Tracking Economic and Financial Policies During COVID-19: An Announcement-Level Database
Author:
Mr. Divya Kirti
Search for other papers by Mr. Divya Kirti in
Current site
Google Scholar
PubMedClose
https://orcid.org/0000-0001-6636-6687
,
Yang Liu 0000000404811396 https://isni.org/isni/0000000404811396 International Monetary Fund

Search for other papers by Yang Liu in
Current site
Google Scholar
PubMedClose
,
Soledad Martinez Peria
Search for other papers by Soledad Martinez Peria in
Current site
Google Scholar
PubMedClose
,
Ms. Prachi Mishra
Search for other papers by Ms. Prachi Mishra in
Current site
Google Scholar
PubMedClose
, and
Jan Strasky
Search for other papers by Jan Strasky in
Current site
Google Scholar
PubMedClose
Publication Date:
03 Jun 2022
eISBN:
9798400213069
ISBN:
9798400213069
Language:
English
Keywords:
IMF Policy tracker; prudential measure; capital injection; IMF article IV report; central bank swap lines; COVID-19
Download PDF (805 KB)

We introduce a new comprehensive announcement-level database tracking the extraordinary fiscal, monetary, prudential, and other policies that countries adopted in response to Covid-19. The database provides detailed information, including sizes where available, for 28 granular policies adopted by 74 countries during 2020. About 5,500 policy measures were announced during this period. Importantly, the database is organized and presented in a format easy for researchers to use in empirical analyses. Announcements were highly correlated across the broad fiscal, monetary, and prudential categories and at more granular levels. Advanced economies (AEs) introduced larger fiscal measures than emerging and developing economies (EMDEs) and relied primarily on large unconventional monetary policies. Bank capital requirements were relaxed widely in both AEs and EMs, while relaxation of provisioning requirements was more common among EMs. Supervisory expectations and reporting requirements were widely relaxed.

Abstract

We introduce a new comprehensive announcement-level database tracking the extraordinary fiscal, monetary, prudential, and other policies that countries adopted in response to Covid-19. The database provides detailed information, including sizes where available, for 28 granular policies adopted by 74 countries during 2020. About 5,500 policy measures were announced during this period. Importantly, the database is organized and presented in a format easy for researchers to use in empirical analyses. Announcements were highly correlated across the broad fiscal, monetary, and prudential categories and at more granular levels. Advanced economies (AEs) introduced larger fiscal measures than emerging and developing economies (EMDEs) and relied primarily on large unconventional monetary policies. Bank capital requirements were relaxed widely in both AEs and EMs, while relaxation of provisioning requirements was more common among EMs. Supervisory expectations and reporting requirements were widely relaxed.

I. Introduction

The Covid-19 pandemic began as a health shock in early 2020, but rapidly triggered a deep economic crisis. In April 2020, the IMF revised its global GDP growth forecast down by an unprecedented 6.3 percentage points (IMF 2020). Governments across the world rapidly introduced fiscal, monetary, prudential, and other measures to address the economic consequences of the pandemic. Many advanced economies (AEs) and emerging and developing countries (EMDEs) soon announced large transfers to households, including the Emergency Family Income Program in Argentina, the JobKeeper Payment Scheme in Australia, the Emergency Aid Scheme in Brazil, and the Economic Impact Payments in the U.S. Central banks in AEs, notably the European Central Bank and the U.S. Federal Reserve Board announced large-scale asset purchase programs. In EMDEs, central banks relied more on lending programs, such as subsidized lending programs for small and micro enterprises in China and the Funding for Growth Scheme in Hungary. Finally, 25 AEs and 22 EMDEs in our dataset either released their macroprudential capital buffers or postponed planned buffer increases. While there have been several initiatives to collect information on economic and financial policies implemented in response to the Covid-19 pandemic around the world, existing sources vary in breadth, granularity, and frequency of coverage, and are often presented in formats that cannot be easily used by researchers.1

This paper presents a new comprehensive announcement-level panel dataset which tracks fiscal, monetary, prudential, and other policy responses to Covid-19 at a high frequency and a granular level. The database provides detailed information for 28 different policies adopted by 30 AEs and 44 EMs during 2020 (See Appendix Table A.1). Taking together, the 74 economies account for about 92 percent of world output and 81 percent of world population.

