1930)): APPs will have no impact on the yield curve under a “no-arbitrage condition.” - Liquidity preference theory (Hicks 1939): public preference for shorter securities implies prompt market adjustment rendering interventions impotent. - Segmentation and preferred habitat theories (Culbertson (1957); Modigliani and Sutch (1966)): APPs can have at least temporary impact because buyers/sellers focus on specific maturities, limiting substitution. - Reality likely lies between these theories; APPs may have more lasting impact if they change the information set (including conveying propriety information) or reduce asymmetric information and uncertainty. ### Channels Through Which APPs Affect Economy and Yield Curve - Portfolio rebalancing channel: - APPs influence duration risk and relative prices, prompting institutional investors to rebalance toward other maturities and asset classes. - Lower treasury yields can cause capital/wealth gains and potentially boost willingness to accept more credit risk, though extremely risk-averse savers could increase precautionary savings and suppress aggregate demand. - Liquidity channel: - APPs can provide additional liquidity and ease liquidity constraints depending on degree of sterilization; even fully sterilized purchases can alleviate slope and bumpiness (for example, via “operation twist”). - Expectation—signaling/confidence channel: - APPs influence the information set and can reduce risks and uncertainties, complementing traditional policy rules and lowering long-term rates. - The better the authorities’ communication track record, the more sustainably APPs can reduce risks and uncertainties. - Intermediation channel: - Affects intermediaries’ ability to on-lend to liquidity-constrained entrepreneurs, households, and governments; effectiveness depends on sterilization and complementary lending facilities and collateral frameworks. - APPs can generate capital gains on trading books for long bonds valued mark-to-market, easing capital constraints. - Affects willingness to lend via firmer expectations; magnitude depends on existing cushions and intermediaries’ risk aversion. - New evidence suggests the slope of the yield curve may matter more for bank lending than previously thought. - Market microstructure effects: - APPs reduce securities available for trading (future turnover) but the presence of a committed large buyer improves price discovery and diminishes informational frictions, possibly narrowing bid-ask spreads. ### Fiscal Channel and Perceptions - The fiscal-channel importance hinges on (1) the purpose of the APP and (2) the perception of government efficiency. - If markets perceive APPs as mitigating financial instability and benefiting most participants and future growth, APPs can lower interest payments and rollover risks for the government. - If lacking trust in government efficiency and fairness, APPs could be perceived as monetization of unsustainable budget deficits and quickly result in higher risk premiums and inflation. ### Country-Specific Note (Croatia example) - Croatia: - Croatian Central Bank was the first in the region to introduce an APP during the pandemic. - The APP aimed to (1) maintain exchange rate stability, (2) increase domestic liquidity to ensure continued bank lending at low interest rates, and (3) support the stability of the government securities market. - The Croatian Central Bank conducted its first auction on March 13, 2020. - Initially, only banks could participate; procedures were quickly amended in the context of investment funds facing large outflows. *Source: Conclusions and Policy Lessons, Annex 1. Conceptual Underpinnings of Central Bank Asset Purchase Programs (appeemea - Conclusions and Policy Lessons).