We begin our data collection effort with the IMF’s Policy Tracker, which draws on regular IMF staff surveillance activities and provides an account of the main policies that countries adopted in response to Covid-19. This is combined with information from a series of alternative sources including other existing trackers, government websites, news reports, and reports from government agencies or the private sector.2 Overall, bringing together information from and cross-checking across many sources helps us provide a more comprehensive and accurate description of policy announcements in response to Covid-19 for the 74 countries in our dataset. One limitation, however, is that the database covers the measures announced in 2020 exclusively.

We collect information on 28 policies, covering granular fiscal measures (grants, tax reliefs, tax deferrals, equity participations, loans, and guarantees), conventional (interest rate and reserve requirement changes as well as financing operations) and unconventional (asset purchases) monetary policies, capital and non-capital prudential regulations targeting the banking sector, and other policies such as moratoria, prudential policies affecting non-banks, and market-based measures.3 For all policies, we collect announcement dates, details of the measures, and sources. Wherever available we collected information on the size of announced policies, too. The database, therefore, also includes information on the size of the measures adopted for 13 policies. Importantly, the database is organized and presented in a format easy for researchers to use in empirical analyses.

The rest of the paper is organized as follows. Section II describes the construction of the database. Section III provides a general overview of the measures countries adopted, while sections IV, V, and VI provide details on the use of fiscal, monetary, and prudential measures, respectively. Section VII concludes.

II. Database Construction

A. Sources

The main source we use to construct our database is the IMF Policy Tracker, which summarizes the key fiscal, monetary, and prudential measures countries adopted in response to the crisis. The IMF Policy Tracker, however, does not systematically provide dates for all policy announcements or information on the size of the measures adopted. Hence, when it comes to fiscal policies, we complement the information in the IMF Policy Tracker with information from the Fiscal Monitor Database of Country Fiscal Measures in Response to the Covid-19 Pandemic produced by the IMF Fiscal Affairs Department. This database provides cumulative sizes of fiscal measures in broad categories.

We combine the information in the IMF databases, which is not exhaustive and often lacks announcement dates, with data from the European Systemic Risk Board (ESRB) Covid-19 Policy Measures database and Yale Program on Financial Stability (YPFS) Covid-19 Financial Response Tracker. We streamline the granular classification in the ESRB and combine it with specific information on non-European countries from the YPFS. In order to expand coverage and improve accuracy, we use a number of additional publicly available sources.

We collect information on 5,462 individual policies. For countries covered in the ESRB database, we add or substantially revise 562 policies in addition to the 2,254 measures we obtain from either the ESRB or YPFS database. For countries only covered in the YPFS tracker but not the ESRB database, we add or substantially revise 1,195 measures in addition to the 1,451 measures we obtain from the YPFS tracker.

For fiscal policies, first, we check government websites, budgetary amendments, news reports, and search engines extensively to look for policy announcements. We carefully address the problem of double counting by comparing announcements containing similar measures and pay special attention to large policy announcements to understand their exact components. Second, wherever possible, we refine our size measures using reports by finance ministries and government agencies. For EU member states, we also use the European Commission’s decisions on state aid schemes, which typically specify operational details and estimated sizes. Note that we use announced sizes and ex-ante estimates, not actual expenditure since this information is much harder to find across countries. Third, we incorporate information from the Bruegel tracker by Anderson et al. (2020), which provides comprehensive and granular size estimates for 12 advanced economies. Finally, we occasionally rely on sources from regional organizations, including (i) the Covid-19 Policy Database by the Asian Development Bank, which covers many emerging markets; (ii) an OECD (2021) report, which provides a timeline of fiscal policy responses in 60 jurisdictions; and (iii) the European Bank for Reconstruction and Development Transition Report 2020–21, which features a brief summary of policy responses to Covid-19 for 37 countries.

For monetary and prudential polices, we consult the following sources: (i) central bank announcements, annual reports, and balance sheet data, which typically document important policy interventions as well as announced or actual sizes; (ii) a timely book edited by English, Forbes and Ubide (2021), which includes detailed narratives of policy responses to Covid-19 by policymakers from 16 central banks; and (iii) a cross-country database of monetary responses by the Bank for International Settlements (Cantú et al. 2021), which collects announcements in 39 countries, though often lacks information on sizes. We also look at other cross-country sources, such as the database by Federico, Vegh and Vuletin (2014) updated in 2022, which records various reserve requirements in 65 countries up to Q1 2021, and the database by Feyen et al. (2020), which tracks financial sector policies across 154 jurisdictions.

Beyond fiscal, monetary, and prudential policies, under “other policies” our database tracks moratoria, prudential policies targeted to non-bank financial institutions (NBFI measures), and market-based measures. Table 1 lists all the sources we used to put together our tracker.

Table 1.

List of Main Sources

article image

B. Criteria For Recording Policy Measures

To determine which policy measures to include in our database and how to record them we follow the criteria below:

  • a) Relevance. We only record extraordinary policy actions announced during 2020 and directly or indirectly related to Covid-19.

  • b) Timing. As the tracker is at the announcement level, a measure must be announced to be recorded. When a measure has multiple related announcements, we use the earliest official announcement that offers details on the nature of the measure. For this reason, a tentative proposal with significant uncertainties is not recorded as an announcement. By contrast, if a measure is officially announced but only with general information on its nature and size, with operational details announced later in a subsequent announcement, we use the first announcement to determine timing and use the second one as supplementary information. For instance, if a measure was announced on March 20 but its size was announced one week later, we use the first date as the announcement date and impute its size using the size announced later.

    The criteria we apply to decide whether to record announcements has important implications for the treatment of pre-existing measures, automatic stabilizers, and unannounced measures. First, preexisting measures are not counted unless there are later announcements updating or expanding these measures in response to Covid-19. In the latter cases, we record the relevant changes rather than the pre-existing policies. For instance, if a government changes the terms of a pre-existing lending program without changing its total size, we record the change with a missing size since there is no actual size change announced. Second, automatic stabilizers, which generally come into effect without an announcement, are not counted unless the government makes a specific announcement. For instance, if the spending on unemployment benefits increases only because there are more applicants, we do not record the increase in spending as a grant. By contrast, if the government announces a discretionary action, such as an extraordinary increase in the level of unemployment benefits or coverage in response to Covid-19, then the action is recorded as a grant because it is associated with a clear announcement of a change in policy. Third, unannounced actions are not recorded in the tracker. For instance, foreign exchange interventions not connected with any announcement or news report are not included.

  • c) Implementing authority. We include measures implemented at the central government level. Therefore, measures implemented by local governments or foreign entities are generally not included. See Appendix A for additional details regarding how we map announcements by different authorities to policy types.

  • d) No actions, extensions, and rollbacks. An announcement of no action does not constitute a policy measure for our purposes (e.g., the central bank decides to keep interest rates unchanged). Extensions and discretionary rollbacks are coded as new measures, and we construct dummies in the dataset to flag them, as “+1” and “-1”, respectively. In coding sizes of policy extensions, to avoid double counting, we only incorporate incremental sizes. Automatic expirations are not counted because they are not discretionary measures.

C. Classification, Coding, and Size of Policy Measures

Our classification of policies consists of four broad categories, namely fiscal, monetary, prudential, and other measures. Table 2 lists all individual policy types under each category. See Appendix Table A.2 for exact definitions.

Table 2.

List of Policy Types

article image
Notes: Y (N) indicates that size information is (not) available for the corresponding policy. Y* denotes that both announced and actual sizes are recorded. For lending operations, only the credit facility subset has announced and actual sizes.

European-level measures require particular attention. We allocate European-level measures, either at the European Union (EU) level or the Euro Area (EA) level, to individual member states since our database is a country-level database. For instance, if the ECB announces a policy measure, we duplicate it in our database for all EA member states. If the measure has size information, we allocate the size to all member states involved. As a result, European-level measures are repeated for all 16 EA member states and 24 EU member states in our tracker. The Next Generation EU is recorded as a fiscal announcement, but without sizes as almost no information was available in 2020 about how the money would be spent. See Appendix B for additional details.