* ### Annex 2. Details of Country ### References ### References ### Scope and themes - Bibliographic list covering literature on unconventional monetary policy, quantitative easing, term structure of interest rates, central bank operations, spillovers to emerging markets, and policy responses during the COVID-19 pandemic. - Includes working papers, journal articles, central bank press releases, IMF policy papers and working papers, NBER working papers, and publications from national central banks and international institutions. ### Major topics and recurring focuses - Unconventional monetary policy and quantitative easing: conceptual foundations, transmission channels, market effects, duration and signaling, and exit strategies. - Term premia, term structure of interest rates, and yield-curve-based forecasting. - International spillovers of advanced-economy monetary policy to emerging markets and frontier economies. - Central bank communication, transparency, and operational instruments during crises, notably the COVID-19 pandemic. - Macroprudential policy relaxation and reserve adequacy in the face of volatile capital flows. - Empirical event studies assessing asset purchase announcements and market liquidity effects. ### Prominent authors, institutions, and publications (selected, with exact identifiers) - Bauer, Michael, and Thomas M. Mertens. 2018. “Economic Forecasts with the Yield Curve.” FRBSF Economic Letter 2018–07, Federal Reserve Bank of San Francisco, San Francisco, CA. - Bernanke, Ben S. 2014. “Central Banking after the Great Recession: Lessons Learned and Challenges Ahead, A Discussion with Federal Reserve Chairman Ben Bernanke on the FED’s 100th Anniversary.” Excerpt, The Brookings Institution, Washington, DC. - Bernanke, Ben S. 2020. “The New Tools of Monetary Policy.” American Economic Review 110 (4): 943–83. - Bhattarai, Saroj, and Christopher Neely. 2016. “A Survey of the Empirical Literature on U.S. Unconventional Monetary Policy.” Working Paper 2016–021A, Federal Reserve Bank of St. Louis, St. Louis, MO. - Brandão-Marques, Luís, Marco Casiraghi, Gaston Gelos, Güneş Kamber, and Roland Meeks. 2021. “Negative Interest Rates: Taking Stock of the Experience So Far.” IMF Departmental Paper 21/03, International Monetary Fund, Washington, DC. - Christensen, Jens H. E., and Glen D. Rudebusch. 2012. “The Response to Interest Rates of U.S. and U.K. Quantitative Easing.” The Economic Journal 122: F385–F412. - Eggertsson, Gauti B., and Michael Woodford. 2003. “The Zero Bound on Interest Rates on Optimal Monetary Policy.” Brookings Papers on Economic Activity 1: 139–211. - Fratto, Chiara, Brendan Harnoys Vannier, Borislava Mircheva, David de Padua, and Hélène Poirson. 2021. “Unconventional Monetary Policies in Emerging Markets and Frontier Countries. IMF Working Paper 21/14, International Monetary Fund, Washington, DC. - Gagnon, Joseph, Matthew Raskin, Julie Remache, and Brian Sack. 2011. “The Financial Market Effects of the Federal Reserve’s Large-Scale Asset Purchases.” International Journal of Central Banking 7 (1): 3–43. - Krishnamurthy, Arvind, and Annette Vissing-Jorgensen. 2011. “The Effects of Quantitative Easing on Interest Rates: Channels and Implications for Policy.” Brooking Papers on Economic Activity (2): 215–80. - Vayanos, Dimitri, and Jean-Luc Vila. 2009. “A Preferred-Habitat Model of the Term Structure of Interest Rates.” NBER Working Paper 15487, National Bureau of Economic Research, Cambridge, MA. - Vayanos, Dimitri, and Jean-Luc Vila. 2021. “A Preferred-Habitat Model of the Term Structure of Interest Rates.” Econometrica 89 (1): 87–112. - Numerous central bank sources and press releases, including Magyar Nemzeti Bank (MNB) press releases from 2020 and 2021, Central Bank of the Republic of Turkey 2020 press release, National Bank of Romania 2020 Board decisions, and Croatian National Bank 2020 Macroeconomic Developments and Outlook. ### Observations on methodological and empirical approaches represented - High representation of event-study analyses, working papers with policy implications, and empirical surveys. - The list includes both theoretical foundations (e.g., Modigliani and Sutch 1966; Hicks 1939; Fisher 1930) and contemporary empirical assessments of central bank tools during episodes of market stress. - Cross-references to IMF institutional views and policy papers on capital flows management, reserve adequacy, central bank transparency, and macroprudential policy. *Source: appeemea - References (PDF chapter/section) — https://www.imf.org/-/media/files/publications/dp/2021/english/appeemea.pdf* --- _Source: https://www.imf.org/-/media/files/publications/dp/2021/english/appeemea.pdf_