For each policy type, a dummy indicates whether a measure falls under this type and its direction. The dummy takes the value of +1, 0, or -1. “0” means the announcement does not correspond to the policy type. “+1” means the announcement loosens policy (as in the vast majority of announcements we track), and “-1” corresponds to a tightening measure. While it is generally straightforward to distinguish loosening and tightening measures, there are a few important exceptions.

  • a) We assign “+1” to the imposition of dividend restrictions. We consider those restrictions as loosening measures because they are intended to positively impact credit through forcing banks to preserve capital.

  • b) Foreign exchange interventions (FXIs) do not follow this sign rule. Instead, we assign “+1” to measures intended to strengthen or stabilize exchange rates, and “-1” to measures explicitly intended to weaken exchange rates.

  • c) Discretionary rollbacks of existing measures are given the opposite signs of the signs of the original measures. For example, a rollback of a loosening measure is considered a tightening measure.

  • d) For NBFI and market-based measures under the “Other” category, we do not distinguish loosening and tightening measures. Thus, they only take the value of +1 or 0.

When recording the size of policy measures where available, we pay special attention to the following issues:

  • a) Announced sizes. All sizes are announced sizes or early estimates. When announced sizes are not available, we use official or unofficial estimates published after the announcements, and we try to minimize reliance on information not available at the time of announcement.

  • b) Missing values. We record the size of a measure as missing either because no information is available, or because the size cannot be quantified. For instance, if a country delays a previously scheduled increase in the capital conservation buffer while the current buffer requirement remains unchanged, we set the size as missing because it lacks a clear measure.

  • c) Budgetary impact vs program sizes. For above-the-line fiscal measures (grants, tax reliefs, and tax deferrals), we record budgetary impact. For below-the-line (equity participations and public loans) and contingent (public guarantees) fiscal measures, we record program sizes, i.e., amounts of capital injections for equity measures and loan amounts for public loans and guarantees. In a few cases where a loan (guarantee) program also involves a capital injection, for instance, a government capitalizes a guarantee fund to launch a new Covid guarantee program, we keep track of both the loan (guarantee) program and the equity injection under their respective policy types so that we do not miss any policy. Nonetheless, this treatment also means that one should be careful when adding sizes across different policy types.

  • d) Actual sizes. When announced sizes are not available, we collect actual sizes for two policies: asset purchase programs and credit facilities (which is a subset of lending operations). The actual size of a measure is calculated as the maximum increase4 between its announced starting date and the earlier of (i) its next extension, or (ii) the end of Q1 2021. For example, if an asset purchase program was announced in March 2020, extended once in September 2020, and extended again in February 2021, we set the actual size of the March 2020 announcement as the maximum increase from March to September 2020, and the actual size of the September 2020 announcement as the maximum increase from September 2020 to February 2021. Conversely, we do not record the February 2021 announcement since it falls out of our sample period.

  • e) Horizon. We focus on the near-term size of a given measure. For fiscal measures, we define near term as expenditure in 2020 and 2021. Therefore, if a long-term policy is announced, we estimate the portion allocated to 2020 and 2021 as its size. For asset purchases and credit facilities, near term includes up to Q1 2021, which in practice only affects our calculation of actual sizes.

Beyond sizes, we record additional details for several policies, such as the durations of moratoria and selected terms of guarantees (see Appendix Tables A.2 and A.3).

Finally, in the database, each line generally corresponds to one policy. However, for some fiscal announcements which include multiple fiscal policies, we know the combined size of these policies but are unable to allocate the total size to each policy. In such cases, we record these policies in the same line and report the combined size, while the sizes of individual policies are recorded as missing.

III. Overview of the Tracker

Our database contains 5,462 policies announced by governments throughout 2020 (Figure 1). For both AEs and EMDEs, policies were highly concentrated in March and April 2020. Over this period, the median number of policies announced was 76 for AEs and 28 for EMDEs, while the median dropped to 29 and 17 from May onwards, respectively. March 12 (18) was the single day with the most fiscal, monetary, and prudential policies announced among AEs (EMDEs). On March 12, the ECB announced a comprehensive package including, among other elements, additional longer-term refinancing operations and asset purchases, as well as a separate regulatory package providing temporary capital and operational relief. On the same day, Bank of Canada expanded its bond buyback program5 and term repo operations, and in Australia, the government announced a $17.6 billion economic stimulus package. On March 18, EMDEs including Indonesia, Panama, Romania, and Turkey announced their first fiscal stimulus packages. Central banks in Brazil, Ghana, and Oman cut policy rates on the same day, and Colombia, Rwanda, and Ukraine launched sizable lending operations to stabilize long-term credit flows. Seven EMDEs relaxed prudential regulations.

Figure 1:
Figure 1:

Overview of Policies

Citation: IMF Working Papers 2022, 114; 10.5089/9798400213069.001.A001

A key stylized fact that emerges from our data is that countries tended to introduce different types of policies simultaneously, leading to positive and statistically significant correlations across policy types (Table 3). Standalone policy announcements are rarely observed in the data. Therefore, it would likely be challenging to isolate the impact of specific policies put in place to respond to Covid-19; and understanding the nature of countries’ policy responses would require coming to grips with how different types of policies were combined.

Table 3:

Correlations Between Broad Policy Categories

article image
Notes: Correlations are at the weekly level (* p<0.05). Rollbacks are not included. We sum up the number of announcements to get the count of each policy type in each period for each country. Then we calculate correlations of counts at the country-period level.

IV. The Use of Fiscal Policy During COVID-19

Figure 2 shows daily fiscal announcements by policy type for AEs and EMDEs, respectively. Among both AEs and EMDEs, grants were the most common fiscal policy measure adopted throughout 2020. Public loans and guarantees to support households and firms were the next most frequently used instruments by governments in AEs, while EMDEs seemed to rely mostly on tax reliefs after grants (Figure 2, Panel C). Unsurprisingly, the size of fiscal measures was larger among AEs relative to EMDEs (Figure 2, Panel D).6 At the country level, the median size of all fiscal measures combined in AEs was 15.2 percent of GDP, while for EMDEs it was only 4.1 percent. Nonetheless, variance is large within both groups. The standard deviation for AEs is 7.5 percentage points, and for EMs is 5.6 percentage points.

Figure 2:
Figure 2:
Figure 2:

Fiscal Policies During COVID-19

Citation: IMF Working Papers 2022, 114; 10.5089/9798400213069.001.A001

Notes: In panel D, each box ranges from the 25th percentile to the 75th percentile with the median in between. The whiskers outside of the box show the adjacent values, and the dots, if any, are potential outliers.

V. The Use of Monetary Policy during COVID-19

Figure 3 shows the use of different monetary measures by AEs and EMDEs, respectively. As expected, because in many advanced economies policy rates were already at or close to the effective lower bound, these economies relied more on asset purchases, while EMDEs introduced changes in interest rates and reserve requirements. Credit and liquidity injections were widely used by both groups of countries (Figure 3, Panel C).

Figure 3:
Figure 3:
Figure 3:

Monetary Policies During COVID-19

Citation: IMF Working Papers 2022, 114; 10.5089/9798400213069.001.A001

Notes: Reserve requirements in the figure include both local currency requirements and foreign currency requirements. FXI refers to foreign exchange interventions. In Panel E, the median size of reserve requirements in advanced economies is not calculated due to the small sample size. For asset purchases, we use actual amounts when announced amounts are not available. A positive number means a rate cut.

Wherever used, asset purchases were strikingly larger among AEs than in EMDEs (Figure 3, Panel D). The median size of cumulative asset purchases among AEs was 17.7 percent of GDP with a standard deviation of 6.5 percentage points, while that for EMDEs were 1.3 percent with a standard deviation of 2.7 percentage points, respectively. While policy rate cuts were undertaken more often by EMDEs than AEs, surprisingly the cumulative sizes of rate cuts were similar among AEs and EMDEs (Figure 3, Panel E). The median (mean) policy rate cut by AEs was 1.5 (1.1) percentage points in 2020,7 and that for EMDEs was 1.25 (1.9) percentage points. The positive correlation across monetary measures for both AEs and EMDEs indicates that these policies were applied in tandem during the pandemic (Table 5).

Table 5:

Correlations Between Monetary Policy Announcements

article image
Notes: Correlations are at the weekly level (* p<0.05). Rollbacks are not included.

VI. The Use of Prudential Policy during COVID-19

Figure 4 shows the use of capital and non-capital prudential measures in AEs and EMDEs, respectively. On average, AEs used prudential measures more widely than EMDEs. In 2020, AEs announced 400 capital-related prudential measures and 348 non-capital prudential measures, and EMDEs announced 181 and 202 such measures, respectively. Among capital-related prudential measures, although changes in buffers and dividend restrictions were introduced, the relaxation of other capital related measures was more prevalent, especially among AEs. Changes in provisioning requirements were frequently adopted among EMDEs. In terms of noncapital measures, countries, in particular AEs, frequently relaxed supervisory expectations and reporting requirements. As with monetary and fiscal policies, different types of prudential policies were introduced simultaneously, leading to positive and significant correlations across policies (Tables 6, 7, and 8).

Figure 4:
Figure 4:

Prudential Policies during Covid-19

Citation: IMF Working Papers 2022, 114; 10.5089/9798400213069.001.A001

Table 6:

Correlations Between Prudential Policy Announcements

article image
Notes: Correlations are at the weekly level (* p<0.05). Rollbacks are not included.
Table 7:

Correlations Between Capital-Related Policy Announcements

article image
Notes: Correlations are at the weekly level (* p<0.05). Rollbacks are not included.
Table 8:

Correlations Between Non-Capital Policy Announcements

article image
Notes: Correlations are at the weekly level (* p<0.05). Rollbacks are not included.

VII. Conclusions

We present a new announcement-level policy tracker for 74 countries. We believe this tracker is more comprehensive and accurate, and, hence, represents an improvement over existing trackers for the countries in our dataset during 2020. Our dataset shows that most policy announcements took place early on in the pandemic. Moreover, announcements were highly correlated across policy types, suggesting that isolating the impact of specific policies may be challenging. The size of fiscal measures was significantly larger in AEs than in EMDEs, while dispersion within AEs and EMDEs is equally large. In response to the pandemic, central banks in EMDEs primarily cut interest rates and reserve requirements, while those in AEs relied on asset purchase programs to provide monetary stimulus. Wherever used, asset purchases were strikingly larger among AEs than in EMDEs. Both AEs and EMDEs announced capital and non-capital related prudential measures. Among the capital measures, while buffers were relaxed in some AEs, the relaxation of other capital measures was more widely implemented among both AEs and EMDEs. The relaxation of provisioning requirements was more common among EMDEs. In terms of non-capital measures, both AEs and EMs relaxed supervisory expectations and reporting requirements.

References

  • Anderson, Julia, Enrico Bergamini, Sybrand Brekelmans, Aliénor Cameron, Zsolt Darvas, Marta Domínguez Jíménez, and Catarina Midões. 2020. “The Fiscal Response to the Economic Fallout from the Coronavirus”. https://www.bruegel.org/publications/datasets/covid-national-dataset/

    • Search Google Scholar
    • Export Citation

  • Asian Development Bank. 2020. ADB Covid-19 Policy Database. https://covid19policy.adb.org/s

  • Cantú, Carlos, Paolo Cavallino, Fiorella De Fiore, and James Yetman. 2021. “A Global Database on Central Banks’ Monetary Responses to Covid-19,” BIS Working Papers No 934.

    • Search Google Scholar
    • Export Citation

  • English, Bill, Kristin Forbes, and Ángel Ubide (Ed.). 2021. Monetary Policy and Central Banking in the Covid Era. Centre for Economic Policy Research.

    • Search Google Scholar
    • Export Citation

  • European Systemic Risk Board. 2020. Policy Measures in Response to the Covid-19 pandemic. https://www.esrb.europa.eu/home/search/coronavirus/html/index.en.html

    • Search Google Scholar
    • Export Citation

  • Federico, Pablo, Carlos A. Vegh, and Guillermo Vuletin. 2014. “Reserve Requirement Policy over the Business Cycle,” Working Paper 20612.

    • Search Google Scholar
    • Export Citation

  • Feyen, Erik, Tatiana Alonso Gispert, Tatsiana Kliatskova, and Davide S. Mare. 2020. “Taking Stock of the Financial Sector Policy Response to COVID-19 around the World,” Policy Research Working Paper No. 9497. World Bank, Washington, DC. http://hdl.handle.net/10986/34941

    • Search Google Scholar
    • Export Citation

  • International Monetary Fund. 2020. World Economic Outlook: The Great Lockdown. Washington, DC, April.

  • International Monetary Fund. (2020). Database of Country Fiscal Measures in Response to the COVID-19 Pandemic, https://www.imf.org/en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response-to-COVID-19

    • Search Google Scholar
    • Export Citation

  • International Monetary Fund. (2020) “Policy Response to COVID-19,” https://www.imf.org/en/Topics/imf-and-covid19/Policy-Responses-to-COVID-19

    • Search Google Scholar
    • Export Citation

  • Organisation for Economic Cooperation and Development. 2021. “An In-Depth Analysis of One Year of SME and Entrepreneurship Policy Responses to Covid-19: Lessons Learned for the Path to Recovery,” OECD SME and Entrepreneurship Papers, No. 25, OECD Publishing, Paris, https://doi.org/10.1787/6407deee-en.

    • Search Google Scholar
    • Export Citation

  • Yale Program on Financial Stability. 2020. Covid-19 Financial Response Tracker, https://som.yale.edu/centers/program-on-financial-stability/covid-19-crisis

    • Search Google Scholar
    • Export Citation

Appendix A: Implementing Authorities

For above-the-line measures (grants, tax reliefs, tax deferrals), we generally only consider announcements by central governments. While we do not directly cover local governments, we indirectly account for activities conducted by local governments because we include extraordinary transfers from central governments to local governments. In a few cases where regional governments play a substantial role—Belgium, Bosnia and Herzegovina, and Switzerland—we combine both central governments and regional governments.

Equity participations, public loans, and public guarantees can be provided either by governments directly or by state-owned entities such as public banks. By contrast, support from central banks or international organizations is not included. The only exception is the European Guarantee Fund managed by the European Investment Bank because the guarantee fund is an explicit response to Covid-19 directly financed by national governments.

Moratoria can be implemented either by the government or by the private sector.

Monetary measures should be implemented by central banks. For this reason, if the government purchases equity securities (or debt securities) on secondary markets, we classify the policy as an equity measure (or a public loan) rather than an asset purchase program (or a lending operation). Similarly, credit from central banks, either in the form of direct lending or indirect on-lending, is classified as a lending operation instead of a public loan.

Prudential measures are implemented by competent authorities. Note that in the case of Basel III-related measures, because the Basel Committee is not a regulator, we only consider announcements in which national authorities take actual actions.

Appendix B: Special Classification Issues

A. Separation of Similar Policies

First, within central bank lending operations, we further distinguish credit facilities and market liquidity measures. However, some market liquidity measures may be similar to asset purchases or credit facilities. For example, the central bank may purchase securities to maintain liquidity in secondary markets, which would be similar to asset purchases at first glance. As we emphasize in the definition in Appendix Table A.2, our classification draws on the intended effect, i.e., whether the stated aim of the policy mainly pertains to improving short-term market liquidity. In the case of lending operations, we also supplement our reading of policy intention with the duration of the measure. In practice, lending operations maturing in one month, such as 7-day repos, are generally classified as market liquidity measures.

Appendix Table A.1:

List of Countries

article image
Appendix Table A.2:

Policy Definitions

article image
article image
article image
Notes: Y (N) indicates that size information is (not) available for the corresponding policy. Y* denotes that both announced and actual sizes are recorded. For lending operations, only the credit facility subset has announced and actual sizes.

Second, central bank lending can be intertwined with public loans, especially in the case of on-lending through public banks, rediscounting of public loans, and government lending managed by the central bank. Here we differentiate between policies implemented by different authorities. We classify on-lending8 as a credit facility regardless of whether it is channeled through public banks. This is because ultimately the program is initiated and conducted by the central bank. If the central bank decides to rediscount loans granted under a public loan program, we record the rediscounting measure as a stand-alone credit facility in addition to the public loan measure because the two measures are implemented by different entities. Finally, if the government establishes a fund to grant public loans and the fund is housed in the central bank, these loans are classified as public loans instead of a credit facility.

Third, many credit-related measures also include equity injections from the government. These equity injections are recorded separately under the “equity” variable. E.g., if the government provides 10 million dollars to capitalize a guarantee fund and the fund can guarantee loans up to 100 million dollars, we record a guarantee of 100 million dollars and an equity measure of 10 million dollars. Similarly, if the central bank sets up a special purpose vehicle (SPV) to conduct asset purchases and the SPV is capitalized by the government, we record an asset purchase measure and an equity measure. This approach ensures that we include full details of these measures.

Finally, we treat the non-refundable portion of public loans as grants if applicable for most recipients. However, public guarantees with non-refundable clauses are not treated as grants because guarantees should be used only in the case of non-payment. While non-refundable clauses change the expected costs of guarantees, we do not track costs in our tracker due to the lack of information. For example, the Paycheck Protection Program in the U.S. is a guarantee program with non-refundable clauses, and we do not classify the non-refundable part as a separate grant.

B. European-Level Measures

We allocate European-level measures, either at the European Union level or the Euro Area level, to individual countries with special caution.

  • a) For the announced amount of asset purchases by the European Central Bank, we allocate the amount to each EA member state using the Eurosystem capital key as of February 1 2020.

  • b) As the ECB reports actual uses of lending operations by country, no additional adjustment is needed here.

  • c) Prudential measures announced by authorities at the EA level (or the EU level) are duplicated for all EA (or EU) member states.

  • d) The European Guarantee Fund is included in two steps. First, the equity contribution by each national government is recorded as an equity measure for each country. Second, if the guarantee fund announces a guarantee to entities in one country, we record a public guarantee for that country.

  • e) The Next Generation EU program is recorded as a fiscal announcement for each EU member state. However, virtually no information was available in 2020 about how the money would be spent. Given this ambiguity, we do not include any size information for this measure.

Appendix Table A.3:

List of Variables in the Database

article image
article image
article image
Notes: When converting currency units and calculating sizes as percent of GDP, we use nominal GDP and end-of-period exchange rates in 2019 from the October 2021 World Economic Outlook Database.
1

Existing cross-country sources include: the Covid-19 Policy Database by the Asian Development Bank, the monetary policy tracker by the Bank for International Settlements (Cantú et al. 2021), the Covid Policy Tracker by the IMF, the Country Policy Tracker by the OECD, the Financial Sector Policy Database by Feyen et al. (2020) at the World Bank, the Covid-19 Policy Measures Database by the European Systemic Risk Board, and the Covid-19 Financial Response Tracker by the Yale Program on Financial Stability.

2

See Section II and Table 1.

3

See Section II and Table 2.

4

For a given sequence of sizes {yt,y2, ...,yT}, the maximum increase from its initial value yt is defined as max{y1,y2,...,yT}-y1

5

The bond buyback program, first implemented in 1998, aims to maintain liquidity and smooth the issuance of treasury bills. The expansion on March 12 was intended to add market liquidity and support price discovery.

6

This same fact was mentioned in the October 2020 Fiscal Monitor.

7

These AEs are: Australia, Canada, Czech Republic, Hong Kong, Israel, South Korea, Norway, the United Kingdom, and the United States.

8

On-lending explicitly asks financial intermediaries to lend out a certain portion (typically 100 percent) of credit they receive from the central bank.

9

Data reflect information obtained from public sources and thus may not reflect all official stimulus measures taken by the authorities in response to COVID-19.

Same Series

Other IMF Content

Other Publishers

  • Collapse
  • Expand
Cover IMF Working Papers
Tracking Economic and Financial Policies During COVID-19: An Announcement-Level Database
Author:
Mr. Divya Kirti
,
Yang Liu
,
Soledad Martinez Peria
,
Ms. Prachi Mishra
, and
Jan Strasky
Volume/Issue:
Volume 2022: Issue 114
Publisher:
International Monetary Fund
ISBN:
9798400213069
ISSN:
1018-5941
Pages:
29
DOI:
https://doi.org/10.5089/9798400213069.001
Figure 1:

Overview of Policies
Figure 2:

Fiscal Policies During COVID-19
Figure 2:

Fiscal Policies During COVID-19
Figure 3:

Monetary Policies During COVID-19
Figure 3:

Monetary Policies During COVID-19
Figure 4:

Prudential Policies during Covid-19